1

Wealth gender differences the changing role of explanatory factors over time

Eva Sierminka1 Daniela Piazzalunga2 and Markus M Grabka3

1 Luxembourg Institute of Socio-Economic Research DIW Berlin and IZA

2 University of Turin and CHILD-Collegio Carlo Alberto 3 DIW Berlin and TU Berlin

8 June 2015

PRELIMINARY AND INCOMPLETE PLEASE DO NOT CIRCULATE WITHOUT PERMISSION

Abstract

In this paper we set out to investigate the explanatory factors of that have contributed to changing wealth levels covering the period of the Great Recession Given the large labor market changes which have taken place during this time in Germany we expect the process of accumulating wealth to have changed In particular we investigate the role of labor supply (which has largely increased for women in the years considered) the portfolio composition and changes in marital status We find the increased participation of women in the labor market and particularly their occupation structure had an increasing role in wealth accumulation We also find an important role of marital transitions particularly for those never married The changing role of explanatory factors has altered the wealth gap in Germany Consequently in the second part of the paper we extend the existing literature and investigate the changing gender wealth gap over time which has resulted from these changes both at the mean and across the wealth distribution We use micro data from the German Socio-Economic Panel (years 2002 2007 2012) which uniquely

provides wealth information at the individual level Preliminary results show that the mean wealth

levels have been decreasing since 2002 for both women and men However while the decrease was

similar for both men and women in the period 2002-2007 men lose more than women in 2007-2012

and the wealth gender gap shrank from 30000euro to 27000euro but this trend hides heterogeneities

along the wealth distribution

JEL D31 D13 Keywords Wealth differences Gender SOEP decomposition

2

1 Introduction Given the growing reliance of economic well-being on private assets including old age provision in the form of pensions and retirement income there is a growing interest in the study of private wealth However until recently the information about private wealth were scarce or not existent in many European countries With the Household Finance and Consumption Survey (HFCS) this gap is signicficantly shortened But still there arises the problem that private wealth is typically only surveyed at the household level by a reference person - as has been done in the HFCS or even also in the SCF - which does not allow for any intra-household analyses of private wealth Though past literature has stressed the importance of looking at intra-household inequalities (Deere and Doss 2006) and it has shown to lead to substantial differences in wealth levels among women and men (Sierminska et al 2010 Grabka et al 2013) Moreover Germany is also an interesting subject for wealth analyses as it was one of the OECD countries which had been hit hardest by the financial crisis ndash the so-called Great Recession in 20082009 GDP dropped by more than 5 in 2009 which was the strongest recession in Germany after World War II Other West European countries were not affected as much by the Great Recession1 It can be assumed that a financial crisis usually should have an impact on financial assets and net worth of private households However given that women are believed to be less risk prone (Cartwright 2011) thus reducing their expected return on a portfolio ndash at the same time higher risk aversion may shelter from unexpected asset fluctuations that had place during the Great Recession Thus we also intend to test whether this lower risk aversion allowed women to lose less (or maybe allowed to gain) during the Great Recession in Germany compared to men With three observations years of the German Socio-Economic Panel we are able to describe the gender specific wealth accumulation before the crisis (2002-2007) and during the crisis (2007-2012) It can be assumed that absolute wealth changes might be more prevalent for men given the higher wealth levels of men preceding the crisis their higher risk behavior in financial investments and income losses as the manufacturing and engineering industry in Germany ndash which by the majority employ men ndash shrank by almost 20 during the Great Recession in Germany and heavily made use of short-term compensation2 With unique micro-data about private wealth at hand we are able to analyze the gender difference in wealth levels and wealth accumulation in Germany given that the German Socio-Economic Panel (SOEP) is one of the very rare data-sources which collects wealth information at the individual level Besides this methodological precondition Germany also provides an interesting context for the analyses of the gender wealth gap womenrsquos attachment to the labor market has been significantly growing While the female labour force participation rate was only 53 in 1993 this share has been markedly increased by 10 percentage points to almost 63 in 2013 (Brenke 2015) and is now clearly above the Euopean level With entry into employment this allows women to earn their own money and to accumulate wealth Thus we like to answer the question how this changed labor market participation had any affects on the wealth distribution of women and men Given that women on average live a few years longer than man (Eurostat 2015) in combination with lower entitlements from public pension systems (the gender pension gap for mandatory pensions is

1 The Great Recession in Germany last for 12 months between Q2-2008 until Q1-2009 measured by quarter-on-quarter changes of seasonally adjusted real GDP For example in 2009 GDP in Spain contracted by only 38 in France by 32 and even in Portugal only by 29 (Eurostat 2014) 2 The basic idea of short-time compensation is that a firm with financial difficulties can apply for financial aid from the Federal Employment Agency to prevent the need for layoffs In return the firm has to reduce working hours and pay The replacement rate is 60 for single workers and 67 for workers with dependents

3

about 34 in OECD countries ndash OECD 2012) there is pressure for women to take care of their own private wealth This is reinforced by the growing number of single-headed female households (UNECE 2014 US Census Bureau 2014) and in particular elderly females which suffer more from old-age poverty than men (OECD 2012) In this paper we set out to investigate the explanatory factors of that have contributed to changing wealth levels covering the period of the Great Recession Given the large labor market changes which have taken place during this time in Germany we expect the process of accumulating wealth to have changed Women have increased their labor market participation while men have had it more difficult We compare the role of factors for women and men Consequently in the second part of the paper we extend the existing literature and discuss the changing gender wealth gap which has resulted from this changes both at the mean and across the wealth distribution The rest of the paper is organised as follows The following section outlines the literature and conceptual backgrounds Section 3 describes the methods Section 4 presents the data The descriptives and empirical results are in Section 5 and 6 respectively followed by a summary and conclusions in Section 7 2 Literature and Conceptual Background As commonly known wealth is accumulated according to the standard life-cycle model where the stock of assets in the current period is the outcome of past decisions regarding investment labor market outcomes savings and consumption As discussed in Sierminska et al (2010) differences in any of these factors will give rise to a different accumulation pattern and consequently a different portfolio structure Consequently any type of macro-economic or life-shock will have a differential impact on individual portfolios according to this structure Among possible causes that have been shown to affect wealth accumulation differently for women and men are women and men save differently women have a weaker attachment to the labor market and women and men invest differently with diverging levels of returns The literature indicates that women and men differ in the risk attitudes with women being less risk tolerant and more risk-averse (Cartwright 2011) which would lead to less risky portfolios and lower rates of return Addionally financial literacy (eg Lusardi and Mitchell 2008) influence investment decisions It is shown that men and women differ in their financial knowledge which also leads to more conservative investments done by women Apart from having differential returns a more risk-loving individual having invested in risky assets will be more exposed to fluctuations in the (stock) market Similarly a person whose majority of assets are invested in real estate property will be very susceptible to changing house prices One of the the more important factors that has been shown in Sierminska et al (2010) to explain male-female differences in wealth accumulation is labor market differences It is not only the lower labor market participation rate of women the lower working hours as women commonly work part-time compared to men with the standard pattern of a continuous full-time employment but even also the still existing gender pay gap which impairs the wealth accumulation for women (Warren et al 2001) Even when holding savings rates constant women are thus expected to accumulate lower levels of wealth (eg Blau and Kahn 1997 2000)

4

Other factors which may affect wealth accumulation and in particular its differences by gender are for example inheritance and bestowals if women are men have differential access to these One factor important for this paper would also be occupations Women and men ten to cluster in different occupation that have different perspectives in advancement as well as different exposures to labor market flucturations and thus could be differentially exposed to the consequences of the Great Recession During the last financial crisis in Germany the labor market has been differentially affected across occupations Over the last two decades the labor market in Germany has observed some major changes Not only women have substanitally increased their participation rates but numerous labor market reforms have been implemented (Hartz reforms) (Dustmann et al 2014 Brenke 2015) which as we will show could have had an impact on the wealth accumulation of women and men Based on the overview of the literature several hypothesis come to mind that could be tested within our framework Hypothesis 1 A more risk-loving individual (men) having invested in risky assets will be more exposed to fluctuation in the stock market Similarly a person whose majority of assets are invested in real estate property will be very susceptible to changing house prices (men) and this will result in the changing wealth levels Hypothesis 2 Womenrsquos increased attachment to the labor market will improve their position in the wealth distribution Hypothesis 3 Higher prevelance of part-time work will have a negative (decreasing ) effect on wealth Hypothesis 4 The changing impact of labor market attachment will have a stronger impact on womenrsquos wealth Hypothesis 5 Labor market factors are less important for older individuals We will consequently address these in our results section 3 Methods As discussed in the introduction we will first focus on what were the causes of the changes in wealth over this period and then we will investigate the changing wealth gap

31 Change in wealth

To analyse the changes of wealth over time we starting by estimating a wealth equation by OLS considering the change in variables instead of the level ones The equation is

(1)

Where is the change in the level of (inverse hyperbolic real) net wealth ie

is a vector of control variables observed in the second period (ie in 2007 for the comparison

2002-2007 and in 2012 for the comparison 2007-2012) migratory background age age squared living in West or East Germany number of children below age 5 in the household number of

5

marriages length of current marriage number of months spent in fulltime and in parttime work in the previous 5 years long term unemployement in the previous 5 years (inverse hyperbolic real) permanent income household inheritances bequestgifts lottery winning and their value over the previous 5 years household savings in the previous 5 years and share of financial assets

is a vector of control variables observed in the first period (ie in 2002 for the comparison 2002-

2007 and in 2007 for the comparison 2007-2012) level of education (controlling also if shehe was still in education) occupational status (inverse hyperbolic real) permanent income tangible assets own property debts and position in the wealth distribution

is a vector of changes in control variables between and change in marital status change

in own property assets change in consumer credits is a random error normally distributed

are the parameters to be estimated by OLS

Equation 1 is estimated separately for men and women and for the period 2002-2007 2007-2012 In addition we also estimate the same equation for married individuals In that case we also control

for the permanent income of the spouse in and for the bargaining power in

32 Gender wealth gap in different periods

The second contribution of our paper is to examine the evolution of the gender wealth gap over time particularly before and after the crisis In the first instance we only consider those individuals that were married in period 13 We focus on married people to see clearly the changing role of explanatory factors for a more homogenous group which nevertheless comprises about 50 of the population and because we have the unique opportunity of doing so with our individual level data In this case we first apply the Oaxaca-Blinder decomposition (Blinder 1973 Oaxaca 1973) at the mean and then Firpo Fortin Lemieux decomposition (Firpo et al 2009) for the whole distribution The Oaxaca-Blinder decomposition relies on the estimation of the following equation for men and women similar to equation 1 but where the dependent variables is the (inverse hyperbolic real) net worth and not its change over time

(2)

The OB decomposition is the standard one (3)

Where =male =female The first component captures the effect which can be attributed to differences in characteristics while the second one captures the effect of differences in returns

33 Detailed decomposition of the gender wealth gaps

3 Throuout the paper with ldquoperiod 1rdquo we indicate 2002 in the comparison 2002-2007 and 2007 in the

comparison 2007-2012

6

In the final section of the paper we focus in performing a detailed decomposition of the gender wealth gaps over time This technique introduced in Firpo et al (2009) allows to decompose differences between two distributions of a variable It allows to have the individual contribution of each explanatory variable considered in the analysis via the characteristics and returns components In our case it will allow us to identify the explanatory factors and how the differences in their distribution and returns change due to the financial crisis and the changing labor market circumstances (characteristics and returns) in order to understand what is contributing to the gender wealth gap in Germany

The Firpo Fortin Lemiuex method is regression based which allows to apply it in a simiar way as the OB method The technique relies on the estimation of a regression where the dependent variable is replaced by a recentered influence function (RIF) transformation and so any distributional statics can be decomposed based on the regression results

In our case we will focus on the differences in quantiles

)()(

^^__^____

QF

QM

M

QM

FM

Q XXX (4)

Where Q is the difference in quantile (or other statistic) τ of the wealth distribution

FM

XX____

are the average observed characteristics for men (M) and women (F) FM

Q

^

are the

coefficients obtained from the regression of the RIF variables of quantile Qτ on the set of explanatory variables for women (F) and men (M) The first terms in equation (4) captures the effect on the differential between the distributions caused by differences in characteristics (explained component) The second term corresponds to the effect of the coefficient in which the contribution of each individual explanatory factor can be distinguished 4 Data We use the German Socio-Economic Panel (G-SOEP) a representative longitudinal survey on individuals in private household The survey started in 1984 in West Germany and extended to states in East Germany before the reunification in 1990 Every year about 15000 household are interviewd (25000 people) In 2002 there was a boost of higher-income people to better capture the upper margin of the income and wealth distribution There were other refreshment samples in 2006 2009 and 2011 while we couldnrsquot include the refreshment of 2012 since they were not asked information on wealth Although the SOEP has a special high-income sample there is still the problem of an undercoverage of the very rich The person with the highest net worth in the sample only holds almost 63 mio euro in 2002 Thus all multi-millionaires above that threshold and even also billionaires are missing although they had a significant impact of the wealth distribution (see Westermeier and Grabka 2015) Every year information on the sociodemographic variables are collected as well as information on education labour market and employment earnings and income household composition health and satisfaction In addition there are topic modules which are replicated about every 5 years and inquiries on specific topic We use mostly data from 2002 2007 and 2012 which contain information on individual wealth The initial sample has more than 23000 observations in 2002 almost 21000 in 2007 and slightly more than 18000 in 2012

41 Sample

7

Along the paper we use two different samples the first sample is composed by the three pooled cross-sections (2002 2007 and 2012) Using cross-sectional weights this sample is representative of the German population at every year We consider the pooled cross-sectional sample in presenting the mean and median wealth and for the descriptive The second sample is twofold one part is composed by people present in the survey in 2002 and in 2007 (Panel sample 2002-2007) the second part is composed by people present in the survey in 2007 and 2012 (Panel sample 2007-2012) These panel sample have the advantage to allow us to follow people over 5 years and track their changes however we lose some observations because of attrition In addition the refreshments which happened during the 5 years of observation which have the objective to cover demographic changes in the underlying population canrsquot be used either since those individuals were not present in our starting point Hence the main disadvantage from the panel sample comes from the attrition and from the fact that we have two samples for 2007 one comparable with 2002 and the other one comparable with 2012 We do not consider only people who are included in the survey from 2002 to 2012 because we would lose to many observations For the panel sample we use panel weights The maximum number of observations is shown in the Appendix table B1

42 Outcome variable

G-SOEP contains information on individual wealth in particular the following assets are details own property (and relative share and relative debt) other real estate (and relative share and relative debt) financial assets (and relative share) business assets tangible assets building loan and private insurance and consumer credits The main dependent variable is net total wealth in 2010 real price In G-SOEP the missing values for wealth are corrected for applying a multiple imputation technique (see Frick et al 2010) for the moment we are only using implicate A but we will later include also the other implicates We apply a 01 top coding and the inverse hyperbolic sine transformation useful to mitigate the effect of the outliners and to deal with skewed variables (with respect to the log transformation the inverse hyperbolic transformation is possible with negative and 0 value) In the first part of the analysis our dependent variables is the difference in net wealth between 2 periods while in the second part is the level in wealth and the gap between men and women The control variables are detailed in Appendix A 5 Descriptives We find that during the period of analysis the real mean wealth levels have been decreasing since 2002 for both women and men (Table 1) The wealth gender gap in 2002 hovered around 30 000 euros at the mean (for a discussion of the sources of the wealth gap in Germany in 2002 see Sierminska et al 2010) it was similar in 2007 and declined to about 27 000 in 2012 The decline in wealth 2002-2007 was similar for women and men (about 6000) and in the subsequent period 2007-2012 it was almost double for men as it was for women (6500 versus 3500) The effect at the median is even more striking A reduction in wealth for men in the pre-crisis period of about 4000 with only a decline of about 400 for women In the subsequent period an increase of 2000 for men and only 500 for women At the median the wealth gender gap increases in 2012 suggesting that there have been different changes for men and women along the wealth distribution Hence the effect of the Great Recession had a differential effect on men compared to women in terms of their wealth level We explore the sources of this differential change

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

2

1 Introduction Given the growing reliance of economic well-being on private assets including old age provision in the form of pensions and retirement income there is a growing interest in the study of private wealth However until recently the information about private wealth were scarce or not existent in many European countries With the Household Finance and Consumption Survey (HFCS) this gap is signicficantly shortened But still there arises the problem that private wealth is typically only surveyed at the household level by a reference person - as has been done in the HFCS or even also in the SCF - which does not allow for any intra-household analyses of private wealth Though past literature has stressed the importance of looking at intra-household inequalities (Deere and Doss 2006) and it has shown to lead to substantial differences in wealth levels among women and men (Sierminska et al 2010 Grabka et al 2013) Moreover Germany is also an interesting subject for wealth analyses as it was one of the OECD countries which had been hit hardest by the financial crisis ndash the so-called Great Recession in 20082009 GDP dropped by more than 5 in 2009 which was the strongest recession in Germany after World War II Other West European countries were not affected as much by the Great Recession1 It can be assumed that a financial crisis usually should have an impact on financial assets and net worth of private households However given that women are believed to be less risk prone (Cartwright 2011) thus reducing their expected return on a portfolio ndash at the same time higher risk aversion may shelter from unexpected asset fluctuations that had place during the Great Recession Thus we also intend to test whether this lower risk aversion allowed women to lose less (or maybe allowed to gain) during the Great Recession in Germany compared to men With three observations years of the German Socio-Economic Panel we are able to describe the gender specific wealth accumulation before the crisis (2002-2007) and during the crisis (2007-2012) It can be assumed that absolute wealth changes might be more prevalent for men given the higher wealth levels of men preceding the crisis their higher risk behavior in financial investments and income losses as the manufacturing and engineering industry in Germany ndash which by the majority employ men ndash shrank by almost 20 during the Great Recession in Germany and heavily made use of short-term compensation2 With unique micro-data about private wealth at hand we are able to analyze the gender difference in wealth levels and wealth accumulation in Germany given that the German Socio-Economic Panel (SOEP) is one of the very rare data-sources which collects wealth information at the individual level Besides this methodological precondition Germany also provides an interesting context for the analyses of the gender wealth gap womenrsquos attachment to the labor market has been significantly growing While the female labour force participation rate was only 53 in 1993 this share has been markedly increased by 10 percentage points to almost 63 in 2013 (Brenke 2015) and is now clearly above the Euopean level With entry into employment this allows women to earn their own money and to accumulate wealth Thus we like to answer the question how this changed labor market participation had any affects on the wealth distribution of women and men Given that women on average live a few years longer than man (Eurostat 2015) in combination with lower entitlements from public pension systems (the gender pension gap for mandatory pensions is

1 The Great Recession in Germany last for 12 months between Q2-2008 until Q1-2009 measured by quarter-on-quarter changes of seasonally adjusted real GDP For example in 2009 GDP in Spain contracted by only 38 in France by 32 and even in Portugal only by 29 (Eurostat 2014) 2 The basic idea of short-time compensation is that a firm with financial difficulties can apply for financial aid from the Federal Employment Agency to prevent the need for layoffs In return the firm has to reduce working hours and pay The replacement rate is 60 for single workers and 67 for workers with dependents

3

about 34 in OECD countries ndash OECD 2012) there is pressure for women to take care of their own private wealth This is reinforced by the growing number of single-headed female households (UNECE 2014 US Census Bureau 2014) and in particular elderly females which suffer more from old-age poverty than men (OECD 2012) In this paper we set out to investigate the explanatory factors of that have contributed to changing wealth levels covering the period of the Great Recession Given the large labor market changes which have taken place during this time in Germany we expect the process of accumulating wealth to have changed Women have increased their labor market participation while men have had it more difficult We compare the role of factors for women and men Consequently in the second part of the paper we extend the existing literature and discuss the changing gender wealth gap which has resulted from this changes both at the mean and across the wealth distribution The rest of the paper is organised as follows The following section outlines the literature and conceptual backgrounds Section 3 describes the methods Section 4 presents the data The descriptives and empirical results are in Section 5 and 6 respectively followed by a summary and conclusions in Section 7 2 Literature and Conceptual Background As commonly known wealth is accumulated according to the standard life-cycle model where the stock of assets in the current period is the outcome of past decisions regarding investment labor market outcomes savings and consumption As discussed in Sierminska et al (2010) differences in any of these factors will give rise to a different accumulation pattern and consequently a different portfolio structure Consequently any type of macro-economic or life-shock will have a differential impact on individual portfolios according to this structure Among possible causes that have been shown to affect wealth accumulation differently for women and men are women and men save differently women have a weaker attachment to the labor market and women and men invest differently with diverging levels of returns The literature indicates that women and men differ in the risk attitudes with women being less risk tolerant and more risk-averse (Cartwright 2011) which would lead to less risky portfolios and lower rates of return Addionally financial literacy (eg Lusardi and Mitchell 2008) influence investment decisions It is shown that men and women differ in their financial knowledge which also leads to more conservative investments done by women Apart from having differential returns a more risk-loving individual having invested in risky assets will be more exposed to fluctuations in the (stock) market Similarly a person whose majority of assets are invested in real estate property will be very susceptible to changing house prices One of the the more important factors that has been shown in Sierminska et al (2010) to explain male-female differences in wealth accumulation is labor market differences It is not only the lower labor market participation rate of women the lower working hours as women commonly work part-time compared to men with the standard pattern of a continuous full-time employment but even also the still existing gender pay gap which impairs the wealth accumulation for women (Warren et al 2001) Even when holding savings rates constant women are thus expected to accumulate lower levels of wealth (eg Blau and Kahn 1997 2000)

4

Other factors which may affect wealth accumulation and in particular its differences by gender are for example inheritance and bestowals if women are men have differential access to these One factor important for this paper would also be occupations Women and men ten to cluster in different occupation that have different perspectives in advancement as well as different exposures to labor market flucturations and thus could be differentially exposed to the consequences of the Great Recession During the last financial crisis in Germany the labor market has been differentially affected across occupations Over the last two decades the labor market in Germany has observed some major changes Not only women have substanitally increased their participation rates but numerous labor market reforms have been implemented (Hartz reforms) (Dustmann et al 2014 Brenke 2015) which as we will show could have had an impact on the wealth accumulation of women and men Based on the overview of the literature several hypothesis come to mind that could be tested within our framework Hypothesis 1 A more risk-loving individual (men) having invested in risky assets will be more exposed to fluctuation in the stock market Similarly a person whose majority of assets are invested in real estate property will be very susceptible to changing house prices (men) and this will result in the changing wealth levels Hypothesis 2 Womenrsquos increased attachment to the labor market will improve their position in the wealth distribution Hypothesis 3 Higher prevelance of part-time work will have a negative (decreasing ) effect on wealth Hypothesis 4 The changing impact of labor market attachment will have a stronger impact on womenrsquos wealth Hypothesis 5 Labor market factors are less important for older individuals We will consequently address these in our results section 3 Methods As discussed in the introduction we will first focus on what were the causes of the changes in wealth over this period and then we will investigate the changing wealth gap

31 Change in wealth

To analyse the changes of wealth over time we starting by estimating a wealth equation by OLS considering the change in variables instead of the level ones The equation is

(1)

Where is the change in the level of (inverse hyperbolic real) net wealth ie

is a vector of control variables observed in the second period (ie in 2007 for the comparison

2002-2007 and in 2012 for the comparison 2007-2012) migratory background age age squared living in West or East Germany number of children below age 5 in the household number of

5

marriages length of current marriage number of months spent in fulltime and in parttime work in the previous 5 years long term unemployement in the previous 5 years (inverse hyperbolic real) permanent income household inheritances bequestgifts lottery winning and their value over the previous 5 years household savings in the previous 5 years and share of financial assets

is a vector of control variables observed in the first period (ie in 2002 for the comparison 2002-

2007 and in 2007 for the comparison 2007-2012) level of education (controlling also if shehe was still in education) occupational status (inverse hyperbolic real) permanent income tangible assets own property debts and position in the wealth distribution

is a vector of changes in control variables between and change in marital status change

in own property assets change in consumer credits is a random error normally distributed

are the parameters to be estimated by OLS

Equation 1 is estimated separately for men and women and for the period 2002-2007 2007-2012 In addition we also estimate the same equation for married individuals In that case we also control

for the permanent income of the spouse in and for the bargaining power in

32 Gender wealth gap in different periods

The second contribution of our paper is to examine the evolution of the gender wealth gap over time particularly before and after the crisis In the first instance we only consider those individuals that were married in period 13 We focus on married people to see clearly the changing role of explanatory factors for a more homogenous group which nevertheless comprises about 50 of the population and because we have the unique opportunity of doing so with our individual level data In this case we first apply the Oaxaca-Blinder decomposition (Blinder 1973 Oaxaca 1973) at the mean and then Firpo Fortin Lemieux decomposition (Firpo et al 2009) for the whole distribution The Oaxaca-Blinder decomposition relies on the estimation of the following equation for men and women similar to equation 1 but where the dependent variables is the (inverse hyperbolic real) net worth and not its change over time

(2)

The OB decomposition is the standard one (3)

Where =male =female The first component captures the effect which can be attributed to differences in characteristics while the second one captures the effect of differences in returns

33 Detailed decomposition of the gender wealth gaps

3 Throuout the paper with ldquoperiod 1rdquo we indicate 2002 in the comparison 2002-2007 and 2007 in the

comparison 2007-2012

6

In the final section of the paper we focus in performing a detailed decomposition of the gender wealth gaps over time This technique introduced in Firpo et al (2009) allows to decompose differences between two distributions of a variable It allows to have the individual contribution of each explanatory variable considered in the analysis via the characteristics and returns components In our case it will allow us to identify the explanatory factors and how the differences in their distribution and returns change due to the financial crisis and the changing labor market circumstances (characteristics and returns) in order to understand what is contributing to the gender wealth gap in Germany

The Firpo Fortin Lemiuex method is regression based which allows to apply it in a simiar way as the OB method The technique relies on the estimation of a regression where the dependent variable is replaced by a recentered influence function (RIF) transformation and so any distributional statics can be decomposed based on the regression results

In our case we will focus on the differences in quantiles

)()(

^^__^____

QF

QM

M

QM

FM

Q XXX (4)

Where Q is the difference in quantile (or other statistic) τ of the wealth distribution

FM

XX____

are the average observed characteristics for men (M) and women (F) FM

Q

^

are the

coefficients obtained from the regression of the RIF variables of quantile Qτ on the set of explanatory variables for women (F) and men (M) The first terms in equation (4) captures the effect on the differential between the distributions caused by differences in characteristics (explained component) The second term corresponds to the effect of the coefficient in which the contribution of each individual explanatory factor can be distinguished 4 Data We use the German Socio-Economic Panel (G-SOEP) a representative longitudinal survey on individuals in private household The survey started in 1984 in West Germany and extended to states in East Germany before the reunification in 1990 Every year about 15000 household are interviewd (25000 people) In 2002 there was a boost of higher-income people to better capture the upper margin of the income and wealth distribution There were other refreshment samples in 2006 2009 and 2011 while we couldnrsquot include the refreshment of 2012 since they were not asked information on wealth Although the SOEP has a special high-income sample there is still the problem of an undercoverage of the very rich The person with the highest net worth in the sample only holds almost 63 mio euro in 2002 Thus all multi-millionaires above that threshold and even also billionaires are missing although they had a significant impact of the wealth distribution (see Westermeier and Grabka 2015) Every year information on the sociodemographic variables are collected as well as information on education labour market and employment earnings and income household composition health and satisfaction In addition there are topic modules which are replicated about every 5 years and inquiries on specific topic We use mostly data from 2002 2007 and 2012 which contain information on individual wealth The initial sample has more than 23000 observations in 2002 almost 21000 in 2007 and slightly more than 18000 in 2012

41 Sample

7

Along the paper we use two different samples the first sample is composed by the three pooled cross-sections (2002 2007 and 2012) Using cross-sectional weights this sample is representative of the German population at every year We consider the pooled cross-sectional sample in presenting the mean and median wealth and for the descriptive The second sample is twofold one part is composed by people present in the survey in 2002 and in 2007 (Panel sample 2002-2007) the second part is composed by people present in the survey in 2007 and 2012 (Panel sample 2007-2012) These panel sample have the advantage to allow us to follow people over 5 years and track their changes however we lose some observations because of attrition In addition the refreshments which happened during the 5 years of observation which have the objective to cover demographic changes in the underlying population canrsquot be used either since those individuals were not present in our starting point Hence the main disadvantage from the panel sample comes from the attrition and from the fact that we have two samples for 2007 one comparable with 2002 and the other one comparable with 2012 We do not consider only people who are included in the survey from 2002 to 2012 because we would lose to many observations For the panel sample we use panel weights The maximum number of observations is shown in the Appendix table B1

42 Outcome variable

G-SOEP contains information on individual wealth in particular the following assets are details own property (and relative share and relative debt) other real estate (and relative share and relative debt) financial assets (and relative share) business assets tangible assets building loan and private insurance and consumer credits The main dependent variable is net total wealth in 2010 real price In G-SOEP the missing values for wealth are corrected for applying a multiple imputation technique (see Frick et al 2010) for the moment we are only using implicate A but we will later include also the other implicates We apply a 01 top coding and the inverse hyperbolic sine transformation useful to mitigate the effect of the outliners and to deal with skewed variables (with respect to the log transformation the inverse hyperbolic transformation is possible with negative and 0 value) In the first part of the analysis our dependent variables is the difference in net wealth between 2 periods while in the second part is the level in wealth and the gap between men and women The control variables are detailed in Appendix A 5 Descriptives We find that during the period of analysis the real mean wealth levels have been decreasing since 2002 for both women and men (Table 1) The wealth gender gap in 2002 hovered around 30 000 euros at the mean (for a discussion of the sources of the wealth gap in Germany in 2002 see Sierminska et al 2010) it was similar in 2007 and declined to about 27 000 in 2012 The decline in wealth 2002-2007 was similar for women and men (about 6000) and in the subsequent period 2007-2012 it was almost double for men as it was for women (6500 versus 3500) The effect at the median is even more striking A reduction in wealth for men in the pre-crisis period of about 4000 with only a decline of about 400 for women In the subsequent period an increase of 2000 for men and only 500 for women At the median the wealth gender gap increases in 2012 suggesting that there have been different changes for men and women along the wealth distribution Hence the effect of the Great Recession had a differential effect on men compared to women in terms of their wealth level We explore the sources of this differential change

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

3

about 34 in OECD countries ndash OECD 2012) there is pressure for women to take care of their own private wealth This is reinforced by the growing number of single-headed female households (UNECE 2014 US Census Bureau 2014) and in particular elderly females which suffer more from old-age poverty than men (OECD 2012) In this paper we set out to investigate the explanatory factors of that have contributed to changing wealth levels covering the period of the Great Recession Given the large labor market changes which have taken place during this time in Germany we expect the process of accumulating wealth to have changed Women have increased their labor market participation while men have had it more difficult We compare the role of factors for women and men Consequently in the second part of the paper we extend the existing literature and discuss the changing gender wealth gap which has resulted from this changes both at the mean and across the wealth distribution The rest of the paper is organised as follows The following section outlines the literature and conceptual backgrounds Section 3 describes the methods Section 4 presents the data The descriptives and empirical results are in Section 5 and 6 respectively followed by a summary and conclusions in Section 7 2 Literature and Conceptual Background As commonly known wealth is accumulated according to the standard life-cycle model where the stock of assets in the current period is the outcome of past decisions regarding investment labor market outcomes savings and consumption As discussed in Sierminska et al (2010) differences in any of these factors will give rise to a different accumulation pattern and consequently a different portfolio structure Consequently any type of macro-economic or life-shock will have a differential impact on individual portfolios according to this structure Among possible causes that have been shown to affect wealth accumulation differently for women and men are women and men save differently women have a weaker attachment to the labor market and women and men invest differently with diverging levels of returns The literature indicates that women and men differ in the risk attitudes with women being less risk tolerant and more risk-averse (Cartwright 2011) which would lead to less risky portfolios and lower rates of return Addionally financial literacy (eg Lusardi and Mitchell 2008) influence investment decisions It is shown that men and women differ in their financial knowledge which also leads to more conservative investments done by women Apart from having differential returns a more risk-loving individual having invested in risky assets will be more exposed to fluctuations in the (stock) market Similarly a person whose majority of assets are invested in real estate property will be very susceptible to changing house prices One of the the more important factors that has been shown in Sierminska et al (2010) to explain male-female differences in wealth accumulation is labor market differences It is not only the lower labor market participation rate of women the lower working hours as women commonly work part-time compared to men with the standard pattern of a continuous full-time employment but even also the still existing gender pay gap which impairs the wealth accumulation for women (Warren et al 2001) Even when holding savings rates constant women are thus expected to accumulate lower levels of wealth (eg Blau and Kahn 1997 2000)

4

Other factors which may affect wealth accumulation and in particular its differences by gender are for example inheritance and bestowals if women are men have differential access to these One factor important for this paper would also be occupations Women and men ten to cluster in different occupation that have different perspectives in advancement as well as different exposures to labor market flucturations and thus could be differentially exposed to the consequences of the Great Recession During the last financial crisis in Germany the labor market has been differentially affected across occupations Over the last two decades the labor market in Germany has observed some major changes Not only women have substanitally increased their participation rates but numerous labor market reforms have been implemented (Hartz reforms) (Dustmann et al 2014 Brenke 2015) which as we will show could have had an impact on the wealth accumulation of women and men Based on the overview of the literature several hypothesis come to mind that could be tested within our framework Hypothesis 1 A more risk-loving individual (men) having invested in risky assets will be more exposed to fluctuation in the stock market Similarly a person whose majority of assets are invested in real estate property will be very susceptible to changing house prices (men) and this will result in the changing wealth levels Hypothesis 2 Womenrsquos increased attachment to the labor market will improve their position in the wealth distribution Hypothesis 3 Higher prevelance of part-time work will have a negative (decreasing ) effect on wealth Hypothesis 4 The changing impact of labor market attachment will have a stronger impact on womenrsquos wealth Hypothesis 5 Labor market factors are less important for older individuals We will consequently address these in our results section 3 Methods As discussed in the introduction we will first focus on what were the causes of the changes in wealth over this period and then we will investigate the changing wealth gap

31 Change in wealth

To analyse the changes of wealth over time we starting by estimating a wealth equation by OLS considering the change in variables instead of the level ones The equation is

(1)

Where is the change in the level of (inverse hyperbolic real) net wealth ie

is a vector of control variables observed in the second period (ie in 2007 for the comparison

2002-2007 and in 2012 for the comparison 2007-2012) migratory background age age squared living in West or East Germany number of children below age 5 in the household number of

5

marriages length of current marriage number of months spent in fulltime and in parttime work in the previous 5 years long term unemployement in the previous 5 years (inverse hyperbolic real) permanent income household inheritances bequestgifts lottery winning and their value over the previous 5 years household savings in the previous 5 years and share of financial assets

is a vector of control variables observed in the first period (ie in 2002 for the comparison 2002-

2007 and in 2007 for the comparison 2007-2012) level of education (controlling also if shehe was still in education) occupational status (inverse hyperbolic real) permanent income tangible assets own property debts and position in the wealth distribution

is a vector of changes in control variables between and change in marital status change

in own property assets change in consumer credits is a random error normally distributed

are the parameters to be estimated by OLS

Equation 1 is estimated separately for men and women and for the period 2002-2007 2007-2012 In addition we also estimate the same equation for married individuals In that case we also control

for the permanent income of the spouse in and for the bargaining power in

32 Gender wealth gap in different periods

The second contribution of our paper is to examine the evolution of the gender wealth gap over time particularly before and after the crisis In the first instance we only consider those individuals that were married in period 13 We focus on married people to see clearly the changing role of explanatory factors for a more homogenous group which nevertheless comprises about 50 of the population and because we have the unique opportunity of doing so with our individual level data In this case we first apply the Oaxaca-Blinder decomposition (Blinder 1973 Oaxaca 1973) at the mean and then Firpo Fortin Lemieux decomposition (Firpo et al 2009) for the whole distribution The Oaxaca-Blinder decomposition relies on the estimation of the following equation for men and women similar to equation 1 but where the dependent variables is the (inverse hyperbolic real) net worth and not its change over time

(2)

The OB decomposition is the standard one (3)

Where =male =female The first component captures the effect which can be attributed to differences in characteristics while the second one captures the effect of differences in returns

33 Detailed decomposition of the gender wealth gaps

3 Throuout the paper with ldquoperiod 1rdquo we indicate 2002 in the comparison 2002-2007 and 2007 in the

comparison 2007-2012

6

In the final section of the paper we focus in performing a detailed decomposition of the gender wealth gaps over time This technique introduced in Firpo et al (2009) allows to decompose differences between two distributions of a variable It allows to have the individual contribution of each explanatory variable considered in the analysis via the characteristics and returns components In our case it will allow us to identify the explanatory factors and how the differences in their distribution and returns change due to the financial crisis and the changing labor market circumstances (characteristics and returns) in order to understand what is contributing to the gender wealth gap in Germany

The Firpo Fortin Lemiuex method is regression based which allows to apply it in a simiar way as the OB method The technique relies on the estimation of a regression where the dependent variable is replaced by a recentered influence function (RIF) transformation and so any distributional statics can be decomposed based on the regression results

In our case we will focus on the differences in quantiles

)()(

^^__^____

QF

QM

M

QM

FM

Q XXX (4)

Where Q is the difference in quantile (or other statistic) τ of the wealth distribution

FM

XX____

are the average observed characteristics for men (M) and women (F) FM

Q

^

are the

coefficients obtained from the regression of the RIF variables of quantile Qτ on the set of explanatory variables for women (F) and men (M) The first terms in equation (4) captures the effect on the differential between the distributions caused by differences in characteristics (explained component) The second term corresponds to the effect of the coefficient in which the contribution of each individual explanatory factor can be distinguished 4 Data We use the German Socio-Economic Panel (G-SOEP) a representative longitudinal survey on individuals in private household The survey started in 1984 in West Germany and extended to states in East Germany before the reunification in 1990 Every year about 15000 household are interviewd (25000 people) In 2002 there was a boost of higher-income people to better capture the upper margin of the income and wealth distribution There were other refreshment samples in 2006 2009 and 2011 while we couldnrsquot include the refreshment of 2012 since they were not asked information on wealth Although the SOEP has a special high-income sample there is still the problem of an undercoverage of the very rich The person with the highest net worth in the sample only holds almost 63 mio euro in 2002 Thus all multi-millionaires above that threshold and even also billionaires are missing although they had a significant impact of the wealth distribution (see Westermeier and Grabka 2015) Every year information on the sociodemographic variables are collected as well as information on education labour market and employment earnings and income household composition health and satisfaction In addition there are topic modules which are replicated about every 5 years and inquiries on specific topic We use mostly data from 2002 2007 and 2012 which contain information on individual wealth The initial sample has more than 23000 observations in 2002 almost 21000 in 2007 and slightly more than 18000 in 2012

41 Sample

7

Along the paper we use two different samples the first sample is composed by the three pooled cross-sections (2002 2007 and 2012) Using cross-sectional weights this sample is representative of the German population at every year We consider the pooled cross-sectional sample in presenting the mean and median wealth and for the descriptive The second sample is twofold one part is composed by people present in the survey in 2002 and in 2007 (Panel sample 2002-2007) the second part is composed by people present in the survey in 2007 and 2012 (Panel sample 2007-2012) These panel sample have the advantage to allow us to follow people over 5 years and track their changes however we lose some observations because of attrition In addition the refreshments which happened during the 5 years of observation which have the objective to cover demographic changes in the underlying population canrsquot be used either since those individuals were not present in our starting point Hence the main disadvantage from the panel sample comes from the attrition and from the fact that we have two samples for 2007 one comparable with 2002 and the other one comparable with 2012 We do not consider only people who are included in the survey from 2002 to 2012 because we would lose to many observations For the panel sample we use panel weights The maximum number of observations is shown in the Appendix table B1

42 Outcome variable

G-SOEP contains information on individual wealth in particular the following assets are details own property (and relative share and relative debt) other real estate (and relative share and relative debt) financial assets (and relative share) business assets tangible assets building loan and private insurance and consumer credits The main dependent variable is net total wealth in 2010 real price In G-SOEP the missing values for wealth are corrected for applying a multiple imputation technique (see Frick et al 2010) for the moment we are only using implicate A but we will later include also the other implicates We apply a 01 top coding and the inverse hyperbolic sine transformation useful to mitigate the effect of the outliners and to deal with skewed variables (with respect to the log transformation the inverse hyperbolic transformation is possible with negative and 0 value) In the first part of the analysis our dependent variables is the difference in net wealth between 2 periods while in the second part is the level in wealth and the gap between men and women The control variables are detailed in Appendix A 5 Descriptives We find that during the period of analysis the real mean wealth levels have been decreasing since 2002 for both women and men (Table 1) The wealth gender gap in 2002 hovered around 30 000 euros at the mean (for a discussion of the sources of the wealth gap in Germany in 2002 see Sierminska et al 2010) it was similar in 2007 and declined to about 27 000 in 2012 The decline in wealth 2002-2007 was similar for women and men (about 6000) and in the subsequent period 2007-2012 it was almost double for men as it was for women (6500 versus 3500) The effect at the median is even more striking A reduction in wealth for men in the pre-crisis period of about 4000 with only a decline of about 400 for women In the subsequent period an increase of 2000 for men and only 500 for women At the median the wealth gender gap increases in 2012 suggesting that there have been different changes for men and women along the wealth distribution Hence the effect of the Great Recession had a differential effect on men compared to women in terms of their wealth level We explore the sources of this differential change

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

4

Other factors which may affect wealth accumulation and in particular its differences by gender are for example inheritance and bestowals if women are men have differential access to these One factor important for this paper would also be occupations Women and men ten to cluster in different occupation that have different perspectives in advancement as well as different exposures to labor market flucturations and thus could be differentially exposed to the consequences of the Great Recession During the last financial crisis in Germany the labor market has been differentially affected across occupations Over the last two decades the labor market in Germany has observed some major changes Not only women have substanitally increased their participation rates but numerous labor market reforms have been implemented (Hartz reforms) (Dustmann et al 2014 Brenke 2015) which as we will show could have had an impact on the wealth accumulation of women and men Based on the overview of the literature several hypothesis come to mind that could be tested within our framework Hypothesis 1 A more risk-loving individual (men) having invested in risky assets will be more exposed to fluctuation in the stock market Similarly a person whose majority of assets are invested in real estate property will be very susceptible to changing house prices (men) and this will result in the changing wealth levels Hypothesis 2 Womenrsquos increased attachment to the labor market will improve their position in the wealth distribution Hypothesis 3 Higher prevelance of part-time work will have a negative (decreasing ) effect on wealth Hypothesis 4 The changing impact of labor market attachment will have a stronger impact on womenrsquos wealth Hypothesis 5 Labor market factors are less important for older individuals We will consequently address these in our results section 3 Methods As discussed in the introduction we will first focus on what were the causes of the changes in wealth over this period and then we will investigate the changing wealth gap

31 Change in wealth

To analyse the changes of wealth over time we starting by estimating a wealth equation by OLS considering the change in variables instead of the level ones The equation is

(1)

Where is the change in the level of (inverse hyperbolic real) net wealth ie

is a vector of control variables observed in the second period (ie in 2007 for the comparison

2002-2007 and in 2012 for the comparison 2007-2012) migratory background age age squared living in West or East Germany number of children below age 5 in the household number of

5

marriages length of current marriage number of months spent in fulltime and in parttime work in the previous 5 years long term unemployement in the previous 5 years (inverse hyperbolic real) permanent income household inheritances bequestgifts lottery winning and their value over the previous 5 years household savings in the previous 5 years and share of financial assets

is a vector of control variables observed in the first period (ie in 2002 for the comparison 2002-

2007 and in 2007 for the comparison 2007-2012) level of education (controlling also if shehe was still in education) occupational status (inverse hyperbolic real) permanent income tangible assets own property debts and position in the wealth distribution

is a vector of changes in control variables between and change in marital status change

in own property assets change in consumer credits is a random error normally distributed

are the parameters to be estimated by OLS

Equation 1 is estimated separately for men and women and for the period 2002-2007 2007-2012 In addition we also estimate the same equation for married individuals In that case we also control

for the permanent income of the spouse in and for the bargaining power in

32 Gender wealth gap in different periods

The second contribution of our paper is to examine the evolution of the gender wealth gap over time particularly before and after the crisis In the first instance we only consider those individuals that were married in period 13 We focus on married people to see clearly the changing role of explanatory factors for a more homogenous group which nevertheless comprises about 50 of the population and because we have the unique opportunity of doing so with our individual level data In this case we first apply the Oaxaca-Blinder decomposition (Blinder 1973 Oaxaca 1973) at the mean and then Firpo Fortin Lemieux decomposition (Firpo et al 2009) for the whole distribution The Oaxaca-Blinder decomposition relies on the estimation of the following equation for men and women similar to equation 1 but where the dependent variables is the (inverse hyperbolic real) net worth and not its change over time

(2)

The OB decomposition is the standard one (3)

Where =male =female The first component captures the effect which can be attributed to differences in characteristics while the second one captures the effect of differences in returns

33 Detailed decomposition of the gender wealth gaps

3 Throuout the paper with ldquoperiod 1rdquo we indicate 2002 in the comparison 2002-2007 and 2007 in the

comparison 2007-2012

6

In the final section of the paper we focus in performing a detailed decomposition of the gender wealth gaps over time This technique introduced in Firpo et al (2009) allows to decompose differences between two distributions of a variable It allows to have the individual contribution of each explanatory variable considered in the analysis via the characteristics and returns components In our case it will allow us to identify the explanatory factors and how the differences in their distribution and returns change due to the financial crisis and the changing labor market circumstances (characteristics and returns) in order to understand what is contributing to the gender wealth gap in Germany

The Firpo Fortin Lemiuex method is regression based which allows to apply it in a simiar way as the OB method The technique relies on the estimation of a regression where the dependent variable is replaced by a recentered influence function (RIF) transformation and so any distributional statics can be decomposed based on the regression results

In our case we will focus on the differences in quantiles

)()(

^^__^____

QF

QM

M

QM

FM

Q XXX (4)

Where Q is the difference in quantile (or other statistic) τ of the wealth distribution

FM

XX____

are the average observed characteristics for men (M) and women (F) FM

Q

^

are the

coefficients obtained from the regression of the RIF variables of quantile Qτ on the set of explanatory variables for women (F) and men (M) The first terms in equation (4) captures the effect on the differential between the distributions caused by differences in characteristics (explained component) The second term corresponds to the effect of the coefficient in which the contribution of each individual explanatory factor can be distinguished 4 Data We use the German Socio-Economic Panel (G-SOEP) a representative longitudinal survey on individuals in private household The survey started in 1984 in West Germany and extended to states in East Germany before the reunification in 1990 Every year about 15000 household are interviewd (25000 people) In 2002 there was a boost of higher-income people to better capture the upper margin of the income and wealth distribution There were other refreshment samples in 2006 2009 and 2011 while we couldnrsquot include the refreshment of 2012 since they were not asked information on wealth Although the SOEP has a special high-income sample there is still the problem of an undercoverage of the very rich The person with the highest net worth in the sample only holds almost 63 mio euro in 2002 Thus all multi-millionaires above that threshold and even also billionaires are missing although they had a significant impact of the wealth distribution (see Westermeier and Grabka 2015) Every year information on the sociodemographic variables are collected as well as information on education labour market and employment earnings and income household composition health and satisfaction In addition there are topic modules which are replicated about every 5 years and inquiries on specific topic We use mostly data from 2002 2007 and 2012 which contain information on individual wealth The initial sample has more than 23000 observations in 2002 almost 21000 in 2007 and slightly more than 18000 in 2012

41 Sample

7

Along the paper we use two different samples the first sample is composed by the three pooled cross-sections (2002 2007 and 2012) Using cross-sectional weights this sample is representative of the German population at every year We consider the pooled cross-sectional sample in presenting the mean and median wealth and for the descriptive The second sample is twofold one part is composed by people present in the survey in 2002 and in 2007 (Panel sample 2002-2007) the second part is composed by people present in the survey in 2007 and 2012 (Panel sample 2007-2012) These panel sample have the advantage to allow us to follow people over 5 years and track their changes however we lose some observations because of attrition In addition the refreshments which happened during the 5 years of observation which have the objective to cover demographic changes in the underlying population canrsquot be used either since those individuals were not present in our starting point Hence the main disadvantage from the panel sample comes from the attrition and from the fact that we have two samples for 2007 one comparable with 2002 and the other one comparable with 2012 We do not consider only people who are included in the survey from 2002 to 2012 because we would lose to many observations For the panel sample we use panel weights The maximum number of observations is shown in the Appendix table B1

42 Outcome variable

G-SOEP contains information on individual wealth in particular the following assets are details own property (and relative share and relative debt) other real estate (and relative share and relative debt) financial assets (and relative share) business assets tangible assets building loan and private insurance and consumer credits The main dependent variable is net total wealth in 2010 real price In G-SOEP the missing values for wealth are corrected for applying a multiple imputation technique (see Frick et al 2010) for the moment we are only using implicate A but we will later include also the other implicates We apply a 01 top coding and the inverse hyperbolic sine transformation useful to mitigate the effect of the outliners and to deal with skewed variables (with respect to the log transformation the inverse hyperbolic transformation is possible with negative and 0 value) In the first part of the analysis our dependent variables is the difference in net wealth between 2 periods while in the second part is the level in wealth and the gap between men and women The control variables are detailed in Appendix A 5 Descriptives We find that during the period of analysis the real mean wealth levels have been decreasing since 2002 for both women and men (Table 1) The wealth gender gap in 2002 hovered around 30 000 euros at the mean (for a discussion of the sources of the wealth gap in Germany in 2002 see Sierminska et al 2010) it was similar in 2007 and declined to about 27 000 in 2012 The decline in wealth 2002-2007 was similar for women and men (about 6000) and in the subsequent period 2007-2012 it was almost double for men as it was for women (6500 versus 3500) The effect at the median is even more striking A reduction in wealth for men in the pre-crisis period of about 4000 with only a decline of about 400 for women In the subsequent period an increase of 2000 for men and only 500 for women At the median the wealth gender gap increases in 2012 suggesting that there have been different changes for men and women along the wealth distribution Hence the effect of the Great Recession had a differential effect on men compared to women in terms of their wealth level We explore the sources of this differential change

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

5

marriages length of current marriage number of months spent in fulltime and in parttime work in the previous 5 years long term unemployement in the previous 5 years (inverse hyperbolic real) permanent income household inheritances bequestgifts lottery winning and their value over the previous 5 years household savings in the previous 5 years and share of financial assets

is a vector of control variables observed in the first period (ie in 2002 for the comparison 2002-

2007 and in 2007 for the comparison 2007-2012) level of education (controlling also if shehe was still in education) occupational status (inverse hyperbolic real) permanent income tangible assets own property debts and position in the wealth distribution

is a vector of changes in control variables between and change in marital status change

in own property assets change in consumer credits is a random error normally distributed

are the parameters to be estimated by OLS

Equation 1 is estimated separately for men and women and for the period 2002-2007 2007-2012 In addition we also estimate the same equation for married individuals In that case we also control

for the permanent income of the spouse in and for the bargaining power in

32 Gender wealth gap in different periods

The second contribution of our paper is to examine the evolution of the gender wealth gap over time particularly before and after the crisis In the first instance we only consider those individuals that were married in period 13 We focus on married people to see clearly the changing role of explanatory factors for a more homogenous group which nevertheless comprises about 50 of the population and because we have the unique opportunity of doing so with our individual level data In this case we first apply the Oaxaca-Blinder decomposition (Blinder 1973 Oaxaca 1973) at the mean and then Firpo Fortin Lemieux decomposition (Firpo et al 2009) for the whole distribution The Oaxaca-Blinder decomposition relies on the estimation of the following equation for men and women similar to equation 1 but where the dependent variables is the (inverse hyperbolic real) net worth and not its change over time

(2)

The OB decomposition is the standard one (3)

Where =male =female The first component captures the effect which can be attributed to differences in characteristics while the second one captures the effect of differences in returns

33 Detailed decomposition of the gender wealth gaps

3 Throuout the paper with ldquoperiod 1rdquo we indicate 2002 in the comparison 2002-2007 and 2007 in the

comparison 2007-2012

6

In the final section of the paper we focus in performing a detailed decomposition of the gender wealth gaps over time This technique introduced in Firpo et al (2009) allows to decompose differences between two distributions of a variable It allows to have the individual contribution of each explanatory variable considered in the analysis via the characteristics and returns components In our case it will allow us to identify the explanatory factors and how the differences in their distribution and returns change due to the financial crisis and the changing labor market circumstances (characteristics and returns) in order to understand what is contributing to the gender wealth gap in Germany

The Firpo Fortin Lemiuex method is regression based which allows to apply it in a simiar way as the OB method The technique relies on the estimation of a regression where the dependent variable is replaced by a recentered influence function (RIF) transformation and so any distributional statics can be decomposed based on the regression results

In our case we will focus on the differences in quantiles

)()(

^^__^____

QF

QM

M

QM

FM

Q XXX (4)

Where Q is the difference in quantile (or other statistic) τ of the wealth distribution

FM

XX____

are the average observed characteristics for men (M) and women (F) FM

Q

^

are the

coefficients obtained from the regression of the RIF variables of quantile Qτ on the set of explanatory variables for women (F) and men (M) The first terms in equation (4) captures the effect on the differential between the distributions caused by differences in characteristics (explained component) The second term corresponds to the effect of the coefficient in which the contribution of each individual explanatory factor can be distinguished 4 Data We use the German Socio-Economic Panel (G-SOEP) a representative longitudinal survey on individuals in private household The survey started in 1984 in West Germany and extended to states in East Germany before the reunification in 1990 Every year about 15000 household are interviewd (25000 people) In 2002 there was a boost of higher-income people to better capture the upper margin of the income and wealth distribution There were other refreshment samples in 2006 2009 and 2011 while we couldnrsquot include the refreshment of 2012 since they were not asked information on wealth Although the SOEP has a special high-income sample there is still the problem of an undercoverage of the very rich The person with the highest net worth in the sample only holds almost 63 mio euro in 2002 Thus all multi-millionaires above that threshold and even also billionaires are missing although they had a significant impact of the wealth distribution (see Westermeier and Grabka 2015) Every year information on the sociodemographic variables are collected as well as information on education labour market and employment earnings and income household composition health and satisfaction In addition there are topic modules which are replicated about every 5 years and inquiries on specific topic We use mostly data from 2002 2007 and 2012 which contain information on individual wealth The initial sample has more than 23000 observations in 2002 almost 21000 in 2007 and slightly more than 18000 in 2012

41 Sample

7

Along the paper we use two different samples the first sample is composed by the three pooled cross-sections (2002 2007 and 2012) Using cross-sectional weights this sample is representative of the German population at every year We consider the pooled cross-sectional sample in presenting the mean and median wealth and for the descriptive The second sample is twofold one part is composed by people present in the survey in 2002 and in 2007 (Panel sample 2002-2007) the second part is composed by people present in the survey in 2007 and 2012 (Panel sample 2007-2012) These panel sample have the advantage to allow us to follow people over 5 years and track their changes however we lose some observations because of attrition In addition the refreshments which happened during the 5 years of observation which have the objective to cover demographic changes in the underlying population canrsquot be used either since those individuals were not present in our starting point Hence the main disadvantage from the panel sample comes from the attrition and from the fact that we have two samples for 2007 one comparable with 2002 and the other one comparable with 2012 We do not consider only people who are included in the survey from 2002 to 2012 because we would lose to many observations For the panel sample we use panel weights The maximum number of observations is shown in the Appendix table B1

42 Outcome variable

G-SOEP contains information on individual wealth in particular the following assets are details own property (and relative share and relative debt) other real estate (and relative share and relative debt) financial assets (and relative share) business assets tangible assets building loan and private insurance and consumer credits The main dependent variable is net total wealth in 2010 real price In G-SOEP the missing values for wealth are corrected for applying a multiple imputation technique (see Frick et al 2010) for the moment we are only using implicate A but we will later include also the other implicates We apply a 01 top coding and the inverse hyperbolic sine transformation useful to mitigate the effect of the outliners and to deal with skewed variables (with respect to the log transformation the inverse hyperbolic transformation is possible with negative and 0 value) In the first part of the analysis our dependent variables is the difference in net wealth between 2 periods while in the second part is the level in wealth and the gap between men and women The control variables are detailed in Appendix A 5 Descriptives We find that during the period of analysis the real mean wealth levels have been decreasing since 2002 for both women and men (Table 1) The wealth gender gap in 2002 hovered around 30 000 euros at the mean (for a discussion of the sources of the wealth gap in Germany in 2002 see Sierminska et al 2010) it was similar in 2007 and declined to about 27 000 in 2012 The decline in wealth 2002-2007 was similar for women and men (about 6000) and in the subsequent period 2007-2012 it was almost double for men as it was for women (6500 versus 3500) The effect at the median is even more striking A reduction in wealth for men in the pre-crisis period of about 4000 with only a decline of about 400 for women In the subsequent period an increase of 2000 for men and only 500 for women At the median the wealth gender gap increases in 2012 suggesting that there have been different changes for men and women along the wealth distribution Hence the effect of the Great Recession had a differential effect on men compared to women in terms of their wealth level We explore the sources of this differential change

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

6

In the final section of the paper we focus in performing a detailed decomposition of the gender wealth gaps over time This technique introduced in Firpo et al (2009) allows to decompose differences between two distributions of a variable It allows to have the individual contribution of each explanatory variable considered in the analysis via the characteristics and returns components In our case it will allow us to identify the explanatory factors and how the differences in their distribution and returns change due to the financial crisis and the changing labor market circumstances (characteristics and returns) in order to understand what is contributing to the gender wealth gap in Germany

The Firpo Fortin Lemiuex method is regression based which allows to apply it in a simiar way as the OB method The technique relies on the estimation of a regression where the dependent variable is replaced by a recentered influence function (RIF) transformation and so any distributional statics can be decomposed based on the regression results

In our case we will focus on the differences in quantiles

)()(

^^__^____

QF

QM

M

QM

FM

Q XXX (4)

Where Q is the difference in quantile (or other statistic) τ of the wealth distribution

FM

XX____

are the average observed characteristics for men (M) and women (F) FM

Q

^

are the

coefficients obtained from the regression of the RIF variables of quantile Qτ on the set of explanatory variables for women (F) and men (M) The first terms in equation (4) captures the effect on the differential between the distributions caused by differences in characteristics (explained component) The second term corresponds to the effect of the coefficient in which the contribution of each individual explanatory factor can be distinguished 4 Data We use the German Socio-Economic Panel (G-SOEP) a representative longitudinal survey on individuals in private household The survey started in 1984 in West Germany and extended to states in East Germany before the reunification in 1990 Every year about 15000 household are interviewd (25000 people) In 2002 there was a boost of higher-income people to better capture the upper margin of the income and wealth distribution There were other refreshment samples in 2006 2009 and 2011 while we couldnrsquot include the refreshment of 2012 since they were not asked information on wealth Although the SOEP has a special high-income sample there is still the problem of an undercoverage of the very rich The person with the highest net worth in the sample only holds almost 63 mio euro in 2002 Thus all multi-millionaires above that threshold and even also billionaires are missing although they had a significant impact of the wealth distribution (see Westermeier and Grabka 2015) Every year information on the sociodemographic variables are collected as well as information on education labour market and employment earnings and income household composition health and satisfaction In addition there are topic modules which are replicated about every 5 years and inquiries on specific topic We use mostly data from 2002 2007 and 2012 which contain information on individual wealth The initial sample has more than 23000 observations in 2002 almost 21000 in 2007 and slightly more than 18000 in 2012

41 Sample

7

Along the paper we use two different samples the first sample is composed by the three pooled cross-sections (2002 2007 and 2012) Using cross-sectional weights this sample is representative of the German population at every year We consider the pooled cross-sectional sample in presenting the mean and median wealth and for the descriptive The second sample is twofold one part is composed by people present in the survey in 2002 and in 2007 (Panel sample 2002-2007) the second part is composed by people present in the survey in 2007 and 2012 (Panel sample 2007-2012) These panel sample have the advantage to allow us to follow people over 5 years and track their changes however we lose some observations because of attrition In addition the refreshments which happened during the 5 years of observation which have the objective to cover demographic changes in the underlying population canrsquot be used either since those individuals were not present in our starting point Hence the main disadvantage from the panel sample comes from the attrition and from the fact that we have two samples for 2007 one comparable with 2002 and the other one comparable with 2012 We do not consider only people who are included in the survey from 2002 to 2012 because we would lose to many observations For the panel sample we use panel weights The maximum number of observations is shown in the Appendix table B1

42 Outcome variable

G-SOEP contains information on individual wealth in particular the following assets are details own property (and relative share and relative debt) other real estate (and relative share and relative debt) financial assets (and relative share) business assets tangible assets building loan and private insurance and consumer credits The main dependent variable is net total wealth in 2010 real price In G-SOEP the missing values for wealth are corrected for applying a multiple imputation technique (see Frick et al 2010) for the moment we are only using implicate A but we will later include also the other implicates We apply a 01 top coding and the inverse hyperbolic sine transformation useful to mitigate the effect of the outliners and to deal with skewed variables (with respect to the log transformation the inverse hyperbolic transformation is possible with negative and 0 value) In the first part of the analysis our dependent variables is the difference in net wealth between 2 periods while in the second part is the level in wealth and the gap between men and women The control variables are detailed in Appendix A 5 Descriptives We find that during the period of analysis the real mean wealth levels have been decreasing since 2002 for both women and men (Table 1) The wealth gender gap in 2002 hovered around 30 000 euros at the mean (for a discussion of the sources of the wealth gap in Germany in 2002 see Sierminska et al 2010) it was similar in 2007 and declined to about 27 000 in 2012 The decline in wealth 2002-2007 was similar for women and men (about 6000) and in the subsequent period 2007-2012 it was almost double for men as it was for women (6500 versus 3500) The effect at the median is even more striking A reduction in wealth for men in the pre-crisis period of about 4000 with only a decline of about 400 for women In the subsequent period an increase of 2000 for men and only 500 for women At the median the wealth gender gap increases in 2012 suggesting that there have been different changes for men and women along the wealth distribution Hence the effect of the Great Recession had a differential effect on men compared to women in terms of their wealth level We explore the sources of this differential change

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

7

Along the paper we use two different samples the first sample is composed by the three pooled cross-sections (2002 2007 and 2012) Using cross-sectional weights this sample is representative of the German population at every year We consider the pooled cross-sectional sample in presenting the mean and median wealth and for the descriptive The second sample is twofold one part is composed by people present in the survey in 2002 and in 2007 (Panel sample 2002-2007) the second part is composed by people present in the survey in 2007 and 2012 (Panel sample 2007-2012) These panel sample have the advantage to allow us to follow people over 5 years and track their changes however we lose some observations because of attrition In addition the refreshments which happened during the 5 years of observation which have the objective to cover demographic changes in the underlying population canrsquot be used either since those individuals were not present in our starting point Hence the main disadvantage from the panel sample comes from the attrition and from the fact that we have two samples for 2007 one comparable with 2002 and the other one comparable with 2012 We do not consider only people who are included in the survey from 2002 to 2012 because we would lose to many observations For the panel sample we use panel weights The maximum number of observations is shown in the Appendix table B1

42 Outcome variable

G-SOEP contains information on individual wealth in particular the following assets are details own property (and relative share and relative debt) other real estate (and relative share and relative debt) financial assets (and relative share) business assets tangible assets building loan and private insurance and consumer credits The main dependent variable is net total wealth in 2010 real price In G-SOEP the missing values for wealth are corrected for applying a multiple imputation technique (see Frick et al 2010) for the moment we are only using implicate A but we will later include also the other implicates We apply a 01 top coding and the inverse hyperbolic sine transformation useful to mitigate the effect of the outliners and to deal with skewed variables (with respect to the log transformation the inverse hyperbolic transformation is possible with negative and 0 value) In the first part of the analysis our dependent variables is the difference in net wealth between 2 periods while in the second part is the level in wealth and the gap between men and women The control variables are detailed in Appendix A 5 Descriptives We find that during the period of analysis the real mean wealth levels have been decreasing since 2002 for both women and men (Table 1) The wealth gender gap in 2002 hovered around 30 000 euros at the mean (for a discussion of the sources of the wealth gap in Germany in 2002 see Sierminska et al 2010) it was similar in 2007 and declined to about 27 000 in 2012 The decline in wealth 2002-2007 was similar for women and men (about 6000) and in the subsequent period 2007-2012 it was almost double for men as it was for women (6500 versus 3500) The effect at the median is even more striking A reduction in wealth for men in the pre-crisis period of about 4000 with only a decline of about 400 for women In the subsequent period an increase of 2000 for men and only 500 for women At the median the wealth gender gap increases in 2012 suggesting that there have been different changes for men and women along the wealth distribution Hence the effect of the Great Recession had a differential effect on men compared to women in terms of their wealth level We explore the sources of this differential change

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

8

We investigate the following possible reasons -different levels of wealth to begin with -risk aversion affecting portfolio composition -more risky portfolio of men than women -more men are self- employed ndash lost businesses -change in labour market attachment of men vs women -more men unemployed then women in post-crisis period In fact our descriptive statistics in Table 5 indicate that women have a lower preference for risk than men (although increasing over time ndash just like men) As shown in Tables 7 and 8 men are more likely to own their home (by about 4 ppt) than women and to have debt either home or consumer debt More generally men are more likely to hold every type of assets (property other real estate financial business tangible building loan and private insurance consumer credits) Interestingly there has been a slight reduction in the percentage of men with business assets and the value of business assets (conditional on owning) is decreasing In addition the percentage of men with building loans and private insurance is decreasing while the same percentage is increasing among women When it comes to married men and women in principle comparable structures could be found although more wealth components are held Women are more likely to have low education but this has decreased by about 6 ppt since 2002 They still have a slight disadvantage in terms of high education but the gap remains at about 4 ppt compared to men This is a striking results since in most other developed countries women are now more educated than men (and this is also true for young German people ndash see OECD 2015) In terms of labour market attachments the number of month in full-time employment for men in the last 5 years has remained about the same with a slight increase for 2007 and subsequent decline while for women the numbers have been steadily increasing although on a slow pace There is still a 1 year difference between the sexes Women are more likely to work part-time but this is declining The number of women not in employment has also been decreasing and is down by 6 ppt since 2002 In terms of the type of employment we can say that the unemployment rate is quite similar for the two groups The share of pensioners has been increasing for men and decreasing for women What is striking is the large 8 ppt increase in white collar employment for women since 2002 which now exceeds that of men by 9 ppt As a result permanent income has been increasing for women and decreasing for men Both groups save regularly at a similar base but the levels for inheritances and gifts received have decreased by 20000 on average compared to the previous period Contrary to previous periods women average levels are now about 20000 more than menrsquos (in 2012) Another aspect that could be affecting this differential changes in wealth levels could be the demographic dimension and more specifically marital status transitions among men and women Marriage is a wealth-enhancing institution because married couples benefit from a joining of assets dual incomes and lowered expenses from economies of scale (Vespa and Painter 2008) See also Vespa and Painter 2011 ldquoOver time marriage positively correlates with wealth accumulation [hellip] We conclude that relationship history may shape long-term wealth accumulation and contrary to existing literature individuals who marry their only cohabiting partners experience a beneficial marital outcomerdquo

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

9

However in contrast marriage has negative effects on human-capital development for women as they usually suspend their employment career or at least reduce working-hours which impairs individual wealth accumulation for women It is well know that there exists a marriage premium for men in wages and a penalty for women and that substantial income and wealth loses occur due to divorce (Jarvis and Jenkis 1999 Zagorsky 2005) In our sample we find women less likely to be married than men and less likely to be single (by about 5ppt) Women are more likely to be widowed by about 7ppt and slightly more likely to be divorced (2ppt) When it comes to marital changes over the last 5 years we find that there is a 8 ppt gap in 200207 who remained married in favour of men Men are also more likely to remain never married (23-25 versus 18-21) and substantially less likely to remain single after a separation widowhood or divorce (11 versus 22) (even though part of the explanation is that women are overall more likely to be widowed) Thus womenrsquos labour market attachment education and participation in white collar employment could be driving their increase in permanent income and thus affecting the way their wealth has been changing Another aspect could be their changes in marital status which could also affect wealth levels and the composition of the portfolio which changed over the period we are considering 6 Results

61 Changes in wealth the role of explanatory factors

In the next step we isolate in more detail the factors that could affect changes in wealth levels and compare these for women and men We regress the inverse hyperbolic sine differences in wealth in the two periods on multiple explanatory factors In the overall sample (Table 9)4 the determinants of the (inverse hyperbolic sine) difference in the wealth vary over time and across genders Being a migrant has a significant negative effect for both men and women Usually one can argue that this has something to do with disrimination at the labor force however we certainly control for this aspect Thus one may argue that there is a migrant ldquodrawbackrdquo in wealth accumulation An strong factors is whether a person is currently living in East Germany Given the still existing lower earnings level and the general lower wealth level in that fromerly socialist region this still also impairs absolute wealth changes A High educational level (university) is associated with higher asolute wealth changes Besides better labor market changes and thus higher earning profiles for higher educated people this covariate may also point to a better financial literacy which facilitates better investment decisions In addition having a higher educational level is more benefical more women than for men In terms of employment variables we see that full-time employment has a positive effect on changes in wealth which definitely enables individuals to save on a regular basis Having said that long-term

4 In Table A2 we present the results of similar regressions where non-married people are separated into living alone and

cohabiting

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

10

unemployment has a negative effect on wealth accumulation Income losses due to job losses often offset by dissaving Being self-employed and in a white collar profession has an increasing positive effect only for men

Marital status changes When it comes to marital change it has been found that divorce usually has a negative effect on wealth In our sample this covariate show a negative sign however the effect is not significant The results for our sample indicate that for those that became widows in the past 5 years this increased their wealth as expected B but this effect is stronger for men than for women One potential interpretation is that men usually the one in a partnership who is responsible for investment decissions thus he is more able to handle and invest additional assets Usually one would also a positive effect for those getting married as both partners now profit from joining of assets dual incomes and lowered expenses from economies of scale However no significant effect can be found for our sample One potential explanation for this finding could be that marriage usually coincide with childbirth and finding a new home these additional costs may interfere the wealth accumulation process It can also be that the number of marital status changes in the 5 years is to small to identify significant changes

Portfolio effect The change in wealth will also depend on the performance of the portfolio For example if the portfolio consists of risky assets with a high expected return it will most likely perform better during an expansion compared to a financial crisis While a less risky portfolio with low expected returns would exhibit lower losses as well as lower gains In our results it seems that a preference for risk has a significant effect in the first period People that save have higher gains in wealth with a stronger effect for women In the following we examined whether people were purchasing or selling property and whether they were getting themselves into consumer debt in other to smooth consumption or finance their consumption needs The overall conclusion seems to be that people with property or those that purchased property are better off than In terms of debt those that get rid of consumer credits or property debts show a positive wealth change as they are forced to pay back their loan on a regular basis However those who newly borrowed have a negative effect on wealth change A very strong negative effect can also observed for those how changed their tenure status from owner occupiers to tenants Although in case of a sale one would assume that net worth Would not change while there is a shift from property wealth to financial assets in many cases bestowals to children or other relatives may the causal explanation

Quintile

People at every other quintile of net wealth lose with respect to people in the first quantile Those in the group of having debts in t0 are forced to pay back their loans on a regular basis and thus show positive wealth changes Those who had zero wealth in t0 are often young people who are at the beginning of the typical life cycle process of saving Thus with increased earnings the possibility to save also increase for this group On the other side those at the top are often elderly who hand over their wealth (by bestowals) to the next generation Additionally the higher the wealth the higher is also the probability to loose in absolute amounts

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

11

Married only When we perform the regression only on those that were married we observe some changes (Table 10) We also include additional variables such as the lagged income of the partner and lagged bargaining power First of all the labour market and employement variables lose significance for women that are married and the partner income is positive and statistically significant for women Being a migrant only matters in the first period With respect to the changes in marital status widowhood has a larger positive impact for men more than women Divorce has a negative impact in the first period significant only among women The number of marriages has a negative effect on changes in wealth but only in the second period for both genders The risk variable is no longer significant for women Saving still has a positive effect on both groups

62 Comparing the gender wealth gap in different periods

621 Regressions As a second step we investigate the gender wealth gap between married men and women in 2007 and 2012 In this case our dependent variable is the IHS of wealth (not the difference) To perform the OB decomposition we first of all run separated regressions for men and women in 2007 and in 2012 The effect of the control variables is similar to the previous specification but some changes can be noticed Being a migrant has a negative effect on the wealth of women also in the second period If we look back to the descriptive it can be noticed that there is an increase of 1st generation migrants among women while the percentage is stable among men Both the descriptive and the regressions point toward a different migration patter of migrant men and women to Germany Age has a positive effect larger in 2012 The level of education has a larger and effect for women than for men in 2007 while in 2012 the positive effect of education is similar for men and women Full time and part-time employment has a stronger effect in 2007 than in 2012 for both men and women On the other hand the long term unemployment has a stronger negative effect in 2012 for men wile for women it is similar in the two periods With respect to the occupations being self-employed has a positive and significant effect on wealth for men only larger in the first period the regression is reflecting what we have already seen in the descriptive that the value of business assets was decreasing over time Becoming widowed has no significant impact on the level of wealth while having a divorce in the previous 5 years has a negative impact on men wealth stable over time For women the effect of a divorce is negative and significant only in the first period On the other hand the negative effect of the number of marriages is stronger in 2012 Savings have a larger positive effect for women

622 OB decomposition

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

12

The OB decomposition is presented in Table 12 The gender wealth gap at the mean decreases from 2007 to 2012 Moreover the explained gender wealth gap is smaller and the unexplained component which is negative is larger In other words the wealth gap decreases as a consequence of smaller differences in characteristics between men and women but also due to slightly better returns for women which partially compensate The level of education which had no significant effect on the explained component in 2007 in 2012 has a positive effect on the explained gap indeed the returns to education in 2012 are very similar but men are more likely to have higher education than women On the other hand the occupational status is no longer significant in explaining the level of wealth and thus does not explain wealth differences among the two groups Permanent income has a smaller effect in 2012 while the income of the partner contributes more in reducing the explained gap It is worth noting that changes in property debt and in consumer credits have both a smaller (absolute) effect on the gap Both property and consumer credits no longer have a significant effect on the unexplained component in 2012 The largest effects come from permanent income and from the partner income the permanent income has a large and significant effect on the unexplained gap while the permanent income of the partner has a significant effect which contributes to reducing the unexplained gap The coefficients on these work in the opposite direction and even though they are still very large they are smaller in absolute terms in 2012 In addition the level of education (which had a negative effect on the unexplained gap) and the labour market status (positive) are not significant anymore indeed full time and part-time unemployment have no significant impact on the level of wealth in 2012 while the higher return to education for women which were present in 2007 are now at the same level for both sex Overall it seems that the smaller return to permanent income of women largely affect the gap as can be seen from the regressions as well the permanent income has a positive effect for men but almost no effect for women compensated by the higher effect for women wealth of partner income (Still ndash men do not compensate enough since the gwg among married is larger than in the total population)

623 FIRPO decomposition The results from the detailed decomposition serve to complement the above OB decomposition at the mean We examine these at the 10th 50th and 90th percentile in Table 13 As in the mean decomposition the wealth gap has decreased from 2007 and 2012 but the largest change is observed at the median There has not been much change at the top of the distribution either for women nor for men The role of education in explaining the gap is more or less constant throughout the distribution Though labour market status differences did not significantly explain the wealth gap at the mean they do explain a sizeable portion of the gap at the median and the upper half of the distribution In fact the differences have a diminishing effect on the gap as the share of men working full-time changed and the share of women working full-time and part-time increased in the first period (from 2002 to 2007) (Table 6) Subsequently the number decreases a bit from 2007 to 2012 and the effect is no longer significant in the second period

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

13

Differences in occupational status consistently contribute to the gender wealth gap over time As in our previous work differences in permanent income play a sizeable role ndash although in the second period this declines at the median Permanent income of the partner is not significant in explaining the gender wealth gap throughout the distribution Another factor that plays a sizeable role in the first period but then diminishes in size is the status of property ownership It explains about one-half and almost all the gap in the second period at the median Risk is only significant in the second period for the top half of the distribution In terms of returns Differences in labour market returns are sizeable and contribute to the gap only in the first period Overall returns in permanent household income have a diminishing effect on the gap in the first period and an effect twice as large in the second period Seems as if womenrsquos income works to diminish the gap The differences in the return to the share of financial assets in the portfolio also seem to diminish the gap 7 Summary and conclusions To be completed

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

14

References Blau FD and Kahn LM (1997) Swimming Upstream Trends in the Gender Wage Differential in the

1980s Journal of Labor Economics 15 1-42

Blau FD and Kahn LM (2000) Gender Differences in Pay Journal of Economic Perspectives 14 75-99

Blinder AS (1973) Wage discrimination reduced form and structural estimates Journal of Human Resources 8(4) 436-455

Brenke K (2015) Wachsende Bedeutung der Frauen auf dem Arbeitsmarkt DIW Wochenbericht Nr 5-2015 S 75-86

Cartwright Edward (2011) Behavioral econimcs Routledge New York Deere CD and Doss CR (2006) The gender asset gap What do we know and why does it matter

Feminist Economics 12(1-2) 1-50 EUROSTAT (2014) Wachstumsrate des realen BIPmdashVolumen (httpeppeurostateceuropaeutgm tabledotab=tableampinit=1ampplugin=1amplanguage=deamppcode=tec00115) Accessed 29 Oct 2014 Eurostat (2015) Life expectancy at birth by sex

httpeceuropaeueurostattgmtabledotab=tableampinit=1amplanguage=enamppcode=tps00025ampplugin=1

Firpo S Fortin N M and Lemieux T (2009) Unconditional Quantile Regressions Econometrica 77(3) 953-973 Frick JR Grabka MM and Marcus J (2010) Editing und multiple Imputation der Vermoumlgensinformation 2002 und 2007 im SOEP DIW Data Documentation No 51 Berlin Grabka M Marcus J and Sierminska E (2013) Wealth distribution within couples Review of

Economics of the Household online first 101007s11150-013-9229-2 Jarvis S and Jenkis SP (1999) Marital splits and income changes Evidence from the British

Household Panel Survey Population Studies 3(2) 237-254) Lusardi A amp Mitchell O S (2008) Planning and financial literacy How do women fare American

Economic Review Papers amp Proceedings 98(2) 413ndash417

Oaxaca R L (1973) Male-Female Wage Differentials in Urban Labor Markets International Economic Review 14(3) 693-709

OECD (2012) Closing the Gender Gap Act Now OECD Publishing httpdxdoiorg1017879789264179370-en

OECD (2015) Education and employment What are the gender differences Education Indicator in Focus n 30 Paris OECD

Sierminska E Frick JR and Grabka M (2010) Examining the gender wealth gap Oxford Economic Papers 62(4) 669-690

UNECE (2014) Statistical Database Fertility families and households httpw3uneceorgpxwebdatabaseSTAT30-GE02-Families_householdslang=1

US Census Bureau (2014) Americarsquos Families and Living Arrangements 2014 Family groups httpwwwcensusgovhhesfamiliesdatacps2014FGhtml

Vespa J and Painter MA (2008) ldquoThe Paths to Marriage Cohabitation and Marital Wealth Accumulationrdquo Paper presented at the Population Association of America Meeting New Orleans Vespa J Painter MA (2011) Cohabitation history marriage and wealth accumulation Demography 48(3)983-1004 doi 101007s13524-011-0043-2 Warren T Rowlingson K and Whyley C (2001) Female finances Gender Wage Gaps and Gender

Assets Gaps Work Employment and Society 15 465-488

Westermeier C and Grabka MM (2015) Significant Statistical Uncertainty over Share of High Net Worth Households Economic bulletin no 14+15 p 210-219

Zagorsky JL (2005) Marriage and divorce impact on wealth Journal of Sociology 41(4) 406ndash424

15

FIGURES

Fig 1 Wealth density over time (truncated)

16

TABLES Tab 1 Mean wealth and wealth difference over time by gender Cross-sections

Mean Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 106642 135845 40281 127015 77432 162938 52449 35659

Men 2007 100767 128069 45380 169715 81052 82013 59860 42036

Men 2012 94347 119209 37326 158985 103543 114782 54273 33813

Women 2002 76959 93159 29518 117536 39725 38089 23448 30395

Women 2007 70972 85897 29239 123674 40931 45809 29171 29084

Women 2012 67483 83038 28327 105268 49630 69216 28027 28478

Difference

Overall Married Single Widowed Divorced Separated Cohabiting Living alone

Men 2007-2002 -5875 -7776 5099 42700 3620 -80924 7411 6377

Men 2012-2007 -6420 -8860 -8054 -10730 22491 32769 -5587 -8223

Women 2007-2002 -5987 -7262 -279 6138 1206 7721 5723 -1311

Women 2012-2007 -3489 -2859 -911 -18405 8699 23406 -1144 -606

17

Tab 2 Mean wealth and wealth difference over time by gender Panel

Mean Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 108041 136933 138010 101088 72908 32174 44429 86185 126959 97629 138912

Men 2007 106958 131553 154391 80779 48649 72641 58316 79899 123838 69120 110566

Men 2007 116191 152250 121065 53710 79199 67663 66682 49240 201724 86226 73577

Men 2012 108804 143022 163213 52282 119929 72166 60708 48744 139996 91828 80364

Women 2002 79022 97702 106478 54484 63599 16445 38789 27902 118592 41276 42162

Women 2007 79011 92635 141341 37755 43067 42080 45874 41214 116356 41376 48398

Women 2007 64339 84901 80961 36430 46164 36587 27811 36217 99759 47696 78474

Women 2012 66837 87957 109402 60689 38392 36447 33606 42035 86805 48788 83456

Difference

Men 2007-2002 -1083 -5380 16381 -20309 -24259 40467 13887 -6286 -3120 -28509 -28346

Men 2012-2007 -7387 -9228 42148 -1428 40731 4503 -5974 -496 -61728 5601 6786

Women 2007-2002 -11 -5067 34863 -16729 -20532 25635 7085 13313 -2235 100 6236

Women 2012-2007 2498 3056 28441 24259 -7772 -141 5794 5818 -12953 1092 4982

18

Tab 3 Median wealth and wealth difference over time by gender Cross-sections

Median Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 24831 61492 2032 33860 9029 6772 5643 1129

Men 2007 20812 52029 2445 70239 8012 8325 6764 2081

Men 2012 23055 57253 1921 65322 11239 24015 9606 961

Women 2002 11851 37246 1129 18059 3386 2257 3386 677

Women 2007 11446 32258 2081 22893 4006 3642 4006 1852

Women 2012 12008 36023 2402 33622 2882 3362 4803 1921

Difference

Overall Married Single Widowed Divorced Separated Cohabiting Living alone

Men 2007-2002 -4019 -9463 414 36379 -1017 1553 1120 952

Men 2012-2007 2243 5224 -524 -4918 3227 15691 2842 -1121

Women 2007-2002 -405 -4988 952 4834 620 1385 620 1175

Women 2012-2007 561 3765 320 10729 -1124 -280 797 69

19

Tab 4 Median wealth and wealth difference over time by gender Panel

Median Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 28217 62077 62077 32731 8465 2257 3386 19752 36117 11287 16930

Men 2007 33819 62145 105099 5338 0 17690 7804 31493 25160 10406 3122

Men 2007 23933 54110 54631 28096 32570 10198 3434 23933 109261 8012 0

Men 2012 30740 54755 95900 6590 48031 14697 5764 8646 43708 3410 0

Women 2002 15801 42325 33860 8465 2409 1354 1693 5643 28217 4515 0

Women 2007 20499 41623 52029 1041 3122 9573 7284 8325 20812 4006 1041

Women 2007 12487 36420 43704 5723 6243 8325 2185 4162 27055 4891 8949

Women 2012 14409 40826 20173 0 3362 10951 4995 3842 25456 3362 3842

Difference

Men 2007-2002 5602 68 43022 -27393 -8465 15433 4418 11742 -10957 -881 -13808

Men 2012-2007 6806 645 41270 -21506 15460 4500 2330 -15288 -65553 -4602 0

Women 2007-2002 4698 -702 18169 -7424 713 8219 5591 2681 -7405 -508 1041

Women 2012-2007 1922 4406 -23532 -5723 -2881 2626 2810 -320 -1599 -1529 -5107

20

Tab 5 Descriptive statistics (overall sample)

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 10664183 10076712 9434735 7695914 7097231 6748333

IHS wealth 786 762 762 742 711 720

German 084 082 080 084 081 079

Migrant I 012 012 012 012 012 014

Migrant II 004 006 008 004 006 008

Age 4782 4855 4993 4975 4946 5021

Married 061 057 055 054 051 051

Single 027 030 031 021 024 026

Widowed 004 004 004 014 013 011

Divorced 007 008 008 009 010 011

Separated 002 002 002 002 002 002

Num of marriages 085 082 081 091 089 086

Length 1544 1506 1498 1348 1319 1340

N children in hh 010 008 008 010 010 009

Low educated 015 014 013 024 021 018

Lower voc edu 051 049 049 050 049 050

Upper voc edu 014 014 014 011 012 012

University 018 020 022 012 015 018

In education 002 003 002 002 002 002

East Germany 019 021 020 019 020 020

FT in previous 5y 2491 2729 2484 1106 1259 1288

PT in previous 5y 187 202 146 825 922 786

Unempl in previous 5y 009 012 009 007 010 008

Not employed 001 001 002 015 011 009

Unemployed 007 006 005 005 006 005

Trainee 006 006 006 005 005 005

Pensioner 025 026 027 029 028 027

Self-employed 007 007 007 003 003 004

White collar 023 022 026 027 029 035

Blue collar 024 023 020 011 010 009

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 001 002 002

IHS Perm income 1043 1020 1015 903 914 925

p10000 040 038 040 036 034 036

t10000 009 006 006 009 006 006

p10000x 020 019 019 017 016 017

c10000 014 018 019 009 014 014

Risk 469 482 523 373 402 440

Inheritance HH 005 008 008 005 007 008

Giftbequest HH 003 007 009 003 006 008

Lottery HH 000 001 001 000 001 001

HH saves regularly 093 087 087 093 085 086

Value inheritancesgift 7938461 5623754 3624136 7946933 5242972 5476717

Fin Assets share 021 020 020 020 027 022

Always married 058 054 050 048

Married gt widowed 001 001 003 002

Married gt divorcedsep 003 003 003 002

NM gt married 003 004 004 003

Always NM 023 025 018 021

Single gt married 001 001 001 002

Other 010 011 022 022

21

Tab 6 Descriptive statistics married people

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 13584507 12806926 11920905 9315925 8589698 8303831

IHS wealth 918 893 908 852 807 843

German 083 081 080 082 079 076

Migrant I 015 015 015 015 016 018

Migrant II 003 004 005 003 005 006

Age 5346 5515 5696 5042 5152 5306

N children in hh 014 012 012 014 014 013

Low educated 014 012 011 022 019 017

Lower voc edu 051 049 048 054 053 053

Upper voc edu 015 016 016 011 012 012

University 020 023 025 012 015 019

In education 000 000 000 000 000 000

East Germany 017 019 019 017 018 018

FT in previous 5y 2729 3092 2769 993 1134 1098

PT in previous 5y 153 178 132 1117 1254 1075

Unempl in previous 5y 008 010 005 008 010 006

Not employed 001 001 001 024 019 016

Unemployed 005 004 003 005 005 004

Trainee 000 000 000 001 001 001

Pensioner 032 034 035 023 025 026

Self-employed 008 008 008 004 004 005

White collar 022 022 026 028 030 035

Blue collar 025 024 020 013 012 011

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 002 002 002

IHS Perm income 1084 1079 1072 851 880 907

Partner IHS Perm Inc 847 878 901 1084 1080 1071

Bargaining power 074 072 070 026 028 030

N of marriages 114 115 117 113 114 116

Length 2481 2587 2649 2442 2510 2560

p10000 055 054 057 050 049 052

t10000 010 007 007 011 007 008

p10000x 029 027 027 027 026 026

c10000 014 020 020 010 015 014

Risk 448 462 498 361 387 424

Inheritance HH 005 008 008 005 008 008

Giftbequest HH 003 007 009 003 008 009

Lottery HH 000 001 001 000 001 001

HH saves regularly 094 088 089 094 088 089

Value inheritancesgift 8282325 6254586 3913133 8725124 6097312 5128218

Fin Assets share 016 016 016 015 020 017

Always married 093 091 091 091

Married gt widowed 006 007 007 006

Married gt divorcedsep 002 002 002 003

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

16

TABLES Tab 1 Mean wealth and wealth difference over time by gender Cross-sections

Mean Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 106642 135845 40281 127015 77432 162938 52449 35659

Men 2007 100767 128069 45380 169715 81052 82013 59860 42036

Men 2012 94347 119209 37326 158985 103543 114782 54273 33813

Women 2002 76959 93159 29518 117536 39725 38089 23448 30395

Women 2007 70972 85897 29239 123674 40931 45809 29171 29084

Women 2012 67483 83038 28327 105268 49630 69216 28027 28478

Difference

Overall Married Single Widowed Divorced Separated Cohabiting Living alone

Men 2007-2002 -5875 -7776 5099 42700 3620 -80924 7411 6377

Men 2012-2007 -6420 -8860 -8054 -10730 22491 32769 -5587 -8223

Women 2007-2002 -5987 -7262 -279 6138 1206 7721 5723 -1311

Women 2012-2007 -3489 -2859 -911 -18405 8699 23406 -1144 -606

17

Tab 2 Mean wealth and wealth difference over time by gender Panel

Mean Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 108041 136933 138010 101088 72908 32174 44429 86185 126959 97629 138912

Men 2007 106958 131553 154391 80779 48649 72641 58316 79899 123838 69120 110566

Men 2007 116191 152250 121065 53710 79199 67663 66682 49240 201724 86226 73577

Men 2012 108804 143022 163213 52282 119929 72166 60708 48744 139996 91828 80364

Women 2002 79022 97702 106478 54484 63599 16445 38789 27902 118592 41276 42162

Women 2007 79011 92635 141341 37755 43067 42080 45874 41214 116356 41376 48398

Women 2007 64339 84901 80961 36430 46164 36587 27811 36217 99759 47696 78474

Women 2012 66837 87957 109402 60689 38392 36447 33606 42035 86805 48788 83456

Difference

Men 2007-2002 -1083 -5380 16381 -20309 -24259 40467 13887 -6286 -3120 -28509 -28346

Men 2012-2007 -7387 -9228 42148 -1428 40731 4503 -5974 -496 -61728 5601 6786

Women 2007-2002 -11 -5067 34863 -16729 -20532 25635 7085 13313 -2235 100 6236

Women 2012-2007 2498 3056 28441 24259 -7772 -141 5794 5818 -12953 1092 4982

18

Tab 3 Median wealth and wealth difference over time by gender Cross-sections

Median Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 24831 61492 2032 33860 9029 6772 5643 1129

Men 2007 20812 52029 2445 70239 8012 8325 6764 2081

Men 2012 23055 57253 1921 65322 11239 24015 9606 961

Women 2002 11851 37246 1129 18059 3386 2257 3386 677

Women 2007 11446 32258 2081 22893 4006 3642 4006 1852

Women 2012 12008 36023 2402 33622 2882 3362 4803 1921

Difference

Overall Married Single Widowed Divorced Separated Cohabiting Living alone

Men 2007-2002 -4019 -9463 414 36379 -1017 1553 1120 952

Men 2012-2007 2243 5224 -524 -4918 3227 15691 2842 -1121

Women 2007-2002 -405 -4988 952 4834 620 1385 620 1175

Women 2012-2007 561 3765 320 10729 -1124 -280 797 69

19

Tab 4 Median wealth and wealth difference over time by gender Panel

Median Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 28217 62077 62077 32731 8465 2257 3386 19752 36117 11287 16930

Men 2007 33819 62145 105099 5338 0 17690 7804 31493 25160 10406 3122

Men 2007 23933 54110 54631 28096 32570 10198 3434 23933 109261 8012 0

Men 2012 30740 54755 95900 6590 48031 14697 5764 8646 43708 3410 0

Women 2002 15801 42325 33860 8465 2409 1354 1693 5643 28217 4515 0

Women 2007 20499 41623 52029 1041 3122 9573 7284 8325 20812 4006 1041

Women 2007 12487 36420 43704 5723 6243 8325 2185 4162 27055 4891 8949

Women 2012 14409 40826 20173 0 3362 10951 4995 3842 25456 3362 3842

Difference

Men 2007-2002 5602 68 43022 -27393 -8465 15433 4418 11742 -10957 -881 -13808

Men 2012-2007 6806 645 41270 -21506 15460 4500 2330 -15288 -65553 -4602 0

Women 2007-2002 4698 -702 18169 -7424 713 8219 5591 2681 -7405 -508 1041

Women 2012-2007 1922 4406 -23532 -5723 -2881 2626 2810 -320 -1599 -1529 -5107

20

Tab 5 Descriptive statistics (overall sample)

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 10664183 10076712 9434735 7695914 7097231 6748333

IHS wealth 786 762 762 742 711 720

German 084 082 080 084 081 079

Migrant I 012 012 012 012 012 014

Migrant II 004 006 008 004 006 008

Age 4782 4855 4993 4975 4946 5021

Married 061 057 055 054 051 051

Single 027 030 031 021 024 026

Widowed 004 004 004 014 013 011

Divorced 007 008 008 009 010 011

Separated 002 002 002 002 002 002

Num of marriages 085 082 081 091 089 086

Length 1544 1506 1498 1348 1319 1340

N children in hh 010 008 008 010 010 009

Low educated 015 014 013 024 021 018

Lower voc edu 051 049 049 050 049 050

Upper voc edu 014 014 014 011 012 012

University 018 020 022 012 015 018

In education 002 003 002 002 002 002

East Germany 019 021 020 019 020 020

FT in previous 5y 2491 2729 2484 1106 1259 1288

PT in previous 5y 187 202 146 825 922 786

Unempl in previous 5y 009 012 009 007 010 008

Not employed 001 001 002 015 011 009

Unemployed 007 006 005 005 006 005

Trainee 006 006 006 005 005 005

Pensioner 025 026 027 029 028 027

Self-employed 007 007 007 003 003 004

White collar 023 022 026 027 029 035

Blue collar 024 023 020 011 010 009

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 001 002 002

IHS Perm income 1043 1020 1015 903 914 925

p10000 040 038 040 036 034 036

t10000 009 006 006 009 006 006

p10000x 020 019 019 017 016 017

c10000 014 018 019 009 014 014

Risk 469 482 523 373 402 440

Inheritance HH 005 008 008 005 007 008

Giftbequest HH 003 007 009 003 006 008

Lottery HH 000 001 001 000 001 001

HH saves regularly 093 087 087 093 085 086

Value inheritancesgift 7938461 5623754 3624136 7946933 5242972 5476717

Fin Assets share 021 020 020 020 027 022

Always married 058 054 050 048

Married gt widowed 001 001 003 002

Married gt divorcedsep 003 003 003 002

NM gt married 003 004 004 003

Always NM 023 025 018 021

Single gt married 001 001 001 002

Other 010 011 022 022

21

Tab 6 Descriptive statistics married people

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 13584507 12806926 11920905 9315925 8589698 8303831

IHS wealth 918 893 908 852 807 843

German 083 081 080 082 079 076

Migrant I 015 015 015 015 016 018

Migrant II 003 004 005 003 005 006

Age 5346 5515 5696 5042 5152 5306

N children in hh 014 012 012 014 014 013

Low educated 014 012 011 022 019 017

Lower voc edu 051 049 048 054 053 053

Upper voc edu 015 016 016 011 012 012

University 020 023 025 012 015 019

In education 000 000 000 000 000 000

East Germany 017 019 019 017 018 018

FT in previous 5y 2729 3092 2769 993 1134 1098

PT in previous 5y 153 178 132 1117 1254 1075

Unempl in previous 5y 008 010 005 008 010 006

Not employed 001 001 001 024 019 016

Unemployed 005 004 003 005 005 004

Trainee 000 000 000 001 001 001

Pensioner 032 034 035 023 025 026

Self-employed 008 008 008 004 004 005

White collar 022 022 026 028 030 035

Blue collar 025 024 020 013 012 011

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 002 002 002

IHS Perm income 1084 1079 1072 851 880 907

Partner IHS Perm Inc 847 878 901 1084 1080 1071

Bargaining power 074 072 070 026 028 030

N of marriages 114 115 117 113 114 116

Length 2481 2587 2649 2442 2510 2560

p10000 055 054 057 050 049 052

t10000 010 007 007 011 007 008

p10000x 029 027 027 027 026 026

c10000 014 020 020 010 015 014

Risk 448 462 498 361 387 424

Inheritance HH 005 008 008 005 008 008

Giftbequest HH 003 007 009 003 008 009

Lottery HH 000 001 001 000 001 001

HH saves regularly 094 088 089 094 088 089

Value inheritancesgift 8282325 6254586 3913133 8725124 6097312 5128218

Fin Assets share 016 016 016 015 020 017

Always married 093 091 091 091

Married gt widowed 006 007 007 006

Married gt divorcedsep 002 002 002 003

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

17

Tab 2 Mean wealth and wealth difference over time by gender Panel

Mean Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 108041 136933 138010 101088 72908 32174 44429 86185 126959 97629 138912

Men 2007 106958 131553 154391 80779 48649 72641 58316 79899 123838 69120 110566

Men 2007 116191 152250 121065 53710 79199 67663 66682 49240 201724 86226 73577

Men 2012 108804 143022 163213 52282 119929 72166 60708 48744 139996 91828 80364

Women 2002 79022 97702 106478 54484 63599 16445 38789 27902 118592 41276 42162

Women 2007 79011 92635 141341 37755 43067 42080 45874 41214 116356 41376 48398

Women 2007 64339 84901 80961 36430 46164 36587 27811 36217 99759 47696 78474

Women 2012 66837 87957 109402 60689 38392 36447 33606 42035 86805 48788 83456

Difference

Men 2007-2002 -1083 -5380 16381 -20309 -24259 40467 13887 -6286 -3120 -28509 -28346

Men 2012-2007 -7387 -9228 42148 -1428 40731 4503 -5974 -496 -61728 5601 6786

Women 2007-2002 -11 -5067 34863 -16729 -20532 25635 7085 13313 -2235 100 6236

Women 2012-2007 2498 3056 28441 24259 -7772 -141 5794 5818 -12953 1092 4982

18

Tab 3 Median wealth and wealth difference over time by gender Cross-sections

Median Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 24831 61492 2032 33860 9029 6772 5643 1129

Men 2007 20812 52029 2445 70239 8012 8325 6764 2081

Men 2012 23055 57253 1921 65322 11239 24015 9606 961

Women 2002 11851 37246 1129 18059 3386 2257 3386 677

Women 2007 11446 32258 2081 22893 4006 3642 4006 1852

Women 2012 12008 36023 2402 33622 2882 3362 4803 1921

Difference

Overall Married Single Widowed Divorced Separated Cohabiting Living alone

Men 2007-2002 -4019 -9463 414 36379 -1017 1553 1120 952

Men 2012-2007 2243 5224 -524 -4918 3227 15691 2842 -1121

Women 2007-2002 -405 -4988 952 4834 620 1385 620 1175

Women 2012-2007 561 3765 320 10729 -1124 -280 797 69

19

Tab 4 Median wealth and wealth difference over time by gender Panel

Median Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 28217 62077 62077 32731 8465 2257 3386 19752 36117 11287 16930

Men 2007 33819 62145 105099 5338 0 17690 7804 31493 25160 10406 3122

Men 2007 23933 54110 54631 28096 32570 10198 3434 23933 109261 8012 0

Men 2012 30740 54755 95900 6590 48031 14697 5764 8646 43708 3410 0

Women 2002 15801 42325 33860 8465 2409 1354 1693 5643 28217 4515 0

Women 2007 20499 41623 52029 1041 3122 9573 7284 8325 20812 4006 1041

Women 2007 12487 36420 43704 5723 6243 8325 2185 4162 27055 4891 8949

Women 2012 14409 40826 20173 0 3362 10951 4995 3842 25456 3362 3842

Difference

Men 2007-2002 5602 68 43022 -27393 -8465 15433 4418 11742 -10957 -881 -13808

Men 2012-2007 6806 645 41270 -21506 15460 4500 2330 -15288 -65553 -4602 0

Women 2007-2002 4698 -702 18169 -7424 713 8219 5591 2681 -7405 -508 1041

Women 2012-2007 1922 4406 -23532 -5723 -2881 2626 2810 -320 -1599 -1529 -5107

20

Tab 5 Descriptive statistics (overall sample)

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 10664183 10076712 9434735 7695914 7097231 6748333

IHS wealth 786 762 762 742 711 720

German 084 082 080 084 081 079

Migrant I 012 012 012 012 012 014

Migrant II 004 006 008 004 006 008

Age 4782 4855 4993 4975 4946 5021

Married 061 057 055 054 051 051

Single 027 030 031 021 024 026

Widowed 004 004 004 014 013 011

Divorced 007 008 008 009 010 011

Separated 002 002 002 002 002 002

Num of marriages 085 082 081 091 089 086

Length 1544 1506 1498 1348 1319 1340

N children in hh 010 008 008 010 010 009

Low educated 015 014 013 024 021 018

Lower voc edu 051 049 049 050 049 050

Upper voc edu 014 014 014 011 012 012

University 018 020 022 012 015 018

In education 002 003 002 002 002 002

East Germany 019 021 020 019 020 020

FT in previous 5y 2491 2729 2484 1106 1259 1288

PT in previous 5y 187 202 146 825 922 786

Unempl in previous 5y 009 012 009 007 010 008

Not employed 001 001 002 015 011 009

Unemployed 007 006 005 005 006 005

Trainee 006 006 006 005 005 005

Pensioner 025 026 027 029 028 027

Self-employed 007 007 007 003 003 004

White collar 023 022 026 027 029 035

Blue collar 024 023 020 011 010 009

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 001 002 002

IHS Perm income 1043 1020 1015 903 914 925

p10000 040 038 040 036 034 036

t10000 009 006 006 009 006 006

p10000x 020 019 019 017 016 017

c10000 014 018 019 009 014 014

Risk 469 482 523 373 402 440

Inheritance HH 005 008 008 005 007 008

Giftbequest HH 003 007 009 003 006 008

Lottery HH 000 001 001 000 001 001

HH saves regularly 093 087 087 093 085 086

Value inheritancesgift 7938461 5623754 3624136 7946933 5242972 5476717

Fin Assets share 021 020 020 020 027 022

Always married 058 054 050 048

Married gt widowed 001 001 003 002

Married gt divorcedsep 003 003 003 002

NM gt married 003 004 004 003

Always NM 023 025 018 021

Single gt married 001 001 001 002

Other 010 011 022 022

21

Tab 6 Descriptive statistics married people

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 13584507 12806926 11920905 9315925 8589698 8303831

IHS wealth 918 893 908 852 807 843

German 083 081 080 082 079 076

Migrant I 015 015 015 015 016 018

Migrant II 003 004 005 003 005 006

Age 5346 5515 5696 5042 5152 5306

N children in hh 014 012 012 014 014 013

Low educated 014 012 011 022 019 017

Lower voc edu 051 049 048 054 053 053

Upper voc edu 015 016 016 011 012 012

University 020 023 025 012 015 019

In education 000 000 000 000 000 000

East Germany 017 019 019 017 018 018

FT in previous 5y 2729 3092 2769 993 1134 1098

PT in previous 5y 153 178 132 1117 1254 1075

Unempl in previous 5y 008 010 005 008 010 006

Not employed 001 001 001 024 019 016

Unemployed 005 004 003 005 005 004

Trainee 000 000 000 001 001 001

Pensioner 032 034 035 023 025 026

Self-employed 008 008 008 004 004 005

White collar 022 022 026 028 030 035

Blue collar 025 024 020 013 012 011

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 002 002 002

IHS Perm income 1084 1079 1072 851 880 907

Partner IHS Perm Inc 847 878 901 1084 1080 1071

Bargaining power 074 072 070 026 028 030

N of marriages 114 115 117 113 114 116

Length 2481 2587 2649 2442 2510 2560

p10000 055 054 057 050 049 052

t10000 010 007 007 011 007 008

p10000x 029 027 027 027 026 026

c10000 014 020 020 010 015 014

Risk 448 462 498 361 387 424

Inheritance HH 005 008 008 005 008 008

Giftbequest HH 003 007 009 003 008 009

Lottery HH 000 001 001 000 001 001

HH saves regularly 094 088 089 094 088 089

Value inheritancesgift 8282325 6254586 3913133 8725124 6097312 5128218

Fin Assets share 016 016 016 015 020 017

Always married 093 091 091 091

Married gt widowed 006 007 007 006

Married gt divorcedsep 002 002 002 003

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

18

Tab 3 Median wealth and wealth difference over time by gender Cross-sections

Median Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 24831 61492 2032 33860 9029 6772 5643 1129

Men 2007 20812 52029 2445 70239 8012 8325 6764 2081

Men 2012 23055 57253 1921 65322 11239 24015 9606 961

Women 2002 11851 37246 1129 18059 3386 2257 3386 677

Women 2007 11446 32258 2081 22893 4006 3642 4006 1852

Women 2012 12008 36023 2402 33622 2882 3362 4803 1921

Difference

Overall Married Single Widowed Divorced Separated Cohabiting Living alone

Men 2007-2002 -4019 -9463 414 36379 -1017 1553 1120 952

Men 2012-2007 2243 5224 -524 -4918 3227 15691 2842 -1121

Women 2007-2002 -405 -4988 952 4834 620 1385 620 1175

Women 2012-2007 561 3765 320 10729 -1124 -280 797 69

19

Tab 4 Median wealth and wealth difference over time by gender Panel

Median Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 28217 62077 62077 32731 8465 2257 3386 19752 36117 11287 16930

Men 2007 33819 62145 105099 5338 0 17690 7804 31493 25160 10406 3122

Men 2007 23933 54110 54631 28096 32570 10198 3434 23933 109261 8012 0

Men 2012 30740 54755 95900 6590 48031 14697 5764 8646 43708 3410 0

Women 2002 15801 42325 33860 8465 2409 1354 1693 5643 28217 4515 0

Women 2007 20499 41623 52029 1041 3122 9573 7284 8325 20812 4006 1041

Women 2007 12487 36420 43704 5723 6243 8325 2185 4162 27055 4891 8949

Women 2012 14409 40826 20173 0 3362 10951 4995 3842 25456 3362 3842

Difference

Men 2007-2002 5602 68 43022 -27393 -8465 15433 4418 11742 -10957 -881 -13808

Men 2012-2007 6806 645 41270 -21506 15460 4500 2330 -15288 -65553 -4602 0

Women 2007-2002 4698 -702 18169 -7424 713 8219 5591 2681 -7405 -508 1041

Women 2012-2007 1922 4406 -23532 -5723 -2881 2626 2810 -320 -1599 -1529 -5107

20

Tab 5 Descriptive statistics (overall sample)

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 10664183 10076712 9434735 7695914 7097231 6748333

IHS wealth 786 762 762 742 711 720

German 084 082 080 084 081 079

Migrant I 012 012 012 012 012 014

Migrant II 004 006 008 004 006 008

Age 4782 4855 4993 4975 4946 5021

Married 061 057 055 054 051 051

Single 027 030 031 021 024 026

Widowed 004 004 004 014 013 011

Divorced 007 008 008 009 010 011

Separated 002 002 002 002 002 002

Num of marriages 085 082 081 091 089 086

Length 1544 1506 1498 1348 1319 1340

N children in hh 010 008 008 010 010 009

Low educated 015 014 013 024 021 018

Lower voc edu 051 049 049 050 049 050

Upper voc edu 014 014 014 011 012 012

University 018 020 022 012 015 018

In education 002 003 002 002 002 002

East Germany 019 021 020 019 020 020

FT in previous 5y 2491 2729 2484 1106 1259 1288

PT in previous 5y 187 202 146 825 922 786

Unempl in previous 5y 009 012 009 007 010 008

Not employed 001 001 002 015 011 009

Unemployed 007 006 005 005 006 005

Trainee 006 006 006 005 005 005

Pensioner 025 026 027 029 028 027

Self-employed 007 007 007 003 003 004

White collar 023 022 026 027 029 035

Blue collar 024 023 020 011 010 009

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 001 002 002

IHS Perm income 1043 1020 1015 903 914 925

p10000 040 038 040 036 034 036

t10000 009 006 006 009 006 006

p10000x 020 019 019 017 016 017

c10000 014 018 019 009 014 014

Risk 469 482 523 373 402 440

Inheritance HH 005 008 008 005 007 008

Giftbequest HH 003 007 009 003 006 008

Lottery HH 000 001 001 000 001 001

HH saves regularly 093 087 087 093 085 086

Value inheritancesgift 7938461 5623754 3624136 7946933 5242972 5476717

Fin Assets share 021 020 020 020 027 022

Always married 058 054 050 048

Married gt widowed 001 001 003 002

Married gt divorcedsep 003 003 003 002

NM gt married 003 004 004 003

Always NM 023 025 018 021

Single gt married 001 001 001 002

Other 010 011 022 022

21

Tab 6 Descriptive statistics married people

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 13584507 12806926 11920905 9315925 8589698 8303831

IHS wealth 918 893 908 852 807 843

German 083 081 080 082 079 076

Migrant I 015 015 015 015 016 018

Migrant II 003 004 005 003 005 006

Age 5346 5515 5696 5042 5152 5306

N children in hh 014 012 012 014 014 013

Low educated 014 012 011 022 019 017

Lower voc edu 051 049 048 054 053 053

Upper voc edu 015 016 016 011 012 012

University 020 023 025 012 015 019

In education 000 000 000 000 000 000

East Germany 017 019 019 017 018 018

FT in previous 5y 2729 3092 2769 993 1134 1098

PT in previous 5y 153 178 132 1117 1254 1075

Unempl in previous 5y 008 010 005 008 010 006

Not employed 001 001 001 024 019 016

Unemployed 005 004 003 005 005 004

Trainee 000 000 000 001 001 001

Pensioner 032 034 035 023 025 026

Self-employed 008 008 008 004 004 005

White collar 022 022 026 028 030 035

Blue collar 025 024 020 013 012 011

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 002 002 002

IHS Perm income 1084 1079 1072 851 880 907

Partner IHS Perm Inc 847 878 901 1084 1080 1071

Bargaining power 074 072 070 026 028 030

N of marriages 114 115 117 113 114 116

Length 2481 2587 2649 2442 2510 2560

p10000 055 054 057 050 049 052

t10000 010 007 007 011 007 008

p10000x 029 027 027 027 026 026

c10000 014 020 020 010 015 014

Risk 448 462 498 361 387 424

Inheritance HH 005 008 008 005 008 008

Giftbequest HH 003 007 009 003 008 009

Lottery HH 000 001 001 000 001 001

HH saves regularly 094 088 089 094 088 089

Value inheritancesgift 8282325 6254586 3913133 8725124 6097312 5128218

Fin Assets share 016 016 016 015 020 017

Always married 093 091 091 091

Married gt widowed 006 007 007 006

Married gt divorcedsep 002 002 002 003

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

19

Tab 4 Median wealth and wealth difference over time by gender Panel

Median Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 28217 62077 62077 32731 8465 2257 3386 19752 36117 11287 16930

Men 2007 33819 62145 105099 5338 0 17690 7804 31493 25160 10406 3122

Men 2007 23933 54110 54631 28096 32570 10198 3434 23933 109261 8012 0

Men 2012 30740 54755 95900 6590 48031 14697 5764 8646 43708 3410 0

Women 2002 15801 42325 33860 8465 2409 1354 1693 5643 28217 4515 0

Women 2007 20499 41623 52029 1041 3122 9573 7284 8325 20812 4006 1041

Women 2007 12487 36420 43704 5723 6243 8325 2185 4162 27055 4891 8949

Women 2012 14409 40826 20173 0 3362 10951 4995 3842 25456 3362 3842

Difference

Men 2007-2002 5602 68 43022 -27393 -8465 15433 4418 11742 -10957 -881 -13808

Men 2012-2007 6806 645 41270 -21506 15460 4500 2330 -15288 -65553 -4602 0

Women 2007-2002 4698 -702 18169 -7424 713 8219 5591 2681 -7405 -508 1041

Women 2012-2007 1922 4406 -23532 -5723 -2881 2626 2810 -320 -1599 -1529 -5107

20

Tab 5 Descriptive statistics (overall sample)

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 10664183 10076712 9434735 7695914 7097231 6748333

IHS wealth 786 762 762 742 711 720

German 084 082 080 084 081 079

Migrant I 012 012 012 012 012 014

Migrant II 004 006 008 004 006 008

Age 4782 4855 4993 4975 4946 5021

Married 061 057 055 054 051 051

Single 027 030 031 021 024 026

Widowed 004 004 004 014 013 011

Divorced 007 008 008 009 010 011

Separated 002 002 002 002 002 002

Num of marriages 085 082 081 091 089 086

Length 1544 1506 1498 1348 1319 1340

N children in hh 010 008 008 010 010 009

Low educated 015 014 013 024 021 018

Lower voc edu 051 049 049 050 049 050

Upper voc edu 014 014 014 011 012 012

University 018 020 022 012 015 018

In education 002 003 002 002 002 002

East Germany 019 021 020 019 020 020

FT in previous 5y 2491 2729 2484 1106 1259 1288

PT in previous 5y 187 202 146 825 922 786

Unempl in previous 5y 009 012 009 007 010 008

Not employed 001 001 002 015 011 009

Unemployed 007 006 005 005 006 005

Trainee 006 006 006 005 005 005

Pensioner 025 026 027 029 028 027

Self-employed 007 007 007 003 003 004

White collar 023 022 026 027 029 035

Blue collar 024 023 020 011 010 009

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 001 002 002

IHS Perm income 1043 1020 1015 903 914 925

p10000 040 038 040 036 034 036

t10000 009 006 006 009 006 006

p10000x 020 019 019 017 016 017

c10000 014 018 019 009 014 014

Risk 469 482 523 373 402 440

Inheritance HH 005 008 008 005 007 008

Giftbequest HH 003 007 009 003 006 008

Lottery HH 000 001 001 000 001 001

HH saves regularly 093 087 087 093 085 086

Value inheritancesgift 7938461 5623754 3624136 7946933 5242972 5476717

Fin Assets share 021 020 020 020 027 022

Always married 058 054 050 048

Married gt widowed 001 001 003 002

Married gt divorcedsep 003 003 003 002

NM gt married 003 004 004 003

Always NM 023 025 018 021

Single gt married 001 001 001 002

Other 010 011 022 022

21

Tab 6 Descriptive statistics married people

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 13584507 12806926 11920905 9315925 8589698 8303831

IHS wealth 918 893 908 852 807 843

German 083 081 080 082 079 076

Migrant I 015 015 015 015 016 018

Migrant II 003 004 005 003 005 006

Age 5346 5515 5696 5042 5152 5306

N children in hh 014 012 012 014 014 013

Low educated 014 012 011 022 019 017

Lower voc edu 051 049 048 054 053 053

Upper voc edu 015 016 016 011 012 012

University 020 023 025 012 015 019

In education 000 000 000 000 000 000

East Germany 017 019 019 017 018 018

FT in previous 5y 2729 3092 2769 993 1134 1098

PT in previous 5y 153 178 132 1117 1254 1075

Unempl in previous 5y 008 010 005 008 010 006

Not employed 001 001 001 024 019 016

Unemployed 005 004 003 005 005 004

Trainee 000 000 000 001 001 001

Pensioner 032 034 035 023 025 026

Self-employed 008 008 008 004 004 005

White collar 022 022 026 028 030 035

Blue collar 025 024 020 013 012 011

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 002 002 002

IHS Perm income 1084 1079 1072 851 880 907

Partner IHS Perm Inc 847 878 901 1084 1080 1071

Bargaining power 074 072 070 026 028 030

N of marriages 114 115 117 113 114 116

Length 2481 2587 2649 2442 2510 2560

p10000 055 054 057 050 049 052

t10000 010 007 007 011 007 008

p10000x 029 027 027 027 026 026

c10000 014 020 020 010 015 014

Risk 448 462 498 361 387 424

Inheritance HH 005 008 008 005 008 008

Giftbequest HH 003 007 009 003 008 009

Lottery HH 000 001 001 000 001 001

HH saves regularly 094 088 089 094 088 089

Value inheritancesgift 8282325 6254586 3913133 8725124 6097312 5128218

Fin Assets share 016 016 016 015 020 017

Always married 093 091 091 091

Married gt widowed 006 007 007 006

Married gt divorcedsep 002 002 002 003

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

20

Tab 5 Descriptive statistics (overall sample)

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 10664183 10076712 9434735 7695914 7097231 6748333

IHS wealth 786 762 762 742 711 720

German 084 082 080 084 081 079

Migrant I 012 012 012 012 012 014

Migrant II 004 006 008 004 006 008

Age 4782 4855 4993 4975 4946 5021

Married 061 057 055 054 051 051

Single 027 030 031 021 024 026

Widowed 004 004 004 014 013 011

Divorced 007 008 008 009 010 011

Separated 002 002 002 002 002 002

Num of marriages 085 082 081 091 089 086

Length 1544 1506 1498 1348 1319 1340

N children in hh 010 008 008 010 010 009

Low educated 015 014 013 024 021 018

Lower voc edu 051 049 049 050 049 050

Upper voc edu 014 014 014 011 012 012

University 018 020 022 012 015 018

In education 002 003 002 002 002 002

East Germany 019 021 020 019 020 020

FT in previous 5y 2491 2729 2484 1106 1259 1288

PT in previous 5y 187 202 146 825 922 786

Unempl in previous 5y 009 012 009 007 010 008

Not employed 001 001 002 015 011 009

Unemployed 007 006 005 005 006 005

Trainee 006 006 006 005 005 005

Pensioner 025 026 027 029 028 027

Self-employed 007 007 007 003 003 004

White collar 023 022 026 027 029 035

Blue collar 024 023 020 011 010 009

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 001 002 002

IHS Perm income 1043 1020 1015 903 914 925

p10000 040 038 040 036 034 036

t10000 009 006 006 009 006 006

p10000x 020 019 019 017 016 017

c10000 014 018 019 009 014 014

Risk 469 482 523 373 402 440

Inheritance HH 005 008 008 005 007 008

Giftbequest HH 003 007 009 003 006 008

Lottery HH 000 001 001 000 001 001

HH saves regularly 093 087 087 093 085 086

Value inheritancesgift 7938461 5623754 3624136 7946933 5242972 5476717

Fin Assets share 021 020 020 020 027 022

Always married 058 054 050 048

Married gt widowed 001 001 003 002

Married gt divorcedsep 003 003 003 002

NM gt married 003 004 004 003

Always NM 023 025 018 021

Single gt married 001 001 001 002

Other 010 011 022 022

21

Tab 6 Descriptive statistics married people

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 13584507 12806926 11920905 9315925 8589698 8303831

IHS wealth 918 893 908 852 807 843

German 083 081 080 082 079 076

Migrant I 015 015 015 015 016 018

Migrant II 003 004 005 003 005 006

Age 5346 5515 5696 5042 5152 5306

N children in hh 014 012 012 014 014 013

Low educated 014 012 011 022 019 017

Lower voc edu 051 049 048 054 053 053

Upper voc edu 015 016 016 011 012 012

University 020 023 025 012 015 019

In education 000 000 000 000 000 000

East Germany 017 019 019 017 018 018

FT in previous 5y 2729 3092 2769 993 1134 1098

PT in previous 5y 153 178 132 1117 1254 1075

Unempl in previous 5y 008 010 005 008 010 006

Not employed 001 001 001 024 019 016

Unemployed 005 004 003 005 005 004

Trainee 000 000 000 001 001 001

Pensioner 032 034 035 023 025 026

Self-employed 008 008 008 004 004 005

White collar 022 022 026 028 030 035

Blue collar 025 024 020 013 012 011

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 002 002 002

IHS Perm income 1084 1079 1072 851 880 907

Partner IHS Perm Inc 847 878 901 1084 1080 1071

Bargaining power 074 072 070 026 028 030

N of marriages 114 115 117 113 114 116

Length 2481 2587 2649 2442 2510 2560

p10000 055 054 057 050 049 052

t10000 010 007 007 011 007 008

p10000x 029 027 027 027 026 026

c10000 014 020 020 010 015 014

Risk 448 462 498 361 387 424

Inheritance HH 005 008 008 005 008 008

Giftbequest HH 003 007 009 003 008 009

Lottery HH 000 001 001 000 001 001

HH saves regularly 094 088 089 094 088 089

Value inheritancesgift 8282325 6254586 3913133 8725124 6097312 5128218

Fin Assets share 016 016 016 015 020 017

Always married 093 091 091 091

Married gt widowed 006 007 007 006

Married gt divorcedsep 002 002 002 003

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

21

Tab 6 Descriptive statistics married people

Variables Men 2002 Men 2007 Men 2012 Women 2002 Women 2007 Women 2012

Wealth 13584507 12806926 11920905 9315925 8589698 8303831

IHS wealth 918 893 908 852 807 843

German 083 081 080 082 079 076

Migrant I 015 015 015 015 016 018

Migrant II 003 004 005 003 005 006

Age 5346 5515 5696 5042 5152 5306

N children in hh 014 012 012 014 014 013

Low educated 014 012 011 022 019 017

Lower voc edu 051 049 048 054 053 053

Upper voc edu 015 016 016 011 012 012

University 020 023 025 012 015 019

In education 000 000 000 000 000 000

East Germany 017 019 019 017 018 018

FT in previous 5y 2729 3092 2769 993 1134 1098

PT in previous 5y 153 178 132 1117 1254 1075

Unempl in previous 5y 008 010 005 008 010 006

Not employed 001 001 001 024 019 016

Unemployed 005 004 003 005 005 004

Trainee 000 000 000 001 001 001

Pensioner 032 034 035 023 025 026

Self-employed 008 008 008 004 004 005

White collar 022 022 026 028 030 035

Blue collar 025 024 020 013 012 011

Low civil service 002 002 002 001 001 001

High civil service 003 003 003 002 002 002

IHS Perm income 1084 1079 1072 851 880 907

Partner IHS Perm Inc 847 878 901 1084 1080 1071

Bargaining power 074 072 070 026 028 030

N of marriages 114 115 117 113 114 116

Length 2481 2587 2649 2442 2510 2560

p10000 055 054 057 050 049 052

t10000 010 007 007 011 007 008

p10000x 029 027 027 027 026 026

c10000 014 020 020 010 015 014

Risk 448 462 498 361 387 424

Inheritance HH 005 008 008 005 008 008

Giftbequest HH 003 007 009 003 008 009

Lottery HH 000 001 001 000 001 001

HH saves regularly 094 088 089 094 088 089

Value inheritancesgift 8282325 6254586 3913133 8725124 6097312 5128218

Fin Assets share 016 016 016 015 020 017

Always married 093 091 091 091

Married gt widowed 006 007 007 006

Married gt divorcedsep 002 002 002 003

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

22

Tab 7 Portfolio composition (1) Percentage of people who have the following assets (and own

share when relevant info)

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 040 038 040 036 034 036

Own share 6216 6396 6505 6342 6384 6426

Property debt 051 050 047 047 047 046

Other real estate 011 011 012 009 009 009

Own share 7106 7356 7289 6805 6949 6758

Other real estate debt 045 045 043 042 042 040

Financial assets 048 050 049 043 046 045

Own share 7691 8133 8205 7960 8330 8402

Business assets 007 006 006 002 002 002

Tangible assets 009 006 006 009 006 006

Buil Loan and Pr Insurances 054 056 053 043 048 048

Consumer credits 014 018 019 009 014 014

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 055 054 057 050 049 052

Own share 5717 5794 5822 5373 5387 5415

Property debt 053 051 047 054 053 050

Other real estate 014 014 015 012 012 012

Own share 6699 6935 6801 6246 6394 6258

Other real estate debt 045 044 043 044 044 042

Financial assets 054 055 054 046 047 046

Own share 6689 7085 7148 6525 6858 7016

Business assets 008 007 007 003 003 003

Tangible assets 010 007 007 011 007 008

Buil Loan and Pr Insurances 059 062 059 048 051 052

Consumer credits 014 020 020 010 015 014

conditional on having propertyother real estate

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

23

Tab 8 Portfolio composition (2) Share of each wealth component over total gross wealth

Total sample Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 030 028 029 029 027 029

Other real estate 005 005 005 004 004 004

Business assets 003 002 002 001 001 001

tangible assets 002 001 001 002 001 001

Buil Loan and Pr Insurances 022 024 022 019 021 021

Financial assets 018 018 018 019 020 019

Property debts 006 006 006 005 006 006

Other real estate debts 001 001 001 001 001 001

Consumer credits 008 010 009 006 008 008

Married Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Own property 041 039 042 040 039 041

Other real estate 005 005 005 005 005 005

Business assets 003 003 003 001 001 001

tangible assets 001 001 001 002 001 002

Buil Loan and Pr Insurances 021 022 020 017 018 018

Financial assets 016 016 016 015 016 015

Property debts 008 009 009 009 010 009

Other real estate debts 001 002 002 001 001 001

Consumer credits 007 008 008 005 007 006

Tab 8b Portfolio composition (2) Business assets value

Men 02 Men 07 Men 12 Women 02 Women 07 Women 12

Total sample 18438 16913 13027 3720 4332 2387 Conditional on having business assets 289830 290928 222793 160091 212800 105889

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

24

Tab 9 Regression on the HIS difference in wealth overall sample

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -097 035 -088 -048

(033) (040) (030) (035)

Age 003 001 005 013+

(007) (007) (006) (007)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -066 -068 -030 -066

(040) (049) (036) (043)

Lagged lower voc edu 020 044 089 031

(035) (043) (027) (033)

Lagged upper voc edu 082+ 126 057 107

(044) (053) (037) (044)

Lagged university 169 222 193 170

(043) (050) (037) (042)

Lagged in educ -008 144 202 196+

(115) (136) (089) (105)

East Germany -191 -074 -152 -111

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -110 -016 -123 044

(044) (057) (039) (047)

Lagged Not empl 103 043 -025 106

(111) (126) (045) (053)

Lagged Trainee -007 084 054 208

(066) (081) (059) (074)

Lagged Pensioner 031 153 -002 -022

(056) (063) (051) (056)

Lagged Self empl 119 232 -024 145

(047) (054) (057) (062)

Lagged White collar 075 126 -039 075+

(033) (039) (033) (039)

Lagged civil serv low -009 071 047 190

(085) (103) (115) (126)

Lagged civil serv high -044 -001 034 106

(061) (069) (068) (077)

Lagged asint perm income -013 010 008 008

(010) (011) (005) (006)

Asint perm income 052 004 -002 002

(015) (012) (006) (007)

Lagged wealth quant II -248 -304 -233 -248

(039) (047) (032) (040)

Lagged wealth quant III -622 -636 -619 -679

(039) (047) (034) (040)

Lagged wealth quant IV -926 -956 -1045 -984

(048) (054) (044) (049)

Lagged wealth quant V -1434 -1462 -1417 -1364

(053) (061) (049) (055)

Married gt widowed 275 479 183 337

(117) (131) (078) (086)

Married gt divorcedsep -119 011 -111+ 061

(075) (093) (063) (083)

NM gt married 045 066 -117+ 153

(065) (076) (061) (071)

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

25

Always NM -009 -018 -062 -009

(059) (068) (054) (061)

Single gt married -015 007 -129 073

(098) (107) (089) (092)

Single -094+ 009 -080 -067

(057) (062) (049) (054)

Num of marriages -041 -052+ -023 -024

(027) (029) (024) (026)

Length -003+ -001 -004 -002

(001) (002) (001) (001)

Sell own property -921 -1003 -1002 -974

(064) (075) (048) (063)

Acquired own property 313 274 282 389

(055) (067) (052) (059)

Always tenant -461 -533 -584 -561

(044) (048) (042) (045)

No more conscredits 336 446 517 435

(055) (057) (060) (058)

Acquire conscredits -296 -151 -144 -262

(052) (058) (056) (059)

No conscredits 225 282 345 247

(045) (047) (050) (048)

Lagged tangible assets -124 -037 -049 -095

(037) (052) (033) (044)

Lagged property debt 135 055 126 113

(036) (040) (034) (038)

Lagged risk 010 003 008+ -005

(005) (006) (005) (005)

Inheritances 062 053 177 -008

(080) (094) (072) (079)

Lottery -222+ 135 -056 -033

(121) (142) (117) (131)

Saving 125 140 173 176

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 069 047 079 078

(012) (014) (009) (011)

Lagged Worried for finacial reasons (cont) -021 -030 -043 -073

(017) (020) (016) (018)

Mis empl -175+ 040 004 -219+

(098) (136) (089) (133)

Mis bequests -092 040 -010 -138

(084) (098) (077) (084)

Mis personal -121+ 038 -054 091

(069) (088) (057) (070)

Constant 227 223 435 356

(256) (277) (195) (222)

Adj R2 022 019 025 022

N 7562 5882 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

26

Tab 10 Regression on the IHS difference in wealth people married in period 1

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant -096 061 -113 -038

(043) (053) (040) (049)

Age 004 012 011 009

(010) (012) (010) (012)

Age squared -000 -000 -000 -000

(000) (000) (000) (000)

Num of children -071 -069 -011 -114+

(057) (071) (056) (068)

Lagged lower voc edu 066 -011 111 046

(047) (061) (037) (047)

Lagged upper voc edu 098+ 032 082 125

(057) (071) (050) (062)

Lagged university 198 144 188 142

(055) (067) (050) (059)

East Germany -227 -064 -169 -081

(039) (044) (037) (041)

Full time empl 003 005 000 001

(001) (001) (001) (001)

Part time empl 002 001 001 002

(002) (002) (001) (001)

Long term unempl -076 -031 -126 062

(063) (083) (054) (071)

Lagged Not empl 148 141 -063 079

(168) (191) (057) (070)

Lagged Trainee 115 -329 020 138

(217) (407) (138) (221)

Lagged Pensioner 022 146+ 012 -070

(070) (079) (066) (072)

Lagges Self empl 114+ 236 -019 102

(059) (070) (072) (082)

Lagged White collar 049 137 -053 039

(043) (051) (044) (052)

Lagged civil serv low -015 096 092 -038

(109) (136) (158) (178)

Lagged civil serv high -019 026 068 062

(075) (083) (089) (100)

Lagged Asint perm income -001 018 003 006

(027) (029) (007) (009)

Asint perm income 084 -001 -005 000

(027) (021) (007) (009)

Lagged Asint perm income partner -004 010 054 078

(006) (007) (019) (023)

Lagged bargaining power -273 166 203+ 164

(115) (131) (121) (140)

Lagged wealth quant II -281 -417 -273 -194

(061) (075) (050) (064)

Lagged wealth quant III -625 -618 -570 -644

(057) (068) (050) (061)

Lagged wealth quant IV -961 -959 -1055 -992

(065) (076) (059) (069)

Lagged wealth quant V -1479 -1522 -1429 -1351

(071) (084) (065) (077)

Married gt widowed 249+ 547 198+ 158

(138) (158) (105) (121)

Married gt divorcedsep -132 089 -119+ 047

(081) (107) (070) (092)

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

27

Num of marriages -060 -124 -017 -108

(040) (044) (038) (043)

Length -004+ -001 -003+ -006

(002) (002) (002) (002)

Sell own property -987 -1124 -1043 -1125

(077) (096) (058) (080)

Acquired own property 301 306 273 377

(070) (088) (065) (076)

Always tenant -502 -541 -526 -585

(053) (060) (052) (058)

No more conscredits 270 284 405 353

(070) (073) (080) (081)

Acquire conscredits -257 -173 -125 -218

(067) (076) (076) (084)

No conscredits 232 190 272 242

(056) (060) (067) (068)

Lagged tangible assets -187 049 -042 -109+

(046) (066) (041) (056)

Lagged property debt 110 009 154 122

(042) (047) (041) (046)

Lagged risk 010 002 008 -004

(006) (007) (006) (007)

Inheritance 019 -086 084 061

(113) (136) (102) (122)

Lottery -152 028 -296+ -006

(174) (185) (161) (171)

Saving 097 147 144 137

(044) (052) (041) (049)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 064 027 068 071

(018) (021) (015) (019)

Lagged Worried for finacial reasons -026 -034 -048 -084

(023) (027) (022) (025)

Mis empl -124 -003 -017 -256

(119) (165) (112) (173)

Mis bequests -108 -080 -135 -033

(118) (141) (107) (127)

Mis personal -131 002 035 155

(105) (150) (093) (119)

Lagged in educ -627

(931)

Constant 032 097 -045 -364

(416) (482) (348) (429)

Adj R2 020 018 023 020

N 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

Note when we do not control for lagged partner income and bar Power the positive effect of having a university degree is

slightly larger for men and larger for women (eg 212 in 0207) In addition working parttime in 0712 has an impact of 002

significant at 10 and the value of inheritances is sig at 10 in 0712 for women The results on other variables are very

similar

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

28

Tab 11 Regression on (IHS) wealth levels people married in period1 (in the appendix Used for OB

decomposition)

y = IHS wealth Men 07 Men 12 Women 07 Women 12

Migrant -122 -039+ -121 -086

(020) (023) (018) (022)

Age 015 021 015 017

(005) (005) (005) (005)

Age squared -000 -000 -000 -000+

(000) (000) (000) (000)

Num of children -022 -018 -064 -028

(026) (032) (026) (031)

Lagged lower voc edu 017 040 101 041+

(022) (027) (017) (021)

Lagged upper voc edu 050+ 076 112 057

(026) (032) (023) (028)

Lagged university 025 097 122 097

(025) (030) (023) (027)

East Germany -007 -037 -007 -007

(017) (019) (017) (018)

Full time empl 002 001+ 001+ -000

(001) (001) (001) (001)

Part time empl 002+ 000 000 001

(001) (001) (000) (000)

Long term unempl -062 -105 -089 -084

(029) (037) (025) (032)

Lagged Not empl 076 117 024 054+

(077) (085) (027) (031)

Lagged Trainee 084 -352+ 073 154

(100) (181) (064) (100)

Lagged Pensioner 072 -007 040 -021

(032) (035) (031) (033)

Lagges Self empl 147 098 050 051

(026) (031) (033) (037)

Lagged White collar 039 014 031 050

(020) (023) (021) (024)

Lagged civil serv low -019 053 164 -017

(050) (061) (073) (081)

Lagged civil serv high 025 -001 025 -049

(034) (037) (041) (045)

Lagged Asint perm income 020 015 003 005

(012) (013) (003) (004)

Asint perm income 022+ 016+ 003 007+

(012) (009) (003) (004)

Lagged Asint perm income partner 003 010 037 036

(003) (003) (009) (011)

Lagged bargaining power 003 109+ 059 063

(052) (058) (056) (063)

Married gt widowed -029 080 -030 -044

(063) (070) (049) (055)

Married gt divorcedsep -080 -090+ -095 026

(037) (047) (033) (042)

Num of marriages -030+ -068 -027 -079

(018) (020) (018) (019)

Length -001 -000 -001 -002

(001) (001) (001) (001)

Sell own property -479 -566 -563 -572

(035) (043) (027) (036)

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

29

Acquired own property -076 -010 -067 -035

(029) (036) (026) (030)

Always tenant -526 -513 -577 -531

(019) (020) (018) (020)

No more conscredits 526 433 654 485

(032) (032) (037) (036)

Acquire conscredits 173 127 249 -005

(030) (033) (035) (037)

No conscredits 531 462 648 486

(025) (026) (030) (030)

Lagged tangible assets 006 108 041 068

(021) (029) (019) (025)

Lagged property debt -004 020 -003 027

(019) (020) (018) (020)

Lagged risk 000 -000 007 -002

(003) (003) (003) (003)

Inheritance 015 079 050 032

(052) (060) (047) (055)

Lottery -004 039 -082 030

(080) (082) (075) (078)

Saving 173 238 196 234

(020) (023) (019) (022)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 067 069 074 061

(008) (009) (007) (008)

Lagged Worried for finacial reasons -058 -038 -040 -063

(010) (012) (010) (011)

Mis empl 018 -078 020 -030

(055) (074) (052) (078)

Mis bequests -038 063 -021 037

(054) (063) (050) (057)

Mis personal 017 -012 019 -018

(048) (067) (043) (054)

Lagged in educ -963

(433)

Constant -383 -684 -554 -498

(190) (214) (161) (194)

Adj R2 044 045 048 048

Obs 4969 3698 5154 3822

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001

Standard errors in parentheses

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

30

Tab 12 Oaxaca Blinder decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

Variables Means SE Means SE changes

overall Men 9665 0085 9710 0097

Women 9018 0083 9240 0095

difference 0647 0119 0470 0135

explained 1228 0253 1072 0262 sectsect

unexplained -0581 0256 -0602 0265 sect

explained groupmig 0008 0009 0002 0003

groupage 0119 0034 0120 0040

groupkids 0000 0001 0000 0001

groupeduc 0028 0023 0089 0025 sect

grouploc -0000 0001 0002 0004

grouplm 0140 0212 0042 0188

groupoccup 0082 0027 0034 0030 sectsect

groupinc 0850 0261 0545 0228 sectsect

grouppinc -0066 0059 -0190 0062 sect

groupbarg 0014 0237 0457+ 0244 sect

groupmarstat 0012 0014 -0015 0019

groupmarr -0011 0010 -0008 0012

groupprop 0280 0052 0212 0060 sectsect

groupcdebt -0232 0033 -0222 0036

grouport -0002 0006 -0016+ 0010 sect

grouprisk 0001 0026 -0003 0027

groupsav 0005 0012 0008 0020

groupvalsav -0000 0002 -0002 0004

groupfinsh -0012 0011 0001 0013

groupworried 0013 0008 0019 0008 sect

groupother -0001 0003 -0002 0006

unexplained groupmig 0001 0043 0061 0042

groupage 0204 1883 0505 2269

groupkids 0026 0023 0005 0025

groupeduc -0670 0214 0020 0279 sectsect

grouploc -0001 0057 -0078 0067

grouplm 0574+ 0319 0179 0291

groupoccup 0130 0166 -0070 0207

groupinc 3163 1181 1759 1223 sectsect

grouppinc -3744 1033 -2789 1212 sectsect

groupbarg -0159 0211 0133 0250

groupmarstat 0005 0038 0014 0044

groupmarr 0013 0554 0664 0635

groupprop 0238+ 0125 0087 0128

groupcdebt -1091 0371 -0132 0363 sectsect

grouport -0050 0088 0013 0094

grouprisk -0251 0154 0059 0180

groupsav -0235 0255 0094 0295

groupvalsav 0005 0009 0019 0018

groupfinsh -0013 0020 0014 0021

groupworried -0329 0265 0479 0313

groupother -0158 0647 0220 0743

_cons 1762 2493 -1857 2886

+ plt010 plt005 plt001 plt0001

sect the group variable was not significant in 2007 ndash or smaller magnitude and became significant in 2012 (or larger abs

value) sectsect the group variable was significant in 2007 and it is not anymore or it has a smaller magnitude

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

31

Tab 13 Firpo decomposition (2007 and 2012) married in period 1 y = as Wealth

2007 2012

10th percentile 50th percentile 90th percentile 10th percentile 50th percentile 90th percentile

Variables Coef SE Coef SE Coef SE Coef SE Coef SE Coef SE

overall Men 0085 0006 12247 0055 13472 0034 0092 0007 12309 0063 13477 0041

Women 0078 0005 11823 0060 13205 0032 0083 0005 12050 0071 13211 0037

difference 0007 0008 0424 0082 0267 0047 0009 0009 0259 0095 0266 0055

explained -0021 0021 0669 0146 0805 0115 -0014 0020 0715 0156 0744 0126

unexplained 0029 0021 -0245+ 0144 -0538 0120 0023 0021 -0457 0154 -0478 0132

explained groupmig 0000 0000 0002 0003 0000 0001 0000 0000 0000 0001 0000 0001

groupage 0008 0003 0032+ 0018 0053 0016 0008 0003 0054 0022 0020 0018

groupkids 0000 0000 0000 0001 0000 0000 0000 0000 0000 0001 0000 0000

groupeduc -0004 0002 0048 0013 0046 0011 0003 0002 0062 0015 0057 0013

grouploc 0000 0000 -0001 0006 -0001 0004 0000 0000 0004 0009 0002 0005

grouplm 0009 0018 -0260 0118 -0266 0100 0018 0015 -0074 0108 -0184 0094

groupoccup 0005 0002 0056 0015 0072 0016 -0001 0002 0029+ 0017 0065 0019

groupinc -0028 0022 0668 0145 0749 0123 -0050 0018 0504 0131 0672 0114

grouppinc 0000 0005 -0014 0033 0032 0028 -0011 0005 -0033 0035 0038 0031

groupbarg 0010 0020 -0086 0131 0064 0110 0042 0019 0004 0140 0037 0121

groupmarstat 0000 0001 0000 0008 -0004 0007 0001 0002 -0001 0011 -0026 0010

groupmarr -0001 0001 -0003 0005 -0013 0006 -0001 0001 -0008 0007 -0006 0006

groupprop 0006 0002 0285 0052 0073 0013 0002 0002 0214 0057 0074 0016

groupcdebt -0027 0004 -0055 0010 -0010+ 0005 -0024 0004 -0069 0012 -0017 0007

grouport 0002 0001 -0011 0005 -0027 0008 0001 0001 -0017 0006 -0032 0011

grouprisk -0003 0002 0009 0014 0033 0012 -0003 0002 0029+ 0015 0031 0013

groupsav 0000 0000 0001 0003 0000 0002 0000 0001 0002 0005 -0001 0007

groupvalsav 0000 0000 -0000 0000 -0001 0004 -0000 0000 -0001 0002 -0004 0007

groupfinsh -0000 0000 -0007 0007 -0001 0001 0000 0000 0001 0009 -0000 0000

groupworried 0001 0000 0007 0004 0004 0003 0001 0001 0014 0006 0015 0006

groupother -0000 0000 -0001 0002 0000 0001 -0000 0000 -0001 0003 0003 0005

unexplained groupmig -0001 0003 0022 0026 0009 0020 0000 0003 0018 0026 0021 0021

groupage 0331 0142 -1199 1131 0917 0880 0081 0168 0893 1401 -0351 1111

groupkids 0002 0002 0000 0014 0002 0011 -0001 0002 0012 0015 0007 0012

groupeduc -0026 0016 -0086 0128 -0115 0100 0005 0021 -0159 0170 0027 0137

grouploc 0006 0004 -0093 0035 0036 0027 -0006 0005 -0085 0042 0032 0033

grouplm 0015 0025 0475 0186 0282+ 0150 -0006 0022 0032 0174 0191 0144

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

32

groupoccup 0023+ 0012 0007 0101 -0111 0078 -0004 0015 0109 0129 0073 0101

groupinc -0065 0097 2016 0665 3307 0552 -0294 0096 2229 0716 3344 0606

grouppinc 0044 0068 -2285 0667 -3324 0483 0056 0082 -2695 0801 -3887 0585

groupbarg 0006 0016 -0030 0129 -0040 0099 0028 0018 -0283+ 0157 -0069 0122

groupmarstat 0001 0003 0040+ 0023 -0024 0018 -0004 0003 -0011 0027 0007 0022

groupmarr -0066 0042 0353 0335 -0230 0259 0013 0047 -0156 0395 0109 0311

groupprop -0020 0009 0163 0076 0140 0059 -0010 0009 0335 0079 0062 0062

groupcdebt -0002 0027 -0245 0228 -0122 0173 0056 0026 -0261 0228 -0094 0177

grouport 0004 0007 -0130 0053 -0001 0041 0010 0007 -0123 0058 0001 0046

grouprisk -0008 0012 -0055 0093 0100 0072 -0012 0013 0139 0112 0133 0088

groupsav 0003 0019 -0191 0154 0127 0119 0046 0022 -0241 0182 0025 0144

groupvalsav -0000 0001 -0000 0006 0008+ 0005 0001 0001 0011 0011 0023 0009

groupfinsh 0002 0002 -0030 0012 0004 0009 0002 0002 -0028 0013 -0011 0010

groupworried -0043 0020 0133 0160 -0143 0124 0008 0023 0242 0194 -0157 0153

groupother -0022 0049 -0068 0389 -0389 0302 0001 0055 0239 0459 -0333 0364

_cons -0154 0190 0957 1489 -0971 1165 0054 0214 -0674 1781 0368 1413

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

33

APPENDIX

A Control variables The control variables are the following ones Sociodemographic variables migration status (German first generation second generation migrants) age and age squared number of children under 5 years old in the household a dummy equal one if living in East Germany Level of education low educated (ISCED 0 1 2) lower vocational (ISCED 3) upper vocational (ISCED 4 5) university (ISCED 6) plus a dummy equal 1 if the individual is still in education Marital history number of marriages length of current marriage Employment history months spent in full-time employement in the previous 5 years months spent in part-time employment in the previous 5 years and a dummy for long term unemployement equal 1 if the person spent 12 months or more in unemployement Occupational status categorical variables not employed unemployed trainee (=1 if military apprentice or trainee) pensioner self-employment white collar(employee) blue collar (=1 if untrained trained or semi-trained worker foreman) low civil servants (low and middle) high civil servants (high and executive) Permanent income (inverse hyperbolic sine transformation) of 5-years average of individual total income (individual labour earnings unemployment benefits old age or other pensions subsistence allowance maternity benefit student grants alimony company or private pension Risk self-defined answering to the question ldquoAre you generally a person who is fully prepared to take risks or do you try to avoid taking risksrdquo and the possible answer are from 0 to 10 Since the preference for risk is not collected every year we use information from 2004 for 2002 and from 2008 for 2007 (or 2006 if missing in 2008) Inheritance giftbequest lottery takes value 1 if the household had inheritances (giftlottery) in the previous 5 years (for the previous 3 years in 2002) We also have the value of inheritancesbequestsgift but we are not able to distinguish among the 3 source of unexpected wealth Savings is a dummy variable equal to 1 if the HH was able to save regularly at least for one year in the previous 5 Financial assets share is constructed using the ratio among financial assets and non-financial assets interacted by a dummy which takes value 1 if the HH have stocks (because fin assets can also be savings accounts) When we used the ldquolaggedrdquo variable which means that we are exploiting the information of 5 years before (eg information from 2002 in 2007 and from 2007 in 2012) We also control for the individual position in the wealth distribution in year 1 (5 quintiles) Lagged Worried for financial reasons every year individuals are asked to answer ot answer to ldquoare you concerned with your own economic situationrdquo The variables in our setup takes value 1 if the person is not concerned at all 2 if she is somewhat concerned 3 if very concerned Missing values are imputed with value 2 We also define changes in marital status own property status consumer credits which are defined comparing the status in year 1 (eg 2002 or 2007) and year 2 (2007 or 2012) Marital status the individual can be always married (ref category) can become widowed from married divorced or separated from married she can be always never married or got married if she was previously never married or previously single (widoweddivorcedseparated) She is considered ldquosinglerdquo if she remained widowed divorced or separated or if she had any change among there 3 categories or from never married into widowed divorced or separated

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

34

Property the individual can be always owner (ref category) sell the property acquire property or be always tenant Consumer credits the individual can have consumer credits in both periods (ref category) get rid of them take out consumer credits or never have consumer credits In addition we also include some ldquomissing variablesrdquo ldquomissing employmentrdquo takes value 1 if the employment history or the occupational category was missing ldquomissing bequestrdquo takes value 1 if the inheritance gifts lottery (or their value) saving variable is missing ldquomissing personalrdquo takes 1 if variables for education marital status or marital history migration background risk preference are missing In the regression for married couples we also include the permanent income of the spouse and a variable controlling for the bargaining power this variable is constructed as the ratio among the personal permanent income and the permanent income of the couple (partner permanent income added to the personal one) When we perform the Oaxaca-Blinder decomposition and the Firpo decomposition the explanatory variables are grouped in the following way Groupmig migratory status Groupage age and ande squared Groupkids number of kids Groupeduc lagged level of education or lagged still in education Grouploc residence in East Germany Grouplm labour market (full time parttime unemployment not employed) Groupoccup lagged occupational status (trainee pensioner self empl white collar civil servants low or high) Groupinc personal permanent and lagged permanent income Grouppinc partner lagged income Groupbarg bargaining power Groupmarstat changes in marital status Groupmarr number of marriages and length Groupprop changes in own property Groupcdebt changes in consumer credits Grouport lagged real estate lagged tangible assets Grouprisk lagged risk Groupsav inheritances lottery savings Groupvalsav value of inheritances Groupfinshare financial assets share Groupworried Lagged Worried for financial reasons Groupother missing variables

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

35

B Additional tables

Tab B1 Sample size

Cross-sections Year Overall Married Single Widowed Divorced Separated Single

Cohabiting Living alone

Men 2002 11127 7163 2885 333 578 167 581 2304

Men 2007 9909 6140 2671 340 606 152 610 2085

Men 2012 8603 5318 2221 343 595 122 517 1719

Women 2002 11899 7214 2540 1066 879 199 628 1912

Women 2007 10859 6226 2460 1004 955 213 658 1825

Women 2012 9610 5434 2075 956 952 192 566 1521

Panel Year Overall Always married

Married gt Widowed

Married gt Divorced

Married gt Separated

NM gt Married

Always NM

Single gt Married Widowed Divorced Separated

Men 2002 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 7144 4639 69 79 68 276 1319 96 194 363 41

Men 2007 5559 3285 56 45 35 250 1187 103 167 368 62

Men 2012 5587 3285 56 45 35 259 1202 103 167 371 63

Women 2002 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 7958 4683 177 100 99 296 1190 105 648 610 50

Women 2007 6220 3317 147 43 43 274 1076 120 518 601 80

Women 2012 6246 3317 147 43 43 281 1092 120 519 602 81 Notes

1 Weighted

2 To conduct the analysis we focus on the changes for 2002-2007 and 2007-2012 To do so we need to keep the same individuals over the 2 periods We avoid doing so for the 3

periods (keeping the same individuals from 2002 to 2012) because the attrition would cause the loss of too many observations The mean and median wealth gap are presented both

considering the 3 samples for the 2002 2007 2012 cross-sections (ldquoCross-sectionsrdquo) and the 2 samples one for 2002-2007 and the other one for 2007-2012 (ldquoPanelrdquo)

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

36

Tab B2 Regression on the IHS difference in wealth overall sample (controlling for cohabiting people)

y = IHS wealth difference Men 0207 Men 0712 Women 0207 Women 0712

Migrant I -089 049 -087 -093

(038) (049) (034) (043)

Migrant II -114 008 -090+ 017

(055) (060) (050) (055)

Age 002 -000 006 012+

(007) (007) (006) (007)

Age squared 000 000 -000 -000

(000) (000) (000) (000)

Num of children -070+ -074 -044 -068

(041) (050) (036) (044)

Lagged lower voc edu 018 041 089 031

(035) (044) (027) (033)

Lagged upper voc edu 079+ 123 059 113

(044) (053) (037) (045)

Lagged university 166 223 198 183

(043) (050) (037) (042)

Lagged in educ 007 170 240 239

(116) (136) (090) (107)

East Germany -194 -076 -162 -125

(028) (031) (025) (028)

Full time empl 003 005 002 002

(001) (001) (001) (001)

Part time empl 002 002 002 002

(002) (002) (001) (001)

Long term unempl -113 -023 -132 034

(044) (057) (039) (047)

Lagged Not empl 106 046 -021 112

(111) (126) (045) (053)

Lagged Trainee -002 106 057 210

(066) (082) (060) (074)

Lagged Pensioner 035 153 -003 -011

(057) (063) (051) (056)

Lagges Self empl 120 226 -018 153

(047) (054) (057) (063)

Lagged White collar 081 124 -033 083

(033) (039) (033) (039)

Lagged civil serv low 000 070 069 215+

(084) (103) (115) (126)

Lagged civil serv high -032 007 053 132+

(061) (068) (068) (077)

Lagged IHS perm income -012 010 007 008

(010) (011) (005) (006)

IHS perm income 052 006 -002 001

(015) (012) (006) (007)

Lagged wealth quant II -253 -308 -234 -247

(039) (047) (032) (040)

Lagged wealth quant III -622 -633 -613 -673

(039) (047) (034) (040)

Lagged wealth quant IV -925 -944 -1036 -965

(048) (054) (044) (048)

Lagged wealth quant V -1431 -1445 -1407 -1338

(053) (060) (049) (055)

Married gt divorcedsep -108 012 -121+ 078

(074) (091) (064) (084)

Married gt widowed 260 454 157 312

(119) (133) (078) (087)

Cohabit gt married 112 114 -026 027

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

37

(082) (090) (074) (082)

Always cohabit 033 -022 -028 061

(087) (096) (078) (087)

Cohabit gt singleNM -108 047 015 -121

(131) (151) (112) (138)

NM gt married -053 -051 -065 240

(090) (120) (083) (107)

NM gt cohabit 009 101 116 065

(091) (105) (081) (095)

Always NM 001 -078 -091 -083

(066) (076) (061) (069)

Single gt married 004 031 -277 165+

(097) (104) (092) (098)

Single -099+ 010 -103 -083

(057) (063) (049) (054)

Num of marriages -036 -063 -013 -038

(029) (031) (025) (027)

Length -003+ -002 -004 -002+

(001) (002) (001) (001)

Sell own property -925 -1001 -1006 -980

(064) (075) (048) (063)

Acquired own property 310 278 286 392

(055) (067) (052) (059)

Always tenant -463 -527 -585 -555

(044) (048) (042) (045)

No more conscredits 339 445 514 436

(055) (057) (060) (058)

Acquire conscredits -292 -147 -147 -260

(052) (058) (056) (059)

No conscredits 231 288 348 258

(045) (047) (050) (048)

Lagged tangible assets -124 -039 -043 -087

(037) (052) (033) (044)

Lagged property debt 133 052 126 108

(036) (040) (034) (038)

Lagged risk 011 003 008+ -004

(005) (006) (005) (005)

Inheritance 088 060 147 099

(073) (087) (068) (073)

Gift 135+ 030 058 031

(072) (085) (066) (074)

Lottery -189 129 -093 025

(118) (138) (115) (129)

Saving 130 147 183 188

(033) (038) (028) (033)

Value inheritancesgifts 000 000 000 000

(000) (000) (000) (000)

Fin assets share 070 047 078 076

(012) (014) (009) (011)

Mis empl -178+ 039 005 -220+

(098) (136) (089) (133)

Mis bequests -038 032 -069 -066

(076) (089) (070) (077)

Mis personal -147+ 019 -075 092

(075) (094) (065) (077)

Constant 120 192 353+ 133

(253) (271) (190) (218)

R2 022 019 025 022

N 7562 5883 8373 6564

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

38

Tab B3 Regression on the (asint) wealth levels people never married in period 1 Y=as Wealth

Men 07 Men 12 Women 07 Women 12

Migrant -127 -045+ -049 -059

(037) (023) (036) (043)

Age -003 021 009 015

(010) (005) (007) (009)

Age squared 000 -000 -000 -000

(000) (000) (000) (000)

Num of children -050 -005 -072 -043

(040) (031) (030) (038)

Lagged lower voc edu 017 042 039 026

(039) (027) (037) (046)

Lagged upper voc edu 063 085 034 038

(057) (031) (048) (058)

Lagged university 129 102 136 167

(058) (029) (051) (059)

Lagged in educ 017 158 258

(081) (067) (085)

East Germany -020 -037 -039 -102

(031) (018) (029) (033)

Full time empl 004 001 002 002

(001) (001) (001) (001)

Part time empl 001 -000 003 002

(002) (001) (001) (001)

Long term unempl -143 -107 -147 -040

(043) (036) (043) (054)

Lagged Not empl -094 114 -053 178

(109) (084) (065) (078)

Lagged Trainee -041 -337 114 214

(048) (168) (047) (058)

Lagged Pensioner 052 -006 165 -010

(120) (034) (102) (136)

Lagges Self empl 097 111 276 144

(064) (030) (082) (091)

Lagged White collar 049 019 119 149

(040) (022) (041) (050)

Lagged civil serv low 120 096 -039 195

(107) (059) (131) (143)

Lagged civil serv high 001 001 -011 015

(110) (037) (083) (103)

Lagged Asint perm income 004 019+ -001 001

(007) (010) (006) (007)

Asint perm income 031 016+ 043 026

(011) (009) (012) (012)

NM gt married 232 137 166

(170) (172) (159)

Single 206 127 149

(207) (219) (189)

Num of marriages -257+ -073 -159 -044

(148) (018) (155) (137)

Length 015 -001 016 -016

(019) (001) (017) (021)

Sell own property -227+ -574 -375 -263+

(130) (041) (111) (144)

Acquired own property -198 -001 -029 -076

(079) (036) (080) (091)

Always tenant -482 -516 -471 -459

(068) (020) (070) (076)

No more conscredits 757 455 1034 687

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

39

(078) (032) (084) (077)

Acquire conscredits 136 142 335 026

(066) (033) (070) (071)

No conscredits 817 486 1071 738

(061) (025) (064) (062)

Lagged tangible assets 057 101 044 021

(055) (029) (055) (065)

Lagged property debt -006 023 103 -064

(085) (020) (086) (094)

Lagged risk -003 -001 010+ -005

(006) (003) (006) (007)

Inheritance 217 073 134 063

(076) (061) (065) (068)

Lottery -145 033 252 127

(121) (081) (126) (151)

Saving 208 238 216 203

(037) (023) (034) (040)

Value inheritancesgifts 000 000+ 000 000

(000) (000) (000) (000)

Fin assets share 071 067 052 062

(011) (009) (008) (009)

Lagged Worried for finacial reasons 004 -037 -035+ -043+

(020) (012) (019) (022)

Mis empl -036 -071 -342 -324

(204) (073) (216) (227)

Mis bequests 069 055 130+ -031

(085) (063) (074) (076)

Mis personal -098 -002 -077 -020

(060) (063) (060) (070)

Constant -304 -592 -1021 -549

(273) (204) (228) (264)

Adj R2 041 046 048 043

Obs 1781 3831 1667 1315

+ p-value lt010 p-value lt 005 p-value lt 001 p-value lt 0001 Standard errors in parentheses

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect

40

Tab B4 Oaxaca Blinder decomposition (2007 and 2012) never married in period 1 Y=as Wealth

2007 2012

Variables Means SE Means SE changes

overall group_1 6329 0165 6589 0179

group_2 6419 0164 6286 0183

difference -0090 0233 0303 0256

explained -0071 0188 -0001 0215

unexplained -0019 0207 0305 0236

explained groupmig -0031+ 0018 -0006 0014 sectsect

groupage -0013 0049 0002 0029

groupkids 0033 0027 -0018 0022

groupeduc -0045+ 0027 -0106 0047 sect

grouploc -0002 0004 -0015 0021

grouplm 0207 0090 0159 0098 sectsect

groupoccup -0046 0066 0031 0067

groupinc 0073 0033 0031 0025 sectsect

groupmarstat -0086 0070 -0205 0102 sect

groupmarr 0079 0067 0173+ 0095 sect

groupprop 0049 0051 0137 0062 sect

groupcdebt -0155 0095 -0038 0103

grouport 0005 0006 -0012 0025

grouprisk -0020 0039 -0028 0045

groupsav -0049 0042 -0019 0046

groupvalsav 0001 0003 -0070 0060

groupfinsh -0068 0034 -0054+ 0028

groupworried -0001 0006 0026 0020

groupother -0000 0017 0009 0017

unexplained groupmig -0098 0067 -0054 0075

groupage -2064 2446 -2957 2754

groupkids 0057 0101 0142 0098

groupeduc 0090 0370 1262 0446 sect

grouploc 0046 0110 -0032 0130

grouplm 0446 0496 0309 0542

groupoccup -0443 0332 -0599 0376

groupinc -0513 1628 -0579 1721

groupmarstat 0211 0510 0578 0487

groupmarr -0228 0499 -0724 0468

groupprop -0231 0897 -0140 0919

groupcdebt -2369 0843 0716 0765 sectsect

grouport -0046 0069 0085 0079

grouprisk -0621 0387 0012 0429

groupsav -0042 0435 0217 0516

groupvalsav -0008 0026 0106 0083

groupfinsh 0071 0059 -0072 0064

groupworried 0814 0524 0394 0644

groupother -0563 0996 0853 0961

_cons 5471+ 3288 0788 3525 sectsect