JEL classification: O1; N0; J16; J12Keywords: Marriage patterns; Female labour; Income inequality; Economic development
1. Introduction
“Whether the law of marriage be instituted or not,the dictate of nature and virtue seems to be anearly attachment to one woman.” (Malthus, 1798:pp. 15).
☆ This research was conducted while I was a PhD candidate at theUniversity of Warwick. I have also benefited from the comments ofparticipants at the Conference on “Marriage Patterns, Household For-mation, and Economic Development” at Utrecht 7–10 October 2010.Special thanks to Maarten Prak who provided further and valuable in-sight as the main discussant. The errors and views expressed in thispaper are mine.⁎ Departament d'Anàlisi Econòmica, Universitat de València, Av.
This paper looks at the relationship between econom-ic development and mating (marriage) patterns, thushousehold formation, in the long run. As the economicenvironment evolves, we expect culture, social norms,and also institutions to reflect these changes (North,1991). Households are no exceptions, and the economicrole of women and the distribution of resources withina society have been changing over time. This paperprovides a general overview of the development andevolution of mating and marriage patterns over time.The process of household formation entails both, menand women. As a result, the changing economic role ofwomen observed across societies and over time couldhave influenced this process. Additionally, with theoccurrence of economic development the distribution ofresources within a society could have also affected theincentives structure of the average man (groom) andwoman (bride).
4 This would not be the case for wealthy men, whose decision of mar-riage would be independent of economic constraints. However wealthymen have been far from the average man throughout history. If this was
313A. Díez Minguela / History of the Family 16 (2011) 312–330
To simplify matters, we concentrate on two specificaspects related to the formation of a household: (a) matingsystems (monogamous as opposed to polygynous unions),and (b) the ‘European’ marriage pattern which results inlate monogamous attachments and sizeable groups ofpermanent celibates (Hajnal, 1965).1 In this regard, theeconomic role of women within a society and the distribu-tion of resources will be at the core of our main discussion.For example, it has been recently argued that increasingmale wage inequality in cities could induce women tosearch longer, and hence postpone marriage (Gould &Paserman, 2003; Loughran, 2002). However, searching iscostly and women earning no income may not be able towait for “Mr. Right”. What women want, and whatwomen can achieve, today or in the distant past deservesfurther study and discussion.
Our reasoning follows a sequence of events. First, weobserve hunter–gatherers at the earliest stage of economicdevelopment. Lovejoy (1981: pp. 346) argued thatMiocene critical ecological conditions “were prime forthe establishment of male parental investment and amonogamous mating structure”. Ethnographic evidencesupports the view that among hunter–gatherers monogamypredominates (Murdock, 1967).2 Nonetheless, it is widelyobserved within sedentary societies, mostly in sub-SaharanAfrica, that polygynous unions have been somewhatfrequent.3 Secondly, we look at the ‘European’ marriagepattern. The high age at marriage recorded in some North-west European countries provides further evidence thatmarriage patternsmay reflect economic conditions. Late at-tachments are costly. If women earn no income, theirnatal household will bear the costs associated with late at-tachments. Parents may be keen on waiting, and spendingtime and resources to reduce uncertainty about the traits ofa potential spouse. Nevertheless, the costs of living, andother costs associated with pre-marital sex or socialnorms regarding gender specific roles, could encouragewomen to marry young. Additionally, men might be
1 Hajnal (1965: 101): “The marriage pattern of most Europe as itexisted for at least two centuries up to 1940 was, so far as we can tell,unique or almost unique in the world. There is no known example of apopulation of non-European civilisation which has had a similar pat-tern. The distinctive marks of the ‘European pattern’ are (1) a highage at marriage and (2) a high proportion of people who never marryat all. The ‘European’ pattern pervaded the whole of Europe except forthe eastern and south-eastern portion.”2 The Ethnographic Atlas shows that 61.3% of nomadic or semi-
nomadic societies were monogamous and independent nuclear, while38.2% were polygynous—mostly with co-wives sharing habitation—and 0.5% were polyandrous.3 Dorjahn (1959), Chamie (1986), United Nations (1990) and the
Demographic and Health Surveys present empirical evidence on theincidence of polygyny across countries.
required to postpone marriage in order to acquire enoughwealth to provide for the newly formed household.4
The paper is structured as follows. Section 2 willdiscuss the Goldin curve, which associates female labourparticipation and economic development (Goldin, 1994).While Section 3 will explore the relationship between theincome distribution within an economy and economicdevelopment orKuznets curve (Kuznets, 1955). Section 4will combine both results and further discuss matters.Finally, Section 5 will offer some concluding remarks.
2. The Goldin curve
Goldin (1994) explored the relationship between eco-nomic development and the female labour force, andfound that the labour participation rates of marriedwomen first declines and then rises as countries devel-op, hence the U-shaped relationship or Goldin curve.5
Goldin (1994) argued that the ensuing U-shaped rela-tionship could be the result of a strong income effectand a weak substitution effect accompanied by a changein the location of production, from home to the factory.In conjunction with the process of economic develop-ment, female labour participation decreased becausefamilies often implicitly bought women's work, andconsequently women retreated into the home.6 More-over, a decrease in the demand for women's labour inagriculture, and social stigmas associated with manuallabour-intensive work, could aggravate the initialdecline. As traditional societies transform into modernones and more skilled labour is demanded, the opportunitycost of staying at home increases and women move backinto the labour force, which is captured by the movealong the rising portion of the U-shaped relationship.7
the case, we would have more equal societies which, given the Kuznetscurve, would occur among hunter–gatherers or advanced economies.5 For the purpose of this study female labour is defined in terms of
the total contribution of women to their household. This includeshousehold work and market work. Household work leads to the pro-duction of household goods, while market work implies either the pro-duction of food in subsistence economies or earning an income inmodern ones. Goldin (1994) made a clear distinction between marriedand unmarried women. We do not make such distinction and refer toall adult women.6 As Goldin (1994) explains income may rise following an expansion
of the market (Internal) or the introduction of a new technology (External).7 The process suggests an initially strong income effect combined with
a small own-substitution effect. At some point the substitution effect in-creases while the income effect declines. During the falling portion of theU the income effect dominates, but during the rising portion of the U thesubstitution effect dominates (Goldin, 1994).
ARG
AUSAUT
BDI
BEL
BENBFA
BGD
BOL
BRA
BWACAF
CAN
CHE
CHL
CHN
CIV
CMR
COL
CRI
DEU
DNK
DOM
DZAECU
EGYESP
FIN
FRAGBR
GHA
GNB
GRC
GTM
HKG
HND
HTI
HUN
IDNIND
IRL
IRN
ISR
ITA
JAM
JOR
JPN
KEN
KOR
LKA
LSO
MAR
MDG
MEX
MLI
MOZ
MRT
MWI
MYS
NER
NGA
NIC
NLD
NOR
NZL
PAK
PAN
PER
PHL
PNG
PRT
PRY
ROU
RWA
SEN
SGP
SLV
SWE
THA
TTO
TUN
TUR
URY
USA
VEN
ZAF
ZMB ZWE20
4060
8010
0
Wom
en e
cono
mic
ally
act
ive
per
100
men
, 198
0 (W
EA
)
6 7 8 9 10
Log of GDP per capita, 1980
Goldin curve
Fig. 1. WEA and Economic development, 1980. Source: PRED Bank, Version 3.0; Heston, Summers and Aten (2002). Notes: WEA stands for the ratioof women economically active for every 100 men. GDP per capita stands for real gross domestic product ($ current prices). Fig. 1: WEA=749.4679−174.6234∗Log(GDP per capita)+10.831∗Log (GDP per capita)2, N=88, R2=0.3533.
9 For a good overview of the reasons and causes of polygyny seealso Bergstrom (1994), Clignet (1970), Ember (1974, 1984), Gould etal. (2004), Kanazawa and Still (1999), Osmond (1965), Tertilt (2006),White (1988), and White and Burton (1988).10 In this paper, we distinguish between African and Harem polygynyaccording to the prevalence or incidence of polygynous unions within a so-ciety. African polygyny refers to a society with many polygynous unions(many men and many women), whereas Harem polygyny implies fewsuch unions(few men and few/many women). Harem polygyny can alsobe referred to as monogamy with occasional polygyny. Harem polygynycan be observed in some Arab countries (Chamie, 1986). Betzig (1992a,
314 A. Díez Minguela / History of the Family 16 (2011) 312–330
Fig. 1 illustrates the Goldin curve in 1980. We usegender differences in labour participation, as opposed tolabour force participation rates for women 45–59 yearsold.8 Gender differences in labour participation havebeen captured with the ratio of women economicallyactive for every 100 men (WEA). Economic developmentis proxied by the log of gross domestic product per capita.
The Goldin curve turns out to be a convenientinstrument to explore the evolution of mating patterns inthe very long run. Countries that appear in the upper leftside of Fig. 1 are mostly sub-Saharan African countrieswhere polygynous unions remain prevalent. Boserup(1970) observed that women did most of the agriculturalwork in sparsely populated regions, where extensiveagriculture predominated, such as sub-Saharan Africa.Grossbard (1976) noticed a positive relationship betweenfemale contribution to subsistence and polygyny. White(1988) claims that polygyny is more prevalent in societiesin which women have more to offer to the householdeconomy. Finally, Jacoby (1995), using a large householdsurvey from Côte d'Ivoire, found that conditional on
8 Goldin (1994) used the labour force participation rates for women 45to 59 years old and the log of GDP/capita, both for 1985. Moreover, thecentrally planned economies and countries of theMiddle East (except Is-rael) were excluded. Also, Goldin (1994) omitted twenty countries forwhich the labour force data differed widely from those given for employ-ment status. The regression line is quadratic in the log of GDP per capita.
wealth, men have more wives when female labour con-tributes a larger share to household income.9
Overall, female contribution to subsistence seems to bea strong predictor of African polygyny.10 Therefore, weexpect that African polygyny arises if and only if co-wives (married women) contribute to the wealth of thehousehold. In other words, it seems that female labour isa necessary yet not a sufficient condition for Africanpolygyny.11
1992b, 1995) provides further discussion regarding Harem polygyny inthe past.11 Table A.1 in the appendix illustrates the prevalence of African andHarem polygyny. While the percentage of married men in polygynousunions exceeds 15–20% in some sub-Saharan African countries, theincidence of polygyny in Pakistan or Nepal is very low. Bangladeshwith a reported 11.3% of married men in polygynous unions providesthe highest figure in Eurasia during the twentieth century (UnitedNations, 1990).
315A. Díez Minguela / History of the Family 16 (2011) 312–330
Fig. 2 illustrates female contribution to subsistenceand intensity of agriculture.12 We have computed theaverage Female Contribution by the type of agriculture,which is described in the Ethnographic Atlas, variable28: Intensity of Agriculture,13 Fig. 2 shows that, onaverage, female contribution to subsistence is higherwhen horticulture and extensive or shifting agriculturepredominates. Agricultural intensification would implylower female contribution to subsistence as Boserup(1970) noticed. The Ethnographic Atlas also illustratesthat horticulture and extensive agriculture were (are)predominant in Melanesia, Micronesia, Polynesia andsub-Saharan African societies.14 On the contrary,intensive agriculture has been widely observed in mostEurasian societies since antiquity. In this way, Fig. 2relates economic development (agricultural intensifica-tion) to female labour, which in turn can be associatedwith mating patterns.
When agriculture is casual or incidental, female con-tribution to subsistence seems to be moderatelylow. The economic conditions associated with hunter–gatherers and pastoralists–who are mostly nomadic orsemi-nomadic-, will be crucial in understanding theevolution of mating patterns. Among hunter–gatherers, there is complete specialisation within thehousehold–men hunt, while women gather-, and thusfemale contribution to subsistence is driven by therelative importance of hunting/fishing with respect to
12 Female contribution to subsistence (Female Contribution) reflects thepercentage total contribution of women to subsistence. To computeFemaleContribution from the Ethnographic Atlas, we use the followingmethodol-ogy. First, values on economic dependence by activity were taken fromvariables 1–5. There are five main subsistence activities described for eachsociety: gathering, hunting, fishing, animal husbandry and agriculture.These values range from 0 to 9. Table A.2 in the appendix shows the cod-ification. Variables 50–54 describe gender specific differences or sex dif-ferences in participation for each activity. To obtain a measure of femalecontribution, we first assign percentages to the codified values of variables50–54. Table A.3 in the appendix introduces these values. For example,code 1 implies that “Males only, or almost alone” participate in a particularactivity, i.e. hunting. Therefore, female participation in hunting activitieswill be 0%. Once we have converted codes into percentages, we computeda weighted average with the values of economic dependence by activity.The variable will be denoted as Female Contribution, and is a percentage.13 Table A.4 in Appendix shows the codification (variable 28:Intensity of Agriculture).14 More than 70% of sedentary societies in sub-Saharan Africa wereclassified as having extensive or shifting agriculture. In Melanesia,Polynesia and Micronesia more than ninety percent were describedas horticultural. While in Eurasia two thirds of societies were de-scribed as having intensive agriculture. Most societies in the NewWorld were nomadic or semi-nomadic. The Ethnographic Atlas refersmostly to indigenous ethnic groups observed during the nineteenthand twentieth centuries.
gathering. This, in turn, depends on the availability ofwild food, and factors such as geography and climate.
Table 1 illustrates how Female Contribution varieswith latitude among foraging societies. Although the vari-ation in the economic roles of men and women amonghunter–gatherers remains questionable, it seems thatamong foragers African polygyny was infrequent andpolygynous unions conventionally involve co-wives shar-ing habitation, mostly sororal-polygyny.15 White (1988)observed that in nomadic societies, sisters assist one anoth-er in household activities, i.e. packing, unpacking, cookingor gathering food.
The long run trajectory of female labour remainsunclear for two reasons. First, the worldwide variationamong hunter–gatherers, mostly related to their mainsubsistence activity, which in turn, is closely associatedto the availability of wild plants and game, and therebylatitude.16 Second, differences in the way agriculturehas been practised across societies (Boserup, 1970)have clouded the trajectory of female labour. Shiftingagriculture has conventionally been related to the“hoe”, whereas intensive agriculture was driven by the“plough” that requires significant animal and humanstrength. Alesina, Giuliano, and Nunn (2011) foundthat the descendants of pre-industrial societies thatwere dominated by the “plough” have currently lowerrates of female labour participation. Fig. 3 illustrates ahypothetical ‘augmented’ Goldin, which in turn pro-vides a long-term overview of the economic role playedby women over time.17
This leaves us two fundamental research questions:(1) Why monogamy? (2) Why the ‘European’ marriagepattern? The transition from a nomadic lifestyle to asedentary one could have an impact on the organisationof the household. Cultural traits or habits can betransmitted vertically. However, we argued above that
15 Murdock and Provost (1973) provide a good review of the divisionof labour by sex across societies. Out of 844 societies in the Ethno-graphic Atlas with available data, there are only 14 cases in whichwomen contribute to hunting activities. Regarding fishing, in 79.2%of societies males fish alone or appreciably more than females. Onthe other hand, gathering activities are predominantly dominated byfemales; in 80.2% of societies females gather alone or appreciably mo-re than males. Among pastoralists, women seem to contribute on aver-age less than men. In this case, the size of domestic animals may alsohave an impact on the division of labour (Murdock & Provost, 1973)See also Schlegel and Barry (1986).16 For a good survey of hunter–gatherers see Lee and Daly (1999).17 Female labour is defined in terms of the total contribution of wom-en to their household. This includes household work and market work.Household work leads to the production of household goods, whilemarket work implies the production of food in subsistence economiesor earning an income. Traditionally, women have been in charge ofhousehold tasks such as food preparation and children/elderly care.
3035
4045
Fem
ale
Con
trib
utio
n to
Sub
sist
ence
(%
)
No Agriculture Casual Extensive Horticulture Intensive Intensive*
Intensity of Agriculture
Fig. 2. Female contribution to subsistence and intensity of agriculture. Source: Ethnographic Atlas (Obs.=616 societies). Notes: (i) No Agriculture;(ii) Casual: Casual agriculture, incidental to other subsistence modes; (iii) Extensive: Extensive or shifting agriculture, long fallow, and new fieldscleared annually; (iv) Horticulture: Horticulture, vegetal gardens or groves of fruit trees; (v) Intensive: Intensive agriculture, using fertilisation, croprotation, or other techniques to shorten or eliminate fallow period; (vi) Intensive*: Intensive irrigated agriculture (World Cultures, 1999).
Table 1Female contribution to subsistence by latitude among hunter–gatherers.Ethnographic Atlas
Latitude 0–25° 25–50° N50°
316 A. Díez Minguela / History of the Family 16 (2011) 312–330
African polygyny was rare among foraging societies.Then, why did African polygyny arise and persist?African polygyny could have existed in Eurasia too.18
Unfortunately, the available evidence is scant and fur-ther research will be required to offer a more rigorousanalysis. Herodotus pointed out that among the Thra-cians, who inhabited Eastern and Southern Europe, itwas customary for a man to have more than onewife.19 Tacitus described the peoples of Germaniain the first century AD as “…almost unique amongbarbarians in being content with one wife…”20 Thisdescription suggests that Germanic barbarians weremonogamous, whereas non-Germanic barbarianswere essentially polygynous, as the Thracians de-scribed by Herodotus.
Theoretically, we expect that small changes in con-sumption at low levels of economic development willbe accompanied by sizeable gains/losses in utility. Forany woman who has to decide whether to enter anestablished household (polygyny) or form a new one(monogamy), the potential level of consumption withinthe household will play a critical role. If her contribu-tion is restricted to the production of household goods(food preparation, child care) then it could be
18 The transition from hunting–gathering to permanent settlementand agriculture is commonly referred as the Neolithic revolution.19 Herodotus (Book V: 5).20 Tacitus (Germania: 18).
economically unsustainable for a married man to takea second wife. If, on the other hand, her contributionto the household wealth increases (food production)then a married man would be less discouraged from tak-ing a second wife. As a result, a woman may prefer apolygynous union when (i) the average level of con-sumption within a society is very low, near subsistence,and (ii) the potential level of consumption of an estab-lished household is larger than that of a newly formedone.
Boserup (1965) observed that when population den-sity increases in agricultural economies, and hence theamount of natural resources per head is reduced, exten-sive agriculture could no longer survive if new openfields were not available. This could have partlyexplained why the demand for women's labour inagriculture sharply decreased since antiquity in some
23 Hajnal (1982: 452): “Single household system associated with…A) Late marriage for both sexes (mean ages at first marriage are,
Fig. 3. The ‘augmented’ Goldin curve.
317A. Díez Minguela / History of the Family 16 (2011) 312–330
parts of Eurasia.21 During the Sung period in China, theeconomic role of women was not closely associatedwith agricultural activities (Ebrey, 1993; pp. 132).22
“Yet whatever women contributed to generalagriculture, in the imagination of the Chinese scholarswomen's work largely lay elsewhere. Their work wasthe slow and tedious production of textiles, one largelywithin the confines of the home. Symbolically womenwere associated with cloth, and since ancient timesthe sexual division of labour had been epitomized bythe saying that men plough and women weave.”
To what extent the changing economic role of womenwithin a society and throughout history have shaped mat-ing patterns and household formation remains unclear.However, it seems that African polygyny is the result ofa very unique economic environment in which womenplay a significant economic role. This is in line with thetheory that explains human monogamy as a result ofincreasing paternal investment, given our long period ofoffspring dependence (Marlowe, 2000).
21 Moreover, large populations could better afford investments in ba-sic infrastructure. This in turn would have improved the standards ofliving (Boserup, 1981).22 Nevertheless the production of textiles involved costs that not allfamilies would have been able to afford, “Not all women in peasantfamilies would have worked at making cloth. Differences in climateand soil made some areas unsuited for the production of cloth, or sosuited to something else like tea that the family specialized in that ac-tivity, purchasing whatever cloth they needed. The very poorest fami-lies might not be able to secure the land and equipment needed toproduce cloth…” Ebrey (1993: 133).
In Eurasia, monogamy has prevailed as the ‘preferred’mating system but as Goody (1983) pointed out: howwas it that after around AD 300 certain general featuresof European patterns of marriage came to take a differentshape from those of ancient Rome, Greece, Israel andEgypt, and from those of the societies of theMediterraneanshores of the Middle East and North Africa that succeededthem? Hajnal (1965, 1982) drew a broad historicaldistinction between the “Northwest European” pattern orsimple household family system and; the “Non-European”,“Eastern” pattern or joint household family system associ-ated mainly with India and China. Hajnal (1982) observed,amongst other things, that men and womenmarried later insome Northwest European countries during the seven-teenth and eighteenth centuries, whereas in India andChina men married relatively earlier and women ratherearlier than in Northwest Europe.23
The classical theory of marriage would argue thatmarriage takes place if, and only if, the gains from it
say over 26 for men and over 23 for women), B) After marriage a cou-ple are in charge of their household (the husband is head of house-hold), C) Before marriage young people often circulate betweenhouseholds as servants.” …“Joint household systems… A) Earliermarriage for men and rather early marriage for women (mean agesat first marriage are under about 26 for men and under 21 for wom-en), B) A young married couple often start life together either in ahousehold of which an older couple is and remains in charge or ina household of which an unmarried older person (such as a widoweror a widow) continues to be the head. Usually the young wife joins herhusband in the household of which he is a member, C) Householdswith several married couples may split to form two or more house-holds, each containing one or more couples”.
27 Goldin (2006) divides the rising portion of the U-shaped relation-ship into four stages: (i) Late Nineteenth century to 1920s; (ii) Transi-tion, 1930–1950; (iii) ‘Roots of the Revolution’ 1950–1970s; (iv) TheQuiet Revolution, 1970–today. This is also a useful timeline when
318 A. Díez Minguela / History of the Family 16 (2011) 312–330
outbalance the costs (Becker, 1973, 1974, 1991). Thebenefits from marriage will be greater when differencesbetween men and women are more acute. Moreover, un-certainty about the traits of the potential spouse and howthe gains from marriage are distributed will induce menand women to spend time and other resources searchingfor the most appropriate partner. As a result, the age atmarriage may depend both on the expected gains fromthe union and on the costs that are involved in finding asuitable spouse. The female age at first marriage was rel-atively high in Northwest Europe.24 De Moor and VanZanden (2010) argued that the EuropeanMarriage Pattern(EMP) arose as a result of the preaching of the CatholicChurch promoting marriage based on consensus and‘girl power’ achieved through labour markets. This is inline with our main discussion, and could be reflected intheGoldin curve too. The ‘girl power’ hypothesis impliesthat a group of countries, mostly Northwest European,were pulling the Goldin curve up.
Figure A.1 in the appendix shows the Goldin curvefor a balanced selection of countries, for the years 1950and 2000. This figure illustrates how the Goldin curvehas pivoted upwards around the less developed countries.Generally, gender differences in labour participation havebeen gradually reduced. This ‘girl power’ has been partic-ularly acute in developed economies, which have also ex-perienced a rapid increase in the female age at firstmarriage. However, what happened before 1950? FigureA.2 in the appendix illustrates the long-term patterns(1850–2000) of gender differences in labour participation(WEA) or ‘girl power’ for a selection of countries.25
Among the Northern and Western European countries,we observe on average a WEA value around fifty womeneconomically active for every hundredmen in the late nine-teenth century and early twentieth century. Southern Euro-pean and New World countries exhibited lower levels.26
These levels, in turn, are rather similar to those observedfor developed countries in Figure A.1 in 1950. FigureA.2 offers empirical evidence on the varying degrees of‘girl power’ across countries. Whether ‘girl power’ partlyexplains the ‘European’ marriage pattern as De Moor andVan Zanden (2010) suggest deserves further discussionand research. Nonetheless, this seems like a good bet.
24 See Wrigley, Davies, Oeppen, and Schofield (1997) for a long-term overview of marriage patterns in England, 1580–1837.25 All of these countries are currently developed economies where the av-erage man and woman postpone marriage until their early thirties. We alsoinclude the sex ratios, given the total male and female populations.26 The only exception was Italy, which presents a comparable patternto Northwest European countries. The unification of Italy occurred in1861 and data were first available in 1871 (Bairoch, Deldycke,Gelders, & Limbor, 1968).
Furthermore, as traditional societies grew into modernones and skilled labour became highly demanded, the op-portunity cost of staying at home increased and womenmay have moved back into the labour force (Goldin,1994).27 In addition, social stigmas could be reducedonce manual labour-intensive work gave way to moreskilled work. Empirically, female labour appears to reacha historical low with the Industrial Revolution. Pinchbeck(1930) argued that during the Industrial Revolution femaledependence on male wages increased. Horrell andHumphries (1995) support this view and explain that thedecline in participation was caused not only by supplyshifts. In this regard, supply factors such as the introduc-tion of contraceptive methods, Goldin and Katz (2002);household technologies, Greenwood, Seshadri, andYorukoglu (2005); cultural beliefs about the long runpayoff to working, Fernandez (2007); have accompanieddemand factors such as the rise of the clerical sector, therate of unemployment or social stigmas associated withemployers. Notwithstanding the decline caused by the ad-vent of the Industrial Revolution, the varying degrees of‘girl power’ remain unaffected within Europe in 1950.
3. The Kuznets curve
Kuznets (1955) observed that along the process of eco-nomic development, the income distribution firstbecame more unequal, and then more equal, generatingan inverted U-shaped relationship or Kuznets curve.28
The hypothesis of an inverted U-shaped relationshipbetween income inequality and economic developmenthas been well documented. Barro (2000) argued that,although the Kuznets curve does not explain the size ofdifferences of inequality across countries and over time,it emerges as a clear empirical regularity. Banerjee andDuflo (2003) using non-parametric methods showthat the growth rate is an inverted U-shaped functionof net changes in inequality. In the first stage of
looking at marriage patterns.28 Robinson (1976) further developed the original hypothesis some yearslater. Wealth inequality can be composed of labour income (wage) and/ornon-labour income.We assume that in traditional societies wealth inequal-ity is mainly caused by non-labour income. Milanovic, Lindert, and Wil-liamson (2007) found that inequality in pre-industrial times was largelydriven by the differences between the rural poor and the landed elite.Piketty (2003) concludes that in France and possibly among other countriesin the twentieth century, wage inequality has actually been extremely stablein the long run, and the decline in income inequality is for the most part acapital income phenomenon.
319A. Díez Minguela / History of the Family 16 (2011) 312–330
economic development (hunting–gathering) wealthinequality may not be substantial. This is because hunter–gatherers are not able to accumulate and transmit wealthacross generations. Similarly, if ownership and propertyrights are not well established, we would expect thatwealth inequality remains relatively low.
As a result, agricultural intensification could also have asignificant impact on land tenure systems. Property rights,or at least the need of a more restrictive system of land ten-ure, emerge when land becomes scarce. The establishmentof a system of property rights affect ownership and the dis-tribution of land within a society, among other movable orimmovable assets. It also makes possible the accumulationand transmission of wealth across generations. Under thesecircumstances, households may be able to implicitly buywomen's work. Fig. 4 illustrates the Kuznets curve in1980.29 We use the Gini index to capture incomeinequality.30 Economic development is proxied by thelog of gross domestic product per capita.
The accumulation and transmission of wealth could, inturn, change the existing customs or mating (marriage)patterns. Married women may retreat into the home anddevote their whole effort to the production of householdgoods. Nevertheless, it also occurs that in traditional soci-eties, where the demand for skilled labour is relativelylow, manual labour-intensive work may be regarded asinappropriate for women, and hence social stigmascould arise.31 As a result, women as daughters or wivesinvoluntarily retreat into the home. Therefore economicdevelopment via increased population density will havetwo main effects on the economy: (i) a decrease on thedemand for agricultural work as we previously exploredin Section 2, and (ii) the development of institutionsregarding ownership and the distribution of land andcapital, which in turn, will allow households to accumu-late and transmit wealth across generations.
When wealth inequality remains low within an econo-my, polygynous unions may arise if a woman prefers toenter an established household rather than to form a newone. As we discussed above, polygyny may occur if anestablished household with a married couple and their
29 We estimate the Goldin and Kuznets curves for the same sample ofcountries. Our objective will be to compare both results.30 We are aware that the Gini coefficient or index does not fully cap-ture income inequality, in particular within less developed economicsi.e. sub-Saharan Africa. Table A.5 in the appendix also show the Giniindex of land concentration that is in line with Fig. 4. Given the avail-able and existing data on income inequality, the Gini index still re-mains a reasonable proxy.31 These social stigmas could arise if parents, husbands or more sig-nificantly potential employers regard women as inappropriate for thespecific work. In this regard, social stigmas may be strongly and pos-itively associated with income levels.
respective children offer a higher level of consumptionthan a newly formed household. When, on the otherhand, inequality increases, African polygyny could breakdown. This, in fact, is a relevant feature for the understand-ing of mating patterns over time. The distribution ofwealth will depend upon class rather than age, and thosemen and women belonging to wealthy households willbe targeted as ‘optimal’ marriage prospects. Under thesecircumstances Harem polygyny could arise.32
When wealth inequality is high, some women wouldhave incentives to become the wife, concubine or mis-tress of a very wealthy man. This may increase the pres-sure on poorer men to provide their wives with enoughwealth to offset those incentives. To do so, some menwould delay or postpone marriage to accumulate the op-timal level of desired wealth. In sub-Saharan Africa, tra-ditional land tenure systems have prevailed in manyregions.33 A common feature of these systems, popular-ly known as “communal land”, is that access to land isassured to most households (Place & Hazell, 1993).Therefore, if land is accessible, income inequality be-tween household is driven either by the size of thehousehold or the age of the household members. Menwould like to form a large households, but not at anycost. In this way, access to resources and female labourappear to be essential in understanding the incidenceand prevalence of polygyny.
Would women wait or marry early thus creating a largespousal age gap? As Hajnal (1965; pp. 134) pointed out,“The uniqueness of the European pattern liesprimarily in the high age at marriage of women (oftenwith a relatively small difference between the age ofhusband and wife), rather than in a high age at marriagefor men”. Ultimately, women would also have a say, butwaiting is costly, and for some poor households untenable.Therefore, the number of households capable ofwaiting will be partly determined by the level and thedistribution of income within a society.
Following this line of thinking, suppose that Harempolygyny occurs at the early stages of economic develop-ment. As standards of living increased, the incentives anywoman has to become a second wife or mistress arereduced.34 Then, Harem polygyny can only be sustained
32 Previously, we discussed two possible equilibriums, African po-lygyny and monogamy. In this section we introduce a third equilibri-um Harem polygyny that involves few wealthy men mating ormarrying with many women.33 Boserup (1970) argued that land tenure systems are one of themain factors explaining the contrasting mating systems of Eurasiaand Africa.34 We assume that the farther away women are from the minimal sub-sistence income, the smaller will be the potential gain provided by anextra unit of consumption.
ARG
AUS
AUT
BDI
BEL
BEN
BFA
BGD
BOL
BRA
BWACAF
CANCHE
CHL
CHN
CIVCMR
COL
CRI
DEU
DNK
DOM
DZA
ECU
EGY ESP
FIN
FRA
GBR
GHA
GNB
GRC
GTM
HKG
HND
HTI
HUN
IDN
IND
IRL
IRN
ISR
ITA
JAM
JOR
JPN
KEN
KOR
LKA
LSO
MAR
MDGMEX
MLI
MOZ
MRTMWI
MYSNER
NGA NIC
NLD
NOR
NZL
PAK
PAN
PER
PHL
PNG
PRT
PRY
ROU
RWA
SENSGP
SLV
SWE
THA
TTOTUN
TURURY
USA
VEN
ZAF
ZMBZWE
.2.3
.4.5
.6
Gin
i Ind
ex
6 7 8 9 10
Log GDP per capita, 1980
Kuznets curve
Fig. 4. Gini index and economic development, 1980. Source: Heston et al. (2002); Deininger and Squire (1996). Notes: GDP per capita stands for realgross domestic product ($ current prices). Gini index: a value of 0 represents absolute equality, and a value of 1 absolute income inequality; Countries arebalanced for both figures. Fig. 3: Gini Index=−2.04520+0.6874∗Log(GDP per capita)−0.0467∗Log (GDP per capita)2, N=88, R2=0.3508.
320 A. Díez Minguela / History of the Family 16 (2011) 312–330
when wealth inequality within a community growsconsiderably.35 Along theKuznets curve,Harem polygynywould progressively turn into monogamy withoccasional polygyny, and then monogamy will prevailwithin modern societies. Milanovic et al. (2007) providesound evidence of the unequal but richer England (1688;1801) and Holland (1571; 1732) as opposed to the unequalbut poorer Old Castile (1752), Kingdom of Naples (1811),India (1750; 1947), China (1880), Byzantine and Romanempires.36 Therefore, we should expect more Harem po-lygyny and less waiting in the latter regions.
Nevertheless, whywaiting? In extremely unequal econ-omies where few wealthy households enjoy most of thewealth, poorer households may find difficulties in waiting.When resources can be accumulated and transmitted acrossgenerations family workmay be determined by the amountof resources within your own household. Those house-holds endowed with more resources will be able to bearthe costs associated with waiting. As a result, the
35 Wealthy men—who might also be powerful political figures—would have incentives to compromise to monogamous marriage in or-der to avoid social conflict. This is also called Socially Imposed Mo-nogamy, (Alexander, 1987; Alexander, Hoogland, Howard, Noonan,& Sherman, 1979; and Betzig, 1986).36 Milanovic et al. (2007) introduce the concept of the InequalityPossibility Frontier (IPF) in which the maximum attainable inequalityis an increasing function of mean income. Bearing this in mind, Milanovicet al. (2007) observe that while Gini coefficients are large for suchstratified societies as England and Holland, both countries remainedfurther away from the IPF than others.
distribution of resources within a society can determinemarriage patterns. Households will be able to implicitlybuy women's work by employing them.37 In this regard,waiting should be less likely in unequal societies inwhich households are close to the subsistence frontier,i.e. Old Castile (1752), Kingdom of Naples (1811).
4. The Big Picture
In general, the Goldin and Kuznets curves are twodistinctive empirical regularities that provide us with acomprehensive picture of the evolution of mating (mar-riage) patterns in the very long run. In the beginning, hom-inids were very similar to apes. Most apes are polygynous,but somehow conditions changed radically and a monoga-mous mating structure emerged (Lovejoy, 1981). Theexisting ethnographic evidence also supports the viewthat foragers are predominantly monogamous. Then, theNeolithic revolution transformed the world economy, andagriculture became the main economic activity. However,convergence in mating patterns seemed far from probable.Monogamous unions could have already been the norm,but African polygyny has also been widely
37 Our views attempt to supplement Galor and Weil (1996, 2000)who argued that the gender gap or relative wages have an effect on fer-tility decisions, population growth, capital intensity and economicgrowth. Nevertheless, there is little evidence to support the hypothesisthat high age at marriage implied lower fertility rates in the eighteenthand nineteenth centuries.
321A. Díez Minguela / History of the Family 16 (2011) 312–330
observed.WhetherAfrican polygyny arose recently or notremains questionable. Some authors argue that slavery en-couraged polygyny, (Dalton & Cheuk Leung, 2011;Manning, 1990). Sadly, the Ethnographic Atlas only pre-sents evidence for the post-slave trade period.38 Never-theless, some polygynous societies in sub-SaharanAfrica were not affected by the slave trade.
In any case, African polygyny prevails in less devel-oped economies characterised by the significant economicrole played by women, and the absence of property rightsand hence more egalitarian. Theoretically, we have dis-cussed that along the process of economic development Af-rican polygyny would break down. In this regard,women's empowerment or the so-called “girl power” hy-pothesis partly explains the second major pattern: the ‘Eu-ropean’marriage pattern. Among “plough” societies wherewomen traditionally retreated into the home, somehowwomen were empowered and marriage postponed (DeMoor & Van Zanden, 2010). Fig. 5 illustrates the com-bined Kuznets and the Goldin curves. On the whole, itseems that “girl power” or an increasing economic roleplayed by women within a society can be closely associat-ed with deviations from the conventional “early attach-ment to one woman” (Malthus, 1798).
Finally, Figures A.3, A.4 and A.5 in the appendixillustrate the Goldin and Kuznets curves for a balanced se-lection of countries in 1980. In addition, we also showwage employment as a percentage of the totallabour force. Figure A.3 presents all available countries,while Figure A.4 depicts those countries where African po-lygyny is prevalent. Figure A.5 shows the Northwest Euro-pean countries. These figures provide an illustration of therelationship between economic development and mating(marriage) patterns. First, we observe that polygynous mat-ing remains significant at very low levels of economic de-velopment. Secondly, Northwest European countries arepulling up (down) theGoldin (Kuznets) curve. Bothmating(marriage) patterns are unique and distinctive, a deviationfrom the norm but with a common feature: the economicrole of women.
5. Conclusion
In this paper we have introduced and briefly discussedtwo important long-term relationships. First we looked at
38 Slavery was first abolished in British colonies in 1807. “The slavetrade was declared illegal in Venezuela and Mexico in 1810, in Chilein 1811, and in Argentina in 1812. In 1817 Spain signed a treaty withBritain agreeing to abolish the slave trade in 1820, but the trade con-tinued to the remaining Spanish colonies until 1880. Chile freed itsblack slaves in 1823; Mexico abolished slavery in 1829, and Peru in1854” (Encyclopaedia Britannica, 2008).
the Goldin curve that relates female labour participationand economic development. Then, we explored therelationship between income inequality and economicdevelopment, or Kuznets curve.
Overall, the economic role of women and incomeinequality, have followed opposite paths. The changingeconomic conditions over time may reflect mating(marriage) patterns, and hence household formation.The transition from the nomadic settlement pattern toa sedentary one brought changes in the formation andorganisation of the household. Sedentary societieswere accompanied by institutional change, i.e. estab-lishment of property rights. As income rose, increasingincome inequality was accompanied by a diminishingeconomic role of women within a society. In these con-ditions, African polygyny could not exist, and Harempolygyny and ultimately monogamy predominated.With a higher level of income, families may be able toimplicitly buy women's work. With the establishmentof property rights, employers could be disinclined tohire women and prefer men instead for tasks that are es-sentially physical.
The second major change regarding household forma-tion is the ‘European’ marriage pattern. Late marriage, es-pecially for women, was unprecedented in the latemediaeval period. Nevertheless, the changing economicconditions after the Black Death in Northwest Europecould have partly induced marital decisions. Conventional-ly, womenwere expected tomarry youngwhile menmarryearly or late depending upon their own economic circum-stances. A high age at first marriage for women raises thefollowing question: who bears the cots associated withwaiting? Whether unmarried women work or not, and thedistance between the natal household income and subsis-tence income seems crucial, and could determine the extentto which women wait. Further increases in the standard ofliving within a society, falling income inequality and morelabour opportunities for women, would in turn increase theage at marriage.
This paper looks mainly at the main divergent mat-ing (marriage) patterns across regions. It provides a uni-fied and linear narrative that encapsulates the worldeconomy. This paper is not in itself a strong resultfrom which a general theory can be built, however itis an encouragement and motivation to further analysethe economic conditions and structural changes thatshape human behaviour. If monogamy, at last, becomesthe norm, should we expect convergence in marriagepatterns across countries in the near future? If polygynypersists, what explains this persistence? Perhaps, we aremore alike than we think. Perhaps, our economic cir-cumstances are more distinct than we think.
Fig. 5. The ‘augmented’ Goldin curve and the Kuznets curve.
322 A. Díez Minguela / History of the Family 16 (2011) 312–330
Appendix A
Table A.1Percentage of married men/women in polygynous union by residence.
Source: Demographic and Health Surveys (1985–2005).Notes: Countries selected were those with available data for men and women.Surveys presented are the latest available.
Table A.2Economic dependence by activity (variables 1–5).
Source: Ethnographic Atlas.Notes: There are five activities described in the Ethno-graphic Atlas: (i) variable 1: Gathering, (ii) variable2: Hunting, (iii) variable 3: Fishing, (iv) variable 4: An-imal Husbandry and (v) variable 5: Agriculture.
Table A.3Sex differences (variables 50–54).
Code Description Percentage
0 Missing data –1 Males only or almost alone 0%2 Males appreciably more 25%3 Differentiated but equal participation 50%4 Equal participation, no marked differentiation 50%5 Females appreciably more 75%6 Females only or almost alone 100%7 Irrelevance of gender, especially industrialised production –8 Activity present:: Sex participation unspecified –9 Absent or unimportant activity –
Source: Ethnographic Atlas.Notes: There are five activities described in the Ethnographic Atlas: (i) variable 50: Gathering, (ii) variable 51: Hunting, (iii) variable 52: Fishing,(iv) variable 53: Animal Husbandry and (v) variable 54: Agriculture.
Table A.4Intensity of agriculture.
Code Description
0 Missing data1 No agriculture2 Casual agriculture, incidental to other subsistence modes3 Extensive or shifting agriculture, long fallow, and new fields cleared annually4 Horticulture, vegetal gardens or groves of fruit trees5 Intensive agriculture, using fertilisation, crop rotation, or other techniques to shorten or eliminate fallow period6 Intensive irrigated agriculture
Source: Ethnographic Atlas.
323A. Díez Minguela / History of the Family 16 (2011) 312–330
ARGAUS
AUT
BEL
BOL
BRA
CAN
CHE
COL
CRI
DNK
EGY
ESP
ETH
FIN
FRA
GBR
GTMHND
IND
IRL ISL
ISRITA
JPN
KEN
LKA
LUX
MAR
MEX
MUS
NGA
NIC
NLDNOR NZL
PAK PAN PER
PHL
PRT
SLV
THA
TTO
TUR
UGA
URY
USA
VEN
ZAF
1030
5070
90
Wom
en e
cono
mic
ally
act
ive
per
100
men
, 195
0 (W
EA
)
4 5 6 7 8Log of GDP per capita, 1950
ARG
AUS
AUTBEL
BOL
BRA
CANCHE
COL
CRI
DNK
EGY
ESP
ETH
FIN
FRAGBR
GTM
HND
INDIRL
ISL
ISR
ITA
JPN
KEN
LKALUX
MAR
MEX MUS
NGA NIC
NLD
NOR
NZL
PAK
PAN
PER
PHL
PRT
SLV
THA
TTO
TUR
UGA
URY
USA
VEN
ZAF
1030
5070
90
Wom
en e
cono
mic
ally
act
ive
per
100
men
, 200
0 (W
EA
)
7 8 9 10 11Log of GDP per capita, 2000
Fig. A.1. WEA and economic development, 1950 and 2000. Sources: PRED Bank, Version 3.0. and Penn World Tables 6.1. Notes: WEA stands forthe ratio of women economically active for every 100 men. GDP per capita stands for real gross domestic product ($ current prices) This figure il-lustrates a balanced selection of countries. Year, 1950: WEA=539.31−155.46∗Log(GDP per capita)+11.84∗Log (GDP per capita)2, N=50,R2=0.4018. Year, 2000: WEA=434.70−90.69∗Log(GDP per capita)+5.39∗Log (GDP per capita)2, N=50, R2=0.3469.
324 A. Díez Minguela / History of the Family 16 (2011) 312–330
050
100
050
100
1850 1900 1950 2000 1850 1900 1950 2000
Belgium France
Netherlands Switzerland
WEA Sex ratio
1850 1900 1950 2000 1850 1900 1950 2000
WEA Sex ratio
Western Europe
050
100
050
100
Denmark Great Britain
Norway Sweden
Northern Europe
Fig. A.2. WEA and sex ratios, 1850–2000. Sources: Bairoch et al. (1968), PRED Bank, Version 3.0.
325A. Díez Minguela / History of the Family 16 (2011) 312–330
Source: Author's; FAO (1995) Programme for the World Census of Agriculture 2000. Notes: Gini (or Lorenz) index of concentration is a measure ofconcentration of agricultural areas. It varies from 0, when all holdings have the same area, to 1, when all agricultural land in a country is in oneholding. Polygynous countries refer to those countries where the incidence of polygyny is above 20%.
326 A. Díez Minguela / History of the Family 16 (2011) 312–330
Fig. A.3. Goldin and Kuznets curves, 1980 (All countries). Fig. A.4. Goldin and Kuznets curves, 1980 (African polygyny).
327A. Díez Minguela / History of the Family 16 (2011) 312–330
Fig. A.5. Goldin and Kuznets curves, 1980 (‘European’ marriage pattern). Sources: PRED Bank, Version 3.0, Penn World Tables 6.1, United Nations(2000). Notes: WEA stands for the ratio of women economically active for every 100 men. GDP per capita stands for real gross domestic product($ current prices). Figures A.3, A.4 and A.5 illustrate a balanced selection of countries for 1980. Goldin curve: WEA=792.914−186.889∗Log(GDP per capita)+11.654∗Log (GDP per capita)2, N=76, R2=0.3895. Kuznets curve: Gini=−256.313+81.644∗Log(GDP per capita)−5.465∗Log (GDP per capita)2, N=76, R2=0.3805. Labour market: Percentage of employees=−133.887+23.491∗Log(GDP per capita),N=76, R2=0.7914. African polygyny: Benin, Botswana, Central African Republic, Côte d'Ivoire, Cameroon, Ghana, Guinea, Gambia, Lesotho,Mali, Mozambique, Malawi, Nigeria, United Republic of Tanzania, and Zambia. These are countries where polygynous unions are highly prev-alent (See Table A.1). Nigeria (Demographic and Health Survey, 2008), Botswana, the Gambia, and Lesotho also showed abnormal ratios of cur-rently married women to currently married men (United Nations Demographic Yearbook, 1978, 1987 and 2000). ‘European’ marriage pattern:Austria, Belgium, Denmark, Finland, France, Ireland, Netherlands, Norway, Poland, Sweden, Switzerland, and United Kingdom of Great Britainand Northern Ireland. These are countries within the Saint Petersburg-Trieste line (Hajnal, 1965).
328 A. Díez Minguela / History of the Family 16 (2011) 312–330
329A. Díez Minguela / History of the Family 16 (2011) 312–330
References
Alesina, A., Giuliano, P., & Nunn, S. (2011). On the origins of genderroles: Women and the plough. National Bureau of Economic Re-search Working Paper No. 17098.
Alexander, R. D. (1987). The biology of moral systems. Hawthorne,New York: Aldine de Gruyter.
Alexander, R. D., Hoogland, D. J., Howard, R., Noonan, K., & Sherman,P. (1979). Sexual dimorphisms and breedings systems in pinnipeds,ungulates, primates and humans. In N. Chagnon, & W. Irons (Eds.),Evolutionary biology and human social behaviour: An anthropolog-ical perspective (pp. 402–435). North Scituate, MA: Duxbury Press.
Bairoch, P., Deldycke, T., Gelders, H., & Limbor, J. M. (1968). Theworking population and its structure. Bruxelles: Institut de Socio-logie de l'Université Libre de Bruxelles.
Banerjee, A. V., & Duflo, E. (2003). Inequality and growth: What canthe data say? Journal of Economic Growth, 8(3), 267–299.
Barro, R. J. (2000). Inequality and growth in a panel of countries.Journal of Economic Growth, 5, 5–32.
Becker, G. S. (1973). A theory of marriage: Part I. Journal of PoliticalEconomy, 81, 813–846.
Becker, G. S. (1974). A theory of marriage: Part II. Journal of PoliticalEconomy, 82, S11–S26.
Becker, G. S. (1991). A treatise on the family. London: Harvard Uni-versity Press.
Bergstrom, T. (1994). On the economics of polygyny. Department ofEconomics, UCSB Available at: http://works.bepress.com/ted_bergstrom/75
Betzig, L. (1986). Despotism and differential reproduction: A Darwinianview of history. New York: Aldine Publishing Co.
Betzig, L. (1992a). Roman monogamy. Ethology and Sociobiology,13, 351–383.
Betzig, L. (1992b). Roman polygyny. Ethology and Sociobiology, 13,309–349.
Betzig, L. (1995). Medieval monogamy. Journal of Family History,20, 181–216.
Boserup, E. (1965). The conditions of agricultural growth: The economicsof agrarian change under population pressure. London: G. Allen andUnwin.
Boserup, E. (1970). Woman's role in economic development. London:Earthscan Publications Limited.
Boserup, E. (1981). Population and technological change: A study oflong-term trends. Chicago: The University of Chicago Press.
Chamie, J. (1986). Polygyny amongArabs.Population Studies, 40, 55–66.Clignet, R. (1970).Many wives, many powers: Authority and power in
polygynous families. : Northwestern University Press.Dalton, J., & Cheuk Leung, T. (2011). Why is polygyny more preva-
lent in Western Africa? An African Slave Trade Perspective Avail-able at SSRN. http://ssrn.com/abstract=1848183
De Moor, T., & Van Zanden, J. L. (2010). Girl power: The Europeanmarriage pattern and the labour markets in the North Sea in the latemedieval and early modern period. The Economic History Review,63(1), 1–33.
Deininger, K., & Squire, L. (1996). A new data set measuring incomeinequality. World Bank Economic Review, 10(3), 565–591.
Dorjahn, V. R. (1959). The factor of polygyny in African demography.Continuity and change in African cultures (pp. 87–112). Chicago:University of Chicago Press.
Ebrey, P. B. (1993). The inner quarters: Marriage and the livesof Chinese women in the Sung period. Berkeley: University ofCalifornia Press.
Ember, M. (1974). Warfare, sex ratio and polygyny. Ethnology, 13,197–206.
Ember, M. (1984). Alternative predictors of polygyny. Behavior Sci-ence Research, 19, 1–23.
Encyclopaedia Britannica (2008). 2008 Encyclopaedia Britannica,London.
FAO (1995). Programme for the World Census of Agriculture 2000.Rome: FAO Statistical Development Series No. 5, United Nations.
Fernandez, R. (2007). Culture as learning: The evolution of femalelabor force participation over a century. National Bureau of Eco-nomic Research. Working Paper 13373.
Galor, O., & Weil, D. N. (1996). The gender gap, fertility and growth.The American Economic Review, 86(3), 374–387.
Galor, O., & Weil, D. N. (2000). Population, technology, and growth:From Malthusian stagnation to the demographic transition and be-yond. The American Economic Review, 90(4), 806–828.
Goldin, C. (1994). The U-shaped female labor force function in eco-nomic development and economic history. National Bureau ofEconomic Research Working Paper No. 4707.
Goldin, C. (2006). The quiet revolution that transformed women's em-ployment, education and family. The American Economic Review,96(2), 1–21.
Goldin, C., & Katz, L. (2002). The power of the pill: Oral contracep-tives and women's career and marriage decisions. Journal of Polit-ical Economy, 110(4), 730–770.
Goody, J. (1983). The development of the family and marriage inEurope. Cambridge: Cambridge University Press.
Gould, E. D., Moav, O., & Simhon, A. (2004). The mystery of monog-amy. Hebrew University Economics Working Paper No. 2003.
Gould, E. D., & Paserman, M. D. (2003). Waiting for Mr Right: Ris-ing inequality and declining marriage rates. Journal of Urban Eco-nomics, 53, 257–281.
Greenwood, J., Seshadri, A., & Yorukoglu, M. (2005). Engines of lib-eration. The Review of Economic Studies, 72, 109–133.
Grossbard, A. S. (1976). An economic analysis of polygyny: The caseof Maiduguri. Current Anthropology, 17, 701–707.
Hajnal, J. (1965). European marriage patterns in perspective. In D. V.Glass, & D. E. C. Eversley (Eds.), Population in history. London:Arnold.
Hajnal, J. (1982). Two kinds of pre-industrial household formationsystem. Population and Development Review, 8, 449–494.
Heston, A., Summers, R., & Aten, B. (2002). Penn World Table Ver-sion 6.1, Center for International Comparisons at the Universityof Pennsylvania (CICUP), October.
Horrell, S., & Humphries, J. (1995). Women's labour force participa-tion and the transition to the male-breadwinner family, 1790–1865. The Economic History Review, 48(1), 89–117.
Jacoby, H. G. (1995). The economics of polygyny in Sub-Saharan Africa:Female productivity and the demand for wives in Côte d'Ivoire. Jour-nal of Political Economy, 103, 938–971.
Kanazawa, S., & Still, M. C. (1999). Why monogamy? Social Forces,78, 25–50.
Kuznets, S. (1955). Economic growth and income inequality. TheAmerican Economic Review, 45, 1–28.
Lee, R., & Daly, R. (Eds.). (1999). The Cambridge encyclopaedia ofhunters and gatherers. Cambridge: Cambridge University Press.
Loughran, D. S. (2002). The effect of male wage inequality on femaleage at first marriage. The Review of Economics and Statistics, 84,236–250.
Lovejoy, C. O. (1981). The origin of man. Science. New Series, 211(4480), 341–350.
330 A. Díez Minguela / History of the Family 16 (2011) 312–330
Malthus, T. R. (1798). An essay on the principle of population. Oxford:Oxford University Press.
Manning, P. (1990). Slavery and African life: Occidental, Oriental, andAfrican slaves trades. Cambridge: Cambridge University Press.
Marlowe, F. (2000). Paternal investment and the human mating sys-tem. Behavioural Processes, 51, 45–61.
Milanovic, B., Lindert, P. H., & Williamson, J. G. (2007). Measuringancient inequality. National Bureau of Economic Research Work-ing Paper 13550.
Murdock, G. P. (1967a). Ethnographic atlas: A summary. Ethnology,6, 109–236.
Murdock, G. P., & Provost, C. (1973). Factors in the division of laborby sex: A cross-cultural analysis. Ethnology, 12(2), 203–225.
North, D. C. (1991). Institutions. Journal of Economic Perspectives,5, 97–112.
Osmond, M. W. (1965). Toward monogamy: A cross-cultural study ofcorrelates of type of marriage. Social Forces, 44, 8–16.
Piketty, T. (2003). Income inequality in France, 1901–1998. Journalof Political Economy, 111(5), 1004–1042.
Pinchbeck, I. (1930). Women workers and the Industrial Revolution1750–1850. London: Virago.
Place, F., & Hazell, P. (1993). Productivity effects of indigenous landtenure systems in Sub-Saharan Africa. American Journal of Agri-cultural Economics, 75, 10–19.
Robinson, S. (1976). A note on the U-hypothesis relating income in-equality and economic development. The American Economic Re-view, 66(3), 437–440.
Schlegel, A., & Barry, H., III (1986). The cultural consequences of fe-male contribution to subsistence. American Anthropologist, NewSeries, 88, 142–150.
Tertilt, M. (2006). Polygyny, women's rights and development. Jour-nal of the European Economic Association, 4, 523–530.
United Nations (1990). Patterns of first marriage: Timing and prev-alence. New York: Department of Economic and Social Affairs.
United Nations (2000). World marriage patterns 2000. New York:Department of Economic and Social Affairs.
White, D. R. (1988). Rethinking polygyny: Co-wives, codes, and cul-tural systems. Current Anthropology, 29, 529–572.
White, D. R., & Burton, M. L. (1988). Causes of polygyny: Ecology,economy, kinship, and warfare. American Anthropologist, NewSeries, 90, 871–887.
World Cultures (1999). World Cultures CD Data Disk (EthnographicAtlas). Journal of Comparative and Cross-Cultural Research, 10(1).
Wrigley, E. A., Davies, R. S., Oeppen, J. E., & Schofield, R. S.(1997). English population history from family reconstitution1580–1837. Cambridge Studies in Population, Economy and So-ciety in Past Time.
Further Reading
Goody, J. (1976). Production and reproduction: A comparative studyof the domestic domain. New York: Cambridge University Press.
Goody, J. (1990). The Oriental, the ancient, and the primitive: Sys-tems of marriage and the family in pre-industrial societies in Eur-asia. Cambridge: Cambridge University Press.
Wooldridge, J. M. (2002). Econometric Analysis of Cross Section andPanel Data. MIT Press.
