Date post: | 03-Jun-2018 |
Category: |
Documents |
Upload: | flavio-falcioni |
View: | 223 times |
Download: | 0 times |
of 26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
1/26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
2/26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
3/26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
4/26
The observation hypothesis: Because children
closely observe the outcomes of their parents
choices, these outcomes will affect the childrens
choices. As a consequence, if children observe
their parents as self-employed, there is a higher
chance that they will choose to become self-
employed themselves.
While observation of the parents behaviour is of
central importance according to social-learning the-
ory, this influence of the parents behaviour is not
constant. It can vary with the extent to which children
interact or identify with their parents. Hence, it is
argued that children imitate same-sex models more
than opposite-sex models, because they recognize that
they share a larger set of common attributes with the
parent of the same sex (Hetherington 1965). Further-
more, children often spend more time in activitiessimilar to those of the role model of the same sex
(Slaby and Frey1975), and children might also learn
faster from parents of the same sex (Bonke and
Esping-Andersen2009). This leads us to formulate the
following two refinements of the observation
hypothesis:
The interaction hypothesis:The more the parents
interact with their children, the more the children
observe of (and learn from) the parents behaviour.
Thus, there is a stronger role-model effect of
observing parents as self-employed when the childis living and/or working with his or her parents.
The same-sex hypothesis:Boys identify with and
spend more time with their fathers, and girls
identify with and spend more time with their
mothers. Therefore, the father will be the most
important role model for boys, while the mother
will be most important for girls.
An empirical implication of the observation hypoth-
esis is that the parents experience from self-employ-
ment should positively affect the probability that thechild eventually becomes self-employed him- or
herself, and we should expect this probability to
increase with the amount of experience the parents
have from self-employment. Furthermore, the inter-
action hypothesis implies that the importance of the
parents self-employment experience should be
greater if it was obtained while the child was still
living at home or if the child has worked with the
parents while they were self-employed. Finally, an
empirical implication of the same-sex hypothesis is
that the experience of the father should have a larger
effect on sons, while the experience of the mother
should have a larger effect on daughters.
We test these empirical implications in the
following sections. However, some of these impli-
cations are not specific to the preference channel,but can also be expected under the two other
transmission channels. First, the financial channel
may also be consistent with an observed positive
effect of the parents self-employment experience
(the observation hypothesis)and in particular a
successful experience. Hence, a positive effect of
parental self-employment experience cannot imme-
diately be ascribed to the preference channel. In
order to eliminate the potential effects working
through the financial channel, we can control for
the wealth of the parents. Furthermore, we candisregard individuals who take over the family
business. In the empirical analysis below, we do
both.
Second, the human-capital channel obviously also
predicts a positive effect of having worked for the
parents while they were self-employed. Furthermore,
this is in turn likely to be correlated with the parents
experience from self-employment. Hence, it may be
difficult to disentangle the effects running through the
human-capital and preference channels. However, if
an effect of parental experience remains after control-ling for work experience in the family firm, this is a
strong indication of an effect working through the
preference channel. This is the approach that we
follow below.
Third, if parents make gender-specific investments
in their childrens human capital, in the sense that boys
tend to work in their fathers firms, while girls work in
their mothers firms, this may give rise to observed
gender-specific effects as predicted by the same-sex
hypothesis above but driven by the human-capital
channel; see also Lindquist et al. (2013). We try toeliminate this effect by controlling for the amount of
experience the children have from working with each
parent.
Finally, as we do not expect genetic factors to run
within gender lines, finding support for the same-sex
hypothesis would be a strong indication that the
preference channel (also) works through role models,
and it would furthermore point to a very specific way
A. Hoffmann et al.
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
5/26
in which the preferences are passed on to the offspring.2
In the existing literature, there are different bits of
evidence that point to a positive effect of parental role
models. Van Auken et al. (2006) study 318 American
and Mexican undergraduate students preferences for
self-employment. Using a questionnaire to ask directlyabout the importance of role models, they find that the
father is the most important role model in general and
that role models who own a business have a greater
influence on the respondents career intentions than
role models who do not own a business. Scherer et al.
(1989) and Matthews and Moser (1996) also use a
questionnaire approach and find that university stu-
dents with self-employed parents have stronger pref-
erences for becoming self-employed. While these
studies point to an effect of parental role models on
entrepreneurship intentions, they do not find theeffects (qualitatively or quantitatively) on actual
outcomes nor do they show how the preferences are
transmitted.
In a recent study of 292 Dutch entrepreneurs,
Bosma et al. (2012) find support for the importance of
role models more generally (not just parental role
models); 54 % of the entrepreneurs had role models
before or shortly after firm start-up, and these were
viewed by one-third to be important for the start-up
decision.
Srensen (2007) finds that self-employment ratesamong children whose parents were only self-
employed during the childs adolescence are almost
as high as those where the parents were self-employed
both in adolescence and adulthood. He takes this as
evidence of the importance of role models, rather than
children taking advantage of parents financial and
social capital. Chlosta et al. (2012) study 461 alumni
from German universities and find that both a self-
employed father and a self-employed mother affect thepropensity to become self-employed. They interpret
this as a parental role-model effect, but they do not
control for influences through the other potential
transmission mechanisms.
The only existing papers to distinguish between
different effects on sons and daughters of self-
employed parents are Tervo and Haapanen (2009),
Andersson and Hammarstedt (2011) and Lindquist
et al. (2013). Tervo and Haapanen (2009) analyse
gender differences in self-employment in Finnish
regions. In doing this, they estimate separate equationsfor men and women and hence indirectly find evidence
for the same-sex hypothesis, as the coefficient of a
dummy for self-employment experience of the father
(mother) is largest for men (women). But as for
Chlosta et al. (2012), they do not control for the
influence through the other channels. Dunn and Holtz-
Eakin (2000) also report (but do not show) results for
women that point in the same direction.
Andersson and Hammarstedt (2011) study the
intergenerational transmission of self-employment
abilities among Swedish immigrants. As in Tervoand Haapanen (2009), separate regressions for men
and women are conducted, and they also find similar
results, but again, they do not control for the potential
effects working through the other channels.
Finally, Lindquist et al.s (2013) study is the most
closely related study to ours. Using data on both the
biological and adoptive parents of adopted children,
they are able to show that most of the intergenerational
transmission mechanism is due to post-birth factors.
They explore the reasons for this and find a stronger
same-sex transmission mechanism, which they inter-pret as evidence of a role-model effect. They go
further than the other studies in controlling for the
effects working through the other channels by, e.g.
including a measure for parental income and a proxy
for their wealth (the financial channel). They also try
to test for the importance of the human-capital channel
by analyzing whether a larger number of siblings (and
thus less parental time for each child) reduce the
intergenerational transmission effect.
2Whether preferences for entrepreneurship or entrepreneurial
human capital are transmitted from parents to children via
observing and interacting with the parents can be difficult to
discern. However, in the current literature, the human-capital
channel refers to the specific skills that are acquired by working/
participating in the activities of the parents firm, whereas the
values, ideas, tricks, etc.,communicated through daily interac-
tion belong to the preference channel and is interpreted as a role-model effect. If we instead choose to define human capital more
broadly as all accumulated experience since birth, the human-
capital channel should also include social-learning effects.
Under this interpretation of the human-capital channel, the
hypotheses above cannot discriminate between the human-
capital and the preference channel, but the hypotheses then point
to a neglected aspect of the standard human-capital model,
namely gender-specific effects in the intergenerational transfer
of human capital. Furthermore, if children work for their parents
without being formally employed by these,we will not be able to
distinguish the effects of this from the effects working through
the preference channel.
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
6/26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
7/26
repeatedly. However, to check the robustness of our
results, we include estimations where we use these
alternative dependent variables.
To test the observation hypothesis from the previ-
ous section, we include as explanatory variables
different measures of parental experience from self-
employment; separately for the mother and the father.The simplest measures are just two dummies, Father
self-employed and Mother self-employed, capturing
whether the father and the mother, respectively, were
self-employed between 1980 and year t. However, we
also use four dummies for each parent to measure the
intensity of the self-employment experience, where
the intensity is measured as the percentage of the
annual observations between 1980 and year tin which
the parent in question was self-employed.
To test the interaction hypothesis, we construct
dummy variables capturing whether the parents wereself-employed while the individual was still living at
home. These dummies are named Father self-
employed while child living at home and Mother
self-employed while child living at home. We also
construct variables measuring the number of years the
child has worked for the mother (Years employed with
mother) and father (Years employed with father),
respectively. However, while positive effects of these
variables are expected according to the interaction
hypothesis, they will (as explained above) also directly
reflect effects working through the human-capitalchannel. Hence, we include them primarily to control
for these latter effects.
Finally, to test the same-sex hypothesis, we include
interaction terms between the above variables and the
gender of the individual, captured by a dummy
variable, Male, which takes the value 1 for men and
the value 0 for women. For the variables measuring the
number of years that the offspring has worked with
each parent, these interaction terms will capture same-
sex interaction effects, but they will also control for
parental same-sex investments working through thehuman-capital channel.3
Throughout, we control for a number of other
individual characteristics that may be expected to
affect the self-employment choice, including the
wealth of the parents. Summary statistics of these
and the variables mentioned above are presented in
Table1.
From the first part of the Table, we observe thatclose to 1.0 % of the observations involve transitions
to self-employment, i.e. on average, 1 out of 100
persons become self-employed each year. From the
second part of the Table, it follows that 29 and 10 % of
the children have fathers and mothers, respectively,
with self-employment experience. The next rows
present the self-employment intensities of the parents.
The majority (80 %) of the self-employed mothers
(i.e. those with a strictly positive intensity) have
intensities below 50 % meaning that they have been
self-employed less than half of the years since 1980.Among the self-employed fathers, close to 50 % have
been self-employed more than half of the time
indicating that males have higher self-employment
intensities than females. Most of the children with
self-employed parents were still living at home when
their parents were self-employed, and the average
work experience from the fathers firm is considerably
higher than the average work experience from the
mothers firm.
In the final part of the Table, the characteristics of
the children are shown. The average age in the sampleis 29.9 years, and the children have on average
12.7 years of schooling and 8.7 years of labour market
experience.4 Furthermore, 64 % have a partner, 41 %
have children themselves and 33 % are living in the
Copenhagen area. The measure of the parents total
wealth is from 1996, the last year where it is observed
in the data.
Before we turn to the multivariate analyses and the
more formal tests of the hypotheses, let us consider
some simple correlations between the main variables
3Lindquist et al. (2013) try to test for the importance of parental
same-sex investments by including the number of sisters
(brothers) interacted with the self-employment status of the
mother (father). Their idea is that if the same-sex correlations
were driven by parental same-sex investments, then a larger
number of sisters should have a negative effect, as the
investment has to be divided between all sisters. However, a
negative effect would also be consistent with social-learning
Footnote 3 continued
theory, as we would then expect less interaction between the
mother and the child and therefore a smaller role-model effect.
Hence, controlling for the number of years that the offspring has
worked with each parent seems a more direct way to control for
parental same-sex investments.4
In the estimations below, we rely on a more flexible
specification of schooling, using both Years of schooling and
dummies for nine different categories of education programmes
based on length and type.
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
8/26
of interest to get a first impression of whether the data
support the various hypotheses.
Table2 shows the annual transition rates to self-
employment conditional on the self-employment
experience of the parents. The transition rates are
presented separately for males and females. As the
transition rates are small, the odds ratios are basically
equal to the relative transition rates, i.e. the transition
rate for those with a self-employed father (or mother)
relative to the transition rate for those without a self-
employed father (or mother). Not surprisingly, tran-
sition rates are generally higher for males than for
females, and having a parent with self-employmentexperience raises the probability of becoming self-
employed (all odds ratios significantly exceed one).
More interestingly, having a self-employed mother
doubles the probability of becoming self-employed for
females, while it raises the probability for males by
less than 60 %. Similarly, having a father with
experience from self-employment also roughly dou-
bles the transition probability for males, while it adds
less than 60 % to the probability for females. This is a
strong indication that gender-specific effects are
present in the data.Table3considers only those observations where at
least one of the parents has been self-employed. The
left part of Table 3focuses on observations where the
father has been self-employed for at least one year in
the period considered, while the right part of the Table
presents similar figures for observations where the
mother has been self-employed for at least one year.
We can see that a higher intensity of the self-
employment experience of the father significantly
raises the probability that the male offspring
becomes self-employed, while there does not seemto be an effect of a higher intensity on the female
offspring. If we disregard the relatively few obser-
vations where the mother has been self-employed
during the entire period (the last column in the
Table), it also appears that a higher intensity of the
maternal self-employment experience raises the
transition rate for the offspring, but now relatively
more for females. As in Table2, this points to a
gender-specific effect of the parents experience
from self-employment.
Using the same observations as in Table 3, Table4divides the observations according to whether (some
of) the self-employment experience of the parents was
obtained while the child was living at home. Despite
the fact that the effect of the father and the effect of the
mother look very similar for males (and also for
females), the differences in odds ratios are again
important and significant. For males, it increases the
transition probability by 80 % to have had a father
who was self-employed while he (the son) was still
Table 1 Summary statistics
Mean SD Min Max
Dependent variable
Transition to self-
employment
0.008 0.090 0 1
Parental characteristicsFather self-employed 0.290 0.454 0 1
Mother self-employed 0.101 0.301 0 1
Fathers self-employment intensity:
0 % 0.711 0.454 0 1
150 % 0.159 0.366 0 1
5199 % 0.086 0.281 0 1
100 % 0.044 0.205 0 1
Mothers self-employment intensity
0 % 0.899 0.301 0 1
150 % 0.080 0.272 0 1
5199 % 0.017 0.129 0 1
100 % 0.004 0.060 0 1
Father self-employed
while child living at
home
0.249 0.432 0 1
Mother self-employed
while child living at
home
0.075 0.264 0 1
Years employed with
father
0.261 1.219 0 28
Years employed with
mother
0.038 0.441 0 28
Individual characteristics
Male 0.536 0.499 0 1
Age 29.863 6.073 15 44
Partner 0.642 0.479 0 1
Children 0.409 0.492 0 1
Years of schooling 12.660 2.400 9 18
Years of labour market
experience (as
employee)
8.730 5.606 0 29
Parents wealth (1996),
DKK millions
0.416 2.733 -328.31 909.60
Copenhagen area 0.334 0.472 0 1
See text for definitions of variables. Number of observations:
6,263,246
A. Hoffmann et al.
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
9/26
living with his parents. The similar effect is less than
50 % for females. This picture is reversed when we
look at the effects of a mother who was self-employed
while the child was living at home. Here, it raises thesubsequent transition probability by around 90 % for
females, while the effect for males is less than 50 %.
Finally, Table5 splits the data according to
whether the individual has previously been employed
in the parents firms. This is in general seen to be
associated with a markedly higher subsequent transi-
tion rate for both males and females, and again, there is
a relatively strong indication of gender-specific
effects.
Table 4 Transition rates by gender and parental self-employment experience while child was living at home
Father self-employed while child living at home Mother self-employed while child living at home
Yes No Odds ratio Yes No Odds ratio
Male 0.0172
(0.00014)
0.0095
(0.00019)
1.83
(0.02)
0.0161
(0.00025)
0.0110
(0.00006)
1.47
(0.03)
Female 0.0056
(0.00009)
0.0039
(0.00004)
1.45
(0.03)
0.0076
(0.00019)
0.0040
(0.00004)
1.88
(0.05)
See text for definitions of variables and Table2 for calculation of odds ratios. The left (right) part of the table only includes
observations where the father (mother) has been self-employed for at least 1 year since 1980. Standard errors in parentheses
Table 5 Transition rates by gender and work experience from
the parents firms
Employed with father Employed with mother
Yes No Yes No
Male 0.0224
(0.00027)
0.0103
(0.00006)
0.0223
(0.00074)
0.0112
(0.00006)
Female 0.0068
(0.00018)
0.0041
(0.00004)
0.0111
(0.00049)
0.0042
(0.00004)
See text for definitions of variables. The left (right) part of the
table only includes observations where the father (mother) has
been self-employed for at least 1 year since 1980. Standard
errors in parentheses
Table 2 Transition rates by gender and self-employment experience of the parents
Father self-employed Mother self-employed
Yes No Odds ratio Yes No Odds ratio
Male 0.0172
(0.00013)
0.009
(0.00006)
1.92
(0.02)
0.0168
(0.00022)
0.0108
(0.00006)
1.57
(0.03)
Female 0.0058
(0.00008)
0.0037
(0.00004)
1.57
(0.03)
0.0079
(0.00016)
0.0039
(0.00004)
2.04
(0.05)
See text for definitions of variables. Odds ratios are calculated as follows: ORMale,Father = (PMale,FatherSE/(1-PMale,FatherSE))/
(PMale,FatherNSE/(1-PMale,FatherNSE)), where PMale,FatherSE and PMale,FatherNSE are the transition probabilities for a male with a self-
employed father and a non-self-employed father, respectively. Standard errors in parentheses
Table 3 Transition rates by gender and the intensity of parental self-employment experience
Fathers self-employment intensity Mothers self-employment intensity
150 % 5199 % 100 % 150 % 5199 % 100 %
Male 0.0148(0.00017)
0.0192(0.00026)
0.0220(0.00038)
0.0167(0.00025)
0.0176(0.00055)
0.0150(0.00109)
Female 0.0058
(0.00011)
0.0059
(0.00015)
0.0053
(0.00020)
0.0076
(0.00018)
0.0096
(0.00044)
0.0076
(0.00086)
See text for definitions of variables. The left (right) part of the table only includes observations where the father (mother) has been
self-employed for at least 1 year since 1980. Standard errors in parentheses
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
10/26
Table 6 Transition probability and parental experience, main specification (probit estimation)
(1) (2) (3) (4) (5) (6) (7) (8)
Individual characteristics
Male 0.402**
(0.004)
0.403**
(0.004)
0.403**
(0.004)
0.398**
(0.004)
0.384**
(0.006)
0.384**
(0.006)
0.384**
(0.006)
0.385**
(0.006)
Parental characteristics
Mother self-employed 0.142**
(0.005)
0.203**
(0.009)
Father self-employed 0.187**
(0.004)
0.128**
(0.007)
Mothers self-employment intensity
150 % 0.141**
(0.006)
0.151**
(0.010)
0.144**
(0.010)
0.184**
(0.010)
0.189**
(0.017)
0.179**
(0.017)
5199 % 0.195**
(0.012)
0.208**
(0.015)
0.178**
(0.016)
0.291**
(0.019)
0.298**
(0.025)
0.248**
(0.027)
100 % 0.171**
(0.026)
0.185**
(0.028)
0.154**
(0.029)
0.246**
(0.044)
0.253**
(0.048)
0.199**
(0.050)
Fathers self-employment intensity
150 % 0.140**
(0.005)
0.162**
(0.008)
0.155**
(0.008)
0.112**
(0.009)
0.136**
(0.014)
0.130**
(0.015)
5199 % 0.227**
(0.006)
0.256**
(0.010)
0.205**
(0.010)
0.151**
(0.011)
0.182**
(0.019)
0.147**
(0.019)
100 % 0.271**
(0.007)
0.301**
(0.012)
0.226**
(0.012)
0.150**
(0.016)
0.182**
(0.022)
0.137**
(0.023)
Mother self-employed while child living at home -0.014
(0.011)
-0.013
(0.011)
-0.007
(0.019)
-0.007
(0.019)
Father self-employed while child living at home -
0.030**(0.009)
-
0.034**(0.009)
-
0.032*(0.016)
-
0.035*(0.016)
Years employed with mother 0.022**
(0.003)
0.030**
(0.004)
Years employed with father 0.029**
(0.001)
0.027**
(0.003)
Interactions
Male 9 Mother self-employed -0.087**
(0.011)
Male 9 Father self-employed 0.083**
(0.008)
Male 9 Mothers self-employment intensity150 % -0.062**
(0.013)
-0.055**
(0.020)
-0.050*
(0.020)
5199 % -0.142**
(0.024)
-0.134**
(0.032)
-0.102**
(0.033)
100 % -0.108*
(0.054)
-0.099
(0.059)
-0.064
(0.061)
Male 9 Fathers self-employment intensity
150 % 0.040**
(0.011)
0.037*
(0.017)
0.035*
(0.017)
A. Hoffmann et al.
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
11/26
4 Results
Section4.1contains our main results, while Sect.4.2
adds a number of robustness checks. In Sect. 4.3, we
compare the importance of role models in self-
employment with the importance of role models for
the propensity to become a teacher.
4.1 Main results
Table6 contains a number of probit estimations in
order to formally test the hypotheses laid out in Sects.
2. In all estimations, we use a dummy for the transition
from wage employment into self-employment as the
dependent variable. In addition to the variables of
interest for the hypotheses in question, all modelsinclude the control variables from Table1as well as
year dummies, eight industry dummies and four
dummies for ethnicity. To save space, the coefficient
estimates for these control variables are omitted from
Table6, but they are generally as expected.
The first two columns of Table 6test the observa-
tion hypothesis, i.e. that parental experience from self-
employment positively influences the self-employ-
ment choices of the children. Consistent with the
existing literature, we find strong support for this
hypothesis. In column 1, we have included the two
dummies for the parents self-employment experi-
ence. Both dummies are found to have a significant
positive effect on the transition probability. As in
Table2, the effect of a self-employed father is found
to be strongest.
In column 2, each dummy is replaced by three
dummies reflecting the intensity of the self-employ-
ment experience of the parents (no experience is the
reference category). The dummies are defined as in
Table3above. As can be seen, the positive effect of
the fathers self-employment experience increases
with the intensity of the experience. This is also the
case for the self-employment intensity of the mother
except when it is increased to 100 %, where the effectis slightly lower (although not significantly) than with
an intensity between 51 and 99 %. However, we
should remember that rather few individuals have a
mother who has been self-employed during the entire
period considered, cf. Table1.
In sum, columns 1 and 2 provide strong support for
the observation hypothesis. As we control for parental
wealth throughout, the results are unlikely to reflect an
effect running through the financial channel, but they
Table 6 continued
(1) (2) (3) (4) (5) (6) (7) (8)
5199 % 0.107**
(0.013)
0.103**
(0.022)
0.084**
(0.023)
100 % 0.166**
(0.018)
0.162**
(0.026)
0.126**
(0.027)
Male 9 Mother self-employed while child living at home -0.009
(0.024)
-0.007
(0.024)
Male 9 Father self-employed while child living at home 0.004
(0.019)
0.003
(0.019)
Male 9 Years employed with mother -0.015**
(0.006)
Male 9 Years employed with father 0.000
(0.003)
Observations 6,263,246 6,263,246 6,263,246 6,263,246 6,263,246 6,263,246 6,263,246 6,263,246
Log likelihood -270,392 -270,201 -270,193 -270,045 -270,313 -270,090 -270,083 -269,951
The dependent variable is a dummy for the transition into self-employment between year t-1 and t. Control variables included but
not reported are: age, age squared, years of schooling, eight dummies for education programmes defined by length and type, years of
labour market experience, a dummy for having a partner, a dummy for having children, a dummy for living in the Copenhagen area,
parents total wealth in 1996, seven year dummies, eight industry dummies and four dummies for ethnicity. **, * indicate significance
at the 1 % level and 5 % level, respectively. Standard errors clustered at the individual level in parentheses
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
12/26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
13/26
Fig. 1 Partial effect of fathers self-employment
Fig. 2 Partial effect of mothers self-employment
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
14/26
transition probability with close to 2.5 % points for
females and by approximately half of this for males.
Taken together, the Figures show that there is not only
statistically significant evidence in favour of the same-
sex hypothesis, the magnitudes of the gender-specific
effects are also sizeable.
An alternative way of illustrating the quantitativeimportance of having a self-employed father is to plot
therelativepartial effects (i.e. the partial effects from
above divided by the estimated initial transition
probability, Uxb) against the estimated initialtransition probability. This is done in Fig. 3, while the
effects of having a self-employed mother are illus-
trated in Fig.4.
These figures show that the relative effects are
largest at small initial transition probabilities,
where having a self-employed father can double the
transition probability for males, while increasing it by50 % for females, and vice versa in the case of a self-
employed mother. At higher levels of the transition
probabilities, the relative effects are smaller but still
sizeable. Hence, at initial transition probabilities
around 0.05, having a self-employed father raises the
transition probability by more than 50 % for men and
by around 30 % for women. Similar but opposite
effects are found in the case of a self-employed
mother.
Columns 68 in Table 6 contain further analyses
of the same-sex hypothesis. In column 6, we use the
same specification as in column 2, but with inter-
action terms between the gender and the threedifferent dummies for the intensities of the parents
self-employment experience. As in column 5, the
results are strongly supporting the same-sex hypoth-
esis. Being a male significantly increases the
importance of the fathers self-employment inten-
sityand more at higher intensitieswhile it
decreases the importance of the mothers self-
employment intensity.
In column 7, we also include interaction terms
between the gender and the dummies for the parents
being self-employed while the child was living athome. Although the signs of the estimated coefficients
again support the same-sex hypothesis, the results are
not statistically significant.
Finally, in column 8, we again control for
experience from the parents firms and interact
these variables with the gender dummy. In this
Fig. 3 Relative partial effect of fathers self-employment
A. Hoffmann et al.
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
15/26
way, we should be able to control for parental
same-sex investments as argued above. In the raw
data, we observe that 1.6 and 6.9 % of the women
have worked for their mother and father, respec-
tively, while 1.2 and 9.0 % of the men have
worked for their mother and father, respectively.
Hence, women (men) are relatively more likely tohave worked for their mother (father)an indica-
tion that parental same-sex investments are indeed
taking place. Furthermore, the average number of
years that a woman has worked for her mother
(father) is 2.9 years (2.9 years), while the average
number of years that a man has worked for his
mother (father) is 2.8 years (3.5 years). Again, this
is an indication that same-sex parental investments
in human capital are made (especially by the
fathers).
As can be seen from column 8, the effects ofhaving worked for the mother and the father are
rather similar for females, while for males the
effects of having worked for the mother are
weaker. However, controlling for this only has a
small influence on the other coefficients, leading us
to conclude that although parental same-sex invest-
ments seem to be present they can only explain
a small share of the same-sex correlations.
This supports that the evidence in favour of the
same-sex hypothesis is not driven by human-capital
effects.6
Taken together, the evidence in Table 6 provides
strong support for two of our three hypotheses: the
observation hypothesis and the same-sex hypothesis,
while the evidence for the interaction hypothesis is
more mixed. We also argued above that theevidence in favour of the observation and same-
sex hypotheses is likely to reflect an effect through
the preference channel as we controlled for both
parental wealth and work experience from the
parents firms, whereas the evidence in favour of
the interaction hypothesis can reflect both a human-
capital and a preference effect. Somewhat surpris-
inglyand in contrast to the predictions of social-
learning theorywe did not find evidence that
preferences for self-employment are more effec-
tively passed on to the children if the childrenexperience their parents as self-employed while they
are still living at home.
Fig. 4 Relative partial effect of mothers self-employment
6Furthermore, we have also tried to run the regressions in
Table6 without those observations where the individual has
worked for his/her parents. The results (not reported) are very
similar.
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
16/26
4.2 Robustness analyses
In this section, we consider a number of robustness
checks of the findings from the previous section.
First, Table7contains estimates similar to those in
Table6, but for males and females separately. The
estimation results in columns 14 are for males, while
the results in columns 58 are for females. This
corresponds to estimating the original model from
Table6 allowing for different gender-specific coeffi-
cients of all variables, not just the variables that relate to
the self-employment experience of the parents.
The parameter estimates for females in Table7
(columns 58) should be compared to those in the upper
part of columns 58 in Table6, while those for males
(columns 14) should be compared to the sum of the
coefficients of the original variables and of the interac-
tion terms in columns 58 of Table 6. As can be seen,
Table 7 Transition probability and parental experience, males versus females (probit estimation)
Males Females
(1) (2) (3) (4) (5) (6) (7) (8)
Parental characteristics
Mother self-employed 0.116**
(0.006)
0.200**
(0.009)
Father self-employed 0.209**
(0.005)
0.128**
(0.007)
Mothers self-employment intensity
150 % 0.122**
(0.007)
0.134**
(0.012)
0.129**
(0.012)
0.183**
(0.011)
0.188**
(0.017)
0.178**
(0.017)
5199 % 0.150**
(0.014)
0.165**
(0.019)
0.147**
(0.020)
0.286**
(0.019)
0.292**
(0.026)
0.247**
(0.027)
100 % 0.134**
(0.032)
0.150**
(0.035)
0.132**
(0.035)
0.241**
(0.045)
0.248**
(0.049)
0.199**
(0.050)
Fathers self-employment intensity150 % 0.153**
(0.006)
0.176**
(0.010)
0.169**
(0.010)
0.108**
(0.009)
0.123**
(0.015)
0.118**
(0.015)
5199 % 0.256**
(0.007)
0.287**
(0.012)
0.233**
(0.013)
0.148**
(0.011)
0.168**
(0.019)
0.135**
(0.019)
100 % 0.305**
(0.009)
0.336**
(0.014)
0.254**
(0.014)
0.168**
(0.016)
0.189**
(0.022)
0.146**
(0.023)
Mother self-employed while child living at home -0.017
(0.014)
-0.015
(0.014)
-0.007
(0.019)
-0.006
(0.020)
Father self-employed while child living at home -0.032**
(0.011)
-0.036**
(0.011)
-0.021
(0.016)
-0.023
(0.016)
Years employed with mother 0.015**(0.004)
0.026**(0.004)
Years employed with father 0.028**
(0.001)
0.025**
(0.003)
Observations 3,358,379 3,358,379 3,358,379 3,358,379 2,904,867 2,904,867 2,904,867 2,904,867
Log likelihood -194,900 -194,720 -194,714 -194,434 -74,405 -74,382 -74,381 -74,316
The dependent variable is a dummy for the transition into self-employment between year t-1 and t. Control variables included but
not reported are: age, age squared, years of schooling, eight dummies for education programmes defined by length and type, years of
labour market experience, a dummy for having a partner, a dummy for having children, a dummy for living in the Copenhagen area,
parents total wealth in 1996, seven year dummies, eight industry dummies and four dummies for ethnicity. **, * indicate significance
at the 1 % level and 5 % level, respectively. Standard errors clustered at the individual level in parentheses
A. Hoffmann et al.
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
17/26
Table 8 Transition probability and parental experience, excluding start-ups in parents firms and industries (probit estimation)
Excl. individuals taking over the parents firm Excl. individuals starting up in the parents
industry
(1) (2) (3) (4) (5) (6) (7) (8)
Individual characteristics
Male 0.385**
(0.006)
0.385**
(0.006)
0.385**
(0.006)
0.385**
(0.006)
0.387**
(0.006)
0.387**
(0.006)
0.386**
(0.006)
0.386**
(0.006)
Parental characteristics
Mother self-employed 0.195**
(0.009)
0.163**
(0.010)
Father self-employed 0.121**
(0.007)
0.096**
(0.007)
Mothers self-employment intensity
150 % 0.180**
(0.011)
0.183**
(0.017)
0.176**
(0.017)
0.158**
(0.011)
0.160**
(0.018)
0.162**
(0.018)
5199 % 0.262**
(0.020)
0.266**
(0.026)
0.234**
(0.027)
0.190**
(0.022)
0.193**
(0.028)
0.200**
(0.029)
100 % 0.239**
(0.045)
0.242**
(0.049)
0.209**
(0.050)
0.135**
(0.050)
0.138*
(0.054)
0.146**
(0.054)
Fathers self-employment intensity
150 % 0.112**
(0.009)
0.135**
(0.015)
0.132**
(0.015)
0.100**
(0.009)
0.126**
(0.015)
0.126**
(0.015)
5199 % 0.137**
(0.011)
0.166**
(0.019)
0.147**
(0.019)
0.097**
(0.012)
0.131**
(0.019)
0.132**
(0.020)
100 % 0.131**
(0.016)
0.161**
(0.023)
0.136**
(0.023)
0.080**
(0.017)
0.116**
(0.023)
0.118**
(0.024)
Mother self-employed while child living at home -
0.004(0.020)
-
0.004(0.020)
-
0.004(0.020)
-
0.003(0.020)
Father self-employed while child living at home -0.031
(0.016)
-0.032*
(0.016)
-0.036*
(0.016)
-0.036*
(0.016)
Years employed with mother 0.020**
(0.005)
-0.006
(0.006)
Years employed with father 0.017**
(0.003)
-0.001
(0.004)
Interactions
Male 9 Mother self-employed -0.080**
(0.012)
-0.051**
(0.012)
Male 9 Father self-employed 0.071**
(0.009)
0.017*
(0.009)
Male 9 Mothers self-employment intensity
150 % -0.059**
(0.013)
-0.051*
(0.021)
-0.047*
(0.021)
-0.045**
(0.013)
-0.038
(0.022)
-0.036
(0.022)
5199 % -0.129**
(0.025)
-0.118**
(0.033)
-0.093**
(0.034)
-0.082**
(0.027)
-0.072*
(0.035)
-0.067
(0.036)
100 % -0.100
(0.055)
-0.089
(0.060)
-0.062
(0.061)
-0.019
(0.060)
-0.008
(0.065)
-0.004
(0.066)
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
18/26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
19/26
Table 9 Alternative dependent variables, self-employment status and self-employed between 1995 and 2009
Self-employed in year t Self-employed between 1995 and 2009
(1) (2) (3) (4) (5) (6) (7) (8)
Individual characteristics
Male 0.599**
(0.007)
0.596**
(0.007)
0.596**
(0.007)
0.596**
(0.007)
0.625**
(0.008)
0.625**
(0.008)
0.625**
(0.008)
0.626**
(0.008)
Parental characteristics
Mother self-employed 0.255**
(0.012)
0.279**
(0.013)
Father self-employed 0.189**
(0.009)
0.189**
(0.010)
Mothers self-employment intensity
150 % 0.223**
(0.013)
0.215**
(0.021)
0.202**
(0.022)
0.259**
(0.014)
0.251**
(0.022)
0.236**
(0.022)
5199 % 0.413**
(0.024)
0.401**
(0.032)
0.334**
(0.033)
0.384**
(0.030)
0.374**
(0.038)
0.292**
(0.041)100 % 0.368**
(0.053)
0.355**
(0.059)
0.284**
(0.060)
0.555**
(0.081)
0.542**
(0.086)
0.448**
(0.088)
Fathers self-employment intensity
150 % 0.139**
(0.011)
0.141**
(0.019)
0.133**
(0.019)
0.161**
(0.012)
0.191**
(0.019)
0.183**
(0.019)
5199 % 0.244**
(0.014)
0.247**
(0.024)
0.195**
(0.024)
0.228**
(0.015)
0.269**
(0.025)
0.222**
(0.026)
100 % 0.277**
(0.020)
0.280**
(0.028)
0.215**
(0.029)
0.270**
(0.027)
0.313**
(0.034)
0.243**
(0.035)
Mother self-employed while child living at home 0.013
(0.025)
0.012
(0.025)
0.012
(0.026)
0.014
(0.026)Father self-employed while child living at home -0.003
(0.021)
-0.007
(0.021)
-0.043*
(0.021)
-0.045*
(0.021)
Years employed with mother 0.040**
(0.006)
0.035**
(0.006)
Years employed with father 0.041**
(0.004)
0.027**
(0.003)
Interactions
Male 9 Mother self-employed -0.143**
(0.015)
-0.098**
(0.016)
Male 9 Father self-employed 0.191**
(0.011)
0.121**
(0.012)
Male 9 Mothers self-employment intensity
150 % -0.094**
(0.016)
-0.083**
(0.026)
-0.077**
(0.026)
-0.074**
(0.018)
-0.031
(0.028)
-0.023
(0.028)
5199 % -0.201**
(0.029)
-0.186**
(0.039)
-0.146**
(0.041)
-0.144**
(0.038)
-0.083
(0.048)
-0.034
(0.051)
100 % -0.236**
(0.067)
-0.218**
(0.073)
-0.176*
(0.075)
-0.423**
(0.107)
-0.355**
(0.112)
-0.299**
(0.115)
Male 9 Fathers self-employment intensity
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
20/26
strengthened in our conclusion that the support for theobservation and the same-sex hypothesis doesnot stem
from an effect through the financial channel.7 Instead,
we feel comfortable concluding that it reflects a
preference effect, as controlling for work experience
from the parents firms (column 4) still has a very
limitedeffecton the coefficient estimates of the intensity
variables. In other words, sons do not only follow in the
footsteps of their self-employed fathers when they
inherit (or work in) the family firm.
Takeovers of the family firm are identified through
the workplace identifier in our data set. In some cases,this identifier may change when a firm is inherited
by the next generation. In this case, the approach taken
above may not identify all the cases in which a firm is
handed over to a son or a daughter. To make sure thatwe eliminate all such cases, we now throw out all
observations where the individual becomes self-
employed in the same industry as the parents. A
similar approach has been used in Srensen (2007) and
Lindquist et al. (2013). Distinguishing between 111
industries in the economy, this eliminates more than
4,000 additional individuals. The results are presented
in columns 58 of Table8.
Despite the large number of observations elimi-
nated, we still find considerable support for both the
observation and the same-sex hypothesis. Some of theinteraction effects are quantitatively smaller and
slightly less significant, but this is not very surprising
given that we have excluded more than 5,000 (suc-
cessful) individuals compared to Table6and
probably those where we would expect the parental
influence to be strongest, namely the cases where the
children choose an occupation within the same
industry as the parents. Still, five out of the six
coefficients of the interaction terms with the self-
Table 9 continued
Self-employed in year t Self-employed between 1995 and 2009
(1) (2) (3) (4) (5) (6) (7) (8)
150 % 0.077**
(0.013)
0.061**
(0.022)
0.057*
(0.022)
0.083**
(0.014)
0.075**
(0.023)
0.072**
(0.023)
5199 % 0.265**
(0.016)
0.244**
(0.028)
0.211**
(0.028)
0.168**
(0.018)
0.156**
(0.030)
0.128**
(0.031)
100 % 0.329**
(0.022)
0.307**
(0.033)
0.250**
(0.034)
0.200**
(0.032)
0.187**
(0.041)
0.135**
(0.043)
Male 9 Mother self-employed while child living at home -0.017
(0.030)
-0.012
(0.030)
-0.068*
(0.033)
-0.067*
(0.033)
Male 9 Father self-employed while child living at home 0.022
(0.025)
0.017
(0.025)
0.012
(0.025)
0.009
(0.025)
Male 9 Years employed with mother -0.019*
(0.008)
-0.017*
(0.008)
Male9
Years employed with father 0.001(0.004)
0.001(0.004)
Observations 7,134,207 7,134,207 7,134,207 7,134,207 486,452 486,452 486,452 486,452
Log likelihood -826,590 -822,620 -822,614 -820,305 -132,210 -132,068 -132,059 -131,891
The dependent variable is a dummy for being self-employed in yeart(columns 14) and a dummy for having been self-employed at
least once between 1995 and 2009 (columns 58). In columns 14, a panel for the years 19952009 is used, whereas in columns 58,
a cross section from 2009 is used. Control variables included but not reported are: age, age squared, years of schooling, eight
dummies for education programmes defined by length and type, years of labour market experience, a dummy for having a partner, a
dummy for having children, a dummy for living in the Copenhagen area, parents total wealth in 1996, seven year dummies, eight
industry dummies and four dummies for ethnicity. **, * indicate significance at the 1 % level and 5 % level, respectively. Standard
errors clustered at the individual level in parentheses
7Leaving out the observations where the individuals take over
the firm of the father or the mother also has only a small negative
effect on the odds ratios from Table 2. The odds ratios for males
decrease from 1.92 and 1.57 to 1.82 and 1.56, respectively,
while the odds ratios for females decrease from 1.57 and 2.04 to
1.54 and 1.99, respectively.
A. Hoffmann et al.
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
21/26
Table 10 Transition probability and parental experience, assortative mating of parents (probit estimation)
Controlling for self-employment of both
parents
Excl. observations with both parents self-
employed
(1) (2) (3) (4) (5) (6) (7) (8)
Individual characteristics
Male 0.384**
(0.006)
0.385**
(0.006)
0.384**
(0.006)
0.385**
(0.006)
0.384**
(0.006)
0.385**
(0.006)
0.385**
(0.006)
0.386**
(0.006)
Parental characteristics
Mother self-employed 0.230**
(0.013)
0.230**
(0.013)
Father self-employed 0.138**
(0.008)
0.138**
(0.008)
Mother and Father self-
employed
-0.056**
(0.019)
-0.049**
(0.019)
-0.049**
(0.019)
-0.052**
(0.019)
Mothers self-employment intensity
150 % 0.208**
(0.014)
0.212**
(0.019)
0.204**
(0.019)
0.203**
(0.015)
0.214**
(0.023)
0.205**
(0.023)
5199 % 0.315**
(0.021)
0.320**
(0.027)
0.272**
(0.028)
0.329**
(0.026)
0.344**
(0.036)
0.290**
(0.037)
100 % 0.262**
(0.045)
0.269**
(0.049)
0.216**
(0.050)
0.285**
(0.051)
0.302**
(0.058)
0.244**
(0.060)
Fathers self-employment intensity .
150 % 0.122**
(0.010)
0.145**
(0.015)
0.141**
(0.015)
0.123**
(0.010)
0.149**
(0.016)
0.144**
(0.016)
5199 % 0.159**
(0.011)
0.190**
(0.019)
0.155**
(0.019)
0.160**
(0.012)
0.193**
(0.021)
0.158**
(0.021)
100 % 0.156**(0.016)
0.188**(0.022)
0.143**(0.023)
0.148**(0.017)
0.183**(0.025)
0.135**(0.025)
Mother self-employed while child living at home -0.007
(0.019)
-0.007
(0.019)
-0.017
(0.028)
-0.018
(0.028)
Father self-employed while child living at home -0.032*
(0.016)
-0.035*
(0.016)
-0.035
(0.018)
-0.039*
(0.018)
Years employed with mother 0.030**
(0.004)
0.034**
(0.006)
Years employed with father 0.027**
(0.003)
0.030**
(0.003)
Interactions
Male9Mother self-employed -0.081**
(0.016)
-0.081**
(0.016)
Male 9 Father self-employed 0.083**
(0.009)
0.082**
(0.009)
Male 9 Mother and Father self-
employed
-0.009
(0.023)
0.007
(0.023)
0.007
(0.023)
0.007
(0.023)
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
22/26
employment intensities of the mother and the father in
column 8 are significant at (at least) the 10 % level.8 In
other words, the importance of parental role
models seems to extend beyond the family firm andindustry.
As argued in Sect.3, other studies of the intergen-
erational transmission mechanism have used
alternative dependent variables. To test the robustness
of our results towards this, Table9 repeats the
estimations from columns 58 of Table6 using two
alternative dependent variables: (1) a dummy forbeing self-employed in yeart(columns 14); and (2) a
dummy for having been self-employed at least once
between 1995 and 2009. In the latter case, we use a
cross section from 2009.
As can be seen, the results are very similar to those
in columns 58 of Table6. In fact, when using Self-
employed in year t as the dependent variable, the
estimated coefficients of the interaction terms tend to
be slightly more significant than in Table 6. Thus, we
Table 10 continued
Controlling for self-employment of both
parents
Excl. observations with both parents self-
employed
(1) (2) (3) (4) (5) (6) (7) (8)
Male 9 Mothers self-employment intensity
150 % -0.064**
(0.017)
-0.057*
(0.023)
-0.052*
(0.023)
-0.053**
(0.018)
-0.044
(0.029)
-0.036
(0.029)
51 %99 % -0.144**
(0.026)
-0.136**
(0.033)
-0.105**
(0.035)
-0.178**
(0.034)
-0.166**
(0.045)
-0.116*
(0.046)
100 % -0.109*
(0.055)
-0.100
(0.060)
-0.065
(0.061)
-0.144*
(0.065)
-0.130
(0.074)
-0.077
(0.075)
Male 9 Fathers self-employment intensity
150 % 0.037**
(0.011)
0.035
(0.018)
0.032
(0.018)
0.031**
(0.012)
0.031
(0.020)
0.028
(0.020)
5199 % 0.105**
(0.013)
0.101**
(0.022)
0.082**
(0.023)
0.109**
(0.014)
0.109**
(0.025)
0.088**
(0.025)
100 % 0.164**
(0.018)
0.160**
(0.026)
0.124**
(0.027)
0.173**
(0.019)
0.173**
(0.029)
0.135**
(0.030)
Male 9 Mother self-employed while child living at home -0.009
(0.024)
-0.007
(0.024)
-0.013
(0.034)
-0.013
(0.034)
Male 9 Father self-employed while child living at home 0.004
(0.019)
0.003
(0.019)
-0.000
(0.022)
-0.000
(0.022)
Male 9 Years employed with mother -0.014*
(0.006)
-0.027**
(0.009)
Male 9 Years employed with father 0.000
(0.003)
-0.001
(0.003)
Observations 6,263,246 6,263,246 6,263,246 6,263,246 5,970,452 5,970,452 5,970,452 5,970,452Log likelihood -270,293 -270,080 -270,073 -269,715 -248,630 -248,437 -248,428 -248,105
The dependent variable is a dummy for the transition into self-employment between year t-1 and t. Control variables included but
not reported are: age, age squared, years of schooling, eight dummies for education programmes defined by length and type, years of
labour market experience, a dummy for having a partner, a dummy for having children, a dummy for living in the Copenhagen area,
parents total wealth in 1996, seven year dummies, eight industry dummies and four dummies for ethnicity. **, * indicate significance
at the 1 % level and 5 % level, respectively. Standard errors clustered at the individual level in parentheses
8Furthermore, if we estimate the model with just Years of
Schooling as a control for education instead of Years of
Schooling and nine dummies for different types of education
programmes, four of these five interaction terms become
significant at the 5 % level.
A. Hoffmann et al.
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
23/26
still find considerable support for both the observation
and the same-sex hypothesis, while the evidence in
favour of the interaction hypothesis remains rather
mixed.
In order to compare more directly with the
estimates in Lindquist et al. (2013), we have also
estimated the model in column 5 of Table 9 using a
linear probability model. In this case, we get partial-
effect estimates for males of 0.043 and 0.071 of having
a self-employed mother and father, respectively.Given that around 15 % of the males have been self-
employed, relative partial effects can be calculated as:
0.043/0.15 & 0.29 and 0.071/0.15& 0.47. For
females, the corresponding relative partial effects are
approximately 0.51 and 0.30. These effects are very
similar in size to those that can be computed from the
estimates for non-adopted children in Lindquist et al.
(2013) in the lower part of their Table7, using the fact
that 25 % of the non-adopted sons and 14 % of the
non-adopted daughters have been entrepreneurs
(Table2a). This results in relative partial effects for
males of approximately 0.43 and 0.52 of having an
entrepreneurial mother and father, respectively, while
the corresponding numbers for females are 0.49 and
0.30.
A final issue to consider is assortative mating.The raw data reveal that for 16.2 % of the observations
with a self-employed father, the mother has also been
self-employed at some point. Similarly, for 46.4 % of
the observations with a self-employed mother, the
father has also been self-employed. This is a strong
indication that assortative mating is taking place in the
data. Could this be driving the gender-specific effects?
Imagine that the effect of a self-employed father is in
general larger (for both boys and girls), but that girls
have a higher threshold to cross before becoming self-
employed. In other words, for girls to become self-employed, it requires that both parents have been
self-employed. Since fathers are in general more likely
to be self-employed, this could show up as a larger
effect of the mothers on the daughters and hence
explain the observed gender-specific effects even
without gender-specific role models.9
To test the importance of this for the intergenera-
tional transmission mechanism, Table10 contains two
sets of re-estimations of columns 58 from Table6.
First, we have tried to include (1) a dummy taking the
value one if both parents have been self-employed;and (2) the interaction of this variable with our gender
dummy. The coefficient of the former variable
becomes significantly negative (but small), while the
coefficient of the interaction term becomes insignif-
icant. This indicates diminishing effects of both
parents being self-employedeffects which are sim-
ilar for boys and girls. Hence, it does not support the
above hypothesis, and the inclusion of these extra
controls only marginally affects the other coefficients.
Second, in columns 58, we drop all the observa-
tions where both parents have been self-employed.This results in estimates very similar to those in
columns 14 and therefore does not change the overall
conclusion. In sum, although assortative mating
between self-employed persons is observed in the
data, it cannot explain the gender-specific effects and
hence the support for the same-sex hypothesis.
Table 11 Transition probability, schoolteacher (probit
estimation)
(1)
Individual characteristics
Male 0.13**
(0.007)Parental characteristics
Mother teacher 0.20**
(0.010)
Father teacher 0.15**
(0.010)
Interactions
Male 9 Mother teacher -0.11**
(0.014)
Male 9 Father teacher -0.01
(0.015)
Observations 6,212,530
Log likelihood -155,718
The dependent variable is a dummy for the transition into
becoming a school teacher between year t-1 and t. Control
variables included but not reported are: age, age squared, years
of schooling, eight dummies for education programmes defined
by length and type, years of labour market experience, a
dummy for having a partner, a dummy for having children, a
dummy for living in the Copenhagen area, parents total wealth
in 1996, seven year dummies, eight industry dummies and four
dummies for ethnicity. **, * indicate significance at the 1 %
level and 5 % level, respectively. Standard errors clustered at
the individual level in parentheses
9We are indebted to one of the referees for suggesting this
potential mechanism.
Parental role models in entrepreneurship
1 3
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
24/26
8/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
25/26
parental role models therefore seem to reach beyond
the family business and industry.
A number of studies in the literature have docu-
mented gender differences in the propensity to become
an entrepreneur; see, e.g. Fairlie and Meyer (1996) and
Verheul et al. (2006). A number of explanations have
been suggested to explain these differences; seeVerheul et al. (2006) for an overview. Our paper
offers a supplementary explanation for the observed
phenomenon, namely that the historical lack of female
entrepreneurs induces fewer women to become entre-
preneurs today. In other words, the lack of female
entrepreneurs in a country like Denmark may partly
reflect a lack of female role models. Changing this
pattern requires a concerted political effort building on
attentive ways of providing female entrepreneurial
role models.
The recommendation of such an effort should,however, also take into account how women fare as
entrepreneurs compared to men. A recent analysis
conducted by the Danish Business Authority (2010)
shows that male and female entrepreneurs differ in a
number of ways. The share of high-growth entrepre-
neurs is, e.g. higher among male entrepreneurs.
Women, on the other hand, have more education but
less labour market experience than men when starting
their own firm. Whether these differences translate
into higher or lower productivity and lifetime earnings
for female entrepreneurs is yet an unexplored issue.
Acknowledgments We thank the editor, Mirjam Van Praag,
and two anonymous referees for many useful comments and
suggestions, and we are grateful to Jan Holst Hansen and Philip
Willerslev-Olsen for excellent research assistance.
References
Andersson, L., & Hammarstedt, M. (2011). Transmission of
self-employment across immigrant generations: The
importance of ethnic background and gender. Review ofEconomics of the Household, 9, 555577. doi:10.1007/
s11150-010-9102-5.
Bennedsen, M., Nielsen, K. M., Perez-Gonzalez, F., & Wol-
fenzon, D. (2007). Inside the family firm: The role of
families in succession decisions and performance. Quar-
terly Journal of Economics, 122, 647691. doi:10.1162/
qjec.122.2.647.
Blanchflower, D. G. (2009). Minority self-employment in the
United States and the impact of affirmative action pro-
grams. Annals of Finance, 5, 361396. doi:10.1007/
s10436-008-0099-1.
Bonke, J., & Esping-Andersen, G. (2009). Family investment in
childrenproductivities, preferences, and parental child
care. European Sociological Review, 27, 4355. doi:10.
1093/esr/jcp054.
Bosma, N., Hessels, J., Schutjens, V., van Praag, M., & Verheul,
I. (2012). Entrepreneurship and role models. Journal of
Economic Psychology, 33, 410424. doi:10.1016/j.joep.
2011.03.004.
Chlosta, S., Patzelt, H., Klein, S. B., & Dormann, C. (2012).
Parental role models and the decision to become self-
employed: The moderating effect of personality. Small
Business Economics, 38, 121138. doi:10.1007/s11187-
010-9270-y.
Danish Business Authority. (2010). Kvindelige Ivrksttere
2010et Statistisk Portrt. Copenhagen: Report prepared
by the Danish Business Authority.
Dryler, H. (1998). Parental role models, gender and educational
choice.British Journal of Sociology, 49, 375398. doi:10.
2307/591389.
Dunn, T., & Holtz-Eakin, D. (2000). Financial capital, human
capital, and the transition to self-employment: Evidence
from intergenerational links.Journal of Labor Economics,18, 282305. doi:10.1086/209959.
Fairlie, R. W., & Meyer, B. D. (1996). Ethnic and racial self-
employment differences and possible explanations. Jour-
nal of Human Resources, 31, 757793. doi:10.2307/
146146.
Fairlie, R.W. & Robb, A. (2007). Families, human capital, and
small business: Evidence from the characteristics of busi-
ness owners survey. Industrial and Labor Relations Sur-
vey,60, 225245.http://www.jstor.org/stable/25249072.
Georgellis, Y., & Wall, H. J. (2005). Gender differences in self-
employment.International Review of Applied Economics,
3, 321342. doi:10.1080/02692170500119854.
Hetherington, E. M. (1965). A developmental study of the
effects of the sex of the dominant parent on sex role pref-erence, identification and imitation in children.Journal of
Personality and Social Psychology, 2, 188194. doi:10.
1037/h0022374.
Hout, M. (1984). Status, autonomy and training in occupational
mobility.American Journal of Sociology, 89, 13791409.
http://www.jstor.org/stable/2779187.
Hout, M., & Rosen, H. (2000). Self-employment, family back-
ground and race. Journal of Human Resources, 35,
670692. doi:10.2307/146367.
Laspita, S., Breugst, N., Heblich, S., & Patzelt, H. (2012).
Intergenerational transmission of entrepreneurial inten-
tions.Journal of Business Venturing, 27, 414435. doi:10.
1016/j.jbusvent.2011.11.006.
Lentz, B. S., & Laband, D. N. (1983). Like father, like son:
Toward an economic theory of occupational following.
Southern Economic Journal, 50, 474493. doi:10.2307/
1058220.
Lentz, B. S., & Laband, D. N. (1990). Entrepreneurial success
and occupational inheritance among proprietors.Canadian
Journal of Economics, 23, 563579. doi:10.2307/135648.
Lindquist, M. J., Sol, J. & Van Praag, M. (2013). Why do
entrepreneurial parents have entrepreneurial children?.
Journal of Labor Economics, Working Paper
(forthcoming).
Parental role models in entrepreneurship
1 3
http://dx.doi.org/10.1007/s11150-010-9102-5http://dx.doi.org/10.1007/s11150-010-9102-5http://dx.doi.org/10.1162/qjec.122.2.647http://dx.doi.org/10.1162/qjec.122.2.647http://dx.doi.org/10.1007/s10436-008-0099-1http://dx.doi.org/10.1007/s10436-008-0099-1http://dx.doi.org/10.1093/esr/jcp054http://dx.doi.org/10.1093/esr/jcp054http://dx.doi.org/10.1016/j.joep.2011.03.004http://dx.doi.org/10.1016/j.joep.2011.03.004http://dx.doi.org/10.1007/s11187-010-9270-yhttp://dx.doi.org/10.1007/s11187-010-9270-yhttp://dx.doi.org/10.2307/591389http://dx.doi.org/10.2307/591389http://dx.doi.org/10.1086/209959http://dx.doi.org/10.2307/146146http://dx.doi.org/10.2307/146146http://www.jstor.org/stable/25249072http://dx.doi.org/10.1080/02692170500119854http://dx.doi.org/10.1037/h0022374http://dx.doi.org/10.1037/h0022374http://www.jstor.org/stable/2779187http://dx.doi.org/10.2307/146367http://dx.doi.org/10.1016/j.jbusvent.2011.11.006http://dx.doi.org/10.1016/j.jbusvent.2011.11.006http://dx.doi.org/10.2307/1058220http://dx.doi.org/10.2307/1058220http://dx.doi.org/10.2307/135648http://dx.doi.org/10.2307/135648http://dx.doi.org/10.2307/1058220http://dx.doi.org/10.2307/1058220http://dx.doi.org/10.1016/j.jbusvent.2011.11.006http://dx.doi.org/10.1016/j.jbusvent.2011.11.006http://dx.doi.org/10.2307/146367http://www.jstor.org/stable/2779187http://dx.doi.org/10.1037/h0022374http://dx.doi.org/10.1037/h0022374http://dx.doi.org/10.1080/02692170500119854http://www.jstor.org/stable/25249072http://dx.doi.org/10.2307/146146http://dx.doi.org/10.2307/146146http://dx.doi.org/10.1086/209959http://dx.doi.org/10.2307/591389http://dx.doi.org/10.2307/591389http://dx.doi.org/10.1007/s11187-010-9270-yhttp://dx.doi.org/10.1007/s11187-010-9270-yhttp://dx.doi.org/10.1016/j.joep.2011.03.004http://dx.doi.org/10.1016/j.joep.2011.03.004http://dx.doi.org/10.1093/esr/jcp054http://dx.doi.org/10.1093/esr/jcp054http://dx.doi.org/10.1007/s10436-008-0099-1http://dx.doi.org/10.1007/s10436-008-0099-1http://dx.doi.org/10.1162/qjec.122.2.647http://dx.doi.org/10.1162/qjec.122.2.647http://dx.doi.org/10.1007/s11150-010-9102-5http://dx.doi.org/10.1007/s11150-010-9102-58/12/2019 Running in the Family Parental Role Models in Entrepreneurship Hoffmann Small Business Economics 2014
26/26
Maccoby, E. E. (1992). The role of parents in the socialization of
children: An historical overview. Developmental Psy-
chology, 28, 10061017. doi:10.1037/0012-1649.28.6.
1006.
Matthews, C. H., & Moser, S. B. (1996). A longitudinal inves-
tigation of the impact of family background and gender on
interest in small-firm ownership.Journal of Small Business
Management, 34, 2943.
Nicolaou, N., & Shane, S. (2009). Can genetic factors influence
the likelihood of engaging in entrepreneurial activity?
Journal of Business Venturing, 24, 122. doi:10.1016/j.
jbusvent.2007.11.003.
Nicolaou, N., & Shane, S. (2010). Entrepreneurship and occu-
pational choice: Genetic and environmental influences.
Journal of Economic Behavior & Organization, 76, 314.
doi:10.1016/j.jebo.2010.02.009.
Parsons, J. E., Adler, T. F., & Kaczala, C. M. (1982). Sociali-
zation of achievement attitudes and beliefs: Parental
influences.Child Development, 53, 310321. doi:10.2307/
1128973.
Scherer, R., Adams, J., Carley, S., & Wiebe, F. (1989). Role
model performance effects on development of
entrepreneurial career preferences. Entrepreneurship
Theory and Practice, 13, 5371.
Slaby, R. G., & Frey, K. S. (1975). Development of gender
constancy and selective attention to same-sex models.
Child Development, 46, 849856. doi:10.2307/1128389.
Srensen, J. B. (2007). Closure and exposure: Mechanisms in
the intergenerational transmission of self-employment.
Research in the Sociology of Organizations, 25, 83124.
doi:10.1016/S0733-558X(06)25003-1.
Tervo, H., & Haapanen, M. (2009). The nature of self-
employment: How does gender matter? International
Journal of Entrepreneurship and Small Business, 9,
349371. doi:10.1504/IJESB.2010.031926.
Van Auken, H., Stephens, P., Fry, F. L., & Silva, J. (2006). Role
model influences on entrepreneurial intentions: A com-
parison between USA and Mexico. International and
Entrepreneurship Management Journal, 2, 325336.
doi:10.1007/s11365-006-0004-1.
Verheul, I., Van Stel, A., & Thurik, R. (2006). Explaining
female and male entrepreneurship at the country level.
Entreprenurship and Regional Development,18, 151183.
doi:10.1080/08985620500532053.
A. Hoffmann et al.
http://dx.doi.org/10.1037/0012-1649.28.6.1006http://dx.doi.org/10.1037/0012-1649.28.6.1006http://dx.doi.org/10.1016/j.jbusvent.2007.11.003http://dx.doi.org/10.1016/j.jbusvent.2007.11.003http://dx.doi.org/10.1016/j.jebo.2010.02.009http://dx.doi.org/10.2307/1128973http://dx.doi.org/10.2307/1128973http://dx.doi.org/10.2307/1128389http://dx.doi.org/10.1016/S0733-558X(06)25003-1http://dx.doi.org/10.1504/IJESB.2010.031926http://dx.doi.org/10.1007/s11365-006-0004-1http://dx.doi.org/10.1080/08985620500532053http://dx.doi.org/10.1080/08985620500532053http://dx.doi.org/10.1007/s11365-006-0004-1http://dx.doi.org/10.1504/IJESB.2010.031926http://dx.doi.org/10.1016/S0733-558X(06)25003-1http://dx.doi.org/10.2307/1128389http://dx.doi.org/10.2307/1128973http://dx.doi.org/10.2307/1128973http://dx.doi.org/10.1016/j.jebo.2010.02.009http://dx.doi.org/10.1016/j.jbusvent.2007.11.003http://dx.doi.org/10.1016/j.jbusvent.2007.11.003http://dx.doi.org/10.1037/0012-1649.28.6.1006http://dx.doi.org/10.1037/0012-1649.28.6.1006