Family structure and entrepreneurship: Evidence from Swedish siblings

ABSTRACT

Family background matters for entrepreneurship, but why do siblings differ in their propensity to become entrepreneurs and the type of ventures they pursue? I draw on family socialization and resource allocation theories to develop hypotheses about the differential effects of family structure – comprising birth order, family size, and sibling gender – on (growth-oriented) entrepreneurship. Using firm incorporation as a marker of growth orientation, I test these hypotheses in a sample of Swedish siblings. Relative to older siblings, later born children are more likely to become unincorporated entrepreneurs, partly because they occupy family niches corresponding to lower educational attainment and higher labour market frictions. Moreover, children in very large families are less likely to pursue incorporation, with stronger effects for men and earlier born children, in line with their limited ability to develop self-efficacy due to resource dilution. Growing up with an opposite-gender sibling does not influence entrepreneurship, indicating that sibling sex composition may not foster stronger gender norms or sex-typing in this domain in Sweden. This study integrates existing findings and offers novel insights, broadening our understanding of family background in entrepreneurship.

KEYWORDS:

• Entrepreneurship
• incorporation
• family structure
• birth order
• family size
• sibling sex composition

Acknowledgments

I thank the editor (Ulla Hytti), the three anonymous reviewers, audiences at Copenhagen Business School, Max Planck Institute for Innovation and Competition, Yale SOM, Oxford Residence Week for Entrepreneurship Scholars, SEI Doctoral Consortium, DRUID, Academy of Management, Spanish Economic Association Meeting, and Mircea Epure, April Franco, Fabian Gaessler, Orsola Garofalo, Matthew Lindquist, Joeri Sol, Mirjam van Praag, Cesare Righi, Henry Sauermann, Markus Simeth, Olav Sorenson, and Friederike Welter for helpful comments. I also thank Kristian Koerselman for excellent technical assistance. I gratefully acknowledge financial support from the Swedish Research Council (Vetenskapsrådet 2017-01941) and the Spanish Agencia Estatal de Investigación (AEI), through the Severo Ochoa Programme for Centres of Excellence in R&D (Barcelona School of Economics CEX2019-000915-S). Part of this research was done while I was a Fox International Fellow at Yale University. This paper previously circulated as ‘Same, but Different? Birth Order, Family Size, and Sibling Sex Composition Effects in Entrepreneurship’ (SOFI Working Paper 8/2018) and ‘Intra-Family Differences in Entrepreneurship: Evidence from Swedish Siblings’.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08985626.2023.2243465

Notes

1. These entrepreneurs’ family background is described based only on publicly available information. Discussions of family structure in a public forum include Wadhwa et al. (2009) and Tech Crunch (2012), among others.

2. Lower-quality entrepreneurs use unlimited liability as an ex ante commitment not to default excessively on debt; higher-quality projects have larger future value and lower default incentives, so they enjoy limited liability, linking growth aspirations and incorporation beyond tax reasons (Edmark and Gordon 2013; Herranz, Krasa, and Villamil 2009). As riskier projects, incorporated firms are more likely to go bankrupt (Glover and Short 2018; Harhoff, Stahl, and Woywode 1998)

3. Not all mismatches lead to subsistence ventures, nor are all unincorporated firms strictly due to necessity.

4. Under primogeniture preference, business inheritance may favour firstborns. Children in Scandinavia become entrepreneurs in the same industry and around the time when parents exit in at most 8% of cases (Dahl and Sorenson 2012; Lindquist, Sol, and van Praag 2015; Sørensen 2007); 3% of U.S. nascent entrepreneurs enter via take-overs (Campbell and De Nardi 2009), 9% of Dutch entrepreneurs enter via family firm take-over (Parker and van Praag 2012), and less than 1.6% of children take over family firms in Canada and U.S (Aldrich, Renzulli, and Langton 1998; Fairlie and Robb 2007; Lentz and Laband 1990). Moreover, Gimenez-Jimenez et al. (2021) find no difference in family firm succession intentions for firstborns relative to later born children. Inheritance is thus unlikely to explain the effects.

5. For educational achievement, early descriptive evidence matches this pattern (Blake 1981), but recent work using instrumental variables produces ambiguous and noisy results (Åslund and Grönqvist 2010).

6. Fletcher and Kim (2018) find no differential effects for first and second born children’s personality; however, their analysis focuses on twin births only at the second and third birth order, and does not fix the rank of twin births and family size when comparing across birth orders, complicating their test. By contrast, Blake (1989) provides descriptive evidence that earlier born children’s education suffers most when families become very large; Mogstad and Wiswall’s, 2016 recent instrumental variable estimates also match this pattern.

7. By contrast, Cyron et al. (2017) find positive effects of sisters on men’s cognitive ability.

8. See Regeringens Proposition 2004/05:85 – Ny aktiebolagslag, page 212, available in Swedish at https://www.riksdagen.se/sv/dokument-lagar/dokument/proposition/ny-aktiebolagslag_GS0385 (accessed January 2023).

9. Where appropriate, I report Kleibergen-Paap F-statistics, adjusting for the family clusters (Cameron and Miller 2015). Bhalotra and Clarke (2019) recently show that twin births may be correlated with maternal attributes. To alleviate this concern, I control for mother’s birth year, education, and income as proxies of maternal health and later use alternative instruments based on the gender of the first two children, with similar results.

10. Under son preference (Dahl and Moretti 2008), the family continues to have children if the first child is a daughter, but despite this selection effect on family size, the second child’s gender is orthogonal to the first child’s gender (Peter et al. 2018). Online Appendix A shows covariate balance for firstborn children with opposite- and same-sex next siblings; as family size differs slightly between groups, I control for it in the regressions.

11. While the split sample analyses present the results transparently, some families cannot contribute identifying variation (e.g. those with two children – a boy and a girl). To maintain broad comparisons, I interact gender and birth order indicators in family fixed effects models: the results support the main conclusion (Online Appendix B).

12. Including indicators for birth order is thus preferable to a linear specification, although I include the latter in Online Appendix B; note that the hypotheses do not assume or require a priori linear effects.

13. Education is indeed negatively correlated with unincorporated entrepreneurship, especially for men, but has a curvilinear relationship with incorporation, increasing, then decreasing (Online Appendix A). Birth order effects in education and ability are negative and significant economically and statistically (Online Appendix B).

14. Building directly on the insights in this paper, Wang et al. (2022) also find that education partly explains the link from birth order to solo entrepreneurship in China, attributing remaining effects to unobserved risk taking.

15. In a sample of young U.S. adults, Fletcher and Kim (2018) find negative family size effects on personality traits – agreeableness, extroversion, and openness – likely correlated with self-efficacy.

16. Sibling spacing is related to sibling interaction intensity (Steelman et al. 2002), but this family influence appears less important for entrepreneurship, regardless of gender (Vladasel et al. 2021; Wang, Wang, and Mu 2022).

17. See, e.g. Blanchflower and Oswald (1998); Delmar and Davidsson (2000); Dunn and Holtz-Eakin (2000); Hout and Rosen (2000); Sørensen (2007); Laspita et al. (2012); Lindquist et al. (2015); Yu et al. (2023).

Additional information

Funding

The work was supported by the Vetenskapsrådet [2017-01941] and the Agencia Estatal de Investigación (CEX2019-000915-S).