Skip to main content

Advertisement

SpringerLink
Log in

Does education enhance entrepreneurship?

  • Published:
Small Business Economics Aims and scope Submit manuscript

Abstract

Formal education is correlated with entrepreneurial activity and success, but correlation does not indicate causation. Education and entrepreneurship are both influenced by other related factors. The current study estimates the causal effects of formal education on entrepreneurship outcomes by instrumenting for an individual’s years of schooling using cohort mean years of maternal schooling observed decades prior. We differentiate self-employment by industry employment growth and firm incorporation status. Policymakers are especially interested in entrepreneurship with the potential to create substantial employment growth. We find that an additional year of schooling increases self-employment in high-growth industries by 1.12 percentage points for women and by 0.88 percentage points for men. Education reduces the probability of male self-employment in shrinking industries. Education also increases incorporated self-employment for women and men and reduces unincorporated self-employment among men but not women. The overall probability of self-employment increases with education for women but is unaffected by education for men. The results suggest that formal education enhances entrepreneurship.

Plain English Summary

Education has long been correlated with entrepreneurial success; we provide the first evidence for a causal effect of education on high-growth industry entrepreneurship in the USA. Education also reduces self-employment in shrinking industries. Our analysis leverages differing education trends across states and ancestry groups to predict education levels for recent workers and provide causal estimates of the effect of education on entrepreneurship. We account for numerous other factors and conduct several tests to confirm that our results are stable across alternatives. Our study has important implications for researchers, policymakers, and broader society. The benefits of education are widely debated, and some worry that education is mostly about signaling and not skill development. Policymakers are also interested in how to facilitate innovative entrepreneurship that creates new products, new jobs, and economic prosperity. Our analysis indicates that education confers valuable knowledge and skills that enhance entrepreneurship and fuel societal well-being.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Notes

  1. On the importance of human capital, see also Glaeser et al. (1995), Simon (1998), Simon and Nardinelli (2002), Moretti (2004, 2013), Shapiro (2006), Hanushek (2013), Winters (2013, 2014, 2018), Hanushek and Woessmann (2015), Hanushek et al. (2017), and Ehrlich et al. (2018). Additional studies on the importance of entrepreneurship include Acs and Storey (2004), Wennekers et al. (2005), Acs (2006), Baumol and Strom (2007), Van Praag and Versloot (2007), Stephens and Partridge (2011), Stephens et al. (2013), and Glaeser et al. (2015).

  2. Surveys of related empirical literature are provided by Le (1999), Van der Sluis et al. (2008), Unger et al. (2011), Marvel et al. (2016), Simoes et al. (2016), Parker (2018), and Hogendoorn et al. (2019).

  3. Our framework is intentionally simplified and is not the only possible model that could be presented. For example, our model has partial conceptual overlap with some others including the Roy (1951) selection model that focuses on how differential skills affects observed earnings in different occupations via occupational self-selection. Our model is chosen to focus on the effect of education on an individual’s decision to enter entrepreneurship rather than the effect of education on entrepreneurial earnings. Our empirical analysis uses instrumental variables to estimate causal effects of education on the decision to enter entrepreneurship.

  4. Our model assumes that there is no differential schooling requirement for paid-employment vs. entrepreneurship. This assumption is generally true in the USA during our time period, but there are some modest exceptions such as real estate agents/brokers. Other countries can have especially higher educational requirements for entry into entrepreneurship (Rostam-Afschar, 2014) that raise the cost of entering entrepreneurship and discourage some potential entrepreneurs. Rostam-Afschar (2014) finds that lowering education requirements for entrepreneurship among craftsmen in Germany increased self-employment.

  5. The idiosyncratic cost of venturing depends on a number of factors including personality, risk aversion, impulsivity, access to capital, exposure to other entrepreneurs, developed skills, and innate ability (Blanchflower & Oswald, 1998; Lazear, 2005; Åstebro & Thompson, 2011; Guiso & Schivardi, 2011; Wang, 2012; Lindquist et al., 2015; Orazem et al., 2015; Wiklund et al., 2017; Levine & Rubinstein, 2018; Hamilton et al., 2019; Hvide & Oyer, 2019; Guiso et al., 2021). A notable literature has also documented non-monetary benefits as important factors influencing decisions to enter and persist in self-employment (Hamilton, 2000; Hurst & Pugsley, 2011; Acs et al., 2016). In particular, entrepreneurs have greater ability to control their own work conditions and schedule and many people value the freedom and control of being their own boss. Our intentionally simplified model does not explicitly incorporate non-monetary benefits, but one could think of them as reducing the idiosyncratic cost of venturing.

  6. Specifically, the owner’s personal assets are not generally recoverable for debts of the corporation, except in cases of fraud or when the owner has explicitly assumed personal responsibility or used personal assets as collateral.

  7. The Current Population Survey (CPS) has largely similar information except it does not include birth state, which prevents one from using the CPS to conduct a similar analysis. As discussed below, our identification strategy relies on linking cohort level maternal education via birth state, birth year, and ancestry group from the 1980 and 1990 decennial censuses to the 2006–2019 ACS.

  8. We have no information on biological parents living outside the child’s household.

  9. The decennial censuses and ACS are independent random samples, so we are generally not observing the same individuals over time. A small percentage of individuals are included multiple times, but we have no way to identify or link them over time. With 50 states, 28 birth-years, and 12 ancestry groups, there are 16,800 potential unique combinations of birth state, birth year, and ancestry group, but 20 cells are empty in either the census or ACS data and we actually observe 16,780 unique cohort combinations in the merged data. While a few cells are relatively small, most are large. In the 1980 and 1990 Census data, the median cohort size is 1260 child observations, and more than 90 percent of cohorts have at least 100 child observations. Large cohorts allow us to compute reasonably precise averages for cohort parental characteristics.

  10. Survey age and survey year effects are included based on expectation that they explain some portion of the variation in self-employment and improve efficiency. However, survey age and survey year effects are minimally important for our 2SLS identification strategy discussed below because they do not significantly and systematically vary across ancestry groups within birth-state × birth-year or across birth-states within ancestry × birth-year. We present robustness checks in the appendix showing that our main results are robust to excluding survey age and survey year effects.

  11. We link industries over time using the IND1990 variable in IPUMS. See Table 6 of the Appendix for more details.

  12. We also report 2019 employment levels and 2006–2019 percentage growth for these. We also computed the correlation between initial employment levels and employment growth by industry; the correlation is 0.19, which weakly suggests that larger industries grow moderately faster than initially smaller ones during this period.

  13. Table 7 of the Appendix reports mean years of individual schooling (from the ACS) and mean cohort maternal schooling (from the 1980 and 1990 censuses) by ancestry group along with means for self-employment (in any industry), any employment, and annual earnings. The table shows that there are differences in mean schooling and other outcomes across ancestry groups.

  14. We use the terms shrinking and negative growth interchangeably.

  15. Results for the parental characteristic control variables in the 2SLS specification are reported in Table 8 of the Appendix. These are not our focus. Most of these coefficients are not significant. The main results are qualitatively robust to excluding the parental characteristic control variables.

  16. For all 2SLS models in Table 2, less than one percent of predicted values are below zero and none exceeded one.

  17. Some of the coefficient estimates are very close to zero (and not statistically significant), especially OLS results in column (6) and 2SLS results in column (8) of Table 2. These very small coefficients reflect a combination of factors including that the dependent variable means are relatively small (see Table 1) and that the true effects are likely zero for these particular relationships. The OLS results also include some very small standard errors, which also reflect the large sample sizes.

  18. The ivreghdfe program uses an endogeneity test from ivreg2 (Baum et al., 2002), which is defined as the difference of two Sargan-Hansen statistics: one that treats individual schooling as exogenous and one that does not.

  19. Individuals with very low reported schooling levels may be influenced by mental or physical impairments or other unique circumstances that we cannot observe or account for. Individuals with very high levels of schooling may also be outliers in many respects, including a willingness to incur high opportunity costs via foregone earnings while pursuing education. Ideally, our results should not be driven by individuals in the tails of the education distribution.

  20. A few previous studies have used instrumental variables to examine effects of education on measures of entrepreneurial success including entrepreneurial income and typically find positive coefficients, but these also do not fully account for selection into self-employment (Parker & Van Praag, 2006; Iversen et al., 2011; Block et al., 2012; Fossen & Büttner, 2013; Van Praag et al., 2013; Kolstad & Wiig, 2015).

References 

Download references

Author information

Authors and Affiliations

  1. Department of Economics, Iowa State University, 273 Heady Hall, 518 Farm House Lane, Ames, IA, 50011-1054, USA

    Kunwon Ahn

  2. Department of Economics, Iowa State University, 460B Heady Hall, 518 Farm House Lane, Ames, IA, 50011-1054, USA

    John V. Winters

  3. Center for Agricultural and Rural Development (CARD), Ames, USA

    John V. Winters

  4. Program for the Study of Midwest Markets and Entrepreneurship (PSMME), Ames, USA

    John V. Winters

  5. Global Labor Organization (GLO), Essen, Germany

    John V. Winters

  6. Institute of Labor Economics (IZA), Bonn, Germany

    John V. Winters

Authors
  1. Kunwon Ahn
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. John V. Winters
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to John V. Winters.

Ethics declarations

Competing Interest

Not applicable.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 11.

Table 12.

Table 13.

Table 14.

Table 6 List of High-Growth Industries and IPUMS IND1990 Codes
Full size table
Table 7 Analytical sample means for selected variables by ancestry group
Full size table
Table 8 Parental characteristic variable 2sls results for self-employment and industry growth by gender
Full size table
Table 9 Parental characteristic variable 2SLS results for incorporated and unincorporated self-employment
Full size table
Table 10 Additional analysis for 2SLS effects of schooling on self-employment and industry growth
Full size table
Table 11 2SLS effects of schooling on alternative industry self-employment measures
Full size table
Table 12 Effects of schooling on paid and non-employment by gender
Full size table
Table 13 Summary of 2SLS estimates for self-employment, paid-employment, and non-employment
Full size table
Table 14 Effects of Schooling on Log Annual Earnings of the Self-Employed
Full size table

Rights and permissions

Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ahn, K., Winters, J.V. Does education enhance entrepreneurship?. Small Bus Econ 61, 717–743 (2023). https://doi.org/10.1007/s11187-022-00701-x

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11187-022-00701-x

Keywords

JEL Classification

Search

Navigation