Impact of earnings and self-employment opportunities on overeducation: evidence from occupations in the United States labor market 2002-2016

ABSTRACT

This is an applied econometric analysis of labour market data for the United States. We study the impact of several factors on overflow of overeducated employees into various job categories. We use panel data regression analysis with fixed and random effects. We also use data visualisation to investigate the overeducation trends during 2002–2016 for various occupation categories. Our dataset consists of seven sets of annual data for 704 occupations. We investigate this phenomenon at two levels: 1) overflow of university graduates into occupations that do not require a university degree, and 2) overflow of Masters and PhD degree holders into occupations that require a bachelor’s degree or less. We observe that the overeducation has increased in most occupations and it causes a crowding out effect; an adequately educated worker might be outcompeted by an overeducated worker. While the income premium of a university education has decreased over time, the income advantage of university education over a high school degree has persisted. Furthermore, our regression analysis has shown that the overeducation ratio has a positive correlation with the median earnings of an occupation and its opportunities for self-employment. The results hold for both college graduates and holders of graduate degrees.

KEYWORDS:

• Higher education
• overeducation
• self-employment
• United States labour market
• credentialism
• crowding out

Acknowledgments

We like to thank Joshua Goodman, Daniel Bergstresser, Judy Dean, Blake LeBaron, Anne Laski, and Steve Sass for their valuable feedback and suggestions. We are also grateful to Andrew O’Bar for assistance with access to the BLS data for this project.

Disclosure statement

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

Availability of data and material

The authors have used open access data from the Bureau of Labour Statistics as our raw data. Our final dataset is available at https://doi.org/10.7910/DVN/VO4ANT.

Notes

1. In this article, non-college occupation refers to occupations for which a university degree is not required. In other words, these are the occupations for which the job advertisements do not require the applicants to have a university degree. We use the term college and university interchangeably in this article. Hence, a college degree is equivalent to a 4-year undergraduate university degree.

2. In most empirical studies of overeducation the dataset consists of panel or longitudinal data on earnings and economic activities of University graduates. For a survey of this literature see Leuven and Oosterbeek (2011).

3. The Bureau of Labour Statistics reports the appropriate education level for each occupation based on expert evaluation of the required knowledge and skills.

4. A detailed review of the literature on overeducation that uses this approach is available in Capsada‐Munsech (2017).

5. The pathbreaking earliest studies on the US labour market are Berg (1970) and Freeman (1976). These were followed by a large number of studies in the 1980s and 1990s. The social and political consequences of overeducation in modern market economies have been discussed in Burris (1983) and Vaisey (2006).

6. The wage penalty is measured as the gap between the wage of an individual in an occupation for which she is overeducated and the potential wage that she would have earned in an occupation that matched her education.

7. Data for the most recent year is available at: https://www.bls.gov/emp/tables.htm. We obtained the data for earlier years by email from the Employment Projections Office of the BLS.

8. We were not able to obtain data for 2010.

9. See https://www.bls.gov/soc/ for details.

10. The few occupational codes that were created or deleted, whether from old codes being lumped together or old codes being refined into new categories, were ignored.

11. The education requirements for occupations in the US labour market are updated and reported annually by BLS in www.onetonline.org, commonly known as O*NET. Using an expert-determined education requirement for each occupation such as this source enables us to construct an objective and unbiased measurement of overeducation and it has a long precedent in the United States. Originally it was introduced by Eckaus (1964) and later it was also adopted by Rumberger (1981) and Burris (1983). Most recently O*NET data on education requirements was used by Arvan et al. (2019).

12. As reported in the BLS 2010 Standard Occupational Classification System, https://www.bls.gov/soc/2010/2010_major_groups.htm.

13. For a detailed analysis of the self-employment trends in the United States see Hipple (2010).

14. Note that the BLS began reporting the number of workers with graduate degrees separately (instead of one measurement for bachelor’s or higher) in 2008. Hence, our analysis of this segment of the data is limited to 2008–2016.

15. Verhaest et al. (2018) page 1.

16. With the exception of talented individuals that select a low-income occupation, such as art or social work, because of personal passion, as observed by Rose (2017).

17. While there is extensive literature on crowding out of high school graduates by university graduates, only a small number of articles such as Vedder, Denhart, and Robe (2013) and Ben-David (2009) look at the role of income and other characteristics of an occupation on its crowding-out ratio.

18. While our dataset includes detailed data about the educational attainment of employees in each occupation, it does not include other important factors such as race/ethnicity and gender distribution of the employees in each occupation.

19. Detailed tables of these regression results (in STATA log format) are available upon request from the authors.

20. Note that the coefficients of the self-employment ratio are negative in the quadratic models (S8 and S9 in Table 4). This result, however, is consistent with a positive correlation because the values of self-employment are standardised (mean = 0) and the observations that are included in the regression are all negative (since they are all below the median of the entire observations). Hence the negative coefficients indicate a positive marginal effect. We also observe similar results for the employment size variable in these two sets of models (no significance for the sample with self-employment ratio above the median and positive correlation for the sample below the median).

21. For our total sample we observe a negative correlation between these two variables.

22. The BLS data for 2002, 2004, and 2006 that we were able to obtain did not report the distribution of educational attainment separately for Bachelor’s, Master’s and PhD degrees. Instead a single statistic was provided for the percentage of employees with Bachelor’s or advanced degrees.

23. Our analysis revealed that the employment size of an occupation also had a positive correlation with the overeducation ratio, but the results were sensitive to sample selection. Hence we cannot draw definitive conclusions about the effects of occupation size.

24. The significance of passion in occupation selection has been studied by Rose (2017). Evaluating the occupation of men with at least a Bachelor of Arts (BA) degree, he found that 10% of men with a BA degree were working voluntarily in low-paying professional occupations because of their passion. This category included school teachers, artists, and social workers.

25. Blázquez and Budría (2012) offer a comprehensive analysis of how differences in personality are correlated with likelihood of being overeducated.