September 2022
ABSTRACT
Most entrepreneurs learn entrepreneurial skills while working as employees. Thus, employees’ opportunities to learn entrepreneurial skills are expected to substantially influence the economy’s entrepreneurship rate and entrepreneurial performance. However, previous studies have assessed such learning opportunities using indirect, rough measurements. This study utilizes a more direct and comprehensive measure. I first analyze entrepreneurs’ tasks by using information on 47 tasks from 31 countries and find that entrepreneurs perform more autonomous and diverse tasks, financial and managerial tasks, and fewer clerical tasks than employees. Next, using individual-level data from 23 countries in 2012–2017, I demonstrate that employees’ greater learning opportunities for entrepreneurial tasks increase individuals’ self-perceived entrepreneurial skills and their probability of becoming an entrepreneur. These relationships are generally robust to alternative learning opportunity indices, instrumental variable estimations, omitted variable bias, and learning-environment-level regressions. My bound estimates imply that a 1-standard-deviation increase in an employee’s opportunities to learn entrepreneurial tasks increases the probability of an individual becoming an entrepreneur by 0.5–1.5 percentage points, which is equivalent to 4%–12% of the mean entrepreneurship rate. By contrast, having more opportunities to learn entrepreneurial tasks does not result in a higher level of employment or innovativeness in entrepreneurs’ businesses.
Keywords: Entrepreneur, Entrepreneurial skills, Entrepreneurship, GEM, PIAAC
JEL classification: J24, L26, M13, M50
PDF available at http://hdl.handle.net/2344/00053486
Please note that discussion papers are works in various stages of progress and most have not been edited and proofread and may contain errors of fact or judgment. Revised versions of these papers may subsequently appear in more formal publication series. The views expressed in this publication are those of the author(s). The IDE does not guarantee the accuracy of the data included and accepts no responsibility for any consequences arising from its use.