Unpaid internships and equality of opportunity: a pseudo-panel analysis of UN data

One of the most pressing arguments against unpaid internships is that they impose a barrier to upward economic mobility. Intern experience leads to better employment outcomes, yet only individuals from relatively advantaged backgrounds can afford an unpaid stint of 3 to 6 months, resulting in unequal access for individuals of more modest means. This could be particularly relevant in an international organization context, where individuals from a broad spectrum of nationalities are employed. Using a unique intern dataset, I test this hypothesis in a pseudo-panel linear probability model of paid versus unpaid internships, conditional on two proxies for equality of opportunity: parents education and developing country origin. I find that having more highly educated parents leads to a 30% higher chance of taking an unpaid internship, confirming the unequal access hypothesis; however, more surprisingly, I find that individuals from developing countries are much more likely to take an unpaid internship (75% higher) than their developed country counterparts. These results are robust to several different specifications, yet contradict cross-section OLS estimates, suggesting that the influence of unobserved time-invariant heterogeneity is substantial.