Is the allocation of human capital between entrepreneurs and workers a key determinant of aggregate productivity and income? How pervasive are its implications for macro-development? Analyzing international survey data unveils a new empirical fact: there is a strong positive link among the adjusted rate of entrepreneurship for higher educated individuals and output per worker or total factor productivity (TFP). Further focus on the U.S. reveals an asymmetric U-shaped pattern between schooling and selection into entrepreneurship, while average business outcomes are strictly increasing for more educated firm owners/managers. I rationalize these findings in a versatile heterogeneous agent model with occupational and educational choices. Under the hypothesis that entrepreneurial human capital may enhance productive capacities via costly technology adoption, the entrepreneurship–education nexus has first-order aggregate and distributional consequences. Quantitative explorations suggest sizeable and persistent misallocation losses due to inadequate complementarity between idiosyncratic talent and human capital. This novel channel can often account for a major share of cross-country income differences vis-à-vis the U.S., as it drastically affects both factor accumulation and endogenous TFP formation. 

Is promoting entrepreneurship always conducive to long-run growth? To what extent should policymakers strive to ensure entrepreneurial risk away? We explore these questions in an endogenous growth model with occupational choice, where individuals are heterogeneous in their risk attitude and entrepreneurial ability. Less risk-averse and sufficiently productive agents become entrepreneurs and contribute to growth by expanding product variety. More risk-averse agents become workers and contribute to growth by enhancing capital formation. As risk tolerance leads to misallocation on the intensive margin, encouraging entrepreneurship may hinder aggregate productivity. The interplay of these forces results in a non-monotone relationship between the rate of entrepreneurship and balanced growth. Decentralized equilibria entail constrained-suboptimal allocations with either too few or too many active producers, even in the absence of financial frictions. Ensuring some entrepreneurial risk away is almost always growth-enhancing, but it is never optimal to provide full insurance.

Using international survey data I document that active entrepreneurs are substantially more educated relative to workers in richer than in poorer countries. I integrate this insight into a heterogeneous agent economy with occupational choice and entrepreneurial human capital as a factor of production. In an extended development accounting exercise, the average share of cross-country income differences due to total factor productivity–a measure of our ignorance–drops by 20-30% relative to the standard approach. Results highlight a distinct pathway along which human capital can emerge as the predominant source of discrepancies in output per worker, and complement alternative explanations such as quality-adjusted education and country-specific returns to experience.