The Causal Impact of Human Capital on R&D and Productivity: Evidence from the United States

This paper contributes to the empirical literature on the impact of human capital on technology adoption and the production structure of the economy by using census micro data aggregated at the state level data for US cohorts born between 1915 and 1939. We test the impact of secondary and tertiary schooling in the US at the state-cohort level on R&D and TFP growth across industries in 1970. While we follow the literature in using the variation in the timing of compulsory schooling laws across states to instrument secondary schooling, we propose a novel instrument for tertiary enrollment. In particular, we exploit, as in Acemoglu, Autor and Lyle (2004), the differences across states and cohorts in World War II mobilization rates. While Acemoglu, Autor, and Lyle (2004) used this variation as an exogenous shift in female labor supply, we exploit the fact that WWII veterans were benefited by the GI Bill Act (1944), which granted them free college education once they were discharged from service. This provides a clean source of variation in the costs of attending college, which allows us to exploit differences in college enrollment across states and cohorts. Our results suggest that, consistent with the initial discussion, different types of human capital are associated to different effects on the productive structure of the economy. Two-stage least squared regressions find no effect of the share of population with secondary schooling over outcomes such as R&D per worker or TFP growth. On the other hand, the share of population with tertiary education has a significant effect on both R&D per worker or TFP growth. In particular, a 1% increase in the share of workers with tertiary education increases R&D per worker by 1.8 percentage points, and annual TFP growth by 1% for 17 years. Creation-Date: 2015