Many Types of Human Capital and Many Roles in U.S. Growth: Evidence from County-level Educational Attainment Data

We utilize county-level data to explore the different roles of different types of human capital accumulation in U.S. growth determination. The data includes over 3,000 cross-sectional observations and 39 demographic control variables. The large number of observations provides enough degrees of freedom to obtain estimates for the U.S. as a whole and for 32 states in and of themselves. This data contains measures of educational attainment for four distinct categories: 1) 9 to 11 years; 2) High school diploma; 3) Some college; and 4) Bachelor degree or more. These variables represent human capital stocks for each and every county. This is a departure from much of the economic growth literature which has (at least in part) relied on extrapolation of stocks from flows, e.g. school enrollment data. We use a consistent two stage least squares estimation procedure. We find that: 1) The percentage of a county's population with less than a high-school education is negatively correlated with economic growth; 2) The percentage obtaining a high school diploma is positively correlated with growth; and 3) The percentage obtaining some college education has no clear relationship with economic growth, but 4) The percentage that obtains a bachelor degree or more is positively correlated with growth. Further, we find that: 5) There is significant qualitative heterogeneity in estimated coefficients across states for the 9 to 11 years and high school diploma categories, but 6) no qualitative heterogeneity for the college level categories. The most consistent conclusion across samples is that the percent of a county's population obtaining a bachelor degree or higher level of college education has a positive relationship with economic growth. Oddly enough, despite findings 2), 4), and 6) above, we find that the percentage of a county's population employed in educational services is negatively correlated with economic growth.