Working Paper: Social Influences on Human Capital Investment: Evidence from a Continuing Education Program in the US Army

Acknowledgements: The views expressed in this paper are those of the author and do not purport to reflect the positions of the U.S. Military Academy, the Department of the Army, or the Department of Defense. I am grateful to my advisor, Ben Castleman, for advice and encouragement at every stage of this research. David Lyle, Amalia Miller, Luke Miller, Bruce Sacerdote, Bill Skimmyhorn, Sarah Turner, Carl Wojtaszek, Jim Wyckoff, and seminar participants at AEFP provided helpful comments and suggestions. I thank the Office of Economic and Manpower Analysis (OEMA) and the Army Continuing Education System (ACES) for providing the data for this analysis. Errors in the paper are my own. Please contact me at [email protected]. Human capital investment represents a complex and far-reaching individual decision that may be influenced by the educational choices made by others, yet we know far less about peer effects at this extensive margin than we do at the intensive margin of education production itself. In this paper, I rely on a unique source of exogenous variation in which individuals randomly receive exposure to different levels of peer investment in human capital and then make their own education participation decisions. Specifically, I study new US Army soldiers who are randomly assigned to companies that vary substantially in their existing participation rates in a subsidized continuing education program. I find that a new soldier assigned to a high-participation company is far more likely to take classes than a soldier assigned to a low-participation company. Building on prior work examining neighborhood and peer effects, I decompose this overall impact into neighborhood, leadership, and peer influences. The decomposition suggests that differences across Army locations and other common shocks are largely responsible for the impacts I observe, though I also find a modest peer effect on participation. EdPolicyWorks