Knowledge Spillovers from Research

We estimate the local spillovers from research university activity in a sample of urban counties. We use the fact that universities tend to follow a rigid endowment spending policy based on the market value of their endowments to identify the causal effect of university activity on labor income in the non-education sector. Our instrument for university expenditures is based on the interaction between each university’s lagged endowment level and the variation in stock market shocks over time. We find statistically significant spillover effects from university activity, and the magnitude of the spillover is significantly larger when local universities are more research intensive or when firms are technologically closer to universities, as measured by labor market pooling and patent citations. The findings provide a rationale for place-based university policies, so long as they focus on industry fundamentals. The results also suggest that the longer-term effects that universities have on their local economies may grow over time as the composition of local industries evolve to take advantage of the knowledge spillovers we identify. We thank the TIAA-CREF Institute Research Grant Program and NSF grant SES-0851788 for financial support. We are grateful to Julie Cullen, Martin Dooley, Enrico Moretti, and three anonymous referees for helpful comments and discussions. Seminar participants at Chapman University, Rensselaer Polytechnic Institute, UC Berkeley, UC Davis, and the University of Hawaii provided many useful suggestions and insights. We also thank Michael Greenstone, Richard Hornbeck and Enrico Moretti for providing their CPS employment transitions data and Ken Redd of the National Association of College and University Business Officers for data on endowment portfolio holdings. Justin Hicks and Nolan Noble provided outstanding research assistance. The views expressed herein are those of the authors and do not necessarily reflect the views of TIAA-CREF or the NSF. 2 Knowledge Spillovers from Research Universities: Evidence from Endowment Value Shocks