Seeing is Believing: The Impact of Local Economic Conditions on Firm Expectations, Employment and Investment

I show that managers overweight observations of local economic conditions at firm headquarters (HQ) when forming their macroeconomic expectations. This implies that HQ local economic conditions have an excessive impact on firm investment and employment growth. Using an empirical strategy identifying the impact of local economic conditions at HQ on employment outside the HQ, I find that a 1 percentage point (p.p.) higher local unemployment rate at HQ leads to 2 p.p. lower employment growth at non-HQ establishments. I consider a number of alternative explanations such as internal capital markets reallocation or local financing, and rule these out using placebo tests and by testing the key implications of the explanations. Then, I present evidence that HQ local conditions are overweighted in managers’ expectations. Worse HQ local conditions lead to more pessimistic sales forecasts and more negative macroeconomic sentiment. These findings, along with results from tests comparing firms with different sensitivities to the macroeconomic cycle, support the notion that local economic conditions bias managers’ macroeconomic expectations. Finally, I show that this bias can explain differences in county economic outcomes and may lead to significant investment misallocation. ∗Harvard University, Department of Economics. Email: [email protected]. I am indebted to Jeremy Stein, Andrei Shleifer, Sam Hanson and David Scharfstein for their encouragement, guidance, and support. I thank Malcolm Baker, Laura Blattner, John Campbell, Gabriel Chodorow-Reich, David Choi, Lauren Cohen, Benjamin Friedman, Xavier Gabaix, Robin Greenwood, Larry Katz, Nihar Shah, Larry Summers, Adi Sunderam, David Thesmar, Peter Tu, Yuhai Xuan, and seminar participants at Harvard finance and macroeconomics lunch for helpful discussions and suggestions. I thank Jim Davis for assistance with the Census microdata. Any opinions and conclusions expressed herein are those of the author and do not necessarily represent the views of the U.S. Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed.