Capital flows, real estate, and local cycles : evidence from German cities, banks, and firms

Capital flows and real estate are pro-cyclical, and real estate has a substantial weight in economies' income and wealth. In this paper, we study the role of real estate markets in the transmission of bank flow shocks to output growth across German cities. The empirical analysis relies on a new and unique matched data set at the city level and the bank-firm level. To measure bank flow shocks, we show that changes in sovereign spreads of Southern European countries (the so-called GIPS spread) can predict German cross-border bank flows. To achieve identification by geographic variation, in addition to a traditional supply-side variable, we use a novel instrument that exploits a policy assigning refugee immigrants to municipalities on an exogenous basis. We find that output growth responds more to bank flow shocks in cities that are more exposed to tightness in local real estate markets. We estimate that, during the 2009-2014 period, for every 100-basis point increase in the GIPS spread, the most exposed cities grow 15-2 basis points more than the least exposed ones. Moreover, the differential response of commercial property prices can explain most of this growth differential. When we unpack the transmission mechanism by using matched bank-firm-level data on credit, employment, capital expenditure and TFP, we find that firm real estate collateral as measured by tangible fixed assets plays a critical role. In particular, bank flow shocks increase the credit supply to firms and sectors with more real estate collateral. Higher credit supply then leads firms to hire and invest more, without evidence of capital misallocation.