No Collateral Channel: How Real Estate Shocks Do Not Affect Corporate Investment
47 Pages Posted: 4 Jun 2020 Last revised: 29 Jun 2020
Date Written: June 27, 2020
My paper shows there is no association between real-estate prices and corporate investment, as claimed by Chaney, Sraer, and Thesmar (2012). First, most of their coefficients are mechanical. They explain changes in normalization (lagged PP&E) used in both their independent variable (real-estate holdings) and their dependent variable (corporate capital expenditures). This is not cured by an additive control. Appropriately purged of shared normalization changes, the estimated point coefficients turn negative. Second, their real-estate prices are persistent. They cannot identify the timing of real-estate “shocks.” Third, firms without real-estate started out observably different. They were smaller and had higher initial capital expenditures (subsequently declining). Merely allowing firms without real-estate a different average investment trend is enough to render information about their actual real-estate holding values, real-estate prices, real-estate price changes, or real-estate shocks unimportant even in their best specification. An extended analysis shows no real-estate collateral channel during and after the Great Recession of 2008, either.
Keywords: collateral, real-estate, corporate investment
JEL Classification: D22 G31 R30
Suggested Citation: Suggested Citation