Response to Welch (2020): Real Estate Collateral Does Affect Corporate Investment
21 Pages Posted: 8 Jun 2020
Date Written: May 12, 2020
This short note is a response to Welch (2020), who claims that our results in Chaney, Sraer and Thesmar (2012) are not robust. We show that none of his findings invalidate our results. Welch makes three major points. First, he correctly points out that our baseline specification uses a common scaling factor (lagged capital stock) for our dependent (investment) and independent (real estate collateral) variables, creating a mechanical correlation between left- and right-hand side variables. We show in this note that, while this point is formally correct, our results are robust to controlling for or removing entirely this mechanical correlation. Second, Welch correctly, stresses that real estate prices are serially correlated, so that identification of a real estate collateral channel is potentially complex. We show in this note that our results are robust to controlling for the serial correlation in real estate prices. Third, Welch correctly worries about the fact that real estate prices are driven not just by local shocks (MSA or State), but also by common shocks (national). We show in this note that our results are robust to controlling for common national real estate shocks. In other words, while we recognize that Welch raises several important points, we argue that none of those results invalidate the baseline findings in Chaney, Sraer and Thesmar (2012). Yet, some of these objections suggest interesting leads for further analysis on corporate investment. We describe these leads in the note, hoping that they will inspire future research.
Keywords: Corporate Investment, Collateral Channel
JEL Classification: G31, D22
Suggested Citation: Suggested Citation