Credit Fluctuations and 'Neglected Crash Risk' in U.S. Bank Returns

Abstract

Using U.S. quarterly data from 1960, the paper studies the interaction between bank stock returns and aggregate credit fluctuations on a set of economic dimensions. First, I investigate the source of Neglected Crash Risk in U.S. bank returns using a new deviation measure of aggregate loans per capita called ltd. A one standard deviation increase in ltd decreases bank stock returns by 5%, and their dividend growth by almost 6% over the following year. This variable embeds important information about both future returns (discount rate news) and cash flow growth (dividend news); yet a decomposition of future unexpected bank returns shows a higher incremental effect associated with the variance of news about the discount rate. Second, I quantify the size of Neglected Crash Risk and find it is economically and statistically significant at different forecast horizons and for different degrees of credit boom intensity. I interpret its size as the outcome of differences in perceived disaster probabilities by investors across credit regimes. Lastly, changes in monetary policy activate crash risk in the perception of commercial bank investors. A straightforward application of the local-projection methods of Jorda (2005) illustrates that a contractionary-type monetary policy action decreases bank returns more when the economy had experienced credit booms in the past.