Failures and FDIC losses for community banks during the banking crises of the late 1980s and late 2000s are compared. Despite increases in risky commercial real estate (CRE) lending and more severe economic shocks in the recent crisis, failure rates were lower. We find that other changes in bank characteristics, like higher capital, made community banks more resilient to shocks. In contrast, FDIC losses on failed banks were higher. These are not explained by changes in CRE exposure or economic shocks. We find that an interest-receivable variable is predictive of failures and FDIC losses. Implications for prompt corrective action are discussed.

We examine whether corporate financing activities (CFA) in aggregate convey information about the macroeconomy. Using statement of cash flow information to construct a bottom-up measure of CFA, we find that it has significant predictive power for future economic activity when we exclude a small set of firms whose external financing is largely insulated from macroeconomic conditions. This CFA index has predictive power beyond that of the Gilchrist-Zakrajsek credit spread, aggregate earnings, and other macroeconomic indicators in predicting future GDP in both in-sample and out-of-sample forecasting tests. Impulse responses from a structural vector autoregression show that unexpected decreases in this CFA index lead to a large and persistent contraction in economic activity for up to four quarters. Our results suggest that a simple portfolio-based CFA measure helps capture supply-of-capital effects from the financial accelerator mechanism and hence has significant incremental predictive power for real economic activity. Data Availability: Data are available from the public sources cited in the text.