In the mood for a loan: The causal effect of sentiment on credit origination

Wednesday, 15 October 2014: 9:00 AM
Douglas Evanoff, Ph.D , Economic Research, Federal Reserve Bank of Chicago, Chicago, IL
One of the recent trends in financial markets has been an increasing replacement of human agents with automatic algorithms in a wide spectrum of financial decisions, ranging from stock trading and market making to credit card approvals. The advocates of this shift argue that it helps remove human emotions from financial decisions and protects firms from possible judgment errors. If human emotions indeed matter in financial decisions, they can also have potentially important economic consequences that extend beyond a single firm.  For example, correlated changes in mood may generate cascading effects such as a market panic or a bank run, which affect entire regions and often invoke a policy response. Despite the potential significance of these effects, we know relatively little about how human emotions affect corporate financial decisions, whether they improve or impair judgment, and what consequences they have on real economic activity. Our paper seeks to provide evidence in this direction by examining the role of emotions in one of the fundamental economic decisions – credit origination.

We study how loan officers’ mood affects their decisions on mortgage applications. Motivated by psychological evidence, we use three mood proxies: (1) outcomes of key sporting events such as the Super Bowl, (2) outcomes of the American Idol competition, and (3) days around major holidays. Our identification exploits the variation in daily loan approvals in each county, while observing the volume and quality of reviewed applications. Positive sentiment events are associated with a significantly higher loan approval rate in affected counties, and negative sentiment events have the opposite effect of a smaller magnitude. The effect is stronger for marginal quality applications, where loan officers have more discretion. The extra loans approved on high-sentiment days experience higher defaults. Overall, our evidence suggests that mood fluctuations affect decisions of financial experts and generate long-lasting real effects.