Can feedback from the jumbo-CD market improve off-site surveillance of community banks?

We examine the value of feedback from the jumbo-certificate-of-deposit (CD) market in the off-site surveillance of community banks. Using accounting data, we construct proxies for default premiums on jumbo CDs. Then, we produce rank orderings of community banks -- defined as institutions holding less than $500 million in assets (constant 1999 dollars) -- based on these proxies. Next, we use an econometric surveillance model to generate rank orderings based on the probability of encountering financial distress. Finally, we compare these rank orderings as tools for flagging emerging problems. Our comparisons include eight out-of-sample test windows during the 1990s. We find that feedback from the jumbo-CD market would have added little value in community-bank surveillance during our sample period. Specifically, rank orderings based on output from the econometric model significantly outperformed rank orderings based on jumbo-CD default premiums. More important, the jumbo-CD orderings improved little over a random ordering. Other attempts to extract risk signals from the jumbo-CD data yielded similar results. Taken together, our findings validate current surveillance practices. We conclude by arguing that the robust economic environment of the 1990s probably plays a large role in our results.