The Federal Reserve’s off-site surveillance system includes two econometric models that are collectively known as the System for Estimating Examination Ratings (SEER). One model, the SEER risk rank model, uses the latest financial statements to estimate the probability that each Fed-supervised bank will fail in the next two years. The other component, the SEER rating model, uses the latest financial statements to produce a “shadow” CAMELS rating for each supervised bank. Banks identified as risky by either model receive closer supervisory scrutiny than other state-member banks. Because many of the banks flagged by the SEER models have already tumbled into poor condition and, hence, would already be receiving considerable supervisory attention, we developed an alternate model to identify safe-and-sound banks that potentially are headed for financial distress. Such a model could help supervisors allocate scarce on- and off-site resources by pointing out banks not currently under scrutiny that need watching. It is possible, however, that our alternative model improves little over the current SEER framework. All three models—the SEER risk rank model, the SEER rating model, and our downgrade model—produce ordinal rankings based on overall risk. If the financial factors that explain CAMELS downgrades differ little from the financial factors that explain failures of CAMELS ratings, then all three models will produce similar risk ratings and, hence, similar watch lists of one- and two-rated banks. We find only slight differences in the ability of the three models to spot emerging financial distress among safe-and-sound banks. In out-of-sample tests for 1992 through 1998, the watch lists produced by the downgrade model outperform the watch lists produced by the SEER models by only a small margin. We conclude that, in relatively tranquil banking environments like the 1990s, a downgrade model adds little value in off-site surveillance. We caution, however, that a downgrade model might provide useful in more turbulent bank times.
Due to the size and confidentially of the data, questions regarding data and programs may be directed to Andrew P. Meyer, email@example.com. Data on CAMEL ratings are confidential and will not be released.