* Timothy Curry and Gary Fissel are senior financial economists in the Division of Insurance and Research of the Federal Deposit Insurance Corporation (FDIC). Peter Elmer was a senior economist in the FDIC’s Division of Research and Statistics when this work was being conducted. The authors would like to thank John O’Keefe, Daniel Nuxoll, Gerald Hanweck, and Richard Bogue for helpful comments, and Audrey Clement and Justin Combs for extensive research assistance.
1 This note focuses on the academic literature. Flannery (1998) summarizes this literature through the late 1990s. More recently, Berger and Davies (1998) use event-study methodology to find that the equity market anticipates upgrades in regulatory ratings but follows downgrades. Berger, Davies, and Flannery (2000) find that regulators acquire information sooner than the equity markets and bond rating agencies do, but the regulatory assessments are generally less accurate than either stock or bond market indicators in predicting the future performance of bank holding companies. Elmer and Fissel (2000) find that equity market variables can be used to augment accounting-related information to predict bank failure. Krainer and Lopez (2001) find that equity market variables such as stock returns and equity-based default frequencies are useful to bank regulators for assessing the condition of bank holding companies. Gunther, Levonian, and Moore (2001) find that a measure of financial viability based on stock prices (expected default frequency) helps predict the financial condition of bank holding companies as reflected in their supervisory ratings. Curry, Elmer, and Fissel (2001) find that incorporating market data into traditional off-site monitoring models helps identify downgraded and upgraded banks and thrifts that were not affiliated with multibank holding companies. Curry, Fissel and Hanweck (2003) find that market-indicator variables add value to models in predicting bank holding company supervisory ratings.
2 Tanoue (2001); Greenspan (1998); Meyer (1998). The term market discipline generally refers to the ability of the market to price or impose costs on institutions based on their risk. The costs, for example, might take the form of higher issuance costs in the bond markets and/or lower equity prices.
3 The three pillars include minimum capital requirements, supervisory review and market discipline.
4 Federal law mandates that all federally insured banking institutions be examined at least every 12 to 18 months, depending on the size and condition of the institution. Weaker institutions are often subject to more frequent scrutiny. For evidence that bank examinations may age quickly, see Cole and Gunther (1998).
5 Levonian (2001) has shown that equity market information and debt market information should produce similar results.
6 The acronym “CAMEL” stands for Capital, Assets, Management, Earnings, and Liquidity, five components of a bank’s financial operation that are examined by the regulators. In the late 1990s a sixth component was added to the CAMEL rating system, recognizing bank and thrift Sensitivity to interest-rate or market risk (CAMELS). CAMELS ratings are assigned on a scale of 1 to 5 with 1 being the highest and 5 the lowest. Because the empirical portions of our analysis relate to ratings assigned before the late 1990s, we reference the five-component rating system in effect at that time.
7 It should be noted that for the largest U.S. banks, in recent years the Comptroller of the Currency and other regulators (including the FDIC) have established supervisory programs with continuous on-site presence.
8 Informal enforcement actions may require institutions to make changes, such as raising new equity capital, limiting the origination of certain types of loans, or increasing loan-loss reserves. Although regulators vary in their practices, the most common type of informal action accompanying a downgrade to 3 is a “memorandum of understanding” (MOU), which is written by bank supervisors and signed by bank officials and supervisors.
9 The Financial Institutions Reform, Recovery, and Enforcement Act of 1989 (FIRREA) mandated that formal enforcement actions become part of the public record.
10 The sample population was drawn from a universe of all banks and thrifts from 1988 to 1995 that were publicly traded, as reflected in the availability of stock price information from the Center for Research in Security Prices (CRSP). To obtain stock price information for individual commercial banks and thrifts, we matched CRSP data against bank quarterly reports going back to 1986. We then matched the firms against bank examination ratings to obtain the historical CAMEL ratings. Within the group for which all this information was available, we identified all institutions that were downgraded to a 3, 4, or 5 during our period. To form the sample in our study, we reduced this group by the additional restrictions discussed in the next paragraph of the text. The sample of CAMEL 1- or
2-rated, or “healthy,” banks against which the downgraded groups were matched was also taken from this universe of publicly traded institutions (see note 14).
11 Analysis of multibank holding company equity securities carries disadvantages (as well as advantages) compared with analysis of non-affiliated banks and thrifts and one-bank holding companies. For example, multibank holding companies tend to be large institutions that are widely traded and rated by nationally recognized rating agencies. Although one-bank holding companies and banks not affiliated with holding companies tend to have the opposite characteristics, their quarterly financial data nevertheless correspond directly to the institution that is publicly traded, and the quarterly financial data are far more extensive than financial data released at the holding-company level. Moreover, the many activities of holding company subsidiaries cannot be separated from the aggregated data reported at the holding-company level, and this lack of separability obscures the extensive information released by individual banks. Market signals at the holding-company level may or may not correspond to the performance of the bank subsidiary. The potential disconnect between the performance of individual banks and the market signals of their holding companies may widen as holding companies respond to passage of the Gramm-Leach-Bliley Act of 1999 by diversifying into additional nonbank activities.
12 Examinations that lead to rating downgrades can last from several weeks to a month or more, depending on the severity of the case. They conclude with a notification to management that the institution’s rating will be downgraded. Thus, the zero quarter can be regarded as contemporaneous with the notification quarter or the quarter of the rating change.
13 The cumulative quarterly return is calculated by multiplying unity plus the daily return for each stock i on day t(1+rit) across all trading days in each quarter, then subtracting unity.
14 As mentioned above, the control sample of healthy banks was also selected from the universe of CAMEL-rated banks and thrifts that were publicly traded over the 1988–1995 period. To be eligible for inclusion in the control sample, these institutions had to have a 1 or 2 CAMEL rating for two consecutive years and had to maintain that rating at their first on-site examination after the two consecutive years were completed. When these criteria were satisfied, the control sample selected contained 151 institutions.
15 The “critical probability” refers to the cutoff level, which determines which institutions fall into the predicted downgrade group and which do not. The logistic regression equation calculates a probability for each observation. The institutions whose calculated probability is 50 percent or more are considered likely to be downgraded and are placed into the “predicted downgrade” category.
16 An out-of-sample test was not conducted because of the limited number of observations for the sample groups. An out-of-sample test requires a “holdout” sample of 20 to 30 percent of the original observations. Holding out that many observations would have significantly reduced the size of the sample available for the analysis.