THIS PAPER PRESENTS a multivariate statistical analysis of the balancesheet and income-statement characteristics of problem banks.' Newly identified problem banks are matched with nonproblem (control) banks and multiple discriminant analysis (MDA) is used to test for group mean differences, to describe the overlap between groups, and to construct rules to classify observations (banks). into either the problem-bank group or the nonproblem-bank group. The empirical findings indicate that measures of banking factors such as asset composition, loan characteristics, capital adequacy, sources and uses of revenue, efficiency, and profitability are good discriminators between the groups (i.e., group mean differences exist). The chi-square measures of group overlap indicate that the distributions of the individual bank's characteristics (i.e., vector of variables) overlap substantially. Given the high degree of overlap between groups, the "descriptive" classification results are "better" than might be expected.The first section of this paper deals briefly with the definition and importance of problem banks. In the next section, a description of the problem-bank group and its control group are presented. The data and methodology are given in the third section, while the empirical findings are presented in the fourth section. In the final section, the summary and conclusions and implications for future research are presented. I. PROBLEM BANKS: DEFINITION AND IMPORTANCEA problem bank is one that, in the eyes of the Federal banking agencies, has violated a law or regulation or engaged in an "unsafe or unsound" banking practice to such an extent that the present or future solvency of the bank is in question.2 Problem banks are identified during bank examinations. The purposes of a bank examination are:3 * Financial Economist, Economic Research Unit, Federal Deposit Insurance Corporation. The views expressed in this paper are the responsibility of the author and in no way represent a policy determination endorsed by the FDIC. I thank Robert A. Eisenbeis for his constant and generous help in the preparation of this paper; William D. O'Brien for his research and computerprogramming assistance; and Edward J. Kane and especially Edward I. Altman for their helpful comments. This is a revised version of a paper presented by the author at the American Statistical Association meetings in New York, December, 1973. Parts of an earlier version of this paper appear in the 1973 Proceedings Volume of the ASA and are reproduced here with the permission of the Association. 1. A descriptive analysis of the characteristics of these banks can be found in Sinkey [23]. See also Sinkey and Walker [24] for a preliminary problem-bank study. 2. The Comptroller of the Currency supervises national banks, the Federal Reserve System regulates state member banks, and the FDIC examines insured-nonmember commercial banks and federally-insured mutual savings banks. Furthermore, all state banks (insured and noninsured) are 21 22The Journal of Finance 1. To dete...

No abstract

Net loan chargeoffs and nonperforming loans reflect realized credit risks for banks. These risks arise from either external factors such as depressed economic conditions (e.g., the energy and farm belts of the United States in the 1980s) or internal factors such as poor lending decisions (including fraudulent ones) or both. For large commercial banks in 1987, we find that almost 94 percent of the variation in loss rates within regions was due to banks having different loss rates on the same types of loans. Our regression results indicate that loan-loss rates in 1987 were positively associated with loan rates, volatile funds, and loan volume from the preceding three years. In contrast, banks with "adequate capital" in the preceding three years tended to have lower loss rates.

In recent years, commercial banking in the United States has experienced a decline in its traditional business of financing loans by issuing deposits. Simultaneously, banks have become more involved in nontraditional activities that provide financial services and generate fee income. As a result, real aggregate noninterest income has risen relative to income from traditional activities. This paper examines features common to banks that are heavily engaged in nontraditional areas. The empirical analysis suggests that these banks tend to be larger, have smaller net interest margins, have relatively fewer core deposits, and exhibit less risk. These findings have intuitive appeal and conform to conventional wisdom; while larger banks have fewer core deposits and face more competitive interest rate conditions, resulting in narrow spreads from traditional intermediation, they have more diverse sources of revenue and greater access to financial markets, which reduces risk.

No abstract