Establishing On‐Site Bank Examination Priorities: An Early‐Warning System Using Accounting and Market Information

THE ON-SITE EXAMINATION is and has been the backbone of bank supervision in the United States. Both federal and state banking agencies attempt to perform on-site inspections every twelve to twenty-four months for banks under their jurisdiction. Banks with known supervisory or financial problems are given top priority in the examination process, whereas small banks with satisfactory management, adequate capital, and acceptable earnings are given the lowest priority.' At present, on-site inspection priorities are established using historical examination and accounting reports. The purposes of this paper are to suggest an earlywarning technique using accounting and market information which may be useful in creating a more efficient method of scheduling bank examinations and to test the proposed dual-screening technique on a population of banks which have recently failed. If a bank is designated or flagged as a potential problem bank by both the market screen and the accounting screen, it is considered to be in immediate need of an on-site inspection. If a bank is flagged by only one system, the early-warning signal is considered to be less urgent. Banks not flagged by either screen and with "clean" histories are placed at the end of the examination queue. The proposed dual-screening technique applies only to banks with activelytraded securities, such banks are the largest ones and present the greatest risk to the deposit-insurance fund and to the stability of the banking system.2 For banks without active markets for their securities, a single screen based upon accounting data is recommended to establish examination priorities. Thus, the suggested early-warning system should be used to order the examination queue. In establishing any early-warning system, the bank regulator's primary desire is to minimize the misclassification of problem banks as nonproblem banks (typeI error). To obtain classification information early, regulators should be willing to accept a higher type-II error (classifying nonproblem banks as problem banks) to gain a lower type-I error. Additionally, the type-II misclassification error only