Page images

adequate evidence to justify sanctions, and the most tangible evidence is a demonstrated deleterious effect on a bank's financial condition.

Given that fact, the only way to deal with management weaknesses before they adversely affect a bank is to use the informal techniques as effectively as possible. Accordingly, it is more constructive to offer suggestions rather than to wait, as regulators do now, until the bank has deteriorated.

By making nonspecific recommendations to bankers about solving their problems, agencies miss a useful opportunity to help correct weaknesses at an early stage. Although the regulators should, of course, avoid managing the banks themselves, they can avoid doing that and still make nonbinding but more useful recommendations. The regulators themselves have pointed out that these are uniquely challenging times for bankers, especially those in small banks. Assistance by banking experts with a broad perspective should be welcome if it is not presented in the form of an ultimatum.

We are not persuaded that bankers would respond to this help by launching litigation. As our survey of commercial bankers showed, they are the least satisfied with the agencies' efforts to help find useful solutions for bank problems. Presumably they would welcome a properly conceived policy to aid them.


We recommend that the Comptroller of the Currency, the Chairman of the Board of Governors of the Federal Reserve System, and the Chairman of the Federal Deposit Insurance Corporation require examiners to make nonbinding but specific written recommendations to banks in examination reports or related correspondence on ways to improve management weaknesses.


All three Federal bank supervisory agencies maintained that their current practices are adequate. They all restated their beliefs that most institutions should be able to manage without detailed recommendations, and the regulators reiterated their desire to avoid unnecessary encroachment into bank management. The Comptroller stated that our recommendation crossed the line between the regulators' accepted role and that of bank management. The Comptroller also interpreted our report as advocating the use of formal enforcement procedures sooner and took exception to this. (See app. IV, p. 91; app. v, p. 94; and app. VII, p. 99.)

We have not asserted a role for bank regulators that usurps bank officers' prerogatives. We recognize, for example, that regulators must use their formal powers only when ample evidence of serious problems develops; that is why we recommended using a better informal tool to address management weaknesses.

We agree that unnecessary interference with bank managers should be avoided and that most banks are adequately managed. But special supervisory situations, such as the ones we reviewed, require more active agency involvement; earlier, more useful guidance could help.

In many cases, as the Federal Reserve and the Comptroller point out, supervisors may give suggestions in meetings and correspondence. But the results of the questionnaires we sent to bank executives indicate they find both the examination process and examination reports to be lacking in this area, confirming our own observations.

The bank examination process by its nature is an incursion into the realm of bank management.

Bank officers deserve the benefit of examiners' constructive suggestions as well as their criticisms. When an examiner does note management weaknesses in a bank, he/she has specific criteria in mind when doing so. In essence, our recommendation only requires them to share these criteria with bankers.




The Federal bank regulatory agencies have designed, developed, and implemented computer-based systems to monitor banks using periodically reported data. Though this effort began in the earlyto mid-1970s, there is as yet no clear-cut role defined for these systems in the supervisory process. Degrees of use and acceptance vary among agencies. As pointed out in chapter 3, the systems' value in detecting bank problems at an early stage is limited. System reports are often not timely. Peer groups of banks used for comparative analyses have been criticized as being nonrepresentative. Moreover, the systems are not being used to allocate agency examination resources.

The reason for these uncertainties is that the systems were not developed or evaluated properly. They grew piecemeal from what was considered a good idea, one used by private firms for evaluating the financial condition of banks. The agencies did not formally specify or quantify benefits and compare them to the costs of developing and operating the systems. Moreover, the Federal Reserve and FDIC did not adequately train their staffs to utilize and accept the systems.


The concept of computerized surveillance systems as an adjunct to onsite examination has evolved since the mid-1970s. The Comptroller's system was created in the most organized fashion, but even that system has never undergone a rigorous cost-benefit analysis. Consequently, though each agency's staff believes that the systems are beneficial, there is no structured body of evidence to prove this contention.

The Comptroller's system, the NBSS (see p. 9 for description), grew from a recommendation in a study made in 1974 and 1975 by the public accounting firm of Haskins & Sells (now Deloitte, Haskins & Sells). To implement the recommendation, the Comptroller established a small staff of examiners and Haskins & Sells employees in September 1975. They generally followed the concepts in the study report and by mid-1976 began operating a limited system with a 5-year data base. By October 1976 the system was fully operational, and since then data has been added to the system.

At the time of our 1976 task force study, the Federal Reserve Board of Governors had developed three major computer systems that screened and monitored functions of banks and bank holding companies. They became operational in late 1976 and early 1977, but since then the Board has made changes to the systems. At the same time, various Federal Reserve banks developed individual computerized monitoring systems. the years, the Board has centralized the computer monitoring function at its headquarters.

in addition to its own systems, the Federal Reserve receives NBSS Bank Performance Reports from the Comptroller on Federal Reserve State member banks, and the Federal Reserve also has access to FDIC's data base through on-line computer terminals.


Also, at the time of our 1976 task force study, FDIC had in operation and under development several systems to monitor bank performance for adverse trends and to identify banks with potential problems. These systems were designed to utilize different methodologies and were, to some extent, competing. After some evaluation, FDIC combined and rearranged the pilot monitoring systems into what it felt was a workable package. The early warning segment of FDIC's system ended up conceptually different from the Comptroller's. The latter system focuses on banks exhibiting significant changes and on banks whose conditions differ from their peers, while the former focuses on banks that have financial ratios exceeding certain tolerance limits.

No cost-benefit studies made

None of the Federal banking agencies performed, either prior to or during the development of their surveillance systems, any evaluations of the costs and benefits of the systems that normally should be part of a sound system design and development plan. Agency personnel feel the systems have been helpful and that their utility will increase, but they believe that no quantitative evidence could be developed.

Proper system development procedures require an analysis that compares costs to potential benefits and an organized method to validate that results are attainable. The typical computer system proposal enumerates expected savings, quantifies tangible benefits, and lists intangible benefits without attaching a price. It is recognized that the intangibles may well be the most compelling reason for approving the proposal. The next step is estimating the probability that the system will achieve the

benefits listed. 1. Finally, the benefits and their expected achievement should be measured against the costs of operating the system. This process should include the consideration of alternative systems capable of producing the same or similar benefits.

Agency officials have acknowledged that their systems were not developed using the structured approach described above. The Comptroller's system development effort, though the most organized and structured, never included a cost-benefit justification. It grew logically from a perceived need to better monitor an increasingly fast-paced industry, but it did not always consider alternative solutions. For example, one subsystem of NBSS, the Action Control System, grew from the perceived need to ensure that regional offices follow up on bank problems in a consistent and timely fashion. The team that designed the Action Control System a lways assumed that it should have been an automated system. But in a separate report, GAO has questioned that reasoning in light of the fact that the system is not being fully used, yet regions do appear to be adequately tracking bank problems without it. 2/

FDIC officials admitted their system grew up piecemeal in that they simply began building it around the data they were already receiving in periodic reports from banks. They also have performed no formal justification or cost-benefit studies. They expressed the belief that if FDIC saves one bank because of surveillance, then the system is justified.

The Federal Reserve had a special development problem in that several of the relatively autonomous Federal Reserve banks were developing their own systems while the Board of Governors headquarters staff worked on what eventually became the Federal Reserve-wide system. No overall cost-benefit or justification studies have been made, even though Board staff believes that surveillance is cheaper than examination and has improved the quality of supervision.

In fact, officials at all three agencies believe that surveillance systems are beneficial and justifiable, but they have no objective evidence. And an official at the Comptroller's office pointed out that surveillance will become even

1/The Successful Computer System, Joseph Orlicky, McGraw-Hill

Co., 1969, pp. 72-73.

2 / "The Comptroller of The Currency Should decide The Extent

To Which His Action Control System Is Needed" (GGD-81-93,
Sep. 28, 1981).

« PreviousContinue »