TEXT OF TREASURY REPORT ON THE U.S. FINANCIAL SYSTEM IS AVAILABLE.

SUMMARY:

A. INTRODUCTION

Regardless of their financial condition, all FDIC-insured banks (thrifts) ¹ pay the same statutory rate for deposit insurance and share proportionately in any premium rebates. ² As a result, deposit insurance rates do not vary with the level of risk that a bank poses to its insurance fund. This system of flat-rate premiums has been criticized on the grounds that it encourages excessive risk- taking and that it inequitably distributes the burden of insurance losses among banks. ³

There are good reasons to review the FDIC's pricing policies. First, there have been substantial changes in the banking industry during the last decade. Changes in the regulatory, economic, and technological environment have created new incentives and opportunities for risk-taking. Thus, although the Bank Insurance Fund (BIF) has been adequate to handle insurance losses thus far, the past may not be a good indicator of the appropriateness of our current pricing system. Recent FDIC estimates suggest that the BIF will decline to $9 billion at the end of 1990, and will suffer further losses in 1991. Moreover, insurer solvency only means that, on average, insurance premiums have been sufficient to cover losses. Allocative inefficiencies still could exist. In effect, more conservatively run banks may be paying for the excesses of others, and banks that elect to take advantage of risk subsidies will grow relative to those that do not.

This chapter, primarily drafted by the FDIC and nearly identical to parts of the FDIC's risk-based study recently prepared for Congress, will discuss the implications of mispriced deposit insurance premiums in Section B. Section C will point out a number of existing factors which may at least partly compensate for the effect of mispriced premiums. Section D will then discuss the obstacles to implementing a system which accurately prices risk. Sections E and F discuss the major proposals for risk-related premiums, and the arguments for and against the proposals. Finally, Section G briefly summarizes public comments received regarding risk-related premiums generally.

B. IMPLICATIONS OF MISPRICED DEPOSIT INSURANCE

1. PREMIUMS AS A SUBSIDY OR TAX

At the industry level, the deposit insurance premium can act as a subsidy or tax, depending on whether the premium is below or above the premium that would be set in a competitive market. The provision of a credible guarantee to pay off insured depositors in the event of a bank's insolvency allows insured institutions to attract deposits at a risk-free rate (or at some rate less than the proper risk- adjusted rate) and, thus, gives them a competitive advantage over uninsured institutions. If this advantage is not offset by charging insurance premiums (either explicitly or implicitly through supervision and regulation) sufficient to cover potential insurance losses, depository institutions will have competitive advantages over other providers of financial services. Thus, the subsidy would allow the industry to grow beyond the size that would result from a purely competitive process, growth that would come at the expense of uninsured providers of financial services. Conversely, setting insurance prices too high would act as a tax on the industry. Banks would be at a competitive disadvantage relative to uninsured institutions, and banks would either drop their insurance or resources would be diverted from banking to other financial service providers.

2. MORAL HAZARD

Mispriced deposit insurance is most often discussed in terms of its implications for the risk-taking behavior of depository institutions. The current flat-rate system has been criticized because it creates incentives for banks to increase their portfolio risk. Market participants are normally confronted with a risk-return trade-off: higher yields can only be obtained at the expense of greater risks. In the absence of deposit insurance, the gains that stockholders may realize from moving to riskier positions would be limited by depositors, who would demand additional compensation for increased risk-taking by the bank.

However, with the introduction of deposit insurance, insured depositors no longer require risk premiums commensurate with the level of risk since their investment is safe and, under a flat-rate premium structure, banks' insurance costs will be the same regardless of their risk position. As a result, banks may take on additional risk without having to pay higher interest rates on deposits or higher insurance premiums. The risk-return trade-off has been altered such that the price of assuming greater risk has been reduced and, consequently, the bank is likely to move to a riskier position. ⁴⁵ This problem is referred to as the "moral hazard" problem throughout this study.

Thus, there are two aspects to the mispricing of deposit insurance, both of which result in a misallocation of credit resources. First, if the overall level of insurance pricing is not equal to the price that would be set in a competitive market, deposit insurance will act as an industry subsidy or tax, and insured institutions will be at a competitive advantage or disadvantage relative to uninsured institutions. Second, the flat-rate pricing system provides incentives toward greater risk-taking, with the result that some risky investment projects will be undertaken that would not otherwise have been undertaken. As a consequence, bank failures are likely to be more numerous and more costly than if insurance prices varied with the level of risk.

C. COUNTERBALANCES TO INCREASED RISK-TAKING UNDER CURRENT SYSTEM

1. MARKET DISCIPLINE

It should be noted that under a flat-rate system there still may be important counterbalances to increased risk-taking. To the extent that uninsured liabilities are at risk, these debt-holders will exert some discipline on bank risk-taking. In addition, stockholders (or owners) have an important stake in the survival of an institution. Provided that they have sufficient capital to lose (adequate capital levels are maintained) and are sufficiently averse to risk, stockholders will place limits on management's risk-taking activities. Thus, while a flat-rate system generally will lead to greater risk (an exception is noted in footnote 4), it does not necessarily imply that institutions will seek the riskiest portfolios.

2. REGULATORY DISCIPLINE

In practice, risk-taking also is limited by the fact that, in addition to the statutory premiums, banks incur other costs under the current system. The provision of deposit insurance requires that insured institutions submit to federal supervision and regulation. Regulations limit insured institutions from engaging in certain financial activities and set minimum capital requirements. Regulators periodically examine banks to determine if they are engaged in safe and sound banking, and undesirable behavior may be penalized through issuance of cease-and-desist orders, removal of bank officers or directors for certain violations, and/or the levying of fines. In addition, the FDIC, as insurer, has the exclusive authority to suspend temporarily or to terminate permanently an insured institution's deposit insurance if the FDIC determines that the institution is engaged in an unsafe and unsound banking practice. These regulations and supervisory sanctions limit the ability of some banks to engage in overly risky activities and they represent an implicit cost of obtaining federal guarantees. To the extent that these implicit costs vary with the riskiness of the bank, they act as a system of risk-related premiums and constrain risk-taking. ⁶

D. GENERAL PROBLEMS IN PRICING BANK RISK

1. EX ANTE VS. EX POST RISK

Nearly all insurance settings are characterized by asymmetric information concerning the insured's risk type. That is, the insured possesses better information about his or her risk type than does the insurer. For example, automobile drivers know their own driving patterns and behavior better than the insurer and, if they were honest with themselves, could better assess their own risk than could the insurer. However, high-risk drivers have incentives to hide their true risk characteristics and to pose as low-risk types. In order to overcome this problem, insurers will attempt to bridge the information gap by using actuarial information to make ex ante judgments about a driver's risk type based on such characteristics as age and sex. The insured's driving record (e.g., traffic tickets and accidents) can be used to obtain ex post information about the driver's risk type. Of course, even with this information the insurer will not know the driver's true risk type with certainty.

Although automobile insurance differs from deposit insurance in many respects, the example helps to illustrate the general problems associated with asymmetric information. Just as in the case of drivers, banks possess more information about their risk type than does the FDIC. Moreover, determining a bank's risk type ex ante is arguably more difficult than in most insurance settings. A major function of banks (as well as other intermediaries) is to assess the risks of lending to idiosyncratic borrowers (borrowers who are obtaining credit for information-intensive projects). For many of these borrowers, public information on their economic condition and prospects is so limited and expensive that the alternative of issuing marketable securities is not economically viable (Goodhart (1987), p. 86). Thus, banks specialize in obtaining information about the very events (credit risks) that are most likely to result in a loss to the insurer. ⁷ Because of this specialized knowledge, the ex ante information gap between the insurer and the insured is perhaps larger than in most other insurance settings. ⁸

2. ADVERSE SELECTION

Asymmetrical information regarding the insured's risk type results in two problems common to insurance settings: the difficulty of correctly classifying a client's risk type resulting in an overabundance of risky clients in the insurance pool (sometimes referred to as the adverse selection problem) and the problem of controlling the insured's risk-taking once insurance is granted (moral hazard).

The insurer can reduce the adverse selection problem by obtaining more information about the client. Of course, the benefits of greater information (more appropriately priced insurance and lower insurance losses) would have to be weighed against the costs of obtaining that information (costs of additional resources needed to obtain information).

Another solution to the adverse selection problem is to offer incentive-compatible contracts. ⁹ For example, automobile insurers offer varying amounts of deductible insurance in combination with different premium rates. If a driver feels that he or she is a particularly safe driver, he or she probably will opt for a relatively high-deductible, low-premium contract, and vice versa for a high-risk driver. By allowing insurance contracts to vary by more than one characteristic, for example, price and coverage, the incentive-compatible contract is designed to induce insurers to signal their true risk type.

An incentive-compatible deposit insurance contract could involve offering banks the choice of various price/capital combinations. Banks that choose higher capital levels (these could be adjusted for loan quality) would pay lower insurance premiums, and vice versa. The idea is that obtaining additional capital would be less expensive for low-risk banks than for high-risk banks. Thus, low-risk banks would prefer to select a high-capital/low-premium combination, while the opposite would be true for high-risk banks. The goal would be to adjust the price/capital combinations so that the long-run revenues of each risk category would be sufficient to cover long-run costs. In doing so, each risk category would be paying an actuarially fair premium and cross-subsidization between risk classes would be eliminated.

In banking, the difficulty is determining when the revenues of any particular category are sufficient to cover expected costs. In casualty insurance, this is relatively easy since the events being insured against are normally occurring events that are fairly evenly distributed over time. As a result, an automobile insurer will learn in rather short order whether the premium revenues are sufficient to cover the long-run costs of any risk category. However, bank failures are not evenly distributed over time. Instead, they tend to be associated with the business cycle or economic shocks. In this environment, adjusting the price/capital combinations so that the long-run revenues are sufficient to cover the long-run costs of each risk category would be a lenghty learning process. ¹⁰ /11/

3. MONITORING THE INSURED

After granting insurance, the insurer must guard against the client taking actions that increase the insurer's potential loss. The moral hazard problem will vary depending on the extent to which the insured has incentives (normally financial incentives) to take actions that increase his or her risk and the extent to which these actions are unobservable by the insurer.

In many insurance settings, moral hazard often is controlled by making the insurance payout contingent on the insured party acting in a specified manner. For example, an insurance company will not pay off on fire damage if the insured party commits arson. However, payouts to depositors contingent on bank behavior would not be feasible, since it would reintroduce the problem of bank runs. Alternatively, the moral hazard problem may be dealt with by monitoring bank behavior (examinations) and imposing penalties on managers and owners when undesirable behavior is observed.

E. PROPOSALS FOR RISK-RELATED PREMIUMS

There is widespread acceptance that a flat-rate premium structure, by itself, creates perverse incentives toward greater risk-taking and penalizes more conservatively run institutions. There is less agreement whether a more explicit risk-related pricing system could be developed that would be a significant improvement over the current system. A number of proposals for establishing risk-related premiums have been made; each has some advantages and disadvantages when compared to the current system. These proposals generally can be categorized into those that try to incorporate the market's assessment of bank risk and those that rely on the public insurer's assessment of risk.

1. USING MARKET INFORMATION TO ASSESS RISK

Several methods that rely on the use of market information to price deposit insurance are found in the literature. For example, the use of interest rates on uninsured deposits, private reinsurance of deposits, and option pricing theory have been advanced as means of correcting for governmental mispricing of deposit insurance.

INTEREST RATES ON UNINSURED DEPOSITS

Deposit insurance provides explicit coverage for deposits of $100,000 or less, leaving uninsured those deposits greater than $100,000. It has been proposed that insurance premiums could be based on the market rates paid on these uninsured deposits (Peltzman (1972), Thompson (1987)). This approach is based on the idea that depositors will demand a risk premium if they perceive that their uninsured deposits are at risk. Since depositors could place their uninsured funds in an alternative investment with the same level of risk (e.g., a money market or bond fund), there should exist a similar risk premium with either investment option.

There are, however, several limitations to this approach that stem from market imperfections. First, investors may perceive that large banks will not be allowed to fail. This expectation of de facto coverage for uninsured depositors may obviate the need for uninsured depositors to demand an appropriate risk premium, especially in the case of large banks. Second, in addition to differences in risk, rate differentials between insured and uninsured deposits may also reflect market imperfections, such as transaction costs or less than perfectly competitive markets. Many large corporations, for example, maintain large bank account balances that are technically uninsured. This observed behavior is likely due to the impracticality of parceling these deposits into insured accounts as well as the perception that some banks are "too big to fail." In addition, some sophisticated (and uninsured) depositors may feel that they always will have sufficient warning to withdraw their funds prior to failure.

Thus, the rate paid on uninsured deposits may not accurately reflect the risk premium that should be charged. If so, risk premiums on these deposits may not be appropriate for setting insurance premiums.

PRIVATE REINSURANCE

Some combination of public and private insurance has been suggested as a way to overcome the shortcomings associated with purely public or private deposit insurance systems. Under one such proposal (Baer (1985)), production and pricing would be separated: government would provide most of the insurance, while private insurance companies would determine market-based prices for both public and private insurance. ¹² For example, the FDIC could insure 95 percent of the $100,000 limit, but utilize the prices established by private insurers who would be insuring the remaining 5 percent. In the event of a bank failure, private insurers would be responsible for paying off their portion of the bank's insured deposits, and would share losses on a pro rata basis with the federal insurer.

In order for such a system to be successful, however, the private insurer must be able to survive systemic risk. This suggests that the private insurer would need to hold a high percentage of reserves against insured deposits and would need to be prohibited from canceling insurance when bank failure appears imminent. In addition, there may be instances where the objectives of the public insurer may conflict with those of the private insurer, particularly in the areas of closure and failure resolution policies. The extent to which private insurers would be willing to provide such insurance under terms consistent with public policy objectives is unclear. This topic is discussed more fully in Chapter VII, "Alternatives to Federal Deposit Insurance."

OPTION PRICING

Option pricing theory has been suggested as a method of determining the value of deposit insurance to a bank. In this literature, deposit insurance is shown to be analogous to a put option. Options, as financial contracts, have been popular because they confer on the holder the right, but not the obligation, to buy or sell specified property at a fixed price and on some fixed future date. There are two basic types of option contracts. The call option gives the holder the right to buy an asset at a specified price, called the exercise or strike price, on some future date. The put option, in contrast, gives the holder the right to sell an asset at the exercise price on some future date.

The value of the put option at maturity depends on the current value of the underlying asset relative to the contract's exercise price. If, at the option's expiration or maturity date, the asset price is greater than the exercise price, the option is not worth exercising and therefore the value of the option is zero. In this case, the put is termed "out-of-the-money." However, if the asset price is less than the exercise price, the option is termed "in-the- money." It will be exercised, since the asset can be sold at a price that is greater than the asset's current market value. The option holder will realize a profit equal to the difference between the exercise price and the asset price. Therefore, the value of the put option at maturity is equal to the maximum of the difference between the exercise price and the asset price, or zero. Similarly, the value of an option prior to its maturity or expiration date will depend on the probability of the option being in-the-money.

Essentially, in purchasing deposit insurance, the bank has purchased a put option, and has the right to sell (put) its assets at a price equal to its insured liabilities. If the value of the bank's assets falls below the bank's obligations to insured depositors, the insurer will appropriate the bank's assets and, in turn, pay off insured depositors. This option to sell its assets to the insurer at a price equal to the value of the bank's insured liabilities has value to the bank because it makes insured deposits perfectly safe and allows the bank to attract deposits at a risk-free rate. ¹³

Merton (1977) was the first to suggest that option pricing theory could be used to determine the value of deposit insurance to a bank. Using the option pricing framework developed by Black and Scholes (1973), Merton derives an option pricing formula for valuing deposit insurance. When the option pricing framework is applied to the problem of pricing deposit insurance, the relationship between the value of the put (and in turn the "fair price" of deposit insurance to the bank) and the probability of insolvency is underscored. Notably, changes in the capital position of the bank lead to changes in the value of the deposit insurance contract. For example, if the value of the bank's assets were to decrease relative to the value of its liabilities, the value of the put (or deposit insurance) to the bank's owners would increase. Similarly, an increase in the variability or volatility of the bank's return on assets would increase the probability of insolvency which would be reflected in an increase in the value of the put and deposit insurance to the bank's owners. (The put option analogy also reveals other factors that influence the value of deposit insurance. Among these are the lifetime of the put option, as measured by the time between bank examinations, and the total amount of insured deposits, referred to as the strike or exercise price of the put option. Additionally, the closure rule followed by the regulators will affect the total amount of liabilities covered by insurance and therefore the exercise price.)

The feasibility of using option pricing theory to price deposit insurance depends on the ability of the insurer to adequately measure the return volatility of bank assets in a timely manner. This requires considerably more information than is available for most banks and, therefore, would be difficult to implement for most institutions. However, even though an option pricing scheme for deposit insurance may be difficult to implement, it is a valuation approach which deserves, and is currently receiving, more study. ¹⁴

2. USING NONMARKET INFORMATION TO ASSESS RISK

When it is not possible or when it is undesirable to utilize the market's assessment of bank risk, the federal insurer would be left with the task of developing its own methods for assessing risk. Various proposals that have been made would permit the FDIC to administratively determine variable-rate premiums, including the FDIC's own proposals (FDIC (1983); Hirschhorn (1986)). Some of these proposals attempt to measure risk ex ante; that is, they attempt to measure the inherent risk of banking activities regardless of the bank's current performance. Most proposals, however, have relied primarily on ex post measures of risk, those that measure risk after it has materially affected the performance of the bank.

Charging banks risk premiums by measuring ex ante risks has the advantage of discouraging risky behavior before it adversely affects the performance of the bank. Not surprisingly, devising such a system is difficult. Ex ante approaches to risk measurement have generally sought to measure various components of risk that are thought to be inherent in the business of banking. These components might include interest rate risk, credit risk, operating risk, liquidity risk, diversification risk, and the risk of fraud or insider abuse. ¹⁵ While there may be acceptable ways to measure some of these individual risk components (most notably interest rate risk, although banks do not now report the kind of information that would be required), attempts to measure and aggregate all of the various components have been largely unsuccessful. ¹⁶

ASSET RISK BASKETS

The risk-based capital guidelines utilized by the bank regulatory agencies are an attempt to apply ex ante measures of perceived credit risk. The plan classifies assets into broad categories according to their perceived risk of default and attaches risk weights to these categories. A risk-based pricing scheme should not be inconsistent with these risk-based capital guidelines.

It would be possible to devise a risk-based premium system using the same approach. The measurement of portfolio risk under this system may be questioned on the grounds that it simply attaches risk weights to individual asset types, while ignoring the composition of assets within the entire portfolio. Furthermore, institutions would be able to increase the risk in their portfolios, without a corresponding increase in their risk measure, by moving to the risky end within each asset category and by having concentrations of assets in particular areas (either sectoral or geographic). Such problems underscore the difficulty in finding acceptable ex ante measures of risk.

Given the existence of the risk-based capital guidelines, a question arises as to the need for a system of risk-based premiums. There are a couple of reasons why it may be desirable to have both systems in place. First, the risk-based capital guidelines focus on credit risk, which is only one form of potential risk. A system of risk-based premiums could further incorporate other forms of risk- taking into the pricing scheme. In addition, because the risk-based capital guidelines are based on an international agreement, changing the guidelines to accommodate other forms of risk-taking or to accommodate changes in the level of risk that these different forms pose to the insurance fund would be more difficult. Second, the explicit pricing of risk utilized under a system of risk-based premiums has some advantages over a regulatory standard or minimum that is established under the risk-based capital guidelines. In the case of a regulatory standard or minimum, an institution either passes or fails the standard: there are no gradations of risk as there might be under a system of risk-based premiums. Moreover, when facing a price, as opposed to a capital standard, an institution has greater flexibility in responding to a bad situation. For example, for an institution that temporarily falls below its desired capital position and faces abnormally high costs in attracting additional capital (capital markets may not always be efficient), the most efficient way to deal with the problem may be to pay higher insurance premiums and temporarily live with the lower capital level, rather than immediately raise capital.

RATINGS BASED ON EXAMINATION INFORMATION

It has been suggested that information derived from the regulatory agencies' onsite examinations could be used as a basis for risk-related premiums. As a result of the examination process, each bank is assigned an overall rating from 1 to 5 (5 being the worst) based on the bank's financial condition. This rating is commonly referred to as the CAMEL rating and is derived from the examiner's evaluation of a bank's capital adequacy, asset quality, management, earnings and liquidity. ¹⁷ Perceptions of ex ante risk play some role in the determination of CAMEL ratings, since examiners evaluate management's policies and practices that influence the bank's future performance. In the areas of capital adequacy, asset quality, and earnings, however, perceptions of risk are largely based on the bank's current performance. Nevertheless, a major argument in favor of using information derived from examinations is that it may contain inside information on a bank's operations that is not obtainable through other means (i.e., offsite monitoring).

A major objection to using examination ratings as the sole basis for assigning risk premiums is that it could have a negative impact on the examination process. One of the advantages of onsite examinations is that they allow examiners to use their experience and judgment to tailor their assessments and solutions to unique situations. However, because of the financial stakes involved with basing premiums on examinations and the likelihood that such ratings would become public information, extreme care would need to be taken to ensure the application of uniform standards and procedures for rating banks. With greater reliance on rules and procedures for assigning premiums, an important attribute of onsite examinations -- examiner discretion -- may be lost. Further, basing premiums on examinations introduces an adversarial relationship into the examination process, and the flow of information that normally occurs during an examination probably would be reduced. While the examination process can have an adversarial aspect, the purpose also is to provide useful information to bank management and regulators about the soundness of its operation and about how it may be improved. Increasing the financial stakes of the examination outcome could lessen the extent to which an examination could serve this purpose. ¹⁸

FAILURE-PREDICTION MODELS

Some proposals for risk-related pricing schemes have been based on information provided by bank failure-prediction models. ¹⁹ Failure-prediction models utilize historical information to determine the importance of various financial variables (usually taken from the reports which are submitted to the federal bank regulators) in predicting the success or failure of an institution. Those financial variables (e.g., measures of nonperforming loans, earnings, capital levels, etc.) that have been consistent predictors of past failures can then be used as a basis for a risk-related pricing system. That is, pertinent financial date can be used to estimate the likelihood of failure for currently operating institutions, and insurance premiums can be assigned on the basis of each bank's probability of failure. More recently, these types of models have been modified to estimate each bank's expected insurance cost (roughly equal to the probability of failure, multiplied by the FDIC's average cost when a bank fails). The expected cost then can be used as an estimate of the risk-related portion of the insurance premium (Avery, Hanweck, and Kwast (1985)).

Not surprisingly, the financial variables that turn out to be most successful in predicting failures are primarily ex post measures of risk and, as a consequence, the predictive power of these models declines rather rapidly when predicting failures much beyond a year. For example, in one recent FDIC proposal (Hirschhorn (1986)), the financial variables that did the best job of replicating the problem bank list included variables describing a bank's capital level, its earnings performance, and the quality of its loans. Using a model based on December 1983 data and limiting the designation of high-risk banks to roughly 20 percent of all banks, ²⁰ the model classified about 90 percent of all failures in 1984 as high-risk banks. However, using the same model (i.e., based on 1983 data) only about 60 percent of the failures in 1985 were classified as high risk. This profile is common in failure-prediction models, and illustrates the difficulty in detecting and pricing risk in a timely manner. ²¹

ADJUSTED CAPITAL APPROACH

This approach would use a depository institution's capital-asset ratio, adjusted for some measure of asset quality and/or other performance measure(s), as the basis for the institution's deposit insurance assessment rate. One such proposal can be found in FDIC (1983), Chapter II. Capital is important to the federal insurer because it provides a protective cushion against adverse changes in an institution's asset quality and earnings. It is this direct relationship between more capital and a lower probability of failure which serves as the foundation for the adjusted capital approach.

Three issues must be addressed in formulating the adjusted capital measure: (1) the definition of capital; (2) the adjustment(s) to capital; and (3) the definition of total assets. The first issue would involve questions regarding what should be included in the capital measure (e.g., common equity, allowances for loan losses, and subordinated debt). With regard to the second issue, an institution's assets, and therefore capital, might be reduced by an amount equal to the losses expected on its nonperforming assets (i.e., assets categorized as past due or nonaccrual). Expected losses on such assets might be based on the industry's historical relationship between nonper[forming] assets and charge-offs. The third issue would involve whether to include some or all of the "off-balance-sheet" assets in the definition of total assets.

In sum, this approach has advantages in its simplicity, its use only of information currently reported to the federal banking agencies, and its reliance upon the most proximate measure of risk to the insurance fund -- capital. On the other hand, it could be argued that this approach needlessly creates a third definition of capital (in addition to risk-based capital and the leverage ratio), and only attempts to accomplish what loan loss reserves should do in the first place: recognize potential losses already existing on an institution's balance sheet.

EX POST SETTLING UP

A more recent proposal for risk-related premiums involves an ex post settlement for failed banks (Benston, et al. (1986); Merrick and Saunders (1985)). As a condition for receiving federal insurance, banks could be required to establish an escrow account with the FDIC, or bank shareholders could be legally subject to extended liability. In the event of a failure, ex post penalties could be assessed depending on the insurer's actual loss experience. Extended liability would expose the bank to an extended set of negative outcomes resulting from its investment behavior (and thereby lower its expected return), rather than limiting the set of negative outcomes to its initial equity investment. Such a system of ex post settling- up may provide the bank with incentives approaching those that would exist with ex ante measures of risk.

A general problem with this type of ex post settlement proposals is that they may result in increased costs for ALL commercial banks regardless of their current risk position. Extended liability for stockholders will increase the costs of retaining and attracting capital, since stockholders will demand additional compensation for the increase in their potential losses should the bank fail. Requiring banks to maintain escrow accounts is equivalent to increasing capital requirements, while restricting the earnings potential of the added capital. (It seems likely that the bank earnings on the escrow account would be limited to Treasury bill rates.) While these proposals have the potential to reduce the incentives toward risk-taking, they also have the potential to significantly increase banks' cost of capital, regardless of the actual risk position of individual banks, and could overly restrict the growth of the banking industry relative to other financial service providers.

MULTI-TEST RISK-BASED PRICING SCHEMES

More recent suggestions for structuring a risk-related system include the use of combinations of the above approaches. For example, statistical models utilizing bank Call Report data could be used to estimate the risk of failure or the expected cost to the FDIC. Premiums based on these estimates could be double-checked by noting the rates paid on uninsured deposits or other uninsured debt, by comparing them to the most recent CAMEL rating, or by using option pricing techniques. Further, depending on the size of an institution, different risk classification techniques might be used in order to improve risk measurements. Although potentially more complicated, a multi-test risk-based pricing scheme might instill greater confidence in the regulator's risk assessments, and so, avert any serious mismeasurement of an institution's risk.

F. ARGUMENTS FOR AND AGAINST RISK-RELATED PREMIUMS

1. THE USE OF MARKET INFORMATION

Conceptually the advantage of utilizing market information is that it represents the assessment of numerous individuals who have a financial stake in correctly assessing bank risk. Moreover, on a theoretical level, basing insurance premiums on those that would be set in competitive, unregulated markets would result in the optimum risk-return trade-off for the economy (assuming no third-party effects).

Despite these conceptual advantages, basing insurance premiums on some form of market information raises questions regarding the quality of market information that could be obtained and whether a market-based scheme would, in reality, lead to more accurate pricing. With respect to the quality of market information, most market-based approaches face some sort of information problem. For example, basing premiums on the rates paid for uninsured deposits would require well- developed markets for both large and small banks. Even if the FDIC were to abandon its policy of providing 100 percent de facto insurance (for "uninsured" depositors) in purchase-and-assumption transactions, regional interest-rate differentials and imperfect markets for small banks' uninsured deposits would make such an approach difficult to implement.

The informational requirements of option pricing techniques also present problems. In order to provide estimates of the value of deposit insurance for all banks, some estimate of asset returns (market returns) and their volatility over time must be made. Studies which have used option pricing to estimate the value of deposit insurance have typically relied on changes in an institution's stock prices over some historical period to estimate returns and their volatility. But these estimates are based on historical returns and do not necessarily represent the returns that an institution expects to receive based on its current investment decisions. To the extent that expected returns deviate from historical returns, the option price will be incorrect. Moreover, as Pyle (1983) has pointed out, small errors in the estimation of the value of assets or their volatility can have major effects on the value of the option contract (i.e., the insurance premium). A further informational difficulty is knowing the appropriate closure rule. If assumptions concerning closure rules are wrong, the value of the put may be in substantial error. Brickley and James (1986) provide some empirical evidence on this point. They show that for the S&L industry during the early 1980s, the assumption that closure would occur at the point of insolvency resulted in an understatement of the option value of deposit insurance (i.e., insurance would have been underpriced with this assumption).

Another practical problem with using the option pricing model is that stock market information is available only for the largest banking organizations. While a proxy for stock prices can be estimated, it is not clear how well this kind of estimation technique would work. Moreover, where stock price information is available, it only is available for the holding company and not for individual banks.

A more fundamental question is whether the market's assessment of individual banking risks is measurably better than information derived from other sources that are potentially available to regulators. A major reason why borrowers obtain loans from intermediaries rather than issue marketable securities is that public information on their economic condition and prospect is extremely limited and expensive. Thus, with respect to the quality of a bank's loans, the bank possesses information that is generally not publicly available. ²² To some extent, the very existence of banks (and other intermediaries) is explained by the inability of markets to act as efficient devices for valuing these loans. If this is the case, we should not expect markets to be particularly efficient at evaluating credit risks in banking. ²³

Whether a system based on an option pricing model would do a worse job than the current system or some other (nonmarket) risk- related system is not clear, and more investigation is needed. ²⁴ However, while the option pricing model appears to do relatively well at ranking the current financial condition of publicly traded bank holding companies at a point in time, from the studies reviewed it's not clear how well the model assesses risks in an ex ante sense or how well it establishes the appropriate premium level for a particular institution. For example, in looking at changes in bank risk-taking for 98 of the largest bank holding companies from 1981- 86, Furlong (1988) estimated that the value of deposit insurance increased from an average of 2.4 one-hundredths of a basis point per dollar of deposits in 1981 to 2.6 tenths of a basis point in 1986. Even the higher 1986 estimate represents only about 3 percent of the 8 basis points that banks were charged at the time. On the other hand, assuming a less stringent closure rule, a study by Ronn and Verma (1986) estimated the average value of the insurance guarantee in 1983 (again for large bank holding companies) to be roughly equal to the 8 basis points. The magnitude of these differences underscores the difficulties in implementing the option pricing model.

2. THE USE OF NONMARKET INFORMATION

If market information is not used in setting insurance premiums, then it should be recognized that an alternative risk-related scheme amounts to a set of administratively determined prices (either explicit or implicit). The question then turns on how accurately we believe regulators can price risk and whether a system of explicit risk-based premiums has advantages over implicit pricing.

Obtaining accurate ex ante measures of bank risk is perhaps more difficult than in many other areas of insurance. In an ex ante sense, the insured nearly always has better information about the potential risk he or she faces than does the insurer. In the case of banks, assessing the financial risks of making information intensive loans is a central function of the enterprise. As a result of this specialized knowledge, the ex ante information gap between the insured and insurer is perhaps larger than in most other insurance settings.

This large informational asymmetry between the insured and insurer is perhaps one of the reasons for the inability of researchers to find good ex ante measures of risk. Although there are steps that the insurer could take to increase the amount of information concerning the inherent risks of specific institutions (such as becoming intimately familiar with an institution's credits), the costs of acquiring this information may well be prohibitive. On the other hand, using ex ante measures that are not based on highly specific information (specific with respect to an institution's credit risks or other portfolio risks) would likely be ineffective and, more importantly, runs the risk of influencing risk-taking behavior and credit allocation in undesirable ways. Thus, most analyses have concluded that any workable system of risk-related premiums would be restricted to one based largely on ex post measures of risk (e.g., see Avery, Hanweck, and Kwast (1985); Merrick and Saunders (1985), p. 707).

There have been two major criticisms of basing risk-related premiums on ex post measures of risk. First, it is argued that if risk is recognized by a premium system only after an institution's asset quality has deteriorated, then the premium structure has not served its purpose of inhibiting risk-taking (Horvitz (1983), p. 259). This argument, however, fails to recognize that after-the-fact penalties may still provide some deterrent effect. While the best approach may be to levy a higher premium for a higher level of risk regardless of the assets' current performance status, it is the case that if a lender knows that a premium penalty will be charged for poorer asset quality the lender will be forced to internalize this cost into the lending decisions, thereby acting as a deterrent to excessive risk-taking.

The second criticism of ex post measures of risk is that they will penalize banks when they can least afford it, i.e., when they have encountered difficulty. A deterioration in asset quality diminishes a bank's earnings and puts pressure on its capital buffer. A premium penalty which is based on some measure of asset quality will put an additional strain on both earnings and capital. While this premium cost is internalized by the lender, the premium charge must not be so large as to threaten the viability of an otherwise sound institution. In addition, credit quality typically declines during a downturn in economic activity. Increasing premiums during an economic downturn could further aggravate banking problems (Goodman & Shaffer (1984), p. 154), even though loan-quality problems would not necessarily be the result of poor management decisions. ²⁵

Any ex post system (for example, using nonperforming assets as a measure of risk) must balance the need to impose penalties sufficiently large to deter excessive risk-taking, against the possibility that excessive penalties may aggravate banking conditions when banks are already in a weakened condition. Realistically, the use of ex post risk measures places substantial constraints on the size of the penalty that could be levied against a high risk bank. If risk could be detected before a bank's performance has deteriorated, a relatively heavy penalty could be levied that may alter its behavior without jeopardizing its existence. However, levying a large penalty against a bank that is already performing poorly would probably ensure its eventual failure. Such a punitive policy would be analogous to an early closure rule, and so, should not be the basis of a premium policy.

Avery, Hanweck, and Kwast (1985) suggest that one way of dealing with this problem might be to refrain from collecting the full premium penalty from a high risk institution at the time of its difficulties, but retain a contingent claim on the bank's future income if it returns to a healthier condition. During the period when the institution is classified as high risk (but still deemed to be solvent), modest financial penalties could be imposed and supervisory actions taken to reduce the bank's risk profile. However, as the authors note, temporary forgiveness of high premiums reduces the incentives for healthier banks to reduce the chance that they will eventually appear in the high premium category. Regardless, limitations in the ability to detect risk in a timely manner, together with the fact that the FDIC is a public monopolist (for all intents and purposes, banks cannot choose another insurer), argue in favor of assessing a relatively modest risk penalty while a bank is experiencing difficulties.

It was indicated earlier that, to some extent, the current system of supervision may act as a system of implicit risk-related premiums. Conceptually, implicit pricing can accomplish the same ends as explicit pricing. Banks can be dissuaded from having excessive loan concentrations either by charging them higher insurance premiums or by issuing cease-and-desist orders (with appropriate sanctions if the order is not followed). Given this, a question arises as to what a system of explicit prices would add to (or subtract from) the current system or some envisioned system of implicit pricing. (The terms "add to" or "subtract from" are used because some amount of on- site examinations and current supervisory sanctions would be needed even with explicit risk-based premiums.)

While in theory the same ends can be accomplished with either explicit or implicit pricing schemes, there are operational differences in the two approaches. From the regulator's perspective, implicit pricing generally offers some advantages in the form of greater flexibility and discretion. For many of the current forms of implicit pricing, such as letters of agreement and enforcement actions resulting from the examination process, regulators have considerable discretion in tailoring sanctions and solutions to individual cases. Even with a strictly formulated scheme of risk- based capital, regulators would probably be given considerable discretion in setting up compliance timetables for banks that fall below the standard.

Of course, the opposite side of this coin is that implicit pricing would tend to be subjective and sometimes arbitrary. Rules or formulas are often advocated as a way of overcoming these shortcomings and as a way of ensuring that public entities act in an appropriate manner. Thus, explicit pricing formulas would have the advantage of ensuring uniformity and constraining regulators' behavior.

From a bank's perspective, explicit pricing may allow for a more flexible response. There always will be situations where some banks will find it more costly than other banks to meet a given standard. For example, banks that temporarily fall below a capital standard and face relatively high costs in attracting additional capital may find it more cost-effective to pay higher insurance premiums and live with a somewhat lower capital level. With implicit pricing, no such choice exists (except at the regulator's discretion). Thus, an explicit pricing scheme may have the advantage of allowing banks to choose a more efficient means of dealing with a bad situation.

Another operational difference is that a system of risk-related premiums is apt to receive greater scrutiny by regulators, banks, and the public. A system of risk-related premiums would be much more visible than most forms of implicit pricing. Banks would be able to observe directly the price of moving to riskier positions (as defined by the regulator). Because of the directly observable costs, banks may be more likely to scrutinize the formulas used to calculate premiums than they scrutinize the current set of implicit premiums. Moreover, a system of risk-related premiums would provide banks, analysts, and the public with information more suitable for making interbank comparisons of risk. Depending on the method used and its availability to the public, it may be relatively easy for analysts or the media to construct a list of the FDIC's riskiest banks.

There are positive and negative aspects to the increased private and public scrutiny that may accompany explicit pricing. In the short run, the adverse publicity associated with being designated a high risk bank may create liquidity problems and, therefore, may hinder the recovery of potentially viable banks. However, in the long run, the potential for this adverse publicity may increase the deterrent effect of risk-related premiums. This may be particularly important if the financial penalties associated with risk-related premiums are relatively small (initially, this is apt to be the case).

The increased visibility of risk-related premiums also may have a positive effect on the insurer's incentives to correctly assess risks in banking. With an explicit pricing formula, banks and the public would periodically question its appropriateness. While the insurer may be uncomfortable with this increased scrutiny, it would force regulators to continually rethink and revise their risk- monitoring system.

G. PUBLIC COMMENTS ON RISK-RELATED PREMIUMS

Of the comments received concerning the Treasury's Deposit Insurance Report, approximately 25 addressed the subject of risk- related deposit insurance premiums. The concept of an ideal system of risk-related premiums was viewed favorably in the majority of the comments received. That is, the accurate underwriting or pricing of the actual risks undertaken by banks was viewed as desirable, and if feasible, would be advocated as a means of deposit insurance reform.

Of those who commented in detail, reservations concerning the feasibility of a risk-related premium system were expressed. Many of these comments indicated that, while conceptually attractive, the implementation of a risk-related premium system was perceived to be impractical, if not impossible. This perception led a few commentators to reject the concept in its entirety. Others saw the problem of implementing a system of risk-based premiums as currently impractical, and they advised proceeding with caution. An overview of the comments concerning the positive attributes and the shortcomings of a risk-based premium system follow.

Generally, risk-related premiums were thought to enhance the equity and efficiency of the banking system. As well, a system of risk-related premiums was seen as complementary to the risk-based capital or net worth standards. Some commentators preferred a system of risk-related premiums to the current flat-rate system, even when the problems of practical implementation were taken into consideration. For example, some believed that a risk-related system would diminish the perceived inequities of the current system.

Several practical problems concerning the implementation of a risk-related premium system were noted. First, risk-related premiums were seen as being too complex and too costly for practical implementation. In addition, the ability of a risk-related system to effectively alter risk-taking in a timely manner was questioned, as was the regulators' ability to accurately determine levels of risk and their appropriate prices. The comments underscored the concern that a system capable of being implemented would penalize behavior post, rather than modify behavior ex ante.

In some comments, the success of a risk-related system was linked to private sector involvement in the pricing of risk. Suggestions varied from modifying the current system by including a degree of co-insurance (private-sector market discipline) to the establishment of a totally private system of cross-guarantees among banks. Other suggested alternatives to risk-related premiums included increased capital and increased capital combined with co-insurance of deposits.

² As mandated in FIRREA, institutions insured by the Savings Association Insurance Fund (SAIF) were assessed .208 percent of total domestic deposits until December 31, 1990. This assessment will increase to .23 percent from January 1, 1991, through December 31, 1993, and decrease to .18 percent from January 1, 1994, through December 31, 1997. The assessment rate will be set at .15 percent from January 1, 1998, onward. BIF institutions were assessed at .12 percent until December 31, 1990. As of January 1, 1991, the FDIC raised premiums for BIF-insured institutions to .195.

³ The deposit insurance systems of other industrialized countries have premium structures that fall into two categories, neither being a risk-based approach in the sense that an individual institution is assessed a premium on the basis of its risk to the insurance fund. One approach, similar to the current U.S. premium structure, assesses an insured institution at a flat rate of its deposit base. Foreign countries using this approach include West Germany and Japan, whose annual premiums are 0.03 percent of deposits and 0.012 percent of savings deposits, respectively. The other approach involves the assessment of participating institutions based on losses to the insurance fund during the year, with some ceiling on an institution's contribution. France and Italy are countries where this approach is used. There are countries whose approach to funding their deposit insurance systems integrate elements from each of these two approaches. For example, Britain's premium structure requires some initial contribution from its insured institutions, with further assessments being called when necessary.

⁴ Technically, whether or not bankers will move to a riskier position depends on their attitudes toward risk-taking. The introduction of flat-rate pricing reduces the cost of assuming more risk. This price change has the effect of inducing banks to assume more risk. This is referred to as the "substitution effect." However, the price change also creates a wealth effect: banks can earn higher returns at any given level of risk. This increased wealth or income may make some bank managers less willing to accept more risk, even though the price of accepting more risk has been reduced. If this "income effect" dominates, bank managers actually may choose a less risky position. However, most economists believe that the "substitution effect" will dominate over the "income effect."

⁵ The perverse incentives toward risk-taking associated with a flat-rate system will exist regardless of the level of the premium.

⁶ Whether or not the current system of implicit premiums appropriately prices risk, or assesses risk in an ex ante sense, is an open question. The point is that regulation and supervision represent a cost to the depository institution and have the potential to constrain risk-taking.

⁷ In addition to credit risk, banking risks also include interest-rate risk, malfeasance, liquidity risk, and operating risks.

⁸ It should be noted that the difficulties that these information problems present for designing an efficient risk-related pricing system apply equally to a system of explicit or implicit premiums, including the current system of implicit premiums.

⁹ The term incentive-compatible simply means that there are incentives for the insured to choose the premium/attribute combination that is appropriate for their risk class.

¹⁰ The problem here is similar to knowing whether the long-run revenues under the current pricing scheme are adequate to handle the long-run costs. Because of the systemic nature of bank failures, even 57 years of experience cannot tell us with much certainty whether the rate at which the fund is being accumulated is sufficient to meet long-run costs.

¹¹ Some sort of ex post settling-up or extended liability schemes could be termed incentive-compatible as well. These schemes would expose stockholders and management to more of the downside risk associated with alternative investment strategies and their implementation would not depend on accurate actuarial information.

¹² A similar scheme relying on private markets has received some consideration by Congressional staff. In discussing the types of proposals reviewed here, the terms coinsurance and reinsurance sometimes are used interchangeably. While these proposals have some coinsurance attributes, we will refer to them as reinsurance.

¹³ The concept of deposit insurance as a put option could be broadened to include the "right" of the owner to transfer all deposits (insured and uninsured) to the insurer in the event of an insolvency or a failure. This interpretation may be more reflective of current failure policies of the insurer.

¹⁴ For the most recent study on the feasibility of using option pricing to set premiums see Kuester and O'Brien (1990).

¹⁵ These risks are defined as follows:

Credit Risk: The risk that a borrower will be unable to completely fulfill the terms of a debt contract to the lender;

Diversification Risk: The risk to a lender's asset portfolio when assets are concentrated in sectors which are likely to have simultaneous credit risk problems;

Liquidity Risk: The risk that an institution will be unable to retain or attract deposits which are needed to fund outstanding assets;

Interest Rate Risk: The risk to an institution's net interest income and net worth due to adverse movements in interest rates in conjunction with a repricing mismatch between assets and liabilities; and

Operating Risk: The risk to an institution's profitability due to operational inefficiencies.

¹⁶ On December 31, 1990, the OTS published a proposed interest rate risk component for its capital standards for savings associations.

¹⁷ The Office of Thrift Supervision has a comparable rating system, referred to as the MACRO rating, for the thrifts it regulates.

¹⁸ If premiums were based on examination ratings, it would be desirable to examine banks at least once a year. The FDIC now is moving in that direction.

¹⁹ Failure-prediction models can be used for several purposes. Many failure-prediction or problem-bank identification models have been designed primarily as early-warning systems. Early-warning systems assist regulators in identifying potential problems and in better allocating supervisory resources to deal with these problems. Some failure-prediction models also have been designed for the purpose of identifying the causes of past failures, rather than for predicting future behavior (Pantalone and Platt (1987)). Relatively few of these models have been used in a specific risk-related premium proposal. While all of these models may provide useful information for the design of a risk-related pricing scheme, a particular model's applicability will be limited by its intended purpose. Generally speaking, in designing a model for the purpose of setting insurance premiums (versus an early-warning system) one must take greater care to ensure that there is a stable underlying relationship between a particular financial variable and bank risk.

²⁰ Once the parameters of the failure-prediction model have been estimated using historical data, the number of institutions that will be designated as high risk can be varied by simply changing the probability of failure threshold. The threshold level is the dividing line between what would be considered a high-risk bank (or alternatively a potential failure or a problem bank) and a low-risk bank. By lowering the threshold level one can increase the number of actual failures that are designated as high risk, but only at the cost of designating more nonfailures as high risk. In the extreme, one could correctly predict all failures by simply classifying all banks as high risk, but this would defeat the purpose of the model. In the case of the model used in the FDIC's proposal, the ability to correctly classify actual failures was achieved at a cost of rating 20 percent of all banks as high risk.

²¹ Another factor limiting the accuracy of these estimates is the fact that not all banks report accurate data. Examinations often reveal that banks have underestimated the true extent of their problems. Perhaps assessing banks penalties when examinations reveal that they have underreported problems would partially solve this problem.

²² Of course, this will vary from bank to bank. Some banks, particularly large banks, may make a considerable amount of loans to corporate borrowers for which markets generally possess a considerable amount of information, or some banks may have portfolios that are weighted more heavily with marketable securities or loans that are more easily evaluated by markets, such as mortgages.

²³ Of course, if there are externalties or third-party effects that result from bank failures, then the market would underprice risk. But this is another kind of inefficiency than the one being discussed here. With the existence of credible insurance, third party effects are apt to be small.

²⁴ A recent study (not yet published) by Kuester and O'Brien (1990) examines this very question. They conclude that adding market information to accounting data increases the ability to predict future bank performance, and vice versa. However, they also conclude that the option pricing methodology should not be used as the sole basis for setting deposit insurance premiums.

²⁵ Under the current rebate system it is likely that effective premiums also will rise during recessionary periods. However, with the current system the burden of higher premiums is shared evenly by all banks.

REFERENCES

Avery, Robert; Gerald A. Hanweck; and Myron L. Kwast. "An Analysis of Risk-Based Deposit Insurance for Commercial Banks," Proceedings of a Conference on Bank Structure and Competition, Federal Reserve Bank of Chicago, 1985, pp. 217-50.

Baer, Herbert. "Private Prices, Public Insurance: The Pricing of Federal Deposit Insurance," Economic Perspectives, Federal Reserve Bank of Chicago, September/October 1985, pp. 45-57.

Benston, George J.; Robert A. Eisenbeis; Paul M. Horvitz; Edward J. Kane; and George G. Kaufman. Perspectives on Safe and Sound Banking: Past, Present, and Future, Cambridge, MA: MIT Press, 1986.

Bierwag, Gerald O. Duration Analysis: Managing Interest Rate Risk, Cambridge, MA: Ballinger Publishing Company, 1987.

Black, F., and M. Scholes. "The Pricing of Options and Corporate Liabilities," Journal of Political Economy, May 1973, pp. 637-59.

Brickley, James A., and Christopher M. James. "Access to Deposit Insurance, Insolvency Rules and the Stock Returns of Financial Institutions," Journal of Financial Economics 1986, pp. 345-71.

Federal Deposit Insurance Corporation. Deposit Insurance for the Nineties: Meeting the Challenge, Unpublished Draft, Washington, D.C., 1989.

Federal Deposit Insurance Corporation. Deposit Insurance in a Changing Environment, Washington, D.C., 1983.

French, George E. "Measuring the Interest-Rate Exposure of Financial Intermediaries," Banking Review, Federal Deposit Insurance Corporation, Fall 1988, pp. 14-27.

Furlong, Frederick T. "Changes in Bank Risk," Weekly Letter, Federal Reserve Bank of San Francisco, March 25, 1988.

Furlong, Frederick T., and Keeley, Michael C. "Bank Capital Regulation and Asset Risk," Economic Review, Federal Reserve Bank of San Francisco, Spring 1987, pp. 20-40.

Goodhart, C.O.E. "Why Do Banks Need a Central Bank?" Oxford Economic Papers, 1987, pp. 75-89.

Goodman, Laurie S., and S. Shaffer. "The Economics of Deposit Insurance: A Critical Evaluation of Proposed Reforms," Yale Journal of Regulation 1984, pp. 145-62.

Hirschhorn, Eric. "Developing a Proposal for Risk-Related Deposit Insurance," Banking and Economic Review, Federal Deposit Insurance Corporation, September/October 1986, pp. 3-10.

Horvitz, Paul. "The Case Against Risk-Related Deposit Insurance Premiums," Housing Finance Review, July 1983, pp. 253-63.

Kuester, Kathleen A., and O'Brien, James M. "Market-Based Deposit Insurance Premiums: An Evaluation," Paper presented at the Federal Reserve Bank of Chicago Conference on Bank Structure and Competition, May 9-11, 1990 (The proceedings are forthcoming).

Maisel, Sherman J. Capital Adequacy in Commercial Banks, Chicago: The University of Chicago Press, 1981.

Merrick, John J., and Anthony Saunders. "Bank Regulation and Monetary Policy, " Journal of Money, Credit and Banking, November 1985, Part 2, pp. 691-717.

Merton, Robert C. "An Analytic Derivation of the Cost of Deposit Insurance and Loan Guarantees: An Application of Modern Option Pricing Theory," Journal of Banking and Finance, June 1977, pp. 3-11.

Pantalone, C. C., and M. B. Platt. "Predicting Commercial Bank Failure Since Deregulation," New England Economic Review, Federal Reserve Bank of Boston, July/August 1987.

Peltzman, Sam. "The Costs of Competition: An Appraisal of the Hunt Commission Report," Journal of Money, Credit and Banking, November 1972, pp. 100-4.

Pyle, David H. "Pricing Deposit Insurance: The Effects of Mismanagement," Working Paper, Federal Reserve Bank of San Francisco, October 1983.

Ronn, Ehud I., and Auinash K. Verma. "Pricing Risk Adjusted Deposit Insurance: An Option Based Model," Journal of Finance, September 1986, pp. 871-95.

Short, Eugenie D., and Gerald P. O'Driscoll, Jr. "Deregulation and Deposit Insurance," Economic Review, Federal Reserve Bank of Dallas, September 1983, pp. 11-22.

Thompson, James B. "The Use of Market Information in Pricing Deposit Insurance," Journal of Money, Credit and Banking, November 1987, pp. 528-37.