Each depositor insured to at least $250,000 per insured bank

A Two-Window System for Banking Reform
by Frederick S. Carns*

Bank specialness has eroded . . . If you went down the bank-asset balance sheet today, youwould probably say that . . . only about $1.5 -- $2 trillion is special anymore (in the sense thatdeposit insurance protection is justified for the underlying activities). Why is the safety net stillso big (protecting nearly $5 trillion of assets in insured depositories), and shouldn't we thinkabout shrinking it? Robert E. Litan 1

Despite strong earnings, improving asset quality, healthy growth and low failure rates in thebanking industry of late, banking reform remains a popular topic among lawmakers, bankers,scholars and industry observers. This seems appropriate, given the record number of depository institution failures during the 1980s, record losses for the deposit insurance funds and the changing nature of banking due to innovation and increased competition. Traditional bank intermediation, the distinguishing feature of which is the conversion of liquid deposits into illiquid loans, or 'liquidity transformation,' is shrinking as a share of total economic activity (Figure 1 illustrates two components of this shrinkage). As banks seek new ways to remain profitable without repeating the results of the 1980s, it is useful to consider whether the structural design of the banking industry and the accompanying safety net are appropriate for the new environment. Such considerations underlie the continuing interest in banking reform, and this interest has generated a variety of reform proposals.

The common goals driving most reform proposals are to remove unnecessary restrictions on banking organizations that restrain competition for financial services and to eliminate undue risk to the deposit insurance funds (the Bank Insurance Fund (BIF), and the Savings Association Insurance Fund (SAIF)). The proposed reforms vary considerably, due in part to different diagnoses of the structural problems in the industry, different conceptions of banking's role in the economy, and different priorities regarding the policy objectives of federal deposit insurance.

This article discusses one version of a 'two-window' approach to banking reform. 2 The general approach is based upon principles set forth in the FDIC's 1987 study of reform issues, Mandate for Change, and it proceeds from the premise that the industry faces potential long-run problems due to restrictions on the product lines and ownership of banking organizations. 3 The proposed solution is 'two-window' banking, whereby customers could choose between a window offering insured deposits and a window offering various uninsured investments issued by nonbank affiliates. This approach is intended to allow banking organizations to engage in a wide array of 'nonbank' activities without benefit of insured-deposit funding, while strictly circumscribing the deposit insurance safety net to cover only traditional banking activities.

Structural Framework

Basic Elements
There are three primary features of the specific two-window structure discussed in this article: (1) Bank risk-taking is confined to traditional intermediation by restricting the activities that may be funded with insured deposits: illiquid loans and high-quality securities held for liquidity purposes are to be the primary assets of insured entities; (2) activities that are prohibited for insured banks are permitted for uninsured affiliates (perhaps subsidiaries) within the same banking organization, but the insured bank is insulated from nonbank risks through separate capitalization and a set of reinforcing 'firewalls' to maintain legal and financial separation; and (3) all ownership and product-line restrictions now applied to banking organizations are lifted so that new capital can be attracted from other industries and banking organizations can compete more directly with nonbank and foreign firms.

The 'two-window' approach occupies a unique middle ground between the major types of banking-reform proposals. So-called 'narrow-bank' proposals, which are discussed below, typically require that banks invest insured deposits entirely in liquid financial assets so that risk to the deposit insurance funds is minimal. This reflects a presumption that traditional bank lending can occur safely and sufficiently outside of the financial safety net. 4 The two-window proposal does not share this presumption. But unlike the group of modest reform proposals at the other end of the spectrum, there is a presumption underlying the two-window approach that banking reform must proceed beyond a mere 'turning of the dials' on existing policy instruments; for example, adjusting capital standards, accounting standards, or levels of depositor protection. In response to the shrinking role of traditional bank intermediation in today's economy, the two-window proposal offers a redesigned safety net that will shrink in a corresponding fashion.

Stated otherwise, the size and shape of the safety net would be adjusted as necessary under the two-window system to ensure that deposit insurance protects only the 'traditional' type of bank intermediation. 5 Perhaps more important than reducing risk for the deposit insurance funds, this measure is intended to rationalize the scope of safety-net coverage so that any risk exposure of the deposit insurance funds serves a legitimate economic purpose.

In response to the shrinkage of traditional intermediation, the two-window proposal also permits banking organizations to conduct nonbank activities of their choice within a structure that is designed to insulate insured institutions from nonbank risks. This provision likewise occupies a middle ground, between the status quo that features 'consolidated supervision' of the bank holding company and limited entry by holding-company affiliates into nonbank financial activities, and a system of 'universal banking' in which a wide variety of nonbank activities might be conducted within the legal banking entity itself, funded with insured deposits. 6

The two-window proposal confronts policymakers with some fundamental issues that often are neglected in discussions of banking reform. The remainder of this article examines one version of a two-window system and discusses the fundamental issues that arise in connection with these features. 7

Comparisons With the
'Narrow' Bank

The concept of two windows derives from the same philosophy that underlies the so-called 'core-bank' or 'narrow-bank' proposals for banking reform (these will hereafter be referred to as 'narrow-bank' proposals). 8 A common purpose is to reduce the risk exposure of the deposit insurance funds (hence, the taxpayer) to the minimum level that is consistent with adequate safety-net protection for deposits. The common approach is to restrict the permissible uses of insured deposits and to remove what are argued to be unnecessary restrictions on the activities of banking organizations. The major differences between the two-window system and various versions of the narrow-bank proposal relate to the proper means of achieving a largely common set of goals. The two-window approach departs from most narrow-bank proposals in the criteria used for determining acceptable uses of insured deposits, the restrictions imposed on nonbank-affiliate activities, and the structural design of the banking organization as a whole. These three major differences will be considered briefly in turn.

Activities Funded With Insured Deposits. Narrow-bank proposals tend to use risk measures alone for determining what are acceptable uses of insured funds 9. Only the safest activities are permitted in the insured entity, and thus the proposed 'bank' typically resembles a money-market mutual fund: its deposits are transactions accounts that are serviced through investments in Treasury securies and high-grade commercial paper. In the two-window approach, the decision whether an activity should be funded with insured deposits involves more than risk measurement. Also important is whether there exists a legitimate economic rationale for extending the safety net to the activity in question.

For example, the traditional banking business of liquidity transformation arguably fulfills a unique and important economic purpose (the arguments are considered under Risk to the Deposit Insurance Funds, below). Accepting this premise for the moment, and recognizing that such intermediation is intrinsically susceptible to destabilizing bank runs that may impose significant social costs, there is a legitimate economic rationale for extending deposit insurance to this function. 10 It is much more difficult to provide a defensible rationale for extending the safety net to products or services that are traded in established markets with no inherent susceptibility to costly market failures (for example, foreign-exchange or government securities markets). 11

According to this logic, insured banks should be limited to activities that are consistent with the traditional business of banking: issuing deposits and other relatively liquid claims, clearing payments, and investing in short- and intermediate-term illiquid loans such as commercial loans and, perhaps, some consumer loans, as well as some high-quality investments for liquidity. 12 In the two-window system, all other activities would be conducted in separately capitalized subsidiaries or affiliates that are not funded with insured deposits. If there are important synergies between traditional banking functions and certain nontraditional activities, exceptions would be permitted.

Of course, there would be no avoiding some arbitrary judgments and operational difficulties in implementing this type of two-window system. How illiquid must a loan be to qualify for funding with insured deposits, and how should this be measured? Precisely how much investment for liquidity purposes should be permitted? There are no easy answers to these and a host of similarly practical questions. 13 In recognizing that the two-window system would inject some additional arbitrariness and complexity into banking regulation, however, it should be recognized that some would be removed as well. For example, there would be no artificial percentage limits on holding-company investments, and such investments would not be restricted to those activities 'closely related to banking.' Other examples will be discussed in what follows but, in the final analysis, the two-window alternative poses a trade-off of the existing set of arbitrary decisions for a different set. >BR>

Moreover, there are alternative versions of the two-window system that would pose fewer practical difficulties. There is nothing inherent in the basic concept of 'two windows' that requires a strict adherence to a liquidity-transformation test for determining insurable activities; thus, at least three different 'two-window' approaches to reform can be identified:

• Passive Approach or Status Quo. It might be determined to continue to permit all activities that currently are conducted by insured banks. New activities could be placed outside the insured bank and funded through an uninsured window if they are not closely related to the credit-granting process. The safety net still would shrink if traditional intermediation continues to shrink, and it could not grow arbitrarily.
  

• Intermediate Approach. Deposit-taking, lending and the holding of some marketable securities would be permitted for the insured bank, without making judgments about the degree of liquidity transformation that is accomplished. All other activities would be conducted in separately capitalized affiliates and funded through an uninsured window.
  

• Activist Approach Based Upon Liquidity Transformation. This option, emphasized throughout this article, requires a willingness to make judgments about the proper types of loans and deposits for insured banks, i.e., about the degree of liquidity transformation that would justify deposit insurance protection, and a willingness to relocate current activities as needed to uninsured affiliates. Clearly, it poses more practical problems than the other options.
  

Restrictions on Nonbank-Affiliate Activities. In most narrow-bank proposals, affiliates of the bank are restricted in their lines of business to some set of financial activities. 14 The operating principle embodied in the two-window approach is that insured banks should be free to affiliate with both financial and nonfinancial enterprises provided that: The bank is well-capitalized upon completion of any affiliation transaction; nonbank affiliates are separately capitalized; and supervisors are satisfied that the resulting entity will not operate in a manner that is abusive to the bank. Supervisors should have authority to preclude affiliation if there are concerns about the quality or character of management in either the bank or the prospective affiliate. Moreover, after an affiliation occurs, there should exist clear supervisory authority to audit both sides of any transaction between the bank and its affiliate and to require reporting as necessary from both parties to the transaction. 15

The general rule for bank supervision in a two-window system is that all advances of bank funds should be governed by an 'arm's-length' requirement. The two-window framework includes firewalls and regulatory safeguards to ensure legal and financial separation. 16. For the two-window structure to function safely and properly, financial flows from the bank to its affiliates must be strictly limited.

The Structure of the Banking Organization. So long as firewalls exist, capital flows are properly restricted, and proper supervisory measures are taken to ensure arm's-length dealings among insured banks and their nonbank affiliates, any corporate structure is permissible under two-window banking. Narrow-bank proposals typically impose a specific structure (usually resembling a holding company) along with a set of regulations and supervisory measures that apply to the banking organization as a whole. 17 The two-window structure is designed specifically to avoid this type of oversight and control: Consolidated supervision has the potential to impede cost efficiency and artificially reduce the value of the banking franchise, thus inhibiting the industry's ability to attract new capital. In order to enter banking under current law, nonbank firms typically must subject their entire organizations to federal supervision. As noted, the two-window system would remove this deterrent, replacing it with functional supervision of separate corporate units and a system of firewalls (discussed under Supervisory Issues, below).

Risk to the Deposit Insurance Funds Under 'Two Windows'

Again, the two-window system envisioned in this article would restrict the banking entity to traditional activities, except where important synergies with nontraditional activities have been demonstrated. The major risks posed to the deposit insurance funds thus would be the traditional types of banking risks, primarily related to credit quality, as well as the possibility that problems at the nonbank affiliates may adversely affect the insured entity. The effectiveness of firewalls will be discussed in a later section. At this point it is sufficient to note that the risk-reducing effects of a two-window approach are unclear.

However, risk reduction is not the sole aim of the two-window approach. Narrow-bank structures would achieve greater risk reduction for the insurance funds than would the two-window structure. The logic underlying the two-window approach suggests that this extra risk exposure is worthwhile because, as mentioned earlier, the traditional intermediation function of banks is considered to be essential for efficient credit allocation in the economy. More specifically, evidence suggests that an important component of the economy's production depends primarily upon banks for financing. 18 Such production is undertaken by firms that, for various reasons, tend to incur prohibitively high costs in attempting to convey information about their investment projects to potential investors. These firms are likely to find that they cannot successfully borrow by issuing their own securities, even when the projects to be financed are viable and productive.

Similarly, not all such firms with viable projects may be able to borrow from finance companies or other substitute lenders. While these lenders offer commercial loans just as banks do, they do not offer liquid claims to investors. This may limit their fundraising ability, so that it is unclear whether nonbank lenders could finance all of the viable projects that depend upon such lending in our economy.

At first glance, this last point appears counterintuitive. Why could not nonbank lenders attract sufficient funds from savers simply by offering higher rates of return on their claims? Perhaps they could, but there may be limits to the willingness of savers to sacrifice liquidity for return. Benston and Kaufman (1988) cite evidence from the money-market mutual fund experience indicating that those funds originally guaranteeing a return, instead of a share's nominal value, eventually were forced via competition to switch to nominal-value guarantees. The conclusion, as described by Phillips (1994), is that the 'public wishes to have a guaranteed nominal value, even at the expense of return.' If so, it is equally plausible that, at some point, individuals will opt for liquidity, even at the expense of return. Intermediation thus may be reduced if lenders are unable to issue liquid claims. 19

More fundamentally, there are no significant historical examples of strong economies in which illiquid lending has been financed primarily by private institutions issuing illiquid debt. Because such institutions have the distinct advantage over banks of reduced susceptibility to runs, it seems reasonable to believe that fundraising limits may explain their relative obscurity across time and across different economies. While today's information revolution undoubtedly has expanded the fundraising possibilities for such institutions, the foregoing discussion suggests that it is unclear whether they alone could finance all of an economy's viable projects requiring illiquid loans; 20 some liquidity transformation by banks may be necessary. As a result, the threat of bank runs may impose large social costs because it impairs this special banking function and precludes the funding of potentially productive projects.

Under this view of credit markets, there is a presumptive market failure to allocate credit appropriately in the absence of deposit insurance -- or some equivalent protection -- for traditional bank intermediation. 21 To remove safety-net protections from the traditional banking function could easily create more social costs than benefits if this view is correct.

Such a view is controversial, and a thorough treatment is beyond the scope of this article. 22 Most narrow-bank proponents do not share the presumption that the market failure cited above is sufficiently important to justify deposit insurance protection for traditional banking functions. Absent this presumption, it is highly unlikely that the two-window structure could be viewed as optimal, since narrow banks would expose the taxpayer to far less risk.

Regardless of which view is correct, this two-window approach serves as a reminder that there is likely to be no 'free lunch' with deposit insurance reform. Too often when changes in deposit insurance coverage are proposed, there appears to be an implicit assumption that the risk exposure of the insurance funds can be reduced costlessly. The two-window proposal reminds policymakers that real economic costs are likely to be associated with reforms that reduce the potential for bank intermediation. The relevant issue, of course, is whether these costs are likely to be smaller or larger than the benefits of the proposed reforms. Given the difficulty of estimating the benefit/cost ratio with any accuracy, opinions on this issue reflect a high degree of subjective judgment.

Profitability and Diversification of Risks

Expanded powers per se may change the risk exposure of the deposit insurance funds: They will create the opportunity for greater profits as well as greater losses within banking organizations. Expanded powers could reduce risks for banking organizations if appropriate diversification rules are followed, but no reliable prediction is possible for the banking industry as a whole. New investment opportunities will not be exploited identically at different institutions and similar investment choices may influence risk and return differently at different institutions, depending upon the characteristics of existing portfolios, the size of new investments relative to the existing portfolios, and managerial efficiency in capturing any economies of scope offered by new powers. 23

Although the net effect of new powers via two-window banking is impossible to predict with any precision, it also may be largely irrelevant. The more relevant fact is that well-managed banking organizations currently face artificial obstacles to diversifying appropriately and maximizing returns for their shareholders; these institutions are likely to become safer and more competitive with expanded powers, potentially benefiting consumers as well as shareholders and taxpayers. 24 Poorly managed institutions, on the other hand, likely will suffer the consequences of a new set of poor choices if expanded powers become available, and many of these firms may fail.

The deposit insurance funds may benefit under a two-window system to the extent that nonbank risks would be removed from the insured entity and to the extent that scope economies, synergies or innovations resulting from new powers would salvage some banking franchises that now are endangered. Capital also would be free to move to nonbank uses within the organization when banking returns are low, thus stabilizing earnings. More-stable earnings may allow organizations to provide better support for their banking units during difficult times, should the owners wish to do so. 25 Meanwhile, any new risks posed by expanded powers would be borne outside the safety net, if the firewalls and regulatory safeguards embedded in the two-window structure are effective. 26 If effective, this structure dispenses with any need for concern about the effects of new powers. Effectiveness is considered in the following section.

Supervisory Issues Under Two Windows

The supervisory issues that arise under the two-window structure may be grouped into two categories. The first category consists of familiar concerns that apply to any proposed expansion of banking powers. Foremost among these concerns are the possibility of anticompetitive effects, including undue concentrations of power, and the potential problems associated with conflicts of interest, including abusive 'tie-in' sales and improper uses of inside information. Such concerns have been debated extensively and, thus, are relegated here to an Appendix that provides a brief summary. The second category of issues deals with the effectiveness of firewalls in maintaining the safety and soundness of insured banks. Because the two-window proposal relies so heavily upon firewalls to protect the insured bank without impairing the ability of the larger organization to capture economies of scope and other synergies, this category is particularly important for evaluating the relative merits of two-window banking. The remainder of this section focuses on firewalls.

Firewalls for Safety and Soundness. For purposes of deciding the merits of a two-window structure, determining whether insured financial institutions can be effectively insulated from risks posed by their nonbank affiliates is more critical than selecting a particular set of permissible activities for those nonbanks. If corporate separation can be achieved within a banking organization and safety-net protections can be confined to the insured bank as required under the two-window framework, then the selection of activities to be conducted by nonbank affiliates is of second-order importance.

Economic theory and formal empirical evidence are ambiguous regarding the prospects for effective separation. 27 Similarly, the abundant anecdotal evidence on both sides of the question offers no firm conclusions. 28 The following discussion contrasts opposing perspectives on the feasibility of corporate separation under a two-window system.

Affiliations may carry two types of risks for an insured bank and, ultimately, the insurance funds: (1) the bank may endanger its own financial health by directly or indirectly assisting a troubled affiliate, and (2) public confidence may be shaken by problems plaguing a nonbank affiliate, with negative repercussions for the bank (such as higher funding costs, bank runs, etc.). For convenience, these risks may be termed 'financial' and 'market' risks, respectively. 29

As noted earlier, financial separation implies separate funding, no commingling of assets, and an arm's-length requirement for all advances of bank funds to affiliates. It follows that loans or services obtained by an affiliate from an insured bank should be on terms comparable to those available for nonaffiliates. Financial separation also requires that a bank does not unduly transfer assets to, or purchase bad assets from, an ailing affiliate. 'Firewalls' refer to the behavioral rules and restrictions that promote the fulfillment of these requirements for financial separation among affiliates.

Banks currently are subject to lending limits, statutory restrictions on transfers of funds, dividend restrictions, and restrictions on amounts and terms for insider loans. 30 Nonetheless, abuse of an insured financial institution by individual owners to benefit outside activities has resulted in a number of bank failures. While this may occur under the strongest of firewall systems, it is relevant to note that state laws governing corporate separation are uneven. The two-window system would potentially close state law loopholes that facilitate such abuse by requiring a minimal set of uniform firewalls for all insured institutions affiliated in any manner with nonbank firms. Despite this, it is reasonable to expect that the incidence of such abuse will rise if there are expanded opportunities for affiliation.

There is no doubt that strong incentives exist to manage the various components of an organization as an integrated whole. For example, funding costs can be lowered for any given affiliate if creditors believe that the strength of the entire organization stands behind the affiliate's obligations. To the extent that stockholders' returns can be elevated or risks can be better controlled by managing the various affiliates as an integrated entity, there will be strong incentives for management to do so. However, experience confirms that regulators also can create strong incentives that affect the bottom line for shareholders. The firewalls separating holding-company affiliates (Sections 23A and 23B of the Bank Holding Company Act) have insulated many banks effectively against repercussions from the demise of nonbank affiliates, and such firewalls always can be strengthened if necessary with stiffer penalties for noncompliance.

A fair assessment of the anecdotal evidence would suggest that existing firewalls work well in the usual course of business but occasionally fail when there are serious problems in one or more parts of the organization. As Flannery (1986, p. 223) has noted: 'The policy question is therefore whether the possibility of . . . relatively extreme behavior should have an important weight in making regulatory decisions.'

A complicating factor is that the anecdotal evidence comes from a regulatory regime that differs from the one envisioned under a two-window system. Firewalls have, to date, operated within the context of consolidated supervision for bank holding companies by the Federal Reserve Board (see note 6). Because the two-window framework does away with consolidated supervision, it is unclear whether firewalls would be equally effective in that environment.

A counterpoint is that consolidated supervision has served as a signal to the market that regulators expect affiliates to be managed as integrated entities. This could account for those past instances in which insured banks have been punished by the market for problems arising in their nonbank affiliates. Thus, it is unclear whether the market reaction would be similar in an environment where regulators demand corporate separation and show a willingness to strengthen firewalls as necessary to enforce it.

As this discussion indicates, complete financial separation is impossible without 'market' separation. Economic theory and empirical evidence suggest that some amount of market risk is probably unavoidable for banking units within financial holding companies, even if firewalls work well. 31 For example, to the extent that customers seek a complete set of financial products and services from an organization, a nonbank affiliate's performance in delivering one element of the set will also affect the demand for the bank's products and services. 32 Similarly, because empirical evidence indicates that the market uses new information regarding one firm's performance to infer changes in the stock values of unrelated firms in the same industry, it is reasonable to expect that the performance of affiliates in businesses closely related to banking will influence also the market's valuation of a bank. 33

Even if affiliates conduct activities that are unrelated to banking, the market is unlikely to ignore signs of poor management anywhere in the organization. If the parent selects management of questionable quality for a nonbank affiliate, the quality of bank management may also come under suspicion. This will be expressed through higher funding costs for the bank, with or without the presence of effective firewalls. Moreover, entities within any conglomerates are likely to be chosen such that their risk-return characteristics are complementary. 34 While firewalls may prevent direct interrelationships among affiliates, they cannot prevent the indirect connections established through the judicious use of financial theory and statistical evidence in exploiting covariance or other properties of asset returns. The market is likely to recognize these indirect connections and understand their implications for the bank's performance. Correspondingly, a bank's portfolio choices are likely to be influenced by the risk-return characteristics of its affiliates, even in the presence of firewalls.

In short, it seems clear that some exposure to market risk will be unavoidable for banks under a two-window system. Such exposure already exists for holding-company banks, but the two-window framework places no restrictions on the types of affiliates that banks may choose. As a result, the mix of affiliates within a two-window organization could have diverse and complex consequences for a bank's exposure to market risk. More generally, the net effect on the banking industry's safety and soundness is impossible to predict with confidence.

Given this ambiguity, some argue that a reliance upon firewalls to protect the insured bank would be too risky. 35 They note that perfect insulation is not possible, and that attempts to strengthen firewalls may merely negate any economic advantages resulting from expanded powers. It is not certain that firewalls can be simultaneously strong enough to maintain an acceptable level of risk for the deposit insurance funds and flexible enough to allow banking firms to exploit new opportunities or economies of scope associated with expanded powers. It is argued that these uncertainties signal a continuing need for consolidated supervision of banking firms in order to catch any problems that may slip through cracks in the firewalls. Quite apart from the substantive merits of consolidated supervision, such skeptics note that this concept is firmly embraced in the revised Basle Concordat; hence, an important international agreement could be threatened by any reforms that suggest a weakened U.S. commitment to consolidated supervision. 36

The case for two-window banking does not rest upon a denial of these complications and uncertainties. It rests upon a belief that there are crucial flaws in the structural design of the banking industry, including the financial safety net, with potential consequences that justify the risks associated with two-window reforms. More concretely, it appears that the safety net is now and, in the absence of reforms, increasingly will be extended to activities for which it is not essential. This creates artificial incentives in the marketplace and raises questions concerning the future ability of the deposit insurance system to support the safety net. The list of nontraditional activities conducted by insured banks (foreign-exchange operations, bond underwriting, off-balance-sheet activities) continues to grow virtually without regard to the design and purpose of federal deposit insurance. No legitimate economic rationale exists for extending safety-net protections in this manner. This argument, and the corollary concern over undue risk to the insurance funds, motivate the redesign of the safety net to protect only traditional intermediation in the two-window system.

An additional 'two-window' argument for structural reform is that restrictions on bank ownership and affiliations currently restrain competition in the financial marketplace. Consolidated supervision appears as an integral part of this problem. It potentially deters entry into banking by nonbank firms and narrowly restricts activities that may be conducted in nonbank units of a holding company; thus, it reduces the flexibility of banking organizations to adapt to a rapidly changing environment and otherwise meet the diverse needs of borrowers and savers.

Conclusions

The two-window proposal, like the recent versions of the narrow-bank proposal, was motivated by the experience of the 1980s. Enactment of the Federal Deposit Insurance Corporation Improvement Act of 1991 (FDICIA) has changed the landscape significantly.

The FDIC's loss exposure has been reduced by the least-cost failure-resolution requirement of FDICIA and, perhaps, by the incentives created through prompt corrective action for undercapitalized institutions (PCA). As well, the Budget Act of 1993 amended FDICIA by providing that domestic-office depositors of federally insured U.S. financial institutions would occupy a preferred position in the failure-resolution process with respect to foreign-office depositors and general creditors. Taken together, these provisions may reduce significantly the maximum potential failure-resolution costs of the FDIC.

Moreover, FDICIA may have greatly reduced the potential magnitude of the deposit insurance subsidy by addressing the 'too-big-to-fail' problem. This aspect of FDICIA has been little noticed and apparently underappreciated. The least-cost resolution requirement generally prohibits the FDIC from extending protection to any failed-bank creditors other than insured depositors, unless such action reduces FDIC costs. An explicit process is created for systemic-risk exceptions, requiring agreement among the FDIC, Federal Reserve, and the Secretary of the Treasury in consultation with the President, and FDICIA provides for a special assessment on the industry to pay for any losses exceeding the least-cost amount. Thus, it can be argued that the process for determining whether large banks get special treatment is similar to, but more systematic than, the process that might be used for other large firms. 37 It would now be difficult to argue that the deposit insurance system per se is responsible for any perception in the marketplace that some banks are 'too big to fail.'

On balance, then, FDICIA appears to weaken substantially the case for those reforms aimed primarily at reducing risk to the deposit insurance fund and eliminating the 'too-big-to-fail' doctrine.

Despite this, and despite the strong performance of the banking industry recently, it can be argued that structural reform is necessary because of an overextension of the safety net and unnecessary restrictions on bank ownership and affiliations. On these bases, the two-window proposal merits serious consideration.

In particular, a two-window system avoids potential problems posed by reforms that rely upon 'narrow banks' and/or consolidated supervision of banking organizations. Narrow-bank proposals ignore social costs that may be associated with removing safety-net protections for traditional bank intermediation and do not address concerns that consolidated supervision deters entry into banking and restricts the flexibility of banking organizations to meet the needs of customers.

There are legitimate concerns regarding the effectiveness of firewalls in a two-window system. Given expanded opportunities for bank affiliation with nonbanks, it is reasonable to expect new temptations that could lead to abuse of insured banks. The effectiveness of firewalls in preventing such abuse has not been tested in an environment of functional supervision. On the other hand, to the extent that FDICIA successfully limits the costs associated with bank failures, reliance on firewalls may pose less risk to the deposit insurance funds and the taxpayers. Moreover, it should be considered that such a risk may be worth taking in order to rationalize the financial safety net, bring it under control, and reduce barriers to entry in the financial-services industry.

APPENDIX

Competition, Concentration of Power, and
Conflicts of Interest With Expanded Powers

Decreased Competition and Conflicts of Interest. A longstanding fear associated with expanded product powers and nationwide branching is that a comparatively small number of large organizations may ultimately dominate the financial marketplace. This raises the specter of localized monopolies, destabilization of the financial system in the event of a single bank failure, and disproportionate control over the nation's financial resources on the part of a few large institutions.

Such concerns are real but difficult to assess. While banking organizations may be larger and fewer under a two-window system, they also may be better diversified. Thus, while any given bank failure under this system may be more costly and destabilizing, there may be fewer failures. Similarly, a decline in the number of banks need not mean that fewer banks will be competing in any given market. Technological advances in information processing and communications are constantly reducing the costs associated with entry into new markets. This suggests that if inadequate competition creates excess profits in a given market, new entrants are likely to be attracted until above-normal profits are competed away. To the extent that expanded powers would increase either actual or potential competition, this would provide a safeguard against the costly excesses typically associated with concentrations of power. 38

For these reasons, it appears unlikely that concentrations of power would produce any serious economic problems. Under a two-window system, however, other legitimate concerns may arise in connection with this issue, including potential effects of financial concentration on the political process. Before committing to any structure that allows for large-scale consolidation in financial services, lawmakers may wish to consider whether remedies would be available to offset any untoward implications of concentration. 39

Another longstanding fear associated with expanded powers is that conflicts of interest will multiply unmanageably as product lines grow, leading to widespread abuses. The primary concern for bank supervisors is that insured institutions may be jeopardized by attempts to combat financial problems elsewhere in the organization. For example, banks may be called upon to make unsound loans, capital injections or asset purchases for the benefit of a nonbank affiliate. This concern is addressed with firewalls and other regulatory safeguards in the two-window system, as discussed in the text. The remainder of this Appendix examines two conflicts that may harm consumers: 'tie-in' sales and improper uses of inside information.

Tie-ins occur when a business entity attempts to condition the sale of a particular product or service upon the purchase of another of the entity's products or services. Typically, problems with tie-ins can be traced to inadequate information or weak competition. For example, extensions of credit to bank customers may be conditioned upon the customers obtaining additional services from a bank, its parent company, or one of its affiliates. In such cases, customers may enter into undesirable tie-in arrangements if uninformed of the consequences of their actions or if unaware of their alternatives. If no viable alternatives exist, customers may feel compelled to purchase products or services they do not want. The fear is that expanded powers will create more opportunities and stronger incentives for tie-in sales, so that problems of this type will become more widespread.

Implicit in the two-window approach is the view that tie-in problems resulting from inadequate information are best controlled by requiring full disclosure of costs, alternatives, and other pertinent facts. Likewise, tie-in arrangements arising from inadequate competition are best addressed through policies that will strengthen competition in the financial marketplace. Under the two-window system, there would be no geographic or product-line restrictions to protect firms that saddle their customers with undesirable tie-in arrangements. The potential competition created by these new ground rules would act as a deterrent to tie-in abuses.

A related concern is that violations of a bank's fiduciary responsibilities might increase along with the number of products and services offered by affiliates. As banking organizations obtain new powers, bankers are likely to gain access to new types of confidential information. Skeptics argue that such information is likely to be misused, because there is a conflict of interest whenever one entity acts as both a promoter of services or products and a disinterested financial advisor.

Such arguments historically have centered on bank involvement in securities underwriting and, in part, served as the basis for passage of the Glass-Steagall Act. Recent evidence indicates, however, that, far from harming investors, issues underwritten by bank affiliates prior to 1933 were, on average, of higher quality and better performing than those underwritten by independent investment banks (Krozner and Rajan, 1994).

More generally, potential conflicts of this type within banking typically have been neutralized successfully in the past. Because banks generally have succeeded in creating an effective 'Chinese Wall' between their commercial lending and trust departments, for example, it would seem that similar steps could be taken if banking organizations are permitted to engage in activities that grant them access to other types of confidential information. Should the level of abuse prove unacceptable under the two-window system, however, lawmakers could enact additional safeguards and stiffer penalties. Given the changing dynamics of competition in the market for financial services, product-line restrictions should probably be viewed as a last resort.

REFERENCES

Aharoney, Joseph and Itzhak Swary. Contagion Effects of Bank Failures: Evidence from Capital Markets. Journal of Business (July 1983): 305-22.

Becketti, Sean and Charles Morris. Are Bank Loans Still Special? Federal Reserve Bank of Kansas City Economic Review (Third Quarter 1992): 71-84.

__________. Reduced Form Evidence on the Substitutability Between Bank and Nonbank Loans. In Federal Reserve Bank of Chicago Proceedings: A Conference on Bank Structure and Competition. Chicago: n.p., 1994.

Benston, George J. The Separation of Commercial and Investment Banking: The Glass-Steagall Act Revisited and Reconsidered. New York, NY: Oxford University Press, 1990.

__________ and George G. Kaufman. Risk and Solvency Regulation of Depository Institutions: Past Policies and Current Options. Federal Reserve Bank of Chicago Staff Memoranda, #88:1.

Bernanke, Ben S. Nonmonetary Effects of the Financial Crisis in the Propagation of the Great Depression. American Economic Review (1983): 257-76.

__________ and Mark Gertler. Agency Costs, Net Worth and Business Fluctuations. American Economic Review (June 1989): 14-31.

__________. Banking and Macroeconomic Equilibrium. In New Approaches to Monetary Economics, William A. Barnet and Kenneth J. Singleton eds. Cambridge, England: Cambridge University Press, 1987.

Billett, Matthew Thayer, Mark J. Flannery and Jon A. Garfinkel. The Effect of Lender Identity on a Borrowing Firm's Equity Return. University of Florida Working Paper, June 1993.

Blair, Christine E. Bank Powers and the Separation of Banking from Commerce. FDIC Banking Review (Spring/Summer 1994): 28-38.

Bryan, Lowell L. Bankrupt: Restoring Health and Profitability to Our Banking System. New York, NY: Harper Business, 1991.

Carns, Frederick S. Should the $100,000 Deposit Insurance Limit Be Changed? FDIC Banking Review (Spring/Summer 1989): 11-19.

Carns, Frederick S., Arthur J. Murton and Lynn A. Nejezchleb. No Free Lunch in Deposit Insurance Reform. In The Financial Safety Net for Consumer Savings: Policy Options for the 1990s. Proceedings of a public-policy forum at Pace University, November 1989, pp. 51-59.

Committee on Banking Regulations and Supervisory Practices. Principles for the Supervision of Banks' Foreign Establishments (The Revised Basle Concordat), May 1983.

Corrigan, Gerald E. Financial Market Structure: A Longer View. Federal Reserve Bank of New York Annual Report (February 1987).

__________. Reforming Federal Deposit Insurance, Modernizing the Regulation of Financial Services, and Maintaining the International Competitiveness of U.S. Financial Institutions, testimony before the Committee on Banking, Housing and Urban Affairs, U.S. Senate, May 3, 1990.

Cornyn, A., G. Hanweck, S. Rhoades and J. Rose. An Analysis of Corporate Separateness in BHC Regulation from an Economic Perspective. In Federal Reserve Bank of Chicago Proceedings: A Conference on Bank Structure and Competition. Chicago: n.p., 1986.

Diamond, Douglas W. and Philip H. Dybvig. Banking Theory, Deposit Insurance, and Bank Regulation. Journal of Business (January 1986): 55-68.

__________. Bank Runs, Deposit Insurance, and Liquidity. Journal of Political Economy (June 1983): 401-19.

Federal Deposit Insurance Corporation. Current Issues '94. Division of Research and Statistics. Washington, DC: Federal Deposit Insurance Corporation, 1993.

__________. Deposit Insurance for the Nineties: Meeting the Challenge. (Unpublished Draft.) Washington, DC: Federal Deposit Insurance Corporation, 1989.

__________. Mandate for Change: Restructuring the Banking Industry. Washington, DC: Federal Deposit Insurance Corporation, 1987.

__________. Report to the Congress on the Findings and Recommendations Concerning the Two-Window Deposit System Proposal. Washington, DC: Federal Deposit Insurance Corporation, September 1992.

Flannery, Mark J. Contagious Bank Runs, Financial Structure and Corporate Separateness Within a Bank Holding Company. In Federal Reserve Bank of Chicago Proceedings: A Conference on Bank Structure and Competition. Chicago: n.p., 1986.

French, George. Banking in Transition, a conference on The Financial System in the Decade Ahead, Jerome Levy Economics Institute of Bard College, April 14, 1994.

French, George and Eric Hirschhorn. Insulation of Banking from Nonbanking: An Empirical Investigation. In Federal Reserve Bank of Chicago Proceedings: A Conference on Bank Structure and Competition. Chicago: n.p., 1988.

Goodhart, C.A.E Money, Information, and Uncertainty. Cambridge, MA: MIT Press, 1989, especially Chapter 8.

Gorton, Gary and George Pennacchi. Nonbanks and the Future of Banking. In Federal Reserve Bank of Chicago Proceedings: A Conference on Bank Structure and Competition. Chicago: n.p., 1992.

Kaufman, George G. and Larry R. Mote. Is Banking a Declining Industry? A Historical Perspective. Federal Reserve Bank of Chicago Economic Perspectives (May/June 1994): 2-21.

Krozner, Randall S. and Raghurm G. Rajan. Is the Glass-Steagall Act Justified? A Study of the U.S. Experience with Universal Banking Before 1933. American Economic Review (September 1994): 810-32.

Litan, Robert, E. What Should Banks Do? Washington, DC: Brookings Institution, 1987. @REFER = Lummer, Scott L., and John J. McConnell. Further Evidence on the Bank Lending Process and the Capital Market Response to Bank Loan Agreements. Journal of Financial Economics, 25 (1989): 99-122.

McConnell, C.E. and W.S. Marcias. Bank Loans to REITs: How Serious the Problem? Keefe, Bruyette, and Woods, Inc., May 1975.

Meyer, Paul. Will Expanded Banking Powers Raise Prices and Reduce Efficiency? mimeographed, Federal Deposit Insurance Corporation, 1988.

Murton, Arthur J. Bank Intermediation, Bank Runs, and Deposit Insurance. FDIC Banking Review (Spring/Summer 1989): 1-10.

Phillips, Ronnie J. The New Regulatory Environment and the Functional Approach to Financial Reform: Establishing Separate Monetary and Financial Service Companies, manuscript, Jerome Levy Institute, June 1994.

Pierce, James L. The Future of Banking. New Haven, CT: Yale University Press, 1991. @REFER = Securities Industry Association. Statement before the Subcommittee on General Oversight and Investigations, Committee on Banking, Finance and Urban Affairs, U.S. House of Representatives, May 1, 1990.

Seidman, L. William. Proposals to Establish a `Core' or `Narrow' Bank, testimony before the Committee on Banking, Finance and Urban Affairs, U.S. House of Representatives, June 18, 1991.

Shockley, Richard and Anjan V. Thakor. Information Content of Commitments to Lend in the Future: Theory and Evidence on the Gains from Relationship Banking, Indiana University Working Paper, January 1993.

Swary, Itzhak. Continental Illinois Crisis: An Empirical Analysis of Regulatory Behavior. (Unpublished Manuscript.) Hebrew University of Jerusalem, January 1985.

Tobin, James. The Case for Preserving Regulatory Distinctions. In Restructuring the Financial System, Federal Reserve Bank of Kansas City (1987), pp. 167-83.

U.S. Department of the Treasury. Modernizing the Financial System: Recommendations for Safer, More Competitive Banks, a Report to Congress. Washington, DC: U.S. Department of the Treasury, February 1991.

U.S. General Accounting Office. Bank Powers: Insulating Banks from the Potential Risks of Expanded Activities. Washington, DC: U.S. General Accounting Office, April 1987.