Skip Header

Federal Deposit
Insurance Corporation

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

Home > News & Events > Speeches and Testimony

Speeches and Testimony







10:00 A.M.
MARCH 5, 1997

Mr. Chairman and members of the Subcommittee, I appreciate this opportunity to present the views of the Federal Deposit Insurance Corporation on financial modernization and related issues. I commend you, Mr. Chairman, and Congressman Kanjorski for placing a high priority on the need to modernize and strengthen the nation's banking and financial systems.

On behalf of the FDIC, I also want to express again our gratitude to you, to members of this Subcommittee, and to other members of the Congress for passing legislation providing immediate financial stability to the Savings Association Insurance Fund (the SAIF). The health and stability of the financial industry are in the interest of everyone -- participants, regulators, banks and thrifts. Sound deposit insurance funds contribute to that health and stability.

The Deposit Insurance Funds Act of 1996 (the Funds Act) capitalized the SAIF and solved its immediate financial problems. The Funds Act also recognized the need for a merger of the deposit insurance funds. The FDIC strongly supports a merger of the Bank Insurance Fund (the BIF) and the SAIF as soon as practicable. The SAIF insures far fewer, and more geographically concentrated, institutions than does the BIF, and, therefore, faces potentially greater long-term risks.

A merger of the BIF and the SAIF is a necessary component of a solution to long-term structural problems facing the thrift industry, and consequently the industry's deposit insurance fund. A combined BIF and SAIF would have a larger membership and a broader distribution of geographic and product risks; a combined fund would be stronger than the SAIF alone. Under the Funds Act, Congress has made the merger of the BIF and the SAIF contingent upon there being no more savings associations.

I am pleased to have this opportunity to testify on financial modernization against the backdrop of two fully capitalized deposit insurance funds and record bank earnings. Although final numbers are still being tabulated, we expect to announce next week that annual earnings for commercial banks surpassed $50 billion for the first time in 1996. Average equity ratios are at their highest levels in more than 50 years, and nonperforming assets are well under one percent of total assets, the lowest level in the 15 years that banks have reported nonperforming assets.

Private-sector thrifts have earned more than $6 billion each year since 1991, when the industry returned to profitability. Thrift earnings in 1996 may have exceeded the record $7.6 billion of 1995 if thrifts had not paid a special assessment to capitalize the SAIF. Equity ratios remain near 40-year highs, and nonperforming assets are down to approximately one percent of total assets, the lowest level in the seven years that thrifts have reported nonperforming assets.

Only six insured institutions, with aggregate assets of $220 million, failed in 1996. Also, the number and aggregate assets of institutions on the FDIC's "problem" institution list have declined sharply over the past five years. At the end of 1991, there were 1,426 institutions with total assets of $819 billion on the problem list. This was the highest level of problem-list assets in the history of the FDIC. Since 1991, the problem list has declined steadily. As of September 30, 1996, only 125 institutions, with assets of $15 billion, were on the list -- a fraction of the highest level.

In recent years, banks and thrifts have benefited from continued economic expansion and low inflation. These favorable conditions have produced strong loan demand and have contributed to wider net interest margins. The resultant growth in revenues has enabled banks and thrifts to reduce their inventories of bad assets while boosting profits.

Although banks have been making record profits recently, evidence suggests that increasing numbers have turned to somewhat riskier investments as they have lost business to competitors. Loan-loss rates in today's favorable environment remain significantly higher than in pre-1980 nonrecessionary periods. Bank performance has varied greatly during the past ten years. Figures 1 and 2 illustrate annual returns on assets and net charge-offs as a percentage of average loans since 1960. The volatility of earnings in the 1980s is readily apparent, as is the relationship between recessionary periods and net charge-offs. In the past ten years, the banking industry achieved both its highest annual return on assets (1.20 percent in 1993) and its lowest return on assets (0.10 percent in 1987) since 1934.

As we consider financial modernization, current favorable economic conditions provide both an opportunity and a challenge. We have the opportunity to merge the deposit insurance funds at a time when both funds are capitalized fully. The challenge for us is to recognize that good times may not last forever. We must evaluate any financial modernization proposal by determining whether it will operate effectively during times of stress for financial institutions.

As the deposit insurer, the FDIC brings a unique perspective to the financial modernization question. Events of the past decade have demonstrated how costly bank failures can be for the insurance fund, for communities across America, and for our economy. The BIF and the banking industry spent approximately $36.4 billion to resolve failing banks from 1980 through 1994. The General Accounting Office has estimated that, from 1986 through 1995, the thrift crisis cost an estimated $160 billion to resolve (including tax benefits); approximately $132 billion of this amount was paid by the taxpayers. Thus, it is imperative that we learn from the past and proceed deliberately as we contemplate a substantial expansion of powers available to banking organizations.

Let me turn now to a discussion of the issues before us today. First, my testimony will discuss briefly the need for financial modernization. Second, I will outline lessons the FDIC has learned from studying the banking and thrift crises of the 1980s and early 1990s. I conclude with a discussion of guiding principles for financial modernization, including an analysis of whether permitting nonbanking activities to be conducted through subsidiaries of banks would constitute an unwarranted expansion of the federal safety net.


Modernization of the financial system is necessary to achieve an efficient and competitive financial services industry able to meet current and future challenges. The financial markets have changed dramatically since the 1930s when many of our nation's laws governing financial services were enacted.

To a greater extent than ever before, businesses have been bypassing traditional financial intermediaries to access capital markets directly. Large corporations now frequently meet their funding needs by issuing commercial paper, debt securities and equity, rather than by borrowing from banks. The shrinking role of banks in lending to business is illustrated by the declining proportion that bank loans represent of nonfinancial corporate debt. This share declined from approximately 28 percent in 1975 to 21 percent at year-end 1995.

In addition to their shrinking role as providers of traditional financial intermediation services, banks and thrifts also are experiencing increasing competition from nonbanking firms that now offer financial products that once were the exclusive domain of banks and thrifts. Banks also have grown much less rapidly than other financial intermediaries during the past ten years. For example, from 1986 through 1995, banking assets grew at an average annual rate of 5.7 percent, compared to growth rates of 19.0 percent and 8.5 percent for mutual funds and pension funds, respectively.

This relative decline in market share and relatively slower growth do not paint the complete picture. Traditional market share measures, which are based on asset holdings, generally do not reflect the growing importance of bank income from off-balance-sheet products and services. The rise in the noninterest income share of bank earnings indicates less reliance on traditional lending activities. From 1985 to 1995, banks' noninterest income increased 165 percent, to $82 billion, compared to growth in interest income of only 22 percent, to $302 billion. This also indicates that banks are innovating and adapting to a changing marketplace.

Nevertheless, banks have experienced a relative decline in market share and relatively slower growth. Financial modernization should strengthen banking organizations by allowing diversification of income sources and better service to customers, which would promote an efficient, competitive, and safe evolution of the U.S. financial markets.


When I became FDIC Chairman, I initiated a project that I called the "Lessons of the Eighties" (the History Project) to answer the question, "Did we, as bank regulators, learn the correct lessons from the banking and thrift crises of the 1980s and early 1990s?" At the time the History Project began, the banking and thrift industries were recovering from the worst period of failures since the 1930s. It is essential that we thoroughly analyze and understand the factors that led to those crises in order to be prepared for the problems that could occur in the future.

The lessons we have learned thus far could be instructive to this Subcommittee in its deliberations on financial modernization. My testimony will focus on two broad lessons in particular: (1) geographic and product constraints on insured institutions can result in inadequate diversification of income sources; and (2) rapid expansion of insured institutions into unfamiliar activities, without adequate supervision, can have undesirable consequences. In the context of this hearing, what these lessons demonstrated to us is that we cannot attribute all the losses from the failures of financial institutions in the 1980s and early 1990s to economic events or to poor management of depository institutions. A significant share of the responsibility must be assigned to overly restrictive laws, changes in the law that provided little time for adjustment, poorly planned deregulation and deficiencies in the supervisory process.

Geographic and Product Constraints

Geographic and product restrictions have constrained the activities of U.S. depository institutions for much of their history. Although these restrictions insulated them from competition, at least for a time, they also hindered banks from expanding their sources of income and from developing portfolios that reflected product and geographic diversity.

The impact of product restrictions most notably is seen in the experience of savings and loan associations. For years, thrifts were limited to providing only savings deposits and home mortgages to their customers. This created an inherently unstable situation -- of borrowing short-term deposits to fund long-term mortgages -- that became apparent in the late 1970s and early 1980s when short-term interest rates rose above long-term rates. This was the beginning of the savings and loan crisis that in time led to the demise of the Federal Savings and Loan Insurance Corporation. Although banks provided a broader range of products, commercial lending was their primary focus, and this market also came under pressure during this period. As the commercial loan market declined, and the commercial paper and junk bond markets grew, banks were forced to find new sources of income since they were restricted in their ability to adapt to their customers' needs.

The impact of geographic restrictions is evident in the relatively high failure rates in states where branching was prohibited or severely restricted, such as Texas, Kansas and Illinois. This impact also is evident in the vulnerability of banks to regional economic problems. Because most U.S. banks serve relatively narrow geographic markets, regional and sectoral recessions frequently have had a severe impact on them. There were four major regional or sectoral economic downturns during the 1980s and early 1990s, and each resulted in increased bank failures.

The first economic downturn accompanied the decline in farm prices in the early 1980s. Agricultural prices increased steadily during the boom years of the 1970s. This ended, however, in the late 1970s as interest rates soared, significantly increasing farm operating costs. At the same time, export demand decreased sharply due to worldwide competition. These events contributed to a collapse in real farm income in 1980. Then, as inflation declined, land values collapsed. Ultimately, this downturn took its toll on many agricultural banks. In 1985, these banks accounted for 48 percent of bank failures.

The second downturn occurred in Texas and other major energy-producing states in the Southwest following the collapse of oil prices in 1981 and again in 1985. Texas banks, for example, had rapidly increased their commercial and industrial loans in the 1970s as strong worldwide demand for oil and OPEC-restrictions on supply brought on a sharp rise in oil prices. Following the oil-generated cycle were wide swings in real estate activity that contributed significantly to the downturn in the economies of Texas and other states in the Southwest and the sharp rise in bank failures in this region. The experience in Texas and certain other states was aggravated by the large number of new banks chartered during the 1980s, and by the fact that newer banks failed more frequently than existing institutions.

Boom and bust conditions in real estate activity also contributed to the third downturn, in the Northeast. In this regional recession, mutual institutions that had converted to the stock form of ownership failed with greater frequency than mutuals that had not converted because of the difficulty of employing excess capital successfully. Of the mutuals that converted in the mid- and late-1980s, 21 percent failed during the period 1990 through 1994. This compared with eight percent of all mutuals that existed as of the end of 1989 and had not converted. The final downturn was a recession in California -- a state without geographic branching restrictions -- following defense cutbacks in the early 1990s. In this downturn we found higher failure rates among smaller and newer banks that were tied more closely to their local economies. The large California banks that operated statewide were less affected.

The lesson we draw from these events is that attempts to ensure the safety and soundness of insured institutions by limiting market competition ultimately fail. In the long run, geographic constraints and product restrictions do not insulate depository institutions from competitors, who will eventually find ways to enter markets.

Congress eliminated many geographic constraints by enacting the Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994. Over the years, Congressional action, agency initiatives and court decisions slowly have removed some product constraints. Nevertheless, barriers remain, such as the Glass-Steagall Act, which limit the opportunities for financial institutions to diversify and to respond quickly and efficiently to changes in the marketplace. To maintain the safety and soundness of the financial system, institutions must be allowed to diversify.

Use of Expanded Powers and Supervision

In response to the deepening crisis in the thrift industry, the early 1980s were dominated by actions to deregulate the product and service powers of insured depository institutions. The resultant rapid expansion of insured institutions into unfamiliar activities without adequate supervision resulted in significant losses for the industry. For example, many banks and thrifts adopted a highly aggressive posture with respect to commercial real estate lending. Large increases in the early 1980s in real estate investment produced a boom in commercial construction and in bank and thrift commercial real estate lending. Further stimulus was provided by legislation that greatly enhanced the after-tax returns on real estate investment and by the expansion of nonresidential lending powers of savings and loan associations implemented through banking legislation.

Other factors led to increased risks in the 1980s as well. During this time, chartering standards were lowered, the inappropriate use of brokered deposits increased, and capital standards were reduced for thrift institutions.

Relaxed chartering policies led to approximately 3,300 new banks being chartered from 1980 through 1990. Of these new institutions, 15 percent subsequently failed; this compared with a 7.7 percent failure rate for banks in existence as of year-end 1979. The influx of new charters created markets that were overbanked, which led to more competition for good loans. This, in turn, created incentives for banks to loosen underwriting standards and take on more risk. The increase in charters also diluted the available management talent necessary to operate a sound institution.

Insolvent thrifts were allowed to use brokered deposits to stay in operation and, indeed, to grow their assets or engage in new activities that could not have been funded through traditional sources. At the same time, regulatory accounting standards for thrifts were adopted allowing many to operate with little or no capital. These institutions, with little or no capital on the line, and access to fully-insured brokered deposits, in many cases took extraordinary risks that resulted in large losses to the old Federal Savings and Loan Insurance Corporation fund, which was not managed by the FDIC, and, ultimately, to taxpayers.

While powers were being expanded, insufficient attention was being paid to safeguards against risky behavior. In the late 1970s and early 1980s, regulators increased their reliance on off-site monitoring and prioritized examinations to focus primarily on problem banks. This was attributable in part to efforts to limit the size of the federal government. As a result, intervals between examination cycles for healthy banks increased on average from annually to as long as three years, and even longer for some institutions, and the number of examiners was reduced. From 1979 to 1984, examination staffs declined by nearly 20 percent at the FDIC and the Office of the Comptroller of the Currency, while the Federal Reserve's examiner staff increased slightly. Additionally, state examiner ranks declined 12 percent during this period. These actions ultimately weakened the ability of bank supervisors to detect and to respond to problems as failure rates began to soar.

The lesson we learned from these events is that deregulation must be accompanied by adequate safeguards and strong supervision and monitoring by the regulators. Unfortunately, in the 1980s this did not occur. In addition, during that period, legislation was passed in a crisis situation without a full understanding of the consequences of the changes being undertaken.

Diversification of income sources by depository institutions remains a desirable goal and will contribute to stronger, more competitive financial markets. With these lessons in mind -- and in the absence of crisis conditions -- we have the opportunity to design an appropriate analytical framework that addresses competitive as well as supervisory issues.


Any financial modernization proposal must balance numerous public policy goals. Financial reform must ensure the safety and soundness of insured depository institutions and the integrity of the deposit insurance funds. It also must allow insured depository institutions to generate sufficient returns to attract new capital essential for normal growth and expansion into new areas. To achieve these goals, banking organizations must be able to compete on an equitable basis with other businesses and to evolve with the marketplace, consistent with safety and soundness. Equally important, concerns about the potential for credit judgments to be made on preferential terms to affiliated companies or other conflicts of interest between banking and nonbanking affiliates and the effects of undue concentration in the economy must be addressed. Moreover, any financial modernization proposal must be examined for its effect on small communities, isolated markets, and customers of insured depository institutions.

In this context, it is useful to have a framework for analysis against which to judge the merits of the proposal. Several key questions must be answered. First, what activities should be permitted, including should commercial firms be allowed to own insured financial institutions? Second, where should the activity be housed within the corporate structure? Third, what safeguards are necessary to protect the insured entity and the deposit insurance funds? Finally, how should the activity be regulated?


First, with limited exceptions, a banking organization should be permitted to engage in activities that are financial in nature. The exceptions would consist of those activities that: (1) pose significant safety and soundness concerns; or (2) harm consumers or small businesses.

Easing the broad range of restrictions on activities of banking organizations beyond those that are financial in nature should proceed very cautiously. While affiliations between banking and commercial firms could benefit the financial system and the economy by permitting a wider deployment of capital, two types of concerns argue for a deliberate approach.

First, while banking organizations have expertise in managing financial risks, most have little experience managing some of the activities that would be permissible under various legislative proposals under discussion. Some savings associations have affiliated with commercial firms under the unitary savings and loan holding company structure, but this experience has been limited. Moreover, the insured institutions involved in these relationships have not been broad-based banking intermediaries or active lenders to commercial firms, but have specialized primarily in real estate lending. As such, the potential effects of the combination of a major commercial firm and a major commercial bank remain subject to conjecture. The history of the 1980s discussed above, with respect to expanding the powers of thrifts into areas in which they had no prior experience provides a clear example of the risks of going too far. A dramatic change of affiliations between banks and commercial firms could not easily be undone.

The second area of concern involves the possibility that combinations of commercial firms and banks may result in undue concentrations of economic wealth and political power. Concern about such concentrations has been a major theme in banking legislation since the early years of our nation's history, and is partly responsible for a distinctive feature of the U.S. banking system as compared to those of other nations -- the large number of separately owned banking organizations. As I will discuss later, in times of economic stress affiliations sometime lead to conflicts of interest. These potential conflicts of interest are of particular concern with respect to the affiliation of a large bank with a large commercial firm.

There are limited precedents under the Edge Act, applicable to U.S. banking activities abroad, for noncontrolling investments by subsidiaries and affiliates of U.S. banks in commercial companies outside the United States. These may be a reasonable starting point for evaluating whether a basket of noncontrolling investment opportunities involving banks and commercial firms would permit a more cautious foray into banking and commerce affiliations.


Second, a banking organization should have flexibility to choose the corporate or organizational structure that best suits its needs, provided safeguards protect the insurance funds and prevent expansion of the federal safety net. There are two organizational structures with which we have experience in the United States that can be used to combine traditional commercial banking with new activities. These are: (1) conducting each activity in separate organizations owned and controlled by a common "parent" organization (the "bank holding company" model); and (2) conducting each activity in a separate organization owned and controlled by a commercial bank (the "bank subsidiary" model). A third model -- the conduct of both activities within the same entity (the "universal banking" model) -- has been used in some other developed countries, although not with unmitigated success in recent years. We believe that universal banking is not a model that would best fit the dynamic financial marketplace in the United States or provide sufficient protection for the deposit insurance funds against the effects of potential conflicts of interest between banking and nonbanking functions in an insured entity or prevent the unwarranted expansion of the federal safety net.

The Bank Holding Company Model. Since the adoption of the Bank Holding Company Act of 1956, one of the primary methods of expanding permissible activities beyond those associated with traditional commercial banking has been through formation of affiliated entities under the bank holding company umbrella. Within this framework, banking organizations have been permitted to engage in an increasing array of financial and related services.

The advantages of the bank holding company model include:

    Providing a good framework for monitoring transactions between insured and uninsured affiliates and for detecting transfers of value that could threaten the insured institution; and

    Maintaining a meaningful corporate separation between insured and uninsured affiliates to ensure that the insured bank is not held responsible for the losses from uninsured activities.

The disadvantages of the bank holding company model include:

    There are no cross-guarantee provisions for nonbank affiliates, and in the past banks have failed while there was still value in the holding company. Thus, if a bank were to get into trouble, the earnings from new activities might not accrue to the benefit of the bank or the insurance fund.

    In distressed situations, the parent will have the incentive to transfer or divert value away from the insured bank, leaving greater losses for the FDIC if the bank ultimately fails.

    The holding company model requires bank owners to establish and maintain an additional corporation. This may add costs, inefficiencies, complexity and, in some cases, an additional regulator. This may be particularly burdensome for small banks.

Bank Subsidiary Model. The bank subsidiary model is exemplified by FDIC bona fide subsidiaries. The FDIC has permitted bona fide subsidiaries of insured nonmember banks to engage in securities activities since December 1984 (12 CFR 337.4). A bona fide subsidiary of an insured nonmember bank must meet certain capital standards. The operations of the subsidiary must be physically separate and distinct from the operations of the bank. As well, it must maintain separate accounting and other corporate records and observe corporate formalities, such as separate meetings of board of directors. It must share no common officers or employees with the bank and must compensate its own employees. A majority of the board of directors of the subsidiary must be composed of persons who are neither directors nor officers of the bank. It must conduct business in a way that informs customers that the subsidiary is separate from the bank, and that its products are not bank deposits and are not insured by the FDIC nor guaranteed by the bank. Additionally, restrictions are placed on loans, extensions of credit, and other transactions between an insured bank and its securities subsidiary. From a practical perspective, there has been much less experience with the bona fide subsidiary form of organization than with the bank holding company form.

Analytically, there are several factors that make the bank subsidiary approach different from the bank holding company model. The advantages of the bank subsidiary approach include:

    With appropriate safeguards, having earnings from new activities in bank subsidiaries lowers the probability of failure and thus provides greater protection for the insurance fund than having the earnings from new activities in bank holding company affiliates. The reason for this is that diversification often leads to less volatile earnings. Indeed, as discussed earlier, one of the conclusions of our History Project was that a lack of product and geographic diversification contributed to the bank and thrift failures of the late-1980s and early-1990s. Moreover, the upside from new activities in a bank subsidiary accrues to the bank while the downside is limited to the bank's investment in the subsidiary, which as I will discuss later, with appropriate safeguards in place, can be limited to excess regulatory capital. Thus, on average, allowing a bank to put new activities in a bank subsidiary lowers the probability of failure and provides greater protection to the insurance funds. For the FDIC as deposit insurer, this is an extremely important benefit of the bank subsidiary structure.

    The insured institution, rather than the parent, controls the allocation of excess capital of the organization. This may mean that in making corporate investment decisions, greater weight is given to the needs of the insured institution.

However, on the negative side:

    While corporate separateness could conceivably be achieved regardless of organizational structure, the bank holding company structure may provide a measure of greater separation, although it is possible to develop adequate corporate separateness rules for the bank subsidiary structure.

Bank Subsidiary Structure and Safety and Soundness. While the experience of the FDIC with bona fide securities subsidiaries of insured nonmember banks has been limited, these subsidiaries generally have not posed safety and soundness concerns. Only one FDIC-supervised institution owns a subsidiary actively engaged in the full range of securities activities permitted by the FDIC, but over 400 insured nonmember banks have subsidiaries engaged in more limited securities-related activities. These activities include management of the bank's securities portfolio, investment advisory services, and acting as a broker-dealer. With one exception, none of these activities has given cause for a significant safety and soundness concern.

There has been one failure of an insured institution supervised by the FDIC that conducted securities activities through a subsidiary. While not the sole cause of the failure, the business relationship with the securities subsidiary added to the cost of the failure. The bank made a substantial unsecured loan that was used to benefit the securities subsidiary.

Expansion of the Federal Safety Net. Some have argued that the existence of the federal safety net -- consisting of deposit insurance, and access to the Federal Reserve's discount window and payments system -- provides banks with funding advantages that could be passed on to bank subsidiaries, thereby resulting in an undesirable expansion of the federal safety net to activities for which it was not intended. The presumed existence of such advantages and the ability of banks to pass them to their subsidiaries have led some to express a preference for the bank holding company structure over the bank subsidiary structure.

I will discuss this subsidy issue in some detail later in the testimony. However, the evidence shows that, if banks receive a net subsidy from the federal safety net, it is small, and that both the bank holding company structure and the bank subsidiary structure would inhibit the passing of any net subsidy that does exist out of the insured bank. Thus, the potential expansion of the federal safety net is not a reason to prefer one organizational structure over the other.

Existing Bank Activities. The activities that banks currently conduct should be left undisturbed. To require that these activities be moved to a subsidiary of either the bank or the holding company, in the absence of compelling public policy reasons, could cause unnecessary disruption and contribute to market inefficiencies. Moreover, if banks historically have conducted the activities in the insured institution with minimal negative consequences, there is no compelling reason to require that such activities be conducted in a subsidiary or an affiliate. A combination of flexibility and sound regulation has contributed to the successful development of the U.S. financial system, and these key elements should be present in any proposal for reform.


The third principle of financial modernization is that safeguards should prohibit inappropriate transactions between insured institutions and their subsidiaries and affiliates. If these safeguards are inadequate or the resources are unavailable to enforce them, the deposit insurance funds, the financial system, and the public could suffer. Transactions between an insured institution and a related firm pose several risks. First, an insured institution may be used to benefit a related firm inappropriately. As examples, this could occur through unwarranted fees paid to an affiliate or subsidiary, or through excessive direct equity injections to a subsidiary, or perhaps upstreaming of excessive dividends to a parent that are used to inject equity to an affiliate. Second, when an insured institution is in danger of failure, the owners and creditors of related entities may try to extract value from the insured entity to minimize their own losses, thereby increasing losses to the deposit insurance funds. As I will discuss later, the past decade has provided examples of a number of instances where transactions were proposed or consummated that served to advantage a holding company or an affiliate at the expense of a failing insured bank.

Third, the business relationship between the insured entity and its subsidiary or affiliate may create a misperception that the products of the subsidiary or affiliate are federally insured. Finally, there is the danger that the business and operating relationship will cause the courts to "pierce the corporate veil" -- that is, to hold the insured entity responsible for the debts of a subsidiary or affiliate in the event the subsidiary or affiliate fails.

Sections 23A and 23B of the Federal Reserve Act place certain restrictions on transactions between banks and their affiliates. These restrictions are intended to safeguard the resources of federally insured banks against misuse for the benefit of an affiliate of the bank. Section 23A was designed to prevent a bank from risking too large an amount in affiliated enterprises and to ensure that if a bank extends credit to an affiliate, the collateral behind the extension of credit is sufficient to ensure recovery by the bank. Section 23A, therefore, regulates certain "covered transactions" with affiliates of an insured bank and does so primarily in two ways.

First, Section 23A places limits on the dollar amount of loans a bank may make to, or investments it may make in, any individual affiliates, as well as in investments to or in all affiliates. Second, it requires that the loans or extensions of credit meet certain standards as to collateral. In addition, banks generally may not purchase low-quality assets from affiliates.

Section 23B essentially expands Section 23A. Section 23B requires that certain transactions between a bank and its affiliate must be carried out at arm's length, under terms and conditions comparable to the terms of similar transactions between unaffiliated entities. The transactions subject to this comparability requirement include: certain sales of securities or other assets by a bank to its affiliate; payments or provision of services by a bank to its affiliates under a contract; and certain transactions between a bank and a third party where an affiliate acts as a broker or agent.

Any financial modernization proposal should maintain the principles of Sections 23A and 23B and apply similarly appropriate safeguards to dealings between an insured bank and any subsidiaries of the bank engaged in activities not otherwise permissible to the bank. Exceptions under these safeguards should be allowed rarely. Consideration also should be given to requiring timely reporting of intercompany transactions, as the Securities and Exchange Commission currently requires.

In addition, we believe that further safeguards are necessary. Only banks that are "well capitalized" should be permitted to have subsidiaries that engage in activities that are not permissible to the bank itself. The bank's equity investment in the subsidiary should be deducted from the bank's regulatory capital and assets. Furthermore, the subsidiary should not be consolidated with the bank for regulatory capital purposes. These safeguards will ensure that support provided by a healthy bank to a subsidiary is provided through transfers of excess capital -- beyond that required for a well-capitalized bank.

Effective Regulation

The fourth principle for financial modernization is that regulation should be commensurate with risk -- no more and no less. Just as the banking and financial services industries are evolving and changing, so are the requirements of an effective and efficient regulatory system. Individually, the regulatory agencies must have the tools to oversee their respective portions of the financial services industry. Collectively, the agencies should not impose undue burdens on the regulated, but they should act in sufficient concert to ensure the overall integrity and stability of the financial system. These individual and collective concerns argue for an approach that combines functional regulation with some measure of carefully designed, comprehensive oversight.

Currently, banking and financial services institutions are regulated largely as entities or on the basis of function. I have long believed that a greater degree of functional regulation would result in both less intrusive and more effective regulation. Properly implemented, functional regulation could avoid the redundancy that may result from subjecting multi-tiered financial institutions to the overlapping jurisdictions of several agencies. Functional regulation also ensures that an appropriate and consistent degree of expertise is brought to bear on an activity, regardless of the charter or structure of the entity conducting it, and that there are adequate protections in place for consumers and investors.

Improperly implemented, however, functional regulation can be additive, thus increasing the regulatory burden on financial institutions. Moreover, functional regulation may pose the danger of an artificial restructuring of financial operations and services based primarily on function rather than along strategic or market-based lines. Such artificial restructuring of financial operations would undermine the flexibility in corporate structure that should be among the goals of financial modernization.

In view of the increasing complexity of the financial marketplace, functional regulation alone may not be sufficient to ensure effective and efficient oversight of banks and other providers of financial services. Financial organizations increasingly are moving toward a more comprehensive view of the risks posed by their various activities. Some activities, practices, and intercompany dealings that affect the distribution of risk across the organization may go unnoticed if there is singular reliance on a functional approach. This suggests a need for some coordination and attention to interstitial concerns, such as maintaining accurate information regarding all operations in the organization, and monitoring compliance with the rules on intercompany dealings.

In addition, some oversight of the consolidated financial organization may be necessary to address concerns regarding the stability and liquidity of the financial system. Certain regulators have responsibilities that transcend supervision of a particular function or type of institution. For example, the Federal Reserve Board has responsibility for the integrity and liquidity of the payments system and as lender of last resort, particularly in times of financial stress. In carrying out its responsibilities, the Federal Reserve may require a degree of authority that goes beyond supervision of state-chartered banks that are members of the Federal Reserve System. For all regulators there will be a need to coordinate supervisory efforts and ensure the ready availability of adequate information with which to gauge risks to the financial system. Systemic risk, as a joint responsibility of financial industry regulators, imposes needs beyond those involved for any single regulator in conducting its day-to-day activities. The FDIC also needs to be able to protect the deposit insurance funds.

In light of these considerations, I believe that, as part of the effort to modernize our financial system, it is important to define what the appropriate role is for oversight or supervision that spans the entities within a financial organization. As an example, for financial organizations containing large institutions that have access to the payments system, this role could well be filled by the Federal Reserve, the ultimate provider of liquidity for this system. However, this role need not involve full-scope Federal Reserve examinations of nonbanking companies.

For other organizations, with relatively limited access to the payments system, the oversight could be provided by another regulatory agency and may focus on coordinating supervisory efforts, monitoring transactions among affiliates, and otherwise addressing any gaps that appear in the supervisory network.

In neither case would it appear to be necessary, for safety and soundness purposes, to include investment-by-investment or activity-by-activity regulation as part of the oversight of the consolidated organization, provided that risks to the financial system and to the insurance funds are understood and appropriately limited. To the extent that bank-like supervision is imposed unnecessarily on nonbanking entities, this would serve as a barrier to entry and thereby defeat one goal of financial modernization.

In summary, effective and efficient regulation should be an important component of any financial modernization effort. I believe that functional regulation should play an important role in the regulatory scheme and that a concept of "umbrella" supervision can be developed that is consistent with functional regulation and an orderly evolution of the financial system.


Your letter of invitation asked for my "views on the government benefits that banks receive from FDIC insurance, the availability of the discount window, and access to the payments system." As I mentioned earlier, some have argued that the existence of this federal safety net provides banks with funding advantages that could be passed to bank subsidiaries, thereby resulting in an undesirable expansion of the safety net to activities for which it was not intended. The presumed existence of such advantages and the ability of banks to pass them to their subsidiaries have led some to express a preference for the bank holding company structure over the bank subsidiary structure. Because of the importance of this topic, I have asked FDIC staff to analyze these issues in some detail and they have made substantial progress in this effort. Although the analysis is ongoing, they have reached certain conclusions that I will share with you today.

Let me first highlight several points regarding the subsidy issue. These points also will serve to outline the flow of our analysis.

It has long been widely accepted, and the FDIC agrees, that banks receive a gross subsidy from the federal safety net.

However, banks also incur costs, both direct and indirect, that at least partially offset this gross subsidy.

The relevant question for purposes of this discussion, therefore, is not whether banks receive a gross subsidy, but whether banks receive a net subsidy, or more accurately, a net marginal subsidy, after taking account of offsetting costs and restrictions. That is to say, the gross subsidy from raising additional subsidized funds must exceed the additional regulatory and other costs that result from raising those funds before banks can pass a subsidy to the affiliated organizations, including bank subsidiaries.

It is very difficult to measure directly whether banks receive a net subsidy. However, on balance, the evidence indicates that if a net subsidy exists, it is very small.

Moreover, if a small net subsidy exists, firewalls, such as those that require a bank's equity investment in a subsidiary to be deducted from the bank's regulatory capital, and the requirements of Sections 23A and 23B of the Federal Reserve Act, serve to inhibit a bank from passing a net subsidy to a subsidiary or to an affiliate of the holding company.

These firewalls are not impenetrable under all circumstances, but neither are they in a bank holding company structure. The available evidence indicates that both structures will work equally well in inhibiting a bank from passing a net subsidy to a subsidiary.

Allowing a bank to put new activities in a bank subsidiary diversifies a bank's income stream. The bank benefits from the earnings of the subsidiary and with appropriate firewalls, the downside risk can be limited to excess regulatory capital -- above well-capitalized levels -- invested in the subsidiary. Thus, the bank subsidiary structure lowers the risk to the insurance funds and may actually lower any subsidy that arises from deposit insurance.

Given these facts, we have concluded that allowing banks to conduct financial activities in a bank subsidiary does not represent an undue expansion of the federal safety net. Banking organizations should be able to apply their sound business judgments to choose how best to organize their activities.

Now, having outlined the essentials of my analysis, let me proceed to discuss the subsidy issue in some detail.

Sources of the Gross Subsidy

It is widely recognized that banks receive a gross subsidy from the federal safety net. In terms of funding costs, this means that, for any given level of capital, banks can borrow funds at a lower interest rate than they could absent the safety net. There are three sources of the gross subsidy that commercial banks and thrift institutions enjoy: deposit insurance, the discount window, and access to the payments system. Let me briefly explain how a gross funding advantage arises from each source.

Deposit Insurance. Deposit insurance lowers the cost of insured deposits for banks. If banks were to pay a "market premium" for this insurance, the lower cost of funds would not in and of itself constitute a subsidy. For example, municipalities often purchase municipal bond insurance to enhance municipal bonds. The savings, in terms of lower yields on the bonds, exceed the cost of purchasing the insurance (otherwise municipalities would not buy it). Nonetheless, the municipalities are not receiving a subsidy. By purchasing insurance from a AAA-rated company, they merely are capturing some of the risk premium they would have had to pay to get investors to purchase their riskier securities.

However, deposit insurance differs from market-provided insurance in two important ways. First, the premium is not set by the market. As I will discuss later, it is very difficult to measure what a market rate for deposit insurance should be. Second, there are two parts to deposit insurance: the insurance funds (the BIF and the SAIF), which the industry has paid for, and a call on the full faith and credit of the United States Government. Measuring the value of this call is quite difficult. Since the call is only "in the money" if one of the insurance funds goes bankrupt, its value varies over time with the health of the banking industry and the insurance funds. Since this call is similar to a standby letter of credit provided by the government, the fee for the call would be paid to the U.S. Treasury. There has never been an explicit charge for this call.

The experience during the recent banking crisis was that the FDIC used deposit insurance funds to resolve bank failures until 1991 when it borrowed working capital from the U.S. Treasury. Those funds were repaid in 1993 with interest, resulting in no net cost to the U.S. taxpayer for deposit insurance for banks insured by the FDIC. Nevertheless, the availability of credit from the U.S. Government for deposit insurance purposes results in insured institutions being able to borrow in the marketplace at lower costs than uninsured financial institutions.

The Discount Window. The Federal Reserve's discount window provides credit to solvent but illiquid banks. Although discount window loans must be fully collateralized, its existence in times when other sources of credit may not be available under any terms, means this backup source of credit provides a subsidy to depository institutions. Moreover, it is not necessary for a depository institution to borrow from the window for some benefit to accrue to the institution. Hence, banks may have the ability to fund riskier and less liquid asset portfolios at a lower cost and on a much larger scale than otherwise would be possible. As with deposit insurance, it is extremely difficult to quantify the subsidy provided by access to the discount window, because the value varies with the health of individual institutions and the banking industry.

Payments System Access. The Federal Reserve District Banks operate Fedwire. Through Fedwire, banks and thrifts with reserve or clearing accounts at a Federal Reserve Bank may transfer balances to other institutions that have similar accounts. For many institutions, payments made on a given day may exceed that day's opening balance. When a bank's account goes into a negative position, a daylight overdraft occurs. Because Fedwire transfers are "guaranteed" by the Federal Reserve at the time they are initiated, the Federal Reserve assumes the intra-day credit risk that a participating bank will not have enough funds at the end of the day to discharge its obligations.

Regulatory, Legislative, and Market Developments Have Lessened the Gross Subsidy

While the federal safety net continues to provide banks with a gross subsidy, a number of statutory and regulatory changes during the past decade have lessened the subsidy. Indeed, many changes were designed specifically to reduce the safety net-related advantages that had been accruing to insured depository institutions.

Capital Regulation. Bank capital serves as a cushion to absorb unanticipated losses and shrinkages in asset values that could otherwise cause a bank to fail. Capital levels can be likened to a deductible for federal deposit insurance. As such, the higher the level of capital, the lower the "market" rate for deposit insurance.

Because of concern about declining bank capital levels, the major industrialized nations adopted uniform standards for capital adequacy in 1988. The Basle Accord established total capital to risk-weighted assets as the international capital measure and set eight percent as the minimum acceptable level of risk-based capital. The stated purposes of the Basle Accord were twofold: (1) to promote the stability of global financial markets and consistency in supervisory standards; and (2) to link capital requirements to the riskiness of a bank's activities, including off-balance-sheet risk exposure. Adoption of minimum capital standards and capital requirements tied to the risk profiles of banks has resulted in banks holding more capital -- indeed, U.S. banks as a whole currently have the highest capital levels since 1941 -- and has moved industry capital levels closer to the levels that might be imposed by the market in the absence of the federal safety net. As such, capital standards have reduced significantly the subsidy from the safety net.

The Least-Cost Test. In 1991, the Congress passed the Federal Deposit Insurance Corporation Improvement Act (FDICIA), which, among other things, instituted the "least-cost test." With rare exceptions, the FDIC may meet its insurance obligations by means other than a payout only if the other transaction is "least costly" to the deposit insurance funds of all methods. Prior to this requirement, the FDIC could choose any alternative that was cheaper than the estimated cost of liquidation. Most institutions with over $100 million in assets were resolved through a purchase-and-assumption transaction in which all liabilities except subordinated debt were assumed by an acquirer. The least-cost test caused the FDIC to change the way it structured resolutions, so that uninsured depositors or other general creditors often suffer losses in a resolution.

In the five years leading to the enactment of FDICIA (1987 to 1991), uninsured depositors and other general creditors suffered losses in only 17 percent of the 927 bank failures. In the five years since FDICIA, during which 187 banks failed, uninsured depositors and other general creditors suffered losses in 63 percent of the cases. The market discipline brought about by the greater risk borne by uninsured depositors and other general creditors after the imposition of the least-cost test serves to reduce the subsidy from the safety net.

Risk-Based Deposit Insurance. FDICIA also instructed the FDIC to develop and implement a system of risk-based deposit insurance premiums. Since the market rate for insurance is related directly to the amount of risk an institution takes, flat-rate insurance provided the greatest subsidy to the riskiest institutions. The rationale underlying risk-based premiums is to make the price of insurance a function of an institution's portfolio risk, thus reducing the subsidy to risk-taking and spreading the cost of insurance more fairly across depository institutions. Though the magnitude of the reduction is not quantifiable, the adoption of risk-based insurance premiums has reduced the size of any subsidy that had been accruing to depository institutions because of mispriced deposit insurance. The FDIC currently is undertaking a comprehensive review of the risk-based premium system in order to identify additional information, alternative risk-measurement methods, and other measures that may enhance our capability to charge appropriately for risks posed to the insurance funds. To the extent that this effort is successful, the value of any subsidy conferred through federal deposit insurance will be reduced further.

Changes to Discount Window Policy. In order to enhance market and regulatory discipline in the banking sector and to protect the deposit insurance funds, FDICIA restricted the ability of the Federal Reserve to lend to undercapitalized institutions through the discount window. In particular, FDICIA placed restraints on Federal Reserve lending to institutions that fall below minimum capital standards by setting time periods beyond which the Federal Reserve may not lend to undercapitalized institutions without incurring a potential limited liability to the FDIC. The liability is incurred if an undercapitalized institution borrows for more than 60 days in any 120-day period. Because undercapitalized institutions would have the most difficulty obtaining credit at attractive rates elsewhere, and thus benefit most from access to the discount window, the imposition of restrictions on their access to the discount window reduces the gross subsidy that flows from such access.

Changes to Payments System Policies. Changes also have taken place over recent years reducing the subsidy from the payments system. First, in 1988, the Federal Reserve instituted a system of net debit caps (credit limits) on an institution's daily Fedwire overdrafts. Then, in April 1994, the Federal Reserve started charging fees for daylight overdrafts incurred in accounts at Federal Reserve District Banks. Since April 1995, the fee has been set at an annual rate of 15 basis points of chargeable daily daylight overdrafts. A chargeable overdraft is an institution's average per-minute daylight overdraft for a given day, less a deductible amount equal to 10 percent of its risk-based capital. From April 1995 through December 1995, overdraft charges averaged $27 million at an annual rate. During that same period, about 120 institutions incurred fees regularly, with the largest banks (those with assets of more than $10 billion) accounting for, on average, 92 percent of total charges. Following the implementation of daylight overdraft fees and debt limits, the Federal Reserve observed a dramatic decline in total daylight overdrafts -- averaging 40 percent in the six months following the initial imposition of fees in April 1994. This reduction in daylight overdrafts has reduced the Federal Reserve's intra-day credit risk and liability as guarantor of all Fedwire transactions and thus has reduced the subsidy that previously accrued from the government-operated payments system.

In addition, changes in technology are rapidly transforming the payments system; real-time settlement, as well as alternative means for settling payments, are likely to erode further the subsidy that banks receive from the payments system.

In summary, while banks still receive a gross subsidy from the safety net, this subsidy is significantly smaller than it was a decade ago.

Evidence of a Gross Subsidy Is Limited

Three facts are cited often as evidence that banks receive a gross subsidy from the federal safety net. First, bank holding company debt generally has a lower credit rating than comparable debt of its lead bank. Second, bank capital levels fell after the creation of the Federal Reserve in 1913 and again after the creation of the FDIC in 1933. Third, banks hold less capital than other financial institutions. Each of these arguments is discussed below. It is important to remember, however, that this evidence is really something of a red herring. Whether a particular fact reflects a gross subsidy to the banking industry does not advance the analysis because everyone accepts that a gross subsidy exists. What is important is whether after taking account of offsetting costs, banks have a net subsidy that they can pass on to operating subsidiaries or holding company affiliates. These facts relative to the gross subsidy shed no light on the net subsidy question.

Credit Ratings of Banks and Bank Holding Companies. The debt of a bank holding company generally has a lower credit rating than comparable debt of the bank holding company's lead bank. This rating differential -- which translates into lower funding costs at the bank level than at the bank holding company level -- is considered evidence that the safety net provides a funding subsidy to the bank that is not transferred to its parent holding company. Because it allegedly shows that the holding company structure is more effective in limiting the advantages of the safety net than a bank subsidiary, this observation is considered a particularly important piece of evidence by those who favor the bank holding company structure over the bank subsidiary structure for financial modernization.

However, there is an alternative explanation of the credit-rating differential between bank and bank holding company debt. Bank holding companies get the vast majority of their income from their bank subsidiaries. The primary asset of a bank holding company is usually the stock of its lead bank. In 1995, approximately 93 percent of the dividends received by the 50 largest bank holding companies came from banks, the remaining seven percent came from nonbank subsidiaries. Thus, debt holders of a bank holding company rely heavily on bank dividends to service their debt. In a bank failure, bank holding company debt holders will usually only get paid if there is residual value in the receivership to pay to the holding company -- the bank's equity holder. The debt-service discrepancy is exacerbated with respect to bank holding companies, because regulators can restrict dividends from the bank to the holding company while at the same time the bank holding company is expected by regulators to be a source of strength to the bank. Much as senior debt usually has a higher credit rating than junior debt, the inferior position of bank holding company debt with respect to both debt service and claims in a bank failure could well explain the difference in credit ratings between a bank and its bank holding company.

FDIC staff has analyzed extensively the causes of the ratings differential between bank and bank holding company debt. In particular, the staff has examined three areas: the credit ratings of nonbank holding companies and their subsidiaries, the existing literature, and the debt-rating criteria used by Moody's and Standard & Poor's. We found substantial evidence that the ratings differential is due to the inferior position of holding company debt as compared with bank debt with respect to both debt service and bankruptcy. We found no evidence that the safety net plays a significant role in the ratings differential.

Nonbank holding companies - If the ratings differential between banks and bank holding companies were due to the safety net, one would not expect this differential to exist for nonbank holding companies. On the other hand, if the differential were due to a different priority structure of bank and bank holding company debt, then it should exist for nonbanks as well. However, as a general rule, nonbanking companies do not have holding company structures in which the parent is dependant on a lead subsidiary for most of its income. For example, while Ford Motor Company may have numerous subsidiaries, the first tier parent has productive capacity -- it makes cars. Thus, a direct comparison of bank and nonbank holding companies is not possible. However, the general absence of a holding company structure outside the banking industry would tend to indicate that this structure is not viewed as advantageous by the market. As such, it may be evidence that forcing banks to operate through a holding company structure adds costs and therefore results in some disadvantage for banks relative to their nonbank competitors.

Literature - FDIC staff did not find any literature that directly addresses the ratings differential between bank and bank holding company debt. However, FDIC staff did find two articles that indirectly address the issue. These articles investigate the required rate of return on bank holding company debt when the proceeds are downstreamed to the bank through equity investments, a practice known as double leverage, and when they are downstreamed to the bank as debt. Both articles concluded that the greater the level of double leverage, the greater the required rate of return. In a bank failure, funds downstreamed to a bank as debt have a higher priority than funds downstreamed as equity. Thus, these articles support the hypothesis that the credit priority of bank holding company debt holders affects the cost of bank holding company debt.

Bank-Rating Criteria of Standard & Poor's and Moody's - FDIC staff spoke to analysts at Standard & Poor's and Moody's. Analysts at both rating agencies cited the priority structure in debt servicing (the fact that regulators can restrict dividends to the holding company) and in bankruptcy as the primary reason why bank debt typically has a higher credit rating than bank holding company debt. Put simply, the fact that the bank debt comes before bank holding company debt means that it has a lower default risk and hence a higher credit rating. Both also said that they monitor the amount of double leverage in an individual bank holding company and, that all other things being equal, the greater the amount of double leverage, the greater the default risk and the ratings differential.

Standard & Poor's noted that another reason for the ratings differential is that it views the cross-guarantee provisions of FIRREA and the Federal Reserve's "source of strength doctrine" to mean that a holding company would be expected to downstream most or all of its financial resources to protect a bank that was in risk of failing, thus raising the possibility of default at the holding company level. Moody's noted that access to the payments system does have a positive effect on banks, but that the effect is "negligible" in accounting for the ratings differential between banks and bank holding companies. (See Attachment 1, bank holding company rating criteria of Standard & Poor's and Moody's.) Thus, based on information by the two major rating agencies, there is little reason to believe that the safety net plays a substantive role in the ratings differential between banks and bank holding companies.

Bank Capital Has Decreased Over Time - A second piece of evidence regarding the existence of a gross subsidy from the safety net is that capital levels at banks fell after the creation of the Federal Reserve System in 1913, and again after creation of the FDIC in 1933. These data were presented in a 1991 Treasury study on financial modernization, and updated in a 1995 paper by Berger, Herring, and Szego. As can be seen from Attachment 2, reproduced from Berger et al., bank capital decreased fairly steadily from at least 1840 through 1940. While the pace of the decrease (the slope of the curve) increased slightly after creation of the Federal Reserve and again after the creation of the FDIC, capital also had decreased as sharply before. Moreover, as shown in Attachment 3, drawn from a 1992 paper by Kaufman, the capital of virtually every industry, has decreased since the beginning of the twentieth century. The most likely explanation of this overall decrease in capital levels is a general increase in the efficiency of the U.S. financial system. Commenting on the decrease in capital in the banking industry the 1991 Treasury study states:

Capital ratios were declining long before creation of either the Federal Reserve System or the FDIC. Indeed, much of the decline both before and after the creation of the safety net no doubt reflects the growing efficiency of the U.S. financial system. Nevertheless, the federal safety net is most likely a key factor in explaining why bank capital ratios can remain near their current levels without weakening public confidence in the banking system. It is difficult to believe that many banks and thrifts operating over recent decades could have expanded their assets so much, with so little additional investment by their owners, were it not for the depositors' perception that, despite the relatively small capital buffer, their risks were minimal. (p. II-4)

This appears to be a fair interpretation of the evidence. Banks' ability to expand in the 1980s and early 1990s with very thin capital margins and the ability of thrifts to expand in the 1980s while insolvent is almost definitely a result of the federal safety net. However, given the downward spiral in bank capital that ran for 100 years from 1840 to 1940 and the overall decrease in capital levels in all industries over this period, it is difficult to conclude that the decrease in bank capital after the creation of the Federal Reserve and the FDIC was due to the federal safety net. Moreover, the significant increase in bank and thrift capital since the late 1980s reduces the gross subsidy from the safety net, as discussed above.

Banks Hold Less Capital Than Other Financial Services Companies. A third argument offered to support the contention that banks receive a gross subsidy from the safety net is the observation that banks hold less capital than other financial intermediaries. Attachment 4, drawn from the earlier-mentioned 1992 paper by George Kaufman, shows that in 1989 large commercial banks and bank holding companies did indeed have lower capital ratios than other financial institutions.

These data, however, need to be interpreted cautiously. Capital levels are tied to the volatility of earnings. Thus, the large difference in capital levels between banking organizations and some nonbank financial firms is most likely due to the much higher risk in their businesses, not the safety net. Nonetheless, the fact that capital levels in industries like life insurance, that are not very risky, are higher than those at banks can reasonably be interpreted as evidence of a gross subsidy, but as we have said, the real issue is the net subsidy.

Measurement of the Gross Subsidy

The earlier discussion showed how banks receive a gross subsidy from the federal safety net, but that the subsidy has been significantly reduced in recent years through statutory and regulatory changes. Measuring the magnitude of the remaining subsidy, however, is quite difficult and few estimates exist.

If one believes the argument discussed earlier that the difference in bond ratings between banks and their bank holding companies is due to the safety net, although evidence is to the contrary, then the difference in bond yields that results from the ratings difference offers a measure of the subsidy from the safety net. According to the data collected by the Federal Reserve, in 1990 this difference was 10 to 15 basis points, but since 1994 this difference has been in the four to seven basis point range. In interpreting these data, it is important to remember two things. First, the ratings differential only captures the difference between the bank and its holding company. If the holding company's debt rating is enhanced by the safety net, there is an additional portion of the gross subsidy not captured by the ratings differential. Second, as discussed earlier, there are other very good reasons aside from the safety net for the ratings differential between the bank and its holding company. Indeed, it is not clear that the safety net plays a significant role in the four to seven basis point discrepancy between the cost of bank and bank holding company debt. In any event, under no circumstances can this entire discrepancy be attributed to the safety net. This is an extremely important point. It means that when comparing the funding cost of banks and bank holding companies the difference is less than four to seven basis points, even before taking account of offsetting costs.

Another method of measuring the gross subsidy from the safety net -- or at least the deposit insurance portion of the safety net -- is to attempt to estimate the market rate for deposit insurance. Unfortunately, it is very difficult to measure what that rate should be. The bulk of the studies that attempt to estimate a market rate for deposit insurance have used an option pricing model applied to data from the 1980s. The application of this model to deposit insurance is based on the observation that if a bank is found to be insolvent, depositors can, in effect, "sell" their share of a bank's liabilities to the FDIC in exchange for cash. The appropriate price for insurance, according to this approach, is the value of this option to sell. If insurance premiums are set lower than the option price, then the bank can be said to receive a subsidy.

Unfortunately, there are numerous methodological problems with applying option pricing theory to value deposit insurance. Most notably, option theory deals with finite time contracts that expire in a year or some other finite period of time, while the deposit insurance guarantee is theoretically open-ended. The value of insurance as calculated by these models also depends critically on the timing of bank examinations, where more frequent examinations lower the risks to the insurance funds and, therefore, the value of insurance, and on the actual recovery on the assets of the failed bank. Hence, the estimated fair value of deposit insurance, as computed by these models, varies depending on the model's assumptions. With these caveats in mind, most option based models estimated the average fair value of deposit insurance premiums to be under 10 basis points. It should be noted, however, that the premium rates estimated in most of the studies employing option pricing models, on average, would have been grossly insufficient to cover FDIC losses during the 1980s. Thus, the "market" premium as estimated by option pricing models is probably understated.

The fair value of deposit insurance, however, is not a measure of the subsidy from deposit insurance because banks pay premiums for the insurance. This holds even today, with both funds fully capitalized and most banks paying no explicit premiums, since banks can be considered to have "prepaid" their deposit insurance premiums. Beginning in 1991, banks and thrifts paid higher premiums to recapitalize the insurance funds, with the premium assessment ranging as high as 31 basis points for some institutions in the period 1993 through 1995. In addition, in the third quarter of 1996, SAIF members paid a special assessment of almost 65 basis points to capitalize the SAIF.

Moreover, whenever the FDIC Board of Directors determines that circumstances exist that raise a significant risk of substantial future losses to the insurance funds, the Board can raise the reserve ratio and premium rates to reach the higher ratio well in advance of a severe crisis -- thereby increasing the likelihood that the call on the U.S. Government will not have to be exercised in the future.

Offsets to the Gross Subsidy

The various evidence of the magnitude of the gross subsidy that does exist points to a gross funding subsidy of around 10 basis points. In addition, banks face other costs such as reserve requirements, interest payments on bonds issued by the Financing Corporation (FICO) and regulatory burden expenses that, at a minimum, partially offset the subsidy. These are discussed in turn below.

Reserve Requirements. Banks must hold required reserves against transaction accounts, nonpersonal time deposits, and Eurocurrency liabilities. These noninterest-bearing reserves must be held as vault cash or as a deposit at a Federal Reserve District Bank. Reserve requirements are intended largely as a tool of monetary policy. By contrast, the foregone income on required reserves can be viewed as a tax or fee paid by the banking industry. For the year 1996, FDIC staff has computed a conservative estimate of the pretax cost to banks for holding required reserves to be approximately $840 million. If one spreads this amount over the approximately $3.8 trillion in average deposits held by FDIC-insured institutions over this time period, the pretax cost of required reserves is approximately 2.2 basis points.

FICO Assessment. In 1987, Congress created the FICO to sell bonds to raise funds to help resolve the thrift crisis. The interest payment on FICO bonds is $793 million annually, and the last of the FICO bonds matures in 2019. Beginning in 1997, the annual interest is being paid by all FDIC-insured institutions, not just SAIF-member savings associations. Commercial banks were asked to share the burden of these payments since they also share in the benefits of deposit insurance, and thus their payment of this fee is a direct result of banks' access to the safety net. For the first quarter of 1997, the FICO assessments were 6.4 basis points (annualized) for SAIF members and 1.3 basis points for BIF members. Beginning in 2000, or 1999 if the funds have been merged, all institutions will pay a pro rata share for FICO, which presently is estimated to be 2.4 basis points.

Regulatory Burden. Perhaps the greatest offset to the gross subsidy that banks receive from the safety net is regulatory costs. Unfortunately, good estimates of the cost of regulatory burden do not exist. A 1992 Federal Financial Institutions Examination Council (FFIEC) study reviewed the literature on regulatory burden. The studies reviewed by the FFIEC differed in methodology and scope. Nonetheless, the FFIEC found that "[d]espite differences in methodology and coverage, findings are reasonably consistent: regulatory cost may be 6 - 14 percent of non-interest expenses, not including any measurement of the opportunity cost of required reserves." Some of the studies reviewed by the FFIEC included the cost of deposit insurance (at the time 8.3 basis points), and some aspects of regulatory burden, such as "truth in lending," are not unique to banks. However, even the low end of the FFIEC range -- six percent -- still yields substantial regulatory costs. In particular, if we take six percent of the approximately $150 billion of noninterest expenses incurred by commercial banks during 1995, it yields a cost of regulatory burden of $9 billion dollars. Expressed in terms of average deposits at commercial banks during 1995 of almost $3 trillion, this amounts to 30 basis points.

Thus, the total offset, including reserve requirements, FICO interest payments, and regulatory burden, is estimated to be more than 33 basis points for banks and more than 38 basis points for thrifts. After the FICO payment is equalized, this offset will be approximately 34 basis points for banks and thrifts.

The Net Subsidy

Measuring whether banks actually receive a net subsidy is difficult because it depends on reliable estimates of the gross subsidy and offsetting costs, which as we have stated, are difficult to determine. Nonetheless, with most estimates of the gross subsidy around 10 basis points, the costs would appear to outweigh the subsidy significantly -- by more than threefold. Clearly, with a difference of this magnitude, those who believe a net subsidy exists must bear the burden of proof.

Given the difficulty of obtaining reliable estimates of a net subsidy, it is helpful to look at other indicators to assist us in determining whether banks receive a net funding subsidy from the safety net that affects the business judgments they make. In this regard, it is useful to look at how a banking organization would best organize itself to exploit a net subsidy if one were to exist. Federal Reserve Board Chairman Alan Greenspan addressed this issue in recent testimony before the House Subcommittee on Financial Institutions. Chairman Greenspan stated that "one would expect that a rational banking organization would, as much as possible, shift its nonbank activity from the bank holding company structure to the bank subsidiary structure. Such a shift from affiliates to bank subsidiaries would increase the subsidy and the competitive advantage of the entire banking organization relative to its nonbank competitors."

Yet, banks conduct a wide range of activities through holding company affiliates that could be conducted directly through a bank or in bank subsidiaries without any firewalls. As of September 30, 1996, the 50 largest bank holding companies had 155 mortgage banking affiliates, 98 commercial finance affiliates and 263 consumer finance affiliates. At the same time, the bank subsidiaries of these holding companies had 104 mortgage banking subsidiaries, 24 commercial finance subsidiaries, and 89 consumer finance subsidiaries. (See Attachment 5.) In addition, the banks conduct mortgage, commercial, and consumer finance activities directly, and could conduct these activities nationwide directly through the bank.

Following Chairman Greenspan's logic that a rational banking organization would choose its organizational structure so as to maximize its competitive advantage from the subsidy, there appear to be only three possibilities -- (1) the subsidy is the same whether an activity is conducted in a holding company affiliate or the bank proper, (2) there is no net subsidy, or (3) the net subsidy is so small that it is outweighed by other considerations. Because Sections 23A and 23B make it highly unlikely that if a net subsidy exists, it is the same whether the activity is conducted in a holding company affiliate or the bank, the fact that banking organizations choose all three forms of organization -- holding company affiliate, bank subsidiary, and bank proper -- suggests that there is not a net funding subsidy, or that if a net subsidy exists, it is so small so as to be outweighed by other considerations.


While the evidence shows that if a net subsidy exists at all it is small, reasonable firewalls designed to protect the insured bank serve to inhibit the passing of any subsidy from a bank to its subsidiary. In particular, as discussed previously, we believe that in order for a bank subsidiary to engage in an activity not permissible to the insured bank, the bank should be "well-capitalized," the bank's equity investment in the subsidiary should be deducted from regulatory capital (and assets), the subsidiary should not be consolidated with the bank for regulatory capital purposes, and that "covered transactions" between the bank and its subsidiary should be subject to the restriction of Sections 23A and 23B of a Federal Reserve Act.

These firewalls are not impenetrable. If a bank has excess regulatory capital -- capital above the well-capitalized level -- it may sometimes have an incentive to pass on a portion of the net subsidy to a bank subsidiary or a bank holding company affiliate. In particular, if a bank had excess regulatory capital that it did not want to leverage because it felt that the market demanded that it hold additional capital, it could lower its overall cost of funding by borrowing additional "subsidized" funds at the bank, investing these funds as "equity" in the subsidiary, and using the new funds to pay off market-rate liabilities in the subsidiary. The banks regulatory capital level would fall because the new equity investment in the subsidiary would be deducted from regulatory capital, but where the bank has excess regulatory capital this would not be a problem. The bank's consolidated capital level would remain unchanged because the increase in liabilities in the bank would be exactly offset by the decrease in liabilities at the subsidiary. Finally, the subsidiary's and bank's overall cost of funds would fall because subsidized funds would have replaced unsubsidized funds.

Of course, if it were the consolidated holding company capital that the banking organization wanted to keep high -- and it is, after all, the holding company that is the publicly traded entity -- then the bank similarly could transfer part of a subsidy to a holding company affiliate. The bank could borrow additional subsidized funds at the bank and upstream additional dividends to the holding company. The holding company could then downstream these funds to an affiliate that could use the funds to pay off market-rate liabilities. The consolidated capital ratio of the holding company would remain the same, because the additional funds borrowed at the bank are offset by the liabilities paid off at the affiliate. The cost of funds of the affiliate and the overall cost of funds of the holding company would decline because subsidized funds borrowed at the bank would have replaced market-rate liabilities borrowed at the affiliate. Unlike the bank subsidiary model, however, such a transfer could take place even if the bank were not well-capitalized, since there is no requirement that a bank be well-capitalized to pay dividends to its parent holding company.

There are three important points to remember about these examples. First, under the bank subsidiary structure, Section 23A would restrict the total investment (equity and debt) in a single subsidiary to 10 percent of the bank's equity capital. Thus, aside from the fact that only excess capital could be invested as equity in a subsidiary, there would be another limit to the bank's investment in its subsidiary.

Second, the mere existence of excess regulatory capital is not sufficient for there to be an incentive for a bank to pass a portion of a net funding subsidy to either a subsidiary or an affiliate. If the bank is willing to leverage its excess regulatory capital -- perhaps because the excess regulatory capital is a result of record profits rather than a business strategy to hold additional capital, then it is more profitable for the bank to leverage the additional capital either through reducing capital levels or by growth, than to take advantage of the excess capital to pass a portion of the subsidy to a subsidiary or affiliate. The reason for this -- if the bank is willing to increase its leverage by decreasing capital levels -- is that the bank can lower its overall cost of funds by borrowing additional subsidized funds and paying out an additional dividend. In this case, the subsidized funds replace equity capital. Because of the greater risk borne by equity holders than debt holders and the tax advantages of debt, the bank's cost of equity is going to be higher than the cost of unsubsidized borrowing at a bank subsidiary or holding company affiliate. As such, paying an additional dividend is a superior strategy to using the excess capital to transfer a portion of the subsidy to a subsidiary or an affiliate. Although the explanation is more complicated, the same result holds true if the bank chooses to leverage its excess capital through growth.

Capital levels at banks, especially large banks, currently are near historic levels. What is not clear is why the banks are holding this capital. Certainly, it is possible that some of this capital is being held because the market is demanding it. However, it also is possible that some of this capital is a product of the record profits that banks have experienced in recent years that simply have not been leveraged yet. Banking organizations, as well as other firms, are generally reticent to increase dividends unless they are confident higher earnings will be maintained since they do not want to have to lower dividends when earnings fall. There is evidence that banks are increasing payouts to their parents above current earnings to fuel stock buy-backs. Of the 50 largest bank holding companies, more than 30 have announced stock buy-back programs since the beginning of 1996. (See Attachment 6.) It also is possible that banks are building up excess regulatory capital to fuel future expansion especially with interstate branching looming on the horizon. Given the fact that large banks historically have held as little capital as the regulators would allow, it is too early to tell if their current capital levels reflect new business realities or merely are a temporary phenomenon that is a result of record earnings. Therefore, it is difficult to say if large banks have even a theoretical incentive to use any of their excess capital to subsidize a subsidiary or holding company affiliate.

Third, unless all of the liabilities of a bank subsidiary or holding company affiliate were replaced with equity investments funded by subsidized bank borrowings, only a portion of the subsidy would be passed to the subsidiary or affiliate. Since we are talking about a net subsidy that may not exist at all, or at most is very small, a portion of the net subsidy would be a de minimus amount.

The real question then is does the theoretical possibility of passing on a subsidy make any real world difference? Chairman Greenspan addressed this issue with respect to bank holding company affiliates in his testimony before the Subcommittee on Financial Institutions and Consumer Credit last month. He recognized that banks could theoretically pass a subsidy to affiliates through dividends paid to a holding company, but argued that as an empirical matter such a transfer has not taken place.

Given the firewalls I have outlined above, the underlying incentives to pass a subsidy to a bank subsidiary are quite similar to the incentives to pass a subsidy to a holding company affiliate. If we do not observe a subsidy being passed to holding company affiliates, we would, in most cases, not expect it to be passed to a bank subsidiary.

The available evidence appears to bear this out. As I discussed earlier, the FDIC has allowed bona fide securities subsidiaries of state nonmember banks for just over a decade. These subsidiaries are subject to restrictions to protect the insured bank similar to the firewalls I have outlined above. If there were a substantial net subsidy that could be transferred to a bona fide subsidiary, one would expect that at least some large bank holding companies would conduct their securities activities through bona fide subsidiaries rather than Section 20 subsidiaries of the holding company. This is especially true since 1991, when the U.S. Court of Appeals for the Second Circuit ruled that the Federal Reserve Board does not have jurisdiction under the Bank Holding Company Act over bank subsidiaries of a bank in a bank holding company (Citicorp v. Board of Governors of Federal Reserve System, 936 F.2d 66). The fact that all large bank holding companies continue to conduct their underwriting activities through bank holding company subsidiaries seems to indicate that, if there is a net subsidy, the bona fide subsidiary structure is as effective in preventing it from being transferred out of the bank as the bank holding company structure.

Of course, in times of stress firewalls tend to weaken, and transgressions have occurred both within and outside the reach of the regulators. But our experience with the financial stress of the 1980s and early 1990s indicates that in times of financial stress, pressure can be exerted on a bank from its holding company as well as from subsidiaries. Organizational structure is unlikely to affect this potential problem. The past decade provided a number of instances where "death-bed transactions" were proposed or consummated that served to advantage the holding company or an affiliate at the expense of the insured bank. The transactions often involved sums in the tens of millions of dollars. Not all of these transactions required regulatory approval. The regulators often, but not always, denied those that did. In some instances, regulators concluded that denial might unnecessarily aggravate the plight of banking organizations that might otherwise survive.

Unpaid tax refunds arose as an issue in more than one case. Bank holding companies generally received tax payments from and downstreamed tax refunds to their banking subsidiaries, acting as agent between the bank and the Internal Revenue Service. The FDIC has observed that in some cases unpaid tax refunds tended to accumulate on the books of failing bank subsidiaries, leaving the cash with the holding company. This practice took place without regulatory approval.

Consolidation of nonbank activities at the parent level is another way to transfer value away from insured bank subsidiaries. When service company affiliates carry out data processing or other activities for banks, the issue of intercompany pricing also is raised.

Linked deals involving the sale of purchased mortgage servicing rights have in some cases been used either to subsidize the sale of a holding company asset or to allow the bank subsidiary to book an accounting gain. The effect of a linked deal either may be to transfer value to the parent or delay the closing of a subsidiary without the benefit of needed fresh capital.

Finally, there have been instances of "poison pills" created by interaffiliate transactions. In one case, key bank staff were transferred to the holding company payroll, apparently to reduce the attractiveness of bringing in an outside acquirer. Interaffiliate data processing contracts also have been structured so as to limit the availability of information to the FDIC or an acquirer after the bank was closed, thereby making regulatory intervention more costly.

Contraction of the Safety Net

In examining the question of financial modernization and the possible expansion of the federal safety net, it is important to recognize that modernization may help contract the safety net. In particular, as I have previously discussed, with appropriate safeguards, having the earnings from new activities in bank subsidiaries provides greater protection to the insurance funds than the holding company structure.

Since the fair market price for deposit insurance is tied to expected insurance losses, by reducing the expected losses of the insurance funds, allowing banks to put new activities in a bank subsidiary also lowers the fair market price for deposit insurance. In addition, it lowers the value of the banking industry's access to the full faith and credit of the United States Government. As such, if banks do receive a net subsidy, allowing banks to put new activities in bank subsidiaries would lower any subsidy they receive.

Given the difficulty in measuring the subsidy to begin with, estimating how much this reduction in subsidy would be is virtually impossible. However, it is important to note that since any net subsidy that does exist is clearly small, the reduction in subsidy would not be material enough to warrant favoring the bank subsidiary structure to the exclusion of the bank holding company structure.


Current restrictions on the financial activities of banking organizations are outdated. Their elimination would strengthen banking organizations by helping them to diversify their income sources, and would promote the efficient, competitive evolution of financial markets in the United States. However, we should proceed cautiously in easing the broad range of restrictions on activities of banking organizations beyond those that are financial in nature. Expansion of bank and thrift powers must be accompanied by appropriate safeguards for the insurance funds.

Any financial modernization proposal should permit financial organizations to engage in any type of financial activity, unless the activity poses significant safety and soundness concerns or is potentially harmful to consumers or small businesses. Moreover, financial institutions should have flexibility to choose the corporate or organizational structure that best suits its needs, provided adequate safeguards exist to protect the insurance funds and the taxpayer.

The two organizational structures with which we have experience in the United States -- the holding company model and the bank subsidiary model -- can provide adequate safety and soundness and can inhibit the undue expansion of the federal safety net, provided adequate safeguards are in place to protect insured institutions and the deposit insurance funds.

This conclusion is based on the fact that if banks receive a net subsidy at all, it is small. Moreover, reasonable firewalls will inhibit the passage of any net subsidy that may exist to a bank subsidiary, and the bank subsidiary structure is likely to be as effective as the bank holding company structure in preventing a subsidy from being passed out of the bank. In both cases, any leakage of a net subsidy out of the insured bank would be de minimus. In addition, the bank subsidiary structure of financial modernization would be more effective than the bank holding company structure in reducing any subsidy that does exist because the placement of new activities in bank subsidiaries provides greater protection for the insurance funds, lowers the fair-market price for deposit insurance and lowers the value of the industry's access to the full faith and credit of the United States Government. For these reasons, allowing banks to conduct new financial activities in a bank subsidiary would not lead to an undue expansion of the federal safety net. Banking organizations should be able to choose the organizational structure -- bank subsidiary or bank holding company -- they feel is best for pursuing their individual business strategies.

Any financial modernization proposal should be consistent with the safeguards of Sections 23A and 23B of the Federal Reserve Act and apply them to dealings between an insured bank and any subsidiary of the bank. In addition, we believe that any financial modernization proposal should require that the capital adequacy of an insured institution be determined after deducting the institution's investment in subsidiaries. The experience of the FDIC has been that in times of financial stress, banking organizations may attempt to engage in transactions that transfer resources from the insured entity to the owners and creditors of the parent company, nonbanking affiliates, or to subsidiaries of the bank.

The FDIC believes that a greater degree of regulation along functional lines may be preferable to the current practice of regulating individual banking entities based on charter or corporate structure. We must ensure, however, that functional regulation is seamless and does not result in duplicate regulation or in the artificial restructuring of banking operations and services. We also must ensure that key transactions between insured banks and their affiliates and subsidiaries can be reviewed by regulators as part of the regular examination process for insured banks.

Furthermore, to address concerns regarding the stability and liquidity of the financial system, it may be necessary, as part of the effort to modernize, to initiate some oversight of consolidated financial organizations. With increasingly complex financial products and organizational structures, some activities, practices, and intercompany dealings may go unnoticed if there is too heavy a reliance solely on a functional approach. At the same time, there will be a growing need to coordinate supervisory efforts to ensure the ready availability of adequate information with which to gauge risks. However, such supervision need not involve full-scope examinations of nonbanking subsidiaries nor activity-by-activity or investment-by-investment regulation of nonbanking activities.

The FDIC stands ready to assist the Subcommittee in evaluating how best to reform our financial system.

Last Updated 06/25/1999

Skip Footer back to content