October 1999

IS BANK REGULATION NECESSARY?

Banks in the United States are the said to be the beneficiaries of several government programs that confer advantages not provided to other industries. As the Gramm-Leach-Bliley Act moved through Congress, much of the debate centered on the question of whether those benefits, which constitute a government safety net for banking, could be transferred to subsidiaries or affiliates in nonbanking businesses.

The debate about the safety net raised a deeper question: whether the existence of the safety net—which is perceived to create risks for the government—is the basis for the regulation of banks, or whether other policy reasons justify this special treatment. If the government’s safety-net risks are the principal reason for bank regulation, then bank regulation could be eliminated or substantially reduced if the safety net were eliminated or transferred to the private sector.

Scholars agree that the safety net consists of three elements: deposit insurance and the Federal Reserve’s roles as lender of last resort for banks and as guarantor of intraday overdrafts between banks in the nation’s large-dollar payment system. (An intraday overdraft occurs when a bank—in sending payments through the Federal Reserve System to other banks—sends more funds than it receives in return.)

A conference at AEI brought together a distinguished group of scholars and banking policy experts to explore whether those government-sponsored advantages were the sole reason for the strict bank regulation we have today and whether—if the safety net were eliminated or replaced by nongovernmental alternatives—a basis would exist for substantially relaxing bank regulation.  

Gary H. Stern, president of the Federal Reserve Bank of Minneapolis, gave the keynote address. Stern argued that it would be a mistake to limit or reduce bank regulation on the assumption that government will not intervene to avert the possible failure of an institution considered too big to fail. As long as government intervenes in such cases—or the markets believe it will—reducing regulation will not reduce moral hazard and risk-taking by institutions protected by the safety net.

Stern’s address highlighted the central issue associated with eliminating or reducing bank regulation: the complicated interplay among regulation, the safety net, and systemic risk. Stern argued that since government could not be restrained from intervening when it sees systemic risk in the financial markets, regulation is necessary to address the consequent moral hazard. This raised the question whether a safety net is really necessary to prevent financial system instability.

The History of U.S. Bank Regulation

Charles W. Calomiris, professor of finance at Columbia Business School and an AEI visiting scholar, reviewed the history of banking policy and bank regulation in the United States. He noted that both grew out of a mercantilist tradition in which governments chartered banks as monopolies that could serve as revenue sources or perform specific financial purposes that the government needed. Regulation was directed toward assuring the purposes of the chartering, not for the purpose of maintaining bank safety and soundness. When monopoly privileges ended in the antebellum free-banking period, regulation continued, principally to assure that banks would remain instruments of national economic policy. Safety and soundness regulation and the other safety-net elements were generally left to private facilities; bank clearinghouses and private payment systems performed many of the functions now carried out by government.

In the nineteenth century, national banking policy promoted unit banking—that is, single-location banks—and prohibited interstate banking. Those policies, which reduced the ability of banks to diversify their assets, gave rise to serious and repeated instability in the banking system. Instability in turn created a basis for establishing the Federal Reserve as a source of liquidity for banks threatened with unsustainable cash demands. The Fed itself grew out of the clearinghouses in the main commercial cities of the United States, and in many respects carried on the same functions under government auspices. When large numbers of unstable unit banks failed in the Great Depression, a system of deposit insurance was the price extracted by Congressman Steagall for support of the 1933 banking law. That system, however, only reinforced unit banking and the instability it fostered.

Thus, Calomiris suggested, the key elements of the safety net were required to compensate for shortsighted banking policy—government restrictions on such things as local and interstate branching. Accordingly, changes in these policies or in the economy might obviate the need for a safety net. But does the safety net, in any event, create stability?

Regulation and Stability

On that question, Richard S. Carnell of Fordham Law School noted that the way the elements of the safety net are administered seriously impairs its ability to bring long-term stability to the financial sector. Ideally, regulators should let weak institutions fail, but—especially in a political environment—the regulators have personal and policy incentives to suppress instability. That leads them to intervene when they should not, by assisting banks that should be allowed to fail—and to forbear when they should not, by allowing weak institutions to remain open.

The thrift crisis of the 1980s was an example of the latter, Carnell noted, with the regulators avoiding necessary bank closures. That permitted hopelessly weak institutions to stay alive, to finance their activities with insured deposits, and ultimately to take risks that resulted in even greater losses.

Particularly troubling is the fact that expansive notions of systemic risk are common among bank regulators, providing an intellectual basis for intervention that may not be necessary and can only further weaken market discipline. To maintain long-term stability, Carnell argued, bank regulators must be willing to tolerate some short-term instability. But the tendency of regulators to suppress instability wherever it occurs suggests that the existence of the safety net might have an adverse effect on long-term stability in the financial system.

Consultant Carter H. Golembe, discussing the Calomiris and Carnell papers, agreed that the structure of the U.S. banking system—which until recently was characterized by large numbers of unit banks—was bound to create instability. That made a safety net seem necessary, but the regulators’ “obsessive fear of instability”—most recently demonstrated in the Long-Term Capital Management (LTCM) failure—suggests that a safety net cannot be administered without creating moral hazard.

With that background, the conference turned to the question of whether bank regulation has a basis other than the need to minimize the risks the government has assumed through the creation of the safety net itself.

Policy Grounds for Bank Regulation

In his presentation, George Benston of Emory University reviewed those aspects of his recent monograph Regulating Financial Markets (AEI Press, 1999), which focused on bank regulation. He noted that historically the regulation of banks has been justified on a number of grounds: (1) the role of banks as an effective means of taxation; (2) the control of banks’ centralized power; (3) the adverse effect of bank failure on the economy as a whole; (4) government’s desire to control the money supply; (5) an effort to suppress competition from banks in other sectors of the economy; and (6) consumer protection. Because of changes in the economy and technology, he argued, none of these rationales for bank regulation is now persuasive.

In Benston’s view, there is little reason to be concerned about the failure of an individual bank. If systemic problems arise or are feared, they can be addressed by the Federal Reserve through open market operations. Accordingly, assuming that the large-dollar payment system can be privatized, protection of the government’s risk in the deposit insurance system is the only sound basis for government regulation. That can be effectively addressed with a regular financial reporting system and higher capital requirements.

Myron Kwast, associate director of research and statistics for the Federal Reserve Board, commented on Benston’s paper. Kwast argued that a high capital requirement and increased financial reporting, while reducing to some degree the ability of banks to use the put option inherent in federal deposit insurance, could not effectively substitute for supervision and regulation. Other initiatives, such as a mandatory subordinated debt requirement, were promising but unproven, and in any event would depend on bank supervisors for enforcement.

Benston’s view that systemic problems associated with individual bank failures could be addressed solely through the Fed’s open market activities raised an issue that was not addressed in Gary Stern’s keynote or in the presentations of Charles Calomiris and Richard Carnell—whether the government’s efforts to stem systemic risk could be administered in a way that did not foster moral hazard or weaken market discipline.

The Federal Reserve’s Function as Lender of Last Resort

That question was addressed directly by George Kaufman, a professor of finance at Loyola University of Chicago, who argued that the Federal Reserve could avert systemic risk more effectively through open market operations than through lending to individual banks at the discount window.

Kaufman noted that the motive behind the Fed’s lender-of-last-resort operations, at least for large banks, is to prevent systemic disruptions caused by the inability of large institutions to obtain liquidity in their normal markets. However, when the lender-of-last-resort system is administered through the discount window, on a bank-by-bank basis, it has the potential to cause substantial losses to the Federal Deposit Insurance Corporation.

Because the Fed receives collateral for a discount-window loan, it takes no risk of loss. If the bank that applies for a discount-window loan is insolvent, however, and subsequently fails, the Fed’s loan in effect causes a loss to the FDIC, since it strips the bank of some of the good assets that could have been sold when the bank was finally closed. If the Fed were to carry out that aspect of its functions solely through open market operations, it would reduce the likelihood of this result and would accordingly reduce the need for safety-and-soundness regulation by the FDIC.

In open market operations, the Fed would sharply increase its purchases of U.S. government securities, thus flooding the markets with additional money, making banks more liquid and hence credit more available. The Fed has occasionally followed that policy, Kaufman noted, when significant financial shocks have caused funding problems for market participants—specifically, in connection with the 1987 and 1989 stock market declines and the Russian default–LTCM debacle in the fall of 1998. Those interventions were successful in restoring market stability and should be contrasted with assistance to banks that were later found to be insolvent—as occurred in the case of the Franklin National Bank in 1974, the Continental Illinois Bank in 1984, and the Texas banks in 1987–1989.

Kaufman also pointed out that open market operations permit the Fed to address systemic risk—and hence the too-big-to-fail problem—without creating moral hazard. In effect, open market assistance ameliorates the danger of systemic collapse but does not assist the weak or failing institution.

Commenting on Kaufman’s paper, Alton Gilbert, a vice president at the Federal Reserve Bank of St. Louis, challenged the view that the Fed should carry on its lender-of-last-resort functions—at least insofar as they involve systemic risk considerations—by acting solely through open market transactions. He argued that this could not be the basis for reducing the Fed’s regulatory and supervisory role because the Fed requires the information it receives in that capacity to carry on its payment system functions. There, he noted, the Fed assumes substantial risks with respect to banks’ intraday overdrafts, and it could not responsibly do so without the information on the financial condition of banks that it receives as a regulator.

Alton Gilbert’s concern about the Fed’s liabilities for the payment system highlights the underlying validity of Kaufman’s point—that the Fed could avoid risk by carrying out its lender-of-last-resort function through open market operations. But do the other elements of the safety net—deposit insurance or the payment system—represent risks to government, and if so can those risks be eliminated? Those questions were addressed in later sections of the conference.

Deposit Insurance

Whether the government bears any risk for the deposit insurance system was the subject of a paper by Bert Ely, a banking consultant. Ely argued that, after the Federal Deposit Insurance Corporation Improvement Act of 1991, the federal government is no longer at risk for the FDIC’s losses. Ely noted that the act required the FDIC to levy an assessment on all insured banks and savings and loans sufficient to build up a reserve fund equal to 1.25 percent of all insured deposits. If that fund is depleted by future deposit insurance losses, the FDIC is authorized to replenish the fund by making emergency special assessments on all insured banks. To the extent that the FDIC has to draw upon its $30 billion line of credit at the Treasury in order to meet emergency liquidity needs, it must repay those borrowings with interest out of its special assessments on insured banks.

Accordingly, Ely pointed out, any liability for FDIC losses today is not a government liability. Ultimately, deposit insurance is a liability of the entire banking industry, even though the FDIC administers the system. Thus, although there is a form of moral hazard associated with deposit insurance, the FDIC is not at risk and has no incentive to reduce it.

In commenting on the Ely presentation, FDIC General Counsel William F. Kroener agreed with Carter Golembe’s point that there is a strong aversion among regulators to any instability in the banking system, although FDICIA attempted to remove much of the discretion regulators previously had either to intervene or to forbear. Kroener also agreed that the effect of the deposit insurance system in its current form is to subsidize smaller, less stable institutions at the expense of large ones, and that as the deposit insurance system is currently structured the FDIC bears no financial risk for bank losses.

The Ely presentation suggested that the deposit insurance system, because it posed no direct risk of loss to the FDIC, did not create a policy basis for federal regulation of banks. However, since any insurance system risks the creation of moral hazard, the banks themselves—since they ultimately bear the system’s risks—should ultimately control the administration of the deposit insurance system.

If the Fed’s lender-of-last-resort function could be performed without financial risk to the federal government, and deposit insurance should ultimately be the risk of the banking industry itself, the remaining question is whether regulation and supervision could be justified because of the risks the Fed bears in its operation of the large-dollar payment system.

The Payment System

The Federal Reserve assumes a risk in the payment system because all interbank fund transfers are netted and cleared at the end of each day, but the Fed in effect guarantees to all recipient banks that transfers they received intraday are final—that is, that they will not be reversed if the sending bank is unable to meet its clearing obligations at the end-of-day settlement. Although payment finality is a considerable benefit to the efficiency of commercial and financial transactions, the Fed clearly bears the risk of a paying bank’s insolvency.

Randall Kroszner of the University of Chicago Graduate School of Business discussed how private payment systems manage risk, suggesting that a private payment system could be developed to operate without risk to the Fed. Focusing on the contractual arrangements that currently exist in the futures and derivatives markets, Kroszner showed that it is possible for a private system to achieve payment finality without government involvement.

The clearing and settlement system at the Chicago Board of Trade, Kroszner noted, effectively creates payment finality. In that system, the Board of Trade Clearing Corporation (BOTCC) has been established as a credible counterparty for each of the members engaged in a clearing and settlement. Members are required to post collateral, or margin; BOTCC itself has capital contributed by the members; and the corporation has the authority to assess the members for any losses it suffers in the clearing process.

All transactions are settled between members and the BOTCC, and if a member fails to make a required payment at settlement, the BOTCC makes the necessary payment to the other party to the trade. Its financial resources come from the defaulting member’s margin, BOTCC’s capital or reserve fund, and—if necessary—an assessment on the other members. This mutualization of risk both creates payment finality and reduces systemwide risk by preventing the failure of one member from affecting the others.

Because BOTCC and similar clearinghouses have to be concerned about adverse selection and moral hazard, they take on a regulatory role, imposing requirements on members with respect to liquidity, capital, and activities. Kroszner pointed out that, historically, clearinghouses of this kind have been extremely stable, surviving the Great Depression, the Second World War, failures of major firms such as Barings Securities, and high levels of volatility in the last decade.

Kroszner also showed that before the establishment of the Federal Reserve System, bank clearinghouse associations performed many of the functions that the Fed performs today, including lender-of-last-resort functions. Those clearinghouses monitored their member institutions, and in some cases established explicit rules on liquidity and capital. In other words, bank clearinghouses were incipient forms of what the BOTCC is today. That suggests that an interbank payment system could be established and operated safely and soundly by banks themselves, without the intercession of or risks to the Federal Reserve.

Commenting on Kroszner’s presentation, Ronald D. Reading of First Manhattan Consulting Group described a clearing and settlement system for currency trading recently established by a number of major foreign currency trading banks. Although the system does not place the clearinghouse in the position of a counterparty—the party on the opposite side of a contract—it does create a system that eliminates settlement risk. Nevertheless, although Reading agreed that this structure could be organized as a counterparty—and thus create some degree of payment finality—the size of the risks involved is so great that the Fed’s role as a lender of last resort could not be entirely eliminated.

Conclusion

The presentations and discussion strongly suggested that the federal government’s risks in maintaining the bank safety net are the sole basis for the strict regulation of banks. The safety net was created to address the banking system instabilities that derived from faulty banking policies—particularly unit banking and prohibitions on interstate banking. However, the government’s aversion to instability in the financial markets has resulted in the excessive use of the safety net by bank supervisors, thus increasing moral hazard and ultimately contributing to—rather than allaying—long term financial system instability.

Ironically, today, after changes in the economy and banking policy have eliminated the causes of instability—particularly unit banking and restrictions on interstate banking—the desire to shield the government from the risks it is assuming with the bank safety net seems to be the underlying rationale for maintaining strict bank supervision. Under these circumstances, several questions arise: If the safety net does not promote long-term financial system stability, should it be retained? Does the safety net represent a risk to the government? If it represents a risk, are there ways that the government can eliminate the risk?
When reviewed individually, it appears that the various elements of the safety net—deposit insurance, the Fed’s function as lender of last resort, and the Fed’s role in the large-dollar payment system—either do not to represent a risk to the federal government (deposit insurance), could be administered by the Fed without risk (open market operations for its lender-of-last-resort responsibilities), or could be carried out effectively by the banks themselves without government involvement (a private payment system). In that case, a strong argument could be made that government regulation of banks would be unnecessary.