G. G. Johnson
Until the 1930s, banking crises occurred with disturbing frequency, starting with problems in a few institutions, but then spreading to others and disrupting financial intermediation. Since then there have been no major banking crises. Individual bank failures have, of course, occurred, but the fact that these have not resulted in general financial panic bears witness to the success of banking “safety nets”—the mechanisms for avoiding bank failures and for containing the effects of the isolated failures that inevitably continue to occur.
The failures of the Penn Square National Bank in the United States and Banco Ambrosiano in Italy in 1982 and the spread of debt servicing problems of major international borrowers, such as Brazil and Mexico, have raised new questions about the continuity of international financial intermediation by banks. Are the safeguards that have evolved over the past several decades sufficient to cope with the complexity and the increasingly international nature of banking?
This article is drawn from a longer study, Aspects of the International Banking Safety Net, by G. G. Johnson, with Richard K. Abrams, published as IMF Occasional Paper No. 17. See page 9 for ordering details.
Current banking problems
Debt crises in Latin America and Eastern Europe have been the most visible and dramatic of international banking’s recent problems. A number of countries that have borrowed heavily from banks are having difficulty meeting their debt service obligations and, in some cases, bank exposures to some of the largest sovereign borrowers are not far short of bank capital. But the amount that banks might have to write off for individual countries, in the event of default, would be unlikely to be more than a small proportion of their exposures, though they might well experience a prolongation of amortization. At the aggregate level, moreover, loans to developing countries amount to only around 7 per cent of the total assets of banks in the industrial countries. Though aggregate figures should be used with care, and with due regard for individual problem areas, they do suggest that the more alarmist versions of the vulnerability of the banking system to international problems are overdrawn.
Problems for international banking can also originate in domestic banking difficulties. In the 1930s, the domestic problems banks faced in a number of countries induced them to pull back the funds they had placed abroad, mainly with other banks, thus spreading the problems internationally. Certainly there are problems today with some part of the 93 per cent of bank assets that consist of loans to domestic clients (or to clients in other industrial countries, for whom “country risk” is largely absent).
In a number of countries banks recently faced some “tiering” of interest rates—that is, lending in the interbank market has taken place at different rates for different banks, depending on market perceptions of their creditworthiness. This is the first instance of widespread tiering since 1974, though so far its magnitude is much less than it was then. To some extent it may reflect new institutional developments, such as the growth of money market funds in the United States. Since the Penn Square failure, money market funds have become extremely cautious about where they place their money, and with the current uncertainty, movements of deposits have been large and abrupt. Some tiering is, of course, normal, and the recent tendency is, arguably, a healthy development, reflecting the operation of market mechanisms to prevent imprudent banking. Whatever the reason for tiering, its current extent suggests that some depositors are concerned with whether present safeguards—the safety net—will ensure bank stability.
The “safety net” concept
The term “safety net” is used here in a very broad sense. It includes not only mechanisms for financial support but also the various safeguards that bankers, bank supervisors, and lenders of last resort have developed to avoid the need for financial support in the first place. The aim of safety nets is, first, to avoid bank failure, though in a market system it is neither feasible nor desirable to prevent bank failure altogether; second, to prevent “contagion” from one bank’s failure from causing problems for other banks; and, third, where such problems do arise, to contain them so that the economy is not damaged by a loss of the intermediary services that banks provide.
What are the elements of a safety net for banking—domestic or international? These can be thought of as a series of defenses: prudential measures to protect bank solvency and bank liquidity; official assurances to depositors that their deposits are safe, even with troubled institutions; provision of liquidity to permit solvent institutions to keep functioning in the face of a loss of depositor confidence; and reliable and known procedures for the orderly resolution of the problems of banks in difficulties.
How does the safety net function for international banking? International banking is essentially an extension of domestic banking; in general, large banks whose main activities are domestic also have substantial international involvement. By the same token, a safety net for international banking is an extension of domestic safety nets. An evaluation of the strength of the international banking safety net is thus primarily a question of evaluating its domestic counterparts. There is an important caveat, however: the links between safety nets are not quite fine enough to ensure that some banks will not fall between them.
The first line of defense for the stability of banking systems is that individual banks themselves act to reduce the risk of failure. Acceptance of risk is inherent in financial intermediation, but banks must limit their risks to prudent levels.
Beyond such basic requirements as the development of techniques of credit evaluation and internal controls to limit the potential damage from incompetence or fraud, prudential standards for banks include diversification of assets and liabilities to ensure that the problems of one or a few customers of a bank cannot be critical. They also include limitations on the extent to which banks carry out maturity transformation, that is, the extent to which banks can make long-term loans funded by short-term deposits. Perhaps most important, prudential standards include maintenance of bank capital at such a level that the bank can absorb substantial losses without becoming insolvent.
Since no system of prudential control can eliminate the possibility of bank failure, depositors can never be certain that banks will be able to repay their deposits; and if doubts arise about the solvency of a bank, depositors will attempt to withdraw their funds. To protect themselves against unexpected withdrawals, banks maintain a prudent degree of liquidity. They keep part of their assets in liquid form (which, incidentally, may also reduce the risk of loss from interest rate fluctuations) and arrange for lines of credit with other banks. As with other prudential guidelines or controls, banks need to strike a balance between prudence and profitability: more liquid assets have lower yields, while charges apply to agreed lines of credit with other banks.
An important aspect of prudential controls in international banking is the progress that has been made in the international coordination of banking supervision. The main organ for coordination has been the Basle Committee on Bank Regulations and Supervisory Practices, which operates under the auspices of the Bank for International Settlements (BIS). (Often this is referred to as the “Cooke Committee,” after its Chairman, Peter Cooke, head of banking supervision in the Bank of England.) The Committee, which was set up in response to the near-crisis of 1974, serves as a forum for international discussion of bank supervisory issues. The work of the Committee appears to have substantially strengthened prudential practices, particularly through its efforts to ensure that no institution slips through the cracks of banking supervision. The Committee’s “Concordat” of 1975 addresses the question of sharing responsibility for the supervision of bank solvency. (Bank supervisors are quick to point out that the Concordat applies to supervision only; it does not apply to the question of who might provide financial support to a troubled bank.) The Concordat provides for
… some sharing of the responsibility for supervision between host and parent authorities, with the emphasis varying according to the type of establishment concerned. For foreign subsidiaries and joint ventures, primary responsibility rests with host authorities; but, in addition, parent authorities must take account of the exposure of their domestic banks’ foreign subsidiaries and joint ventures because of those parent banks’ moral commitments to those foreign establishments. For foreign branches, solvency is indistinguishable from that of the parent bank as a whole. It is therefore essentially a matter for parent supervisory authorities.
The somewhat ambiguous phrasing of the Concordat with respect to foreign subsidiaries and joint ventures reflects the fact that the various national authorities were not entirely in agreement on the partitioning of supervisory responsibility.
Since the adoption of the Concordat, the Basle Committee has attempted to give it more precise operational definition. The most important step in this direction came in 1978, when the BIS governors endorsed the Committee’s proposals that the evaluation of the adequacy of bank capital by supervisors should be carried out on the basis of consolidation of the positions of subsidiaries, as well as of branches, with the position of the parent bank. While not diminishing the responsibility of the host authority in the case of domestic subsidiaries of foreign banks, this did increase the responsibility of the authorities in the country where the bank had its headquarters. Some countries whose banks are active internationally, such as the United States, have long evaluated solvency on a consolidated basis; most others have adopted the principle in the last few years. A recent revision of the Concordat, in May 1983, explicitly recognizes that the need to examine the totality of banks’ worldwide business on a consolidated basis increases the reponsibility allocated to parent supervisory authorities.
The partitioning of responsibility, which seems well established in principle, still leaves some gaps in practice. In any case, whatever the precautions taken to protect bank solvency, problems are bound to arise from time to time. It is a truism that the “next banking crisis” will come from an aspect of banking that was not considered a source of difficulty, since otherwise prudential measures would have been adopted to deal with the problem. Savings and loan associations in the United States, for example, would presumably not have undertaken maturity transformation to the extent they did, if they had foreseen the levels to which interest rates would climb or the extent to which institutional innovations, such as money market funds, would generate competition for deposits from their traditional sources of funds.
By the same token, no degree of care in maintaining bank liquidity could make banks completely immune to contagion from problems of other banks. While it is essential to a competitive banking system that badly managed banks be allowed to fail, a strong case can be made for official support for institutions that face difficulty purely because of contagion from the difficulties of other banks.
One aspect of official support, deposit insurance, is largely irrelevant to international banking. This is not because of any discrimination against international depositors; in fact, most schemes apply equally to resident and nonresident deposits. The problem is that limits of coverage are small relative to the size of international deposits.
This raises the question of the role played by lenders of last resort, which are normally central banks. Lenders of last resort both provide financial support to banks that are basically solvent, but face temporary liquidity problems, and work out procedures for the orderly resolution of the affairs of banks that are actually failing. In fulfilling these functions lenders of last resort face the problem of “moral hazard.” This expression originated with the economics of insurance, where it initially referred to the danger of fraud where insurance applied. In banking it has acquired a more general connotation. Any measure that reduces the extent to which the market penalizes imprudent behavior, or rewards prudent behavior, may well encourage institutions to behave imprudently. The dilemma faced by lenders of last resort has been described by J. Spero (The Failure of Franklin National Bank) as “the problem of trying to make depositors feel confident without making bankers feel complacent.”
Thus, beyond the official support, which is almost automatically available in many banking systems through, for example, central bank discount windows, the extent to which the authorities are prepared to assist an institution is purposely left vague. Generally, however, they indicate their intention to intervene when they judge it to be necessary, and confidence in their capacity to do so has been generated by the success of their past interventions.
While the role of lenders of last resort in domestic banking systems seems by now to be well established, some uncertainties remain in international banking. One area of uncertainty is the treatment afforded to foreign creditors in the case of failed banks. In the 1973 failure of the U.S. National Bank of San Diego, and again in the failure of Bank-haus Herstatt in the Federal Republic of Germany in 1974, there was some initial uncertainty among foreign creditors that they would receive equitable treatment. Ultimately, these fears proved not to be justified, and, by and large, equitable treatment of foreign creditors now seems well established. A recent example is the failure of Banco Ambrosiano, where both foreign and domestic depositors with the domestic offices of the parent bank received full repayment.
Treatment of subsidiaries
The treatment of the creditors of foreign subsidiaries of Banco Ambrosiano, however, was quite different. That treatment provides an example of another area of uncertainty about the role of lenders of last resort—that is, who takes responsibility for a branch or subsidiary whose headquarters lies in another jurisdiction? In the case of branches, any support that the parent authority is prepared to extend to the parent bank extends virtually automatically to foreign branches, since for most legal purposes the parent bank is fully responsible for meeting the obligations of the branch.
The situation of subsidiaries is less clear. While the parent bank does not, in general, have formal legal responsibilities, it could normally be expected to support the subsidiary in order to preserve confidence in itself. Host authorities try in a number of ways to reinforce the parent bank’s acceptance of “moral responsibility.” Authorities in the United Kingdom, for example, originated the practice, which Canada and others have also adopted, of requiring “letters of comfort.” The U.K. letters define “moral responsibility” as “responsibility to support those investments beyond the narrow limits laid down by laws of limited liability and, above all, as responsibility to protect depositors with those banks.”
The acceptance of moral responsibility by bankers means that, in most cases, a troubled subsidiary will be supported by the parent. But what happens in the case where the parent itself is in difficulty? Lenders of last resort for parent banks have not, in general, acknowledged that they incur any obligations as a result of comfort letters or other informal understandings between parents and subsidiaries.
In the case of Banco Ambrosiano Holding, the Banco Ambrosiano subsidiary in Luxembourg, the Italian authorities declined to take responsibility. There were a number of institutional complications in that case; Banco Ambrosiano Holding was not a bank under Luxembourg law, and the role of guarantees by the Vatican Bank was unclear. The controversy nonetheless pointed up the uncertainties. Some observers questioned the propriety of the Italian authorities protecting depositors with the parent bank but failing to do so for the subsidiary; on the other hand, Professor Klaus Kohler, a director of the German Bundesbank, noted in a Reuters interview that “no central bank has written a blank check for what happens somewhere in the world where one cannot look in.” Whatever the merits of the case, it is clear that there is no concordat allocating responsibilities among lenders of last resort.
The importance of this gap, however, should not be exaggerated. The Banco Ambrosiano case was a special one involving fraud and problems of supervision, and the authorities in both Luxembourg and Italy have taken steps to correct these problems. When both the subsidiary and the parent are located in industrial countries, the two authorities can usually be expected to arrive at a satisfactory division of their responsibilities as lenders of last resort.
Questions still remain about the status of subsidiaries of banks based outside the industrial countries, as the parent authority might lack the ability to provide support. The status of subsidiaries operating in many of the offshore centers might also be in question, as there might be neither a local lender of last resort nor a strong understanding on the moral obligations of the parent bank and its authorities. Such banks are in a sense on the fringes of the system and it is unlikely that failure of a few of them would, in itself, lead to major problems for banks in general.
Another question in the international area arises from considerations of the balance of payments effects of bank problems. The lender of last resort can generally provide whatever liquidity is needed in domestic currency; the problem is what to do if deposit withdrawals lead to capital outflows that affect exchange rates or deplete foreign exchange reserves. A communique following the September 1974 monthly meeting of the Central Bank Governors of the Group of Ten and Switzerland raised this issue:
The governors had an exchange of views on the problem of lender of last resort in the Euromarket. They recognized that it would not be practical to lay down in advance detailed rules and procedures for the provision of temporary liquidity. But they were satisfied that means are available for that purpose and will be used if and when necessary.
While typically vague, this statement, which has been reaffirmed on a number of occasions, was designed to provide assurance of the availability of liquidity to deal with system-wide banking disturbances.
Despite the general strength of the international banking safety net, in one respect it falls short of what is expected of domestic safety nets. The ultimate purpose of such safety nets is to ensure that production and income are not threatened by a loss of the intermediary services of banks. There is little doubt that the international banking system, as such, can be preserved; there is still a question, however, whether banks will continue to play their necessary role in intermediating between surplus and deficit countries. Banks have been insufficiently cautious in the past about their international lending, as the seriousness of country debt problems indicates. But at the moment the slowdown in bank lending may reflect an excess of caution.
Banks’ recent problems should, however, lead to a stronger system in the long run. A more discriminating approach by depositors should, in general, be salutary. The moral hazard pendulum had perhaps swung too far, with depositors thinking they had no risk of loss, thus facilitating incompetent behavior in some institutions. A more cautious attitude on the part of depositors will make it more difficult for institutions with less than solid credentials to expand rapidly in risky or questionable areas of activity. At the same time bank supervision is likely to be strengthened, and there is likely to be further clarification of international understandings in that area. There is also likely to be a better delineation of the relative responsibilities of the various national authorities for the orderly resolution of the problems of failing banks. Taken together, these developments should reduce the extent to which banks can evade the intent of regulation and supervision of prudential behavior by locating in centers where monitoring is lax.
We are pleased to announce the publication of a selection of articles from Finance & Development in Chinese. This selection, translated and produced by the China Financial & Economic Publishing House of Beijing, will be distributed primarily in China.