Current Legal Issues Affecting Central Banks, Volume IV.

Chapter 27 International Banking Capital Standards for Market Risk: Recent Developments and Possible New Directions

Robert Effros
Published Date:
April 1997
  • ShareShare
Show Summary Details

This chapter addresses the emerging international capital standards, particularly the standards for market risks arising from the trading activities of banks. The basic framework under consideration for structuring new capital requirements is examined. The April 1993 proposals of the Basle Committee on Banking Supervision and some of the thoughts of the Committee regarding new standards are reviewed. The Committee was considering two risk-measurement techniques: one based on a so-called standard approach that applies risk weights to various trading positions, and another based on the results of a bank’s own internal models.

Basle Capital Accord

The existing capital standard that applies to internationally active banks headquartered in many of the world’s largest countries was created in late 1988 under the auspices of the Bank for International Settlements, in Basle, Switzerland.1 That standard, known as the Basle Capital Accord, was deemed necessary by the Committee for several reasons: (i) to make regulatory capital requirements more sensitive to differences in risk profiles among banks; (ii) to introduce off-balance-sheet exposures into the assessment of capital adequacy; (iii) to minimize disincentives to holding liquid, low-risk assets; and (iv) to achieve greater consistency in the evaluation of the capital adequacy of major banks throughout the world.

In large part, the Basle Capital Accord has accomplished these goals. By design, however, it addressed only credit risk—the risk that a borrower will default on its obligation and not repay the bank. The Basle Capital Accord did not address other important risks that banks face, particularly the risks that evolving market conditions, such as changing interest rates or changes in the prices of equity instruments that banks trade, will directly and adversely affect a bank’s financial strength. Consequently, soon after the Basle Capital Accord was introduced, the Committee established several working subgroups to address these acknowledged shortcomings and to enhance the standard with capital requirements for interest rate risk and for risks in trading equities and foreign exchange.

Initially, these working subgroups worked independently in their respective efforts to develop adequate capital standards that could be used by bank regulators worldwide. Although the focus was only on activities of “internationally active banks,” it was recognized that the standards would likely cover far more than just the world’s largest financial institutions. In some countries, many relatively small banks are considered to be internationally active because they finance international commerce and trade foreign exchange. Competitive conditions within countries have also led some regulators to extend international standards to banks with little or no truly international operations. In the United States, for example, requirements of the Basle Capital Accord have been applied to all commercial banks.

Such wide-ranging coverage and the diversity of banking industry structures among countries dissuaded the working subgroups from developing standards that only the largest and most sophisticated institutions could implement. Simple or “shorthand” rules were thought to be useful, so that banks with limited internal modeling abilities could calculate their capital requirements without investing heavily in new computer systems and additional personnel.

The first objective of the Interest Rate Risk Subgroup (IRR Subgroup) was to develop a technique for measuring the exposure of a consolidated banking institution to interest rate risk, and the subgroup made some progress toward that end. Soon, however, the subgroup’s priorities were changed because of events in the European Community (EC). In order to accommodate its planned economic integration, the EC needed common capital standards for banks and securities firms, so that those two industries could compete throughout the EC on an equitable and sound basis. This pressing need of the Europeans required the IRR Subgroup to shift its attention to the trading activities of banks.

Consistent with that purpose, the work on traded equities and foreign exchange was transferred to the IRR Subgroup, which was also temporarily expanded to include representatives of several securities regulators. It was hoped that the Basle Committee on Banking Supervision would be able to agree with the International Organization of Securities Commissions on capital standards for trading activities of both the banking and securities industries.

In 1992, the EC, facing its own time pressures, reviewed the work in progress of the IRR Subgroup, made slight revisions to those draft documents, and adopted as European law tentative unified standards for the two industries. The resulting Capital Adequacy Directive2 took effect at the beginning of 1996 and represents capital standards that EC banks and securities firms must meet. Subsequently, in April 1993, the Committee issued its own (and slightly different) proposal for a public comment period that extended through the end of 1993.3

Initial Market Risk Proposal

The Committee’s initial market risk proposal covered traded debt, equity, and foreign exchange, whether reported on or off the balance sheet.4 It also included an important provision to expand the netting of counterparty obligations beyond that originally permitted by the Basle Capital Accord, which recognized only netting by novation.5 This latter procedure is typically limited to foreign exchange contracts under which obligations between banks and their counterparties to deliver a given currency on a given date are combined, thus legally substituting a single amount for the previous gross obligations. The proposal permitted other forms of legally enforceable bilateral netting under specified conditions.

The consultative document of the Committee also discussed the IRR Subgroup’s efforts to measure interest rate risk in the nontrading activities of a bank. This part of the document, however, is more accurately described as a status report on the working subgroup’s thinking, rather than as a specific proposal. It is also viewed only as a risk measure for supervisory use, rather than as a formal capital standard.

Traded Debt

The proposal dealing with traded debt is the most complex of the April 1993 proposals. When developing this proposal, the Committee relied heavily on the concept of duration, which is a useful technique for estimating an instrument’s price sensitivity to changing interest rates. This measure is determined by the timing of an instrument’s principal and interest cash flows, is relatively easily calculated, and is commonly used by financial institutions and investors. In effect, it represents the percent change in market price for a given percentage point change in market rates. Duration has several important limitations and is most accurate only when estimating the effect of small market rate shifts. However, as a simple and low-cost indicator of an instrument’s price sensitivity, it has significant appeal.

Using this duration technique, banks can distribute their trading assets, liabilities, and off-balance-sheet interest rate contracts (for example, interest rate swaps, futures, and forward rate agreements) among a variety of time bands on the basis of the instrument’s maturity or next repricing period. The supervisor provides the duration risk weight for each time band by calculating the duration on a hypothetical instrument that has cash flows presumably representative of the instruments that a bank will report in each time band and assuming a change in market interest rates for that band. By multiplying the amount in each time band by its risk weight, the institution can derive a total net weighted position for its entire trading account.

Stopping there, however, ignores important risks. First, the actual shifts in the market yield curve will probably never be exactly those assumed by the risk measure, so the actual changes in the market value of the trading portfolios will be different from those that were estimated. This risk is called yield curve risk. Second, opposite positions within a given time band in different instruments (for example, a long position in U.S. treasury bonds offset by a short position in corporate obligations) are not likely to respond to a given market rate change in exactly the same way. This risk is called basis risk.

The range of time within each time band presents another source of possible measurement error. The risk weight is derived by assuming that the maturities or repricing periods of the bank’s instruments are distributed evenly throughout a time band and that duration risk weights can be determined by using the midpoints. In practice, however, instruments that mature or reprice at opposite ends of a time band (which can range from one to three years) could have significantly different characteristics of interest rate risk. If positions are not distributed evenly within the time band, the measure will overestimate or underestimate the risk. Finally, in the interest of minimizing the reporting burden, the Committee proposed to collect no information about an instrument’s coupon. Although that information is typically less important than an instrument’s maturity or next repricing period, the coupon rate affects the instrument’s price sensitivity to changing market interest rates.

To address these possible measurement errors, the Committee proposed the imposition of a series of add-ons to the initial calculation of the net weighted position that have the effect of “disallowing” part of the initial netting. The details of calculating these add-ons are beyond the scope of this chapter and are somewhat complicated. They are, however, similar in concept to techniques used by some securities regulators outside the United States.

Foreign Exchange

Because some smaller institutions (especially in Europe) trade foreign exchange, the Committee developed both simple and sophisticated risk-measurement approaches. The simple approach requires capital equal to 8 percent of the sum of the larger figure of the institution’s net longs or net shorts in each currency. The sophisticated approach simulates potential losses to the institution’s existing portfolio of foreign exchange positions by using the actual daily exchange rate movements experienced during the past five years (assuming a two-week holding period). The capital required should cover 95 percent of the simulated losses. It was expected that most large institutions and those with significant foreign exchange trading activities would use the sophisticated method.


The risk measure proposed by the Committee for equities addressed separately the two fundamental causes of movements in market prices: specific and general market risk. Specific risk refers to events that are relevant only to an individual firm—in this case, the issuer of the equity shares. Market risk refers to adverse effects on a company’s shares in response to developments in the general market. The total capital charge for equities is the sum of the amounts assessed for the two risks.

A capital charge of 4 percent can be applied against the net position in any specific security (for example, IBM or Toyota), whether that net position is long or short. Capital is required against either net position because the trading institution can lose on both types of exposures at the same time: the share prices for which the trading institution has a net short position can rise in value, while those for which it has a net long position can decline.

General market risk, meanwhile, relates to the overall position of the portfolio, which can be either net long or net short. For this risk, the Committee proposed a capital requirement equal to 8 percent of the bank’s net equity trading position.

Off-Balance-Sheet Contracts and Netting

One of the more important features of the 1988 Basle Capital Accord was its recognition of counterparty credit risk arising from off-balance-sheet transactions, such as those related to foreign exchange and interest rates. The level of credit risk arising from these types of transactions can change rapidly in response to market conditions, as a counterparty’s obligation increases or declines.

In order to incorporate this fluctuating credit risk, the Basle Capital Accord requires institutions to hold capital against their current exposure arising from these transactions plus an add-on for their potential future exposure. The current exposure is equal to the market value of the contract if it is a positive value or equal to zero if the market value is zero or negative. Current exposure represents the replacement cost that an institution would incur if its counterparty were to default. The potential future exposure is estimated by multiplying the notional value of the contract by a credit conversion factor ranging from zero to 5 percent, depending upon the type and remaining maturity of the contract; it measures the amount by which the current credit exposure may be expected to increase over the remaining life of the contract.6

When the original Basle Capital Accord was created, supervisors recognized that institutions, especially those actively involved in off-balance-sheet transactions as dealers, would likely have multiple contracts with the same counterparty and that the positive values of some of these contracts could be partly or entirely offset by contracts having negative market values. When the two parties to the transactions have legally binding agreements to settle their obligations on a net basis, the actual credit exposure of each institution becomes that net amount. However, because of their concerns about the legal enforceability of many netting agreements between and among counterparties, the Basle Committee recognized netting only under the limited case of novation. Institutions were encouraged to pursue other bilateral and multilateral netting agreements, but the benefits of these agreements were ignored for capital purposes.

This issue of netting relates to credit risk and is not integral to measuring market risk. The matter is, however, of substantial importance to those institutions most affected by any new market risk standards because off-balance-sheet transactions are significant to their trading activities.

For this reason, and to acknowledge and further encourage the risk-reducing use of netting agreements, the Committee addressed the issue of bilateral netting in its April 1993 market risk proposals.7 Nevertheless, the Committee remained concerned about the ability of institutions to enforce their bilateral netting agreements in many jurisdictions. Therefore, it proposed to recognize those agreements only when the following conditions have been met:

  • The bank has a legally enforceable netting agreement with its counterparty that permits the bank to receive or pay only the net value of the sum of unrealized gains or losses on the included transactions if the counterparty defaults.
  • The bank has written and reasoned legal opinions stating that, in the event of a legal challenge, the courts and authorities in the relevant jurisdictions would support settlement on a net basis. In this context, if a supervisor in the home or host country of either counter-party is dissatisfied about enforceability, neither counterparty may calculate capital requirements on a net basis.
  • The netting agreements do not contain so-called walkaway clauses, which permit non-defaulting parties to make only limited payments or no payment at all to the defaulter, even if the defaulter is a net creditor.8

Recent legislation in the United States (the Federal Deposit Insurance Corporation Improvement Act of 1991) clarified the enforceability of netting agreements between financial institutions in that country; this legislation should provide a further boost to the use of these agreements.9 Continuing growth in these markets and the recognition of netting contracts for capital purposes may lead other countries to take similar steps in the future.

In July 1994, the Committee adopted (in substantially the same form) the April 1993 proposal on bilateral netting for the purpose of calculating current credit exposure.10 At that time, it also issued for comment a proposal that would give limited recognition to the effects of netting agreements when calculating potential future exposure.11

Public Comments on the Market Risk Proposal

Although the proposal dealing with bilateral netting was strongly favored by respondents and adopted by the Committee, most other aspects of the April 1993 proposal were widely criticized. In the United States, especially, criticism was strong, as respondents urged greater use of their internal models for measuring risk. These models, they felt, offered many advantages over the proposed approach: (i) greater accuracy; (ii) greater consistency with existing risk-management techniques; and (iii) more adaptability to new products. In contrast, the proposed method would, for many institutions, be only an additional regulatory burden.

Although respondents in other countries also criticized the proposal and supported the use of the internal models, European banks, in particular, focused principally on the differences between the Committee’s proposal and the Capital Adequacy Directive.12 Nearly all of them called for standards the same as or similar to those in the directive, which—insofar as it differed from the Committee’s proposal—would require less capital. This somewhat disparate response reflects the relevance of the Capital Adequacy Directive to European banks, but not to the others; it may also reflect different practices regarding internal models.

Internal Models

Beyond simply reflecting the preference of many international banks, the use of internal models has other benefits over the April 1993 proposals—not the least of which is an internal, theoretical consistency. Although the key elements of the 1993 proposal were based on empirical analysis, they also reflected multilateral compromises that were necessary to achieve an international consensus among bank regulators and that also held promise for reaching an understanding with securities regulators.

By relying heavily on historical price movements and measuring the risk in diversified portfolios consistently and systematically, internal models sidestep many of the problems created by the earlier proposals. Nevertheless, these models have their own limitations in the context of a capital standard, and supervisors need to learn more about them. In particular, the differences and similarities among the internal models need to be understood better, so that they can be incorporated into a capital measure.

This task is made more difficult because banks have designed these models for another purpose. Supervisors seek to evaluate the adequacy of a bank’s capital under highly stressful market conditions, but bank managements typically use models to help them manage risks in normal times. When the risk measures show rising market volatility and, therefore, a greater likelihood for significant gains or losses, banks tend to advise their traders to reduce their open, proprietary positions. As markets calm and, especially, as trends emerge, positions are once again increased.

This focus on normal conditions and the complexity of many market risk models present obstacles that bank supervisors must overcome if they are to use models to evaluate the capital adequacy of banks. Securities regulators must also be convinced that these models can be used for capital purposes because development of a standard for both securities firms and banks remains an important objective of the Basle Committee.

Fortunately, while understanding each model’s assumptions and peculiarities is important, the models of most major banks are structured in similar ways; they rely on historical price movements and use similar, although not identical, risk-measurement techniques. These similarities range from methods for measuring present values of financial instruments to techniques for estimating the price volatility of instruments with option characteristics. The latter techniques are far from simple but relatively few in number. Therefore, if supervisors can overcome these obstacles, banks could build upon their existing procedures to evaluate capital adequacy rather than be required to design new risk measures that would serve only supervisory purposes.

One of the most common practices of banks, for example, is to decompose their trading positions into a variety of risk factors that influence an instrument’s price. For this purpose, each instrument is classified in several or more ways, beginning with its major product group (for example, interest rate, foreign exchange, equity, or commodity). Long-term rates move differently from short-term rates, and currencies constantly change in value relative to one another. Therefore, information is needed about the instrument’s maturity, currency, and other relevant features. A single instrument may involve only a few risk factors; however, large banks may use hundreds of these factors to manage their entire trading portfolios, depending upon the level of detail that they believe is needed to identify relevant variables and upon the nature and breadth of the instruments that they trade. By classifying and aggregating their positions by risk factors, banks can evaluate risks without constantly processing tens or hundreds of thousands of individual transactions.

Nevertheless, the specific procedures of an individual model cannot be ignored, as they can lead to significant differences in measured risk among banks. As is often said, “The devil is in the details.” Therefore, both bankers and bank supervisors should have knowledge of a model’s key assumptions and measurement techniques in order to understand and appropriately use the model’s results. These models must also reflect similar levels of rigor if their results are to be relied upon in constructing an international capital standard. Therefore, regulators are likely to require that some elements of the risk-measuring process be standardized. The unresolved question is, What parts?

Key Parameters

Several factors, or “parameters,” are especially important for the structure and output of models used by banks to measure their market risk. First, the historical observation period is used to measure the price volatility of traded products and to calculate the correlations of the price movements of instruments within and among product groups (for example, debt, equity, foreign exchange, or commodity contracts). Second, the confidence interval is needed to evaluate the model’s results. Third, the assumed holding period (or “investment horizon”) is relevant for each instrument. Other parameters that deal with risk-measurement techniques include (i) the manner in which the model relies upon price correlations between different product groups; (ii) the number and nature of risk factors that a model employs; and (iii) specific modeling procedures for evaluating explicit or embedded options.

One parameter that would almost surely be standardized in any supervisory framework is the holding period that banks assume. In practice, most trading banks calculate daily volatilities and assume a one-day holding period in managing their trading risks. That approach may be adequate for management purposes, as the turnover of trading portfolios is typically fast—often minutes or hours, rather than days or weeks. That assumption may be too weak, however, for supervisors, who must consider those periods of market turmoil when an instrument’s liquidity can disappear. Reducing exposures at those times may require institutions to recognize large losses. The assumption of longer holding periods is particularly important for instruments that have explicit or embedded options because normal daily volatilities may be too low to “trigger” a price shift that can produce a significant change in market value.

The confidence interval is another parameter that would likely require a standardized approach. Currently, when estimating the range of future price movements, each bank decides what confidence level its management needs. Usually, an interval of 95–99 percent is chosen. At that interval, the coverage of possible losses (resulting from a “one-tail” test)13 translates into roughly 1.7 to 2.3 standard deviations if the observations are distributed “normally,” as many bank models assume. The vast majority of daily market movements, however, are small, and their greater frequency overwhelms that of large, highly unusual market shifts when calculating standard deviations. Therefore, when a bank assumes normal distributions, the unusual market events tend to disappear or to have probabilities that seem much more remote than they are in reality. This result is of obvious concern to supervisors—not because of management’s practices, but because capital requirements should be based on highly stressful situations, not on “normal” levels of market volatility.

At best, some adjustment of standards seems necessary. The easiest part would be to require relatively strict confidence intervals for all banks, say 99 percent. The more troublesome aspect involves the “fat tails”14 of the distribution curve: the large, adverse events that normal distributions suggest will almost never happen, but that seem to occur somewhere every few years. One solution would be to use statistical techniques that do not rely on normal distributions; another would be to expand the bank’s results by some “adequate,” but necessarily arbitrary, multiple. Operationally, this latter approach is probably less disruptive to banking institutions, but it requires determining the “correct” size of the multiplier. Results of true stress tests may help guide that decision.

It is less clear whether other elements need to be standardized. The observation period is an example. In practice, banks seem to use periods ranging from the most recent few months to the past five years or more. Institutions that prefer short time spans believe that the most recent market movements best predict near future volatility; they also want to identify any short-term trends that may exist. Those using long time periods want to capture a wider range of financial market conditions than is possible with short periods; they place less importance on apparent trends.

In some respects, a bank’s selection of observation period can be irrelevant, provided that there is no inherent bias in the model’s results. When volatilities are calculated based on standard deviations, short time periods are not necessarily more or less conservative than long periods. Indeed, the use of short periods will at times predict larger potential market movements than long periods because their observations are not “averaged down” by the large numbers of more normal price changes. They will project greater volatility following periods of market instability and less volatility when markets are calm. Although the differences derived from using short and long sample periods may offset in time, different sample periods can produce widely differing estimates of future price movements for similarly risky portfolios at any given time. That result may be unacceptable to supervisors.

With respect to other parameters, permitting banks to use correlations of price changes between different types of products raises questions similar to those just discussed. Some banks calculate correlations among all trading product groups, while others assume, or force, such correlations to be zero, in the belief that any measured linkage—in the price changes of equities versus exchange rates, for example—is coincidental and should not affect the decision-making process. Still other institutions ignore correlations of rate or price movements among broad risk categories altogether and take the most conservative approach of assuming that the correlations are equal to one. Since these diverse practices can significantly affect the level of measured risk, supervisors will need to address this issue themselves as they finalize their capital requirements.

Output of Internal Models

The purpose of an internal model is to estimate the amount of “value at risk,” which represents the maximum value that management can expect the portfolio to gain or lose during a specific period (such as one day) with a given level of confidence (for example, 99 percent). Management and supervisors can compare the daily estimate with subsequent results when evaluating the accuracy of the model and the institution’s ability to manage risk. While the vast majority of daily results should be within the predicted maximums, one should expect to see results that exceed the estimates more often than the confidence level suggests. As noted above, actual trading gains and losses are not distributed normally, as most models assume; the industry sometimes witnesses extremely good and bad days.

A critical consideration should be the rate at which excesses occur and their distribution among realized gains and losses. If excesses occur much too frequently, one should question the accuracy of the model. If the excesses tend to be mostly losses, one might also question management’s ability to respond to adverse market conditions.

Qualitative Factors

A bank may be required to obtain supervisory approval of its internal model before using it for capital purposes, because such a model can involve complex calculations and assumptions and have substantial informational needs. Supervisors should also ensure that the model’s results are seen by management (and not generated only for examiners) and that the bank’s overall risk-management process is sound.

In July 1994, the Basle Committee on Banking Supervision issued a statement describing sound management practices for derivatives activities.15 That document could guide bankers with respect to the standards that supervisors could be expected to apply when reviewing their trading activities. First, boards of directors and senior management should exercise appropriate oversight. Directors need not be experts in derivatives or trading practices, but they should approve relevant policies and remain informed of the risks that the institution takes. Senior management would, of course, be expected to provide more extensive oversight and to monitor and control operations closely.

Second, an adequate risk-management process should integrate prudent risk limits, sound measurement procedures and information systems, continuous risk monitoring, and frequent reporting to management. This process should be tailored to the bank’s specific activities and staffed with individuals who understand the risks and are independent from the derivatives and trading functions.

Third, there should be comprehensive internal controls and audit procedures. Policies and procedures related to derivatives and trading activities should be fully integrated into routine work flows and consistent with those applied to other operations of the bank.


Developing an international capital standard for trading activities would be a complex and time-consuming process under any conditions. It becomes further complicated, however, when the risk-measurement framework is intended for institutions with significantly different levels of trading activities and expertise. Any agreed-upon approach must be adaptable to the diverse supervisory regimes in many countries and must involve acceptable costs, both to the supervisors and the regulated institutions. The standard must also, of course, produce a capital requirement that seems reasonable to all concerned.

The option of using internal models to determine capital requirements is attractive to many bankers and bank supervisors alike. Many key questions and issues must first be resolved, but the matter may be coming to fruition, partly as a result of industry comments. The pace and direction of the efforts of the Basle Committee on Banking Supervision are also influenced by other regulatory and political pressures caused by the growth of trading and derivatives activities throughout the world, as well as by highly publicized losses in these markets by both banking and non-bank institutions. Interest in developing and implementing capital standards for market risks is high.


In January 1996, the Basle Committee on Banking Supervision issued, subject to completion of appropriate rule-making procedures in member countries, an amendment to its capital standard to address market risk. Most non-European Union (EU) countries are expected to apply the new standards only to their largest institutions, which are relatively heavily involved in trading activities.16 As mentioned above, EU countries have a Capital Adequacy Directive17 that applies the standard more broadly to all banking and securities firms operating within their jurisdictions. Affected banking organizations would be required to comply with the Basle Committee requirements by the beginning of 1998, although participating countries may permit institutions to comply voluntarily at an earlier date.

As structured, the final amendment was highly consistent with the approach outlined in this chapter, although the simulation of foreign exchange positions is available only to institutions using the internal modeling approach; otherwise, the more simple technique would be required. The final ruling included these specific requirements regarding the use of internal models: (i) an institution’s historical observation period for measuring the volatility of past market movements must be at least one year; (ii) an institution’s “value at risk” should be based on a 99 percent confidence level and on an assumed holding period of ten days; and (iii) the modeling of yield curves should consider at least six different points (time bands) on each curve. The results of the modeling process would then be multiplied by three to provide further coverage for extremely adverse market conditions.

That multiplier reflects, in part, an analysis of market volatilities over nearly two decades of market movements and of the amount of capital needed to absorb the largest loss on a hypothetical diversified trading portfolio. The requirement of the ten-day holding period, noted above, is intended only to produce a sufficiently rigorous market movement by, in essence, applying an instantaneous market movement of the size normally experienced over a ten-day period to an institution’s existing portfolio. That treatment does not reflect an expectation by supervisors that institutions would actually hold a given trading portfolio for that extended period of time.

The U.S. federal banking agencies approved the amendment in August 1996 but will permit their institutions to use only the internal model approach. This decision reflects their view that internal models provide the best measure of true market risk and their desire to encourage all institutions to improve their processes for managing and measuring this risk.



This comment focuses on some of the issues that have arisen in the United States in the course of implementing the Basle Committee on Banking Supervision’s proposals on capital adequacy. One general supervisory and regulatory policy concern that pervades the various issues is the matter of regulatory burden. It is, admittedly, sometimes difficult to find a balance on this issue. On the one hand, if a very simple model is used to measure risk, it may not necessarily yield the most accurate information. The question then arises of why the model was proposed in the first place. On the other hand, if a complex and detailed model is required, it may not be suited to institutions that, for example, have small, specialized operations. In fact, such a model may well be burdensome for those institutions, as it could lead to an increase in their operational costs. Supervisors are sensitive to this balancing process. In the United States, the regulation of banking is extensive, and supervisors are accordingly aware of the burden that it imposes on banks. Approximately 14 percent of the banks’ expenses are devoted to complying with regulatory requirements.


The 1994 proposal1 by the Basle Committee to take account of netting arrangements in the calculation of capital adequacy has been implemented in the United States. When the proposed rules2 were being drafted, there were a couple of issues of particular significance. First, to qualify for netting, there must be an enforceable contract that can survive a counterparty’s default, bankruptcy, or insolvency. When the parties are under different jurisdictions—which can be either different state jurisdictions in the United States or very different legal frameworks across countries—it can become difficult to ascertain the effect of bankruptcy proceedings on the enforceability of contracts. Therefore, the requirement that the banks must provide legal opinions on the enforceability of the contracts to which they are parties may, in some cases, be difficult to fulfill. However, there does not seem to be any alternative: if there is no sufficient legal certitude in these matters, netting cannot be a valid option for capital adequacy purposes. The problem is further complicated in the case of bankruptcy, which may involve as counterparties not only financial institutions but also ordinary nonfinancial corporations. Bankruptcy and insolvency rules applicable to ordinary corporations may differ from those applicable to financial institutions.

With respect to the issue of regulatory burden, bank supervisors are aware that forcing a bank to go to an outside law firm every time that a legal opinion concerning the enforceability of a contract is required can become costly and may, in fact, be unnecessary. Another option is for the bank to have its own in-house legal counsel produce the opinion.

The Basle Committee did not expect that there would be an international harmonization of legal rules or opinions, or that similar opinions would be required of every member. Instead, banking supervisors in all member countries have been given the discretion to evaluate the sufficiency of their countries’ legal opinions. Because of the different legal frameworks, this seems like the sensible thing to do. Nevertheless, this requirement will obligate legal counsels and supervisory agencies to determine the sufficiency of these opinions.

Also noteworthy in the netting proposal was the suggestion—similar to that in the Basle Committee’s consultative papers—that, for risk-based capital purposes, contracts involving walkaway clauses be ineligible for netting.3 Moreover, in a bankruptcy case, the bankruptcy court or receiver must not be given option to “cherry-pick” the contract. If the rule were otherwise, the administrator of the bankrupt estate could pick and choose among the contracts, upholding those contracts in which the bankrupt estate was on the receiving side and disavowing those calling for payments by the estate. In some countries it may even be the duty of the administrator to cherry-pick, based on its fiduciary responsibilities. This is a matter that the legal counsel needs to look at when reviewing netting contracts.


The Basle Committee is looking at another increasingly prominent issue: supervision of the financial institutions using derivatives. The Eurocurrency Standing Committee of the Bank for International Settlements is also looking into these matters, as is the International Organization of Securities Commissions. It is possible that these agencies and organizations may eventually coordinate their recommendations, at least on some points.

A banking circular issued by the Office of the Comptroller of the Currency (OCC) highlighted for banks the standards that banks should maintain in their derivatives management practices.4 The OCC has also issued additional guidance in the form of questions and answers about derivatives use.5 In addition, the Basle Committee prepared guidelines on risk management of derivatives.6

The Basle Committee is committed to cooperating with other banking supervisory groups around the world. All concerned are aware of the need to harmonize and cooperate where possible, because banking has become increasingly international. The Basle Committee engages at least once a year in consultations with other supervisory groups. Similarly, the Basle Committee is considering holding annual or more frequent consultations with securities and insurance supervisors because of the issues that are increasingly overlapping their respective jurisdictions. Finally, the International Conference of Banking Supervisors, which has a larger membership than the Basle Committee, holds a conference every two years in which these and similar matters are considered.

    Other Resources Citing This Publication