OCC BULLETIN 1999-2
Subject: Risk Management
Date: January 25, 1999
To: Chief Executive Officers of National Banks, General Managers of Federal Branches and Agencies, Deputy Comptrollers, Department and Division Heads, and Examining Personnel
Description: Risk Management of Financial Derivatives and Bank Trading Activities - Supplemental Guidance
As of January 12, 2012, this guidance applies to federal savings associations in addition to national banks.*
The following guidance summarizes key lessons learned and fundamental control issues reaffirmed by experiences at commercial banks and other financial institutions, both domestic and international, since mid-1997. This issuance supplements the guidance already contained in OCC Banking Circular 277 (October 1993) and the Comptroller's Handbook for National Bank Examiners, Risk Management of Financial Derivatives.
Background, Scope, and Reference
Since the summer of 1997, deteriorating economies in Asian, Eastern European and Latin American countries have created significant volatility in global financial markets and contributed to unexpected declines in trading revenues, as well as counterparty credit losses, for some dealers. In some cases, losses resulted from design weaknesses in risk management systems, particularly with respect to their ability to anticipate and control the financial impact of volatile markets on their trading positions and counterparty credit portfolios. In other cases, such as Long Term Capital Management (LTCM), the problems were exacerbated by placing too much reliance on assumed financial strength and reputation and too much emphasis on meeting competitive pressures. Normal risk management standards were compromised as a result. These events reinforce the importance of developing, implementing and maintaining high quality risk management processes, notwithstanding competitive pressures.
The trading and credit losses experienced by the most affected banks generally did not materially alter their reported capital positions. However, they suffered significant losses in market capitalization as their equity prices declined disproportionately relative to other industry groups. As a result, investors and other external constituents have lingering concerns regarding banks' abilities to manage the credit, market, liquidity, compliance and transaction risks in their trading activities. Thus, market reality demands that banks place strong emphasis on developing and maintaining a superior risk management program for trading activities.
While these market events provided some new lessons, they primarily reaffirmed the importance of the fundamental risk management principles outlined in OCC Banking Circular 277 and the Comptroller's Handbook. This supplemental guidance should be factored into examinations of bank trading
Price Risk Management
Significant market volatility over the past eighteen months highlights the importance of assumptions in banks' price risk measurement systems. Value-at-risk (VaR) [NOTE 1: VaR is an estimate of the potential loss within a specified confidence interval in a portfolio's value over a defined holding period, assuming certain volatility factors and correlations.] is the most common method dealer/trading banks use to measure aggregate price risk.
Although VaR provides useful information, it is important that management and boards of directors understand the method's limitations. VaR is appealing because it reduces multiple price risks to a single number. However, current versions of VaR do not incorporate liquidity, credit, and other non-price risks. Moreover, VaR numbers are merely estimates. The quality of these estimates will depend on the care given to the quality of the assumptions and the appropriateness of the approach used. While an improvement in the quality of the inputs and the modeling approach can improve the confidence in the estimate, it is still an estimate. It is particularly important that management and boards understand that VaR does not provide an estimate for the maximum amount of losses a trading portfolio will ever suffer. For example, a VaR estimate of $5 million, using a confidence level of 99 percent and a one day holding period, means that the trading portfolio will lose less than $5 million over the next day, 99 percent of the time. However, VaR at the 99 percent level does not provide any information as to the magnitude of the potential loss, beyond $5 million, in the 1percent tail.
Notably, firm-wide VaR calculations actually performed quite well at several institutions during the recent volatility. However, once losses occurred, executive management and boards found that they did not have the loss tolerance they thought they had under previously approved VaR limits. Specifically, some managers and directors found that some external constituents viewed their bank's trading losses, even within an acceptable range, as a broad indictment of their corporate risk management and control processes.
Examiners should ensure that management and boards receive sufficient information to assess the strengths and limitations of bank price risk measurement systems in order to determine appropriate risk limits. This information should include a periodic summary of price risk measurement techniques and the results of independent validation and back-testing of price risk measurement models. Boards should also review management's response to strengths and limitations identified through validation and back-testing processes. Management and boards should consider this information, including the potential for indirect effects of downside performance beyond the bank's finances, when they determine and communicate their risk tolerance.
The technological advances in price risk measurement in recent years need to be coupled with the development of a strong and methodical program to stress test exposures. [NOTE 2: Stress tests evaluate portfolio performance under extreme, possibly improbable, scenarios.] Banks may be reluctant to incur expenses to measure events that management believes will never occur. Even where the commitment is present, the development of a sound stress testing methodology is challenging. Banks need to approach this effort with the same seriousness and rigor used to design their VaR systems.
Stress testing of exposures from trading positions is critical to a sound price risk management system, precisely because VaR provides no cap on losses outside the chosen confidence interval, and because the VaR estimate is, by definition, imprecise. Furthermore, the estimation of the probabilities associated with the largest losses "tail events" is even more imprecise. Thus, effective risk managers should ensure that the bank has properly identified, measured, and communicated to senior management scenarios that could threaten the bank's viability or reputation, regardless of the probability associated with those scenarios.
Stress testing should ideally involve both the risk control unit and the trading desk. Traders are generally the most knowledgeable about the portfolio's vulnerabilities, while a good risk manager will attempt to identify the possible market scenarios that can cause the greatest losses for the bank, regardless of probability. Bank management must then decide whether the potential losses identified through stress testing are acceptable. The participation of risk control personnel provides independent oversight and an objective viewpoint to assure the integrity of the process.
Examiners should determine that banks have appropriately considered the following factors in their stress testing. While these factors are described in some detail in existing guidance, we elaborate on each here to reinforce those key lessons reaffirmed by recent events.
Examiners should evaluate how managers use stress tests to understand and control the bank's downside risks. To demonstrate effective oversight, senior management should review regularly the results of stress testing scenarios, including their major assumptions. Senior management should present the results of stress tests to the board of directors as results and/or market circumstances dictate. Effective stress testing will probe for those scenarios that are likely to create the largest portfolio losses. Banks should consider placing either hard limits or management action triggers on the size of losses they will tolerate. Depending on the potential losses revealed by the stress tests and the anticipated likelihood of occurrence, management should consider appropriate risk-reducing actions. Senior management should review and approve decisions not to curtail risk when stress tests identify material exposures.
Credit Risk Management
The recent credit-related losses experienced by some banks arising from trading and dealer activities, in part, reflected management of such exposures in a manner inconsistent with a bank's overall credit standards. For example, some banks viewed the ability to sell cash market instruments as the primary source of repayment in lieu of a formal assessment of issuer capacity. Also, some hedge fund counterparties did not provide, and banks sometimes did not demand, basic financial information necessary to make a fully informed credit decision. With hedge funds, banks often based credit decisions on the reputation and perceived risk management capabilities of the counterparty as well as largely qualitative assessments of the fund's risks. In some cases, the lenders seemingly were too timid or embarrassed to impose the same kinds of collateral and financial disclosure requirements as would normally be pursued with other less "sophisticated" customers. These shortcomings compromised the accuracy and rigor of subsequent parts of the credit process, including the setting of limits and the collateral and margin arrangements required. Recent experiences confirm that there is no substitute for the establishment and implementation of a strong firm-wide credit risk management culture.
Some financial institutions adjusted various elements of standard International Swaps and Derivatives Association, Inc. (ISDA) documentation to liberalize credit terms. For example, some institutions accepted less conservative provisions to certain elements of close-out clauses, such as liberalizing the limits on deterioration in net asset values. These more lenient credit terms limited their ability to take risk-reducing actions when they identified a deterioration in the financial condition of some hedge funds.
While the number of materially compromised credit relationships were not significant at any given bank, several that resulted in losses during the turbulence were either very large in an absolute sense or resulted from exposures to highly visible obligors. As a result, the consequences of the decisions to stray from normal credit administration processes were adverse from both a financial and reputation perspective.
Examiners should ensure that when banks extend credit, either through owning specific issuer obligations in cash markets or to trading counterparties in OTC transactions, they obtain sufficiently comprehensive financial and other information to provide a clear understanding of the obligor's risk profile. A sound credit approval process should contain the following elements, calibrated to the size of the obligor and the nature of its activities:
Examiners should also determine that banks appropriately translate their risk tolerance levels into effective policies and procedures that deal both with individual as well as important classes of obligors. These policies not competitive pressures in the marketplace need to drive the credit decision process. When bank management identifies credit concerns with regard to an obligor, it should take appropriate steps to limit and manage the exposure consistent with the bank's overall underwriting standards and risk appetite. For example, banks should either deny credit or implement tougher credit conditions for those obligors who provide less than complete information about their risk profile or where the total extent of the fund's vulnerability may not be known. Appropriate compensating actions could include the elimination or reduction of loss thresholds, requiring initial and variation margins, limiting the range of assets acceptable as collateral, and more conservative financial covenants including limitations on the counterparty's leverage and/or other financial transactions.
Collateral can be an important and effective vehicle for a bank to control the risk of a credit relationship. However, banks should not compromise key elements of their credit underwriting process simply because the obligor has promised to collateralize the current replacement cost of credit exposures resulting from trading activities. For example, banks should not solely depend upon collateral to compensate for inadequate financial information, nor should banks allow the perceived comfort of collateral to allow for less frequent monitoring of an obligor's financial performance. Collateral can reduce credit risk but at the expense of increased transaction and compliance risks.
The collateral itself poses market risks. Banks must understand the correlation between the value of the collateral and the value of the underlying transaction. Credit exposures increase when the current mark-to-market of a position becomes more favorable for the bank while the collateral value falls. Even with effective systems for daily collateralization of mark-to-market exposure, banks can face potentially significant unsecured exposure in their trading and derivatives activities. Such exposures can occur because of the existence of loss thresholds or when disorderly market conditions create uncertainty regarding both collateral values and contractual exposures.
Examiners should affirm that banks have in place clearly articulated policies for the establishment of collateral arrangements with counterparties. Policies should lay out clear guidelines for the type of collateral arrangements required, based on criteria such as the rating assigned to the counterparty, the quality of information available, and the nature, volatility and liquidity of the transactions. In particular, banks need to have clear internal procedures to ensure control over collateral, and guidelines detailing the types of acceptable collateral and their respective haircuts, as well as under what conditions they will impose initial and variation margin requirements, in order to cover the potential future exposure of trading transactions. Finally, the granting of two-way collateral arrangements, and any re-hypothecation rights given to the counterparty, should be a function of the obligor's credit quality and the bank's own liquidity position. Examiners should ensure that audit and other oversight units regularly evaluate the adequacy of the collateral management function as well as test compliance with established policies and procedures.
Recent events underscore the interconnection between market and credit risks. Risk interconnections are particularly important when dealing offshore. Foreign exchange rates, local bank creditworthiness, the legal environment, and political stability can all move against the bank at once, resulting in large and more problematic risk exposure than originally envisioned.
Similar to price risk management, when managing the risk in individual credit transactions and broader credit portfolios, banks must consider the correlations between market exposures and obligor credit quality. For example, preceding the Russian debt crisis, some banks entered into Russian ruble forward contracts with Russian banks. When the ruble devalued, some of these banks, or other counterparties whose credit quality was linked to the Russian economy, failed and defaulted on their contracts, creating un-hedged exposures and new market risks for their counterparties.
During periods of market turbulence over the past several years, we have observed several instances where counterparties had significant out-of-the-money positions on derivatives contracts (dealer banks were in-the-money). Some of these counterparties pursued legal or public relations efforts to void contracts or seek partial forgiveness on their obligations claiming that the transaction was either inappropriate or that their dealer provided insufficient disclosure about the transaction's risks. Therefore, examiners should ensure that bank management recognizes that there can be a strong linkage between market, credit and compliance risks through the implementation of a careful relationship selection and approval process in addition to the more fundamental price and credit risk processes.
Currently, many banks' procedures for stress testing their credit exposures are significantly less developed than for price risk exposures. Recent events demonstrate that credit exposures, particularly those arising under derivatives transactions, can change rapidly as market volatility increases. Even a reasonably well-margined exposure can quickly become partially unsecured when there are extreme price movements. A vivid example of this risk occurred in derivatives transactions involving counterparties who had short positions in the U.S. dollar and long positions in Asian currencies that experienced devaluations.
Banks need to stress test their counterparty credit exposures to identify individual counterparties, or groups of counterparties, with positions that are particularly vulnerable to extreme or one-way directional market movements. Examiners should determine that banks have effectively stress tested exposures to identify risk issues, such as concentrations in collateral, that warrant further investigation and potential risk reducing actions. Examiners should also evaluate whether banks have appropriately considered the use of discretely selected (rather than historic) volatilities to generate a more meaningful understanding of potentially adverse risk exposures. Substituting selected for historic volatilities can help measure exposure to situations vulnerable to abrupt change; for example, where contract values are linked to a managed currency.
Banks should establish credit limits based on their risk tolerance, taking into consideration the terms and conditions of the credit agreement and the counterparty's financial capacity to perform on its obligations.
Pre-settlement risk (PSR) exposure is the amount by which exposure might grow in the time it takes to terminate a contract and liquidate existing collateral. Therefore, the exposures from collateralized transactions, and the sufficiency of collateral, are determined using measures that reflect the time that it takes to terminate and liquidate as well as the volatility of the transaction value across that time horizon. Such measures should recognize that the stressed market conditions may also be the conditions under which the market is least liquid.
Some institutions set limits based on PSR measures that were not meaningful and that were, therefore, ignored. In those institutions, PSR was calculated by setting the time horizon for the calculation to the full maturity of the contract. The output derived from such a calculation was recognized as conservative and ignored.
Examiners should ensure that a bank's risk measures and limit-setting processes recognize that the actual credit exposure in collateralized relationships is a function both of the volatility of the underlying positions and of the collateral securing the relationship. Examiners should determine that banks with collateralized exposures use supplemental measures calculated over an appropriate time period that reflect stressed market conditions in the collateral maintenance and exposure monitoring processes. The horizon should reflect the time it would take for the bank to terminate the contract and liquidate existing collateral once it recognizes that the counterparty has missed a collateral call the period of exposure.
Transaction Risk Management
Unconfirmed trades are a regular part of trading activities, but banks must properly manage them to control risk. Unconfirmed trades have two major risks: 1) the enforceability of the trade, and 2) transaction errors going undetected in the records of the bank, both of which can lead to unanticipated market risks.
Some institutions also tolerate a backlog of unsigned master agreements. Master agreements reduce counterparty credit exposure through such provisions as netting, and thus represent an effective means of controlling credit risks. Banks should establish a limit on the size and number of transactions with counterparties from whom the bank has not obtained a master agreement. These transactions should be approved through the bank's credit approval process before consummation. Transaction confirmations should contain language that protect the bank in case of default by the counterparty. Commonly, the confirmation would contain the "netting" language from the master agreement. Finally, banks should make future transactions subject to the completion of a master agreement.
Examiners should evaluate bank procedures for addressing unconfirmed trades and unsigned master agreements to ensure both the operational integrity of the trading function and management's commitment to control business risks.
The increasing worldwide competition of the derivatives business has led to lower profit margins. Banks have responded by developing more highly structured and sometimes less liquid transactions. The less liquid the transaction, or components of a transaction, the more likely the back office will not have access to an independent source to value the position. In such cases, banks use internally derived estimates of key pricing parameters (e.g., volatility) to generate both inception profits and daily profit and loss data. Reliance upon internally generated valuations exposes the bank to the prospect of surprise write-downs, including the possible restatement of prior earnings figures, if actual market prices ultimately deviate from these estimates. The types of transactions that necessitate internally generated valuations may generate profit margins that create incentives for an even greater volume of such trades.
Examiners should ensure that bank policies and operating procedures consistently and effectively compensate for the inherent lack of independence in the valuation of some trading positions. Appropriate controls include a firm-wide policy governing the methodology to be used for valuing illiquid positions, the regular identification and reporting of these positions to facilitate senior management awareness and assessment of the reasonableness of the valuation methods used, more frequent audits of such positions, and the deferment of compensation to traders until such time that comfort is gained that the valuations are sound. Depending on the nature and scope of a bank's operations, examiners should determine that policies address curtailment programs to control the risk from less liquid securities positions that become stale. Examiners should also ensure that banks establish and maintain appropriate reserves to compensate for downside valuation adjustments.
Compliance Risk Management
During times of market volatility, counterparties in out-of-the money positions are more likely to dispute the enforceability of contracts. For example, some counterparties (with deep out-of-the money contracts) have disputed in the courts the enforceability of certain OTC contracts with dealer banks. Counterparty selection thus remains a critically important issue for banks, which must satisfy themselves that the counterparty has both the legal capacity to enter into a transaction, and the willingness to perform. The soundness of legal systems and validity of internal opinions may also become more critical during times of stress. Recently, a Russian court ruled that non-deliverable forward contracts were gaming contracts and thus unenforceable. While this decision was overturned on appeal, it serves as a reminder of the necessity to ensure that contracts are enforceable under local laws.
Because the enforceability of many OTC derivative contracts has not been tested in the courts in all jurisdictions, examiners should ensure that banks employ competent legal counsel to review applicable documents before executing transactions, and periodically thereafter. Counsel should be familiar with the economic substance of the transaction, the laws of the jurisdictions in which the parties reside, and laws governing the market in which the instrument was traded. Whenever a bank does not use standardized documents or makes changes to standardized contracts, examiners should determine that bank counsel reviews the documents and/or changes for propriety. When the legal enforceability of netting arrangements is not certain, examiners should also ensure that the bank avoids understating credit exposures by evaluating risk on a gross basis.
Corporate Risk Oversight
Recent events also have highlighted weaknesses in some banks' internal oversight control functions such as market risk control, loan review and audit. Ideally, risk oversight units provide an early warning function for senior management by pointing out negative market trends, undesirable risk/reward relationships, and weaknesses in internal procedures. With such warnings, a bank could scale back activities in certain markets or implement mitigating controls and thereby minimize their vulnerability to surprise adverse performance. In contrast, some institutions did not promptly address policy exceptions identified by risk control units, including deficiencies in the administration of some prominent credits. In some cases, the failure of senior management to take proactive measures was, in part, due to the lack of organizational stature of the risk control units.
A critical difference between the better managed institutions and the rest of the industry lies in the firm-wide culture. One constructive development at the top institutions has been the implementation of self assessment programs where line managers are rewarded for promptly identifying and remedying gaps in the controls surrounding their own businesses. Conversely, penalties are assessed if previously unidentified deficiencies are uncovered by risk control units or bank examiners. Another sound practice has been the increasing use of "guest auditors," where line employees with defined expertise are borrowed to facilitate audits or loan review targets. Such programs have the effect of immediately raising the expertise and stature of the oversight team, and result in a more efficient and effective audit.
Examiners should ensure that senior management is appropriately responsive to oversight units. Situations where repeat deficiencies are identified and reported, but not corrected on a timely basis, should be the focus of examiner discussions with executive management and boards.
Front, middle and back office personnel, as well as oversight units should be sufficiently staffed and trained to accommodate the challenges associated with new lines of business. In particular, trading support functions, such as market and credit risk control units, risk measurement, operations, finance and audit need to keep pace with changes in business risks. Because most support functions do not generate revenue, management may overlook their importance during periods when such businesses generate strong revenues. Even in normal market environments, banks often understaff these functions. As recent events confirm, stressed market environments will highlight breakdowns in operations, such as failure to make collateral calls, trade input errors, and inaccurate valuations all of which can create significant financial losses and damage to the bank's reputation.
Examiners should ensure that bank management's considerations prior to entering new businesses includes comprehensive assessments of both revenues and the expenses required to manage the risks properly. It is critical that such assessments and subsequent resource allocations provide for adequate support staff to complement revenue generation units. Examiners should also determine that product and market liquidity is realistically assessed and understood at inception and on an ongoing basis. Such assessments should recognize that although liquidity might initially be abundant, it may quickly evaporate. Accordingly, examiners should determine that risk management processes include appropriate early warning triggers such as diminished turnover, lack of price quotes and widening of spreads.
An ineffective process for assessing novel and potentially more volatile investments could expose the bank to costly litigation, account terminations, and loss of potential new funds/clients. Bank investment advisory and fiduciary units must determine both whether a proposed hedge fund investment is permissible under the governing instrument and applicable federal/state laws, and is an appropriate vehicle for the client. Additionally, these units should consult with legal counsel about appropriate client disclosures. Prior to purchasing or recommending such investments, investment advisors and fiduciaries should consider:
An effective ongoing monitoring process should be a continuation of the initial analysis. Fiduciary account managers and investment advisors should continually monitor the investment strategy and periodically reassess whether the hedge fund continues to be appropriate for the individual client account. The standard risk measurements reported may need to be supplemented for some hedge fund strategies. For example, because VaR or volatility estimates are generally predicated on assumptions about market relationships that may not hold in practice, stress tests are used to assess the sensitivity of the measurements to the assumptions.
Examiners should determine that appropriate due diligence, ongoing monitoring, and client disclosures are integrated into the investment management process. Examiners should also ensure that the fundamental risk management principles outlined in OCC Bulletin 96-25, Fiduciary Risk Management of Derivatives and Mortgage-backed Securities, are employed with respect to investment products being offered to clients.
Questions or concerns regarding this supplemental guidance should be directed to the Treasury and Market Risk Division, (202) 649-6360.
Michael L. Brosnan
*References in this guidance to national banks or banks generally should be read to include federal savings associations (FSA). If statutes, regulations, or other OCC guidance is referenced herein, please consult those sources to determine applicability to FSAs. If you have questions about how to apply this guidance, please contact your OCC supervisory office.