By Andrew Cornford∗, SUNS # 6480, 23 May 2008
Unforeseen shifts in concentrations of banking risks have been at the centre of the current credit crisis. Such concentrations are now a major item on the regulatory agenda and the subject of a recent paper (‘Cross-sectoral review of group-wide identification and management of risk concentrations’) issued in April by the Basel-based Joint Forum on Financial Conglomerates.
The Joint Forum paper draws on two surveys, one of the views of 15 supervisory bodies from 10 countries and the other of approaches to identifying and managing risk concentrations in 18 financial conglomerates. Its publication coincides with that of two other reports covering actual risk management practices in some financial conglomerates during the credit crisis. ‘Observations on Risk Management Practices during the Recent Market Turbulence’ [March 2008] is a survey conducted by a Senior Supervisors Group from France, Germany, Switzerland, the United Kingdom and the United States of eleven of the largest banking and securities firms. The second, ‘Shareholder Report on UBS’s Write-Downs’ [April 2008] is a review of the key findings of a report by the Swiss bank, UBS, to the Swiss federal Banking Commission on the reasons for its huge losses due to structuring, trading and investment activities, in particular, those due to securities with a value which referenced United States sub-prime mortgages.
The establishment of the Joint Forum in 1996 was the outgrowth of work of the Basel Committee on Banking Supervision and of the International Organization of Securities Commissions (IOSCO) in 1992 on principles for the supervision of financial conglomerates. This work led to the establishment of a working group called the Tripartite Group of Banking, Insurance and Securities Regulators, which was reconstituted as the Joint Forum in 1996 and now comprises the International Association of Insurance Supervisors (IAIS) as well as the Basel Committee and IOSCO. The membership of the Joint Forum is limited to banking, insurance and securities supervisors from 13 industrial countries.
The risks which the Joint Forum is concerned with are those which arise from the mixing of banking, securities and insurance activities in a single business structure. Such mixing creates new connections between the risks of the separate activities. The connections are a potential source of contagion within the financial firms themselves as well as, more broadly, throughout the financial sector. Moreover, the mixing poses new problems for management and supervisors in areas such as conflicts of interest between different parties participating in or affected by the firm’s activities, internal transparency and financial reporting, and setting levels of capital appropriate to the different activities of the conglomerate as well as for the firm overall.
For the purpose of the work of the Joint Forum, a financial conglomerate refers to a group of companies under common control whose activities consist of the provision of services in at least two of the sectors, banking, securities and insurance. Such conglomerates have a history which long antedates the 1990s. In the United Kingdom, for example, in the 1960s, large commercial banks established subsidiaries to provide hire-purchase and other specialized financial services. In continental Europe, conglomeration often took place through internal expansion of the operations of universal banks and through an increase in the number of Bancassurance and Bancaffianz groups which provide both banking and insurance services. In the United States, financial conglomeration received a huge fillip from the 1999 Gram-Leach-Bliley Act which ended the comparmentalisation of banking embodied in legislation of the 1930s (the Glass-Steagall Act) and opened up new possibilities for the integration of banking, insurance and securities business.
The initial focus of the Joint Forum was the internal organization and management of financial conglomerates, and its recommendations responded to concerns regarding the internal transfer of risks which might threaten the stability of the more sensitive parts of conglomerates such as their banking activities. The April 2008 paper was originally intended to be a sequel to earlier technical work on the management and supervision of concentration risk. However, drafting was overtaken by events beginning in summer 2007, and in consequence, the discussion covers features of concentration risk highlighted by the credit crisis and connections between concentration risk and other key financial risks such as liquidity and market risk.
For the purpose of the Joint Forum paper, risk concentrations are defined as exposures with the potential to produce losses large enough to threaten a financial conglomerate. The concentrations may involve the conglomerate’s assets, liabilities, off-balance-sheet positions, and the execution and processing of transactions.
These definitions make it possible for the Joint Forum to maintain in the paper its customary primarily microeconomic focus on internal risk management and firm-level supervision. However, the transmission of risks between financial conglomerates and the broader markets in which they operate nonetheless intrudes owing to the need to include principles for the management of liquidity risk – a subject which necessarily reflects conditions in these markets.
Such definitions also in principle exclude traditional concentration risk in the form of external exposures to a single firm, sector or set of related economic activities. Here too, however, recent events intrude, and the paper itself acknowledges certain concentration risks that arise owing to developments external to the conglomerate in the markets in which it operates.
Nevertheless, the definitions chosen by the Joint Forum mean that the paper does not address the concentration risks which are typically a major concern of regulators of commercial banks and financial conglomerates in emerging-market countries. In Asia, concentration risk is often part of the same set of problems as related-party lending. Identification of risk concentrations due to exposures of financial firms to particular groups can be handicapped by use of multiple legal entities by borrowers. This can take a bewildering number of forms such as lending by a bank to an apparently independent company or individual that on-lends the funds to a related company or arranging for another bank to lend to a related company against a guarantee not recorded in the books of the originating bank.
Moreover, risks due the denomination of the assets and liabilities of banks and their borrowers in foreign currencies, which have proved of great importance in combined banking and currency crises in emerging markets, are also not covered in the Joint Forum paper. Owing to the effect of depreciation of the currency and deflation on the ability of borrowers to meet obligations denominated in foreign currencies, these crises often bring into the open hitherto concealed concentration risks involving financial and non-financial firms and sectors.
The focus of the Joint Forum paper reflects its membership and mandates. This does not imply that the contents of the paper do not merit wider attention. With the development of their own financial markets, financial firms and their supervisors in emerging-market and other developing countries are increasingly likely to confront issues related to financial conglomeration similar to those exercising their counterparts in developed countries. Moreover, subjects raised in the paper are relevant to understanding international financial markets to which, for good and ill, most countries are now exposed in varying degrees.
The first section of the Joint Forum paper reviews traditional approaches to the management of concentration risks. The following section focuses on the increasingly active management of concentration risks in the context of the rapid growth of the market in financial instruments which transfer risks between different parties. This is followed by a discussion of the challenges posed to the corporate governance of financial conglomerates by more integrated approaches to risk management. The next section of the paper takes up a number of technical issues related to the identification and management of concentration risks. These include the measurement of the capital required for concentration risks, and stress testing and scenario analysis. An annex takes up selected features of the regulatory regimes in major industrial countries.
The survey of industry practice discussed in the first section of the paper indicated that risk identification and management in financial conglomerates still tends to be managed within silos corresponding to major risk categories. Under a silo-based approach, personnel, processes and systems are grouped by risk categories such as credit, insurance, liquidity and market risk. Credit risk concentrations are controlled through risk limits on exposures to particular names, economic sectors and sub-sectors, geographic regions and countries, and products (such as real-estate or consumer lending). Such an approach is not conducive to integrated risk management.
The survey also found that securitisation exposures – so important in the current credit crisis – are usually grouped by tranches corresponding to different levels of risk and return for investors. This approach attributes a key role to the credit ratings of different tranches. The Joint Forum’s survey found that only a few conglomerates use a “look through” approach to securitisation exposures which entails assessment of the credit risk of the assets underlying the tranches of the securitisation.
A non-integrated, silo approach to risk management can be particularly problematic with respect to liquidity risk which reflects developments at the interface between the conglomerate and the markets in which it operates. Financial firms encounter three types of liquidity risk: (1) funding mismatch risk, that is the risk that the firm will not have sufficient funds to meet its normal obligations; (2) market liquidity risk, that is the risk that the firm will not be able to convert assets to cash or to access financing on reasonable terms (owing to a lack of buyers or uncertainty over valuation); and (3) contingent liquidity risk which results from difficulties in meeting obligations due to firm-specific or market-wide unexpected events.
All three types of liquidity risk have been pervasive during the credit crisis. Under contingent liquidity risk, banks have had to reconcentrate on their balance sheets liquidity and credit risks from off-balance-sheets entities which the banks originally created or sponsored but for which funding from investors subsequently ceased to be available. The verdict of the Joint Forum is that, in comparison with other risk categories, liquidity risk tends to be less well integrated into systems of conglomerate-wide risk management in spite of its special importance in stressful market conditions.
These conclusions are reinforced by findings in the reports of the Senior Supervisors Group and of the UBS. The Senior Supervisors found that the firms which had been more successful in managing their problems were those which “aligned treasury functions more closely with risk management processes, incorporating information from all businesses in global liquidity planning, including actual and contingent liquidity risk”.
The Senior Supervisors also viewed the disappearance of liquidity as a major cause of the huge losses incurred by some financial firms on their holdings of Super Senior tranches of Collateralized Debt Obligations (securities backed by portfolios of loans, bonds and other assets including mortgages). These tranches were at the very top of the securitisation structures of CDO tranches, not only carrying the highest possible rating from credit rating agencies but also classified as even safer than other tranches with such ratings.
Owing to their low capital requirements and thus low financing costs, Super Senior tranches were an investment favoured by banks for their own portfolios. However, the result of banks’ own investments in these tranches was a heavy concentration of holdings of Super Senior tranches in a single category of institution. During the flight to liquidity in the late summer of 2007, when banks wanted to sell these assets, there were few buyers so that the banks had to take huge losses. The conclusion of UBS, for which 75 per cent of the losses of its CDO desk (or 50 per cent of the bank’s total losses) up to December 2007 were due to Super Senior positions, is “that there does not appear to have been a liquid secondary market and that the business tended to retain the Super Senior tranche”.
As the Joint Forum paper notes in the second section of the paper, in recent years, there has been an enormous growth in the markets for instruments transferring risk. The gross notional value of all positions in over-the-counter (OTC) derivatives (i. e. derivatives not traded on organized exchanges) exceeded $500 trillion in mid-2007. Of this total, more than $40 trillion consisted of credit default swaps, a derivative which transfers credit risk. Financial conglomerates participate in the markets for risk transfer in various ways and for various purposes: to fund assets which they originate; to manage the credit, market and insurance risks of positions retained on their balance sheets; to invest in securitised products; and to generate revenues from trading on their own account as well as from distribution to their customers.
However, according to the Joint Forum, “Often, the process of so-called risk transfer more closely resembles a transformation of a variety of risks into credit exposures to the counter-parties to which they were transferred”.
Moreover, the complexity of the products available through these markets has generated new risk management challenges. Some of these are due to features of the products themselves. Among these are increased systematic risk (i. e. risk which cannot be diversified away) due to the way the assets backing CDOs are pooled, the opacity of the products, and additional leverage which increases the rapidity of the changes in value of a product or portfolio in response to changes in market, credit and liquidity risk.
Moreover, the challenges to integrated risk management are the more formidable, the broader the conglomerate’s participation in different risk-transfer activities. Active risk management or other forms of involvement in risk-transfer activities through the markets for risk-transfer products leads to additional exposures to liquidity risk since the effectiveness of the products for both hedging and income generation depend on their marketability and their value as collateral for financing.
In this context, the report of the Senior Supervisors Group draws attention to the role in the credit crisis of innovative products which “had been created during the period of more benign market conditions” so that “banks and securities firms had not observed how such products would behave during a significant market downturn and found their risk management practices tested to various degrees”.
UBS’s criticisms of its own processes include the absence of a comprehensive view of its exposures to sub-prime mortgages in different parts of the UBS Investment Bank. This view was required for the setting of overall portfolio limits to complement transaction-by-transaction approval for CDO positions and related hedges. Such a view would be an essential part of integrated risk management. However, UBS’s priorities were elsewhere: “the emphasis was generally on speeding up approvals as opposed to ensuring that the process achieved the goal of delivering substantive and holistic assessment of the proposals presented”.
Under the features of good corporate governance required for integrated risk management, the paper returns to subjects also covered in an earlier 2003 report of the Joint Forum. Failings regarding corporate governance, the Joint Forum acknowledges, are still common and are exemplified in the UBS report. In addition to the absence of holistic assessment just mentioned, the UBS report also notes that members of [Investment Bank] Senior Management “did not sufficiently challenge each other in relation to the development of their various businesses”.
Under corporate governance, the focus of the Joint Forum is on organizational structures. Missing from the paper is discussion of the way in which in a financial conglomerate, even a well designed organogram, does not necessarily ensure that senior decision takers have at their disposal the information required for successful risk management. Multiple organizational layers put distance between data gatherers and users in senior management, and information gets lost in the summarizing of data as they are transmitted up the organizational chain.
Failures on this front in major financial firms during the credit crisis raise questions about the still-widely-held assumptions as to the optimum size of financial conglomerates. These questions include how far the drive to expand through the establishment or purchase of new entities in both home markets and across borders, which can still be observed among major international banks, is compatible with the maintenance of effective internal information transmission and controls, even allowing for the aid furnished by the latest technology.
As the complexity of financial transactions and organizations has grown, regulation and supervision have come to rely increasingly on the vetting of systems of internal control of the financial firms. At the same time, an increasingly important part of such systems are procedures which depend on computer simulation. Two of these procedures covered in the Joint Forum paper are the estimation of economic capital and stress testing and scenario analysis.
Economic capital is the amount which a financial firm believes is necessary to absorb potential losses due to major categories of financial risk. Stress testing is a tool used to evaluate through computer simulation the potential impact of an event or movement in a set of financial indicators.
The appeal of economic capital is the common measure of risk which it imposes across different businesses and risk categories. On the other hand, the models used have the limitation that they often fail to take account of indirect and second-order effects. For example, they may not capture the indirect credit exposure of a bank to the collateral of a loan as well as the direct exposure to the borrower, thus missing the potential risk concentration due to the joint exposure.
Moreover, the estimates of risk correlation used in the models are often subject to shortcomings. Partly, these are due to insufficient data, especially for periods characterized by disturbances. But the shortcomings also reflect the way in which risks due to different factors are aggregated. For example, in the case of a loan backed by collateral, a simple correlation between market and credit risks may take account of the effect of an increase in interest rates on the likelihood of default but not of the negative effect of the higher interest rates on the value of the assets serving as collateral which are seized by the bank.
Stress testing and scenario analysis have advantages of flexibility and transparency over models of economic capital. The two techniques can be used to investigate the potential effects of a wide range of risks specified in ways more susceptible to intuitive understanding. These include business risks, reputational risks, legal risks, country/transfer risks, and various other event risks. However, stress testing and scenario analysis are subject to the limitations of the expert judgment used to select them. Moreover, there is a lack of widely recognized procedures for stress tests as a tool for measuring risk concentrations and liquidity risks, especially those which manifest themselves as part of scenarios involving other risks with potentially market-wide effects.
The points raised under stress testing and scenario analysis by the Joint Forum are reinforced by the findings of the Senior Supervisors Group: “Some [firms] found that their stress tests or scenario analyses generally matched the movements in markets, but others found that the actual shocks to credit spreads [and thus to asset prices] tended to be wider and longer lasting than their prior analyses had suggested”. In particular, high credit ratings proved to be poor proxies for low price volatility.
The Joint Forum’s discussion of economic capital and of stress testing and scenario analysis indicate that, useful though they can be, they are neither magic bullets nor quick fixes for the problems for internal control and supervision posed by increasingly complex financial firms.
As shown by the annex of the Joint Forum paper, regulatory control of concentration risk in major industrial countries consists mainly of limits on exposures to particular borrowers or other counter-parties.
In the European Union, rules for large exposures are included in the Capital Requirements Directive (which implements Basel 2) and set limits to such exposures in relation to the capital of financial firms. There are also rules limiting large exposures in relation to capital in Japan and Canada. Rules in the United States reflect the multi-tier character of its regulatory regime with different regulators at both federal and state levels. They include not only limits on large or concentrated exposures in relation to capital but also ceilings on loans in relation to value for exposures to residential and commercial property. The qualitative dimension of risk concentrations in these countries are addressed in rules to be followed by the supervisors of financial firms.
Reflecting the Joint Forum’s consultative character, the paper does not attempt to provide a blueprint for regulatory reform. Each section concludes with Considerations designed to flag the principal points of the preceding diagnostics. The results of the paper’s recommendations will, or alternatively will not, eventually be reflected in new regulation at the national level and in improved risk management in financial conglomerates themselves.
By and large, the trend to financial conglomeration is not as advanced in emerging-market and other developing countries as in the member countries of the Joint Forum. The paper’s value lies in the attention it draws to problems to which such conglomeration can give rise. But there is no reason why the authorities in developing countries should accept the view, implicit or explicit, in much of the literature on the subject that financial conglomeration according to models found in industrial countries is an inevitable feature of financial development.
Both the institutional structures of financial conglomeration and the complex transactions which usually accompany it are the subject of regulatory permission for domestic firms and of procedures for granting market access to foreign financial firms. The latitude for conglomeration of different financial activities in a single firm should reflect decisions by the authorities based on national priorities. Historically, compartmentalization and specialisation were long features of the financial sectors of several industrial countries.
As for any presumption that the complex financial products typically associated with financial conglomeration necessarily bring wider economic benefits, the last word can be left to Paul Volcker, former chairman of the United States Federal System who, in a speech to the Economic Club of New York in April 2008, noted of the move from “a commercial bank centred, highly regulated financial system to an enormously more complicated and highly engineered system” that “It is hard to argue that the new system has brought exceptional benefits to the economy generally”.
This is a contrarian observation coming from this source and one which should be borne in mind by those responsible for financial-system design in emerging-market and other developing countries.
*Andrew Cornford was formerly a senior UNCTAD economist and is currently Research Fellow at the Financial Markets Center.