Revised 15 November 2000
Risk Management and Financial Globalisation
José Braz
TEcFinance, Lisbon
Jorge Braga de Macedo
OECD Development Centre, Paris
With financial globalisation, capital markets become increasingly integrated and the volumes of capital flows increase out of all proportion to the growth in trade. In such an environment, basic notions of risk management such as leveraging and volatility, would lead the market to increase its expectation of the probability of a financial crisis. Indeed, the underlying causes of the 1997-99 crises had been recognised for some time: unhedged foreign currency borrowing, excessive short term external debt and growing bubbles in stock and property prices. What went far beyond the expected, however, was the magnitude of the crisis and the speed with which contagion spread across countries in the same region and to other regions of the globe.
The credibility of the monetary authorities' commitment to defend a given exchange rate position often hinges on the central bank's vulnerability to an attack on its solvency. As frequently happens in a crisis of confidence, decisions by economic agents are made on the basis of perceptions rather than firm information. Government authorities contribute to this by being ambiguous or secretive with the facts.
In particular, the extent of deposit guarantees was left undefined in some cases and, as had happened in many other crises, some central banks were thought to be operating in futures markets and other off-balance sheet areas without being explicit. As a result, investors and creditors, both domestic and foreigners, participate in a frenzied scramble to recover their investments, frequently generating situations of panic which exacerbated the overshooting of variables such as exchange and interest rates. All parties could be better off if their decisions were based on facts rather than perceptions and many analysts and institutions have called for greater transparency of relevant information.
The central issue in the crisis was that of unsound finance, which translated into national balance sheet vulnerability. Balance sheet crises have far more leverage - both in collapsing a country’s financial structure and economy and in spreading contamination to other countries. The core problem is that of a mismatch of maturities and of a mismatch of currency denominations, issues which have long being central to asset-liability management in commercial financial institutions. As Rudi Dornbusch (1998) put it, "the right answer for crisis avoidance is controlling risk".
Controlling risk begins with measurement, and over the last several years, regulators have encouraged financial institutions to use portfolio theory to produce dynamic measures of risk. Because the Value-at-Risk (VaR) of a portfolio is less than the sum of its components, measuring this VaR becomes a crucial management tool within large financial institutions.
In line with the recommendations of the Bank of International Settlements (BIS), VaR methods have become a benchmark for risk measurement. VaR was first developed as a measure of market risk, describing the potential loss incurred by unfavourable market conditions within the considered time horizon and confidence level.
As any product of portfolio theory, it is built for tranquil times and involves judgement when systemic risk is present, as was the case between the August l998 Russian crisis and the near collapse of Long Term Capital Management. Consequently, the method helps short run profit and loss-risk exposure measurement, but it is not without dangers, as described in Helmut Reisen (2000). Meanwhile, the application of VaR has been extended to other risk types, including credit risk.
One of us has argued that the same technological progress that helped to make the crisis more intense, namely sophisticated financial instruments and electronic transmission of information and funds, can also help to reduce the probability and scale of future crisis. VaR analysis, in particular, can be extended to provide a synthetic measure of the vulnerability of central bank accounts. This information, appropriately standardised and provided on a regular basis, can help financial markets to take decisions on capital flows on the basis of better quality information and can help policy makers decide between policy choices which take into account their impact on the risk-adjusted balance sheet of the country.
With the continuous decline in computational cost, there has been a rapid development of computer-intensive risk management technologies, namely in the areas of market risk and credit risk. It is already common practice for commercial banks to measure the exposure of their portfolios to changing market conditions, such as movements in interest rates, yield curves, exchange rates and the prices of stocks, bonds and commodities, as well as the derivative instruments based on these prices. The most advanced techniques are those that use full valuation historical or stochastic simulation to calculate the VaR.
International standards for bank supervision and regulation have recognised and, in some cases, provided incentives for banks to develop internal models for calculating risk-adjusted capital requirements. Parameters for VaR estimation by internal risk measurement models of commercial banks were established by the BIS in 1996, with respect to market risk. More recently, models evolved to include credit risk, such as counterparty risk in swap contracts; further development is under way to allow all bank operations to be included in a bank-wide VaR measure.
Unlike traditional accounting measures of balance sheet items, which look at historical costs and accumulated depreciation, VaR analysis is based on market values as they evolve over time. Also, it looks at all operations, including contingent liabilities and off-balance-sheet items. It is a measure of the vulnerability of the balance sheet to movements in the prices of its components. Since financial crises are precipitated by increased vulnerability, it seems natural to propose the use of the same technology to analyse the balance sheet of the national financial authority or central bank.
As indicated by Mario Blejer and Liliana Schumacher (2000), central banks may engage in derivative operations to provide additionality to incomplete or illiquid markets; to defend a fixed exchange rate regime or an exchange rate band. They may do it to alleviate the conflict between the defence of an exchange rate regime and the stability of the financial system. Derivatives and other contingent liabilities are also used as an automatic stabiliser of a foreign exchange market or as an alternative instrument for monetary management under specific circumstances.
Examples of such interventions include the Bank of Spain operations in the options market during the 1992/3 ERM crisis, Mexico’s stabilisation scheme of August 1996 and the Bank of Thailand sale of forward contracts in 1997. The recent use of foreign exchange swaps by the Reserve Bank of Australia to increase liquidity to deal with the Y2K problem provides another example. More traditional and widespread examples of central bank contingent liabilities include the provision of credit guarantees and of deposit insurance (in cases where there is no autonomous deposit insurance fund).
All standard assets in the central bank portfolio (foreign currency reserves, bonds, gold) can be included in the VaR measure with technology currently in extensive use by commercial banks. Additionally, policy risk can be added in the case of central banks; financing of budget deficits, for example, will have a negative net effect inasmuch as liabilities (base money) increase and the corresponding asset (loan to the treasury, assumed not to be repaid) has an economic value of zero.
The economic valuation of off-balance-sheet contingent positions permits the aggregation of all central bank transactions (or, rather, all financial authority transactions, including treasury positions where relevant). While central bank accounts always balance in accounting terms, this balance ignores the economic valuation of contingent commitments, explicit or implicit, such as forward contracts and deposit insurance guarantees (unless there is a fully funded deposit guarantee scheme). Also, assets are registered at their nominal values, not their economic valuation, which would take into account risks such as repayment probabilities of certain loans (treasury loans, for example, are usually understood never to be repaid), the time value of assets, the credit risk of reserves invested abroad, etc. On the other hand, certain assets such as real estate are frequently registered at an original book value below their current economic valuation.
Considering all off-balance-sheet and contingent liabilities at their economic valuation, a central bank will usually have negative net equity, with analysis having to focus on the magnitude of such negative equity. Forward contracts in foreign exchange make net equity more negative, as do financial sector deposit guarantees, with the cost of such guarantees naturally rising in proportion to the size of commercial bank assets. For any given level of such assets in nominal terms, the economic value is also dependent on the quality of bank loans; the risk-adjusted value of bank assets declines as the quality of bank loans deteriorates, increasing the value of the central bank contingent liability and decreasing its net equity.
Financial crises have caused hardship in individual countries, but to date they have not threatened globalisation. Any tendency to draw back from world financial markets, as in Malaysia, has been isolated and temporary.
The problem is rather how to tailor macroeconomic and financial arrangements to the imperatives of globalisation and in particular to limit the vulnerability of the domestic economy to the crisis problem. The complexity of this problem is evident in the exchange rate dilemma. It has become clear that exchange rate systems involving fixed but adjustable rates, bands and crawls are increasingly crisis prone. The alternatives are to float more freely, as such Latin American countries as Chile, Colombia and Brazil and such Asian countries as the Philippines, Korea and Thailand have begun to do, or to adopt a hard currency peg, as in Argentina and Hong Kong.
Indeed, if there is a difference with the management of the ERM crises, it is the importance of international banking supervision, which is now more explicitly acknowledged than it was then. It includes better risk management along the lines proposed by the BIS and would also support the creation of new networks including major emerging markets, such as the BIS itself or the G-20.
It may be, then, that the crisis encouraged responses that would not have been possible in calmer periods and in this sense served as a co-ordinating device. Bank restructuring which took place during the crisis may not have otherwise happened; as a result, debt structures are in better shape now than if countries had postponed reforms
Grandiose reforms of the international architecture are not in the cards; the existing architecture is deeply embedded in dense networks of market and political relationships. But there have been concrete steps in the direction of strengthening the architecture, such as the creation of the Financial Stability Forum and the G-20, which have been investigating highly leveraged institutions, offshore centres and short-term capital flows. They have tabled a variety of modest recommendations aimed at improving capital-flow statistics and strengthening prudential standards in both lending and borrowing countries. Progress in risk management techniques is also required, as many global players have not yet exploited the potential of their own auditing services.
Thus, after analysing the limits of VaR in the LTCM crisis, Myron Scholes (2000) concludes: "Now it is time to encourage the BIS and other regulatory bodies to support studies on stress-test and concentration methodologies. Planning for crises is more important than VAR analysis".
Beyond the broad agreement, several issues about implementation of transparency requirements remain controversial. Thus Avinash Persaud (2000) documents the shortcomings of the regulatory schemes in place at the moment: "regulators are warned about whether banks in their jurisdiction have exposures that threaten themselves, not whether banks around the world have exposures which together threaten a foreign market and could become contagious".
Among the wide variety of proposals for a new international financial architecture, one of the few areas of agreement has been that more transparency is desirable. The creation of the Special Data Dissemination Standard (SDDS) reflects this consensus. It is a new benchmark for IMF members in providing accurate and timely financial and economic information to markets and the public at large.
The objective of transparency would be better served if central banks were to work towards producing and publishing regular VaR reports of national financial accounts (consolidated central bank accounts plus relevant treasury accounts), in line with current practice of international commercial banks in risk management.
The VaR measure is a synthetic yet comprehensive indicator, incorporating all the assets and liabilities of the national financial system, including contingent liabilities, thus permitting rapid comparison between different countries and the analysis of the evolution over time for any single country. It incorporates all known information on international currency, bond, equity and commodity markets, updated continuously as that information evolves.
Policy-focused VaR-analysis reduces the temptation of central bank officials to try to "beat the market", by making all their operations more transparent, increasing the accountability of officials and promoting good governance in institutions that usually prefer a very broad definition of "independence". From a global perspective, the use of VaR analysis promotes capital efficiency by allowing international investors to utilise available funds in such a way as to maximise returns for a given level of risk, measured uniformly across different markets.
In effect VaR analysis permits regular tracking of the situation under "business-as-usual" conditions and stress-testing to view the maximum likely loss of value under extreme scenarios. It incorporates portfolio effects and the analysis of components, to examine where the greatest risks originate.
An example of the assessment of policy options prior to decision-making would be to inform cabinet ministers of the impact on the national VaR measure of proposals to increase spending and the budget deficit. The use of "what-if" scenarios to study the impact on the VaR number of implementing alternative policies involves of course a great deal of judgement, but it can at least be guided by measurement. At the time of the Long Term Capital Management crisis, there were reports in the financial press to the effect that most financial institutions did not take notice of the results of the stress tests carried out in connection with the August l998 Russian crisis. Those who did suffered far less from the crisis.
The best way to avoid situations like the recent crises, is to make investment and capital flow decisions better informed as to the current risk levels implicit in different countries and regions. This could already be done with existing technology extended and adapted to include all the assets and liabilities of a national balance sheet for each country.
Countries could be encouraged to use the technology of Value-at-Risk (VaR), for which some guidelines already exist as developed by the Bank for International Settlements (BIS). They should provide reports in a common format, under the supervision of an international agency, such as the BIS or the IMF (for example as part of the SDDS).
If a special facility with the nature of "lender-of-last-resort" were established, either at a multilateral level (IMF or BIS) or on a regional or even a bilateral basis, it might mitigate the need for each central bank to have large reserves or expensive hedging instruments to help in asset liability management. In any event, the credit limits available under such arrangements would be one of the items included in the VaR report.
References
BIS. Amendment to the Capital Accord to Incorporate Market Risks. Bank of International Settlements, Basle 1996
Blejer, Mario and Schumacher, Liliana, Central bank vulnerability and the credibility of its commitments: a value-at-risk approach. Journal of Risk, vol. 2, no. 1, November 1999.
Blejer, Mario and Schumacher, Liliana, Central Banks Use of Derivatives and Other Contingent Liabilities: Analytical Issues and Policy Implications. IMF Working Paper, WP/00/66, March 2000.
Braz, José, "Risk-Informed Capital Flows – A Global VaR Approach", in Jorge Braga de Macedo and Tadao Chino (editors) Sustainable Recovery in Asia, Paris: OECD Development Centre, 2000, pp. 151-158.
CGFS (2000), Stress Testing by Large Financial Institutions, Committee on the Global Financial System, Bank of International Settlements, Basle April 2000
Dornbusch, Rudi, "After Asia: New Directions for the international financial system", MIT, July 1998
Macedo, Jorge Braga de, "Financial Crises and International Architecture: a 'Eurocentric' Perspective," OECD Development Centre Technical Paper no. 162, August 2000.
Naranyan, Paul, "Credit portfolio management", presentation at the central bank of Argentina, 24 August 2000
Persaud, Avinash, "Sending the herd off the cliff hedge: the disturbing interaction between herding and market-sensitive risk management practices", State Street, September 2000
Powell, Andrew, "Medicion de riesgo crediticio, requisitos de capital y previsiones: aplicacion de un modelo de portafolio a la central de deudores", presentation at the central bank of Argentina, 24 August 2000
Reisen, Helmut, "Overview", in Ricardo Hausmann and Ulrich Hiemenz (editors), Global Finance from a Latin American Viewpoint, Paris: OECD Development Centre, 2000a, pp. 9-18.
Reisen, Helmut, "Revisions to the Basel accord and sovereign ratings", in Ricardo Hausmann and Ulrich Hiemenz (editors), Global Finance from a Latin American Viewpoint, Paris: OECD Development Centre, 2000b, pp. 71-80.
Scholes, Myron S. "Crisis and Risk Management", American Economic Review, May 2000, pp.17-21.
Wilde, Tom, "Creditrisk+", presentation at the central bank of Argentina, 24 August 2000