Market risk management is no longer the domain of big investment banks. It is becoming a buzz phrase across all areas of business, from smaller banks and hedge funds to insurance companies, pension funds, asset managers and non-financial companies.

Click on the button to see how market risk affects different types of institution.

Spurring on this development is a series of regulatory changes. In the US, for example, the Securities and Exchange Commission's 1997 rules on derivatives disclosure forced financial institutions to sit up and take note. And the switch to "fair value" accounting, introduced by the Financial Accounting Standards Board in 2000, and by the International Accounting Standards Committee this year, is making derivatives fully accountable on the balance sheet.

These changes are piling pressure on financial and non-financial companies alike to confront the need for rigorous market risk management. Market risk is defined most simply as the chance of financial loss caused by unfavourable movements in market prices. It can arise from a straightforward exposure to a single market position. But it's also driven by more complex factors such as the interplay of two or more market positions, and problems of market structure.

For example, "basis risk" is a special kind of market risk arising from a change in the relationship between two risk factors or financial assets - such as two interest rates, or the price of a hedge and its hedged position. And market "liquidity risk" describes the danger of being unable to sell a market position at short notice because the number of buyers has dried up. We'll take a closer look at both these special forms of market risk later on.

The advances in market risk management are part of a much broader trend in financial markets and institutions: the recognition that return is only one side of the performance equation. Information about risk, including relatively unlikely or unexpected risks, is necessary for evaluating trading strategies, investment choices and asset allocation.

But risk information is most meaningful in the context of the other risks a firm faces and the measures it has taken to offset them. This is where value-at-risk (VAR) comes in. Although market risk management didn't begin with VAR, it has revolutionised the tools available to market risk managers.

VAR is a statistical methodology that helps risk managers to aggregate risk numbers across business and product lines in a meaningful way. This helps financial institutions to impose risk limits on market risk exposure, and it helps them to manage and optimise risk across their various portfolios - at a corporate level as opposed to a trading desk level.

Developments outside the financial markets are bringing home the importance of market risk management to a new breed of risk managers, too. In the newly deregulated and privatised markets of electricity and telecoms capacity, investment in risk management technology and expertise is soaring as companies become aware of the volatile market risks in their businesses.

Click on the picture of our first Expert Witness to read about the dramatic volatility of the newly deregulated US power and energy markets - and to see why even the most conservative companies have been obliged to build new strategies for managing market risk.

In the insurance industry too, market risk management is becoming both more widespread and more complex. Many insurance or reinsurance companies are establishing their own financial subsidiaries to offer derivatives products, subjecting them to the same risks as investment banks. They are also offering "blended transactions", which bundle together insurance and financial products. These entail complex risk monitoring and management, as our next Expert Witness explains.

Meanwhile, internet trading, e-commerce and new auctioning mechanisms are developing new markets for any number of goods in different sectors of the economy. As these markets become more liquid and offer transparent prices for goods in which no proper market has existed previously, they are likely to broaden the need for market risk management still further.

Deregulation of markets, advances in financial engineering and increased computing speed over the last 30 years have been the incubator for modern market risk management. Click on our timeline to track through some of the key developments and events.

In the last ten years, market risk management has come a long way. Some argue today that it has even become too efficient. Widespread use of market risk models means that many firms may be prompted to reduce their exposures at the same time, which could have dangerous implications for market liquidity.

This type of "herd mentality" has been seen as a contributing factor in the risk contagion of 1998, which stemmed from Russia's default and the near-collapse of hedge fund Long-Term Capital Management.

The proliferation of off-the-shelf risk management software means that it has become all too easy to create VAR and stress numbers in large quantities, without a correspondingly deep understanding of their meaning. But the key to good risk management lies in what you do with the numbers, not simply in generating them. It is essential to understand how the numbers are produced, to distil them and combine the information they offer with human intuition and understanding.

Even more worryingly, today's measures of market risk are really designed for formal, liquid markets where risk is calculated over short time horizons. In markets where trade is illiquid, prices opaque or risks non-tradable, VAR may be inadequate or even misleading.

Research continues into these tricky areas, with institutions supplementing VAR with additional mechanisms such as stress testing, non-statistical limits, and buffer reserves or capital charges. Research continues too on adapting market risk management for investment management and corporate needs, for example by incorporating longer time horizons.

At the same time, the relative success of the new market risk management techniques and methodologies has led to their partial adoption for measuring and managing other kinds of risk, especially credit risk and operational risk. The market risk manager's mantra, "Identify, measure, monitor and control" is common today among credit and operational risk managers.

The formal quantification of market risk as a key component in investment decisions also helped spawn a more general approach to risk-adjusted return on capital - with all its strategic implications for financial and other businesses.

Value-at-risk - the industry standard
Financial legend has it that in the early 1990s, Dennis Weatherstone, chairman of investment bank JP Morgan, demanded to see the bank's aggregated risk exposure at 4.15 each day. The type of report that Weatherstone received is now a common feature of financial institutions worldwide, and it entails calculating the firm's value-at-risk.

Weatherstone's demand led eventually to the creation of RiskMetrics, a VAR methodology that when it was made public helped create an industry standard approach to risk management - as our next Expert Witness explains.

VAR estimations provide information about the potential for losses in value for a given position or portfolio. Of course, over any time horizon, a portfolio can lose 100% of its value - maybe even more if the portfolio is leveraged. So VAR is an estimate of the possible size of loss over a given time period, and for a given level of confidence.

For example, if the parameters of the model have been set conservatively, a senior executive looking at a VAR number might be able to think, "There's a 99% chance that our losses won't exceed that VAR number within the stated time horizon".

There are three ways to calculate VAR: the variance-covariance approach, historical simulation and Monte Carlo analysis. Click on the button to find out how they work.

Today, numerous companies offer tools to calculate VAR to rival those offered originally by RiskMetrics. Click on the button for a sample of the kind of tools that are on offer.

VAR is an important tool to help decision-making at all levels of management in a financial institution. For members of the board, it provides a highly aggregated view to help evaluate the firm's risk appetite and risk taking. It gives senior management a strategic view of risk taking that can help them manage the firm's risk profile and control the risks it assumes by setting limits.

For traders, it affords the tactical risk information they need if they are to comply with the market risk limits set for them by their business and senior managers - though traders will often also use other risk measures to help them manage the risk and rewards of specific positions and instruments.

By providing a consistent, integrated risk assessment that covers all types of market risk faced by the institution, VAR helps make risks transparent and allows them to be tackled in a consistent manner. As VAR produces a single aggregated risk number, it can be used to set regulatory capital requirements - although the output of VAR models does vary.

But VAR is not perfect. While it provides a powerful tool in normal market conditions and over short timeframes, it can be misleading in times of market turbulence. Extreme market volatility causes assumptions - such as the level of correlation between the price movements in different market positions - to break down and, often, market liquidity to dry up.

As VAR does not take into account differences in liquidity, or provide information on the potential size of losses on days when VAR is exceeded, other tools must be used to explore these risks. Such tools include stress testing and scenario analysis, as well as special methodologies for incorporating liquidity risk into VAR.

Beyond VAR - additional ways to measure risk
In turbulent market conditions, the shortcomings of VAR can become all too obvious. Stress tests, which simulate portfolio performance in abnormal markets, hold some of the answers to this problem. They help companies to understand the effects of large market moves that cannot reliably be estimated statistically, and they can also be used to expose any non-linear effects of these moves on positions in the portfolio.

Stress tests can also provide information on the potential size of losses on days when the VAR figure is exceeded. They do this by means of scenario analysis. By analysing specific scenarios, such as a major shift in interest rates or commodity prices, or by replicating past market shocks such as stock market crashes, the stress test can indicate the size of loss under each scenario. Stress tests may also use worst-case scenarios to discover the maximum possible loss to a portfolio within a specified time horizon. They do not, however, indicate the probability of these losses occurring - unlike VAR.

Another way in which stress tests overcome the shortcomings of VAR is that they can indicate the direction of risk exposure, for example, whether the risk of loss is from a rise or fall in interest rates.

Market risk managers can use stress tests to indicate where they should reduce any unacceptable levels of risk in individual portfolios and whether they need to diversify any concentrations of risk in the aggregated portfolio, as portfolio effects may differ in stressed markets.

In order to provide meaningful results, the scenario should only include events that are plausible - as our next Expert Witness explains. Scenarios should be relevant to the current economic climate and all the movements in risk factors within the scenario (ie, price or rate moves) should make economic sense when taken together. They should also be relevant to the positions and strategies in the portfolio and sufficiently detailed to provide meaningful results for each position.

Within these general principles, there are seven key rules to follow to formulate meaningful stress tests. Click on the button to find out what they are.

Stress tests became widely used for the first time during the Asian crisis of 1997-98, when market risk factors often shifted wildly. Today they are considered by banks and regulators as an essential element of market risk management.

However, neither stress tests nor VAR can calculate potential tail-end risk - the probability of loss at the end of the distribution curve - for which data is severely limited. Extreme value theory is a way of filling this gap. Adaptation of EVT to financial risk management is very recent, although it has been applied to environmental events for several years.

EVT is a stochastic methodology for describing extremes in random phenomena to estimate the behaviour of unusual or rare events. One variation of EVT is well suited to producing estimates of fat-tail probabilities, for example, a one-in-ten year event. Another variation helps assess the expected loss on days when losses exceed VAR. Depending on the application, EVT may entail using explicit analytic formulas or complex simulations.

However, EVT is by no means perfect. Accurate estimates of extreme value distribution parameters require more historical risk factor data than is typically used, or available, for VAR. And EVT is not well adapted for drill-down analysis of portfolios, which is essential for understanding the interplay of risks and the underlying cause of a risk exposure.

Adapting VAR models - liquidity risk and portfolio management
There are some areas of market risk management for which traditional VAR models and stress tests are insufficient.

One of these is liquidity risk - the risk that it will not be possible to transact a deal at the prevailing market price because of a lack of available counterparties. As our next Expert Witness explains, liquidity is a difficult risk to manage.

For example, in stressed markets where major market makers have similar inventories of positions, they may simultaneously attempt to exit or hedge those positions.

When market liquidity dries up in this way - as it did during the third quarter of 1998 with the Russian default crisis and near-collapse of hedge fund Long-Term Capital Management - bid-offer spreads can widen dramatically, especially for large deals. It may be necessary to postpone the deal or trade below the market price.

In addition, periods of illiquidity tend to be associated with an increase in volatility and, in extreme cases, with the breakdown of normal correlations between markets or securities.

Traditional VAR models are so-called "one period" models that do not take into account differences in the trading environment that might affect liquidity. Some risk managers try to take liquidity into account by increasing their volatility estimates, and thus boosting the VAR figure - but this is an artificial procedure that is almost bound either to overestimate or underestimate liquidity effects.

That's why much research has focused recently on adapting VAR to stressed, illiquid markets to create a more dynamic framework, for example by allowing positions to alter as the trading environment changes. But it's proving an uphill task, as the button below explains.

Another way in which current research is developing the VAR framework is in the field of portfolio management. Excessive concentrations of risk to certain types of exposure in the portfolio (say, to shares belonging to a particular industry sector) can threaten an institution in the event of adverse developments in that sector. But measuring portfolio concentration is difficult in a traditional VAR framework as there is no obvious best metric for measuring excess concentration with VAR, as there is with the "standard deviation" measure of traditional portfolio theory.

The relatively new concepts of "incremental VAR" and "delta VAR" respectively indicate which risks have the most adverse effects on the portfolio and measure how much each position contributes to the overall risk of the portfolio. Once the riskiest assets have been identified, suitable hedges may be put in place to help manage the portfolio.

Another key concept in portfolio management is risk-adjusted return on capital (or Raroc) which, in effect, applies VAR to measure the risk-adjusted performance of assets, products and businesses. It allocates to each position a capital charge equivalent to its maximum expected post-tax loss (ie, VAR, with some adjustments) over the course of a year.

Because Raroc offers a framework for balancing VAR against asset returns, it helps institutions to compare risk-adjusted returns within and across a great variety of portfolios. Riskier assets, products or businesses must generate greater returns to compensate for the higher capital charge. If they do not, the bank will not retain or invest in them.

Raroc was developed in the 1970s by Bankers Trust and expanded to include portfolio concepts such as correlation in 1991. There are various other ways to balance market risk and return, as the button above explains.

Asset liability management - from accounting to VAR
Market risk is commonly considered the preserve of a financial institution's risk management function. But another group of managers also has an interest. A bank's asset/liability management team is responsible, as the name suggests, for balancing the assets and liabilities of an institution. In practice, this often means controlling the institution's exposure to fluctuations in interest rates and maximising net income and economic, or market, value of equity (without exceeding risk limits).

Changes in financial markets, notably interest rates, impact on the balance sheet, which the ALM team is responsible for protecting, as well as on the trading book, supervised by risk managers. But the risk measurement models for ALM traditionally used an accounting approach.

These static "gap" and simulation techniques, however, are being complemented today by mark-to-market models based on economic value of equity, and lately, VAR. The switch to market value accounting has helped spur the growth of the mark-to-market approach among asset/liability managers.

Static gap models produce a matrix showing the effect of interest rate shocks on net income. Net interest income simulation provides "what if" scenarios showing the impact of interest rate changes on net income and individual items on the balance sheet.

Newer market value (or duration) models, however, measure the net worth of a financial institution according to a mark-to-market measure of its expected cashflows, rather than an accounts-based estimate of net worth. Their aim is to ensure the market value of assets is in line with the market value of liabilities.

However, there are difficulties in translating the mark-to-market approach to the balance sheet. Models assume the market value of each item can be derived from the present value of its cashflows. As there are no active markets in many balance sheet items, a mark-to-model approach is used. Model estimates of value or interest rate sensitivity are particularly dependent on the model used and the accuracy of the estimated model parameters.

VAR introduces an element of mark-to-market into this process by mapping the model-estimated sensitivity of the item to an actual, historically observable risk factor (two-year swap rates, for example). Using VAR also permits aggregation of risk exposures across the banking and trading books. But as a daily market risk measure, value at risk is perhaps not an obvious choice of methodology for long-term balance-sheet management in an asset/liability management context.

The diagram below helps make clear what each of the four types of ALM model are most suitable for, from a market risk perspective.

Four Types of ALM Model

Criteria	Static gap	Simulation	Market value	VAR
Complexity	1	3	2	4
Foundation	Accrual accounting	Scenario analysis	Present value	Statistics
Critical assumptions	Rate shock	Account growth Product mix	Yield curve Cashflows	Correlations Volatilities
Best for	Simple, stable balance sheetsk	Scenarios Budgeting	Longer terms Hedging	Trading book Raroc

• Net Interest Earnings at Risk - the potential negative change in net interest income resulting from a change in interest rates, usually measured over one year.
• Net Earnings at Risk - includes the indirect impact of interest rate changes on the non-interest rate items, such as fee income.
• Equity Value at Risk - the potential negative change in the present value of total equity resulting from an assumed change in interest rate risk. This is probably the most suitable interest rate risk measure for ALM as it is a long-term strategic measure. It also fits in with the approach taken by the Basle Committee - the international regulators of the banking industry.

Derivatives can be divided into two broad groups: exchange-traded and over-the-counter. On a regulated futures exchange, the exchange itself acts as the counterparty to every deal. Futures contracts are traded in parcels of specified size and for fixed dates, and are usually short term. The last two years have seen a revolution in the field of exchange-traded derivatives as exchanges abandon their trading pits in favour of cheaper screen- and web-based trading.

Hand-in-hand with the move to electronic trading has come a wave of mergers between futures exchanges in different countries to form alliances. For example the Globex alliance links the Chicago Mercantile Exchange, Singapore Exchange, Paris Bourse, Spanish Futures and Options Exchange, Montreal Exchange and Brazil's Bolsa de Mercadorias & Futuros. A trader can now buy futures on an exchange half way round the world, around the clock.

Financial futures traded as standard contracts on an exchange are not always suitable for corporate risk management needs, as the terms of the contracts rarely match a company's risk exposures exactly. Some companies are unwilling to deal with the basis risk that arises from hedging positions with an imperfect match. The need to arrange a separate source of funding to meet calls for margin may also deter corporate users. Commodity companies like oil or gold producers, however, are some of the biggest users of commodity futures.

Over-the-counter derivatives, offered by banks and, increasingly, insurance companies, can be fitted to hedge the underlying exposure more exactly. The current trend among sell-side banks is to treat plain vanilla OTC derivative instruments as "commoditised" products, which they can trade electronically (often on their websites) while concentrating on more innovative "tailor-made solutions".

While derivatives have become an essential tool for managing risk, the very mention of them fills some less savvy institutions with fear. This is because derivatives fulfil a dual role - they can also be used for speculation. Dealing in derivatives can provide enormous exposure for relatively little outlay, increasing the leverage of an institution. This means that if something goes wrong, it goes wrong big time.

The last few years have witnessed several market risk - and risk management -derivatives scandals that have brought companies to their knees. Click on the button to find out how derivatives caused Ashanti Goldfields' downfall, or review the market risk case histories listed in our Case Studies section.

Regulation - a key motivator
Reform by regulators and supervisors has been one of the key factors motivating developments in market risk management, especially in the world of banking. As well as developing a regulatory capital charge that banks must hold against each type of asset, regulatory, self-regulatory and supervisory bodies have also helped to establish wide-reaching best practice guidelines.

In 1996 the Basle Committee on Banking Supervision, the main international banking regulator, introduced its Amendment to the Capital Accord to Incorporate Market Risks. The reforms, implemented in 1998, allowed banks to calculate their regulatory capital for market risk in one of two ways. The "standardised approach" uses constant factors set by the regulator, relating to the type of asset and factors such as its market value and maturity. Alternatively, banks can use their own VAR-based internal models, once these have been approved by a regulator and provided they are regularly back-tested to ensure accuracy. Click on the button for the scope and coverage of the capital charges.

The Basle Committee is currently reviewing its Capital Accord. While this review concentrates mainly on credit and operational risks, some of the changes that are planned will alter the environment for bank market risk management, as well. After the revisions, the revised capital framework will follow a "three pillars" approach. Minimum capital requirements are the first pillar. The other two are a supervisory review of banks' capital adequacy and internal assessment processes; and greater market discipline and transparency to improve banking disclosure as a way towards sound banking practices.

In 1997 the US Securities and Exchange Commission also highlighted the need for registrants to disclose accounting policies for derivatives and qualitative information about the market risk of financial instruments. Click on the button for details of the ruling.

The quality of market risk control structures within financial institutions were a focus of concern through much of the 1990s, starting with a report entitled Derivatives: Practices and Principles from the Group of Thirty in 1993. This industry-led study aimed to define a set of sound risk management and control practices for those involved in derivatives activity. It offered 20 recommendations, including daily marking-to-market of derivatives positions and the need for an independent market risk management function.

Various supervisory and self-regulatory bodies have also issued statements recently relating to market risk management and disclosure for non-bank financial institutions.

For example, in 1999 the US President's Working Group on Financial Markets, which comprises the main US regulatory bodies, issued a report entitled Hedge Funds, Leverage and the Lessons of Long-Term Capital Management, which made recommendations about disclosure by hedge funds. Click on the button for the report's recommendations.

Future developments - integrating risks and the electronic revolution
The art and science of market risk management is now in a relatively mature phase, after the frenetic developments of the 1990s. But while the next few years are unlikely to bring revolutionary change in the basic approach, there are significant problems still to be solved in terms of managing risks in immature and abnormal markets.

More immediately, huge structural changes are occurring in the environment of the market risk manager. Those changes can be summed up as the battle inside banks to integrate risks, and changes in the external trading environment brought about by the market deregulation around the world and the emergence of online trading and internet markets.

Market risk is one of four areas of financial risk management that financial institutions today are focusing on. The others are credit risk (or counterparty risk), operational risk (risks relating to people, processes and technology) and business strategy risk.

Traditionally a bank's market risk function would be separate from its credit and other risk management functions. But increasingly financial institutions are realising the value of combining risk functions, and of transferring many approaches and methodologies developed in market risk management to other areas of the bank.

In the Enterprise-wide piece of the Risk Jigsaw we explain why this makes sense from the perspective of an institution and its shareholders. But one additional reason stems from the markets themselves. The growth of the derivatives market has created a whole range of off-balance sheet assets that combine market risk and credit risk.

Today it is common even for on-balance sheet assets to have derivatives embedded in them - for example, a loan with a cap. In such cases, the transaction's credit risk is dependent on its market value, and it makes little sense to manage these risks as if they were unconnected.

The other key harbinger of change is the Internet. Retail investors and independent traders have already embraced the web. Not only are transaction costs lower, but it provides them with a level playing field for the first time, as traditional trading relationships and monopolies crumble. Electronic trading also provides anonymity, giving the small, independent trader the same clout in the market as the biggest players.

But anonymity also brings risk. It can help the big trader that is trying to move the market, as competitors will not be alerted to its market presence. And the sheer number of different electronic trading systems threatens to split liquidity, leaving traders with markets where spreads are wide and trading inefficient.

The revolution has proved slower burning for those building new wholesale financial markets on the net, but few investment bankers doubt that the effect of electronic markets and market risk management will be profound in 2002 and beyond.

After all, the Internet has already played an important role in the development of market risk management: it was JP Morgan's decision to make its RiskMetrics datasets available on the web that made VAR an industry standard.

In the longer term, the most profound changes may not be the ones that grab the headlines. Internet infrastructure developments likely to shape next-generation developments in market risk management include common languages and legal certainty, as much as front-end trading technology.

An example is the development of Financial Products Mark-up Language (FpML). Developed by JP Morgan and PricewaterhouseCoopers, FpML is an eXtensible Mark-Up Language, or XML-based meta-language for describing data shared between financial derivatives applications.

This might sound unexciting, but Standard FpML specifications for formatting derivatives and other financial data are just one example of the standardisation and digitising of market risk data. The trend should make it easier to gather and consolidate market risk management information for existing markets, and cut the cost and complexity of risk management in the markets of the future.