07/14/97 TAKING THE ANGST OUT OF TAKING A GAMBLE

View items related to this story

TAKING THE ANGST OUT OF TAKING A GAMBLE
It's called value-at-risk management, and it's quickly moving into mainstream finance
In early 1994, trading operations of many big banks and brokerage firms discovered that they had accidentally put all their eggs in one basket. Sure, they had split their money between such assets as commodities, debt, and cash. But when interest rates shot up, they found that every asset class contained a built-in assumption that interest rates would fall. Debt was in long-maturity bonds; commodities were in holdings such as interest-rate futures that would appreciate if rates fell; even cash was sunk in ''structured notes''--derivative securities that paid little or nothing if interest rates got too high. The result: one big splat.
Losses from that hidden concentration of risk helped convince the institutions that there must be a better way to handle risk--and now there is. It's a new risk-management methodology called ''value at risk.'' The idea is to calculate--dispassionately and methodically--the amount of risk inherent in a financial portfolio at any given time. The method spits out a number, known as the value at risk, which is the most money a portfolio could possibly lose in, say, the next 24 hours with a level of confidence of 95%, for example. It looks at how risk exposures in different parts of a portfolio can either amplify each other or (happily) cancel each other out.
Value-at-risk techniques not only make sure that you're truly diversified and have adequate capital on hand. They also tell you when you're leaving money on the table by being unnecessarily cautious. For those reasons, the techniques are moving into mainstream finance with startling speed. In 1994, they were used mainly by a handful of international banks and brokerage firms to manage risks on their trading desks. But now, VAR is expanding in two dimensions: It's being used for risks other than market movements, such as the risk of default by the other party in derivative deals. And it's being used by people other than traders, including mutual-fund managers and even CFOs at nonfinancial companies, including Xerox, General Motors, Enron, and GTE.
HISTORICAL DATA. Some of the momentum comes from regulators, who see VAR analysis as a more efficient means of determining the amount of capital that firms, especially banks, should have in reserve against unexpected losses. But much of the push is from the companies themselves. Enron Capital & Trade Resources, the nation's biggest wholesale marketer of natural gas and electricity, calculates its VAR every day.
A major boost for value-at-risk methods came on July 1, when a benchmark for credit-risk management developed by J.P. Morgan & Co. picked up the endorsement of the world's two biggest credit-rating agencies as well as eight big banks and accounting firms. The benchmark, known as CreditMetrics, contains historical data about credit risks that can be plugged into computer programs that firms use to calculate their value at risk. It's linked to RiskMetrics, a benchmark for market-movement risk that Morgan introduced in October, 1994. Says James E. Satloff, a managing director in charge of product development at Standard & Poor's Corp., one of the new CreditMetrics sponsors: ''The combined analysis of credit and market risk is going to become one of the most important financial-management tools.''
To be sure, the value-at-risk approach won't protect against what's known as operational risk, ranging from outright fraud to costly clerical errors. While Merrill Lynch & Co. calculates its VAR daily, it puts more emphasis on sound personnel practices and structural checks and balances, says Daniel T. Napoli, senior vice-president in charge of global risk management. ''Once you've built your statistical models, there's a temptation to fall in love with them,'' says physicist John D. Breit, a managing director working for Napoli. ''You're diverting a lot of very smart people onto 'angels-on-the-head-of-a-pin' land.''
True enough, but the VAR technique is catching on nonetheless because it does things that older risk-management methods cannot--such as discover ''hot spots'' of risk. If an operation is producing small profits yet requiring a big capital cushion because of its high VAR, it is likely to receive instructions to cut back or hedge. The beauty of VAR is that it allows for comparisons between business units that use different yardsticks for internal purposes, says Lev Borodovsky, executive director of the Global Association of Risk Professionals in Tarrytown, N.Y. ''You can really tell who's got more risk on the books.''
Value-at-risk calculations can't be taken as gospel. They're typically based on historical patterns that aren't always a good guide to the future--especially in times of turmoil. So most firms supplement the analysis by ''stress-testing'' a portfolio to see how it would perform in various worst-case scenarios.
Although regulators like VAR, it poses a dilemma. On the one hand, it's a more precise basis for setting capital requirements than traditional techniques, which ignore the opportunities for a portfolio to be internally hedged. On the other hand, there is no universally accepted way to calculate value at risk. Regulators have to take the word of firms that their rocket scientists have done their VAR math correctly.
BEYOND BANKS. Despite those qualms, the European Commission and the Basel (Switzerland) Bank for International Settlements, an advisory body, have cleared European banks to start using their own VAR models as the basis for calculating capital requirements starting in 1998. The Federal Reserve is postponing its own move to VAR as it considers an even more deregulatory approach. While the Europeans will be required to hold capital reserves equal to three times a conservatively calculated VAR, the Fed is studying whether to let banks decide for themselves how much capital to hold based on their VAR calculations--though they would be subject to fines and public disclosure if they underestimated risk and their capital reserves proved insufficient.
VAR is going beyond banks. U.S. broker-dealers are voluntarily reporting daily calculations in anticipation of a new regulatory scheme for derivatives trading--much of which goes unregulated now. Also, the Securities & Exchange Commission may require all types of firms to report the estimated market value of derivatives on their balance sheets. One way is through VAR analysis. In the next decade, banks may be required to mark all assets to market on a regular basis--even ordinary loans and mortgages. They will need an accepted tool for making those calculations.
This is great news for the pioneers of value-at-risk analysis, including J.P. Morgan, Bankers Trust, Credit Suisse, and Deutsche Bank, which are selling their expertise. Bankers Trust Co. charges $1 million for RAROC 2020, which shows the risk-adjusted return on capital for each element in a company's portfolio, as well as the exposure to each type of risk. ''If you appear to be well-versed in risk management, you get to talk to the fancier clients,'' says Till M. Guldimann, an originator of J.P. Morgan's RiskMetrics who is now executive vice-president of Mountain View (Calif.) Infinity Financial Technology Inc.
For all its imperfections, VAR analysis is a disciplined method for tracking down sources of risk that might otherwise escape notice. Lisa K. Polsky, a Morgan Stanley & Co. managing director and chief derivatives strategist, likes to quote J.R.R. Tolkien: ''It doesn't do to leave a live dragon out of your calculations if you live near him.'' VAR helps find those dragons.
By Peter Coy in New York, with bureau reports

RELATED ITEMS

TABLE: How to Figure How Much Is at Stake