JP Morgan's RiskMetrics has become the leading standard measure of market risk, but measuring credit risk is still uncharted territory. Not surprisingly, Morgan has stepped up to the plate with a new measure of credit risk, which it hopes will become a second standard. The methodology for this long-awaited measure was described in an article entitled "On Measuring Credit Risk" in the most recent issue of Morgan's RiskMetrics Monitor (published on paper and at www.jpmorgan.com/RiskManagement/RiskMetrics/pubs.asp).

Here's a summary of the article for those not brave enough to paw or scroll through the 40 pages of dense text and equations.

The computation of credit exposure is fundamental to risk measurement in a transaction subject to default. Though the RiskMetrics framework is a well-established methodology for measuring risks associated with changing market rates, the risk in a particular transaction depends also on the credit standing of the counterparty. In addition to potential changes in swap rates, the degree of risk of an interest rate swap depends on whether the counterparty defaults before the swap's maturity. The credit exposure in a transaction therefore is the amount subject to risk when there is a change in the credit standing of a counterparty. It is, in effect, the amount that can be lost when a counterparty defaults. It is not a risk measure itself but rather an amount that can be combined with other information to provide a measure of credit risk.

Three methodologies can be employed for measuring the credit exposure of transactions whose values have been marked-to-market. Two of them supply credit exposure measures without relying on simulation, and may be computed with RiskMetrics methodology and data. A third estimates credit exposure by simulating future rates.

A simple interest-rate swap illustrates each of these methods. Credit exposure can be viewed as a function of current and potential exposure. The current exposure, of course, is simply a function of the mark-to-market value of a swap. Potential exposure, however, depends on the values of future swap rates as well as the mark-to-market values. Potential measures of exposure can be classed into worst-case and expected measures.

Why is the current exposure equivalent to the swap's mark-to-market value? A party to a swap with a positive market value will lose that value if the counterparty defaults. A party holding a swap with a negative value has a mark-to-market of zero. This follows from the fact that a party owing money at the time its counterparty defaults incurs no loss. For vanilla swaps, current exposure is equal to the replacement cost at current market rates.

The all-important determination of potential exposure is the exposure that may arise from future changes in interest rates. Unlike current exposure, a risk manager can do no more than estimate potential exposure, given some model on how rates and prices evolve over time. Common measures of potential exposure include maximum, peak, expected and average exposure. These calculations recognize the probability distribution of underlying prices. Worst-case measures provide estimates of exposure in terms of future values, and include maximum and peak exposure. Expected exposures, on the other hand, measure estimated credit exposure in terms of present and future values.

Maximum exposure is an important measure of credit exposure because it can help determine how much credit to allocate for transactions against a general counterparty. Risk managers use maximum exposure for credit risk control. In particular, they will often identify those transactions whose current exposure is greater than the maximum exposure defined when the transaction originated. A byproduct of maximum exposure is peak exposure, which is the maximum of all maximum exposures over a specified time interval. Peak exposure is a useful measure of credit exposure because it tells risk managers the time in the future when the largest losses are expected given that a counterparty defaults.

Expected exposure is the exposure that exists at any point in the future. It measures the amount, on average, that will be lost if a default occurs. In practice one can compute expected exposure at different points in the future over the life of a transaction. These sampling times could be equally spaced but need not be. The exposure measured at each sampling time would be equal to the current exposure. The weighted present value of these exposures is known as average exposure. (The weights correspond to different discount factors to account for averaging over time). Since averaging is performed over time, care must be taken to weigh each expected exposure by the appropriate discount factor.

This overview of potential exposure provides the basis for two analytic approaches for measuring credit exposure that do not rely on simulation. They may be computed with RiskMetrics methodology and data (volatilities and correlation). They employ normal probability models to measure exposure.

Worst-case (including maximum and peak exposures) as well as expected and average measures of credit exposure can be computed quite precisely using normal distribution. The maximum exposure at any given sampling time is an estimate of the maximum credit exposure, with a 5 percent chance that the realized loss is actually greater. In other words, in the event of default by a counterparty, there would be only a 5 percent chance of having to pay more than this amount to replace the outstanding transaction. As a practical matter, the calculation of all four credit exposures-maximum, peak, expected and average-requires expression for the mean and standard deviation of the distribution of present value of cash flows generated between the first sampling and maturity of the swap, discounted back to the first sampling. The standard deviation is a function of the sampling time and the time of the final cash flow, and is based on the volatilities and correlations generated by the swap. It can thus be related to the RiskMetrics daily VAR.

According to the statistical approach, expected and maximum exposures in a three-year USD par swap start off small and increase until they reach a peak, and then decrease as the sampling time nears the swap's maturity. The swap's credit exposure evolves in such a manner because the volatility or standard deviation scales with time, and because there are fewer future cash flows generated by the swap as it nears maturity.

In addition to measuring transaction credit exposure, this method lends itself to calculating portfolio credit exposure. The most misleading method for measuring credit exposure of a swap portfolio would be to aggregate the credit exposures. This is intractable because, among other things, it is not obvious how to report a peak exposure estimate for a portfolio of swaps with different duration. An alternative method is based upon bilateral netting, in which, for any given counterparty, positive market values are offset against negative market values at each sampling time. Such an approach would naturally reduce average exposures relative to simple aggregation. Such netting can have a significant impact on credit exposure estimates. In particular, the average estimate based on netting is lower than the aggregate approach. Moreover, it is now straightforward to compute peak exposure. Bilateral netting is suitable for a portfolio consisting of many counterparties. A company entering into numerous swap arrangements with three different banks would compute its credit exposure on a bilateral basis by first splitting swap arrangements into three groups depending on the swap's counterparty, then netting all cash flows within each group, and then computing credit exposure measures, as suggested above.

A third method allows risk managers to simulate future swap rates in order to measure credit exposure. Such simulation is attractive because credit exposure arises from changes in interest rates occurring after the swap is initiated. In a three-year par swap between a company and a bank, if the bank defaults six months after settlement, the replacement cost for the company-the cost to replace the bank counterparty-would be a function of the difference between the swap rate that prevailed when the swap was first purchased and the two-and-a-half-year par swap rate at the six-month sampling time. To determine the credit exposure to the company at this six-month sampling time, we need to simulate the distribution of the two-and-a-half-year USD par swap rates in six months, since they are the rates that the company will be faced with if default occurs. To do the simulation, we need the volatility of the two-and-a-half-year rates six months hence as well as the two-and-a-half year forward rate six months forward.

The simulation approach can produce quite different peak exposures. The large differences between nonsimulation and simulation results may result from two important factors. First, simulation uses par forward rates rather than zero rates to simulate future rate distributions. Second, the nonsimulation approaches use volatilities and correlations based on zero rates, whereas simulation applies volatilities and correlation on par rates.

Exchange officials who have been anxious about the growth of EFP (exchange of futures for physical) transactions should be able to sleep a little easier, thanks to a recent report from the Chicago-based Catalyst Institute.

EFPS are bilateral transactions in which counterparties privately exchange offsetting cash and futures positions at an agreed price. The exchange of cash commodities for futures has been around for a long time, and has been particularly valuable in the agricultural business as a means of exchanging a physical commodity while cancelling out the opposing futures positions of the two parties. EFP's thus kill two birds with one stone.

Nowadays, EFPs are used to inject flexibility into a trade as the terms of the exchange are agreed between the two parties, as in the OTC market. Volume has grown considerably in recent years alongside the general growth of financial futures and the evolution of 24-hour global markets. For example, in December 1995, more than 324,000 three-year bond contracts were traded via EFP, representing almost one-third of the entire volume in that contract.

Exchanges fear that all this is costing them business. For the most part, the Catalyst Institute finds that these fears are unfounded. It says that most EFP trades would not have occurred if the mechanism had not been available, and that current levels are not detrimental to market liquidity.

As a cautionary note, however, it says that all market participants do not have access to EFP trading information. If this situation persists, the lack of transparency could lead to decreased liquidity, greater transaction risk and greater hedging costs.