Tools for measuring credit risk.

For some time, managers and directors of energy companies have known that their company's risk management programs must do more than simply monitor exposure to market risk within some value-at-risk limit. While market risk remains the largest risk faced by energy companies (particularly for power and gas marketers), credit risk brings up a strong second and looms large enough to create company-killer situations. In just this past year three widely discussed industry-wide credit events (the California crisis, the PG&E bankruptcy, and the Enron bankruptcy) have once again opened the eyes of managers to the need for improved credit risk management practices.

The most recent such event, the Enron bankruptcy, highlights the destructive potential of credit events. At the time of Enron's bankruptcy filing, the aggregate exposure to Enron of all its counterparties was estimated at $6.3 billion.¹ Energy companies held $900 million² of that exposure, and the "Top 10" most exposed publicly traded energy counterparties among this unfortunate group held a combined total of $685 million³ or 76 percent of all energy company exposure. ()

While it is too early to predict how much of this exposure will translate into credit losses, the effect on the market capitalization of our Top 10 was immediately quantifiable. The day after Enron's filing, the market capitalization of these firms was down a total of $4.2 billion (or an average of 10 percent) from the equivalent figure at the end of September; one month later, their share prices had still not recovered and were showing a $2.6 billion total decline⁵ (or an average decline of 12 percent). Each of the companies on this Top 10 list has scaled back growth plans to some degree. Four of the companies on the list, prompted by pressure from the credit ratings agencies, have raised a combined total of $3.3 billion in additional (and expensive) equity capital and convertible debt since the Enron bankruptcy.⁶ Worse, some of these companies must continue to guard against the danger of a sudden loss of investor confidence, as credit analysts and investors remain wary of "the next Enron."

Clearly, the stakes are high. The frequency of occurrence and the size of losses place credit risk management squarely in the "must-act" category. Unfortunately, while risk managers at energy firms acknowledge that they must improve their firm's credit risk management capabilities, most remain focused on current exposure measurement (i.e., current mark-to-market exposure, plus outstanding receivables) and collateral management. The problem with this focus is that it places excessive emphasis on the present and fails to provide an acceptable indication of credit risk at some point in the future. Because losses from credit risk take a relatively long time to evolve, a more useful measure of exposure is .

Unlike current exposure, potential exposure exists in the future and therefore represents a range or distribution of outcomes rather than a single point estimate. Two useful measures of potential exposure are best described by the questions they answer and the applications they support. The first measure, average or Expected Potential Exposure (EPE), answers the question: "How large can I expect my exposure to this counterparty to become over the relevant time horizon?" EPE is useful for applications like capital adequacy and deal pricing where a portfolio of counterparties dampens the effect of the loss from any single counterpart. The relevant time horizon for EPE extends to a year or even longer. The second measure, Maximum Likely Potential Exposure (MLPE) answers the question: "What is the most I could lose to this counterparty with some degree of confidence (e.g., 95 percent) over the relevant time horizon?" MLPE is similar to the market risk metric of value at risk (VaR) and is handy for limit setting and stress testing to contain the effects of "the next Enron." For limit setting the relevant time horizon for MLPE can extend to deal maturity, while stress-testing horizons are typically less than a year but longer than the single-day horizon common to a typical VaR framework.

Some energy companies have begun to recognize the importance of potential exposure in their credit risk management frameworks. When these companies have applied potential exposure methodologies, they have naturally borrowed heavily from the accepted practices for measuring derivative counterparty exposure in the financial markets. This is a step in the right direction, but it is not sufficient. Practices borrowed from the financial markets must be adapted to cope with the unique characteristics of counterparty exposure within the energy sector. A good example of where improvements to this approach must be made is highlighted in Figure 2, where the extreme volatility in natural gas prices during the California crisis last summer rapidly converted potential exposures associated with long-term contracts into actual exposures (and in some cases real losses). Here the price swings erupted so quickly and were so extensive that financial market models failed to predict the extent of the damage (i.e., they severely underestimated potential exposure prior to the crisis). To avoid being bloodied by actual losses in the next crisis, energy companies must adopt forward-looking potential exposure measurement techniques using models that are appropriate to their markets. The remainder of this paper focuses on how potential exposure is measured and describes five key ways that it can be employed to better control credit risk.

Potential Exposure- A Calculation Primer

For a company to protect itself from a credit loss, it must first be capable of assessing its exposure to a given counterparty before default. As we discussed above, this exposure is not simply the netted sum of all the current mark-to-market (MTM) values of the trades with a particular counterparty, plus some outstanding receivable balance. We must also take into account how large a particular exposure might grow to over some relevant time horizon. Figure 3 illustrates this concept of potential exposure with regard to a two-year swap contract for natural gas.

1. At origination of the deal (assuming it is priced at the money) the MTM value is zero.
2. Over time, however, natural gas prices vary and, as a consequence, the MTM value of the contract will change. Each line in Figure 3 represents a possible path that the MTM value could take. In some cases, the points on the lines are positive (representing a positive MTM value on the position), and in some cases the points on the lines are negative. In yet other cases, the MTM value switches back and forth between positive and negative. The more time that passes, the greater the range of possible price movement and the larger the range of possible MTM values.
3. As the deal gets closer to maturity, the MTM value returns toward zero. This is because the MTM value of deal is a function not only of possible changes in market price, but also of the number of remaining contractual cash-flows stipulated by the deal. (Other types of derivatives can give rise to other "shapes" of potential exposure over time.)
4. In the context of credit risk, the potential exposures that are of concern are the positive exposures depicted in Figure 4.
5. When evaluating the amount of credit exposure generated by a deal, future negative MTM values are treated as equal to zero credit exposure.
6. The Expected Potential Exposure (EPE) of the swap contract is the average of the remaining possible MTM values, treating negative values as zero. It is an important measure of potential exposure that figures into the analysis of capital adequacy.
7. The Maximum Likely Potential Exposure (MLPE) of the contract is a "worst case" measure calculated to some acceptable confidence interval.

This measure indicates to executives how certain they can be that the loss from a counterparty will not exceed the MLPE value. For example, a firm might decide to calculate the number to a 95 percent confidence interval, which is consistent with most market risk VaR calculations. In this case, executives can feel confident that in 95 cases out of 100 the loss that they experience will not exceed the figure for MLPE generated by their model.

The accepted approach to calculating potential exposure in the financial derivatives and banking markets must be adapted to take account for significant additional challenges in the energy industry. Specifically:

• Financial derivatives market participants have simpler netting requirements. This is particularly the case for energy companies that trade in both physical and financial contracts.
• Energy markets are shaped by a much larger number of risk factors than the financial derivatives markets.
• Energy market risk factors typically possess troublesome price behaviors (e.g., price jumps) that are often compounded by illiquidity concerns.
• Contract portfolios of energy firms typically contain many highly structured and long-dated trades, often with large doses of embedded optionality written into them. This is particularly true of energy trading and marketing firms, but it is also the case for firms that hold energy assets. Each of these issues requires careful consideration when an energy company seeks to develop its own way to model potential credit exposure.

Potential Exposure Applications

The driving force behind the creation of a potential exposure model at most energy firms is the desire to enhance a set of key management tools and applications.

While each of the applications in Figure 5 can be addressed without an appropriate potential exposure model, the quality of the results is heavily dependent on the exposure input. Here is how a forward-looking potential exposure model improves these applications:

• Most responsible energy firms want to make sure that the risks they are running are in line with the risk capital that they have available. This means making use of a key risk concept known as 'economic capital', which is the amount of capital that a firm should hold to protect itself from insolvency to a given degree of confidence over a specific time interval (typically set at one year). It is an increasingly important risk metric because it creates a common denominator of risk across risk types (e.g., market, credit, business). What matters most when calculating economic capital is the expected level of losses, and the volatility of those losses. To compute the amount of economic capital that a firm needs to cover its credit risk, the firm must be able to predict its expected potential exposure at the one-year horizon, and better still, to do so using a model that also predicts the volatility of that exposure.
• As Enron began its descent into bankruptcy, the risk managers at the energy giant's many counterparties had one question at the top of their minds: "What is my current exposure to Enron?" Measures of current exposure are clearly useful in such a situation, but two problems often exist. First, managers often have a difficulty getting hold of timely exposure measures. Second, even when current exposures are measured, the positions that create them cannot be instantly unwound. Recognizing these two problems, the more alert managers were asking a different question: "How big could my exposure to Enron become over the next month?" When a counterparty's credit quality rapidly begins to deteriorate, a model that is capable of measuring both expected and maximum likely potential exposure over an adjustable time horizon is a valuable tactical asset.
• In the wake of the Enron collapse, liquidity management is being viewed with an increased sense of urgency. It can be complicated to work out the liquidity profile over time for a complex energy portfolio. For example, financial transactions will often have margin requirement provisions, whereas physical transactions generally do not. A model of potential exposure that carefully accounts for the differences in margining requirements between various transactions can be employed to improve the predictability of short-term cash needs. Furthermore, it can offer early warning of any potentially large margin-calls that might threaten a firm's liquidity in the medium term.
• Few would deny that the presence of credit risk in a transaction generates a cost. A model similar to the one that calculates Economic Capital over one-year can be employed to calculate the cost of credit risk over the entire life of a transaction. The key input for this model is the figure for the potential credit exposure of an energy contract, and the volatility of that exposure over the entire life of the transaction. Once the cost of the credit risk has been identified, it can be built into the price of the transaction. Alternatively, management can at least make sure that it is aware of the cost of a deal in terms of foregone risk-adjusted profits if the cost cannot be passed on to the counterparty.
• To be effective, a risk management program must offer a mechanism for limiting the size of a firm's exposure to its counterparties. Traditionally, such credit limits are tied to the credit rating of the counterparty, and are often identified in terms of the amount of exposure. For example, a firm might take on no more than $10 million of credit exposure to a counterparty that is rated BBB by the rating agencies. As we have seen, however, this limit is not sufficient if it is defined simply in terms of current exposure. It might even be dangerously misleading. For example, a swap contract that is struck at the current market price (i.e., at the money) has zero current credit exposure at the moment that the deal is done. Yet common sense, and the swap contract example that we discussed earlier, suggest that the forward-looking credit risk of the contract must be captured if the firm's system of credit limits is to function in an appropriate manner. Again, a model of potential exposure can provide an accurate measurement approach, and can also be adapted to provide advance warning of the need for any corrective action.

The Answer?

The analysis of potential credit exposure is not a substitute for traditional prudent risk management. A potential exposure model will not prevent the occurrence of the "next Enron" bankruptcy. What potential exposure can do is provide a consistent measurement framework that supports a number of vital credit risk management applications that greatly enhance a firm's ability to forecast, control, and respond to crisis events. The Enron horse is already out of the barn. But a potential exposure model will help close the barn door before the next crisis in the energy markets is upon us.

1. "Companies' Enron exposure estimated at $6.3 bln" Reuters News Service, 7 December 2001
2. Ibid.
3. Ibid.
4. As the Enron bankruptcy has not been completely resolved, all figures are preliminary
5. NewPower Holdings Inc. was formed through a subsidiary and partnerships of Enron Corporation and is 44 percent owned by Enron.
6. Ibid.
7. Ibid.

Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.