February 2004

Spreading the Risk of Default

Fueled by the world's economic problems-especially the largest number of corporate loan defaults since the 1930s Depression-a new industry in credit risk trading has been growing over the past several years. Before long, you will probably be in it.

by Kathleen O'Toole

In his job on the trading desk of an investment bank in Singapore, David Ling often finds money managers asking for his bank's latest credit risk offerings. The market is developing new products all the time, but the customers always want more.

"Our clients tend to bypass the older, simpler products," says the 1994 MBA graduate. His job involves figuring out what price the bank should charge for credit risk derivatives—investment products based on the probabilities that a company or country will default on its creditors or, more generally, experience a change in its credit rating.

At a time when interest rates are low and plenty of a-rated American and European companies have declared bankruptcy or defaulted on loans, Ling's mostly Asian-based customers want more credit risk derivatives of Asian companies in their portfolios. "There isn't enough to go around," he says, "and most of the ready-made deals are based on American or European credits."

What is available may have changed by the time you read this article. Invented less than a decade ago, the market in credit risk derivatives produces new products every month. There are collateralized debt obligations—CDOs for short—and credit default swaps, or CDSs. Indeed, a large fraction of CDOs are constructed synthetically from default swaps. By late 2003, institutional investors also could trade "baskets" of credit risk, including indexes of debt derivatives𤸼 companies in Europe, for instance, or 20 in an industry. Unlike the Lehman Bond Index, which investors can't buy, these indexes can be traded and could someday dominate indexed corporate bond funds, including those that individuals buy as they near retirement and seek fixed income vehicles, says Business School finance professor Darrell Duffie.

The credit derivatives market has been growing at the rapid rate of 50 to 100 percent annually since almost a year before the stock market bubble burst in the spring of 2002. It was expected to cover at least $2 trillion of corporate and sovereign debt by the end of 2003 and to reach $4 trillion by the end of this year. Exactly when it slows down may depend on the world economy, say Duffie and finance professor Ken Singleton, the coauthors of a new book, Credit Risk, published by Princeton University Press. Their latest research projects indicate credit markets are strongly affected by the business cycle or, perhaps more precisely, investors' knowledge and fears of the effect of the business cycle on corporate default risk. Ling gives this example: "Because some Asian investors got burnt on u.s. and European defaults and others have read about the problems, customers here now want to buy Asian risk."

"My parents think I'm a stockbroker," Tracey Benford, MBA '95, says with a chuckle. Like others who deal in credit risk derivatives, she says she gets little opportunity to talk about her work in this hot new market. "When the Enron and WorldCom bankruptcies were in the news, people would ask me about what I did, but fixed income investment in general seems to be a mystery to most people," Benford says.

While Ling says he was drawn to this career because of its technical challenges, Benford says she likes the excitement of rapidly changing, innovative products. (The money is good too, according to the Wall Street Journal, which reported last year that four to seven years' experience in credit structuring skills can bring salaries of $800,000 plus bonuses.) Benford sells mostly investment-grade U.S. bonds or credit derivatives in the Chicago office of Goldman Sachs to money managers who are trying to match or beat benchmark yields in their particular niche. "Originally, the main users of these products were banks and hedge funds, but the market has exploded in the last several years," she says.

In the not-so-old days, lenders or companies with outstanding receivables from customers would sometimes buy insurance against not collecting what was due to them, in much the same way a homeowner insures against his or her house burning down or a farmer insures against a hailstorm. Today, the more liquid credit derivatives are replacing credit risk insurance, say School alumni and academics in this field.

Institutions take positions in other, older types of derivatives as well. For instance, as global business increased in the seventies and eighties, companies that did business across borders worried not just about being paid but also about being paid in a currency whose value could drop against their own. That prompted a market in currency derivatives, in which institutional investors would take on the currency risk and charge a premium for the protection. If companies were worried about losses from interest rate volatility, they could also buy risk protection in the form of interest rate derivatives. Over time, standard contracts helped buyers and sellers agree on the conditions under which each had to pay or collect.

The protection available through what are now called credit default swaps, in which a lender sells off default risk and, more recently, baskets of credit risk from multiple bank loans and bonds, is regulated differently than insurance. More important to investors, these derivatives trade on an ongoing basis, so they can go long or short with a goal of making more money or reducing their risk. The International Swaps and Derivatives Association sets contract standards under which the buyers and sellers agree on what type of event is treated as a default, triggering payments from one counterparty to another. Banks that hold corporate loans and large lines of credit to corporations often sell off parts of their risk in pieces. Investors buy partly because, unlike generic equity offerings available through exchanges such as the New York Stock Exchange, derivatives can be tailored to particular institutions' requirements.

The market's initial growth, however, was triggered by the world's recent economic problems. A major reason for the timing is that speculative-grade corporate default rates reached their highest level since the Great Depression and made investors more sensitive to credit risk, says Duffie. Pending changes to banking regulations are also an impetus. International rules known as the Basel 2 Accords are slated for implementation in 2006. They would require banks in participating countries to set aside capital for unexpected losses while continuing to set aside cash to cover anticipated losses.

Most banks will apply a new standard formula, but those with their own credit risk modeling capabilities—the largest investment and commercial banks—might be able to reduce their capital requirements by developing quantitative models of their credit risks, says Duffie. Since there is no final agreed-upon formula yet for the amount of capital that banks would be required to hold as backing for the credit risk in their loan portfolios and other lines of business, the area is a rich subject for academic research. The trend has prompted executives of banks, investment banks, central banks, rating agencies, insurance companies, hedge funds, and pension funds to sign up for the Credit Risk: Pricing and Risk Management Executive Education course, next offered in April, that Singleton and Duffie codirect and teach. "Investors want to understand the latest developments in this market," Singleton says.

One of the hottest topics in investment banking now is how to allow for the added risk of correlated defaults in a portfolio; for example, the failure of a Japanese bank might trigger the default of a Korean derivatives broker in the same portfolio, he says. Alumni/ae in the field agree.

"There is a lot of interchange right now between practitioners and the academic community, and certainly the Stanford team is in the forefront of this," says Leonard Brous, MBA '89, the product manager for structured credit, which includes credit derivatives, at Morgan Stanley in New York, one of the banks where Duffie and Singleton have given recent seminars.

As customers become more sophisticated, they are part of the innovation process, says Singleton. "When customers approach investment banks for price quotes on new products, their research and risk departments have to figure out what to charge. It's a matter of survival and competition."

Some insurance and reinsurance companies have been criticized for taking on too much credit risk in the form of default swaps, but Duffie says they are often among the best positioned to do it. "By the nature of their businesses, some hold a lot of capital against large rare losses. If they sell protection on highly rated borrowers, they can make a hefty average risk premium and plan on incurring large occasional losses. This is not dramatically different from insurance against catastrophes, although insurance firms selling default protection would want to develop some degree of expertise in analyzing business risks, which are, of course, different than natural catastrophe risks."

"The users of credit derivatives have become very diverse," says Brous of Morgan Stanley. "There are no official statistics, but the kinds of players range from global commercial and investment banks to regional and foreign banks, hedge funds, insurance companies, fixed income mutual funds, and institutional investment managers, and more recently, big pension funds and endowments." With wider participation, the liquidity of credit derivatives continues to improve, he adds, and in many cases is now comparable to that of corporate bond and loan markets. Some institutional investors use credit swaps as a "proxy for corporate bonds when derivatives are available on more attractive terms and/or in the timeframe and quantity they want," Brous says. Because it takes time to assemble a fixed income portfolio, a money manager may decide to use the most liquid derivative products like Trac-XSM—a series of credit derivative index products created and sponsored by Morgan Stanley, JP Morgan, and other banks—to park cash temporarily while shopping for other more permanent investments.

Eventually, says Duffie, credit swap index funds are likely to appear in a form similar to bond index funds that companies like Vanguard and Fidelity could market to individuals seeking a fixed income fund. "An index fund of credit derivatives offers roughly the same kind of diversified portfolio risk, but the credit derivatives market is becoming so much more liquid that it probably will be the more efficient way to do it," Duffie predicts.

He explained the liquidity advantage this way:

"Suppose you are a broker. I call you and say I would like to buy corporate bonds of Italia Telecom, the Italian phone company. You might say, 'Well, I don't have any of that particular bond right now, but I'm sure I can find some in a day or two.' If, instead, I call and say that I would like to buy the same effective exposure to Italia Telecom, you'll probably say, 'Sure, no problem.' You don't need to have the bond to sell me the derivative, although you may want to lay off your risk with some other customer, and I don't need to have $50 million in my money market account today to buy it because I don't pay anything in the beginning. In effect, I get paid over time for accepting the risk of default." It is in this sense that a default swap is called an "unfunded'' product.

Some critics of credit risk markets point to the lack of transparency in the market as the biggest untraded risk. Since there is no way for outsiders to know how much risk any given institution has taken on, the fear is that a large investment bank or reinsurance company could experience correlated defaults and cause their own failure, according to one article in Investment Dealers Digest. Famous investor Warren Buffett of Berkshire Hathaway is among those who have expressed concern. On the other side, Federal Reserve Chairman Alan Greenspan lauded derivatives in general last year for making financial institutions "less vulnerable to shocks from underlying risk factors" and "the financial system as a whole … more resilient." From his free-market economist's perspective, a more flexible market for dividing up and trading the risks of currency or interest rate fluctuations or defaults on loans means that the institutions that can most afford to carry the risk are likely to free others of the burden—for a price, of course.

"Given the potential informational asymmetries in credit markets, there could be significant events that surprise many market participants," Singleton says, "but the more sophisticated market participants with some of the largest exposures have introduced risk management systems that allow them to 'stress' their portfolios. So a key question is, are they designing stress tests in the directions from which the next major credit surprises will arrive?" He adds that "the default in Russia shook up the market and demand for credit derivatives dropped off for a while, but the contracts largely served their intended purposes and investors have come back. I'm not in the camp that believes there is an enormous pool of risk that people are completely unaware of."

Adds Duffie: "All investors have an incentive to gather information to get that default swap rate right, and the prices we see are telling us that they are paying attention."

Choice in the Marketplace

Consumers Don't Always Know What They Want

Economic Incentives Inspire New Products

Credit Risk Trading

Leadership: The Art of Rustproofing a Household Icon

Strategy: Rescuing a Part of American History

Collaboration: An Entrepreneurial Journey into the Desert

Researchers Explore Credit Risk Measurement, Pricing

Finance faculty members Darrell Duffie, the James Irvin Miller Professor of Finance, and Kenneth Singleton, the C.O.G. Miller Distinguished Professor of Finance, share their work on credit risk modeling at seminars with institutions in the world's investment capitals and in an executive education course, to be offered April 18-23. They are authors of the book Credit Risk: Pricing, Measurement, and Management, published last year by Princeton University Press. Here is an overview of their recent work.

DARRELL DUFFIE: We deal with two basic types of problems. The first is measuring the credit risk of loaning money to a corporation or country, and the second is pricing. Measurement has two sub-problems: What is the probability of a default—how do you estimate it?—and, in the event of default, how much will you lose?

In the past few years we have learned that default losses are variable not only on a case by case basis but also as the economy changes. Ed Altman at New York University is the grandfather of this area, and his numbers indicate that during 2001 and 2002, the recoveries in defaults were far lower than the longtime average—about half! Perhaps losses were much higher during this period because a lot of distressed assets were coming onto the market at once.

A follow-on dimension is how these risks correlate across different counterparties in the market; so if, for example, you have loans to a Korean bank and a Japanese broker, are those risk correlated? If so, how much? That's one class of problems, which is measuring credit risk.

The other issue is pricing, which Ken and I have been passionately interested in for a few years. Not just how much should you charge outright on a loan to reflect the credit risk but also how much you charge on a credit derivative that protects you against that risk. This includes derivatives with options embedded in them and derivatives that give you exposure to the first to default among multiple names (companies or countries). Here you face the extra pricing dimension: You have to know how much of a risk premium to add to compensate an investor for carrying that risk of correlated default.

In my recent research, I was surprised to find that risk premiums are as big as they are, and I was pleased to see that they change with where we are in the business cycle. They also change based on the sector. For example, if you go to the health care sector, they are much bigger than in the oil and gas sector. You might say that doesn't sound right-oil and gas is really risky. But it's the premium per unit of risk. That is, for every additional 100 basis points [1 percent] of default probability, how much more are you getting in extra return? My preliminary estimates are it is about 50 percent more for the health care industry.

We used to think that losses in the event of default didn't change much across time either, but the annual volatility that I measured in my study is about 100 percent per year. The typical probability of a firm defaulting in a year is under 1 percent, so when I say it varies on a percentage basis by 100 percent a year, it might be going from 0.7 percent this year to 1.4 percent or 0.35 percent next year. It varies a lot with stages in the business cycle.

KEN SINGLETON: Finance research has established that the distribution of interest rates changes with different stages of the business cycle, which is intuitive. That is, whether the economy is in a recession or an expansion influences how yields on bonds of different maturities behave relative to each other and to the underlying macro economy. With some coauthors we have developed a way to integrate these changing distributions into a pricing framework that investment banks can use, not only to price bonds but also derivatives. Our framework allows the volatilities, correlations, and persistence in bond yields to change with the stage of the business cycle. Investment bankers have long recognized these risks, but there has not been a tractable means of incorporating them into pricing models or associated risk management systems. We are attempting to develop a framework that will accommodate shifting distributions as conditions change.

We are in the preliminary stages of empirical analysis, but it appears that market participants do "price" the risks of regime changes into bonds. That is, the market's aversion to these risks is reflected in bond prices. When we look at how these risks move over the cycle and how market participants' attitudes toward the risks change at different stages, an interesting question is to what degree those changes are related to the conduct of monetary policies or underlying macroeconomic sources of risk such as changing output growth or inflation. It's one of the next phases of this research project.

This same conceptual framework is potentially useful for modeling correlated defaults, for it captures the possibility that economic conditions could precipitate a significant deterioration in credit across several counterparties, roughly all at the same time. Moreover, the likelihood of this happening can depend both on economywide and industry-specific factors as well as measures of market liquidity. We hope to explore these ideas in future research.

Finally, I am currently exploring the determination of prices in credit default swap markets; in particular, how do the default swap spreads for different names move relative to each other and why? For instance, under what circumstances does the event of economic or financial distress in one country spill over to the prices of bonds and credit default swaps for other countries? Such spillovers are often referred to as "contagion" effects, but their economic underpinnings are often not what I would call a contagion phenomenon. I am exploring ways of incorporating these spillover effects into models for pricing sovereign bonds and default swaps.

Credit Risk: Pricing, Measurement, and Management
Darrell Duffie and Kenneth J. Singleton, Princeton University Press, 2003