To meet regulatory requirements, banks and other financial institutions regularly measure the total credit risk in their debt portfolios. For internal risk management, however, measuring overall risk is just a first step in understanding how a portfolio’s elements contribute to its risk. By decomposing total risk into risk contributions for each sector, region, line of business or creditor, banks can better manage their risk and make more informed risk-return tradeoffs. But what is the best way to break down overall portfolio risk?
Paul Glasserman studied this problem in a project funded by the Federal Deposit Insurance Corporation (FDIC) and the National Science Foundation. His work has led to better methods for measuring marginal risk contributions — the part of a portfolio’s overall credit risk that can be attributed to each component.
“The industry has developed general principles on which portfolio risk should be decomposed,” says Glasserman, “but actually determining the risk contributions can be difficult in complex portfolios. My FDIC paper develops a practical method for measuring risk contributions in the setting of the industry-standard model for portfolio credit risk.”
A portfolio view of credit risk is essential in comparing the risk and return of individual transactions. Consider a bank’s decision to extend credit to an airline. A naïve analysis would view the transaction in isolation and simply balance the airline’s creditworthiness against the interest and fees the bank could charge. A portfolio view asks how the transaction contributes to the risk and return of the bank’s overall portfolio. The transaction may contribute more risk — and thus appear less attractive — if the bank has already made loans to several other airlines.
Banks gauge portfolio credit risk through the concept of economic capital, a measure of the buffer of capital required to cover losses resulting from defaults. In order to calculate economic capital for a large portfolio of corporate debt, a bank must capture correlations in the creditworthiness of the many creditors to which it is exposed. The process is further complicated by the fact that defaults of highly rated firms are rare events and are thus difficult to predict.
Working with a former doctoral student, Jingyi Li, now at Credit Suisse First Boston, Glasserman developed an efficient Monte Carlo technique that addresses these challenges. “Monte Carlo techniques are widely used in the financial industry for pricing derivative securities and measuring risk,” notes Glasserman. “They take advantage of computing power to capture far more complexity and realism than is possible using simple models.” The Glasserman-Li method identifies scenarios that are most likely to lead to large losses from defaults and then systematically explores those scenarios.
Glasserman’s research on portfolio credit risk builds on derivatives pricing techniques that he developed in the late 1990s with a grant from IBM. More recently, he has focused on credit risk and credit derivatives, areas that have seen dramatic developments both in industry and academic work in just the past few years. His work on decomposing portfolio credit risk applies those two earlier strains of research in a new context.
In work he presented at a Moody’s credit risk conference in London in May 2005, Glasserman showed that related ideas could be used to dramatically improve precision in measuring the marginal risk contributions of portfolio components. His approach separates factors that influence credit risk across broad sectors or regions from idiosyncratic sources. He shows that the key determinants of marginal risk contributions are the individual default probabilities in those scenarios in which the common factors produce higher overall default rates.
“The field of credit risk management and credit derivatives pricing continues to grow rapidly,” Glasserman says, “so the industry hasn’t matured yet to the point where banks are satisfied with the methods they’re using. This is the phase in which academic research can have the greatest influence over how the industry develops, and the industry is quick to absorb valuable new ideas.”
Paul Glasserman is senior vice dean and the Jack R. Anderson Professor of Business at Columbia Business School. In 2004, he received the Wilmott Award for Cutting-Edge Research in Quantitative Finance.
Professor Glasserman's research and teaching address risk management, derivative securities, Monte Carlo simulation, statistics and operations. Prior to joining Columbia, Glasserman was with Bell Laboratories; he has also held visiting positions at Princeton University, NYU, and the Federal Reserve Bank of New York. In 2011-2012, he was on leave from Columbia and working at the Office of Financial Research in the U.S...