Primary navigation:

QFINANCE Quick Links
QFINANCE Topics
QFINANCE Reference

Home > Asset Management Best Practice > Forecasting Default Rates and the Credit Cycle

Asset Management Best Practice

Forecasting Default Rates and the Credit Cycle

by Martin S. Fridson

Executive Summary

  • The benefit to corporate bond investors of staying a step ahead of the credit cycle has stimulated interest in models for forecasting one of the cycle’s best markers, the default rate.

  • A market-based forecasting model can complement actuarial and econometric models, which have inherent limitations.

  • This article describes a market-based default-rate forecasting model, based on the distribution of outstanding high-yield bonds between the distressed and non-distressed categories, and the respective, historical default rates of those categories.

  • The actual default rate tracks the market-based model’s year-ahead forecast fairly closely, although the forecast can overshoot under extreme market conditions.

The Nature and Importance of the Credit Cycle

Cycles play a major role in analysis aimed at achieving superior investment returns. Stock market participants base their valuations on corporate earnings, which fluctuate with the business cycle. The interest rate cycle strongly influences the performance of high-quality fixed-income assets, such as government bonds and mortgage-backed securities. Similarly, investors in corporate bonds, for which the risk of default is a material factor, can benefit from anticipating turns in the credit cycle.

At the beginning of the credit cycle, lenders perceive the risk of default to be low. They gladly extend loans even to low-quality borrowers, and accept small risk premiums (measured by yield differentials over risk-free rates). Inevitably, some borrowers incur more debt than they are able to support when the business cycle turns down. They consequently default on their obligations, which causes lenders to turn more conservative in their credit extension policies. As it becomes more difficult to borrow, other borrowers fail as a result of being unable to refinance their maturing debts. Finally, as the default wave subsides, lenders regain confidence and a new cycle begins.

The link between the comparative liberality of credit extension risk and premiums on debt is illustrated in Figure 1. In a quarterly survey conducted by the Federal Reserve, senior loans officers of major money center banks indicate whether they are currently raising or lowering the quality standards that corporate borrowers must satisfy to obtain loans. As banks make it harder to qualify for loans, the average risk premium rises in the investment grade corporate bond market.

Risk premiums, in turn, are closely connected with default rates. Figure 2 documents this linkage over the past two US credit cycles. The trailing 12-months default rate on speculative grade issuers reached cyclical highs in June 1991, January 2002 and November 2009. Corresponding to these peaks were the cyclical maximum points of the option-adjusted spread on the Merrill Lynch High Yield Master II Index, in January 1991, June 2002 and December 2008.

Approaches to Forecasting the Default Rate

One outgrowth of investors’ interest in understanding the credit cycle is an effort to develop a model for forecasting one of its best markers, the default rate.1 Credit market analysts have worked extensively on this problem since the early 1990s (see More Info for key articles). Three types of default rate forecasting models have emerged from the research—actuarial, econometric, and market-based.1

The actuarial approach derives from empirical data documenting the relationship between bond ratings and the historical incidence of default over stated periods. For example, based on statistics compiled for the period 1983–2008, Moody’s Investors Service reports that on average, issuers rated Aa had a 0.019% probability of defaulting within one year and a 0.247% cumulative probability of defaulting within five years. The corresponding figures for issuers rated Caa are 13.730% and 43.747%. Actuarial models apply the rating-specific default rates to the distribution of speculative grade issuers within the rating categories (Ba, B, Caa, and Ca-C), to generate forecasts of the default rate for the speculative grade category as a whole.

A fundamental limitation of the actuarial method is that rating-specific default rates vary substantially from year to year, as a function of variation in economic and credit market conditions. For instance, the B category had a 1.983% default rate in 1997 and a 9.340% default rate in 2001. Actuarial models do not capture this effect, although they typically take into account another period-related variance, namely that an issue’s probability of default within a given year is partly a function of the number of years elapsed since issuance. (The curve rises for the first three to four years, then declines thereafter.)

The econometric approach models the speculative grade category’s default probability as a function of several variables. These may include indicators of aggregate economic activity, for example, interest rates, measures of credit market conditions, and the variables employed in actuarial models. Generally, the economic indicators employed in such models are forecast, rather than historical, variables. Accordingly, the accuracy of the default rate forecast depends on the accuracy of the forecasts of such items as gross domestic product (GDP) and factory utilization. To put it mildly, errors are not uncommon in macroeconomic forecasting.

Back to Table of contents

Further reading

Books:

  • Fridson, Martin, M. Christopher Garman, and Sheng Wu. “Real interest rates and the default rates on high yield bonds.” In Theodore M. Barnhill, Jr, William Fr. Maxwell, and Mark R. Shenkman (eds). High Yield Bonds: Market Structure, Portfolio Management, and Credit Risk Modeling. New York: McGraw Hill, 1999; pp. 164–174.
  • Moyer, Stephen G. Distressed Debt Analysis: Strategies for Speculative Investors. Boca Raton, FL: J. Ross Publishing, 2005.

Articles:

  • Fridson, Martin S., Kevin P. Covey, and Karen Sterling. “Performance of distressed bonds.” Journal of Portfolio Management 34:3 (Spring 2008): 56–62. Online at: dx.doi.org/10.3905/jpm.2008.706243
  • Helwege, Jean, and Paul Kleiman. “Understanding aggregate default rates of high yield bonds.” Journal of Fixed Income 7:1 (June 1997): 55–61. Online at: dx.doi.org/10.3905/jfi.1997.408202. (Also published in Current Issues in Economics and Finance (May 1996): 1–6. Online at: tinyurl.com/65mqnwo)
  • Jónsson, Jón G., and Martin S. Fridson. “Forecasting default rates on high yield bonds.” Journal of Fixed Income 6:1 (June 1996): 69–77. Online at: dx.doi.org/10.3905/jfi.1996.408166

Reports:

  • Fons, Jerome S. “An approach to forecasting default rates.” Moody’s Special Report. August 1991.
  • Keegan, Sean C., Jorge Sobehart, and David T. Hamilton. “Predicting default rates: A forecasting model for Moody’s issuer-based default rates.” Moody’s Special Comment. August 1999. Online at: ssrn.com/abstract=1020303
  • Metz, Albert, and Richard Cantor. “A cyclical model of multiple horizon credit ratings transactions and default.” Moody’s Investors Service. August 2007.

Back to top

Share this page

  • Facebook
  • Twitter
  • LinkedIn
  • Bookmark and Share