Reverse Stress Testing

Finance Professor Wins Best Paper Award for Creating Insightful, Novel Method of Risk Assessment

Finance professor Yaacov Kopeliovich and his RiXtrema research team colleagues have won the 2015 Peter L. Bernstein Award for Best Paper for their work titled, “Robust Risk Estimation and Hedging: A Reverse Stress Testing Approach.”

The article originally appeared in the Journal of Derivatives in May 2015. It was selected by a three-person review committee and was chosen from a pool of nominations from 11 top financial journals. The judges looked for an original or new approach to the field or subject of study; surprising and/or insightful results or implications; and both practical and academic relevance.

Yaacov Kopeliovich (Nathan Oldham/UConn School of Business)
Yaacov Kopeliovich (Nathan Oldham/UConn School of Business)

Kopeliovich conducted his research from 2013 to mid-2014 together with other research team members at a risk-management software vendor, RiXtrema Inc. in New York. His co-authors included Daniel Satchkov, president of RiXtrema; Barry Schachter of the Courant Institute of Mathematical Sciences at NYU; and Novosyolov Arcady of the Department of Mathematics at the Siberian Federal University in Russia.

Traditional risk modeling, using value-at-risk (VaR) is widely viewed as ill equipped for dealing with tail risks, Kopeliovich explained. As a result, scenario-based portfolio stress testing is increasingly being promoted by experts as central to the risk-management process. A recent innovation in portfolio stress testing and one that is endorsed by regulators, called reverse stress testing, is designed to identify economic scenarios that could threaten a financial firm’s viability. But the goal is to do so without injecting the manager’s cognitive bias into stress scenario specification.

While the idea is appealing, no ‘template’ has been provided to make that idea a reality, he said. Some first steps in developing ‘reverse stress testing’ approaches have begun to be written in financial literature. But complexity and computational intensity appear to be potential stumbling blocks, Kopeliovich said. Also, emerging research challenges the relationship among the concepts of likelihood, plausibility and representativeness.

In their paper, Kopeliovich and his colleagues propose a novel method for reverse stress testing, using a multivariate normal distribution and mixing principal components analysis with Gram-Schmidt applications to determine scenarios leading to a specified loss level.

“We came up with a new way to identify plausible scenarios for given loss level. This is a known problem in risk management, particularly these days when the Fed looks very carefully on possible events that can happen and hurt big financial institutions that are of systemic importance,” he said.

The judges included Gary Gastineau of ETF Consultants, Professor William Goetzmann of the Yale School of Management and Ronald Kahn of BlackRock. Kahn said the analysis in the article offers very practical benefits, including the potential to build better portfolios by identifying the most likely tail-risk scenarios. Robert Arnott, chairman and CEO of Research Affiliates, which funded the award, said the work strengthens the investment industry.

Kopeliovich is a professor-in-residence in the finance department and has taught courses in fixed income and financial risk management. He completed a Ph.D. in finance in 2014 from the EDHEC Business School in France. He also has a Ph.D. in mathematics, magna cum laude, from Hebrew University in Israel, as well as a masters in financial engineering from University of California in Berkeley.

His research interests include portfolio optimization for bonds and technical trading profitability. As winners of the Peter L. Bernstein Award, the authors will receive a $5,000 award to divide.