On February 21, 2012, the Journal of Investment Management announced the winners of the Harry M. Markowitz award for the best paper of 2011 in the Journal of Investment Management. Four Nobel Prize winners, including Robert C. Merton, awarded the prize to John Y. Campbell (Harvard University), Jens Hilscher (Brandeis University and Senior Research Fellow at Kamakura Corporation) and Jan Szilagyi (Hawker Capital LLP) for their paper “Predicting Financial Distress and the Performance of Distressed Stocks.” More than any other event, this award marks the definitive end to the debate about whether or not reduced form models are superior to the Merton model of risky debt. This blog explains why.
For more than a decade, practitioners have debated whether or not the reduced form approach to modeling default was superior to the 1974 Merton model of risky debt in predicting default. The Harry M. Markowitz award to Campbell-Hilscher-Szilagyi for their reduced form default model and its applications to the equity market is definitive. Robert C. Merton himself, joined by fellow Nobel Prize winners Harry M. Markowitz, Myron S. Scholes, and William F. Sharpe, concluded that the CHS paper was the best paper of 2011 in part for its evidence that “our model is… between 49% and 94% more accurate than [a Merton distance to default model].”
For bankers who have believed, without statistical proof, that the Merton approach is the most accurate approach to modeling corporate default, the award to CHS may come as a surprise. In the academic world, however, Professor Robert Jarrow commented that the debate about reduced form credit models versus the Merton model of risky debt ended years ago in the face of overwhelming statistical evidence that the reduced form approach was superior. Campbell, Hilscher and Szilagyi summarized the reason for the performance differential with this key quote in their paper “In Search of Distress Risk” (Journal of Finance, December 2008, page 2915): “If one’s goal is to predict failures, however, it is clearly better to use a reduced-form econometric approach [than the Merton approach] that allows volatility and leverage to enter with free coefficients and that includes other relevant variables.” CHS add, “Bharath and Shumway (2008), in independent recent work, reach a similar conclusion.”
Now, through the Harry M. Markowitz Award to CHS, Robert C. Merton and his three fellow Nobel Prize winners have formally agreed with these conclusions.
What steps should current users of legacy Merton default probabilities take in response to this confirmation of more than a decade of voluminous statistical proof? We suggest these steps:
Step 1: Follow the advice of John Maynard Keynes, who was quoted by Malabre as saying “When the facts change, I change my mind. What do you do, sir?”
Step 2: Contact Kamakura Corporation at info@kamakuraco.com for an immediate start of a trial of the Kamakura Risk Information Services reduced form default probability service, which covers 31,000 public firms in 37 countries, 183 sovereigns, and non-public firms. The KRIS service includes the Merton model as well, so those who retain an affection for the model, despite its flaws, need not be disappointed.
Step 3: Transition away from legacy rating agency products as rapidly as possible.
We close by pointing the reader to helpful links on this topic:
For the original press release by the Journal of Investment Management, see this link:
https://www.joimconference.com/markowitz/markowitz_pr_022112.pdf
For the press release by Kamakura Corporation on the award to Prof. Hilscher, see this link:
http://www.kamakuraco.com/February282012PressRelease.aspx
For a copy of the award-winning 2011 paper by Campbell, Hilscher and Szilagyi, use this link:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1829622
For key statistical studies of credit model accuracy, see these selected papers from the CHS bibliography:
Bharath, S. and Shumway, T. (2008). “Forecasting default with the Merton distance to default model,” Review of Financial Studies 21, 1339–1369.
Campbell J.Y., Hilscher, J., and Szilagyi, J. (2008). “In search of distress risk,” Journal of Finance 63, 2899–2939.
Chava, S. and Jarrow, R.A. (2004). “Bankruptcy prediction with industry effects,” Review of Finance 8, 537–569.
Hilscher, J. and Wilson, M. (2009). “Credit ratings and credit risk,” unpublished paper, Brandeis University and Oxford University.
Shumway, T. (2001). “Forecasting bankruptcy more accurately: A simple hazard model,” Journal of Business 74, 101–124.
Key credit risk model accuracy studies are available from Kamakura Corporation, the most detailed of which are the Kamakura Risk Information Services Technical Guides associated with each version of the Kamakura models. The most recent KRIS Technical Guide (Version 5, September 2010) is authored by Robert A. Jarrow, Sean P. Klein, Mark Mesler and Donald R. van Deventer. It is available to KRIS clients and to regulatory bodies worldwide upon signing of a confidentiality agreement.
Donald R. van Deventer
Kamakura Corporation
Honolulu, February 29, 2012
© Copyright 2012 by Donald R. van Deventer. All rights reserved.