A new study co-authored by Professor Turalay Kenc, a Pembroke Visiting Scholar at Cambridge Judge Business School, has developed a new and more accurate model to measure the risk of bank defaults during highly volatile periods such as experienced during the financial crisis of 2007-2009.
The study is based on the GARCH option pricing model for the time variation of volatility, but with richer dynamics that improve default-risk measurements. (GARCH is the acronym for the generalised autoregressive conditional heteroskedastic process used for modeling volatility in time series.)
The paper is based on a final sample of 138 US banks with total assets exceeding $1 billion at the end of 2006, representing more than 73 per cent of all commercial bank assets at that time. Of those 138 banks, 42 met the criteria for “distress” based on failure, below investment-grade credit rating, or sharp fall in share price coupled with dividend suspension.
The research paper – entitled “Default indicators with volatility clustering” – is co-authored by Professor Turalay Kenc, Dr Emrah Ismail Cevic of Namik Kemal University in Turkey, and Professor Sel Dibooglu of the University of Missouri, St. Louis.
Professor Turalay Kenc is former Deputy Governor of the Central Bank of Turkey, and has been a Pembroke Visiting Scholar at Cambridge Judge during the current academic year, working closely with the School’s Finance subject group.
Pembroke Visiting Scholars have been welcomed at Cambridge Judge since 2012 to allow senior faculty active in finance to visit the business school for periods up to six months. Holders of the post are admitted to Pembroke College, one of the Colleges of the University of Cambridge, where they join in meals and other events at the college, and have use of an apartment rent-free for the period of their residence.
Turalay Kenc discussed some of the paper’s findings and its broader context:
- The financial crisis showed the need for more accurate predictors of bank default risk. There were only 10 bank failures in the US in the five years prior to 2008, but 465 failed banks were closed by US regulators from 2008 to 2012. There’s a real need to more closely monitor the default risk of financial institutions to prevent future crises in advance.
- Existing models for measuring default risk have largely been based on the assumption of constant volatility. The financial crisis of a decade ago showed that volatility can be anything but constant, and the models failed to account for the large shifts in default risk caused by the strong volatility at that time. What we sought to do in our study was to maintain the practicality of the GARCH formula, while employing an approximate pricing mechanism to minimise certain numerical dimensions.
- This type of estimation greatly improved default measures during this period of heightened volatility. The distance to default measure (a widely used measure of corporate default risk) for the study’s 138 sampled banks ranged as high as 28 per cent during the September-December 2008 period using the estimation, compared to nine per cent using the industry standard model. That greater predictive power could prove important in helping to avert future crises.
- There’s scope for further research using our model. While our study looked at individual banks, it also would be useful to capture the systemic component of default risk throughout the broader banking system.