Cyclicality in Catastrophic and Operational Risk Measurements

51 Pages Posted: 18 Feb 2004

See all articles by Turan G. Bali

Turan G. Bali

Georgetown University - Robert Emmett McDonough School of Business

Linda Allen

City University of New York, Baruch College - Zicklin School of Business - Department of Economics and Finance

Multiple version iconThere are 5 versions of this paper

Date Written: February 1, 2005

Abstract

A natural point of departure for all elements of business risk measurement is the past. Future trends and current metrics are often extrapolated from an historical data series. However, this process is fundamentally flawed if there are cyclical factors that impact business measures of risk or performance. Historical data on operational risk gathered during an economic expansion may not be relevant for a period of recession. Estimates of default risk and recovery rates incorporate cyclical components that are correlated to systematic risk factors, such as macroeconomic fluctuations and regulatory shifts. All too frequently, however, researchers and practitioners alike ignore these cyclical factors and blithely extend an unadjusted trend line into the future. The metrics obtained using this methodology are fundamentally flawed. By aggregating across different macroeconomic regimes, these historical estimates do not accurately reflect either time period. It is the goal of this paper to demonstrate the importance of developing models to adjust for systematic and cyclical risk factors in business metrics. This is the first paper, to our knowledge, to test the cyclicality of catastrophic and operational risk measures. Using equity returns for financial institutions we estimate both catastrophic and operational risk measures over the period 1973-2003. We utilize an extreme value approach (Generalized Pareto Distribution, GPD), as well as a generalized distributional approach (Skewed Generalized Error Distribution, SGED) to obtain estimates of catastrophic risk parameters and 1% value at risk (VaR). We find evidence of procyclicality in the catastrophic VaR for financial institutions. We define a new, residual operational risk measure and estimate the risk parameters using both the GPD and SGED. We use these operational risk parameters to determine the 1% operational VaR. Using our measure, we find that operational risk is quite significant, comprising approximately 18% of the total equity returns of financial institutions. This paper presents the first evidence of procyclicality in operational risk measures. Our results are robust to alternative distributional specifications, conditionality in downside risk measures, and simulated databases. Thus, we conclude that macroeconomic, systematic and environmental factors play a considerable role in influencing the risk of financial institutions. Models that ignore these factors are therefore fundamentally flawed. These results provide encouragement for further research into both catastrophic and operational risk measures that are conditioned on cyclical factors.

Keywords: operational risk, catastrophic risk, value at risk, extreme value theory, skewed fat tailed distribution.

JEL Classification: G20, C13

Suggested Citation

Bali, Turan G. and Allen, Linda, Cyclicality in Catastrophic and Operational Risk Measurements (February 1, 2005). Available at SSRN: https://ssrn.com/abstract=501003 or http://dx.doi.org/10.2139/ssrn.501003

Turan G. Bali

Georgetown University - Robert Emmett McDonough School of Business ( email )

3700 O Street, NW
Washington, DC 20057
United States
(202) 687-5388 (Phone)
(202) 687-4031 (Fax)

HOME PAGE: https://sites.google.com/a/georgetown.edu/turan-bali

Linda Allen (Contact Author)

City University of New York, Baruch College - Zicklin School of Business - Department of Economics and Finance ( email )

17 Lexington Avenue
New York, NY 10010
United States
646-312-3463 (Phone)
646-312-3451 (Fax)

HOME PAGE: http://stern.nyu.edu/~lallen

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
752
Abstract Views
3,083
rank
21,330
PlumX Metrics