Regulatory Evaluation of Value-at-Risk Models

32 Pages Posted: 9 Nov 2006

See all articles by Jose A. Lopez

Jose A. Lopez

Federal Reserve Bank of San Francisco

Date Written: December 1997

Abstract

Beginning in 1998, U.S. commercial banks may determine their regulatory capital requirements for financial market risk exposure using value-at-risk (VaR) models i.e., models of the time-varying distributions of portfolio returns. Currently, regulators have available three hypothesis-testing methods for evaluating the accuracy of VaR models: the binomial method, the interval forecast method and the distribution forecast method. These methods use hypothesis tests to examine whether the VaR forecasts in question exhibit properties characteristic of accurate VaR forecasts. However, given the low power often exhibited by these tests, these methods may often misclassify forecasts from inaccurate models as accurate. A new evaluation method that uses loss functions based on probability forecasts, is proposed. Simulation results indicate that this method is capable of differentiating between forecasts from accurate and inaccurate, alternative VaR models.

Keywords: value-at-risk, volatility modeling, probability forecasting, bank regulation

JEL Classification: C52, G2, G15

Suggested Citation

Lopez, Jose Antonio, Regulatory Evaluation of Value-at-Risk Models (December 1997). FRB of New York Staff Report No. 33, Available at SSRN: https://ssrn.com/abstract=943498 or http://dx.doi.org/10.2139/ssrn.943498

Jose Antonio Lopez (Contact Author)

Federal Reserve Bank of San Francisco ( email )

101 Market Street
San Francisco, CA 94105
United States
415-977-3894 (Phone)
415-974-2168 (Fax)

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
714
Abstract Views
7,097
Rank
66,747
PlumX Metrics