Backtesting VaR models: Quantitative and Qualitative Tests
Carlos Blanco and Maksim Oks
This is the first article in a two-part series analyzing the accuracy of risk measurement models. In this first article, we will present an overview of backtesting methods and point out the importance of conducting regular backtests on the risk models being used. In the second article, we will present an alternative to measuring VaR using a top-down or “macro” approach as a complementary tool to traditional risk methodologies.
Should risk models be accurate?
Firms that use VaR as a risk disclosure or risk management tool are facing growing pressure from internal and external parties such as senior management, regulators, auditors, investors, creditors, and credit rating agencies to provide estimates of the accuracy of the risk models being used.
Users of VaR realized early that they must carry out a cost-benefit analysis with respect to the VaR implementation. A wide range of simplifying assumptions is usually used in VaR models (distributions of returns, historical data window defining the range of possible outcomes, etc.), and as the number of assumptions grows, the accuracy of the VaR estimates tends to decrease.
As the use of VaR extends from pure risk measurement to risk control in areas such as VaR-based Stress Testing and capital allocation, it is essential that the risk numbers provide accurate information, and that someone in the organization is accountable for producing the best possible risk estimates. In order to ensure the accuracy of the forecasted risk numbers, risk managers should regularly backtest the risk models being used, and evaluate alternative models if the results are not entirely satisfactory.
VaR models provide a framework to measure risk, and if a particular model does not perform its intended task properly, it should be refined or replaced, and