A Note About Calibration Tests for VaR and ES

This paper assesses the performance of several calibration tests developed in the recent years to evaluate the forecasting ability of two risk measures commonly used in practice, namely, the value-at-risk (VaR) and expected shortfall (ES). The assessment is made through Monte Carlo experiments on models, with the most popular class of volatility models. The size and power of every test are evaluated under different scenarios considering both normal and Student-t innovation distributions. We also illustrate the application of the calibration tests to two time series of daily financial returns.