This code implements the Dynamic Binary (DB) backtest based on non-linear regression models to test for the conditional coverage hypothesis in VaR forecasts. Several specifications are considered for the DB test, including the lagged violations and the lagged VaR. For comparison reasons the DQ (Engle Manganelli, 2005) and LRCC (Christoffersen, 1998) tests are also presented. The code requires a sequence of VaR violations (binary indicator, 1 if violation, 0 otherwise), and the sequence of VaR for the α coverage rate. The number of lags to be used in the test and the coverage rate at which the VaR has been computed are selected in a further step.
Oxford Bulletin of Economics and Statistics (2022)
Matsuoka Hideaki