In sample value at risk and backtesting (Scripts) Publisher's description
I wrote this code because I wanted to incorporate a new distribution in the model
I wrote this code because I wanted to incorporate a new distribution in the model, not availiable in Matlab or OxMetrics. In this sample the input is a return series, it uses a Garch(1,1) model with a constant in mean in the returns, and calculates the maximum likelihood estimator (mle) for the gaussian distribution. Using fmincon with appropriate constraints it minimizes the mle. Then it plots the VaR levels for six cases and calculates the success/failure ratio, the Kupiec-p test and the Christoffersen consitional coverage. The results are exact with OxMetrics, and the user can change the code with his own Garch model or a different distribution at will. The output is a structure with the results.
System Requirements:MATLAB 7.11 (2010b)
Program Release Status: Major Update
Program Install Support: Install and Uninstall