Could GARCH-VaR method measure mutual fund risk in post-crisis era China effectively

Abstract

In order to evaluate mutual fund risk in post-crisis era China, this pa- per constructs two VaR-GARCH models, and estimates the VaR of different mutual funds under t-distribution and generalized error distribution(GED) separately. Then by employing Kupiec back-testing meth- od, we test the accuracy of two VaR- GARCH models. It turns out that the VaR model under GED is better than the other one in reflecting mutual fund risk but nei- ther holds the marked back-testing effect. K eywords: mutual fund; VaR method; GARCH model; t-distribution; GED