Estimating market risk under a wavelet-based approach: Mexican case

This paper aims to estimate value-at-risk and beyond the VaR for the main Mexican stock index (IPC) under a wavelet-based approach. The wavelet approach is used since financial time series have shown to be non-stationary and in most of the times non-normally distributed, and it allows to face the problem when signals have very high frequency components with short time spans, and low frequency components with long time spans. The IPC variance is decomposed into different time-scales using the discrete wavelet transform (DWT) and the least asymmetric (LA) Daubechies wavelet filter. Results show that the decomposed risk measure cannot be rejected at 95% and 99% confidence level.