Conditional autoregressive valu at risk by regression quantile: Estimatingmarket risk for major stock markets

This paper employs a new approach due to Engle and Manganelli (2004) in order to examine market risk in several major equity markets, as well as for major companies listed in New York Stock Exchange and Athens Stock Exchange. By interpreting the VaR as the quantile of future portfolio values conditional on current information, Engle and Manganelli (2004) propose a new approach to quantile estimation that does not require any of the extreme assumptions of the existing methodologies, mainly normality and i.i.d. returns. The CAViaR model shifts the focus of attention from the distribution of returns directly to the behaviour of the quantile. We provide a comparative evaluation of the predictive performance of four alternative CAViaR specifications, namely Adaptive, Symmetric Absolute Value, Asymmetric Slope and Indirect GARCH(1,1) models. The main findings of the present analysis is that we are able to confirm some stylized facts of financial data such as volatility clustering while the Dynamic Quantile criterion selects different models for different confidence intervals for the case of the five general indices, the US companies and the Greek companies respectively.