Value-at-Risk for Greek Stocks

Abstract

This paper analyses the application of several volatility models to forecast daily Value-at-Risk (VaR) both for single assets and portfolios. We calculate the VaR number for 4 Greek stocks, 2 portfolios based on these securities and for Athens Stock Exchange General Index (ASE). We model VaR for long and short trading positions by employing non-parametric methods, such as historical and filtered historical simulation, and parametric ones. Especially for the later techniques we use a collection of ARCH models (GARCH, EGARCH and TARCH) based on three distributional assumptions (Normal, Student-T and Skewed Student-T), while we combine the Extreme Value Theory with a volatility updating technique (via GARCH type-modeling). In order to choose one model among the various forecasting methods, we employ a two-stage backtesting procedure. In the first one, we implement two backtesting criteria (unconditional and conditional coverage) to test the statistical accuracy of the models. In the second stage, we employ standard forecast evaluation methods in order to examine whether the diferences between the models, which have converged suficiently, are statistically significant.