New approach to VAR estimation using pearson's type IV distribution

Abstract

Abstract- This paper constructs a robust Value at Risk (Var) estimation model for equity return time series and tests it extensively on ten South East Asian Stock Indices. Two major characteristics of such series are volatility clustering and non normality owing to fat tails of the return distribution. While volatility dynamics phenomenon has been extensively studied using the GARCH model and its many relatives, the phenomenon of non-normality of the residuals of the GARCH model has not been comprehensively explored. Extreme Value Theory (EVT) has been used to estimate the unexpected losses due to extreme events and hence modify the current methodology of VaR. A combination of EVT and GARCH has also been explored to analyze financial data showing non-normal behavior. This paper explores a combination of the Pearson’s Type IV distribution and the GARCH (1, 1) approach to come up with a robust model with enhanced predictive abilities. A robust model would obviate the need for imposing special ad hoc margins by the regulator in times of extreme volatility. A rule based margin system would increase efficiency of the price discovery process and also the market integrity with the regulator no longer seen as managing volatility. Index Terms- Dynamic VaR, GARCH, Pearson’s Type IV Distribution