Parametric and non-parametric estimation of value-at-risk

Deepak Jadhav and T. V. Ramanathan

Abstract

ABSTRACT

Value-at-risk (VaR) is one of the most common risk measures used in finance. The correct estimation of VaR is essential for any financial institution, in order to arrive at the accurate capital requirements and to meet the adverse movements of the market. We give a brief review of all of the existing parametric and non-parametric methods of estimating VaR. We have introduced some new non-parametric estimators for VaR. Comparison between these estimators are made using in-sample and out-of-sample backtesting techniques. It is found that one of the newly suggested nonparametric estimators works well compared with others, specifically for return data with high variability.