Epanechnikov Kernel Estimation of Value at Risk

Abstract

Anything that deviates from the normal is termed as risk. This definition looks simple but in real sense breaking it down into components is the most difficult thing. Analysis of what is “normal” and what is “abnormal” and also the measure for deviation is what researchers are exploring for years. Over decades different measures for risk have been put up and this area of study is always evolving. One way of quantifying this measure of risk is the VaR (Value at Risk) methodology suing historical simulation method which is what this paper tries to stress on. Risk analysis is generally undertaken on be making assumptions of the distribution of the base element of it. But this paper analyses the non-parametric method of VaR estimation using the kernel approach of historical simulation specifically using the Epanechnikov Kernel.