Abstract. Risk managers use various types of techniques to estimate different kinds of risk and ways to minimize its impact. VaR which stands for Value at Risk is one of those techniques. Various new methods for calculation of VaR have been developed. In this study, four techniques of VaR estimations have been employed: i) Historical Simulation; ii) Variance Covariance Approach; iii) Monte Carlo Simulation, and iv) AR-GARCH method. The purpose of this study is to compare the different VaR estimation methods and draw conclusions based on the Back- Testing methods.

As per the analysis, historical method proved to be the best method for estimating value at risk. This method is widely preferred by risk managers and practitioners in the banking sector. Though the portfolios used in the study was diversified and contained stocks from different sectors, still the historical simulation method came out to be on the top as it was accepted for all the four portfolios. This method does have some limitations as the patterns generated from the past data may not hold true all the time.

Keywords: Value-at-Risk (VaR); simulation model; variance-covariance matrix; Monte Carlo simulation; GARCH approach.