Evaluating var models for the volatile Indian futures market Jitender School of Economics, University of Hyderabad, Hyderabad, India, Email id: jitendergame@gmail.com Online published on 19 December, 2023. Abstract The application of Value at Risk (VaR) methodology in the context of managing market risk has gained widespread acceptance as a standard for quantifying the inherent "market risk" associated with a range of financial instruments. This research seeks to assess the most suitable VaR model for the characterization of risk in the context of the Indian futures market, with a specific emphasis on Nifty futures. Through an examination of data covering the time frame from 2006 to 2015, a series of VaR estimations have been calculated, utilizing a variety of methodologies. These methodologies encompass parametric approaches such as variance-covariance, Exponentially Weighted Moving Average, and Generalized Auto-Regressive Conditional Heteroskedasticity, as well as non-parametric methods including historical simulation and Monte Carlo simulation. The analysis reveals that VaR models based on the normality assumption tend to provide underestimated risk estimates, particularly when dealing with returns that do not follow a normal distribution. In contrast, models that take into consideration the presence of fat-tail behavior in financial returns, such as historical simulation, exhibit better performance in accurately capturing the risk associated with these financial instruments. Top Keywords Value-At-Risk (Var), Nifty Futures, Variance-Covariance, Historical Simulation, Financial Risk Management. Top |