Futures Contracts Margin Setting by General Pareto Distribution VaR

According to the significance of margining process in derivatives contracts in clearing houses, margin setting in futures contracts is noteworthy due to tradeoff between risk management and trading cost. Due to the extensive use of Value at Risk models in margin setting, this study, using gold coins futures price returns, in the period from 2008 to 2016, estimates IME gold coin futures contracts initial margin by parametric and non-parametric Value at Risk models, such as historical simulation, general pareto distribution and adaptive GPD models. For models back testing, it applies Christoffersen conditional coverage likelihood ratio (LRcc) test and Lopez and Blanco-Ihle loss functions. The paper finds that historical simulation model has been outperforming parametric models, as the fat tail empirical distribution of futures return in low confidence level. In high confidence level adaptive GPD had the proper results. As a result findings propose the use of these two models to margin setting in proper confidence level.