Margin Setting to Short and Long Futures Contract Positions by Coherent Risk Measures

This study, using gold coin spot price returns, in the period from 2008 to 2016, estimates and compares IME gold coin futures contracts short and long positions initial margin by coherent risk measures, specially Expected Shortfall and spectral risk measures such as Exponential weighting Function and Power weighting Function. GARCH, EGARCH and GJR GARCH models used for volatility process modeling. Fore models back testing, it applies Christoffersen conditional coverage likelihood ratio (LRcc) test and for models rating used lopez second loss functions and Blanco-Ihle loss functions, and Fore ES models evaluations uses MAE and RMSE loss functions. The paper finds that, GJRGARCH has outperformed the other models that support the asymmetric response of gold coin price to positive and negative shocks. The average margin quantity estimated for short positions with all risk measures, is significantly larger than long positions margin, that confirm asymmetric response of gold coin price to positive and negative shocks.