Comparative Empirical Study on the Margin Setting of Stock Index Futures Calendar Spread Trading

Using the data of Hang Seng index futures, this empirical research investigates exponentially weighted moving average (EWMA) method, auto regression moving average-exponential generalized autoregressive conditional heteroscedasticity (ARMAEGARCH) model, and extreme value theory (EVT) to provide a scientifically prudent and practically available approach for the margin setting of calendar spread trading in China. The results show that the above three models have their own advantages and disadvantages, taking into account the stability and practical feasibility. EWMA is simple in calculation and easy to be put into practice, but it may lead to an underestimation of the market risk because of the inaccurate decay factor in the model. EVT is prudent but not easy to be widely implemented because of the requirement for chronic data accumulation. ARMA-EGARCH has the best performance in both accuracy of risk estimation and feasibility of practical implementation. To a certain extent, the research provides both theoretical and empirical support for the margin setting of the coming CSI300 index futures.