典型事实约束下的沪深300股指期货动态保证金设定研究——基于APARCH-GPD模型的VaR度量

A Study on Dynamic Margin Setting for CSI300 Futures Under Stylized Facts:The VaR Prediction Based on APARCH-GPD Model

本文提出了一种新的设定股指期货动态保证金水平的APARCH-GPD模型。它结合了APARCH模型良好拟合收益序列集丛性和尖峰厚尾分布的能力,以及GPD分布充分拟合尾部残差的特点,可提供准确的VaR风险度量。实证结果表明,使用该模型估计沪深300股指期货的保证金水平,其风险覆盖效果明显优于APARCH-norm模型、APARCH-t模型和APARCH-GED模型。研究还发现,目前股指期货市场的保证金水平偏高,具有下调空间,且空头头寸面临的价格波动风险要小于多头,因此可以对不同头寸设定差异化的保证金水平。

APARCH-GPD model has been proposed in this paper as a new way to set dynamic margin of stock index futures. It combines the advantage of APARCH model to measure the time-varying volatility and leptokurtic distribution of returns, and al- so the advantage of GPD to model the tail distribution of the APARCH processed annotations. In order to provide an accurate measure of VaR, we use APARCH-GPD model to estimate the margin levels of CSI 300 futures, which achieves better perfor- mance of risk coverage than APARCH-norm model, APARCH-t model and APARCH-GED model. From the empirical evi- dence, we also find that the current margin level of CSI 300 futures market is overrated and therefore could be decreased, and the price volatility risk on short positions is less than that on long positions. Thus, differentiated margin levels could be set ac- cording to the different positions.