基于分位点回归和影响因子的动态保证金率

Dynamic margin rate based on quantile regression and impact factors

保证金是金融市场上对交易各方履约的基本保证.为了能够在风险可控的范围内设定更为合理的保证金水平,本文结合VaR方法中的GARCH、TGARCH、EGARCH模型,加入影响因子和基于MCMC的分位点回归方法来拟合模型参数,并利用Cornish-Fisher展开式估计分位数Φ-1q,以找到制定保证金水平更为合理的方法,然后采用沪胶指数日线为数据进行了对比和验证.

Abstract.. Margin is a basic guarantee for the performance of the parties to the transaction on the futures market. To set more reasonable margin levels in the range of risk control, we combine GARCH, TGARCH, EGARCH model in VaR approach, together with impact factors and estimating model pa-rameters based on MCMC quantile regression method, evaluate the quantile φq-1 by using Cornish-Fisher expansion,and use daily data of Hujiao index, strive to develop a more rational approach through com- parison and verification.