统计套利模型研究——基于上证50指数成份股的检验

A Model Study of Statistical Arbitrage——A Test Based on Constituent Shares of Index Shangzheng 50

本文借鉴协整的思想,并采用比协整回归更一般化的方法来研究股票之间的统计套利模型。采用逐步回归法来确定合适的定价子空间与证券组合,并将统计套利模型应用于上证50指数的50个成份股,并使用方差比分析来检验可预测性,其结果表明随机去势后的股票价格序列明显偏离随机游走,存在着可预测成分。联立方程模型的估计结果表明错误定价趋于在短期内形成趋势,而在更长时间内回复。样本外绩效对交易费用水平的变动非常灵敏,机构投资者的年夏普比为1.3。

We adopt the idea of Cointegration and apply the methods that are more generalized than cointegration regression to study the Statistical Arbitrage Models of the securities. We use the method of stepwise regression to identify the appropriate subspace for pricing and the stock portfolios from the Constituent Shares of Index Shangzheng 500 The results of Variance Ratio analysis which tests the predictability show that the detrended stock prices deviate significantly from random walk and contain predictable components. The estimation results of the model of simultaneous e- quations indicate that the mispricing of the stocks tends to trend in the short - term and revert in the longer term. The out - of- sample performance of the Statistical Arbitrage model is sensitive to the movement of cost very much. The institutional investors＇ annualized Sharpe Ratio is 1.3.