基于银行股的配对交易策略及其实证研究

发布时间：2018-02-22 21:19

本文关键词： 配对交易融资融券随机价差均值回复协整检验止损边界夏普比率　出处：《浙江大学》2012年硕士论文　论文类型：学位论文

【摘要】：配对交易策略的研究由来已久。2010年以来,国内市场逐步放开,我国相继推出了融资融券业务和股指期货产品,这两项举措为包括配对交易在内的统计套利模型在中国金融市场中的应用提供了现实的基础。本文首先对配对交易过程进行了深入的分析和总结,主要分为三个方面：一是配对股票的选择,二是配对交易策略,三是收益的度量与风险控制。其次,本文对同质性较强的银行股票进行了实证分析。选用了日数据,30分钟数据,以及1分钟这三种频率的数据。先利用滚动相关系数法和滚动协整法来进行股票对的选择。然后对最优建仓时机与平仓时机进行了探测。针对日数据,本文探索出的最优进出场边界分别是0.85倍及2.8倍标准差,同时采取二日均线止损；针对30分钟高频数据,本文探索出的最优进出场边界分别是0.55倍及3.5倍标准差；针对1分钟高频数据,本文探索出的最优进出场边界分别是0.95倍及2.3倍标准差,同时不采取止损。考虑交易费用后的平均年化收益率分别为6.51%,11.44%,54.31%,远高于同期无风险收益率；夏普比率则均较低。而且还利用探测出的最优进出场边界对样本外数据进行了交易,结果表现良好。此外,本文还对两种交易费用进行了敏感性分析。其中对日数据而言,交易时间跨度较长,融券费率对配对交易的获利能力影响较大,而其它交易费用对收益的影响则略小一些。
[Abstract]:Since 2010, the domestic market has been gradually liberalized, and China has introduced margin trading and stock index futures products one after another. These two measures provide a practical basis for the application of statistical arbitrage models, including pairing transactions, in Chinese financial markets. This paper firstly analyzes and summarizes the process of paired trading, which is mainly divided into three aspects: first, the choice of paired stocks, the second is the strategy of paired trading, and the third is the measurement and risk control of returns. Secondly, this paper makes an empirical analysis of bank stocks with strong homogeneity. And 1 minute data of these three frequencies. First, the rolling correlation coefficient method and rolling cointegration method are used to select the stock pairs. Then, the optimal position setting time and the closing time are detected. The optimal boundary of the field is 0.85 times and 2.8 times standard deviation respectively, and the two day average stop loss is taken, and the optimal exit and exit field boundary is 0.55 times and 3.5 times standard deviation for the high frequency data of 30 minutes. According to the 1 minute high frequency data, the optimal boundary of the field is 0.95 times and 2.3 times standard deviation respectively, and the stop loss is not taken. The average annual rate of return after taking into account the transaction cost is 6.51 and 11.44 / 54.31 respectively, which is much higher than the risk-free return rate in the same period. Sharpe ratio is lower, and the best boundary of incoming and outgoing field is used to trade the data out of the sample, and the result is good. In addition, the sensitivity analysis of two kinds of transaction costs is carried out. For the daily data, the transaction time span is longer, and the margin rate has a great influence on the profitability of the matching transaction. Other transaction costs have a slightly smaller impact on earnings.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51

【参考文献】