【作者】 孔祥哲；

【导师】 杨秀昌；

【作者基本信息】 山西财经大学 ， 金融学， 2016， 硕士

【摘要】 在上证50和中证500指数期货品种推出之前,国内学者关于指数期货套利的研究主要集中在期现套利和跨期套利,新的指数期货推出之后,可以研究的内容得到了很大的扩充,使得进行跨品种套利研究成为可能。本文选择了上证50和沪深300股指期货作为标的,进行跨品种套利交易策略的研究。这不仅丰富了指数期货跨品种套利在理论上的研究,而且也为投资者提供了更多的套利机会,可以进一步促进我国金融期货市场的发展。国内外学者对于跨品种套利有着非常深入的研究,其基本内涵是对不同种类的期货合约开立多仓和空仓,并且是利用价差的波动来获利。根据本文的分析可以看出:上证50与沪深300指数期货5分钟高频数据的价差波动特征主要有两个:长期来看,上证50与沪深300股指期货的价格走势有很强的相关性;短期来看,两种期货合约的价格波动又有所不同。基于这样的价差特点,本文提出了跨品种套利交易的策略,主要是从三个方面进行设计,分别是套利交易的成本、无套利区间的确定和套利交易算法。套利交易的成本主要有手续费、佣金和保证金。对于无套利区间的确定,本文采用的是非线性协整模型,分别利用门限协整模型和门限向量误差修正模型估计无套利区间。本文最大的创新在于套利交易算法的设计,主要是开仓点和平仓点的设定。前人的研究大多采用无套利区间的边界作为开仓和平仓点,但本文考虑到了价差的变动具有惯性,当受到市场力量或者其他因素影响时,价差才会回复。所以在设定开仓点时,引入了价差的滞后一期作为判断条件,这样就可以避免价差的惯性运动,很大程度上提高了套利交易的成功率。对于平仓点的设计,本文将套利交易的收益情况作为判断条件,设定了止盈起始点、最大盈利回撤和止损点等一系列的参数来判断收益的情况。最后本文检验了两种无套利区间的实际效果,结果表明两种模型的套利策略成功率都达到了75%以上,验证了该套利策略的有效性。并且两种模型的套利策略都获得了非常不错的收益率。在股灾之后,中金所大幅提高了股指期货交易的手续费,如果手续费可以降到原来的水平,那么收益率大概可以提高到原来的三倍左右。同时在模拟交易过程中,发现了一个与理论并不十分相符的现象,在文章的最后,本文通过参数设置、样本选择和成本因素进行了解释。更多还原

【Abstract】 Before the SSE 50 and the CSI 500 Index futures release, domestic research on index futures arbitrage focused on arbitrage and inter temporal arbitrage. After the new index futures appearing, the content that can be studied has been greatly expanded and arbitrage cross varieties becomes possible. This paper choose the SSE 50 and the CSI 300 Index Stock Index Futures as a subject to study the cross-species arbitrage trading strategies. This not only enriches the research on index futures arbitrage in cross-species theory, but also provide investors more opportunities for arbitrage. It also can further promote China’s financial futures market development.Domestic and foreign scholars have a very in-depth study on cross-species arbitrage. Cross-species arbitrage’s basic meaning is that different types of futures contracts open long positions and short positions, and the use of spread volatility to make profits. The SSE 50 and the CSI 300 Index futures 5 minutes high frequency data’s characteristic are mainly two. In the long run, the SSE 50 and the CSI 300 Index futures prices have a strong correlation. In the short term, price fluctuations in the two futures contracts is different.Based on this spread, this paper proposes a cross-species arbitrage trading strategies, mainly designed from three aspects, namely the cost of arbitrage trading, determining arbitrage-free interval and arbitrage trading algorithms. Cost arbitrage trading mainly includes fees, commissions and margin. For the no-arbitrage interval, it uses a nonlinear cointegration, respectively threshold cointegration and threshold vector error correction model to estimate the no-arbitrage interval.The biggest innovation of this article is to design the arbitrage trading algorithms, which is mainly to set the opening and closing points. Most previous studies mostly using the no-arbitrage interval border as the opening and closing points. But this article takes into account the changes in spreads have the inertia. When subjected to the market forces or other factors, the spreads would be restored. Therefore, when the opening points set, we introduce the one lag of spreads as the determination condition. As a result, we can avoid the inertial motion of spreads, and greatly improve the success rate of the carry trade. For the design of the opening points, the arbitrage trading earnings will be used as judgment condition. We set a number of parameters, such as the starting point of only profit, maximum profit and stop-loss point to determine the earnings situation.Finally, this paper examines the two non-arbitrage interval’s practical effect. The results show that the success rate arbitrage strategy of both models has reached more than 75%, which prove the validity of the arbitrage strategy. And arbitrage strategies of both models are given a very good rate of return. After the stock market crash, The CICC has a substantial increase in the stock index futures’ trading commission. If the fee can be reduced to the original level, then the income can be increased to the three original times. At the same time in simulated trading process, we discovered that a theory is not very consistent with the phenomenon, and the paper finally gives a reasonable explanation.更多还原