基于GARCH模型的统计套利策略在期货中的应用

发布时间：2018-03-31 23:38

本文选题：统计套利　切入点：协整关系　出处：《浙江工商大学》2013年硕士论文

【摘要】：上个世纪80年代,从摩根士丹利的研究人员所创造的匹配交易开始,人们逐渐把各种计量方法运用于证券交易中,由此开始了统计套利的辉煌时代。但21世纪统计套利遭遇了挑战,它原先所带来的巨额利润慢慢变得越来越薄,人们也开始怀疑它存在的价值,尽管如此统计套利的实际存在意义是不容否定的。统计套利在我国使用和研究的时间并不长,本文以建立协整关系模型和GARCH异方差模型为基础针对豆油和棕榈油这两个期货品种从2013年1月到10月的价格序列进行了统计套利的实证分析。从分析的结果来看,以运用协整关系模型和误差修正模型所计算的均衡价差为基础的统计套利要远远优于按照市场中性的假定所计算的价差为基础的统计套利；以GARCH模型计算所得出的条件标准差为基础的统计套利从稳定性和长期性上来说要优于以固定标准差为基础的统计套利,以固定标准差为基础的统计套利只能在短期内获得比较高的收益,在长期会带来很大的风险性和不确定性。本文得出的最终结论是在确定好两个确实存在很强相关关系的期货品种之后,运用协整关系模型和误差修正模型计算得到均衡价差,再在此基础上建立GARCH模型,得到条件标准差,以条件标准差为基础进行统计套利总是能够获得稳定的、风险较小的套利收益。以不同的标准差的倍数作为交易的信号会带来不同的套利收益,以较大的标准差的倍数作为开仓信号会带来较大的套利的收益,但同时会减少套利的次数；与此相反,以较小的标准差作为开仓信号所带来的套利收益较小,套利次数较多。投资者可以根据不同的偏好和需求选择不同的开仓标准。
[Abstract]:Since the matching trading created by Morgan Stanley researchers in the 1980s, people have gradually applied various measurement methods to securities trading, thus starting the glorious era of statistical arbitrage.However, the statistical arbitrage in the 21st century has been challenged. The huge profits it brought are becoming thinner and thinner, and people begin to doubt the value of its existence, although the practical significance of statistical arbitrage cannot be denied.Statistical arbitrage has not been used and studied for a long time in China.Based on the cointegration model and GARCH heteroscedasticity model, this paper makes an empirical analysis on the price sequence of soybean oil and palm oil from January to October 2013.From the results of the analysis, the statistical arbitrage based on the equilibrium spread calculated by the cointegration model and the error correction model is far better than the statistical arbitrage based on the market neutral assumption.The statistical arbitrage based on the conditional standard deviation obtained by the GARCH model is better than the statistical arbitrage based on the fixed standard deviation in terms of stability and long-term.The statistical arbitrage based on the fixed standard deviation can only obtain higher returns in the short term, which will bring great risks and uncertainties in the long run.The final conclusion of this paper is that after determining two futures varieties with strong correlation relationship, the equilibrium price difference is calculated by using the cointegration model and the error correction model, and then the GARCH model is established on this basis.To obtain conditional standard deviation, statistical arbitrage based on conditional standard deviation can always obtain stable arbitrage income with less risk.Using multiple of different standard deviations as a signal of a transaction will bring different arbitrage returns. Using a multiple of a larger standard deviation as an opening signal will result in a larger arbitrage return, but at the same time reduce the number of arbitrage times; in contrast,Using smaller standard deviation as opening signal brings less arbitrage income and more arbitrage times.Investors can choose different opening criteria according to different preferences and needs.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F724.5;F224

【参考文献】