摘要
本文利用我国2014年12月股指期货连续合约的秒级分时数据,估计了自回归条件久期模型的相关参数,并以此为基础构建了ACD交易策略,实证结果表明:该交易策略在tick(30秒)和1分钟级别数据上均可以获得正的收益,但3分钟及以上级别数据的收益为负。在和传统统计套利交易策略的对比中发现,ACD交易策略有更高的盈亏比,而统计套利策略的胜率相对更高,两种交易策略的风险收益比则比较一致。
This paper used second level time data of stock index future continuous contracts in December 2014 of China to estimate the autoregressive conditional duration model, based on which it constructed ACD trading strategies. The empirical results show that: the trading strategy in tick (30 seconds) and 1 minute level data can get a positive earning, but in 3 minutes and above the level, the income is negative. In the comparison with the traditional statistical arbitrage trading strategies, we found that the ACD trading strategy has higher profit and loss ratio, and the winning probability of the statistical arbitrage strategy is relatively higher. The risk return ratio of the two kinds of trading strategies is close.
出处
《投资研究》
2016年第1期122-136,共15页
Review of Investment Studies
关键词
股指期货
交易策略
ACD模型
Stock index futures
Trading strategy
ACD model