Applying Extensible Classifier System to Inter-market Arbitrage with High-Frequency Financial Data

The most popular arbitrage opportunities detecting methodology is derived from the cost of carry model. Recently, many researches were intent to enhance the accuracy of these arbitrage models using econometrics approach. However, the market behavior is still hard to be known well, especially when inter-market spread trade with intra day one minute tick data. This research is aimed at inter-market arbitrage with high frequency data, and two futures indexes are used for empirical study, including Taiwan Stock Index Futures of Taiwan futures exchange (TAIFEX) and MSCI Taiwan Index Futures of Singapore Exchange Limited (SGX). Moreover, the price of index futures will get close to that of spot products when the futures contract is due. Founded on such property, the spread ratio and the different due days of TAIFEX and SGX, we finally build up an extended classifier based arbitrage system which can gauge the timing of index stock deals.