應用基因演算法於期貨與現貨領先落後關係之行為分析

Abstract

效率市場假說中說明由於市場是『效率的』，因此資訊會在市場價格上做迅速反應。所以不存在套利空間，投資人也無法在從中獲取超額報酬。然諸多的學術及實證研究得知，市場間充滿著不同的差異。也就是完美的效率市場並不存在於現實世界中，因此套利空間將會存在。 投資人是否可藉由交易相同標的物的現貨和期貨市場而打敗大盤呢？並進而從中獲取超額報酬？ 本研究嘗試導入基因演算法，對期貨與現貨在領先落後關係下的行為模式進行實證研究。以期貨市場價格交易資料作為其輸入因子，並對台灣加權指數進行驗證。在初步的實證研究中確認市場並非是有效率的，確可提供投資人較佳的投資策略與規則

Efficient Market Hypothesis states it is impossible to outperform the market because the financial markets are “efficient”, or that all the prices on the traded assets already incorporate and reflect to all the relevant information, hence, no existence of arbitrage. However a large body of evidence supports and argues that the financial markets are not always efficient. Arbitrage opportunities may exist.

Can investors beat the market by trading the same underlying assets in spot and futures markets and realize the excessive profits through different market strategies?

This study is to demonstrate the relationship of price deviation between the traded assets in spot and future markets by use of “genetic algorithm” analysis. In this simulation, the past historical future trading prices were acting as input variables and then compared with the performance of Taiwan Stock Exchange Index (TAIEX). The evidence indicates that the market is not always efficient due to different market trading strategies and investor’s psychology.