Test of the Efficient Market Hypothesis by utilizing statistical arbitrage

Abstract

The purpose of this paper is to test the efficient market hypothesis. The thesis includes an extensive review of the literature on the Efficient Market Hypothesis and tests if it is possible to earn economic profit by trading on an information set known to all market participants. The trading strategy is based on statistical arbitrage pairs. Pairs trading is a concept based on a co-integrating relationship. Co-integration is the long-term stationary relationship between two asset prices. The pairs in this thesis are untraditionally composed of indices. The test for co-integration uses the price of the ETF for S&P500 and the prices of six other indices from the European continent. I find that there is a stationary long-term relationship between the prices of the respective ETF’s of S&P500 and FTSE100 before adjusting for currency. The cointegration test shows a stationary relationship between the two prices and that the spread has mean-reverting properties. A trading strategy based on an out-of-sample period of 85 days does not yield positive profits after adjusting for transaction costs. With USD as the base currency, none of the seven variables co-integrates. In the absence of a co-integrating vector, there is no basis for trading the indices, and the EMH cannot be rejected.