Statistical arbitrage: pair trading for stock exchanges and exchange traded funds in NYSE and NASDAQ

Number of stocks and exchange traded funds has been increasing year by year, so it gets a fertile ground for risk arbitrage strategies. One of highly used risk arbitrage strategy in financial markets is pair trading, which is developed by algorithmic traders for automatic trading. Pair trading is a market neutral strategy, which makes it robust under any economic conditions. This thesis proves that an individual trader can success with a limited computing power and resource, in a game dominated by hedge funds and investment banks. In this study, pair trading strategy is developed exclusively for stocks and ETFs. The strategy is tested paired stocks and ETFs, which traded on NYSE and NASDAQ for the period between 2011 and 2015. First algorithm is coded in MATLAB, after then its performance is tested on 17 different pairs. The backtest program has a capability of requesting data from Yahoo Finance, parsing historical data into components, applying econometric models and finally using statistics to trigger entry and exit signals in an automated way. S&P 500 is taken as a benchmark, and the performance results of pair trading are compared on the basis of cumulative compound rate and Sharpe ratio. It is proven that the process of pair trading requires multi-dimensional skills and complex econometric calculations. Parameter choice and regime shifts can highly affect performance statistics of the strategy. Therefore it is suggested to use adaptive hedge ratios and continuous parameter optimizations. At the end, retail investors can utilize risk arbitrage and generate profits by using pair trading strategy on daily closing prices.