The One-Sided Pair Trading aims to test a version of Pair trading strategy, where short selling is not allowed.
Traditional pair trading provides to: (1) identify pair(s) of stocks (2) go short on undervalued stock and go long on overvalued stock - when paired stocks are relatively far from each other (3) close the position when the paired stocks come back to each other
This project aims to: (1) Test pair trading when short selling is not allowed (2) Optimize the strategy parameters, such as (a) finding 'optimal' trigger distances between the paired stocks as buy/sell signals, (b) test using k>=2 pairs simultaneously (isntead of 1 pair), (c) finding 'optimal' length of 'testing' (when pairs are found) and 'trading' (trading period before re-balancing) periods. (3) Performance is assessed versus benchmark - S&P500 index. Performance assessment indicators: (a) Cumulative return, (b) Annualized return, (c) Sharpe ratio, (d) Jensen's alpha, (e) portfolio Beta.
See available results in the PDF file Code is available in R.