|M.Sc Student||Bagrov Natan|
|Subject||Ranking and Trading Execution of Mean-Reverting|
|Department||Department of Computer Science||Supervisor||PROF. Ran El-Yaniv|
|Full Thesis text|
Statistical arbitrage is an essential pillar of quantitative trading and has long been used by hedge funds. The term statistical arbitrage encompasses a wide variety of investment strategies, which identify and exploit temporal price differences between similar assets using statistical methods.
Historically, statistical arbitrage evolved out of the simpler pairs trade strategy, in which stocks are sorted into pairs (portfolios of two stocks) according to fundamental or market-based similarities. When one stock in a pair outperforms the other, the underperforming stock is bought long and the outperforming stock is sold short with the expectation that underperforming stock will climb towards its outperforming partner.
Mathematically speaking, such portfolios are typically selected using tools from cointegration theory, whose aim is to detect combinations of assets that are stationary and, therefore, mean-reverting.
Having a mean-reverting portfolio, however, may not be enough to ensure profitability. This type of portfolio might show very small volatility, requiring significant leverage to be profitable as well as incur trade and borrow costs (commission).
As a step toward bridging this gap, we present a framework comprising three main stages: (i) constructing time series of mean-reverting portfolios, (ii) ranking these portfolios and balancing the reversion rate together with profitability potential, and (iii) trading these portfolios efficiently, using risk-averse decision-making.
Tested on real stock data, the framework is shown to be market neutral as well as competitive in terms of Compound Annual Growth Rate (CAGR) and Sharpe Ratio.