Year of Graduation
Statistical Arbitrage Opportunities on Financial Markets
Double degree programme in Economics of the NRU HSE and the University of London
This paper presents algorithm to exploit statistical arbitrage opportunities on current financial markets. The first chapter reviewed the attitude of famous theoretical models to ability of investor to earn positive abnormal returns on a systematic basis on the financial markets. The second chapter highlights limitations and operational risks of deterministic arbitrage and offers statistical arbitrage technic, which is based on Avellaneda and Lee (2010) work with some modifications by the author. The third chapter concentrates on empirical work. Weak form efficiency tests were performed using ARIMA model on price of Gold, Silver and share of British Petroleum. All models showed very low R2 and weak forecasting ability, and thus confirming the Efficient Market Hypothesis. Then different potential pairs (Gold and Silver, Brent and Light Sweet Crude Oil, pairs of shares of 9 large oil companies, calendar spreads of Light Sweet Crude Oil futures) were tested for cointegration with Augmented Dickey-Fuller (ADF) test on unit root, and those pairs, which are cointegrated, were used to model statistical arbitrage strategy. Analysis showed that all 6 cointegrated pairs generated positive returns; some pairs have even 30% return per year, confirming that it is possible to earn positive abnormal returns on the financial markets during several years of forecasting period. It was noted that these returns were not riskless due to existence of loss making trading periods. Empirical results confirm statistical arbitrage popularity among investors.