Generalization of Cointegration Arbitrage Based on GARCH Model, Improved by Marginal Probability Distribution

Abstract

As the Russia Ukraine War broke out in February, 2022, a series of international sanctions led to a panic among international hot moneys and slumps in global stock markets. For such a volatile market, hedging strategy is in a position to avoid risks. In finance, the correlation among financial derivative is always a topical issue and investors are concerned about the statistical methods to model the volatility. Recent years, the rapid development in data science makes it possible to record and manipulate massive data of financial derivative series, of which activates statistical arbitrage, a hedging strategy by pairs trading. The most widespread method in practice is cointegration arbitrage and many experiments show that it can always generate positive profits. This paper introduces the backgrounds of statistical arbitrage and relative definitions in cointegration arbitrage like time series, stationary, autocorrelation process, integration and cointegration. Based on the features of financial derivative series, this paper also illustrates error correction model, ARCH model and GARCH model. Finally, this paper shows the current strategy of cointegration arbitrage and raises some improvement by marginal probability distribution.