A Study Of Pairs Trading Strategy Based On Copula-GARCH Model

Quantitative investment strategy of pairs trading is a kind of statistical arbitrage,but also a kind of market neutral investment strategy.Its core idea is to find two in the market which are satisfied at the same time in the long run and will deviate in the short run.The commonly used methods of pair trading are minimum distance method,cointegration method,random price difference method and Copula function method,which have been studied for nearly 10 years.With the in-depth study and application of these methods,pair trading has been greatly developed both in theory and in actual market transactions.On the basis of previous studies,this paper combines GARCH family model with Copula function,and proposes Copula-GARCH model to capture the trading opportunities caused by price deviation due to market inefficiency.This paper is divided into five parts.The first part is the introduction,which mainly explains the background and practical significance of this study,and summarizes the previous research results,and comments on their research,leading to the innovation of this study.The second part is an overview of statistical arbitrage strategy.It elaborates the concept,development and modeling process of quantitative investment,and explains the status,concept and main research methods of statistical arbitrage in the field of quantitative investment.The third part is the most core part of this paper.Firstly,the theoretical basis of the co-integration method pairing trading strategy is put forward,the stock pairs with strong correlation are found,and the co-integration trading strategy model of their spread time series is constructed.Secondly,the edge distribution based on GARCH model and kernel density function is proposed.Then,the Copula-GARCH model is proposed,and the concept of Copula function is further deduced.This paper focuses on the signal indicators of the transaction model.The fourth part is to select banking listed stocks with strong homogeneity as the stock pool,select Agricultural Bank of China and Industrial and Commercial Bank of China as the research object,respectively,using the co-integration method and the CopulaGARCH model established in this paper to test and simulate the sample trading.The last part is the conclusion and Prospect of this paper.The conclusion of this paper is that the GARCH family model based on different distribution assumptions is used to describe financial time series,and then the Copula function is selected to find the joint distribution of two research variables to construct the core model of this paper,Copula-GARCH model,to capture market trading opportunities.Taking 16 listed stocks satisfying the financing and margin trading conditions in the homogeneous banking industry as the research object,the research finds that five of the state-owned holding banks have the best correlation,and then choose ICBC and Agricultural Bank which have the best correlation as the data of the simulation transaction.By using the co-integration method and Copula-GARCH method,the data in and out of the sample are compared and analyzed.Considering the transaction cost,it is much better than the cointegration method in terms of transaction number,annual return,Sharp ratio and maximum withdrawal.