Index Replication Strategy Based On Compliex Network

Index replication is a passive investment method that uses a specific method to design and construct an investment portfolio and copy the trend of a selected target index to obtain a return rate that is as consistent as possible with the target index.Index replication is usually applied to the construction of index fund products.General stock-based fund products or stock investments rely heavily on the operation of fund managers and investors,and often result in poor investment returns,even lower than the average market return.Therefore,more and more investors,based on risk considerations,tend to obtain the average return of the market,and index fund products fit this investment goal.In addition,investors can also use index fund products for hedging transactions and arbitrage,and the products of index replication also have the function of price discovery.Therefore,since the birth of index fund products,the index fund products in the market have been endless.As of December 2019,the size of domestic index funds reached 1.19 trillion yuan,an increase of 440 billion yuan from the end of 2018,an increase of more than 58%.In the context of the prosperity and development of index funds,how to formulate index replication strategies to make index funds have a better market income replication effect has become the focus of scholars.Therefore,the exploration of index replication issues has important theoretical and practical significance..This article first explains the concepts,methods,effects,and influencing factors of exponential replication,and on this basis,explains in detail how to usecomplex networks to solve the problem of exponential replication.In terms of strategy design,this article starts from the idea of stratified sampling and intends to select the most influential component stocks from the different price fluctuation laws of index component stocks to construct an investment portfolio to replicate the index trend.Considering the advantages of complex network topological properties in the selection of cluster center points,this paper uses complex networks to describe the similarity of index component stock price fluctuations,and at the same time uses a multi-level optimization community discovery algorithm to automatically identify the number and rationality of clusters existing between component stocks.The cluster structure of the full mining the differences of different stock clusters,and optimized the stock screening method in the index replication problem.In the stage of portfolio weight allocation,the optimized genetic algorithm is used to solve the global optimal solution to obtain the best index replication effect.In this paper,the Shanghai and Shenzhen 300 index is selected to conduct an empirical study on the index replication strategy based on complex networks.The results show that the tracking error of the out-of-sample interval of the complex network method is only 0.0028 when 25 stocks are also selected for index replication.The layer selection method has been reduced by 22.2% and 36.4%.Through the rolling test of different time periods,it is found that the complex network method has the advantage of stable existence compared to the other two methods.The tracking error of the out-of-sample interval can be controlled within 0.55%.The tracking deviation is controlled within 0.45%,and the increment of the replication accuracy is stable above 0.5%.For the period when the index fluctuates more than 1000 points in half a year,the out-of-sample tracking error of the complex network method decreases by an average of 28% compared with the other two methods.The average replication effect at different periods can basically achieve the replication target of passive index funds in the market.It proves the feasibility and relative advantages of the complex network method insolving the problem of index replication,and provides a useful reference for the construction of index funds.