Development of Financial Networks Based on Cointegration Concept (A Study on Tehran Stock Exchange)

Network theory can be used to better understand the financial markets, which are one of the most complex concepts in today's world. Financial networks are a set of Nodes(vertices) and edges. Each node represents one stock and each edge indicates the relationship between the stocks. Early studies in this area used correlation coefficients to describe short-term relationships between stocks and market modeling. This modeling method applies to data that have symmetry and only determines the presence or absence of a relationship between stocks, and does not specify the type of relationship, the direction and weight of that relationship. In recent years, financial time series modeling has been done using the concept of cointegration. In this study using the time series stationary test,unit root tests of Dickey Fuller and KPSS ,and Engle-Granger test, cointegration based network were designed. In order to analyze this network, the centrality measures such as: centrality degree, between centrality, closeness centrality, page ranking and Bonacich were used. The results of this study show that cointegration networks can provide more complete graph of markets, and also central measures analysis can play an effective role in stock portfolio selection and provide a good model for understanding stock relationships.