Analyzing the impact of global financial crisis on the interconnectedness of Asian stock markets using network science

Abstract

As importance of Asian Stock Markets (ASM) has increased after the globalization, it is become significant to know how this network of ASM behaves on the onset of financial crises. For this study, the Global Financial Crisis is considered whose origin was in the developed country, US, unlike the Asian crisis of 1997. To evaluate the impact of financial crisis on the ASM, network theory is used as a tool here. Network modeling of stock markets is useful as it can help to avert the spillover of crises by preventing the stock markets which are highly connected in the network. In this empirical work, weekly indices data from 2000-2013 for fifteen stock markets is used, which is further partitioned into three periods: pre, during and post crisis. This study shows how 13 important stock markets in Asia namely, India, Bangladesh, Philippines, China, Japan, Indonesia, Malaysia, Singapore, Hong Kong, Pakistan, South Korea and Thailand are connected to each other and how India, Japan, Hong Kong and Korea stock market appeared as the systemically important stock markets from them. Introduction of the US stock market into this network gives insight how the US stock market might had connected to systemically important markets which resulted into spread of crisis in the Asian region. Furthermore, using Kruskal algorithm spread of contagion is explained like how it first hit the Hong Kong stock market and from there it proceeds to the other systemic important stock markets like a virus. Addition to that, we quantified the network behavior in the form of metrics such as adjacency matrix, clustering coefficient, degree of nodes and Minimum Spanning Tree (MST), and on the basis of these some of the important questions like which stock markets are highly connected in Asia which if affected can induce the crises in the other stock markets of region are answered. This study can be used for the portfolio optimization as well as for policy making for which network analysis should be conducted on a regular basis.