Co-movement of European Stock Markets based on Association Rule Mining Youqin Pan, Elizabeth Haran, Saverio Manago Dept. of Marketing and Decision Science Salem State University Salem, USA Emails :{ypan, eharan, smanago}@salemstate.edu Yong Hu Dept. of E-Commerce Guangdong University of Foreign Studies Guangdong, China henryhu200211@163.com Abstract—Due to the fluctuation and complexity of the stock market, it is challenging to capture its non-stationary property and describe its moving tendency. Moreover, globalization increases the interdependence among countries. It is important for investors to understand the co-movement of international stock markets in order to make informed decisions which lead to profit. With the huge amount of data generated by the stock markets, researchers started to explore this problem using different approaches. In this paper, we apply one of the data mining techniques, namely, association rules, to illustrate knowledge patterns and rules of European stock markets. Especially, this paper investigates the co-movement of the European stock market indices with the leading global stock indices. This study shows a strong co-movement between stock market indices of Germany and Unitied Kingdom. Moreover, the European stock markets seem to have strong co-movement with the US stock market. Their co-movement with the Brazil seems to be also strong. However, Brazil stock index does not assume the dominant role, as the US stock index does. This study also shows that there is a weak relationship between European and Japanese stock markets. Keywords-co-movement; association rules; stock index; co- integration. I. INTRODUCTION International stock market linkages are of great importance for financial decisions of international investors. International diversification reduces total risk of a portfolio. Increase co-movement between asset returns can diminish the advantage of internationally diversified investment portfolios [18]. Changes in co-movement patterns call for an adjustment of portfolios [26]. Forecasting stock index is a challenging task due to its dynamic and complex nature. Forecasting stock index plays an important role in developing effective market trading strategies [14]. Stock markets can be influenced by various factors such as the international environment, government policies, political climate, economic growth, war, and natural disasters. Among these factors, some of them have long- term effect on the markets while others have only short–term effect [28]. Recently, globalization adds more complexity to the movement of stock markets. Globalization in finance and trade increases the interdependence among countries. Such relationships further cause the co-movement of the financial markets between countries. Studies have confirmed that most of the world’s stock markets are integrated and associated [22]. Loh [19] claims that understanding the dynamic co- movement between global financial markets plays an important role in predicting stock market returns, allocating assets and diversifying portfolios. This paper extends the existing literature on stock market co-movement between the European stock markets with that of the US, Brazil, and the Japan. The major European markets include UK, Germany, and Turkey stock markets. The rest of the paper is organized as follows. In Section 2, we give the literature review. Section 3 presents data and research technique. Section 4 presents research findings and discussions. Section 5 concludes the paper. II. LITERATURE REVIEW The dynamic interdependence and market integration among major stock exchanges have been investigated by various studies using vector autoregression (VAR) and autoregressive conditional heterosedastic (ARCH) models. Vuran [27] found that the ISE100 index is co-integrated with stock markets of the United Kingdom (FTSE), Brazil (BOVESPA), and Germany (DAX). Floros [9] demonstrated the linkages and co-integration among mature stock indices (such as S&P 500, Nikkei225 and FTSE-100) using a vector error correction model and the Granger- causality approach. Ozdemir and Cakan [23] claimed that there is a strong bidirectional nonlinear causality relationship between the US stock index and the stock market indices of the Japan, France and the UK using non- linear causality tests. Some studies have demonstrated that the U.S stock market has a dominant impact on emerging markets [20] and some developed stock markets such as Japan and France [23]. These studies demonstrated the dynamic causal linkages among international stock market indices. Contrary to these findings, Chan, Gup and Pan [6] concluded that stock markets are not co-integrated, by analyzing 18 stock market indices. Pascual [24] also found that there is no co-integration relationship between the French, German, and UK stock markets, by using quarterly data. Zhu et al. [30] rejected co-integration relationships between market returns in Shanghai, Shenzhen and Hong Kong. Dimpfl [8] further proved that international financial 54 Copyright (c) IARIA, 2014. ISBN: 978-1-61208-358-2 DATA ANALYTICS 2014 : The Third International Conference on Data Analytics