Analysis of dual-listed companies in mainland and Hong Kong

Abstract

This thesis attempts to research price differences in Chinese segmented stock markets. Companies in China can be listed as A-share trading in Shanghai (SH) and H-share trading in Hong Kong (HK). However, the large and persistent price gaps between A-and H-share prices have been observed, raising concerns on market segmentation within China and its implications on the efficiency of price discovery. In the thesis, the factors are studied, which can be used in understanding the dissimilarity between the two markets. The topics are considered to include the use of statistical tests to perform factor analysis, and dynamic time-warping algorithm to conduct a technical analysis about the pattern of both markets.Topic 1, the some analyses cover the main economic and company-specific factors that influence the AH Premium Index, including the fundamental, technical, and market microstructure factors. Data is collected from 50 companies listing in both markets from April 2011 to June 2014, which are grouped into three clusters by the k-means clustering technique. An appropriate factor model is built for each cluster, and statistical tools are applied to test the model. The results demonstrate that different factors can be contributed to explain the price gaps in different clusters.It is noted that large price gaps are mostly related to the shares of small market capitalization. Since the small supply of A-share and information asymmetry are disadvantages for international investors, the trading strategy in relation to small cap stocks are more likely to succeed.On the contrary,the small price gaps account for the large market capitalization. The price gaps have been narrowing in recent years, which may pave the way to convergence after issuing the Shanghai-Hong Kong Stock Connect. Relative price convergence, but not absolute price convergence, is likely to occur Obizhaeva and Wang(2013).Topic 2 is based on the cluster results of topic 1.In line with the results of variance test,the price differences between A-and H-share markets are significantly obvious.In other words,the price differentials in the high and low premium clusters are larger than that in the non-premium cluster. The thesis applies dynamic time-warping algorithm to fit the patterns of two stocks for the same company, using the nine examples to explain this economical phenomenon. The results present that the market factors obviously have impact on the premium cluster, and the non-premium group is influenced by some fundamental factors.