| This thesis focuses on the issue of capital market segmentation (integration) in the North American market. More specifically, it investigates the existence of partial segmentation between Canadian and U.S. capital markets. Testing specifically for partial segmentation, as opposed to complete segmentation, is important because failure to do so could lead to inappropriate inference.;To test for partial segmentation, I first examine the factors that potentially segment the North American market. It is suggested that factors associated with a firm's listing status, firm size, and dividend yield are likely to contribute to the segmentation of Canadian and U.S. markets. Using these instruments (listing status, firm size and dividend yield), I partition the overall sample of Canadian firms into sub-groups which are then used to test the hypothesis of partial segmentation. The tests are run over different time periods, using a univariate and a multivariate framework. I also run a number of sensitivity tests. For instance, I test for different definitions of interlisted firms, the impact of thin trading in Canadian stocks, and the impact of outliers.;The results show that listing status plays a role in the segmentation of Canadian and U.S. markets. More specifically, we find more evidence of integration for the interlisted firms than we do for purely domestic stocks. However, the results also show that both firm size and dividend yield are useful in controlling for the factors segmenting the North American market. In particular, the results provide support for the hypothesis that small (large) Canadian firms are more likely to be priced in a segmented (integrated) market. Similarly, controlling for size, high (low) dividend yield firms are more likely to be priced in a segmented (integrated) North American market. Thus, the results suggest that differential taxacion between Canada and the U.S. and factors related to firm size such as information incompleteness contribute to the segmentation of the North American market. Finally, it is interesting to note that our results help explain why existing studies provide conflicting evidence as to the existence of segmentation in the North American market. |
