The foreign share discount in a segmented market

Existing literature examines price premia on foreign shares under various market conditions. Recently we observe a foreign share discount in some emerging markets. This dissertation studies a segmented equity market allowing foreigners to invest in foreign shares only and locals in domestic shares only. Foreign and domestic shares receive identical dividends, but they trade on separate markets.; The first essay models the pricing of foreign and domestic shares in a segmented market. Political risk and alternative investment vehicles are analyzed to understand the price discrepancy between foreign and domestic shares. The models yield simple analytical solutions. The ratio of foreign share price to domestic share price depends on their expected returns and foreigners' confidence level in the local political environment.; The second essay extends these models to reflect the time varying nature of the arrival rate of political shocks. The resulting pricing formulae are in closed form in terms of Bessel and Lommel functions.; The third essay investigates the effect of market segmentation on the returns of foreign and domestic shares in China's secondary market. Chinese equity markets display the market segmentation described above with foreign shares trading at only 30 percent of domestic share prices. Weekly returns on foreign and domestic shares of the same company are poorly correlated. The foreign shares exhibit properties similar to NYSE and AMEX stocks. But weekly returns on domestic shares are highly correlated, regardless of whether the company issues foreign shares. Therefore, the dynamics of stock prices in China are influenced more by investors' preferences than by issuers' fundamentals.; The fourth essay examines the source of differing returns on Chinese foreign and domestic shares, lower expected returns required by Chinese investors or foreign investors' perceptions of political risk in China. We find no correlation between the Chinese government bond yields and the return difference. This rules out investment opportunity choices as the cause of return difference. We further test the hypothesis that political events cause the return difference, using event study methodology. Over 70 percent of the events support our hypothesis. The empirical results are consistent with our pricing models.