Three essays on international financial markets

My dissertation investigates the market dynamics in international financial markets with a particular reference to equities and exchange-traded funds (ETFs). It contains three chapters.

The first chapter examines the NYSE move to Hybrid trading which was initiated in October 2006. I find that this new trading platform introduced a much larger proportion of electronic transactions relative to floor auction transactions. This migration to electronic transactions is further evidenced by a mirror shift in price discovery from floor trades to trades marked for automatic electronic execution. In addition, I find that the move to Hybrid trading introduced a significant decrease in inventory control costs, as well as a noticeable increase in trade persistence. Finally, the new trading platform has increased the speed with which orders are met, and has also decreased the proportion of executed shares which receive price improvement.

In the second chapter, I analyze return and volatility of Asian iShares traded in the U.S. The difference in trading schedules between the U.S. and Asia offers a unique market setting that allows me to distinguish various return and volatility sources. I find Asian ETFs have higher overnight volatility than daytime volatility, explained by public information released during each local market's trading session. Local Asian markets also play an important role in determining each Asian ETF return. Nonetheless, returns for these funds are highly correlated with U.S. markets, indicative of the effects of investor sentiment and location of trade. Finally, returns in the U.S. market Granger-cause returns in all six Asian markets analyzed.

The third chapter examines the profitability and resiliency of an intraday pairs trading strategy derived solely from historic price dynamics and contrarian principles. I find that the profitability of the strategy hinges on a cointegrated relationship, which Engle and Granger (1987) show also implies an error-correcting relationship. Thus, if two series are cointegrated, then they will never move arbitrarily far apart from one another before reverting back. Additionally, if two series are cointegrated, then movement in one series will Granger-cause movement in the second, which means that an error correction model can be used for forecasting. Simply stated, the most important feature of arbitrage, particularly in terms of how it relates to pairs trading, is the convergence of a temporary pricing flaw back to their expected values. The results of this study show that although a trader would not be able to foresee the movement of the cointegrating leader, that move generates a predictable response from the cointegrating follower, which the pairs trading strategy is able to capitalize on and profit from.