Intraday Spreads and Institutional Trading in the Foreign Exchange Market

This thesis consists of three chapters exploring the foreign exchange market microstructure. This market is the largest global one, and the desire to work out the short-horizon dynamics of its exchange rates is the primary reason for research in this area. Chapter 1 examines the determinants of the spreads in the interdealer market of foreign exchange trading. The foreign exchange spreads behave dramatically differently in comparison to spreads in securities markets. Most notably, foreign exchange spreads vary inversely, rather than positively, with respect to trading activity and volatility. This paper traces these differences to aberrations in the intraday behavior of adverse-selection risk and inventory risk. These dissimilarities, in turn, are traced to the foreign exchange market's absence of an extended overnight period of curtailed trading activity.;Chapters 2 and 3 turn from the perspective of the dealers to that of the customers. Chapter 2 surveys institutional herding in financial markets, focusing on theoretical models and empirical evidence. Theory highlights at least six reasons why institutional investors might trade together. The first three explain why investors might trade in parallel: simultaneous perception of public fundamental information; simultaneous perception of past returns; simultaneous need to rebalance their portfolios. The last three suggest that some investors mimic others. Investors might gain knowledge from the preceding trades of other investors or they may follow the benchmark to preserve their reputation. Additionally, they may follow others when their performance is compared to that of other investors, and their compensation is evaluated as a result.;The survey points out that even though theoretical papers identify many reasons for correlated trading, empirical research on the topic typically employs the Lakonishok et al. (1992) "herding index," which simply identifies whether or not trading is correlated among investors. The empirical research yields weak proof of herding among institutional investors in the average stock at quarterly time horizons with relatively stronger evidence of herding among smaller, less familiar stocks.;As noted in the survey, only one paper examines herding in currency markets, the focus of Chapter 3. This paper provides the first extensive analysis of herding in currency markets, concentrating on the trading of mutual funds in 28 currencies at frequencies ranging downward from one quarter to one day. The results indicate higher levels of herding than previous studies of stocks. The results also reveal that correlated trading is more evident in less frequently traded currencies, consistent with the finding that there is more herding in smaller stocks. The results also indicate that measured herding decreases as the time horizon lengthens. To investigate the sources of correlated trading, Chapter 3 disaggregates funds along several dimensions including actively-managed vs. index funds; level of trading activity, and fund NAV. Regardless of the measure, the results detect no differences across funds. The paper finds that negative-feedback trading explains about half of the correlated trading.