Essays in Market Microstructure

Abstract

This thesis consists of three chapters on the topic of market microstructure, focusing on liquidity and information revelation in markets with asymmetrically informed traders. Market microstructure refers to the study of the mechanics of trading in financial markets, and how those mechanics affect the characteristics of financial markets. One policy proposal that has received some recent attention in the realm of market microstructure is a financial transaction tax. An interesting question then is how such a tax might affect the functionality of financial markets, specifically, the liquidity and informational efficiency of markets, as these could have important welfare implications. In the first chapter, I analyze the effect of such a tax on liquidity and information revelation. I develop an insider trading model in the style of Kyle (1985) in which a single trader privately informed about the value of a financial asset chooses quantities to trade in a series of of auctions in an order-driven market. There are also uninformed traders who demand quantities for liquidity reasons and a risk-neutral market maker who sets the price equal to the expected value of the asset conditional on all public information. I implement a financial transaction tax on quadratic order flow in this market and prove the existence of a linear equilibrium in the model. I then analyze the effect of the transaction tax on liquidity and information revelation in numerical parameterizations of the model. I find that when the tax does not affect the trading demands of uninformed traders, it improves liquidity in the market but reduces the speed of the private information revelation. However, if the tax reduces the amount of trading by uninformed traders, I show that it may be the case that both liquidity and information revelation are worsened by the tax. In chapter 2, I conduct a study on the components of the bid-ask spread in markets with investors who trade based on private information. Recent research has suggested that informed traders may endogenously choose to trade when bid-ask spreads are low and liquidity is high. This would indicate that the bid-ask spread, a traditional measure of liquidity in a market, may then be a poor indicator because of information asymmetries. However, an interesting question is how the components of the spread, which is partly due to asymmetric information costs as well as costs associated with supplying liquidity, evolve in markets with informed trading, and whether a decomposition on the spread can lead to an improved measure of the existence of informed trading. To study such markets, I exploit an SEC disclosure rule to conduct an event study (as in Collin-Dufresne and Fos (2015)). Under the 1934 Securities Exchange Act, Schedule 13D forms (``beneficial ownership reports") must be filed with the SEC within 10 days of an investor acquiring 5\% or more of a publicly traded company with an interest in influencing the management of the firm. SC13D filings frequently precede change in management events such as mergers and takeover. Since the run ups to SC13D filings frequently see abnormal returns above the value-weighted market index and higher share turnover than the periods before or after, and once the form is filed, the information becomes public, they are used as a proxy for private information revelation events. I use a structural model of price changes to decompose the bid-ask spreads of those firms on whom Schedule 13Ds were filed into an asymmetric information component and a component consisting of inventory and order processing costs. The price variance is also decomposed similarly. I study how the components of the spread and the price variance evolve over the period surrounding these filings. I find that the model predicted spreads increase substantially during the period preceding the filing date, with the asymmetric information component and the liquidity cost component increasing by 17\% and 14\%, respectively, from 20 days prior to the filing date to just prior. In both my and Collin-Dufresne and Fos (2015)'s samples of SC13D filers, there appears to be a tendency for volume to be concentrated early on in the period when they possess valuable private information. In chapter 3, I develop another model in the style of Kyle (1985) to propose a possible explanation of this tendency, namely information leakage to the market. I then explore the effects of this possible explanation on liquidity and informational efficiency. In the model, an exogenous noisy signal correlated with the insider's information is gradually revealed to the market maker over the course of the trading period. I describe the necessary and sufficient conditions for the existence of a linear equilibrium in the model. In subsequent parameterizations, I find that this information leakage tends to improve market liquidity by substantially decreasing the price impact of order flow. It also improves the speed at which the insider's information is revealed to the market, thereby improving market efficiency. I find that under certain parameterizations, this can also explain the concentration of volume toward the beginning of trading. Lastly, I examine the effects on liquidity and information revelation in parameterizations where the informativeness of the noisy signal changes over time.