Essays on Trading and Financial Econometrics

Abstract

This dissertation studies trading and investment in financial markets through the lens of financial econometrics. Chapter 1 develops a continuous-time model of the optimal strategies of high-frequency traders (HFTs) to rationalize their pinging activities - defined as rapid submissions and subsequent cancellations of limit orders inside the bid-ask spread. The current worry is that HFTs ping inside the spread to manipulate the market. In contrast, the HFT in my model uses pinging to control inventory or to chase short-term price momentum without any learning or manipulative motives. I use historical message data to reconstruct limit order books, and characterize the HFT's optimal strategies under the viscosity solution to my model. By gauging the model's implications against data, I show that pinging is not necessarily manipulative and is rationalizable as part of the dynamic trading strategies of HFTs.

In Chapter 2, joint with Harrison Hong, we use overdispersed Poisson regression models to study social networks in finance. We count an investor's social connections in different cities as proportional to the number of stocks held by this investor that are headquartered in those cities. When connections are formed in an i.i.d. manner, the count of such connections in any city follows a Poisson distribution. Using data from institutional investors' holdings, we find instead overdispersion for a number of cities like San Jose and San Diego, which suggests that investors have non-i.i.d. propensities to be connected to these cities. Overdispersed cities have a large number of graduates from local universities who work in the fund industry. Managers with relatively high non-i.i.d. propensities to have social contacts significantly outperform other managers.

In Chapter 3, I propose a continuous-time model for the joint stochastic process of asset price and trading volume to study the transmission mechanism from changes in trading volume to price movements at the high-frequency level. A GMM-based estimation procedure is developed based on the model's closed-form moment conditions. I estimate the model on real-world high frequency financial data and find that, jumps in volume have a strong cross-excitation effect on jumps in price. Other implications of the model are also discussed.