Talk Topic: An Order Flow Model and a Liquidity Measure of Financial Markets
Adlar J. Kim, MIT CSAIL & Sloan School
July 16, 2008

Abstract
In the first part of this talk, I demonstrate statistical modeling techniques to model order flow generation, which is a liquidity generation process for financial markets. A simulation of the order flow model successfully replicates various statistical properties of price returns, such as clustered volatility, fat-tailed return distribution, and no predictability of future returns. 

In the second half of this talk, I argue that the change in market liquidity explains how my order flow model satisfies the weak form of the Efficient Market Hypothesis (EMH), which asserts that future price returns cannot be predicted from past returns. This is not an obvious result since the order flow model has a predictable transaction sign process (i.e. buyer or seller initiated trades) as one of its components. A method of quantifying market liquidity from order flow data is introduced to explain this result. I will also present some interesting patters of liquidity measures in the real market data.

The talk will assume no prior financial knowledge.

Bio
Adlar J. Kim is a Postdoctoral Associate in the Laboratory for Financial Engineering (LFE) at MIT. He received a Ph.D. from MIT in Computer Science (2008). Additionally he holds a B.S. degree in Computer Science and Economics (1998), and M.S. degree in Information Systems Management (2000) from Carnegie Mellon University. His research interests include agent-based simulation, artificial intelligence, machine learning in the framework of financial markets, and study of prediction markets. While he was a graduate student at MIT, he spent a year at the Santa Fe Institute as a graduate fellow.