Taking Advantage of Small Predictabilities in Financial Securities

Host: Michael Neely, Assistant Professor, EE-Systems, mjneely@usc.edu, EEB 520, x03505Seminar: Taking Advantage of Small Predictabilities in Financial Securitiesor, as we say in the engineering world,Multistage Mean-Variance Portfolio Selection in Cointegrated Vector Autoregressive SystemsSpeaker: Charlie Rohrs, Massachusetts Institute of TechnologyBased on the PhD work of Melanie B. RudoyFriday, January 23, 2009, 3:30 - 4:30 pm, EEB 248Abstract: Certain combinations of security prices reveal small amounts of intertemporal predictability when modeled as Cointegrated Vector Autoregressive Systems (CVAR). Such predictability can be exploited using tools from dynamic programming and optimal control theory to produce profitable portfolio optimization schemes. Here we use a standard Markowitz optimal mean variance approach on such systems. Certain CVAR systems display converging variance with increasing time horizon and we show how to appropriate exploit this phenomena when a single portfolio with no rebalancing throughout time is required. When rebalancing is allowed, the mean reversion properties of certain CVAR systems allow extra leveraging in intermediate stages to create higher profits. The extra leveraging produces excess profits while the variance trend is reversed by the terminal time to meet the overall variance constraint. The terminal time mean-variance problem does not quite fit with the CVAR models and an approximate solution is proposed.Biography: Charlie Rohrs received his BS degree from Notre Dame in 1976 and his Masters and PhD from MIT in 1978 and 1982 respectively. He served on the faculty at the University of Notre Dame from 1982 until 1997 and a Visiting Professor at MIT from 1997 until 2000. Charlie spent much of his career at The Tellabs Research Center, the research arm of Tellabs Operations, Inc., a manufacturer of telecommunications equipment for public service network providers. From 1985 until 1995, Dr. Rohrs was Director of the Research for Tellabs as the company grew from sales in the tens of millions to over a billion dollars annually. In 1995, he became the first Tellabs Fellow. Since 2001, Charlie has worked at MIT, first as a Principal Research Scientist in the Laboratory for Information and Decision Systems (LIDS) and then in his current position in the DSP group in RLE. While Dr. Rohrs is best known for his early contributions to the study of robustness in adaptive control, he has also contributed work in adaptive control, adaptive signal processing, communication theory, and communication networks and switching systems.He has recently become interested in the problems of financial engineering.

Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248

Audiences: Everyone Is Invited

Contact: Georgia Lum