Lan Liu

Proposed thesis title:

Forecasting return covariance structure and portfolio tracking error management

Supervisor:

Professor Stewart Hodges

Start date:

January 2003

A summary to the research area:

My research is related to the forecast of portfolio tracking error in active fund management. Tracking error measures the difference between a managed portfolio and a benchmark portfolio. It’s a measure of the value added by the professional money managers. Managers also use forecast tracking error to construct portfolios in order to avoid extreme deviations from the benchmark. Our study develops a new forecast model and proposes a new approach to evaluate alternative risk models particularly for portfolios of equities. We also use simulation techniques to investigate the problems of estimation noise in risk forecasting. Together, we wish to better understand the use of risk models and forecast more accurate tracking error.