Influence Functions in Finance: Statistical Analysis of Portfolio Risk Measures and Performance

Shengyu Zhang*

Department of Statistics, University of Washington

05 October 2006, 4:00 P.M., Mary Gates Hall 238

Portfolio risk and performance measures, which are widely used in asset allocation and portfolio management decision, are often based on the estimated statistics from historical data, therefore subject to estimation error and sometimes very sensitive to the perturbation of the underlying distribution and adversely influenced by outliers. Influence function developed in the robust statistics literature provides a useful tool to study the following key aspects of an estimator: (a) the influence of an individual data on the estimator; (b) an approximation to the bias caused by a small fraction of outlier or inliers; (c) asymptotic variance of the estimators. However, very limited studies in the finance literature have been carried out to apply influence functions to finance research. In this study, we attempt to use influence function as a systematic tool to study statistical property and character of various portfolio risk and performance measures.

The slideset from this talk can be downloaded from here.


*Shengyu Zhang is a graduate student in the Department of Statistics.

NOTE: Off-campus professionals interested in quantitative finance are invited. Park in Central Parking via Gate 2 on 15th Ave. See map at http://www.washington.edu/home/maps/northcentral.html.
The University of Washington is committed to providing access, equal opportunity and reasonable accommodation in its services, programs, activities, education and employment for individuals with disabilities. To request disability accommodation contact the Disability Services Office at least ten days in advance at: 206-543-6450/V, 206-543-6452/TTY, 206-685-7264 (FAX), or dso@u.washington.edu.