Tid: 9 februari 2004 kl 1615-1700
Plats : Seminarierummet 3733, Institutionen för matematik, KTH, Lindstedts väg 25, plan 7. Karta!
Föredragshållare: Panagiotis Pavlidis.
Titel: Estimation Risk in Portfolio Selection. (Examensarbete)
Sammanfattning: The classical method for choosing a portfolio is based on the work of H. Markowitz. This method shows how to efficiently choose a portfolio which has a minimal variance for a given return level. The major drawback of this method is its sensitivity to inputs. Ignoring the estimation error inherent in the parameters which are needed for solving the portfolio problem leads therefore to some drawbacks, such as unstable solutions, poor out-of-sample performance, and unintuitive results. This thesis attempts to study some methods which have been developed to deal with these problems. These are purely statistical methods based on resampling and Bayesian techniques for estimating the parameters.