Download this Paper Open PDF in Browser

Alternative Objective Functions for Quasi-Shrinkage Portfolio Optimization

33 Pages Posted: 24 Mar 2010 Last revised: 7 Feb 2011

Andre Guettler

University of Ulm - Department of Mathematics and Economics; Halle Institute for Economic Research

Fabian Trübenbach

EBS Universität für Wirtschaft und Recht - EBS Business School - Department of Finance and Accounting

Date Written: February 7, 2011

Abstract

In this paper we propose a quasi-shrinkage approach for minimum-variance portfolios which does not use a quadratic loss function to derive the optimal shrinkage intensity. We develop two alternative objective functions for linear shrinkage. The first targets the reduction of portfolio variance. The second incorporates returns of assets to improve portfolio performance with respect to mean and variance. We compare the out-of-sample performance of our proposed portfolios to nine benchmark strategies across seven data sets. Our strategies often have lower portfolio variance and higher Sharpe ratios than the benchmark strategies. In particular, we beat the naïve portfolio empirically on all seven and significantly on three data sets.

Keywords: Covariance Estimation, Eigenvalues, Parameter Uncertainty, Portfolio Choice, Shrinkage

JEL Classification: G11, G12, C13

Suggested Citation

Guettler, Andre and Trübenbach, Fabian, Alternative Objective Functions for Quasi-Shrinkage Portfolio Optimization (February 7, 2011). European Business School Research Paper No. 10-07. Available at SSRN: https://ssrn.com/abstract=1576567 or http://dx.doi.org/10.2139/ssrn.1576567

Andre Guettler (Contact Author)

University of Ulm - Department of Mathematics and Economics ( email )

Helmholzstrasse
Ulm, D-89081
Germany

Halle Institute for Economic Research ( email )

P.O. Box 11 03 61
Kleine Maerkerstrasse 8
D-06017 Halle, 06108
Germany

Fabian Trübenbach

EBS Universität für Wirtschaft und Recht - EBS Business School - Department of Finance and Accounting ( email )

Oestrich-Winkel, 65375
Germany

Paper statistics

Downloads
142
Rank
172,422
Abstract Views
1,060