Portfolio Construction and Performance Measurement when Returns are Non-Normal

The foundation of popular approaches to portfolio construction and performance measurement lies in the mean-variance framework of Markowitz (1952, 1959). However, the suitability of such approaches in practice is questionable in light of considerable evidence of non-normalities in returns. This paper explores the potential usefulness of a non-parametric approach to portfolio construction and performance measurement recently proposed by Stutzer (2000). The Portfolio Performance Index (PPI) is based on the notion that investors associate risk with the failure to achieve a target return. Stutzer proposes that portfolio construction and performance measurement be approached by calculating the decay rate in the probability that a given portfolio will underperform its designated benchmark. By comparing the PPI and Sharpe ratio metrics, this paper presents preliminary evidence of the economic significance of non-normalities in Australian equity returns, and documents the impact of such on portfolio construction and performance evaluation practice.