Scaling Issues in Additive Multicriteria Portfolio Analysis

Abstract

We discuss the effects of different utility scales on the results of multi-attribute portfolio problems. In particular, we analyze three effects: (i) the portfolio size effect, that the evaluation might be biased according to the number of items included in a portfolio, (ii) the baseline effect, which requires attention to the consequences of not including an item in a portfolio, and (iii) consistency across different aggregation sequences. We show that these three effects have similar causes related to the use of an interval utility scale, which allows for additive transformation of utilities. We discuss several ways to overcome these problems, and their consequences for utility and weight elicitation. © Springer International Publishing Switzerland 2014.