A Practical Approach to Using Taguchi Loss Functions to Rank Investment Choices

Abstract

We use a series of hypothetical financial situations to describe the manner in which a quantitative technique could assist analysts in making financial decisions. We also provide a detailed example of a portfolio manager's attempt to replicate a given stock with another stock that meets multiple characteristics. The demonstrated quantitative technique is an improvement over the traditional “accept/reject” dichotomy because it imposes an explicit penalty for deviations in the observed values from the desired characteristic values, allows multiple characteristics to be considered simultaneously, and also allows the selected characteristics to be weighted based on their relative importance