Index Fundamentalism Revisited

Abstract

A structured global optimization model budgets a fund's total active risk (tracking error) among its asset classes or across investment managers. Its objective is to maximize total fund expected excess return (over benchmark) for a given level of total fund tracking error (the active risk budget), assuming that total fund and asset class tracking errors are controlled solely through the selection of investment managers with various active risk profiles. The key innovations are, first, that the efficient frontier (expected excess return versus tracking error) of each asset class is fitted to a functional form, and then the optimal active risk level for each asset class is determined using the asset class efficient frontier functions as inputs. The model compares well to a structured global linear model and an unstructured global model that optimizes with respect to all investment managers simultaneously.