Country and Sector Bets: Should They Be Neutralized in Global Factor Portfolios?

Abstract

A perennial question that comes up in the construction of factor portfolios is whether country and sector biases (in the form of active weights) should be neutralized. In this article, we review the assumptions that drive this decision, highlighting that both the portfolio construction framework and the desired portfolio criteria are critical to answering this question. We focus on two empirical examples—a rules-based approach to constructing the portfolio that is reflective of many current indexes in the marketplace and an optimization-based approach that is becoming an increasingly popular path. We corroborate past findings that indicated the answer depends on the portfolio construction framework. Moreover, it depends on the factor in question as well as the criteria—for example, returns, risk, information ratio, and turnover. If the returns and the information ratio are the focus, empirical evidence suggests some form of neutralization across both types of portfolio construction, particularly for value and size. When minimizing risk, maximizing the exposure per unit of tracking error and minimizing the turnover are the primary objectives, so the argument for neutralization is less clear.

TOPICS: Portfolio construction, factor-based models, portfolio theory