ファクター投資の課題：意図せざるファクター・バイアス

Factor-based investing has gained considerable traction over the past decade. Concepts such as ‘factor premiums’ or ‘smart beta’ have become popular buzzwords, and now appear frequently in mainstream financial media. Prominent institutional investors have also publicly embraced allocation to well-documented factors. But despite growing awareness as to the potential benefits, many still struggle with how to put such strategies into practice in their portfolios.

In the previous article in this series, which is dedicated to the major challenges investors face in factor investing, we advocated a comprehensive and balanced approach in terms of exposure to different premiums. Because factors can clash with each other and are virtually impossible to time effectively, investors must avoid any excessive or undesired exposure to individual premiums.

The same is true for geographic regions, countries and business sectors, as well as individual stock exposures. Portfolio construction processes that focus solely on factor premiums can lead to significant, unintended biases, especially in terms of sectors. In a recent whitepaper¹, researchers from the Scientific Beta/EDHEC-Risk Institute showed that portfolios targeting greater factor exposure also tend to focus heavily on a limited number of industries, and concentrate on fewer sectors overall.

For example, a strategy designed to capture the momentum premium without taking into consideration any other element, may rapidly lead to excessive concentration of the portfolio on a small number of industries that may be in vogue at the time. As a result, sector-specific developments can significantly undermine overall performance.

Concentration risk is especially important for generic factor-based strategies, particularly when they rely on the replication of popular smart beta indices. Many of these products do not have explicit concentration limits. The S&P 500 Low-volatility index is a good example. There are no constraints on sector weights, which can lead to huge concentrations. As a result, in December 2012, around 60% of this index was invested in only two sectors: utilities and consumer staples.

Investors take this issue very seriously. In fact, a FTSE Russell survey carried out in 2016 suggested that avoiding unintended sector biases ranked third among investor concerns regarding factor allocation.

Although the focus should remain on optimizing exposure to relevant factors, the merits of broad diversification across a varied selection of securities should not be forgotten. Robeco’s in-house research shows that adding constraints on sector weights to an unconstrained portfolio reduces concentration risk while not significantly altering returns, at least to a degree. At a certain point, however, concentration limits start to have a negative effect on performance.

As a result, there is a converse relationship between the return/risk ratio of a portfolio and concentration levels, as measured by the allowed active weight for regions, countries, sectors, size groups or single stocks. This means that an optimal level of concentration exists that must be taken into account by investors.

Efficient factor strategies should therefore not only focus on maximizing exposure to premiums, but should also prevent unintended geographic or sector biases, as well as undue concentration on some single stocks or sub-segments of the financial markets.