Thought Leadership

Factor-based investing: are other strategies leaving returns on the table?

Many practical investors worry that investing in more than one or two factors is self-defeating, as they fear they will end up simply holding a more complicated version of the index. The new evidence presented here shows that these investors are right to worry.

11.10.2017

Smart beta appeals to many investors drawn by its claim of low fees, transparency of return sources, and attractive historical performance. This is an impressive package. But sophisticated investors are increasingly aware that simple “smart beta” implementations can also suffer some serious drawbacks. A previous article (Schroders, 2016) discussed some of these, including the problem of which factors to choose and the importance of building high-quality exposure to the chosen factors.

This paper provides new evidence of a better way to incorporate factors in a portfolio. Most factor-based strategies target individual factors, such as momentum or value. To obtain diversified exposure to factors, investors can combine individual factor strategies. We call these “single-factor portfolios”, because investment decisions are made for each factor strategy individually, and the portfolio is a sum of these exposures. This is a common and simple way to combine factor exposures, but it may not be the most effective way.

In this paper, we consider a different approach. Instead of combining individual factor strategies without consideration as to how they overlap, we introduce a “multi-factor portfolio” that looks at all of the relevant factor information for each stock and manages exposures for the portfolio as a whole. We prefer this approach because it is
more efficient in capturing diversified exposure to factors. For example, it avoids buying stocks in one factor portfolio that we wish to avoid in another portfolio, and it makes more efficient use of the information available in choosing stocks. 

First, we show how single-factor and multi-factor portfolios work in practice. We then present a framework which explains how a multi-factor approach can deliver superior risk-adjusted returns, compared with a single-factor approach. The advantage gets larger as more factors are used. Indeed, in a simple case study, we show that a four-factor integrated approach can provide twice as much effective factor exposure as its single-factor counterpart. Critically, in the historical cases we study, this leads the multi-factor approach to outperform by twice as much as the single-factor approach.