Turn on Javascript in your browser settings to better experience this site.

Don't show this message again

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Find out more

Quantitative Investing - Not all factors are created equal

  • 06Mar 18
  • Sean Phayre Global Head of Quantitative Investments

Factors are increasingly part of the language used to describe investment processes; however, care is needed to interpret what is meant, as not all factors are created equal.

Not all factors are created equal

Factor investing, where investors seek to generate superior risk-adjusted returns by systematically capturing risk-premia, or factors, has been part of the investment landscape in different forms for many years. The most recent incarnation of this is Smart Beta, a type of factor investing which, due to its combination of active and passive management approaches, has proved attractive to cost-conscious investors. Factor investing has also been seen as a natural way to incorporate industry trends such as embedding ESG considerations and Artificial Intelligence into investment processes. However, not all factor investing is created equal - different approaches to constructing factors and portfolios can produce different outcomes, not all of them intentional or desirable.

What is a factor?

In an investment context, a factor is simply a common theme or characteristic that links a set of companies. It may take the form of country of domicile or economic sector, or more commonly an investment style such as Value or Size. Dividend Yield is an example of such an investment style and income investors often buy stocks which have high dividend yields as a way of targeting regular income. Like other factors, stocks defined by high dividend yields are likely to move together in certain circumstances; for example, rising interest rates often negatively impact this group due to them being viewed as bond proxies. In general, factor investors target themes which are expected to provide outperformance over the long term – those favoured by Aberdeen Standard Investments include Value, Quality, Momentum, Small Cap and Volatility.

Factor portfolios

Factor investors create portfolios based on investment styles. However, there is divergence in approaches to constructing such factor portfolios. The simplest, and most common, way, is to rank stocks on a metric such as Value (defined, for example, by book yield) and buy those stocks with the highest scores while avoiding, or short-selling, those with the lowest scores. Although this process generally results in a high exposure to the desired factor, it may also lead to other unintended factor biases. Some factors constructed this way may have large positions in certain sectors – Value factors are generally overweight banking stocks. Or there may be unintended exposures to styles other than the target factor – a naive Momentum factor is often underweight Value by nature of their relative compositional biases.

Purer factors

One attempt to control for unintended exposures is to make the factor sector or industry neutral. While this may eliminate concerns around economic grouping biases, the resultant factor may still have some sensitivity to overall market moves. A further attempt to purify the factor would be to hedge out this market risk and make it beta neutral. In extremis, the purest factor is one which has exposure only to the targeted factor, while having no exposure to other factors. Such factors are termed orthogonal and are absent all unintended factor exposure (see Table 1).

Caveat emptor

Different factor approaches to construction can lead to different outcomes – Chart 1 shows the exposures of a US Value factor constructed using different methods.

While all the factors are positively exposed to the target, in this case Value, all but the pure factors have other non-targeted exposures in their make-up. These unintended exposures, such as underweight Quality and Momentum and overweight Small Cap, will influence the performance of the factor and could in some cases lead to it performing quite differently from expectations.

Historically, the practical implementation of such factors, and factor portfolios, is to translate into active portfolio weights and then overlay those weights upon a market cap index to produce a tilted long-only active portfolio.

Smart Beta and multi-factor

Smart Beta has become a popular flavour of factor investing in recent years. This particular approach seeks to gain concentrated factor exposure in a systematic way, while avoiding using market capitalisation as a starting point. As with single factors, issues of construction also impact Smart Beta methodologies which seek to simultaneously target multiple factors. Additional considerations arise due to alternative methods of combining multiple factors in one portfolio and the use of optimisation or alternative weighting schemes to assist construction. Chart 2 shows two well-known Smart Beta multi-factor offerings. Despite ostensibly targeting similar factor sets, the resultant portfolio exposures are quite different.

Understanding the details

Factors are increasingly a part of the language used by investment professionals to describe their investment processes. It seems reasonably clear that this will continue to be the case as the growth of Smart Beta continues and investors look to ESG and Artificial Intelligence as areas which could naturally be incorporated within a factor-investing framework. But how these factors are constructed and the contents of those portfolios which purport to target these factors are aspects requiring clear optics. As always, the investor should take care to understand what they are buying as not all factors are created equal.