Allocation to factors has become increasingly popular in recent years, but practical implementation remains a puzzle for many investors. Determining the best strategy or the best combination of strategies is often the first pitfall they must face.
A growing number of academic studies suggest that systematically harvesting a number of well-rewarded factor premiums, such as value, size or momentum, ensures enhanced returns in the long run, both in equity and bond markets. Over the past decade, these findings have led to the emergence of a new kind of investment product, frequently branded as ‘smart’ or ‘alternative’ beta, that have definitely drawn investors’ attention.
However, despite growing awareness, the practical implications of allocating to factors are often still very difficult for newcomers to grasp. There are currently hundreds of smart beta products available in the market, from basic single factor ETFs to sophisticated multi-factor solutions. A FTSE Russell survey carried out in 2016 suggested that determining the best strategy or the best combination of strategies ranked first among investors’ concerns, when looking at factor allocation.
Indeed, finding the right combination of factors is a major challenge. The fact that it is very difficult to successfully predict which factors are going to do well in the near future and which factors are going to lag, supports diversification across factors. Unfortunately, individual factors also have negative exposures to one another. For example, strategies focusing solely on the momentum factor tend to have a very negative exposure to the value premium.
‘Efficient factor investing strategies should avoid risk concentration’
Moreover, a 2016 paper* by David Blitz, Head of Quantitative Equities Research at Robeco, showed that many existing smart beta products do not offer maximum factor exposure. That’s because they are exposed to cap-weighted factor indices, while equally-weighted factor strategies are known to generate higher returns.
Efficient factor investing strategies should therefore be designed to avoid risk concentration and to ensure that premiums do not clash with each other. Limiting unnecessary turnover is also a key concern.
Our research and more than twenty years of practice show it is possible to build well-diversified portfolios selecting securities that provide efficient exposure to one specific factor while avoiding negative exposure to others.
For example, it is possible to find stocks that are attractive not only from a low-volatility perspective but also in terms of valuation and momentum. In a similar way, it is also possible to find attractively valued stocks that are also interesting from a quality and momentum point of view.
As a consequence, investors should ensure they take a multidimensional approach to risk in their quantitative strategies in order to ensure optimal exposure. This is very important, as most generic smart beta products available in the market tend to focus only on one aspect at a time.
Portfolio construction processes should never rely on one single factor and should integrate both backward and forward looking measures of risk, including elements such as past volatility statistics or earnings expectations, for example. They should also mitigate concentration risk by having strict research-based concentration limits for region, country, (sub-)sector, size and single stock weights.
It is also important to note that there is no single ideal approach to factor allocation. In an article** published in September 2016, David Blitz and Joop Huij, Head of Factor Research at Robeco, argued that the optimal factor-investing portfolio depends on investor-specific beliefs and preferences. Depending on its own profile, each investor will seek a specific kind of performance, and will be willing to take on more or less risk.
For example, the low-volatility factor is very attractive for those who want downside protection without sacrificing return potential, such as pension funds aiming for funding ratio stability. However, it can be less attractive for investors who care more about maximizing return than about reducing risk and for investors who dislike the high tracking error involved in low-volatility strategies. Finally, the existing portfolios and factor exposures of an investor must also be taken into account when building an optimal solution. For instance, an investor who already has a significant value tilt in the core portfolio might want to give less weight to that factor in the factor portfolio. To help investors make their choices, Robeco has actually developed a specific tool that determines the factor exposure profile of a portfolio or a benchmark.