Multi-factor investing: the new generation of quantitative processes

Post with image

Just as it can be irritating to hear someone say “things used to be better than they are now”, it can be irritating and even offensive when you have reached a certain age to hear, over and over again, that the new generation outperforms the previous one. Although they have long been at the cutting edge of factor-based investment research, our quantitative investment teams do not suffer from this shortcoming.

Just because we use factor investing strategies, this does not mean that we have dumped the models that we used when launching smart beta in our fund range in 2008. Factor investing and smart beta are, in fact, perfectly complementary. We are confident that the way we apply the factor strategy can generate added value relative to index replication, including for the US S&P 500, an equity index that many regard as unbeatable by active investment strategies.

Our multi-factor investing approach calls on the findings of behavioural finance in exploiting market imperfections, or diverging points of view between different market participants. Moreover, we have chosen to base our strategies on risk calibration, which ensures considerable regularity in performance. Our new generation is both less aggressive than some of its predecessors and less timid than others.

The main principles of multi-factor investing

Multi-factor investing uses algorithms to select stocks on the basis of their ability to outperform regularly over time. There are four factors of recurring outperformance:

  • Value
  • Quality
  • Low volatility
  • Momentum

These factors can be presented differently (see exhibit 1 below).

 Multi-factor investing: the new generation of quantitative processes

Note: Intended only as an illustration. Source: BNP Paribas Asset Management, as of June 2018

“Virtuous” investment management calls on these notions:

  • Justice: if a stock is undervalued for the wrong reasons, it is worth looking into
  • Courage: if a company does its utmost to implement a quality development strategy, it is worth investors taking an interest in it
  • Prudence: overweight less volatile shares
  • Moderation: monitoring trend indicators is tantamount to acknowledging that some factors are unfathomable and that, although the market may not always be right, it is not necessarily wrong

Exhibit 1 may give the impression that the factors are in conflict, even though they are complementary and applying them leads to a consistent universe.

Of course, there are strategies based on just one of these factors. We propose a more reasonable approach that implements risk control at the level of each factor.

There is one more virtue that is important: humility, as we discussed in this article last year. Research has taught us that there is no point is trying to forecast market beta. Admitting that makes the factor approach even more attractive.

Implementing the factors

Details of how the main principles are implemented in our strategies were presented in this article. Below, we feature ideas that are characteristic of our multi-factor investing approach, while the factors chosen are rather consensual.

First of all, factors are neutralised for systematic biases, i.e. sectors, geographical regions and market capitalisation. Our rigorous approach has proven that the resulting portfolio is better immunised against global shocks.

Our risk allocation is constant over time between the various factors. We know that some factors are more efficient in certain markets, but adjusting weightings to reflect this would require determining in advance what configuration is in place. This approach, which would amount to market timing, is incompatible with the humility that guides us.

We have deliberately chosen to limit ourselves to four factors, which are complementary. For each factor, we have reduced the number of indicators under consideration as much as possible (see exhibit 2 below).

Multi-factor investing: the new generation of quantitative processes

Note: Intended only as an illustration. Source: BNP Paribas Asset Management, as of June 2018

We believe the nine indicators chosen offer a good trade-off between stability and efficiency in our model. This selection results from comprehensive research. Our criteria are not always the ones emphasised in more traditional approaches.

Portfolio construction: a chameleon strategy

As you have seen, our multi-factor investing approach is not mere data-crunching or a process in which each factor has been back-tested so that it provides a more flattering picture of past performance.

In the first stage, calculating factors provides a list of stocks that will be finetuned by risk allocation between each factor. On that basis, it is possible to apply specific filters to arrive at a final portfolio that will meet investor constraints, for example environmental, social and governance selection criteria (ESG) for reducing a portfolio’s carbon footprint, without introducing distortions in the process.

The next stage assumes that the starting universe is broad enough, so that factoring in constraints does not make it impossible to find, from among the stocks that emerge from the multi-factor investing process, stocks that meet specific needs.

Correcting the low-vol bias

Technically speaking, the “low-vol” factor introduces a defensive bias into our portfolio as it lowers beta. Since the portfolio is less exposed than the benchmark to equity risk, this undermines its long-term performance accordingly. Exposure to equity index futures can lower overall beta to a neutral level.


To find out more about our multi-factor investing strategies, click here. More posts by Etienne Vincent

Etienne Vincent

Global Head of Quantitative Management

Leave a reply

Your email adress will not be published. Required fields are marked*