From Alexander Weis and Gerd Kommer
This blog post is the second part of our trilogy on factor investing (Smart Beta Investing) and expands on the previous blog post published in May 2019 “Factor investing – the basics”. In May 2020 we started with “The Pains of Factor Investing“The third and final part of the trilogy was published. Interested readers who are not yet familiar with factor investing should read the first part first.
Anyone who has decided to take factor premiums into account in their passively managed stock portfolio and would like to do so with more than just one factor premium is faced with the fundamental question of how to practically implement such multifactor investing. We will address this rather technical question in this second part of our blog series on factor investing. So this is about the specific one Implementation from factor investing or to put it more casually: What is the smartest way to add factor premiums to your portfolio? In our experience, many private investors pay too little attention to this implementation question.
To illustrate the matter visually, we symbolically represent the global stock market as a cube, more precisely as the famous "Rubik's Cube"– see Figure 1 below. Our cube represents the approximately 9,000 stocks that are continuously traded in around 45 countries around the world. These 9,000 stocks represent around 99% of the market capitalization (market value) of the global stock market. You can use the stock market cube like the real one Rubik's Cube, split into 27 sub-cubes (= 3 × 3 × 3 = 27). In terms of the stock market, the 27 subcubes along the three dimensions of the cube - height, width and depth - represent the following three factor premiums: small size (depth), value (width) and quality (height). (We are aware that there are more than just three such “return dimensions”, i.e. factor premiums. For the sake of simplicity, however, we will limit ourselves to these three in this article.) The dimensions of the factor premiums just mentioned are highlighted in color in Figure 1.
Figure 1: The factor premiums Small Size, Value and Quality shown as the three dimensions of height, width and depth of a cube
► Source: Gerd Kommer Invest GmbH
In order to “exploit” the three desired factor premiums as an investor, i.e. to overweight them in their portfolio compared to a benchmark index weighted by market capitalization, the investor could now simply purchase three individual index funds: a global small-cap ETF, a global value ETF and a global quality ETF – i.e. one ETF for each of the three cubes. This method is often referred to as “simple multifactor investing” or “fund-level multifactor investing.” In practice, it is the most common method of taking factor premiums – often referred to as “factors” for short – into account in a portfolio.
However, this traditional route of multifactor investing suffers from a basic problem that many investors overlook: Anyone who combines the three factor premiums in this way sometimes - indirectly and unintentionally - also brings false, negative premiums into their portfolio; i.e. those that you just didn't want. For example, within a small-cap ETF there will inevitably be “growth-heavy” and “low-quality-heavy” small caps, since the stocks from the underlying stock universe were sorted exclusively according to the “small size” dimension for the purpose of forming the small-cap ETF.
This can be illustrated metaphorically with the following situation: Suppose you want to throw a particularly spectacular party in your garden. All guests must meet two criteria: First, each guest must be dressed entirely in red and second, everyone must bring a bottle of champagne. You know in advance that there will probably be more people coming to the party than you can accommodate in your garden; That's why you hire two bouncers, Mario and Luigi. To ensure entry requirements are met, Mario checks arriving guests to ensure they are dressed entirely in red, while Luigi ensures guests have champagne with them. Unfortunately, you have instructed Mario and Luigi separately, so that each of the two bouncers only checks the individual entry criteria assigned to them - there is no coordination between the two. Because this is the case, there will almost inevitably be party guests who are correctly dressed in red but do not bring a bottle and, on the other hand, guests who have champagne with them but do not meet the dress code. You specifically only wanted guests in red and with drink.
As bizarre as our party analogy may be, it illustrates the basic problem with traditional, simple multifactor investing. Several single-factor funds are combined here. The factors (filter criteria) are listed individually Fund level (Fund Level or Index Level) into the investor portfolio (the “Party”). You could say that dress code red corresponds to the small cap factor, the bottle souvenir corresponds to the value factor. However, despite the bouncer selection, there are also unwanted guests at your party, namely people who are not dressed in red and those who have not delivered a drink.
In simple multifactor investing, unwanted party guests are also present: the unwanted negative factor premiums, i.e. the wrong end of the factor spectrum, here “Large Size” (Large Caps) and “Growth”. They weaken and dilute the desired positive factor effect (the expected return increase) for the overall portfolio. This makes factor management of the portfolio less precise and less efficient overall. Nevertheless, this method is still virtually the market standard today.
Let's go back to our dice analogy (and put your spatial imagination to the test a little): If you combine three single-factor funds in a portfolio, you take 19 of the 27 sub-cubes into your portfolio. Each of the nine single-factor dice partially contains the wrong end of the factor spectrum of the two other factor premiums, namely large-cap, growth and low-quality stocks. These are the unwanted guests at the “multifactor party” because the selection process does not include all three factor premiums at the same time has taken into account.
Fortunately, a better mousetrap now exists, often referred to as “integrated multifactor investing” or “stock-level multifactor investing”. This improved multifactor approach works like this: Instead of combining three separate index funds, each with a single factor premium, upfront individual Index funds (or index) are constructed in which only those stocks that meet all three factor criteria at the same time are included from the outset. This means that portfolio construction happens on Individual stock level (Stock level) and not at fund level. This means no uninvited guests come to the factor party.
We have graphically represented the integrated multifactor investing process using the Rubik's cube in the middle of Figure 2. However, if the approach were followed consistently, the resulting portfolio would only consist of a single sub-cube.
Figure 2: Comparison of simple multifactor investing (fund level – the cubes on the far left) and integrated multifactor investing (stock level – cubes in the middle and on the right)
► Source: Gerd Kommer Invest GmbH
This shows the potentially biggest disadvantage of the integrated approach: the number of “super stocks” within the global equity spectrum (said 9,000 stocks) that have all three factor premiums at the same time without also carrying the wrong or negative factor premiums is relatively small. If the integrated approach is implemented too rigidly, the portfolio loses part of its high degree of diversification. We don't want that because broad diversification is just as important to us when investing as overweighting factor premiums.
However, the theoretical diversification disadvantage of the integrated approach can easily be addressed by loosening the individual factor definitions somewhat. This relaxation of the filter criteria is visually expressed in the Rubik's cube on the right in Figure 2, in which a total of seven sub-cubes fulfill all three factor premium definitions simultaneously. As a result, you can realize the advantages of more efficient integrated multifactor investing - a party without uninvited guests - with only minor losses in the degree of diversification.
Another advantage of the integrated multifactor approach is that it tends to result in lower transaction costs than the simple multifactor approach. On the one hand, the investor only has to buy one fund (instead of three) and on the other hand, the fund's internal securities turnover will also be lower in the long term. Both help to reduce costs. The latter costs, which arise from reallocations within a fund, are not directly observable for the investor because they depend directly on the performance of a fund, but that does not make them any less relevant than other types of costs.
In summary, we consider the essential facts from this blog post and the first part “Factor investing – the basics” stipulates: (a) Factor investing on a buy-and-hold basis produces a more attractive return-risk combination in the long term than passive investing without taking factor premiums into account, and this in turn produces a more attractive combination than active investing. (b) Anyone who adds more than one factor premium to their stock portfolio is practicing “multifactor investing”. (c) Anyone who wants to practice multifactor investing is faced with the fundamental decision of whether to do so by way of simple or integrated Multifactor investing should do. (d) We believe that the integrated approach is better. With integrated multifactor investing, the selection process for the stocks to be included in the portfolio takes place at the level of the individual stock (stock level), while in the simple approach it takes place at the level of a fund (fund level). (e) The potential disadvantage of lower portfolio diversification in the integrated approach is sufficiently mitigated by relaxing the factor criteria. (f) Due to the advantages of the integrated approach at the stock level level presented here, we assume that this method produces around 0.2 to 0.3 percentage points more return per year than the fund level approach (simple multi-factor investing) in the long term with approximately the same risk.
literature
Clarke, Roger; de Silva, Harindra; Thorley, Steven (2016): “Factor Portfolios and Efficient Factor Investing”; In: Financial Analysts Journal; Volume 72; No. 6; 2016, Corrected May 2017.
Dimson, Elroy; Marsh, Paul; Staunton, Mike (2019): “Factor Investing”; Credit Suisse Global Investment Returns Yearbook 2019; Long Version, Chapter 4, pp. 65-85.
Fama, Eugene; French, Ken (2018): “Volatility Lessons”; In: Financial Analysts Journal; Volume 74; Issue 3; 2018; pp. 42-53
Fitzgibbons, Shaun; Friedman, Jacques; Pomorski, Lukasz; & Serban, Laura (2017): “Long-Only Style Investing: Don’t Just Mix, Integrate.” In: The Journal of Investing; Winter 2017, Vol. 26; No. 4.
FTSE Russell (2018): "Top-down or bottom-up? Balancing exposure and diversification in multi-factor index construction" (source no longer available)
Innes, Andrew (2017): “The Merits and Methods of Multi-Factor Investing“
Kommer, Gerd; Weis, Alexander (2019): “Factor investing – the basics"; Blog post; May 2019
Kommer, Gerd; Weis, Alexander (2020): “The Pains of Factor Investing"; Blog post; May 2020
Scott, Louis/Cavaglia, Stefano (2017): “A Wealth Management Perspective on Factor Premia and the Value of Downside Protection”; In: Journal of Portfolio Management; Spring 2017; 43; No. 3.