We do not guarantee the accuracy of this transcript.
This podcast is for professional investors only.

Erika van der Merwe (EM): Quant equity investing has enjoyed the resurgence of value since late 2020. Meanwhile, the quest for innovation in capturing returns continues unabated.

Welcome to a new episode of The Robeco Podcast.

EM: Joining us to discuss the evolving priorities on the quant research agenda are Harald Lohre and Matthias Hanauer. Both are quant equity researchers at Robeco. Welcome, both.

Both: Thanks, Erika.

EM: So since the birth of quant investing, there’s been a steady rise in the number of factors that researchers claim can explain alpha. And I believe these factors now number in the hundreds, around 400, probably depending on how you count them. But it turns out that 15 factors are enough to capture most of the drivers of equity market returns. What does that mean, and what can investors do with that?

Matthias Hanauer (MH): Well, let’s get a step back and start [with] what factor models are. Factor models usually try to explain differences in stock returns with a parsimonious set of factors and factors that have a measure, like the return difference between stocks that really score well on certain characteristics versus stocks that score bad on these characteristics. And over time, the academic literature has brought up like more than 150 of these kinds of factors.

And on the other side, like common academic asset pricing models, use only a handful of factors. And so there’s a large gap between having a handful and more than 150. And our research was trying to find out what is really the number of factors that you need to tame this factor zoo. And it turned out to be for the US, that it’s like 15 factors are enough to capture all the available alpha in the zoo.

EM: Right. So we have the special quiz for you. Name as many factors as you can in one minute.

In tandem: The market factor. Value. Quality. Low risk. Investment. Residual momentum. Price momentum. Net share issuance. Earnings momentum. Net operating assets. Return on equity. Return on assets. Net operating assets. Liquidity.

EM: I think I’ve lost count by now.

MH: Debt-to-market-ratio.

EM: I think you are making this up. This is way too long a list. It’s a zoo.
Harald Lohre (HL): It’s a zoo.

MH: There are more than 100.

HL: We even have egg factors.

EM: What is it with the egg? Harold, I have to ask. There’s an egg with Harold.

HL: Let’s not talk about my egg.

EM: And is it a chocolate egg?

HL: It’s not a chocolate egg. It’s a genuine egg.

EM: What is that?

HL: It’s a breakfast egg.

EM: Which you found where?
[Buzzer]

HL: Thanks for the buzzer. [Laughter] Fits perfectly to diversification: Don’t put all your eggs in one basket. You see, I put my one egg in my one basket.

EM: It sounds like the Magnificent One. Egg-cellent. Which is the one factor that you would have chosen to own over the past year?

MH: The Magnificent Seven factor.

EM: That’s just plain cheating, Matthias.

HL: I probably have lots of eggs. [Laughter] So essentially, I mean, it’s all about the four bigger ones: value, momentum, quality, low risk. So a diversified version of those mixtures should make you very happy irrespective of the year.

EM: So you’re playing it safe. Really.

HL: Yeah. Going forward that’s the best you can do. Not just have one egg.

EM: Quite an admission from a quant. What’s the most important skill to have as a quant researcher? I’m going to give you five options. It’s multiple choice: statistical prowess, mathematical genius, people skills, being aware of your own personal biases, and knowing how to work your academic network.

MH: I would say knowing your own biases.

HL: Yeah, these all are super important and if you add curiosity, I think you can go a long way.

EM: So you admit to being a mathematical genius, Harold.

HL: We all are.

EM: What’s the most significant contribution made by quant researchers over the past 30 years?

MH: Translating academic insights into successful solutions.

EM: [Laughter] Are you passing, Harold?

MH: Nothing to add?

HL: It’s too good to improve.

EM: So, Matthias, you made in your very last point. So this is for the US market in particular. Presumably that’s where you had the data for. Would this also apply globally? Harold?

HL: Yep. I mean of course in academics that’s the most of the factors are – you kind of researched in the US and then like you look for it in Europe and in other regions and it’s also something we do and that’s also something we tested in this paper. And I mean this magic number of 15, of course, that’s not to be found everywhere. So if you look elsewhere, there are more nuances to it. But by and large, like the type of factors that are being chosen are very similar. And also the types of factor themes. So we started out value, quality, momentum. So these overarching themes, this is sort of an eternal feature, if you will, in the factor zoo. So these pieces, they kind of are here to stay.

But like what we found in the research is that like the underlying factor definition. So if you really come to ‘how do you measure value, how do you measure momentum?’ This is something that is changing through time. Meaning I mean, and it pretty much speaks to what we do in quant. Right? It’s not just, “Let's pin down these factors early in the 90s and just ride it through.” No, we can’t rest on our laurels. So we have to constantly innovate these factors and make sure that they stay relevant. So this is something we’ve seen in that recent research we’ve done.

EM: And generally speaking then is there agreement amongst academics on what you’ve just said that you can hone it down to 15 factors or 15 groupings of factors? Matthias.

MH: I think the bigger factors groups, there’s quite some consensus having valuation, momentum, quality, low risk, maybe some kind of short-term factors, maybe there’s some disagreement. But I think the favorite factor representation, that can vary among researchers. And maybe one point regarding this innovation, maybe you can think about your smartphone. There’s every year a new model for it. And typically some of these features of your smartphone like the camera, the chip, the battery: they are getting better, they’re getting refined, enhanced year by year. And this is also our work on the factor side.

So we start with one value definition with a momentum definition. And each year we try to slightly enhance these definitions by taking out some unreported risk or enhancing the returns. And this is a bit like the pros that is also done in the automobile industry or for your smartphone. So this is the process that we do in our daily work.

EM: Harald we can’t speak to quant researchers without mentioning value. Why is value so important in quants?

HL: I mean, ultimately what you what you want to have in – you want to come up with a value or like a fundamental sort of valuation anchor for a stock, right, to make your assessment. “Is this worth a buy or is this not worth a buy?” So in a way it’s like the most important factor. It’s been a bit disappointing for quite a while. So it’s been quite a struggle to keep to that concept. Probably a large part of my career at least. But still, it’s like the foundational factor, if you will. So if you lose trust in this one, you’re losing quite some of your foundation in building a quantitative model.

EM: So you refer to the difficult times. So give us a quick recap on that quant winter. Why didn’t value perform in 2018 to, was it, 2020?

HL: Well we are strong believers in the value factor, but this belief is more evidence-based and not beliefs-based. And uh, when a factor like value does bad over a time period like between 2018 and 2020, we really want to understand why. And therefore we looked at various things. But one thing really stood out, namely that the valuations of already expensive factors or expensive stocks, they became really even more expensive.

So for instance, when you look at evaluation of value stock, they typically trade at a forward price earnings ratio of around 10. And growth stocks, by definition they are more expensive. They were trading at a forward earnings surprise ratio of around 20. But then during this time period between 2018 and 2020, the valuation of these growth stocks that even went to levels like 30, 40, 50. And so it was really that the multiple expansion was causing that value did badly, but not that these value stocks had deteriorating fundamentals.

EM: There was a relative story. They just lagged. Right?

HL: Exactly.

EM: They were collapsing. So it’s, you know – when you when you read, listen to podcasts with quant or read comments, you still hear kind of almost the scarring, the emotional scarring from that time and then Harold, that’s what you said, the difficulty sometimes of just maintaining one’s focus.

HL: Yeah.

EM: Can you just elaborate on that? So it is as Matthias said, it’s evidence-based. So you are always going back to the science. It’s not an emotional thing. Forget your biases and fears.

HL: Yeah. I mean of course we’re confronted with this negative performance and the feedback coming from clients of course. And then you always have a very short horizon that you’re looking at. And that’s a bit where as a quant at least you can resort to your evidence, to longer sample and kind of get this confidence and be like, “We’ve done the research and we know these periods can kind of happen and we just have to stay with strong hands and pull through and not kind of leave that trade at the worst possible time.”

EM: So we’ve had this comeback in value since late 2020, which has been good for quant investing. But why haven’t all quant portfolios and quant investors benefited fully from this comeback? Matthias.

MH: Yeah, I think when you look at 2021 and 2022, I think value investors – so outperformance across the market and across our growth stocks. But in 2023 it was a bit of a mixed picture. So when you look at the average stock, value stock was still outperforming the average growth stocks, especially in emerging markets and developed markets as well. But in the US it was a bit mixed because there Magnificent Seven were really dominating everything: so both value and growth stocks. And when you had really high active value solutions, you had to also underweight in long-only portfolios these bigger stocks to achieve this activeness. And therefore you were underperforming maybe the market. But this was not like a value effect but it was more anti-mega cap effect in disguise.

EM: Do you agree Harold?

HL: Yep. Fully, so, and by design we are quants. So we are like, if you will very bad at forecasting single stocks. And the Seven is slightly better but it’s still very difficult of course. So that's a bit futile undertaking. So this is nothing we would try. But still it’s, given like the size of these mega caps, I mean, the name kind of says: these are mega caps, so they make up a large portion of the US market of the benchmark that we are measured against.

So we also cannot kind of put our head in the sand and just ignore them. They’re going to go away So we effectively researched ways to mitigate, become more resilient in our active capabilities with respect to this. We think of skewed benchmarks, right? This large concentration of mega caps. How can we deal with that? And that’s more like thinking not about the alpha side of the equation, but more like the risk portfolio construction side of the equation. How can we become more resilient in portfolio construction?

And I mean maybe sparing all the nitty gritty and details. But effectively what we’ve researched are ways nudge these mega cap stocks. So if they come across and have an expected return of something and some tiny stock is slightly better, I mean, to make the optimizer not ignore the mega cap stock, but go for it. So slightly nudge it, push it in, and thus mitigate the benchmark relative risk that comes from these stocks.

EM: Can you go one layer further, but still in a way that we can understand practically? How do you manage that risk then of a very concentrated portfolio?

HL: Effectively making sure that those Magnificent Seven stocks that make up this high concentration have a higher chance of actually being selected, whilst having very similar good ratings. Of course, we’re not pushing any mega cap in where we’re not convinced in terms of their performance potential. But I mean, the optimizer is just he’s not reading the news, right? He’s just kind of looking at the ranking. And if he sees tiny differences, just goes for the slightly better ones. So if you nudge the mega caps just a little bit, you have a higher chance of ending up with them and thus implicitly reducing the active risk that comes from having them.

EM: You use the word optimizer. So is this the model that you use to manage this kind of risk?

HL: Exactly. And this type of risk is something that you would not be able to capture, like with an off-the-shelf risk model. So you have to be more – think hard how to how to come up with these nudging type of solutions. So that’s a bit, nothing of the usual, if you will. But for sure, portfolio construction and risk research is also high on the agenda. It’s a bit – when we discussed beforehand and whenever I touched this, it’s less enticing, it seems, to people in alpha. Of course, alpha is great. We can all relate to it. But of course, it’s all about taking these factors, these alphas, into portfolio, making sure we have good transfer of the information and of the premium we wish to harvest.

EM: More and more, portfolio optimization is going beyond the traditional dimensions of returns and risk and looking to optimize even multiple variables. And increasingly I see it’s become fashionable to refer to 3D investing. So the implication is that there’s a third variable that you’re optimizing. And we hear that the third variable would be sustainability or impact. Have you done research on this and what does it mean for investors exactly?
HL: We’ve um just recently put out some work on ‘called 3D investing’. What is the, I mean, the two ‘Ds’ we’ve discussed extensively like risk and return, that’s like the first and foremost investment objectives of any investor. But what is the third one? That’s sustainability. Right. And this is front and center with many of our clients these days. As a quant, we’ve been doing this for ages. That’s what we would usually say, we’ve been doing exclusions in the old days, and we still do. You can apply a constraint to kind of get sustainability improved relative to a certain benchmark. But really what is different with 3D investing is taking sustainability and putting it like front and center in the objective function of your portfolio optimization alongside risk, alongside return. And, not treat it as an afterthought, really not.

EM: Not a side effect or a consequence of the other decisions.

HL: Exactly. Really targeting. So if you mention – you really have, if you can formulate an impact objective, really put that in because that’s also from some clients we see that’s really a demand. And people thinking along these lines and of course we're eager to provide solutions for that.

EM: So what examples would that be in terms of sustainability? What is the kinds of metrics that they would be targeting?

HL: I mean, the classic ones, of course, is in terms of getting the climate transition right. So monitoring carbon footprint, that, of course, is an important concern. If you think of why does sustainability, what we’ve investigated is just Robeco SDG so furthering Sustainable Development Goals, of course that’s front and center. And that can be all. That can just be a few selected. These would be like typical considerations. But sustainability is different for anyone. And in that sense it’s important to have something flexible, customizable. And if you can formulate your sustainability belief objective, we can put it in.

EM: So whatever the client imagines, you can do it.

HL: So you can reflect ‘50 shades of green’.

EM: How has AI and machine learning in particular changed the way you approach your work? And I’m not –I don’t mean the kinds of strategies or products you might be building, but just how you go about your daily work. How is it empowered you to work better? Matthias.

MH: I think for daily work. I use AI for sometimes writing papers, not from scratch, but if I have first draft, I use it for proofreading, for instance, and for more enhancements, as we said in the beginning, the enhancement of factors. You can enhance factor combination by using machine learning. So there’s more –there’s the statistical part of machine learning. In the past days, people came up with their final model by picking weights by the human.

But you can also let the machine decide how to combine these factors or signals. And usually it was done more in a linear fashion and with more complex machine learning algorithms. We can also include interactions and nonlinearities in this process. It was also possible before, but then you manually had to insert them. But you can also then find patterns automatically that present these opportunities.

EM: Harald?

HL: Exactly. I mean from a distance you could think “What has changed?” Actually not much. I mean, quant is effectively turning information into investment decisions. And in that sense, the machine learning toolbox is yet another toolbox that we can use to actually get the job done. And as Matthias mentioned, it gives you a way to more quicker come up with relationships that are otherwise hard to model or just to investigate.

But fully data-driven of course, we are still economic quants. We want to understand what’s happening. So we need to rationalize – sort of it’s elevating. It’s a natural evolution if you think of pure machine learning. But in the end it’s still like “What is more important?” Is it the information that we’re feeding, or is it the tools that we use to transfer? And probably it’s still information.

EM: Robeco’s quant equity team manages a range of strategies. So taking your emerging market strategies as an example. So take us through that. How would you apply some of these elements that we’ve just discussed to real life strategies that your clients use?

MH: Maybe let’s go back to the early days of quant emerging market strategies. And actually it started quant emerging markets as an idea generation for our fundamental emerging markets colleagues. So they already had a team in the 1990s investing in emerging markets. And they wanted to have a tool to quickly see which stocks are attractive based on certain factors. And our quant colleagues back then built this, and we saw that these lists actually work quite well, because back then there was quite some discussion if quant investing can work in emerging markets. Some people said you have to be fundamental in emerging markets. You have to be locally on the ground. Maybe you have to read between the lines to be successful in emerging markets.

But it turned out that these quant lists were quite successful. And as there was also client demand for some lower tracking error solutions, actually the first quant emerging market strategy was born, and this was more than 15 years ago. And it turned out that this approach is highly successful. And we see it, that it works in practice, not only on paper.

EM: Thanks, Matthias. Harold, in terms of factors that you use there, or in terms of this optimizer that you refer to, is all of this relevant for emerging markets?

HL: Of course, we’re using the same approach across the board, whatever regions. It’s the right factors and it’s the right transfer mechanism. And that has particularly worked out in EM as Matthias has said. The study we did effectively was then also not just looking into quantitative funds to see like, oh, bragging like, “How well did we do?” Actually, the question was a bit like, “If you have an information ratio of one or even above one, is that actually a good thing or can anyone do it? Is it that easy in emerging markets? And how do fundamentals do? Do you have to be fundamental to be even better?”

Or conversely, you could think like “Are fundamental investors just factor investors in disguise?” So kind of using what we use to begin with. But not telling people.

EM: I see what you did there.

HL: Of course that creates a bit of a tension, but, just as we work together at Robeco and that’s something I mean, I’ve been at other places before. What I’ve never seen is this healthy collaboration from both camps, so fundamental and quantitative, which is I mean: both sides are kind of eager to learn from the other side. So fundamentally using our quant rankings for information, but also us being eager to learn about what should we be looking at and all the practicalities, nitty gritty. So I think that’s great. And it shows in the track record. So ultimately it’s neither one of the two. It’s more like you can actually enjoy the best of both worlds by bringing fundamental and quant funds together.

EM: Matthias, what’s next for you on your research agenda? What frontiers are you pushing?

MH: Getting to the next model release and enhancing the factors. No, I think we work on various areas. We try to improve existing factors like value, momentum, quality, low risk, but you also look into new areas. So Mike Chen, we had him on the podcast. I think he has exciting research projects or we all have it. So looking into textual data, audio data. So I think this is really a new cool frontier for us that we can really – where we can really leverage these new technologies.

EM: And Harald.

HL: If you listen to Matthias, he’s fleshing that well out. And you see, life is better in a factor zoo than in a museum. So we constantly keep on innovating these factors as per our research of course. And next-gen elements, that’s a crucial element, of course, that’s bringing in new information and seeing what is the current – where are factors headed to keep them relevant.

But at the same time, of course, risk portfolio construction is also important. So thinking of sustainability as such for instance, I mean can you come up with risk factors here? This is quite a thorny topic as well because in the past people didn’t really care too much about these sustainability risks. Hence it’s really hard to pin down in the data. And you have to kind of approach it in more like a forward-looking manner and think about like what could be relevant, what should be used. And this is also something we should be investigating.

EM: Harald and Matthias, thank you so much for your time, for your insights. Good to have you here.
MH: Thank you, Erika.

HL: Thank you.

EM: And to listeners, thanks for joining us. Check out the full podcast series. We publish a new episode every month covering a range of investment-related topics. If you subscribe, you’ll receive a notification as soon as the new episode is published. In the meantime, please rate the show and share the show link with a friend. This monthly podcast and Robeco’s bi weekly podcast, In Tune With the markets, are available on all major podcast platforms and on the Robeco website. Until next time.

Thanks for joining this Robeco podcast. Please tune in next time as well. Important information: This publication is intended for professional investors. The podcast was brought to you by Robeco and in the US by Robeco Institutional Asset Management US Inc, a Delaware corporation as well as an investment advisor registered with the US Securities and Exchange Commission. Robeco Institutional Asset Management US is a wholly owned subsidiary of ORIX Corporation Europe N.V., a Dutch investment management firm located in Rotterdam, the Netherlands. Robeco Institutional Asset Management B.V. has a license as manager of UCITS and AIFS for the Netherlands Authority for the Financial Markets in Amsterdam.