In this episode I speak with Michele Aghassi, principal at AQR Capital Management where she serves as a portfolio manager for the firm’s equity strategies.

The conversation spans three major points: optimization, the opportunity in emerging equities today, and factor investing. While these are the headline topics, the underlying theme of the conversation, in my opinion, is the idea of unintended bets.

More specifically, how controlling for unintended bets, whether through optimization or thoughtful consideration, can sharpen your resolve in your conclusions. Whether it is the influence of China in emerging markets, the influence of currency in foreign equity returns, or crowding effects in factors, being aware of the potential for unintended bets can shape the how, where, and even the when of portfolio construction.

Please enjoy my conversation with Michele Aghassi.


Corey Hoffstein  00:00

Okay, are you ready?

Michele Aghassi  00:01

I’m ready.

Corey Hoffstein  00:02

All right 321 Let’s go.

Corey Hoffstein  00:09

Hello and welcome, everyone. I’m Corey Hoffstein. And this is flirting with models, the podcast that pulls back the curtain to discover the human factor behind the quantitative strategy.

Narrator  00:22

Corey Hoffstein Is the co founder and chief investment officer of new found research due to industry regulations. He will not discuss any of newfound researches funds on this podcast all opinions expressed by podcast participants are solely their own opinion and do not reflect the opinion of newfound research. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of new found research may maintain positions and securities discussed in this podcast for more information is it think

Corey Hoffstein  00:53

In this episode, I speak with Michelle Lagasse principal at AQR Capital Management, where she serves as a portfolio manager for the firm’s equity strategies. The conversation spans three major points optimization, the opportunity in emerging equities today, and factor investing. While these are the headline topics, the underlying theme of the conversation, in my opinion, is the idea of unintended bets. More specifically, how controlling for unintended bets, whether through optimization or thoughtful consideration, can sharpen your resolve and your conclusions. Whether it’s the influence of China in emerging markets, the influence of currency and foreign equity returns or crowding effects and factors, being aware of the potential for unintended bets can shape the how, where and even the win of portfolio construction. Please enjoy my conversation with Michel Lagasse.

Corey Hoffstein  01:51

Michelle, welcome to the podcast. Thank you so much for joining me really excited to have you here.

Michele Aghassi  01:57

Well, thank you for having me here. I’m excited to be with you today.

Corey Hoffstein  02:01

Let’s begin at the beginning. It’s always typically the best place to start. I guess with your background. Can you lay it out for the audience, where you got started, how you got interested in the industry and walk us forward until today?

Michele Aghassi  02:15

Sure, so maybe I’ll go backwards. So I’m a partner at AQR. I’ve been with the firm for 18 years. I’m a portfolio manager for the full range of global stock selection strategies. And that ranges in geography, market capitalization long only strategies hedge fund strategies. I’m one of four partners within the global stock selection team. And we set the group’s priorities and we critique and vet all research before any innovations make it into our process. Those have to survive the challenging questions that we pose and ultimately get our seal of approval. Over the course of my career at AQR, I’ve made some noteworthy contributions, I lead the development of our emerging equity strategies. These include large cap and small cap offerings, long only as well as market neutral capabilities. And they’ve become an important part of AQ ours lineup. For a number of years, I co headed our research effort in the stock selection group. And in my earliest years at AQR, I redesigned our approach to portfolio optimization. And that’s where my academic background is. So prior to joining AQR, I had worked in the industry in quantitative finance, previous to joining AQR, but right before I joined AQR, I did a PhD. in a field within applied math called operations research, which centers around large scale optimization problems, I’d really always been interested in finance, I think, largely because the problems are complex and difficult, and they’re not necessarily neat and tidy. And they represent a very rich area in which you can apply mathematical tools. And I find that I find that exciting.

Corey Hoffstein  04:03

I don’t think I’ve ever had a guest give me their background, backwards. I think I prefer it that way. Going forward, I’m gonna ask all my guests to give me their history in reverse chronological order. Well, let’s start with some of the optimization work you’ve done. We’re going to touch on the different parts of your career, we’ll dive into emerging markets a little later. But as you mentioned, your academic training is an applied math and optimization and you started your career really focused on that optimization piece. wanted to start with maybe what is a simple but potentially difficult question, which is, why is optimization an area worth focus? Why is it something that’s important in the first place? Why our heuristic approach is not sufficient, particularly when you look at the literature that suggests that so many of these naive approaches even something just as simple as, say, one over n are incredibly difficult to beat out of sample?

Michele Aghassi  04:55

Yeah, you know, it’s it’s a good question. Certainly, there are some very simple approach is out there, and they can be appealing due to their simplicity. But one thing I would point out is that they’re often compared apples to oranges to more complex approaches, you know, think about something like transaction costs. Often the equal weighted approach that you mentioned, you know, it’s hard. As you mentioned, some papers talk about it being hard to beat a one over N approach and equal weighted approach. But often, that approach is heralded in the context of its performance on paper, gross of transaction costs. And interestingly, maybe not surprisingly, its transaction costs can be several times that of many other rules based or factor based methodologies, methodologies that use optimization. So clearly, one needs to look at performance net of transaction costs. We don’t live in a world without tea costs. But it’s not just costs that are important. There’s also the question of constraints or requirements. So a portfolio may have certain constraints, whether driven by significant preferences of the investor or whether driven by rules or regulations, those constraints may relate to risk limits, they may relate to limitations on exposures. And particularly if you need to juggle a variety of such requirements, simple heuristics are probably not going to cut it, they’re probably going to violate those requirements a lot of the time. And you can try to adapt those heuristics to satisfy multiple requirements. But often that attempt to adapt ends up looking like a game of Whack a Mole, where you make one adjustment, you know, there’s a step in the now brought in turistic, where you make an adjustment to try to satisfy one requirement. And somehow in the process, you end up throwing off kilter all of the other requirements and just sort of end up going around in circles. So optimization is extremely useful, in that it can juggle all of that simultaneously, and over a potentially very large universe. So it can juggle, you know, return and risk specific exposure requirements that might relate to country or industry, among other things. And of course, it can also think about and plan for and build the portfolio cognizant of transaction costs. For all those potential benefits, optimization has a reputation of being particularly fragile. How do you think about those trade offs, the potential benefits that optimization might bring versus the risks that it might accidentally introduce into the process? It’s a very good point, you’re absolutely right, that optimization can indeed be fragile. And it certainly has that reputation optimizers. They’re incredibly precise and powerful machines. And in contrast, in finance, there’s usually an inherent lack of precision, because we’re often working with low accuracy predictions. And unless you tell an optimizer otherwise, it thinks that all of the inputs are known with certainty. Basically, it sort of assumes you’re psychic. So as a result, it makes sense that the outputs often end up being overfit to those inputs. And that’s essentially where the fragility comes from you basically, you know, if you’re not telling the optimizer that there’s uncertainty, you’ve asked for the very best solution that’s entirely tailored to those precise inputs. So you’re, you’re basically asking for overfitting, you’re asking for fragility. And that can have a number of negative side effects on the portfolio, the portfolio might become more concentrated than it should be, essentially an exaggerated distortion of your inputs, or it can end up over trading on what is nothing more than meaningless wiggles in the data from one period to the next. And in addition, it can take on significant extra risk in order to lock in an insignificant amount of extra return. So certainly, there are fragility issues, and they can have very real consequences. And these issues are not to be taken lightly. But the good news is that there are effective ways of mitigating these unwanted behaviors, you can formulate the problem in a way that is conscious of uncertainty. Whereas in contrast, you know, as we talked about, fixing the shortcomings of heuristics tends to be much harder. When there are many requirements at play. We talked about that resulting game of Whack a Mole, dealing with the fragility of optimum

Michele Aghassi  10:00

zation is actually one of the first challenges that I tackled at AQR, I developed a proprietary optimization approach that directly deals with this uncertainty. And I in doing that I drew on concepts for my PhD dissertation, which wasn’t about finance, but was centered around the topic of robust optimization, which is a field within optimization that focuses on how you handle uncertainty in a way that makes sense where you’re not replacing hefty assumptions about what you know, by other hefty assumptions about what you know, it’s actually thinking about what you don’t know. And it’s thinking about how you deal with that uncertainty in a way that’s computationally tractable.

Corey Hoffstein  10:43

Sounds like some of the alpha from this podcast will be to the listeners who can go dig up that PhD dissertation and give it a read. So I want to transition now, towards your work in emerging markets. You recently authored a piece called reemerging equities. But before we get specifically into the nuances of the piece, I’m going to risk asking what is potentially an overly large question because I think it applies to optimization. I think it applies to thinking about different regional markets. And the question is, when you’re thinking about investing us specifically in perhaps more AQR at large, is the approach something that’s global? Or is it something that needs to be specifically tailored to the investment universe you’re working with? And you know, how does the choice of that investment universe impact the approach whether it’s the optimization approach you have to take or perhaps some of the metrics you have to look at within that particular universe?

Michele Aghassi  11:38

Right. So when you look across different equity universes, whether different geographies large or small cap, among other things, I strongly believe we strongly believe at AQR that there are far more similarities than there are differences. And that’s true of the characteristics that predict future relative return. It’s also true about the general overarching structure of the design choices for how to best build portfolios around these ideas. And, you know, the empirics strongly support this notion that the similarities dominate the differences.

Michele Aghassi  12:13

Why is that intuitive? Well, you know, it makes sense intuitively, because if you think about the characteristics for a minute, you know, which characteristics do well at predicting future relative return, many of the characteristics that we consider, and certainly other quantitative investors consider, they’re based on either behavioral biases. And those tend to be fairly universally applicable, or they’re based on some idea of being compensated for a risk. And that also tends to be fairly universally applicable. So there are a lot more similarities than differences. But there can be differences. And I’m not going to go into all of the different things that might drive differences. Some of those maybe are a little bit more intellectual property related, but maybe I’ll point out some of the more basic ones, you know, you might have differences in data availability, you might have legal or structural differences between markets. On the topic of data availability, it’s certainly not uncommon for novel data sources to become available in the US first. And on the one hand, that might be a limitation, it might be just that a limitation for X us. On the other hand, it might actually inspire innovation, right? Necessity is the mother of all invention, it might inspire innovation toward how to find alternate ways to capture that same idea. But in the absence of whatever that special us only data source is. And as an example, that’s opposite in spirit related to regulation. You know, there may be instances in which a higher degree of regulation, say in the US or in developed markets, may render some types of information less helpful in those more regulated environments. And that may actually present more of an opportunity outside of those markets. But to put it succinctly, our approach intentionally tends to be a global one, with some variation reflecting nuances between these different universes. You know, I think that it makes intuitive sense that the similarities would dominate the differences, and certainly the empirics are strongly supportive of that.

Corey Hoffstein  14:26

You recently authored a piece titled reemerging equities, where you argue that the expected premium for emerging markets over developed markets is generally generationally wide today. I was hoping you could lay it out for me, why are you so excited about emerging markets today?

Michele Aghassi  14:44

Yeah, there are a couple of reasons to be very excited about emerging markets today. The first relates to what you mentioned, which is regarding the asset class itself, and its return expectations. The second relates to an opportunity for alpha. Regarding the asset class itself. We estimate that emerging equities are poised to outperform developed equities by more than usual by more than we’ve predicted over the last 25 years. And I should caveat here, that it is very difficult to predict returns for asset classes. That being said, yield based methodologies do demonstrate some predictive ability, over a five to 10 year horizon. And if you take that methodology, that’s where you get that emerging equities over the next five to 10 years, we expect will do better than developed markets by a degree that’s greater than what we’ve seen as that prediction over the last 25 years. So the asset class itself looks especially attractive now, relative to developed markets. And before I get to the alpha potential, you know, let me just take an aside to talk a bit about the risk profile, which I think is important. And I also want to mention one other thing, which is, ironically, anecdotal evidence suggests that many institutional investors may currently be underweight, emerging equities relative to developed markets. And we know maybe that’s intuitive in that probably there’s been some degree of performance chasing over the last decade, certainly developed markets, particularly the US outperformed emerging markets meaningfully. And it happened to be driven really just by multiples expansion. in developed markets, particularly the US, I mentioned risk profile as something you want to consider in the context of this predicted premium, you might wonder whether that premium is coming with a catch in the way of risk. And, of course, historically, emerging equities have been far riskier than develop equities. But they’ve actually in the last five to 10 years become much less risky than they once were. In fact, over the last five years, they’ve realized about the same level of volatility, as developed markets. And that occurrence is not something sudden, basically the realized volatility level has been trending down for some time in emerging equities. In addition, they’ve become somewhat more diversifying of developed markets in the last decade, empirically than they were in the previous decades. So we think that they are poised to outperform by more than usual, and their risk profile has become more attractive in the sense that lower risk and somewhat higher degree of diversification. Now, we’re not just excited about emerging equities because of the expectation for the asset class itself. There’s a second reason why we’re very excited. And that’s that we also see a particularly attractive opportunity for active investing in the space, cheap companies in emerging equities have become far too cheap, versus their same industry, expensive peers. And that’s been true in a way that is not justified by things like recent and predicted earnings, cash flows, dividends, so there’s a dislocation. And if history is any guide, these types of dislocations tend to correct. And that correction tends to bring handsome returns to active value investing. So we think not only that the asset class looks more attractive than usual generationally attractive, we also think that now is an especially good time for active investing in emerging equities.

Corey Hoffstein  18:48

Whenever you talk about emerging markets, it’s hard not to talk about the big hanging idiosyncratic risk that is China, China’s come to compose a huge part of emerging market indices, though, somewhat ironically, we could say the same thing about US equities and global stocks. But particularly in China, I think people perceive more regulatory risk or just other risks that perhaps aren’t similar to the way us dominates global equities. If you were to look at this opportunity with an perhaps without China included within the set, how does that impact the results? How are the conclusions impacted if you ignore China?

Michele Aghassi  19:28

Sure. Well, the answer there is very simple. Removing China affects the results very, very little. We’ve actually done the same analyses that I described just a moment ago, using an emerging X China universe instead. And the conclusions are the same. The numbers that you get are very, very similar. That being said, like you mentioned, China is the largest weight by far of any of the countries in emerging markets, when you look at common emerging indices, and although yes, certainly the US dominates developed markets by far more and as you mentioned, China does pose some unique risks, whether with respect to geopolitical tensions, whether with respect to government regulation or crackdowns on certain industries. You know, I will say, I think there was a lot more rhetoric there than there was actual action, there was some action, but far more rhetoric. There have also been regulatory frictions coming from the Chinese government, as well as from the US actually imposed on us investors. The thing that’s tricky here is that no one knows if or when any of these issues will give rise to larger problems. And given how broad how liquid, the Chinese equity market is, particularly for a tracking error conscious investor, it’s tough to shun China, not knowing if or when that might be the right thing to do. There’s obviously an opportunity costs in the meantime, are a potential opportunity cost in the meantime. And certainly, you know, the other thing I would say here is that these issues, these frictions, if they were to become a bigger problem, in all likelihood, they wouldn’t just be an issue for China. I mean, when you think about geopolitical tensions, for example, if those were to inflame further, it would cause massive global supply chain disruptions, potentially. So it wouldn’t just affect China, it would affect the US it would affect other developed markets. The good news is that many of these frictions have appeared to if not de escalate, then at least decelerate in the last year or so. And I think the big picture takeaway in terms of what investors should do is that certainly an investor will want to understand their investment managers philosophy and approach on such issues. How does that investment manager deal with these types of risks?

Corey Hoffstein  21:52

In the piece, you note that emerging country’s GDP has actually outgrown develop country’s GDP. But GDP growth doesn’t necessarily translate into earnings growth. Doesn’t that challenge your conclusions potentially about the future expected return advantage of emerging equities over developed equities?

Michele Aghassi  22:13

Very good point GDP growth does not necessarily translate into corporate earnings growth. And to be fair, emerging equities have over the past decade or so lagged developed markets, particularly the US on corporate earnings growth, despite the fact that they’ve beaten develop markets on GDP growth. But the important point here is that our asset class performance expectations are not based on aggressive earnings or earnings growth assumptions. They’re not based on the rate of GDP growth that we’ve seen in computing those return expectations. We assume only a 50 basis point annual real earnings growth advantage for emerging equities over developed markets. And we would actually come to similar return premium conclusions, if we dropped that growth advantage down to zero, meaning if we assumed that emerging market companies were not going to grow their earnings any more quickly than developed market companies. So we’re making a very conservative assumption about growth. And in addition, the starting level of earnings that we’re considering in this analysis, and that’s the level to which that growth, that very conservative growth rate is applied, that starting level of earnings is itself another quite conservative assumption, we are using a trailing 10 year average earnings in this analysis, rather than something forward looking rather than something with a shorter trailing window. And so the return premium that we derive or that we compute is estimated based on an in spite of that last decade of emerging equity, earnings being disappointing. So the results that we come to are in no way dependent on a close relationship with GDP growth and corporate earnings growth.

Corey Hoffstein  24:14

You alluded to this point in one of your answers a little bit earlier, but on our pre call, you really explicitly said that there are more similarities than differences between developed and emerging markets. And I was hoping maybe you could pull on that thread a little bit, talk about some of the similarities that listeners may find surprising. And then conversely, some of the differences that remain critically important to be accounted for.

Michele Aghassi  24:41

Sure. So like we talked about a little while ago, theory and empirics. Both support that there are far greater similarities and differences. And we talked about why that’s intuitive given that, you know, a lot of those ideas are based on either behavioral bias or risk based explanations. I have to say we’re not I’m really surprised by the similarities, it’s a little hard to imagine why anyone should be. But, you know, as far as its differences, I would say those tend to be more at the level of beneficial to consider rather than critically important. I’ll give some examples that are maybe more obvious. We talked a little bit about some before having to do with data availability and regulation, let me maybe give a couple more fairly obvious less intellectual property related examples. Think about something pretty basic, which is that there’s a higher degree of event risk in emerging markets in amongst emerging countries than in developed markets. So you want to ask yourself, how do you account for that in the approach that you take, as another example, there’s a broader range of growth and interest rate differentials among emerging countries than developed market countries. And that can skew potentially some of the preferences or views that you form, if you ignore those differences. And at a more nitty gritty level, there are also accounting standard divergence in emerging markets, less than it used to be. But you know, that’s sort of really not an issue in developed markets, but still somewhat present in you also want to think about how you deal with that. Like I said, though, I don’t know that any of these is going to result in something night and day, whether you take it into account versus not, but it is beneficial to think about these sorts of things. Among some other differences that we won’t go into today, we’ve talked quite a bit about sort of the opportunity at the asset class level, you briefly touched upon the opportunity, maybe what I’ll say is the alpha level,

Corey Hoffstein  26:38

I want to dive into that a little bit more deeply, because the paper highlights the opportunity specifically for value investing within emerging markets. I was hoping you could discuss how you think about measuring this opportunity, and why we should have conviction that the opportunity today isn’t the result of just say measurement noise or something unintended bad, but there really is a real broad opportunity for value investing.

Michele Aghassi  27:04

This is a very important question. And you want to you want to be quite certain that you’ve examined this and sort of played devil’s advocate in trying to look for drivers related to noise or measurement issues. So this is something we’ve taken very seriously. You know, over time, the attractiveness of value investing can vary. And we measure that attractiveness through something that we call the value spread. And it’s basically asking the question, how much bigger than usual, is the discrepancy between cheap companies price multiples, and expensive companies, price multiples. And to be more precise, we look at that at a version of that, that is, essentially country and industry neutral. So how different are multiples when looking within the same country and same industry at cheap versus expensive? That discrepancy that value spread is at notably, historically elevated levels? So it suggests a very attractive tactical opportunity for what is actually also a strategically very beneficial investment approach value investing. But how do we know this opportunity is real and not a mirage? That’s an important question, we gained confidence that it’s not a mirage by looking at a number of things, and I’ll highlight a few of them here. First, the fundamentals of cheap companies are not worse than usual. In fact, if anything, they are somewhat better than usual. Think about the growth advantage that expensive companies typically enjoy over their cheaper peer counterparts. Of course, that growth advantage tends to be positive, right? Expensive companies do tend to grow faster than cheap companies. That’s part of the reason why the opposite of value is often referred to as growth even though technically speaking, the opposite of cheap is expensive, not growth, there is a growth advantage, but it’s actually been smaller than usual lately. Me meaning that cheaper, you know, same country, same industry, peers have actually kept pace better with their expensive counterparts on growth than they typically do.

Michele Aghassi  29:24

Second thing that I would highlight is that for today’s prices to be justified, you can actually back out what type of a growth advantage you’d need to see over the next five years. So take it as if those prices are justified. What would be the rate at which you’d need to see expensive companies outgrowing their same industry peer cheap counterparts? That growth advantage that is implied by today’s prices being justified? Is an amount that is double the maximum growth advantage we have ever

Michele Aghassi  30:00

I’ve seen in emerging markets. And that maximum that we’ve ever seen is about double itself, the long term historical average, which happens to be about the same as what analysts are predicting today. So could something like that happened for a handful of stocks that turn out to be unicorns? Yes, absolutely. But it’s, it’s highly unlikely that we would see a growth advantage. That’s four times the long term average, two times more than the largest we’ve ever seen, for a highly diversified basket of many expensive stocks versus their cheap, same industry peers. I think that’s a very compelling point. I should also mention that we’ve looked at intangibles on this topic that the question of intangibles often arises, basically, is this all driven by intangibles are failing to account for intangibles? And the answer there? Two is no. You basically get the same results. If you take out intangibles Heavy Industries, and just look at the remaining universe. If you use measures of value that are adjusted for intangibles, you basically come to the same conclusions.

Corey Hoffstein  31:10

Taking the investment opportunities set a little more global backing away from emerging markets a little over a decade ago, you co authored a paper titled avoiding unintended country bets in global equity portfolios. And I think this is a really critically important area. This idea of unintended bets has been something I focused heavily on in my career. And I think when you start talking about moving from your domestic bias to international markets, these unintended bets are everywhere. So maybe you could talk a little bit about what are some of the unintended bets that you think are the most important, why they’re important and how people can think about dealing with them in the investment process?

Michele Aghassi  31:49

Sure. So I would define unintended bets as essentially accidental exposures that you might take on when building a model or when building a portfolio. And by the way, this is relevant, probably even more so for discretionary managers. As it is for quants, you’re trying to capture a and you accidentally get a bunch of be mixed in with it. These unintended exposures are important to be aware of because they contribute or can even dominate a portfolio’s risk. And they might not be compensated, they could even introduce potentially a headwind into performance. So in that paper, we specifically looked at this from the angle of unintended country exposures, you know, think of an investor seeking to focus solely on stock picking.


Imagine that investor compare stocks on basic characteristics. Without regard to country membership. This approach of ignoring country membership can produce a portfolio that’s dominated by country risk, rather than by idiosyncratic stock risk, which is, you know, supposedly, what this investor was focusing on. And in fact, the extent of that country risk domination can vary pretty wildly over time. In addition, in the paper, we explored the conclusion that the country exposures that fall out of this process can look very different from what you’d get if you were more intentionally directly comparing the countries themselves. So how to deal with unintended exposures? Well, the first step is to define what’s intended versus not. And the second step is being able to measure how much of what you’re getting is intended versus not. More specifically, you need to think about, in what context are certain characteristics, informative versus not. And there, there are lots of examples of this. But I’ll give one example, you know, if you think about differences in legal systems between two countries might drive a tendency toward equity offerings in one, and debt offerings in another. And if that’s the case, levels of corporate leverage might not be comparable, from one country to another. So you need to think about what information is relevant where looking at differences in leverage within a country at individual stocks within an industry inside that country. That may be very informative, but looking at levels of leverage from one country to another, or between a stock that belongs to one country and a stock that belongs to another country might be less informative. So you have to ask yourself, what information is relevant where and then how can you build your model and portfolio so that you avoid using the right kind of information in the wrong place? In other words, what kinds of opinions are giving rise to what kinds of tilts and in our investment process? We try to build things in a modular way so that we’re only pairing the right kinds of information with the right kinds of potential tilts?

Corey Hoffstein  35:04

Well, on that thread of the right kinds of information when it comes to investing internationally, I think one of the unintended bets that can creep in that’s particularly complicated is the idea of currency. For example, I think there’s a broad open question as to what currency even matters? Is it the currency the stock is listed in or the currency of the trading partners of the underlying firm, whether it’s where expenses are generated or revenues generated? I’d love to get your thoughts on this problem.

Michele Aghassi  35:33

It’s a great question. And the answer is that it’s both currencies that matter. Clearly, there’s a mechanical link between the currency in which the shares are denominated and the value of those shares to an investor with a different home currency. On the other hand, there’s an economic exposure to currencies of trading partners, the more the trading partners currency depreciates, the less of your goods they can afford. And conversely, the more the trading partners currency appreciates, the more of your goods they can afford. So we think about this problem by thinking about it modularly, we can predict the currency returns separately from the local effects returns to the stock or basket of stocks. And this, this not only makes the problem more manageable, but it also helps with avoiding unintended bets, like we talked about using the right information in the wrong places.

Corey Hoffstein  36:28

So I want to sort of round out the end of this conversation with a discussion of factors, which AQR has certainly been on the forefront both vocally in the public sphere, discussing factors but also academically. More recently, Cliff has been incredibly vocal about the opportunity and value we talked about the opportunity and value in emerging markets. But just more broadly, globally, as well as in the US. I think Cliff term to the everything bubble is the phrase he’s been using. So we’ve we’ve actually, if you look at the value spread come back a bit from peak levels in 2020. But it certainly hasn’t been a smooth or straight ride back. How do you think about sizing a bet on something like value where you’re at these extreme levels, but the recovery could be incredibly volatile.

Michele Aghassi  37:17

Recoveries of any kind tend to be rocky, right? I mean, stock market recovery is certainly an example of that. And value recoveries are no different. As you said, value has made up meaningful ground from its 2020 lows when measured by returns to the factor or at least our version of value. And yet, when we look at what is now cheap, and now expensive, I should mention in stress that there is still a historical opportunity here, there’s still plenty of dislocation correction left to go. As far as how we think about sizing a tactical bet on value, it’s important to keep in mind that that rockiness is a part of the recovery package, so to speak, to take advantage of this correction, you need to be willing to ride out those bumps. But at the end of the day, how you size that bet? You know, of course, it’s an important question. And the answer is as much judgment and experience as it is science. So of course, there’s a science that we talked about of measuring just how wide value spreads are, how large is that dislocation. But judgment also has to come into play for at least a couple of reasons. We just don’t have a lot of past similar observations to allow the data to decide for us, you know, we have the tech bubble. And to some extent, the value dislocation around the the GFC. Also is another observation, but I would say the comparison there is not as close. So we just don’t have that many observations, can’t let the data determine the decision for us of how to size it also has to depend that sizing of that tactical bet has to depend on context, how you size that in a levered hedge fund will be entirely different from what a benchmark relative traditional long only investor can stomach who might have quite a limited tracking error appetite. So you want to size it in a way that doesn’t prevent the investor sticking with it. Because of course, if you throw in the towel too early, you forego the benefit of that correction. Being at the forefront of the factor space. I know. AQR has a lot of conversations and receives a lot of feedback on factors which led you to co author a new paper called fact fiction and factor investing where you provide evidence for some of the facts and against some of the fictions that frequently arise in the conversations you’re having. In the paper. One of the things you address is the the idea that quote this time is different as it relates to the poor recent performance of many factors, not just within equities, but in a multi asset.

Michele Aghassi  39:59

The landscape as well. So the open question would be, is this time really different? You know why or why not? How is that impacting your outlook? The world is always changing. And in that sense, this time is always different.  But that’s not really the right question. The right question isn’t, is this time different? The right question is, is this time different enough? For it to matter?

Michele Aghassi  40:27

And the answer, there is no, this debate, you know, of course, this time around, it was about value and whether it’s broken, because the world has somehow changed. Certainly, the idea that assets should converge to a fair price that there is some notion of fair price, whether you can measure it or not. And that assets should converge to that fair price. That’s a pretty timeless idea. And I think, quite a theoretically correct idea. It’s never going to make sense to overpay for something that’s worth far less. So the question becomes whether your proxies for fair value have somehow become obsolete. That’s basically what is this time different getting at? No proxy is perfect. And by the way, that’s why you want to use multiple proxies. But what we observed with values struggles in the years leading up to and including 2020, doesn’t seem to have anything to do with imperfection based on what the empirics tell us. You know, imagine for a moment, imagine that you had perfect foresight, and that you invested in favor of cheap price to earnings companies and against expensive price to earnings companies. Like I talked about before, let’s say you do this in a country, neutral and industry neutral way. But you’re making those preference decisions based on versions of P e, that are using an E and earnings

Michele Aghassi  41:57

with perfect foresight of what those earnings are going to be that perfect foresight strategy, of course, totally unrealistic. But I’m using this stick to come to a point. Of course it works smashingly Well, most of the time, but it actually lost money during two distinct periods. Guess which two? The first was the tech bubble in recent decades. And the second was the recent value drawdown. And I think that that simple thought experiment using that totally unrealistic, perfect foresight strategy provides fairly compelling evidence that this value drawdown wasn’t about metrics becoming outdated. It was about a sentiment driven shift in market preferences. And you can think about it from a different angle. We touched on this in passing a little while ago, one claim that’s been put out there is that value metrics have become outdated. Because of intangibles. You know, first of all, if what you’re comparing is cheap versus expensive, within industry, accounting for intangibles is going to be far less of an issue, perhaps still an issue, but certainly far, far less of an issue than if you’re comparing across industries. But we can actually let the empirics speak for themselves, accounting for intangibles. And we use some measures of value that adjust for intangibles. those intangibles adjustments wouldn’t have allowed you to avoid the recent drawdown in value number one, and they don’t explain away how wide value spreads are today. Number two, making those adjustments does add some value over the long run to performance. So you know, it makes sense to include them in an assessment of cheap versus expensive, but it wouldn’t have really changed history as far as the drawdown, and it wouldn’t change the forward looking opportunity. So bottom line, yes, the world is always changing. And that’s part of the reason why you want to continue to innovate. But hopefully that innovation isn’t your only line of defense. You also ideally want to use factors that are robust to changes in the world. But we don’t think that this time was different enough. It’s a very convenient excuse or explanation as to why one sees negative performance. But it just doesn’t line up with the fact pattern in terms of that drawdown and the dislocation that we see today.

Corey Hoffstein  44:32

One of the big push backs you often get with factors, particularly given the decay in factor performance over the last five years is that they simply became crowded. AQR was too good at promoting factor allocations among institutional and retail investors and that they become too crowded and the performance has been crowded out. Now this is almost immediately contradicted by your last answer where you talked about that it was a sentiment driven shift in value. But maybe we can talk about the factors at large. How do you think about trying to measure whether a factor is or is not crowded? How do we know that this performance hasn’t been simply crowded out?

Michele Aghassi  45:15

Right, these investment themes, value momentum, carry defensive, they’ve been known about for some time. And there are actually a variety of ways of measuring their crowdedness. And that can fluctuate through time. One way of measuring, we think probably one of the better ways is actually a method that we already discussed, which is looking at value spreads. So if you think for a moment, just to make it simple, let’s think about the factor as having a long side and a short side, if your factors long side is more expensive than usual. So I’m talking here about a market neutral factory. And obviously, if we’re in a long only benchmark relative context, you could just think about overweights and underweights. But anyway, if your factors long side is more expensive than usual, in a price multiple sense, and if the short side is too cheap, that would suggest crowding, right? There are too many people that like the same long side that you like, and that’s pushed up prices, the opposite would suggest a lack of crowding. So if your factors Long’s are cheaper than usual, versus your shorts, when we measure things in this way, we don’t see meaningful levels of crowding in any of the investment themes that I mentioned a moment ago. And for what it’s worth, value was actually one of the less crowded than usual factors during its period of struggle and leading into its period of struggle.

Michele Aghassi  46:43

Now, still, you might wonder, why is it that the factors aren’t more crowded, if they’re so known, and they’re so effective? First of all, the behavioral biases on which these factors are based, those are extremely widespread, entrenched, persistent, human beings just aren’t becoming more rational enough to get rid of the sizable other side that allows these factors to persist. And in addition, these factors are far from a free lunch, right? I mean, they can go through, we’ve lived through this, they can go through protracted periods of very difficult performance, like we saw in recent years with value. And that potential pain itself actually keeps many investors away and serves as a barrier, if you will, to crowding.

Corey Hoffstein  47:38

In the paper. One of the things you suggested that the factors tend to work well across a variety of macroeconomic conditions. But there’s been a decent amount of literature written and many investors would passionately claim that the factors actually work well in different conditions that, for example, we know that momentum tends to crash during a recession, and that we should actually be potentially timing our factor exposure. Can you sort of reconcile this for me? There’s obviously passionate people on both sides of the debate. Are both conclusions right, is one clearly wrong? Why not try to time or factor exposure based on macroeconomic conditions?

Michele Aghassi  48:25

Yeah, you’re right. Momentum does suffer in distinct environments, I would frame it more as momentum tends to crash in sudden and stark reversals. Those don’t necessarily line up with recessionary periods. But sort of getting at the broader question of how is it that both sides can be right? One side may be saying that the factors tend to work well across many macro economic conditions. And the other saying no, that there’s actually some degree of specificity there. Well, factors go by the same name, but can be implemented in very different ways that can look very different. So you know, what exposures are you neutralizing to what unintended bets? Are you allowing or not allowing to creep in? Are you constructing value in a way that is industry neutral? Or are you not making that neutralization and potentially allowing value to become dominated by comparisons of cheap versus expensive industry to industry? If you think about a market, non neutral, industry, non neutral version of value, just as an example, that’s probably going to have a much bigger exposure to the business cycle than a market neutral, industry neutral version? And as we show in the paper, more risk managed versions of these factors. So this is what I’m trying to motivate here. More risk managed versions of these factors tend to vary in performance across macro environments far less than the stock market and aon market, and what we see is that factor performance tends to be positive. In all those macro environments, the variation in how positive is reasonably minimal. Now, timing can sometimes be worthwhile. But when deciding if or how much to time, one needs to consider a few things. First of all, timing based on market or macroeconomic conditions has a high hurdle. Because you have to be right twice, you have to be right about the relationship between the factors performance and the contemporaneous conditions. And you also have to be right in your prediction of those conditions. So that’s a high hurdle to get both of those things, right. The other thing one wants to consider is that timing is inherently a lower breath exercise. Right, you’re basically comparing only two things now, versus usually, and the fewer things that you have to compare, generally, the lower the return potential. Certainly cross sectional comparisons, by contrast, tend to offer far more breadth. You know, if you’re looking across the universe of individual stocks within emerging markets, let’s say or develop markets, you’re talking about multiple 100, or possibly even multiple 1000 things that you can compare. Lastly, I would point out that any type of timing incurs an opportunity cost, or at least a potential opportunity cost. The more you tilt factor weights, the more you’re potentially sacrificing diversification in the process. If you don’t tilt that much, this is maybe less of an issue. But if you’re tilting a lot, if you’re swinging those weights around a lot, you’re possibly sacrificing meaningful diversification. And that opportunity cost that loss of diversification should be weighed. So big picture, timing is difficult. Sometimes it’s worth it. Certainly we think in massive dislocations like we see now in value, it makes sense to meaningfully lean into that opportunity.

Corey Hoffstein  52:13

Michelle, last question of the podcast for you. It’s the same question I’m asking every guest in this season. It’s that the cover art that I’m generating is inspired by tarot cards, and I’m having each guest pick a tarot card where the imagery or the key words resonate with them. You chose the card, the sun, and I was wondering what drew you to that card?

Michele Aghassi  52:37

Sure, I have to admit, I know very, very little about tarot cards haven’t had much exposure to them. I chose the Sun card. First, just instinctively, I found the aesthetics appealing. But when I clicked through and read some of the description, I found that the apparent meaning behind that card is actually quite relevant to finance and quant investing. And, you know, maybe as an aside, I guess, in tarot card reading, or palm reading, probably there’s some degree to which the descriptions are so universal, that the person on the other side of the table thinks like, Oh, yes, yes, this speaks to my life. So maybe, maybe this is a coincidence. But anyway, I did find that the description has some relevance to finance and quant investing, for example, the card apparently relates to the contrasting ideas of on the one hand, optimism and confidence. And on the other hand, excessive enthusiasm and unrealistic expectations. And, you know, to me, the excessive enthusiasm and unrealistic expectations spoke to what we were discussing a bit earlier about expensive stocks being mispriced relative to their cheap counterparts in a way that implicitly assumes a growth differential that is far larger than anything we’ve ever seen. The card also contrasts confidence with overconfidence, I think it refers to it as conceitedness. But anyway, you can think about that as confidence versus overconfidence, and that to resonated in terms of some of the behavioral biases that we seek to profit from so anyway, it was a nice picture and when I clicked through some of the description, to me seemed relevant to finance and quant investing.

Corey Hoffstein  54:16

Well, thank you so much for joining me. This has been really fantastic. It was a pleasure.

Michele Aghassi  54:19

It was fun. Thanks, Cory.