Welcome back to The Bull, your personal finance podcast. Aside from a few giants like Warren Buffett and Gene Fama, the person most frequently mentioned on this podcast is probably Cliff Asness, the legendary former student of Fama himself and founder of the hedge fund AQR, the third largest in the world. Why is that? First, AQR is the firm that, perhaps better than any other, has managed to bridge academic financial theory with real-world application. Second, Asness and his extraordinary team produce an impressive number of papers which, at least in my case, have probably contributed to about 50% of my overall understanding of finance. In many episodes, I told you that come what may, and even at the risk of getting hit with a restraining order, sooner or later I would manage to bring Asness here with us. At long last, this dream has come true. And the long conversation we had, rightfully earning the title of the longest episode in the history of The Bull, adds immeasurable value to this podcast. So even though you will never listen to this episode, thank you to Cliff for this incredible honor and thanks to all of you because without your presence and without being so many, this meeting would probably have remained just a mirage. With Cliff, we have talked about his years in Chicago under Fama and French, about why markets have become less efficient over time, about momentum and value, expected returns, trend following, and much more. So without further ado, I leave you to my conversation with this extraordinary figure in global finance, Cliff Asness. Enjoy the episode. Cliff, thank you so much for being here today with us.
Oh, I am happy to do it. Happy to do it. Thank you for having me.
Oh, you're gifting us an incredible privilege by attending our podcast. You can't know it, but I've mentioned you and AQR's work so many times here that you already belong to this podcast without even knowing it, long before your first appearance today.
Okay. Well, that's a little frightening, I got to tell you, but it's nice to hear.
One year ago, we hosted here your grandmaster, Gene. So it's kind of a serendipitous occasion to have you here right today. Somehow the stars have aligned.
Well, he is the grandmaster. So you're very lucky to have him. He doesn't do many things like that.
Yeah, it was truly remarkable. If you don't mind, I'd like to start right from there. Would you tell us about your days in Chicago in the 80s and 90s, because back then you had the opportunity to study finance where modern finance was born?
Yeah. Whenever something works out well, you got to admit you got lucky, at least partially. Few things have no luck involved. I started the PhD program in Chicago in 1988, and it was the late 80s and early 90s when Fama and French were really starting what today we might call factor investing. I always laugh because we didn't have these terms back then. You always use today's terms to describe the past, even though I don't even think we called the price-to-book the value factor immediately. I'd have to go check exactly when that started, but I'm going to guess their first few papers didn't. So I got lucky. I was there. I got to work with them. I got to be early in something, and that is a lot of luck. I don't think anyone, certainly including me at that point, thought this is going to develop into a major part of the investment industry. So yeah, it was phenomenal. And it's not just being there, it's studying with those two men. Gene Fama, I remember I actually wrote a fair amount of my dissertation on what today we call the momentum factor. And I remember being very nervous about telling Gene I want to write on this, because everything can be argued, but it's not a very efficient market kind of factor. Some people can try to make an argument that maybe this works for an efficient market reason. It's hard. I think I mumbled, 'I want to study this price momentum factor, professor,' and it works very well, because if it didn't work very well, that is a Chicago Gene Fama dissertation, because a lot of investors look at things like momentum. So to be able to write, tons of people claim to be outperforming, but momentum really doesn't work just would have been beautiful. It would have been like stock splits aren't a miracle. It would have just fit in perfectly. But I said this to Gene, and his response, which I feared, was, 'If it's in the data, write the paper.'
Yeah. Sure. Sure. Yeah. I know the response.
Great. One thing you learn from both Gene and Ken is not that data should drive everything. Theory and intuition, you have to avoid just overfitting, but their respect for data, their willingness to acknowledge, Fama's many times called momentum, I think he calls it like the key embarrassment to the five-factor model, or something close to that. It doesn't matter what you like. It doesn't matter what result you want to see. Respecting the result you do see is something I hopefully learned. I know those guys have it. I hopefully learned it too.
Yeah. I admit that when Fama came here last year, I tried to ask him about momentum and his answer was, 'It's not convenient.' He's honest about it. And their model forms like the scaffolding for modern finance. It's hugely important. But no model is perfect, right?
Yeah. I have written that I think they should include momentum as a six-factor. You're allowed to disagree while still admiring. But Gene is quite honest about it, so I've always appreciated that. Momentum is a challenge for the pure efficient market hypothesis because, as you mentioned earlier, it leans more toward the behavioral side when it comes to explaining the factor premium behind it. Now my question is, in hindsight, what was incomplete in the strong efficient market hypothesis view, in your opinion at least?
Well, I'd like to even back up a little bit. The pure efficient markets hypothesis, even Gene, I don't think he'd say markets are perfectly efficient. I think sometimes people overstate it and they want to, if someone believes the markets are close to efficient, they'll paint Gene as more extreme than he is. I sat through his introductory first-year PhD class three times. Once as a student and twice as the teaching assistant. I went to every class because I was terrified to get something wrong. And you know, sometime after teaching what the efficient market hypothesis is, Gene shocks the class by saying markets are almost assuredly not perfectly efficient, because Gene is brilliant and sane. Perfection is a silly hypothesis. I think Gene probably thinks they're more efficient than I do these days. I probably think they're more efficient than the average day trader, to go to an extreme. So I don't think most of us who are paying attention and trying to get the details right would have ever argued a pure efficient market hypothesis. I do think, and I know you know the literature well, in the ridiculous amount of years since then, I make a lot of comments about how old I feel when we go over these 1980s kind of stories. 37 years ago when I first met Gene, it's a long time. But in that period, a lot of empirical and theoretical results have tested. And you accurately talk about an efficient market explanation versus a behavioral one. We're still fighting about almost all of them. There's always going to be a story in one direction. You have to make a judgment on each one, which explanation makes more sense to you. And I think you're right. Momentum is not impossible, but is much harder than some of the others to tell in an efficient market. If a characteristic says this portfolio of say a thousand stocks around the world should beat this other portfolio, or does beat this other portfolio, there are always three possible reasons that can be true. One is accidental and overfitting. It just happened. It's not going to happen again. Someone wrote a paper on it. If you get past that, and I think, certainly this is self-serving, but I think things like the value effect, the momentum effect now have such overwhelming evidence in so many places. I don't give that a lot of credence, but you always have to have that one. So if something is real, it can work in an efficient market sense because those things that outperform are riskier. And you got to get paid for taking risk. And we could go down a rabbit hole of how to define risk. That is not simple either, and is part of the issue. But the other reason something can work is someone's making a mistake. It's not that these thousand stocks are riskier than these other thousand stocks, or the portfolio of them to be more precise is riskier. It's that on average people overpay or underpay or something like that. And I do think, again to your point, momentum falls. It's very hard to come up with something other than a behavioral explanation for momentum. With say value investing, it's a closer battle. I admit I've drifted more to the behavioral side over my career. But it's still an interesting argument. It's not a gimme either way.
Yeah, definitely. Last year you wrote a widely cited piece, 'The Less Efficient Market Hypothesis,' where you claim that markets have become less efficient over time. Long story short, you do feel markets are still pretty efficient, at least for the average investor, but not as efficient as they used to be. Could you please walk us through all of that?
Yeah, sure. First, I hope Gene's not listening. I'm still scared of him when I say these things.
Do you still feel guilty for trying to beat the market?
Yeah, that was in my bio for a while. I don't know if it's still there. That I'm scared that I'm trying to beat the market because Gene's going to yell at me or something. Let's go back to this idea that everyone who's thought about it would agree, I think, that markets are not perfectly efficient. Once you accept that kind of simple concept, two questions: how inefficient are they, and does this change through time become legitimate? If things are not perfectly efficient, you don't know how inefficient they are, and the idea that they will always stay the same amount of inefficient strikes me as kind of silly. So what I've observed over my career, and I admit doing this stuff live and suffering the slings and arrows of some periods where some of these factors we're talking about were not rewarded. That is a mild way to say viciously punished. Two periods in particular: the what's often called the dot bubble in the late 90s. I say often called because if you just call it the dot-com bubble, you're kind of giving away the story, because you're using the bubble word, not a good Gene Fama word. If you think it was a bubble, we wrote a lot on this at the time, and we revisited it over the years. If you take the basic Fama and French value kind of work, and you could do much more subtle versions of value than just book to price, but you see the effect no matter what you look at. A strategy based on buying cheap and selling expensive got crushed for about 18 months, like two and a half, three standard deviations kind of crushed. Not 10 standard deviations, we're not talking about a nine-sigma black swan. We're just talking about a really ugly period. And one thing we did way back then, and we were the first to do this at least publicly, you never know if tons of people did this privately, so I don't want to steal anyone's credit. But all of the academic work we had seen up till then just sorted stocks on valuations and asked a very interesting question: do low multiples beat high multiples? Are there better or worse ways to measure this? But at least to my knowledge, to that point no one had asked about how different the pricing was. You can imagine if you keep it simple and just do Fama French price to book. Again, I think most people, including us, certainly measure value a little differently these days. But if you sort stocks on price to book and go long the low multiples and short the high multiples, we thought a pretty interesting question was how different are the multiples? Does that change through time? And does that have predictive value? And what we found was, well, first of all, let me say the incredibly obvious. If you do this sort, the expensive ones always look higher multiple than the cheap ones. I like to say your spreadsheet is broken if that's not the case. I mean, you did the sort. But the ratio, I'm doing this from total memory, the price to book in Fama and French's construction, if you did that sort, it kind of varied between three and six over time. Meaning the top third, which is how they usually slice it, of stocks that were more expensive as a portfolio sometimes were about three times the price to book of the cheap stocks, and they got up to about six times. At the peak of the dot bubble, they got up to, again from memory, about 12 times. Yeah, that's crazy.
Nothing similar had ever occurred before in like 50 years of data. It was one of these embarrassing graphs. I hope everyone can see. I'm drawing a graph on my fingers now. It kind of did this. Woo.
And it, you know, so clearly something was going on. Either you could tell an efficient market story where our measurements were missing something that the world had changed, or the world had gone nuts. We also showed in that paper, which we still believe today, that these things are very hard to time, and I wouldn't bet my life on it. Even a few years ago I called it sinning to time these factors, and then I said you should sin occasionally and modestly. But we said, you know, over the whatever years of data, long-term data we had, that in the 3 to 6 range, when it was more extreme, value did better over the next few years and vice versa. I think that you could have an efficient or a behavioral story for that one. And so we said at 12, it's looking pretty good, and that did work out. I did think that was a bubble. I didn't think that was an efficient market story. I wrote a piece called 'Bubble Logic' in early to mid 2000.
You know, I don't know how to define a bubble, but here's my definition. If I can't come up with assumptions that are in the ballpark of what I think are realistic that justify these prices, I'm willing to use the B-word, the bubble word. And in my view at the time, on very diversified portfolios, I would never use the bubble word. I think some people are promiscuous in how they use it. I wrote about this in 2012 in the Financial Analyst Journal. Like calling a single stock a bubble. You might think it's very overpriced or something, but for me, a bubble also has to be a market-wide pervasive thing. It has to be about diversified portfolios for me, and it fit the criteria. But if you had asked me way back then, so roll forward the next few years, 2003, 2004, I'm bragging now, but we turned out to be right. The round trip was actually quite profitable for us, even if we did not enjoy the first 18 months of horror for value. That was also month 2 through 19 of my new firm. So it was not a lot of fun to have one core part. Momentum was of course pretty good over that period, but not nearly enough to offset the disaster that was value. We did meet these arguments that it's going to be great. It did turn out to be great. The round trip was great. But if you had asked me after that full round trip, do you expect to see anything this crazy in your career again? Thank God no one asked that because I think I would have gotten it wrong. I think I would have said highly unlikely. This was the most extreme event in 50 years. The question in your career has a built-in assumption that you and people of your cohort are still around. And if you're still around and survived, you're probably closer to in charge. So how is it going to happen again when you all saw this happen once? You're the senior people now. And so I would have said, nah, very unlikely. And then it happened again. Even pre-COVID, so we're talking about maybe late 2019, that same kind of measure again done simply on price to book, but as we wrote, there's lots of stuff on our website, you could have many different measures that are more diversified. We got very close to tech bubble differences with value suffering in a very similar way. And then maybe this one is exceptional, but then COVID hit, and it all got way worse. We blew past the tech bubble extremes in terms of the spread between cheap and expensive.
Yeah, remembered back then.
There was that kind of stupid six-month period where the only stock you were supposed to own was Peloton and Tesla. And that was not pleasant for people who systematically want to take advantage of the so-called value effect. So it did happen again, and we survived it. We actually ended up net making money round trip, but it was extremely unpleasant for a while. So separately, we think we've made our process more robust if this happens again. Even if we didn't, we love our process, it worked, but it was not an easy thing to live through. It was not a fun thing to live through. So after the second time, you know the old cliche, fool me once, shame on you, fool me twice, shame on me. I got fooled twice. So shame on me. It made me go, what did I miss? If there were two bubbles in my career that were unprecedented in the prior, again, call it 50 years, what did I miss? Why? And then I started accepting again the empirical fact which some might dispute, but I think I lived through two things I would call a bubble that were more extreme than things in the data prior. I started saying why, and the paper I wrote called 'The Less Efficient Market Hypothesis.' Again, I'm not saying they're grossly inefficient or easy to beat, just less than they used to be. First, I want to fix something in the paper because this is what I had in mind, but I don't think I was as clear in it as I could be. You always want to rewrite part of an old paper. I actually try to avoid reading old papers, and don't take this personally, I don't tend to listen to podcasts that I'm on because all I do is think, 'Oh, I could have said that so much better.' The only time I listen to a podcast is if I'm going back on the same podcast again, then it's mandatory to say I don't want to tell Ricardo the exact same stories I told him last time. But I did start asking why. If markets aren't perfect and are prone, what I should have said more in the paper is I don't necessarily mean they're less efficient every single day, but they are more prone to these bouts of craziness. If you believe as I do that these were bubbles, you got to start asking yourself why. And in the piece, I am very forthright that I'm conjecturing here. We have two episodes. I like to joke as human beings we see too many crazy episodes in markets. As statisticians, we see far too few. Give me a few hundred bubbles of similar magnitude, maybe I could start telling you something statistically reliable about them and start really hypothesis testing what caused it. Two is not a very large N to run tests on. But they did happen. So I have basically a few, but I had two of my two favorite hypotheses which I do believe. I'm soft selling it here. I'm trying to admit that I cannot be certain they are guesses, but I do believe in these guesses. The first is a very well-known one: the rise of passive indexing. I'm not a passive hater. There are people who think passive is just Satan and has ruined markets.
Yeah, I know at least one.
I know more than one. Some of them are brilliant people. They just really are quite extreme on it. I think passive has still been a great enhancement for the average individual. Jack Bogle and his like are heroes to individual investors to be able to get the market premium for no cost. But we all know, if well, I think anyone who's thought about this would agree. Let me say it that way. The whole world doesn't know this. We can't all be 100% Jack Bogle market cap style passive. Jack Bogle agreed with this. Like you, we had a podcast for a few years. Because in I don't know about Italy, but in the United States, I believe legally everyone must have a podcast at some point. Well, maybe we'll start this again one day. It was kind of fun. But we would, some of them were internal stuff, but we also had guests. And I was extremely lucky to get to know Jack very well and he agreed to come on our podcast. And that was a little bit of an odd couple thing, right? Long short levered hedge fund quant and Mr. Passive. But we actually agreed on a lot. And he came on and we got into this discussion about the whole market can't be passive. And Jack absolutely agreed with that. He thought, and he's probably right, that for the marginal investor where we are today, the move to passive is still a good one. But he absolutely agreed it can't be 100%. Jack was a very intellectually honest guy. And I asked him, 'So Jack, what fraction could it be?' And he says, '85%.' And I'm like, 'That's fascinating Jack, where'd you come up with that?' And he says, 'I made it up.' Because no one knows. And the point which I think Jack was trying to convey, and I'm trying to convey now, is we know the market gets really weird, and it's even hard to think about what everyone being passive would look like. Literally no one's thinking about prices. Who says the Magnificent 7 are worth more than the corner drugstore if nobody is literally doing it? Passive is a free rider. Passive free rides on people doing the work. And if everyone tries to be a free rider, you get a perpetual motion machine. So I think we know at 100% passive. I steal, finance people like me love to steal analogies from physics because it makes us feel smarter. So I call this a singularity, a black hole. The regular laws of physics break down. We don't know what 100% passive looks like, but we know it's really weird. And it's unlikely that all of the weird happens between 99% passive and 100%. So if you have fewer people thinking about what stocks are actually worth and more people trying to free ride on those who do, the idea that prices could be at least a little less tethered to reality and a little more prone to these episodes strikes me as pretty reasonable. Again, passive is awesome. I don't know exactly where we are on that curve where things get weird, but just the hypothesis. I know we have a lot more passive today than we used to. It's hard to even count how much. There's direct passive, that's not that hard to count, but what about very low tracking error portfolios? If you're doing that, then you're mostly passive.
Yes. Because most of them are closet indexers.
