The Market is Getting Less Efficient: Here is Why That is GREAT for Value Investors | Cliff Asness

And so if that is true and markets are in fact somewhat less efficient, not grossly inefficient, but less efficient than they used to be, yeah, there might be more money to be made from someone who can stick with rational investing, but it will also be harder to stick with. You go back at the end of year three, it's pretty much you and your mom in the fund at that point, and your mom hasn't filed them yet, but she's filled out the redemption papers. Democracy is the worst system for governing ever devised, except for every other one that's ever been tried. And you might say the same about markets. But I've always been terrified that if I ever show up in anything like that, I'll be played by Albert Brooks. Nothing against Albert Brooks, he's hilarious, but that would bring out the pure Jewish angst version of it. If you need five more shares of Nvidia at 2 a.m. on a Saturday, you might reconsider your whole investment plan.

That title was not my creation. That title was a co-author, Toby Moskowitz's creation. Probably wouldn't do it today. The world has got a little more sensitive to off-color humor. But it's funny because of my reputation for saying more outrageous things, everybody assumes that was me, not Toby Moskowitz, ten-year professor at Yale University. So I get either credit or blame, depending on whether you like or are offended by the title. Not me.

Yeah, there's a little bit of humor in there, but it's uncomfortably close to accurate. You know, some of these, it's a great example of correlation versus causation. Those morning routines may be correlated with very highly productive people. The same mindset that can make you wake up and do some wacko, uncomfortable, painful routine, but I don't think it's as causative as people think. You know, Winston Churchill used to wake up at in the afternoon and work in bed till the evening and then stay up till 4:00 a.m. Different things can work for different people.

Well, you should do them. The key to survival right there in a nutshell.

Yeah, I'm starting admittedly from only two observations, with a statistician never enjoying having two observations, but they were giant observations that were very anomalous given 75 years of history. And essentially it was the dot-com bubble in the late 90s, and then, surprising because it happened once and I thought it'd probably be 50 years till it would happen again, a very similar event in 2019 and 2020, where the difference in valuation between so-called growth and so-called value stocks, I prefer to call them expensive and cheap stocks, every once in a while the cheap guys are actually growing pretty well, but the difference in valuations hit easily far and away a record level compared to history in the dot-com bubble, and then actually surpassed that during COVID, and I can't even blame COVID because it was getting right up there pre-COVID. And that led to anyone, you know, quantitative value strategies, but anyone who cared about price, I don't care if you were a concentrated stock picking Graham and Dodd manager, you generally hated those periods to a level that was unprecedented in both cases. They've massively recovered and the round trip has been fine. So the main challenge was holding on. But after the first one, it shook my faith maybe in efficient markets a little bit. I wrote a paper called 'Bubble Logic' during it that was a little sarcastic about some of the valuations. That maybe was an odd title at that point for a Gene Fama disciple. I still love Gene even if I differ at least on the spectrum of how inefficient the market is than he is. But then when it happened again, I sat back and I said, all right, I'm pretty sure there's something out there causing either general levels of inefficiency to be somewhat bigger or occasional massive bouts of mania to be somewhat bigger. So I was reasoning from data points. I didn't predict this, and it's always more valid if you predict something and then it comes true than if you observe it and then try to come up with a story. But with that caveat, I do believe markets have gotten somewhat less efficient. I want to be clear, I don't think they're grossly inefficient. I don't have a better way for society to allocate capital. I'm fond of the Winston Churchill quote, it's my second Winston Churchill reference in six minutes, I didn't mean to do that, but he has this great quote about democracy. It's a good date for this quote actually. Democracy is the worst system for governing ever devised, except for every other one that's ever been tried. And you might say the same about markets. Gene Fama, I sat through his class three times, first as a student, then two more years full-time sat in every lecture as the TA, and he always tells the class about a few weeks in, markets are after introducing the concept that markets are assuredly not perfectly efficient. And Gene, if anything though, super intellectually honest, he admits perfect efficiency is ridiculous, nothing's perfect. But once you admit things aren't perfect, how imperfect they are and how much that very subjective time become a legitimate object to test, and if it's hard to formally test, to even speculate on. I think the intuition you mentioned that markets would get more efficient over time is largely based on technology. The speed of trading, the cost of trading has gone down, the ubiquity of information. But I think most of that is about speed, not about correctly processing information. I think it's almost a certainty that information gets into prices faster than 35 years ago, which is roughly the beginning of my paying attention to this stuff. But I think the difference is 10 minutes versus 10 milliseconds. I'm old but I'm not carrier pigeon old. We had Bloombergs and we had phones, and when news would come out, prices would react. So information itself has to be processed well, and I then advanced a bunch of theories as to why that processing might be less efficacious these days.

Testing market efficiency is historically extremely hard. First, it runs up against something famous in academic circles, which is a bit of an oxymoron, it means famous among 11 of us, but that's called the joint hypothesis problem. The efficient market hypothesis, boiled down to one line, says prices reflect all available information. Not included in the hypothesis is a model of how prices should reflect what information is relevant. Do you pay a lower price and expect a higher return for a more volatile stock even if it's uncorrelated with the market? The capital asset pricing model would say no. The capital asset pricing model is a great example of a famous part of the joint hypothesis. You can test the joint hypothesis that markets are efficient and the capital asset pricing model is what's being used to set returns and prices. That test is wildly rejected. It's one of the great failures. 50 to 100 years of data across every country, the security market line is relatively flat. It is not rewarding beta as the CAPM says. But then you have to step back and go, so what went wrong? I had two hypotheses. Is the market trying to use the capital asset pricing model but there are frictions and biases and mistakes people make that lead them to not succeed? Or is the market quite efficient and the CAPM is just wrong, that there are other elements besides market beta, and maybe market beta isn't that important, and if we had the true correct model? Or are both wrong? We almost never know the answer to that question. Some things, when you find strategies that for some period have incredibly high risk-adjusted returns, you can say maybe it's implausible there is an equilibrium model that in a joint hypothesis would explain a three Sharpe ratio. And those usually go away pretty quickly. But if that exists, maybe you can start to say it's market efficiency that's suffering. But mostly you don't know. But one thing I took great latitude in this paper is this was an invited paper for the 50th anniversary of the Journal of Portfolio Management. I think they just did it by age, whoever published first and was still alive gets to write one of the invited papers. And lucky for me, I'm one of both. But as part of this, the publisher Frank Fabozzi was quite encouraging about retrospectives, opinion, wise old man kind of stuff. I've been around for a while. And I admit at the very front of the paper, I will not be proving this assertion. Tests of the general level of market efficiency over long periods are still highly contentious. It's not an easy thing to prove one way or another. Therefore, tests of a modest to decent size change in that level over your career are going to be even harder. So there is a lot of opinion. I'm not making excuse, it's a strongly held opinion I stand behind. But this is a very long-winded answer of saying, unlike many of my papers that are filled with t-statistics and other formal approaches, I'm not going to be claiming I have a p-value of 0.03 that I'm right. I'm going to be claiming I think I'm right because of what I'm observing, but it is more anecdotal and logical and reasoned than my normal kind of work.

Actually, I don't. And I've heard that in response to my paper, and I haven't gotten around to it, but I plan on writing a little something on this, probably just on my website. The Journal of Portfolio Management issue is just coming out now, and I do find it a little weird to have to write a defensive 'no, it's not this' to critics before the actual paper is out. I can be thin-skinned, but that's taking it a little too far. But here's the problem with that. It's always going to be the case. Bill Sharpe's arithmetic is tyrannical. The average deviation from the market doesn't beat the market, and the average deviation from the cap-weighted, Jack Bogle-style market loses to the market after fees and transactions costs. So the observation that the average person owning an active portfolio doesn't win, if it does win, your measures are broken, you're not doing it right. You can measure subsets. Do mutual funds do Graham and Dodd old school managers? Do quants win on average? Because obviously they can be taking the, that could be negative alpha to others and positive to some. But the general observation that the whole, my argument is an extremely conceited one. It's that this basic valuation-driven approach, it doesn't have to be just quant value, it could be broader than that, it could be multi-dimensional value that considers profitability, changes getting moving in the right direction short term in terms of fundamentals, risk levels, beta, all. But my basic intention is those who follow that rigorously will do better going forward in a world of less efficiency. But you can never forget, and again there's an arrogance to this, any kind of active management has an arrogance because it says I can be on the right side of Bill Sharpe's arithmetic. They are taking someone's money, and now you're serving a purpose, you're making the market more efficient if you're taking the right side. That's what markets work. Everyone deviating from cap weighting is trying to take someone else's money. And that leads to pretty good results if the market stays close to efficient. But the fact that the average doesn't win, nope. And even looking at the subset I like, now it's come way back, had a really rough time in 2018 through 2020, made more than all of it back, which is nice. But the value spreads I focus on, I've been obsessed with since 1999 with just tracking the spread between cheap and expensive. I usually publish very simple versions. I've also looked at kind of quality-adjusted versions. They usually say about the same thing. The way we track it, they set a new record in late 2020, early 2021, blew past the way we measure it, blew past the high of the dot-com bubble, which again I never thought I'd see in my career. I admit I got that wrong. It's come way back, so it set a new 100th percentile. I have made this joke many times, I apologize if your listeners have heard it already, but I took to calling it the 125th percentile, which was always fun to say to an audience, and if nobody laughed, you knew you were in trouble. Because well, actually they might just not have found it funny. I have trouble accepting that. But if they might have just thought, oh, 125th percentile, that's great. No, there is none. It's the new 100th percentile. But the point is, of course, there's a fair amount wider than the prior one. Huge recovery. And I'll brag, much larger in ways we measure value, which is much more balanced within industries, not tech versus everything else, global, not just the US. In the US, the major value indices have continued to be crushed by the major growth indices, and that's a MAG 7 phenomenon. But those are cap-weighted and have huge industry bets. If you do it more like an active manager, and it's not unique to us, and you do it very balanced, not equal weight, that's way too far, but closer to equal weight, many hundreds of stocks long and short around the world, we've seen a huge comeback. But we're only back to low 80th percentiles in that spread versus history. So if maybe we'll stay here forever, but if we eventually head back to historically normal, I still think rational strategies, you mentioned the 'Long Run Is Lying to You' paper earlier. The point of that paper is when valuations are at extremes, asset classes, strategies, relative returns on different geographies, history will give us a poor guide to the future because it's really recording the valuation change, not something fundamental. So I think the wind has still been in the face of what I call rational strategy. So I look forward to the future because I do believe in this. I do believe it'll be more lucrative to be on the rational side for at least the time being. The caveat to that, and I'm very clear about in this paper, is what inspired me to look at this were two wild, unprecedented episodes that were extremely painful for what I call rational investors that ultimately paid off for them. And so if that is true and markets are in fact somewhat less efficient, not grossly inefficient, but less efficient than they used to be, yeah, there might be more money to be made from someone who can stick with rational investing, but it will also be harder to stick with. These periods as we've seen, if my theory is correct, will be you'll have to fight bigger extremes that last for longer. And I say in the paper, no one might care what I think is fair. I'm just throwing you my opinion. But this strikes me as kind of fair. Harder to do but a bigger payoff if you can do it. Is even within an inefficiency argument is some type of efficiency that strikes me as a fair tradeoff. I steal a line from my ex-professor Ken French who labels the market has, what does he say? And he attributes this to someone else, but I'm giving it to him, that markets have an efficient amount of inefficiency. And we could get into that, that gets a little geeky. But this strikes me as something that is fair going forward. I think the payoff to being the rational investor willing to take the other side of extremes will be bigger in the next 20 years than it's been in the last 20 years, but it will occasionally test you more.

So backing up, I vaguely alluded to it earlier, but the chart formally using, in this case it's Fama and French, so it's just their price-to-book measure. I'm not going to diss price-to-book, but I'm also going to say it's not what most quants use. We use a wide variety of measures, some of which are well-known, some of which we've tinkered with for a long time. But as I show in the paper, more modern quant measures come up with very similar results, so why not stick with the classic that's been looked at a ton of times? This goes back to 1999 during the dot-com bubble. We built this measure which people had not looked at yet. I say that people had not publicly talked about. I always want to be careful, someone privately might have been doing it 10 years before us. We've had things at our firm we've been doing for 10 years and someone finally writes about it publicly and we're like, darn, why'd they tell people? So I don't want to scoop people who were ahead of us, but we were the first to write about this. Where the prior tests, almost all the academic and practitioner research had sorted stocks on something like price-to-book, formed a portfolio of the cheaper versus expensive, and just looked at the return differences. To our knowledge, none had asked the question, how big are those differences? Well, if you take the price-to-book of the expensive and divide it by the price-to-book of the cheap, that number better always be more than one. I'm fond of saying if it's not more than one, your code or your spreadsheet is broken. You've sorted the stocks and you took the high divided by the low, and it's not allowed to be negative, so you're going to get a number bigger than one. But from 1950 to 1998, it had looked like a very well-behaved series. Either the straight series or the five-year moving average that you refer to looked, for an economics series, fairly well-behaved. It varied between about three and six, meaning the top 30% of stocks was at the most about six times more expensive on this scale than the bottom 30%, and at the tightest was only about three times more expensive. And then in the dot-com bubble, it exploded to, I don't remember exactly, but 10 something. It looked like a broken chart. Or the famous 'the world has changed.' At least on this measure, the world did change. I don't think ultimately the gravity of the world that would pull this back to normal went away, but there was a change. Then over the next 20 years, you saw it meander around. I think those periods where it spends long periods around the median or below the median, those are kind of the normal periods where maybe even if markets are less efficient in extremes, nothing extreme is going on. What I'm talking about are really these bouts of crazy. But the point of the five-year moving average, and I do it a few different ways, five-year moving average is of course going to be more extreme if the numbers are more extreme, but it also shows things lasting for longer. In the paper, I also quote number of months it's been above median, where when you have big spreads, how long, and that's gotten radically longer. And that's two forms. There are two ways to hurt when you're underperforming: intensity and duration. How bad are your losses, be they absolute losses in a long-short portfolio or relative losses to the market in a traditional portfolio? How big are they and how long do they last? And one thing, in my career, if you ask me at the very beginning of my career, back in the stone ages, which of these two are more important, I'd say, oh, the magnitude. Magnitude is what counts. If you write down formal utility functions, maybe there is one, maybe some genius professor has done it, I don't know any that have duration of annoyance in them or duration of pain in them. In the real world, sticking with the strategy, even if the absolute size of the drawdown is two-thirds of the size smaller, if it lasts three years instead of a year and a half, that's much harder to stick with. After year one, you go back to a client who likes you, who chose you. You're saying this is what happened, this is why we like it going forward. You're usually okay with that. You go back after year two, they're like, you know, that's what you said last year. And you go, yeah, it just happened again, and this is why we like it even more. Some don't buy it, many do. You go back at the end of year three, it's pretty much you and your mom in the fund at that point. And your mom hasn't filed them yet, but she's filled out the redemption papers. So I'm exaggerating of course, but my mom would have never redeemed, but that would have just been a Jewish mama sticking with her son. She would have lost total faith in my abilities. Duration counts too. But if that tether to reality is not gone but weaker, which is pretty much what I mean by bouts of more extreme inefficiency, then they're going to do both. They're going to be bigger and they're going to last longer. I just think of it as gravity. There's a gravity pulling things back to fair, and if that force is weakened, you're going to see bigger and longer deviations.

There are a few different ways to take this. First, I list that as one possible reason. I do think the rise of passive, and by passive I mean true Jack Bogle market cap style passive. Every once in a while someone will define passive as like rules-based, so every ETF no matter how anti-market it is would fall into passive. That makes no sense to me. Passive is owning what everyone else owns. And I do talk about how the rise in passive, and I'll get to why in a moment, could help loosen that tether to reality. I do think the general notion that money going to passive radically distorts things in favor of the big caps is overdone. You ask the exact right question: what's the impact compared to liquidity? And there's some evidence, I cite one paper, this research is in its infancy. But the question comes down to like, if I bought 1% of Nvidia and 1% of a micro cap, do I have a bigger impact on the Nvidia price? And you could tell different arguments. You could tell an argument that it should just be proportionate. You could tell an argument that ultimately dollars matter, that it's not just proportionate. The initial evidence I've seen says dollars matter a little, but not a ton. That probably fits my intuition. But more so, when I'm looking at it through my own lens of course, but when we've done things like recalculate that value spread throwing out the 5, 10, and 20% biggest caps, throwing out just the most expensive stocks, throwing out some version of the Magnificent 7 in a more dynamic way, like always looking for what was the closest thing to it at its day, obviously the Magnificent 7 20 years ago wasn't the Magnificent 7, and the value spread, and I know I'm singing the same song, but the value spread historically comes out looking very similar and value performance comes out looking very similar no matter how you're doing this. So to us at least, looking at it our way, somebody else might have a different way to look at it, what we consider the returns to rational investing, the dislocations in terms of the spreads when rational investing is out of favor, simply are not being driven by only the mega caps. They are part of it. Again, our versions have about 700, 800 stocks long and short around the world, relatively balanced by industry, with nothing having a huge weight. So even if the MAG 7 are grossly out of whack, which you don't come to a quant to answer that question, quants don't answer questions about seven stocks, but even if they are, we knew this result even before running the numbers that it can't be driving the kind of spreads I'm talking about. They're much more pervasive. Now here's where I think the rise of passive can matter. There is one number. Let's ask the question: what percent of the market could actually be Jack Bogle style passive and still have some degree of an efficient market or not? I don't know. Some you have an inflection point where it gets really crazy. Well, there's only one number we know for sure that can't be 100%. This is something I joke in the paper that PhD students drunk at 2 a.m., and then I admit it's one wine cooler at 9:30 p.m. It's almost like a conundrum. How do you even think about a world where everybody is trying to free ride off everybody else? There's literally nobody doing the work of saying, is Nvidia worth more or less than the corner drugstore? We're all just assuming the other person did it and copying them. And so let's just say that would be nuts. I don't know what it would look like. It's hard to think about. We actually once had Jack Bogle himself on a podcast we were doing. We did a podcast for a while because, as you know, legally in America everybody must have their own podcast at some point. And I was lucky enough to be friends with Jack. It was an odd friendship: an 85-year-old godfather of passive and at that point a 50-year-old long-short active quant. But Jack was a wonderful man. I think the fact that his son was a long-short active quant probably softened everything about me to him. If anything one of my kids do, I'm predisposed to maybe think isn't the devil. Actually, one of my kids I would assume it was the devil, but I'll keep that to myself. Had Jack on the podcast, we got into this discussion, and Jack, like Gene Fama, I'm going to say this about another famous person, incredibly intellectually honest guy. He's like, of course 100% of the market can't be passive. I think my marginal recommendation to most investors to be passive is the right one, but 100% everything breaks down. He freely admitted that. And I said, Jack, so how much of the market can be passive before it starts to get weird? And he said, 75%. And I'm like, oh, that's really cool. Where'd you come to that, Jack? And he just looks and goes, he gets a little twinkle in his eye and goes, oh, I completely made it up. And when you're, I don't know, he was 85 at the time and you're Jack Bogle, you can get away with that. I think those of us who didn't create Vanguard probably shouldn't make that joke. But his point, even that has a point. No one really knows. A lot of people are very histrionic about this and think passive is the devil and everything has fallen apart. I think it is unlikely to me that all the craziness that I admit happens at 100% passive happens magically at 99.999% passive. 100% most things in life are some continuum. So the idea that we're much more passive now than we used to be has contributed to maybe some smaller tether to reality, I find fairly intuitive. But assigning a magnitude to it, and particularly again, I think markets are less efficient than they used to be, but I don't think I've proven they're inefficient to where you can drive a truck through and just as the rational investor make free money all the time. I don't think it's that far gone. So I think passive is part of the story for the reasons I outlined, but I think, and I admit this is all is still in the realm of hard to prove, that putting a magnitude on it is difficult, and my instinct is those who are absolutely freaking out about it are overstating its role and have kind of made it their life's work to be kind of anti-passive.

Always Arnold. But your viewers might not remember him well, but I've always been terrified that if I ever show up in anything like that, I'll be played by Albert Brooks. Nothing against Albert Brooks, he's hilarious, but that would bring out the pure Jewish angst version of it. So he, it'll, Albert's a little older now, so maybe it won't happen, but they'll find someone similar if this ever happened.

Fair enough. We'll give you script approval and free range on footnotes.

Yes and no. One of the things it's doing, again, I'll stop giving all the cowardly caveats, but in a paper that's largely opinion, as someone who often writes for more academic things, I always feel like I have to give the caveat every time. If this is right, I wouldn't say it's a different opportunity in kind. I would say it's just in magnitude. Markets have never been perfectly efficient, as Gene told us. They've always had deviations. The factor that I wrote my dissertation on, price momentum, he's always been very generous in calling that the chief, Jegadeesh and Titman should get pride of place for finding, and I was right behind them. But he always says that's the chief flaw to their three and now five factor model, and very hard to reconcile with market efficiency. So in that sense, I think if markets have always been somewhat inefficient, leading to opportunities for some subset of the market, not for all active management, I think it just loosens it and they're generally bigger. There's one I can think of specifically. Rob Arnott, who I've had my run-ins with Rob but I have great respect for him, has recent work on additions and deletions to passive indices that shows some, at least in my quick reading, rather large effects. And that again, I don't know the ins and outs, but it's very plausible that the larger role for passive would make that effect a unique and new thing.

Probably is driven by the specific role, larger role of passive. That's one I can think of, but mostly I think it's just further loosening the same bounds of rationality that have never been perfect, even if they've been pretty good over time.