Jean-Philippe Bouchaud is a French physicist. He is the founder and chairman of Capital Fund Management (CFM) and is a professor of physics at École Polytechnique.¹

Skeptical scrutiny is the means, in both science and religion, by which deep thoughts can be winnowed from deep nonsense.

Carl Sagan

Michael:

I want you to elaborate on your physics background, but bring that into this framework of classical economics. Classical 
economics, the rationality of economic agents, the supposed rationality, the invisible hand, market efficiency . . . But for some reason the idea of empirical data often gets left out of the equation—and I’ve seen this with some other traders that have had success—a physics background is different. It allows you to maybe look at the world through a wider lens.

Jean-Philippe:

Yes, exactly, I’m surprised that you say all of this because that’s more or less what my usual message is and you’ve captured all of it in a few words. Yes, it is true that I’m a physicist by training, and physics is of course learning through doing experiments. And you learn that theories are no good if they’re not able to reproduce observations. And even if your theory is beautiful, if it doesn’t fit, it doesn’t fit. You just have to throw it in the dustbin and start again.

As a physicist approaching economics and finance back in the early ’90s, that’s what struck me most—that it is a lack of a statistical aspect to the way economics and finance theories are built . . . very much axiomatic in imagining how the world could be or should be. Then developing the theories without much care about what’s going on out there.

I guess that for a while it was justified because data was not so easy to access, so the whole academic world has developed without data in a sense. And so people had to maybe supplement the lack of data by actions and by ways of thinking . . . That can happen too in physics, actually, so perhaps I was fortunate enough to enter the field when data became very easy to access. And when looking at data and trying to make sense of data through a kind of vivid light on the failures and drawbacks of efficient market theory and Gaussian statistics, Black–Scholes, all this to me was quite apparent . . . that it was not enough to understand the world.

Michael:

You mention the Black–Scholes model. It’s still in use. So even though that systematic underestimation of risk is well known by people like yourself, it’s still in use.

This is a fundamental truth about any sort of practice: If you never push yourself beyond your comfort zone, you will never improve.

Anders Ericsson

Jean-Philippe:

Yes, I know, I’ve been ranting about that for ages and one problem is students. You have to teach students something, and Black–Scholes is so easy to teach and it’s so beautiful mathematically, that a lot of people just resent the idea of having to put it all down and start again with something more messy. Of course the world is messy, and being messy it’s much harder to teach . . . to focus on the right things. By definition you have to form your intuition on something else than mathematics.

That’s why physics is good at that, because it gives you a lot of examples where you can put your hand in the dirt and try to push on some button and see what happens. But the same should be more and more true with economics and finance now, through two channels. One is the availability of data and the possibility to make experiments on data, simulations, that is, and even without data you can do simulations. You can invent worlds of people trading according to some rule, or funds producing according to some rules, and implement whatever rule of thumb or feature of the world that you think should be there, and then just run the simulation and see what happens.

Then what’s very striking when you do that—plus it’s fun because you kind of play God. And second, very quickly you realize that some of the rules just don’t work. They don’t represent at all what is seen out there, and others seem to capture something that’s very close to reality. So my impression is that by training people more and more with this type of background, this type of experimental background . . . experimenting with simulation is a strange notion, which even in physics it took a little time for people to accept that simulation was a legitimate way to do science.

I don’t know if you know Mark Buchanan, he’s a science writer and he wrote a few years back something I like a lot . . . just after the crisis in October 2008. He said the following:

“Done properly, computer simulation represents a kind of ‘telescope for the mind,’ multiplying human powers of analysis and insight just as a telescope does our powers of vision. With simulations we can discover relationships that the unaided human mind, or even the human mind aided with the best mathematical analysis, would never grasp.” For me, this is the essence of what the physics way of doing things has brought to the game.

Michael:

So people don’t think that you’re giving an interesting marketing story about a physics background for your trading firm, your firm does not hire traders?

We’re rarely rational when we vote because we’re rarely rational, period.

Robert M. Sapolsky

Jean-Philippe:

Yes, it only hires physicists. And okay, people can think that, but we’ve been saying the exact same thing since the mid-90s. The everyday life of CFM is driven by data. It’s banging our heads against data and trying to make sense of what we see and make models inspired by what we see.

[Another reason] that it’s not pure marketing is we’re very strange as a trading firm to have published something like 100 science papers in the last 20 years, all published in academic journals, which shows that it’s really in our DNA to consider science as the right way to do things.

Michael:

Let’s jump right into my primary reason for reaching out to you, which was seeing your paper [“Two Centuries of Trend Following”; see Chapter 20], which really didn’t have any big fanfare. Just all of a sudden it appeared in the Internet ether. . . . I wonder if you might lay out a scenario for how that paper came to be and then we can discuss the specifics inside.

Jean-Philippe:

This particular paper was in the back of our minds for a long time. There are two reasons for it to appear right now. One may fall in what you call marketing, which is that we’re launching . . . a fund called “Institutional Systematic Diversified,” and part of that fund is based on long-term trend following. So it is true that we needed to give some support to why we’re doing that.

When it comes to money, the best investments were probably the ones I did not make.

Marc Faber

The second thing is that we recently in the course of the very last few years have had access to much longer time periods than we had in the past. We’ve been able to go back to the beginning of the nineteenth century on commodities and indices in terms of data. This allowed us to back test quite a number of ideas and in particular trend following. To our surprise we realized that the strategy has been extremely consistent as long as we could go back in the past.

This seemed to us to be a very interesting finding in the year where the Nobel Prize was given to Fama, Shiller, and Hansen. But this debate on the efficient market theory, on which I’ve been pretty vocal myself in the last 10 years, it is ironic that it’s given to Fama who was still arguing there’s no bubbles, there’s no crashes. That the market went down in 2008 in anticipation of the crisis and not the other way around, and that everything is perfect.

Trend following, momentum in general, is something that efficient market theorists have a real difficulty to explain because that’s completely out of the framework. It’s very hard to evoke some kind of risk premium that would be associated with trend following. So it has to mean that markets are not that efficient.

There’s a lot of other clear discrepancies between theories and reality, but this one is a very genuine and clear one which talks to everybody. Just looking at the trend on a long time scale is giving information on the future motion of the market. It really means that all public information is not included in the price right now. For me it’s both from an intellectual and commercial point 
of view a very interesting finding.

Michael:

The statement that jumped out at me was, and this is from your paper, “The existence of trends is one of the most statistically significant anomalies in financial markets.” That’s a powerful statement.

Jean-Philippe:

We’ve been looking at financial markets in the last 20 years and it’s very hard to find extremely significant statistical effects. You can find them on the high-frequency side, but then there’s a lot of murky things around high frequency. First of all, costs are tremendous if you want to trade at high frequency. It’s not clear that all the high-frequency anomalies that have a strong statistical signature are, as economists would say, very relevant from an economics point of view.

On the other hand, these very slow trends where a lot of money can pile in and has piled in, is of course much more mind-boggling in a way, and also has to be taken into account both for academics but also from the point of view of professionals.

It’s a wonderful thing to be optimistic. It keeps you healthy and it keeps you resilient.

Daniel Kahneman

Michael:

Other interesting facts in the paper, ­perhaps this is obvious if data is going back to the early 1800s, but trend predates trend following, which I thought was interesting. It’s actually a very small percentage of traders employing trend following models that make up the volume.

Jean-Philippe:

Yes, I agree. Well, you can see it both ways. I would say that traders on aggregate using trends is probably the reason why trends are there in the first place, and people using trends have been around for 200 years. That’s my interpretation of what we see—there’s a lot of people, even small people, who on 
aggregate play the role of trend followers and therefore create these trends.

Michael:

You mentioned Fama and the split Nobel Prize. I had a chance to speak with Harry Markowitz recently, who’s very lucid [89 years young now], and the point that I made, “Harry, did you find it interesting that when you wrote back in the 1950s this is what we should be doing, that within a few decades, other academics had taken what you said we should be doing and had said this is what we are doing?” His response was, “I think you’re going to have to talk to the behavioral economists about that.” He didn’t want to touch it, but his point is that I never said this is what it is, this is what we should be doing. Other people interpreted him to come up with these, as you might say, hard axioms that became rules—the foundation of the efficient market theory.

Men, it has been well said, think in herds; it will be seen that they go mad in herds, while they only recover their senses slowly, one by one.

Charles Mackay 
Extraordinary Popular Delusions 
and the Madness of Crowds²

Jean-Philippe:

Yes, it is a strange field in the sense there’s clearly interaction between what people do and what people observe, and the use of Black–Scholes in 1987 is rather a vivid example of how things can go wrong when wrong models are used. That’s what makes the subject fascinating for a physicist, because you have to go kind of one step further and try to understand how the models themselves might change the game.

Actually we came up with a simple model on how trend could lead to trend or mean reversion could lead to mean reversion. It’s not clear that we could imagine a world where people would follow mean reversion rather than trends, but it seems that humans have such a propensity to follow trends. There’s a lot of very interesting psychological experiments where you can show that when a small child sees three points aligning on a line, it gives him pleasure. . . . We’re wired in to extrapolate past trends, and that’s probably a way to extrapolate the motion of a tiger jumping on us or something that makes us alive today.

My intuition is that it’s much harder to go against the trend than it is to follow the trend. Again, there’s a lot of very interesting psychological or even biological experiments showing there’s a lot of things [for instance] . . . hormones going in and out of our body in the two situations. One when we’re conforming to the crowd and the other when we’re not conforming to the crowd. There’s a pain associated with not conforming to what’s going on.

Michael:

Classical economics has no framework through which to understand wild markets—and wild is your phrasing. Could you talk to the idea of classical economics not having the ability to have a framework to see those wild markets, to see through them? And when you use the word wild, what does that mean to you?

Jean-Philippe:

Yes, you’re referring to a paper that I wrote “In Nature,” which was just after the unraveling of the crisis, and this made me react very strongly because I felt that this was in the cards. And of course other people had seen it coming, but I was not too happy with the way economists had been dismissing all the attempts to introduce a little more wildness in the description 
of economic systems and financial markets.

Actually, wild is the reference to Benoit Mandelbrot. Mandelbrot introduced fractals, of course. He’s introduced also the idea of 
distributions without the second moment, without variance or infinite variance and distribution with infinite means. That’s his classification of randomness, if you want. He would call benign 
randomness, the ones that the economists love, Gaussian and things . . . where you can replace a heterogeneous system with 
its average. We know for example that this relates to Piketty’s book [Capital in the Twenty-First Century] as well that the 
distribution of anything in economics is so broadly distributed, that very often it just doesn’t make sense at all to replace a collection of people by an average people—or, a representative agent, would be the classical word.

But coming back to Mandelbrot, benign randomness is the one that I just described, whereas wild randomness is the one that is difficult to tame and it’s difficult to tame because it’s hard to speak about averages and variances. That’s really what I was referring to when using the word wild.

Nature is written in mathematical language.

Galileo Galilei 
1564–1642 
Italian Physicist

Now why is economics in general not able to capture these big swings? It’s very strange. It’s because the models are constructed 
to be intrinsically stable. It’s like people insist on the fact that a rational world is a stable world and so your model should be stable. So models in economics came up with equilibrium points which are intrinsically stable. That is if you perturb them by a small amount, they’re going to naturally go back to the equilibrium. This is so much ingrained in the model that it’s by definition impossible to have a crisis.

What’s interesting is that when you remove a few of these rational assumptions and introduce market imperfections, then it’s very easy to find situations where the rational equilibrium of economics, even if it still exists, is actually unstable, and if it’s deterred by small external shocks in such a way the system goes out of whack for a while, this is what we would call a crisis.

Mathematical analysis is there to allow us to not only describe, but anticipate to some extent, or at least make space for, crisis in the economic world. And to me this is a fascinating topic of research on which we’ve been focusing in the last 
few years.

Michael:

You sound like you’re having fun with this subject. You get to wake up every day and have fun.

Jean-Philippe:

Yes, exactly. That’s totally true and I’m happy you say that.

Michael:

What I have loved over the years, talking to you today, and many of your peers, is that when one accepts uncertainty, there’s a certain honesty to it. I feel a lot of discomfort when people are so certain about what’s going to happen.

Mathematics is the science of what is clear by itself.

Carl Gustav Jacob Jacobi 
1804–1851 
German mathematician

Jean-Philippe:

I think that’s the big difference between physicists and economists, and there’s a lot to be written about this. In a sense I would say that we’re privileged as physicists because we don’t have to talk to politicians and we don’t have to make statements about how the world is supposed to work, in the sense that nobody’s relying on us to make political decisions. I think there’s a huge amount of pressure on economists because they’re under the spotlight. They have to come up with stories and decisions and this means that it’s very hard for them to take a step back and say, “Okay, I’m really going to try to understand what’s going on here and maybe it’s going to take me 10 years or 20 years, but at the end of the day we’ll have a better theory for the world.”

Okay, well all this is great, but what am I going to say to my minister of finance when he asks me, “Should I raise tax or should I do this or that?” It’s true that it puts people in a bad situation because they can’t think long-term, and as you’ve just said, as physicists by training, what we love is to be able to think that we understand something.

If we fail, well, it’s okay. There’s nothing wrong in failing. We know that physics has had so many revolutions and so many things that people were absolutely convinced were true, turned out wrong 
in the end. It’s an incredibly good backside against what you said earlier . . . against certainty and some form of arrogance as well.

Michael:

To a degree you’re expecting failure, and maybe the economists, they can’t acknowledge there might be failure.

Jean-Philippe:

Yes, because their theory’s construction is comple­tely different for sociological reasons as well.

Michael:

I appreciate you taking the time.

Mathematics knows no races or geographic boundaries; for mathematics, the cultural world is one country.

David Hilbert 
1862–1943 
German mathematician

Jean-Philippe:

Very happy to meet you someday. Of course we like your book [TurtleTrader]. As you saw, it’s actually referenced in our paper, and I’m very happy to have been able to talk to you.