Nassim Taleb on Antifragility

nassim taleb photoNassim Taleb, author of Fooled By Randomness and The Black Swan, talks with EconTalk host Russ Roberts about antifragility, the concept behind Taleb’s next book, a work in progress. Taleb talks about how we can cope with our ignorance and uncertainty in a complex world. Topics covered include health, finance, political systems, the Fed, your career, Seneca, shame, heroism, and a few more.

Click the below link to listen:

Taleb on Antifragility : Time: 1:13:56



Intro. [Recording date: December 19, 2011.] This book is in process, you estimate it’s roughly a year away. So, we are here to have a conversation that is in progress. I did my first EconTalk on the day of the release of The Black Swan. So nice, so fun. I know a lot of listeners out there are very jealous because I’ve had the privilege of reading the manuscript-not the final book. It’s in process. But for those of you who are out there, excited, this will have to satisfy you for maybe about a year. Nine months. Well, that’s a good gestation period. Now you start off with a very provocative idea. The title of the book is a little bit strange. I don’t think it’s a word in the English language: antifragility. You start off by asking: What is the opposite of fragile? And of course we think we know what that is. The opposite of fragile is robust, you say; it may be unbreakable. But you argue that’s not right way to think about it. It doesn’t capture the essence of fragility. So, why do we need another term? Because if you send a package by mail to your cousin in Australia and it has champagne glasses, you write “Fragile” on it. If it is something that is robust, you don’t write something on the package. You don’t say you don’t care, you can do whatever you want. So the fragile, the upper bound comes back unharmed or [?] and of course the worst is completely destroyed. So, that’s the fragile. The robust has an upper bound of unharmed and a lower bound of unharmed. The empty fragile would be a package on which you’d write: Please mishandle. Because a lower bound would be unharmed. And the upper bound would be improved–you’d get, instead of sending 6 champagne glasses, 8 would arrive. Exactly. Like in mythology. Or they’d be better glasses, stronger somehow. Like the Hydra–you cut one head, two heads grow back. So the robust would be more like the Phoenix–you shoot it and it comes back. So the upper bound and the lower bound are both unharmed; with Hydra, the Hydra wants harm. When I first read that idea I thought: Okay, that’s interesting, true; but why is it relevant. There are not that many Hydras around, life doesn’t consist of many Hydras. But much of the book convinced me and the reader that actually antifragile is a very powerful idea. So, where is that important, where is it relevant in our lives? The first thing, the reason I had that word–I had an equivalent word for something like volatility, but it was not powerful enough to capture it. And it was called long volatility or love volatility, but it didn’t quite capture the idea. But one day I read this book by Guy Deutscher on language, and he reports in a book something that was discovered by the U.K. Prime Minister Gladstone, and that was to the shock of everyone, about 140 years ago–that the Greeks did not have a word for blue. The color of the sky. The wine-dark sea. Homer did not have a word for blue; he did not have the full spectrum of colors. And these developed much later; and ancient Mediterraneans, the Greeks, the Hebrews, the Semites, didn’t have a word for blue or for many similar words. They were not color blind. They were biologically okay. They were just culturally color blind. So, I realized you just give light on something by coining a word; and that turned out to be “antifragility.” And once I wrote it down, I realized, I started seeing it in places I never suspected. It had this property. Give us some examples of things that are anti-fragile. We understand what things are fragile. Actually, I am even going to go beyond. Half the book is about things that love volatility, love stressors, love uncertainty–political life. But we’ll get to that in a minute. The human body, the bones. The bones need stressors, constant stress. They communicate with the environment with stress. If you spend Christmas vacation in a space shuttle, you’ll come back with diminished bone density. Which is weird. Because you’d think it would be great to be in the space shuttle because your bones will get to rest. What could be better than getting them unstressed? Actually, a complex system, a paper that really changed my thinking, by Gerard Karsenty, a 2003 paper in Nature, and of course he had a lot of follow-ups; and in it he showed it’s not aging that causes weakness in the bones, the reverse is equally true. You have a complex system with feedback loops that are not as obvious as in a linear, ordinary system. And therefore the idea that weak bone mass makes you older. And vice versa. Weaker and older. You can see the bone density of females in African villages who carry jugs of water on their heads, between 100 and 200 pounds–they have excellent posture and excellent health. And even male reflective [?] abilities are affected by bone density. Which means if you go to the gym you are wasting your time because you need weight-bearing stressed, not these [?] machines that waste your time. We’ll go into that in more detail later; you bring it up in the book and it’s a very interesting idea: that certain things are good for our body, exercise and weight lifting, actually are not. We’ve talked a little about this with Art De Vany, who I know you are a fan of. Art gave me a lot of ideas and suddenly everything flashed together, when I made the distinction between two types of systems, the organic and the non-organic. The organic has the property that the difference between the living and the dead, the living and the non-living; the living, between living and a machine for example, requires stressors. That’s how the complex systems communicate with their environment. You need a stressor. As with the bones, with your muscles, a lot of things. And usually overcompensate for the stressors–there is a mechanism in biology called hormesis. This table I have in front of me will never get better if I bang on it. Use it and lose it. On the other hand, the human body gets better if it is exposed to the right amount of stressors. Of course, you have to define the stressor and the quantity of stress. But then that makes a difference between two worlds–the organic and the engineered. And now, if you can apply that to economic life–is economic life in the first or second category? If it’s in the first category then we should have bailouts, top-down engineers, everything. If it’s in the second category then sorry, you know, it doesn’t work that way.
9:05 You use a metaphor that I also use myself, which I find deeply provocative and educational, which is the forest fire. So, if the goal of fire policy is to have no fires and you are constantly putting out every fire as quickly as you can–which is to me what the bailout policy of the last 30 years has been about–it’s true that in the short run it looks great because there are no fires. There’s no damage, everybody is fine, it turns out okay; and people even brag–we didn’t have any out-of-pocket, all we did was guarantee these loans; we didn’t even have to pay anything because the guarantee never was invoked. But the problem is, is that the brush starts to build up–in the natural case–and then when the fire does come, it’s the fire of all fires. It is incredibly destructive. Exactly. So, what you are doing is you are not lowering volatility. You are increasing fat tails. Which means that most of the contribution to the damage will come from a small number of events or one single event. It’s the same thing in economic life with Greenspan, or anybody. If you put someone in a germ-free environment for 15 years, then invite him to take a ride in a New York subway at rush hour, I think he will last a few minutes. The Greenspan example is–you are fine-tuning and steering and averting every possible bad event; and then suddenly there comes a bad event you can’t overcome any more. Exactly. Because we had something I noticed in 1997 with barrier options–if you take variations, when you spend a long time without variation, you have a lot of exposure built that would be harmed by it. And it’s linear with time. So if you go 5 or 6 months without a new low, for example, and you make a new low, you have a lot more blowups. You see? So quiet periods aren’t very healthy for financial markets. And I noticed it in dynamic hedging, by showing how the number of barrier options start building up below the lows. It was a metaphor that kept in my mind for 15 years and here it popped up when I started talking about the economy, how it causes it to get weaker. Let me now make one statement about antifragility that’s quite important. Is what benefits from the following phenomena: volatility, stress, uncertainty, error, chaos, variation, time–because time is volatility and things like that. Now these categories I’ve just enumerated are qualitatively and theoretically different. Volatility and stress are different things. But the phenomenology is the same. That is what is quite central to the book; had to explain that something that gains from volatility tends to gain from stress. And something that gains from volatility gains from time. So, this is quite central in the book, to explain it early on. This kind of analogy can be very dangerous. There’s a temptation to say–I’ve fallen prey to this temptation myself; I think it’s true but it is a little bit risky–the temptation to say: Well, this works in nature and because economic systems are organic, then the same phenomenon is happening here. The question is: Is that really true and what are the underlying causes of that? And I think what you get at that is so interesting is really the fundamental nonlinearities in these organic, unconstructed, unengineered, emergent systems. The way I think about it is: there is no reason that Federal Reserve policy should really be like fire policy. It’s really interesting, amazing, that it could be the same phenomenon; the question is, is it? I think the reason it is gets at the nonlinearities you talk about, which is that if I have a bad event happen ten times, it’s not the same as a ten-times worse single event. That massive large event, because the damage is not linear, the effects of the damage get dramatically worse as the size of the damage occurs. You talk about this in many different ways in the book. You talk about it mathematically in terms of convexity and concavity; but the simple way that I understood it is: if you have a bunch of little financial errors, even though they could add up to a relatively large number, they are not very harmful. But one financial error that is equal to all those little ones by itself can be unbelievably harmful in terms of the systemic effects. So, this seems to be the essence of what’s going on in these organic systems. Yes, that’s one thing. Whenever you have nonlinearities, then you have a certain effect, and it so happens that that effect translates into either fragility or antifragility. Nonlinearities implies one or the other. And actually, this idea to detect risks that Taylor has of stress-testing, I am collaborating with him on a paper on Taylor risk, seeing if you have nonlinearities in exposure–for example, these stress tests that we do, they are nice; they work sometimes; sometimes they don’t. Why? Because the number you have to come up with for a stress test is arbitrary. Why are we stress-testing 10% and not 11%? So, someone who has nonlinear exposures and you stress-test him at 10% and he passes the test, as many European banks did before the last rout. But then if you test them at 11% or 12% and see the risk increasing then you realize you have a huge exposure to error from that nonlinearity. Now that one, this phenomenology is identical to the one that allows me to detect physical fragility. Let me give you the idea. You take a car; you drive it against a wall at a 10th of a mile per hour, a hundred times, a thousand times. Are you going to be harmed? No. The car will have some damage, but not you. Now drive the same car once at 100 miles an hour. You are dead. The definition of fragility, universal systemic physical fragility, model fragility, has to have some nonlinear term or what we call short gamma in option trading. It’s quite universal. This coffee cup I have on my desk now has suffered a lot of shocks, with zero material fatigue. You can’t see me do it, but if I let it fall to the floor it will break, it will shatter. It’s not a shock that it gets every few hours on the table. What you have is that if you are harmed increasingly, if 10% harms you disproportionately more than 9%, 11% disproportionately more than 10%, you have accelerated harm, then you are fragile. And vice versa for antifragile.
16:57 Let’s talk about the antifragile side, because that’s the less intuitive side. We understand this idea that if I bump a coffee cup a little bit against the side of the table, I can do that a thousand times and there’s no damage. There might be no cumulative effect, even. It might be relatively harmless. But one large shock, like dropping it three feet rather than an inch 36 times is totally different. It’s not just a little worse–it’s over, it’s dead. Now how does that work on the antifragile side? The argue has to be, to carry it over, it gets stronger and stronger and stronger, so that in fact, the bigger the bump, the better, as long as I stay below some threshold. Exactly. Of course, the second derivative flips in sign at some point. But a very simple example: you go to the gym. What’s better over a week–to lift a tenth of a pound a thousand times, or to lift 100 pounds once? So, that’s my body. You have fast twitch fibers in your body and these are very antifragile. Fast twitch fibers require strain, even more pronounced. But you have low twitch fibers that benefit from say an earlier–it has some convexity that stops very early. So, there’s a chapter in the book on health, which listeners know I’m very interested in, recently I’ve started to eat differently and work out in a rigorous way. But it’s much more important than just about your health. A huge part of the book is about both career advice and about policy advice. Let’s talk a little bit about career advice. You point out, it’s a provocative idea, that the tails are very different from the middle. Essentially again a nonlinearity. So, a person who is working at a minimum wage job, you argue, in a certain dimension is very antifragile, and a person who has a nice steady job at a bank, who is say a clerk, is very fragile. And yet that’s not our common sense. Talk about those differences. There is something I call the barbell in the The Black Swan where I showed that a portfolio that has 90% in Treasury Bills and 10% in extremely risky securities was more robust because your measure of volatility–I mean, provided of course adjusted for inflation or if you have inflation-linked securities for 90%–would be a lot more robust because it doesn’t depend on computation. It doesn’t have model error. It was very risky, you know it’s very risky; so that’s a lot more robust than a regular portfolio. Likewise, if you walk and sprint, as De Vany says, it’s vastly better for your health than jogging. Right. And then I continue: if you want, a company should have a certain amount of its energy in the things that make money, but should have dual strategy of very safe and then very speculative. Just like monogamous birds–the strategy in female kingdom, even in monogamy, is you have the accountant, and once in a while is you get pregnant by the rock star. That’s the barbell strategy. Another barbell strategy–I notice that the great writers did not work as journalists. They had a cushy job that had a barbell; and then when they wrote it was very speculative literature. Like Kafka didn’t write a journal article a day. It’s not like he was a journalist. That’s very common in the French literature–when their parents are rich they become diplomats; when the parents are poor, they become postal workers. And you write on the side, rather than have that middle. Einstein was a clerk; and then did his theories at night instead of being an academic. Faulkner worked in a boiler room somewhere when he was writing his best books. But why is that relevant? Why is it bad to work as a journalist? That seems okay. Things in the middle sometimes–some I can explain, some I can’t explain. But when I go to a restaurant, I like to have my steak and salad first and then the dessert later; I don’t want to mix the steak with the salad and the dessert and bring them to me at once, you see? You want a separation of function, a separation of things. In a lot of domains it confers some robustness on the grounds that to get antifragility, first you have to remove fragility. It’s not symmetric. Antifragile is what has a right tail. Fragile is what has a left tail, but the antifragile requires no left tail. In other words you have to clip your left tail; you have to have so much safety and then be very speculative, a lot of volatility. And underlying that idea, which is very beautiful in the book, is this idea of an option, where the downside comes, but you are not subject to it. You can refuse it. And then you can embrace the upside. Exactly. And I took it to epistemological grounds by saying that in finance you may pay for the option because nobody will give it to you for free and typically an option is highly overpriced because people are scared of them when they are labeled as options. But in real life, people don’t notice the options. Optionality situations [?]. And I took it to epistemological grounds by showing how if trial and error is a rational option–in a sense you keep what you like, nature, actually, thinkers, and we discovered only 40 years ago that nature knows it can’t make a perfect baby, so you have spontaneous abortions 50% of the time without anyone knowing about it. So, trial and error is embedded in my idea of the convex economy. Most mutations are not productive and they get weeded out. They don’t get to reproduce; they are failures and they are just thrown out. Exactly. There are fractal layers of antifragilities. There’s a gentleman who read my draft, top notch physicist turned geneticist; and he uncovered the mechanism within your cells that causes, why premature aging comes from not stressing your system because there’s a competition between molecules in your system. So, what you have as a resolution is an antrifragile process. It likes some volatility, otherwise you don’t have selection bias. There is selection. Likewise in your body you have competing things, and you want to always accelerate the weakness of the weak in order to get rejuvenation of cells. And there’s a mechanism called hormesis that was known by the ancients–if you give someone a little bit of poison he becomes immune, a process called mithridatisation. Mithridates was a king who became immune to poison that way. But they noticed also that it went beyond, was actually stronger with very small doses of exposure. It’s not what you think. It’s definitely not homeopathic. It is something actually that has a lot of domains. Small dose of radiation protects you from skin cancer. It’s universally observed in science. It’s very difficult to find a counterexample. The expression in everyday life is: The dose makes the poison. Meaning in a small dose it’s good for you; in a large dose it could kill you or will kill you. We see it constantly. The discoveries about the benefits of alcohol; obviously in small doses it appears wine is good for your heart. In large doses it destroys you. That’s true, but you have to be careful. That’s what Paracelsus, one of the pioneers of homeopathy, discovered. I’m not talking about homeopathy where it’s a matter of dosage. I’m talking about something that’s evidence based process by which some amount of exposure causes an overreaction on your part. And actually, when you go to the gym, what you have is an overreaction. I lift 100 pounds today; my body will project that I am going to have a stressor slightly worse next time. So it prepares for 105 pounds. Provided you have enough recovery. Homeopathy is something where you have very small things. The same applies to both, but I would be careful. I don’t want my ideas to be used by people in homeopathy to justify their methods, because their methods have not yet shown any scientific empirical validity. And you are not suggesting you should be sprinkling arsenic in your food. You are saying that the right level of stress in small doses–what doesn’t kill you, makes you stronger. Yes. And also but you have to make sure that you don’t have selection bias as well. And I wrote about some illusions we have about things that get people stronger when in fact it’s just selection bias. For example? There are a lot of things in antifragility that are counter common wisdom. You talk about that because of selection bias it looks like it’s making it stronger but in fact it’s just weeded out some of the weak ones. Exactly. Which is effectively how it works within the cells. It’s more a destructive process within the cell than a strengthening process. And it can do the same with populations, you see?
27:52 Let’s talk about via negativa, which is a phrase you use to capture addition by subtraction. Part of this is a reaction to some of the criticism you got from The Black Swan, where people would say: So, every once in a while there’s this really bad thing that happens, so how do I predict that really bad thing? And your insight, which is so profound, is that the whole point of the black swan is that you can’t predict it; so rather than figuring out when it’s going to come so you can be ready for it, the best thing is to create an environment where it can’t hurt you very much. Exactly. And shifting to non-predictive methods. When I started writing this book I had a lot of bitterness in me because people didn’t understand my Black Swan. They were buying it; it sold several million copies. And most of the people didn’t get it. I’m not talking about I’m trying to prevent people from exposure to position errors. And nature is not predictive. And via negativa, is what I call negative advice, acts of omission. And when people ask me what should I do, I tell them: Don’t get exposed to negative black swans. It’s very easy to remove negative black swans from your life. If you are a bank, sell tails; if you are in a business don’t do these things; if you are an individual, don’t smoke, don’t have fructose sugar. So people didn’t understand; they said: I want advice. And they couldn’t get the point that the rookie tries to win; the professional tries to make the other person lose. Or for example, via negativa, by removing things than by adding in. People didn’t want to hear it. By preventing people from smoking you save more lives than by producing insulin cumulatively. Likewise all these things–people are looking for the miracle drug or the miracle cure or the silver pill or anything to make people live longer. The philosophers don’t realize that starvation is the most effective way, and we have enough evidence of that. We even now have evidence and another paper last week that if you starve people for four months, you reverse diabetes and they can gain their weight back and they will still be normal. Reverse–by subtraction. Remove food. The ancients knew that. So, the via negativa is–I’ve got a little more aggressive with via negativa by showing that removing does not have side effects, or no long term side effects. You see? Whereas in a complex system, every time you add something you have side effects. And you can’t predict what they are, by nature of complex systems. So via negativa is not. People ask me what would you do in economics; I tell them: just remove Geithner to start with; everything will be okay. Then we will have less control by certain crowd. Democracy works by via negativa. The idea is not to find good rulers but to be able to remove the bad ones. The American Constitution, when it was put in place, was designed to talk about what government can’t do, and instead government has become more concerned about what it can do.
31:44 You talk in the book very provocatively about our bias toward intervention, which this is all a part of. I want to know the vitamin pill I should take. I don’t want to hear about what I shouldn’t be doing. I want to be proactive. We have this incredible psychological bias toward being proactive rather than passive. Which again is a nice example of your opposite. The opposite of proactive isn’t reactive. It’s doing nothing, it’s passive, doing nothing. Letting things happen. That doesn’t mean nothing is going to happen. It sounds terrible. Fabius, the Cunctator, [?] whose big achievement resisting and making sure Hannibal destroyed himself before attacking. Big general; he was not a coward. He just said: My asset is time. This ties in a little book, well to me a lot, where you have a nice diagram in the book. You have a giant cube, which is the real world. And then you have a little tiny cube, which is what we know about the real world that is scientifically rigorous, experimental, tested through rigorous techniques. And unfortunately it’s a fact that what we understand rigorously is a very small part of what we’d like to understand. But we can’t help ourselves. We want to believe that that real world is a lot more understood than it actually is. How do you explain that? This is where I have the character Fat Tony where he didn’t have knowledge of the small world but he had rigor in the large world. And what matters is rigor in the large world. Last time we spoke I was talking about this fantasy we have. Procrustean bed. This fellow as bad and he would abduct travelers and feed them; and put them in his bed. And if they were too tall, he would amputate their legs; and if they were too short he would stretch them. The bed was always a perfect fit. We have a tendency in a lot of things to try to put people in a Procrustean bed. Fat Tony had the rigor of thinking outside the bed, or outside the box. But it’s not trivial, because there are rules outside the box. And these rules–you can actually catalog them; you can actually change the small world by coming from the large world rather than doing the opposite. The artistic [?] whatever you want to call it. Analytical, scientific. [?] Like trying to impose logic on [?]. Logic works but seems a lot more ambiguous. So, via negativa is a central idea. Harnessing antifragility [?] by trial and error and focusing on payoff, which is non-predicted, rather than focusing on probabilities which have predictive thing in it. Probability is predictive, has a predictive aspect to it. So it has a different paradigm when you work in the large world. And of course we have the invisible what I call logic of rules we have inherited from the elders. And typically, just like you mentioned, the American Constitution, you have the heuristic work, heuristics, inherited from the elders. And typically these are all the negative rules, what not to do. I think that’s probably almost always true. Yes. Like debt. The Babylonians had an interdict against that; so were the Hebrews. The Greeks did not like debt but didn’t have a strong interdict. [?] had a fatua about that. “Neither a borrower nor a lender be”–Shakespeare. And then of course Islam bans debt. But the language used by [?] is vastly stronger than the language used in Islam against debt. So you learn. There’s no reason you don’t understand the logic. There is something in risk management–the largest n you find, you are going to find it in nature or in history. So, if you are talking on statistical grounds. Big sample. And yet we always think we just need to manage it. Debt is bad of course, but I don’t need to worry about it because I’ve got a theory, I’ve got a model, I can make it work. I’ve used the very same convexity effects to show why you need stochastic, why of course hormesis, what you were calling benefits of small doses of something are less harmful, and all the other things. There is the same convexity effect. Size and debt, and overspecialization, fragilization. And leverage. They fragilize. A redundancy makes you robust. And actually necessary even for becoming antifragile is to have large redundancies. It may seem strange, non-optimal, but we have inherited, as I said last time, we have two lungs and two kidneys. An economist would never design a human being with two lungs and two kidneys. It’s wasteful. Deadweight loss. So, the opposite of spare parts would be debt. And nature doesn’t like debt. Nature likes redundancies. This mechanism of overreaction is redundancy. Let me give you a little hint here. When you take a trader or risk manager at a bank, they look at the past for the worst-case scenario. And actually making a mistake does not recursing. They say: Well, the worst day was 22% in the stock market, so let’s use 22%. The human body doesn’t do that. The human body doesn’t think that the worst harm is going to be the next worst harm. The human body builds a margin on top of that. It thinks the worst day was 22% in the market? The human body would say, let’s calibrate and adjust and manage for 27%. Or 50%. Not that much. If I lift 100 pounds, my body will start coding for an ability to be able to lift 105 pounds. And then if you lift 105 pounds, it will code again for a little more. There are limits of course, structural limits, but they will go to these limits.
39:28 The other side of this, which again is incredibly fascinating–it helps you see the way the world is working these days–is that in today’s world, because of policy we’ve imposed already, we allow fragile people to impose–they become antifragile and they push their fragility onto others. You have many examples of this in the book. Talk about some of those and what that means. I went into ethics in the last section of the book, what I call Book V, because I discovered at some point that I had a lot of books in the book. So I call them Book I – Book V, because of their separate topics after the first section where I present antifragility and the last one I call “Skin in the Game: The Ethics of Antifragility.” What happens is that some people in society have the option, namely the bankers, the managers of businesses, they have other people’s skin in the game, the left column. No skin in the game other than they keep the upside and transfer the downside to others. You can see this, the stock market has lost about $5 trillion over the past 10 years, comparatively because a lot of stocks were at a higher level compared to cost of funds. Managers of companies made $400 billion. Why? Because you have the upside and no downside. So they actually own the option and they benefit from volatility. So, no skin in the game–in that category I put bureaucrats, journalists, corporate executives, bankers of course, and other people I call Fragilistas. Now, people with skin in the game are citizens, people who have the upside and downside of their actions. If they don’t pay their mortgage, they lose their house. Or, if I make a mistake, then you have skin in the game. And of course my rule of ethic is what I write about, is to have the corresponding position in the market, so I don’t really care about right or wrong; what matters is the payoff. Anyway, so this is the middle column, people with skin in the game. And of course the right column is an interesting column of people who actually don’t have upsides. They are there to take the downside of others. And they have the highest status in society, traditionally. Compare a banker who has upside and no downside, because they don’t have negative bonuses, to someone in the military. He doesn’t have a bonus and he has his life on the line. It’s a very beautiful insight. Honor is bestowed on those who take the bullet for others. Exactly. These don’t have to be a saint, a knight, a warrior, a soldier, a prophet, or a philosopher in the pre-modern sense. Or a maverick. You can be just like the babysitter who pushes herself and lost her life because she had a responsibility for the baby she was holding. And you argue that modernity is pushing more and more people into the left column, the Fragilistas who impose their downside on others, and we don’t spend as much time on that right column. Our culture doesn’t do that. In fact, we look at those people sometimes as fools–they could have avoided that harm; they are idiots. Exactly. And in fact it’s the first time we have power for people who don’t have courage. It’s the first time in history in which the people on top have power without courage. First time. You cannot find that in any society. Take the knights. The knights were people who, their trade was that they were risking their lives. This is why they or lords were supposed to die first. And of course the United States was supposed to be first in battle. Not someone pushing a button. It changes the incentives. So, the only way you can have a safe society is by moving the first column, the left column, moving these people out of there, making it more accountable. It’s hard to get there, though. You can, with the legal system you can in various ways. But I think society will explode because then you start having a growing wedge between what is ethical and what is legal. And I name names in here; I don’t know if you want me to name them. It’s your choice. I’m willing to take a lawsuit. It’s your book. I name people who to me–for example, some academics can cherry-pick; they can give contradictory advice and retain the one, because they don’t have skin in the game; they don’t go bankrupt from the bad trade, so to speak. They are always going to be there. The problem is a Nobel Prize in economics. People are not penalized for being wrong. You cannot have a proper functioning. I just wrote a paper in a policy journal in which I show that without accountability, risks will keep going because people will hide them, and we had a system in ancient societies for that. That’s Hammurabi’s Law. Hammurabi understood risks very well. And it was as follows: if the house collapses and kills the owner of the house, the architect is put to death. Why? Not because to punish people. As a deterrent, because no inspector, no regulator, nobody will ever know more about what’s in a foundation than the architect himself. So, you can hide risks from society, you can cherry-pick, you can do a lot of things unless you have a direct responsibility for the results of your action. When I put it in the New York Times, it seemed too simplistic. I put that proposal there by saying capitalism is not about incentives; it’s about disincentives. People said it would be too easy; can you implement it? Of course. Society can only survive when everything is based on very simply heuristics, not 2800 page documents. No appendices with lots of Greek letters. No.
46:37 It’s a deep idea because it really gets at this interface between what I would call effectiveness and ethics. I had a conversation this past week with someone; I was suggesting that the Federal Reserve, the Chair over the last 15 years, has encouraged imprudence. Has honored and rewarded malfeasance; has insulated people from recklessness. Relentlessly. And maybe there are some bad incentive problems with the way the Fed was structured. And her reaction was: I don’t like to think that about the Chair of the Fed. I think they are trying the best that they can. That’s nonsense. No society has ever put someone in a position of responsibility without accountability. I said that’s nice; it could be true; but when has there ever been a person who had that much power who didn’t succumb to the dark side of it? It’s nice that there might be such a person, but like you say they don’t really exist in history. So the counterargument is their reputation: their reputation will protect us from their malfeasance. But the problem is it’s very hard to evaluate. It’s true. When you don’t see the link between action and consequence. So, when I talk about fragility in my central chapter, chapter 5, on political systems, I talk about localism. Talk about Switzerland. People think that Switzerland doesn’t have a government, or it has a government at a very localized level, it’s a collection of municipalities and the noises washes out. So their political system is remarkably intelligent; it’s just that it’s small. So the mistakes are made small, and things aggregate up without the mistakes. The other element in it, concerning ethics is if you make a mistake in forecasting or make a mistake, any kind of mistake, it’s not like Alan Greenspan who has never done wrong to you or other victims. You encounter these people Sunday at 10 o’clock at church. So you have this biological skin in the game, that doesn’t exist when you–at Russell’s for example, when you can have this loss of ethics at the level of [?]. Shame. Shame plays a role in everyday life. If the only people you see are the people you are helping. Exactly. This is why in my column to the right I have the artisans. The artisan, the last great artisan we had was Steve Jobs. Where people have their ego in the game. They have their product. You know, Steve Jobs had the inside of the computer, he was just like cabinet makers, unlike these commercial shelves you buy, they look great on the outside but ugly on the inside because their inside is not meant to be displayed, to be seen. It’s the same thing with Steve Jobs–he had the inside of the computers look good. Although you can’t open them. Which is kind of crazy. That’s an artisan; he is not in the game except for to do a little authenticity [?]. That is artisans. You will see that with politicians at a local level. This is one of the great arguments. Now Sweden, you think has some 60% of the GDP in government. It’s not the same as the United States because the bulk of the money is spent locally. It makes a big difference. A huge difference. I’m not sure that’s the only difference between the countries. Size has a lot of effects. I use the same argument–size is visible everywhere. Forecasting errors grow with size. I have a very sinister theory of the size bias, which is–you point out Switzerland works very well, that small mistakes are relatively harmless, they just stay small, they don’t aggregate up. And then you look at the euro, the European Union, and you think it’s obvious that the European Union (EU) is not going to work well; it’s too many people, there’s the accountability, and the feedback loops aren’t there. And yet the people who want to run the EU are going to tell you it’s so much more efficient. Well, of course it is if it could be run by God, but when it’s run by human beings it doesn’t work that way. Exactly. They themselves have a certain principle, this subsidiarity [?] principle, that any problem should be dealt with at the lowest possible level. Only necessary things should go up to Brussels. But given that you have a lot of bureaucrats in Brussels, these people make their job look–of course they are going to create jobs for themselves, and you know, metastatics, bureaucracies have always been metastatic. That’s what destroyed ancient Egypt, by the way, the first centralized nation-state. Metastatic autocracy. Doesn’t work very well.
51:55 Let me ask you something–I’m going to take a shot at you; and then I’m going to defend you; and I want to know if maybe your defense is different. But this is the way I defend your ideas. Some people have said about your work–I would consider this your third work in a trilogy. It starts with Fooled by Randomness, it goes to The Black Swan, and it goes to this one. And some people say it about your first two books–they may say it about this one, too, but it doesn’t really matter–they say: There’s nothing really new in there; we knew it already; we knew there are tails; this is uncertainty; people understand that probability is hard to assess, that we get fooled. And I have a very different take on your work. Basically I see it as: it’s all the same book, and I view that as a plus, not a minus. There’s nothing new under the sun, to take a very old insight. What’s new is how we think about them. And it gets at your distinction between the inventor and the implementer. People love the inventor; they don’t give much honor to the implementer. But the real issue is insight isn’t worth anything unless you absorb it. So, if I tell you don’t put all your eggs in one basket, and you write it down; but if it doesn’t get into your bones, it’s not going to change your life. And the power of your ideas is that they get in your bones, the way you write, the use of metaphor, the use of humor, the use of characters; and you go deeper and deeper into these ideas; and they are very deep. Both your criticism and your answer confirm one thing to me, is that what I’m saying is not wrong. That’s a plus. Exactly. The point is that when people tell me that what you are saying is not new, and they find it completely in predecessors [?], it makes me smile, because what I am saying is exactly the opposite of what they consider not new. Traditionally we have put the most trust in small probabilities, the most risk in small probabilities. And I am trying to stand this argument completely on its head: this is where we don’t understand anything. And it was never written before that small probabilities are completely uncomputable. You see? In proportion to how small they are and in proportion to p, the probability itself. And this is what they are not getting; but the fact that they are saying it is not new means that they are agreeing with me. So, what I am proposing here is a system; and it took me a long time to develop it. I know I’m saying the same thing; I hope I’m saying the same thing–because as I said in the Introduction, I’m not writing close-ended books on closed topics with an expiration date. I am going deeper and deeper into the central element of daily life, or of life–what you do when you don’t know what is going on. That’s a great way to describe it. We face a government, an individual, a corporation, a dentist, anything–what do we do when we don’t know what is going on? That’s the most important question that I am obsessed with, and I am trying to answer it. Now, the best compliment I can ever hear is when someone tells me it’s not new, because I know it’s not written elsewhere, or maybe some portions of the derivation exist elsewhere. But this notion that the smaller the probability, the less we know what’s going on, is exactly in reverse of the common understanding of such a problem. And that’s because…? It’s because of convexity effects, because small probability is very convex to error. Meaning the consequences are so different. No, no, no. It’s the probability itself. Take the Gaussian distribution. And actually in a separate paper I finally proved something that has taken me three years. Take a very thin-tailed distribution such as the Gaussian. Thin-tailed, the normal distribution. You have two inputs, one of which is standard deviation. Standard deviation is very much your error. Now, if you take a remote event, say, 6, 7, 8 sigmas, you increase the standard deviation away from the mean; you increase the sigma by 10%, the probability of that is multiplied by several thousand, several million, several billion, several trillions. So, what you have, you have nonlinearity of remote events to sigma, to the standard deviation of the distribution. And that, in fact if you have uncertainty, the smallest uncertainty you have in the estimation of the standard deviation, the higher the small probability becomes and at the same time, the bigger the mistake you are going to have about the small probability. So, in other words, most of the uncertainty in parameterizing the model, most of the tails. So, you take an event like Fukushima, you see, where they said it should happen every million years; you perturbate probabilities a little bit and one in a million becomes one in thirty. Or the financial crisis. Or anything. So, what I meant is, what I managed to derive is the following: Any small probability is derived with an error rate. You have an error rate. Because only God doesn’t have an error rate. God doesn’t have probabilities. You have an error rate, and that error rate has an error rate. And this in turn has an error rate. You continue. Then you end up with power loss. Depends on the regime, of course, depends on how big the error of the error of the error is going to be. And in the power lies a very fat tail. And of course how to privatize. And of course that was how–you can derive the power of fat tails with some counterfactuals. But what is essential here, what I am saying is essential, in my work, is that you can have certainties; you are never going to get into big trouble in the body of the distribution. Where you are likely to get into trouble is in the tails. And the benefits of being right in the tails are very small. It’s a small imprecision, multiplies a 10 sigma event by several trillion; you shouldn’t be talking about 10-sigma events.
59:13 But I want to come back to the way you formulated a minute ago, because I think it’s very deep and very important. You said: What do you do when you don’t know what’s going on? And I would add, which I think comes right out of your book: and most of the time, you don’t know what’s going on. It’s not like it’s a special case. I think about it a lot of the time with economic policy, because when I suggest that maybe we should do nothing or if maybe the government should get smaller, we should reduce debt across the board, people say: How can you do that? You’ve got to have some positive. The reason, say, Hayek, says it doesn’t matter–I don’t think it’s a fair criticism of Hayek, but people say: He didn’t want to do anything. And one counter is: You don’t know what you are doing. You have no idea what you are doing. You claim you have a scientific basis for it. I may have some more solutions here, based on this convexity effect. It’s not all pessimistic. I know, for example, from convexity effects, that I pretty much can map the extent of the unpredictability. And let me give you an example. A specific, big convexity effect or fragility, is in traffic. You put 80,000 cars on a street, you have no traffic. You move 90,000, now traffic time goes up 10%, 90,000 cars. You go to 100,000 cars, and traffic times doubles. Those are those little perturbations, that each person is struggling to keep up with the person in front of him, and that slows the person behind him down; and congestion is very nonlinear. So, we have an idea. You can apply the same thing to the size of corporations. Nature applies it to the size of animals, and the size of an elephant is speculative–there are negative convexity effects. You can pretty much use it with speed. We do the same thing with speed. We actually do it. We limit speed to 55 miles per hour. Because accidents at 55 miles an hour aren’t a tenth as dangerous as accidents that take place at higher miles an hour. So we can do things using this concept, the convexity effect, in other areas. And how much redundancy do you need? How many mean deviations do you need to be away from your accelerated harm? Heathrow, for example. Those who built Heathrow Airport didn’t realize that the smallest perturbation causes 4-, 5-, 6-hour delays. Which means you have to reduce it by a certain amount. And actually, I’m talking a lot to the Cameron Administration to use these methods to control size. So, the analog, in the financial system: Would you favor limiting the size of banks? Or limiting the size of leverage? Or both? Or neither? I don’t know if we can limit as a society. It happens that companies destroy themselves. Unless we save them. Exactly. So we should not save them, and we should have a pact. And what I proposed to the Cameron Administration and they like the idea. On paper. No, no, they like it. They called me to go there, visibly, because they want to really do something. They want to go after size. Size is their big enemy. Before that they had romantic arguments about small is beautiful, before these convexity effects. And I said: It’s very simple; you take the convexity and you certify whether company sales–you don’t know if it’s going to fail, but you know, should it fail, the taxpayer has to bail it out? Yes/No? If you think the taxpayer would need to bail it out, if it fails, then automatically the employees can no longer can no longer get bonuses. De facto, potential civil servants. And you can’t play the long option game at the expense of taxpayers. Remember, my ethics problem is someone who owns the option at the expense of taxpayers or someone else. That would automatically force companies to be of such size as they won’t be bailed out. It’s a very nice pact you make with a company. You say: You can do whatever you want, you can pay each other as much as you want, we don’t care. Provided we don’t deem that you are to be bailed out. Now, of course, it’s a gray area, a large gray area. For a lot of companies we know it’s very visible. We know we are going to bail them out. Therefore they are civil servants. And now by putting caps on how much banks can pay in bonuses, people are moving to hedge funds, the risk is slow into hedge funds, and these are not to be bailed out. Great; let them go. Exactly. So, I am not asking to regulate society. So, government can use something to protect citizens from large corporations. Not exactly a rent directly in proportion to their size. It’s an interesting way to limit their size organically, in theory at least. It’s a way to discourage them from growing because they realize that if they do they will lose their opportunities for the upside. Exactly. They no longer can use the option, it’s the option of society. Or do what I call the Bob Rubin. Bob Rubin had $5 million in bonuses, retroactively financed by the taxpayer. We have to eliminate the Bob Rubin problem.
1:05:08 When you are talking about modernity, your insights into our culture and the clamor by experts for solutions they claim are scientific–it reminds me of “The Second Coming,” by Yeats, where he says: “The best lack all conviction, while the worst / are full of passionate intensity.” It saddens me how overconfident so many people are who have no grounds for it, and how cautious or lacking in conviction are people who really understand the limits of our knowledge. But you see, at some point, things can change, because you can have–we think that systems can survive for a long time while being weak like that. They can’t. Fragility will get you eventually. Can I mention something, do I have time to talk about modernity? Go ahead. Modernity, to me–I love modernity and a lot of things–the rise of the nation state. That’s the first thing, top-down government. In the past, we had some top-down governments, but the government did not have the means to reach around the place. France, even Colbert, could not reach provinces. Only a few cities along the tax routes. This is why France had 400 [?] and about 72 dialects. The good old days. So, modernity is the rise of the nation state, military state-nation states. Rise of the expert, the predictive methods, pseudo-science, the rise of social science; and of course the no-skin-in-the-game. Diseases all. All dangerous. And of course the bailouts. And of course the thing will destroy itself. And what will replace it? I feel the future very positively. You will have more and more artisans. At no point in time, we’ve never been that rich yet we’ve never been more in debt. So, when the dust settles, we’ll have more robust systems. More and more artisans in the sense of people really liking what they do. That’s sort of my definition of an artisan–it doesn’t have much to do with scale. Again when I talk about scale, scale is specific to the industry, specific to the type of function. You see? You have the death of the nation-state, which we are witnessing; a rise of local government; and of course, the end, the utopia for me, is what we are starting to see in Europe right now. They are all talking about golden rules and the golden rule, Rabbi Hillel’s golden rule, is: No government deficits. That’s a good heuristic. No government deficits makes things a lot less fragile. Very simple heuristic. Again, solutions can only come from very simple heuristics. That’s what we’ve been doing now since civilization started. And of course the codification [?], the first one we have is Hammurabi’s Law.

Here’s a little postscript. You wanted to say something else about Seneca. Go ahead. Most people don’t understand what Stoic is. They think that a stoic wants to sort of be robust, no positive nor negative emotions, get rid of. The attachment from the world. Exactly. Become a vegetable. That’s the impression that for a long time, about 2000 years, of the Stoics, because nobody really read them. People kept commenting on comments. But when I read the best expository of Stoicism, the best two expositors, Marcus Aurelius and Seneca–and probably also to some extent, Cicero–I realized these are not that type of people. Very different. And now, recently I saw some papers confirming my idea. That was what Seneca was, was about being long in options. He wanted to keep the upside and not be hurt by the downside. That’s it. It’s just how to set up his method. Seneca was the wealthiest man in the world. He had 500 desks, on which he wrote his letters talking about how good it was to be poor. And people found inconsistency. But they didn’t realize what Seneca said. He was not against wealth. And he proved effectively that one philosopher can have wealth and be a philosopher. What he was about is dependence on wealth. He wanted the upside of wealth without its downside. And what he would do is–he had been in a shipwreck before. He would fake like he was a shipwreck and travel like he was a shipwreck once in a while. And then he would go back to his villas and feel rich. He would write off every night before going to bed his entire wealth. As a mental exercise. And then wakes up rich. So, he kept the upside. In fact, what he had, my summary of what Stoics were about is a people who really had, like Buddhists, an attitude. One was to have the last word with [?]. And my definition is a Stoic is someone who transforms fear into prudence, pain into transformation, mistakes into initiation, and desire into undertaking. Very different than the Buddhist idea of someone who is completely separated from worldly sentiments and possessions and thrills. Very different. Someone who wanted the upside without the downside. And Seneca proved it. And the way you get there, Seneca is suggesting, is through mental exertion. Through renunciation–some of it’s action, but some of it is the way you look at your life and what you prepare yourself for and how you affect your expectations. Exactly. He understood the hedonic treadmill that Daniel Kahneman rediscovered 2000 years later. He understood it very well. And he understood wealth, debt from others or from fortune. And he wanted to write off debt from fortune and he wanted to remove his dependence on fate, on randomness. He wanted to have the last word–was randomness. And he did. Not a bad goal.

H/T: Value Investing World