Loading summary
Transcript145 lines
- [00:03]
A
Welcome to Econ Talk, part of the Library of Economics and Liberty. I'm your host, Russ Roberts of Stanford University's Hoover Institution. Our website is econtalk.org where you can subscribe, comment on this podcast, and find links and other information related to today's conversation. You'll also find our archives where you can listen to every episode we've ever done going back to 2006. Our email address is mailecontalk.org we'd love to hear from you. Today is August 1, 2013, and my guest is Nassim Taleb. His books include Fooled by Randomness, the Black Swan and Antifragile, and we've done podcasts talking about those before. Our topic for today is Skin in the Game. And as a jumping off point, we're going to use a recent paper Nassim has written with Konstantin Sandis titled the Skin in the Game Heuristic for protection against tail events. Nassim, welcome back to Econ Talk. Hi.
- [01:03]
B
Thank you for inviting me. This is becoming. I mean, I'm a habit. I'm never ever invited by the same person twice.
- [01:12]
A
I'm afraid that says as much about me as it does about you, Nissan, but here we go.
- [01:16]
B
Definitely, yes.
- [01:18]
A
Let's start with this expression, which is familiar to some in America, but maybe not to non native English speakers and to some even English speakers. What do you mean by skin in the game?
- [01:32]
B
I mean that I cannot scale in the game. I cannot take risks entailing, you know, that may harm others without being subjected to them myself. That's what I meant by scaling the game. In other words, that you cannot possibly make a bet on entailing a random variable that can harm others without you yourself being somewhat harmed. It doesn't have to be as harmed as you need to incur some personal harm enough to be deterrent. Now, this is quite potent because it is probably the earliest idea that ever, okay, Disseminated in society. It's the first idea, probably. It's definitely behind the first document. I mean, the oldest document we have, Hammurabi's code. So it is very potent and in it, and we'll explain a few minutes, there's so many things. I mean, it allows society to function for several thousand years, okay? Particularly, I mean, maybe not between societies, but within societies, but unfortunately is gone today from consciousness because of modernity and all these fancy ideas.
- [02:43]
A
How does it show up in Hammurabi's code?
- [02:46]
B
Okay, so in Hammurabi, basically, Hammurabi is probably not the oldest mention, but the oldest One extends, so it says the following. Say the architect or the engineer, depends, I can't translate Babylonian. If he builds a house and the house collapses and kills the owner of the house, the architect shall be put to death Now. And it also continued that if it kills the firstborn son of the owner, the firstborn son of the architect shall be put to death. Now the point is that the Babylonians visibly didn't have much against architects. They liked architects, they have of course, their suspended gardens. But the thing is that they wanted it as a deterrent. They wanted the idea as a deterrent. And behind it lies this very powerful idea is that the architect can hide tail risks, you see, delayed blow ups. Because the cheapest, the place where you can hide the risk is by cutting corner in the foundation. And of course he's not going to be there when a thing collapses. And the Babylonians detected it, of course they understood that no inspector, no regulation will ever, ever, ever outperform that simple rule. And behind that rule, of course, Konstantin and I went back to some, you know, archaeology, some cultural archaeologists at the dig and we found that that rule was directly behind the lex talonis, namely, you know, the eye for eye that of course the Semites spread. An eye for eye, a tooth for tooth. In Exodus we see it. And of course they have love your neighbor as yourself, Leviticus. And of course it kept going. The Silver rule by Hillel and Rabbi Hillel, the one that he was asked to explain the Torah. Standing on one leg, he said, don't do to others what you would not have them do unto you. And the rest is commentary.
- [04:56]
A
That was.
- [04:57]
B
And Isocrates at about the same time came up with his own version of golden rule, but with a difference. And that's neglected because it's almost never mentioned that he applied it to states and he said a state should not do to other states, weaker states, a strong state should not do to weaker states what he wouldn't want a stronger state to do to it, you see.
- [05:22]
A
Who applied it? Who applied it to states? Sorry, who applied it to states?
- [05:27]
B
Isocrates.
- [05:28]
A
Oh, Isocrates. Yeah, yeah, yeah, Isocratus. Around the same time I say, yeah, yeah, like I know Isocrates, but I remember it from your paper. Keep going.
- [05:35]
B
Yeah, Isocrates. I mean we're talking 5th century already, right? And of course we have the golden rules that we see in the New Testament, which is a positive, you know, up until then it was a negative rule. Don't do to others what you don't want them to do. To you. And then the golden rule, of course, do to others. So we had a silver rule till then, and then do to others what you want them to and so on. So here, what we see behind this is a foundation of moral philosophy as a foundation of ethics and foundation of civil society. But in it, we saw something much more potent. We saw the foundation of risk management. This is why you cannot disentangle risk management from ethics, just as you cannot disentangle economics from moral philosophy. Some people try it, they can't. So this is our paper and of course, we put some probability theory, and I'll talk about it, I'm sure, a little bit later, that how Taylor risks are impossible to figure out unless you're the one causing them.
- [06:42]
A
Yeah, I want to go back to Hammurabi and the builder or the architect. And before we do, I just want to mention, because it's just a pet peeve of mine, that in the book of Exodus, when it says an eye for an eye and a tooth for tooth, in its application in Jewish law, as elaborated in the Talmud, it substitutes the physical eye for the economic value of the eye. So if you have an accident, by an accident, you poke out somebody's eye, a Jewish court would not poke your eye out. They would force you to pay damages equal to the loss of an eye in a fairly sophisticated way of how that might be measured and is elaborated in the Talmud. But I just want to mention that. That the eye for an eye, tooth for a tooth in Exodus was not implemented in Jewish law by the literal description of the text. But I want to go back to the builder because.
- [07:31]
B
Yes, sorry, one thing here I just mentioned earlier. I said about the same time, Isocrates and Hillel was not the same time. I mean, there's still several centuries separating them between Isocrates and Hillel. That just came, you know, just realized.
- [07:45]
A
Yeah. And they probably didn't know a lot about each other's work. Hillel came after Socrates. He probably may have known him. I don't know.
- [07:53]
B
No, but there was strain, what people call that great transformation when the thing started spreading. And then finally, of course, when you look at Kant, the modern culmination of this idea is in Kant. Kant, of course, is act only in accordance with that maxim through which you can at the same time, will, you know, make it a universal law. You see?
- [08:16]
A
Yeah. When you decide what would the world be like if everybody behaved as you did.
- [08:22]
B
Exactly. This is a universal law. So it's more generalization of all these laws, culminating in Kant. And we're still stuck on Kant and.
- [08:28]
A
It'S the categorical imperative. And I see that, by the way, as a way to get people to morally overcome the free riding problem. The idea that you can get away with something, which you often can. But what Kant is telling you is that even though it might be rational in some narrow sense, even though it might be in your self interest, it's immoral often to act that way. And Kant is a way of reminding you of that. And just as a side note, Adam Smith's concept of the impartial spectator is another way to think about some of these moral rules. Smith wanted you to think about your behavior as if someone were watching you. He says that's the way we actually do behave. It's a positive theory. It's not so much a normative theory, he says has normative elements, but he's trying to describe the way the world works. He says we act as if someone were watching us. We look at, we step outside ourselves and put ourselves in the shoes of an impartial spectator. And there's an interesting question of whether that influenced Kant or not. They were contemporaries, unlike Isocrates and Hillel. But I'll put a link up to a paper on that as well. So I want to go back to the builder though. So the builder explain why the incentives for the builder are so powerful and how they relate. What are we trying to avoid? And basically you're talking about opacity, that is non transparency and information asymmetry. These are two ideas. I think they're important to generalize. Exactly.
- [10:01]
B
But there's a third one. There's a third one involved here which is what is behind our paper. Because our paper doesn't bring anything new except for one concept behind it. Okay. And that's as follows. I mean, I spent all my life dealing with tail events and is that there's a very simple property to probability distributions that generate tail events, particularly while they're asymmetric. And you know, you make a penny, you make a penny, you make a penny and then you lose dollars. So. And the property is very, very simple. As I mentioned in the mathematical part of the paper. I say if you have a process such as a Pareto 8020, you know that 8020 that people talk about. Okay, no, no.
- [10:42]
A
What do you mean? Explain.
- [10:43]
B
8020 means 80% of the companies generate 20% of the profits and 20% of the company generate 80% of the profits.
- [10:49]
A
Or 80% of your inventory accounts for 20% of your inventory accounts for 80% of your sales. Yeah, yeah, okay, exactly.
- [10:56]
B
So if you look at something that 20% of the days generate 80% of returns and stuff like that. Okay, which is what we effectively observe in many domains. Well, the point is that in any given year you will be 90% of, have 90% of the observation above the true mean. You see, and that's what it's.
- [11:21]
A
Because it's asymmetric. Because it's not.
- [11:22]
B
When it's asymmetric, you don't observe the mean. You typically observe the, you know, you observe profits, profits, profits, profit, profits. And in fact, the mean can be negative. Take the banks for example. The banks, you know, someone was boasting that they had the whole quarter without losing money on a single day. It's totally irrelevant. Banks lost $4.7 trillion during the crisis. So if you make small gains and large losses, you see, you cannot observe the statistical properties from a small sample. You have small sample effects. And I've done something. What we're talking about is an anti fragile, okay? This ethical rule is an anti fragile is behind, you know, what I call fragility of society or someone who trades antifragility at the expense of someone else, where someone has an option against someone else. But the ideas that I've been working on now are the ideas of the black swan trying to mathematize it. And I showed the following things. The most misunderstood thing in society is a law of large numbers. You take, you know, observation and you think and figure out the mean just from observing what's going on. In reality, it's not that way. Most distributions have the slow numbers, if you want to say. In other words, the mean doesn't reveal itself very easily. So people put themselves in a situation where the frequency of profits is very high. So I manage money for someone say, you know, for a fund for whatever it is. If I don't have skin in the game, my optimal strategy is to shoot for small steady gains and rare losses.
- [13:10]
A
This is picking up nickels in front of the bulldozer. So you're just, most of the time, bulldozer is moving slowly. There's plenty of time to pick up the nickels. You look like you're really smart because you've got all these nickels and then one day you get run over, but you say, that was a bad day, it was just bad luck. I didn't, as you point out in the paper, you say things like based on the information, I did the right thing.
- [13:31]
B
Exactly. And there's that property of law of large numbers or something like that it's quite shocking because if you look at fat tails, and that's what, as I said, my new work is about, if you look at fat tails in a fat tailed series, eventually the maximum you will have will be equal to the mean, to the totals.
- [13:53]
A
Explain.
- [13:54]
B
So in other words, let me explain. If you have an option trader or someone involved in a fat tailed game, all right, Say you have, you know, you trade for 10,000 days. Your maximum, your extremum will correspond to the total. So in other words, everything will be dominated by single observation.
- [14:17]
A
Right? So this is what. So basically, in that case, the case we're talking about of the nickels in front of the bulldozer, the outcome is you're wiped out.
- [14:25]
B
Your outcome is the bulldozer is not the nickels.
- [14:28]
A
The nickels are irrelevant.
- [14:30]
B
Irrelevant. In other words, you don't observe. So in a process that we observe commonly, As I said, 90% of the observation will be below the means, which means that nine years out of 10 will not reveal the mean. So you have at least nine years of profits ahead of you. So what happens is you catch a bonus year one, bonus year two, bonus year three, bonus year nine, and then you're a. When you lose all the money and.
- [14:55]
A
You'Re a genius, you have a great genius, of course.
- [14:58]
B
Of course. And I said the thing that if you make money four years in a row in New York, people start saying hello to you, you see, they eventually laugh at your jokes, you see?
- [15:07]
A
So.
- [15:08]
B
So you make money for nine years. Exactly. So you make money nine years and then the 10th year, everything's gone. And you claim what is adverse event. Everybody lost money. It's a difficult year, whatever it is. Like the banks in 1982, you know, Walter Wriston from Citibank said, oh, it's only one bad quarter. What are you guys talking about? Yeah, the one by Porter. They wiped out everything, made a necessary city mask.
- [15:30]
A
Nissan. I thought of you last night. I had a friend. We had a friend over for dinner and she was talking about how scary the roller coaster is at the Santa Cruz boardwalk. It's very rickety and it's built, I think, in 1924. And this example has nothing to do with the actual Santa Cruz roller coaster because for all I know, it's beautifully maintained and it's extremely safe. But when she pointed out that there hadn't been an accident in maybe the whole life of the roller coaster, I naturally thought of you. Because you could have 365 days a year for 90 years with no accidents. And it's obviously very safe. And then one day it's possible for 150 people to be killed. And you say, well, but look how good my safety record is. It's 99.99 out of 100. But that one day is devastating.
- [16:18]
B
That's true, but still, I mean even that's mild because there's no, you don't have as big a problem in modern society.
- [16:26]
A
Well, that's because you don't kill all the people who wrote it on it on all the other days. That would be the analogy. Exactly.
- [16:32]
B
The analogy is it kills back everyone who died, whoever wrote on it. That was exactly. That is what this is. This eventually would be a fat tailed process. So if you take now what was happening today in modern society is you have invisibility of the risks. So someone you know is shooting, of course, for a high probability low impact benefits and low probability high impact, you know, losses. If he doesn't bear the losses, that's the optimal strategy. Plus think about the metaphor I was given earlier. As you're given money to manage, you're going to make money nine years out of row. If you make money for two, three years, you're going to get a lot more capital. So you're going to get bigger and bigger and bigger and bigger until you blow up. So when you blow up, you'll be.
- [17:21]
A
At your maximum, but that's when you have no skin in the game. Exactly. Your point is that if all of your money is in that fund, that picking up nickel strategy is not going to be your natural incentive.
- [17:34]
B
Exactly. Your incentive will be very different and you won't really care about perception, you care about your own money. So what happens is that you won't hear people saying, oh look, he's losing money or he's making money, you don't really care, it's your money in it. So your reputation. So you don't game. The reputations are very easy to game. You should shoot for a high probability low impact event. So what you should require is if you invest with someone, you should tell them, listen, if you lose money, I want you to be harmed a lot more than me. And of course his reputation should be harmed in the reputation but also harm financially because it's not enough of a deterrent.
- [18:18]
A
I'm going to raise a problem now with, with that claim and it goes back to the builder claim as well. You can have too much skin in the game and you don't seem, I don't think you deal with this in the paper and it Seems like an issue you should deal with. Let's start with the builder case, because it's so easy to understand. If I'm going to execute a builder every time that he builds a house that collapses and kills someone, they're going to be a lot of cost to being a builder. And if it's possible, and it is that sometimes houses collapse that aren't the fault of the builder, that they're just bad luck, there's random events and there's a continuum on which we would judge irresponsibility. There's a continuum on which we would judge safety. So if we say to the builder, you have to build the house safe enough so that if someone dies, you'll be killed, so you're going to then build a very, very safe house. It could be that house is so expensive, I would really rather take a little more risk than that. Or if I have a builder who has a different standard than I do of the value of life. And similarly, I mean, there's gotta be some equilibrium.
- [19:23]
B
There gotta be some equilibrium. People will ask the builders who cut corners, otherwise you can't afford the house. And then you end up eventually with the right kind, the right balance.
- [19:34]
A
So if I say that, I mean.
- [19:35]
B
We'Re not, but we're not, we're not hammering in Hammurabi's days. So as I've said at the very beginning, skin in a game doesn't mean matching the exposure of others. Okay? In other words, you don't need to kill the owner of the airline company every time there's a plane crash. But on the other hand, there has to be a painful disincentive beyond, beyond the cosmetic. See, this is what we're talking about. So there has to be some. I mean, look at capitalism. Capitalism has a built in asymmetry in the sense that bankruptcy is zero in it. There's no negative, you see, for a company. But you still can have skin in the game. By forcing people to lose a little bit of money. It doesn't have to be unlimited, so you have unlimited profits and limited losses, but still maintain a skin in the game. I think we are reaching that equilibrium in economic life outside of, of course, government intervention, banking and bailouts. You have that equilibrium. In other words, the builder isn't put to death. There are financial penalties when you go to a doctor, if a doctor amputates, amputates the wrong leg, you don't take the doctor and amputate his leg in return, you see? But there is a penalty, you see. So we're not worried about places in which this equilibrium has been discovered heuristically bottom up. I'm worried about modernity. I'm worried about bureaucrats causing hyperinflation, affecting savers and outright citizens, but not harming them at all. I'm worried about that kind of stuff. You see, we're not worried about contracts between individuals that can find their equilibrium in some way or another.
- [21:15]
A
So we've been mixing together a couple things here and I want to pull them apart a little bit. Skin in the game, it has positive impacts, meaning in the technical sense of economics, meaning it has natural incentives for the people making decisions on the behalf of others. What you're pushing in this paper, among other things, pushing a lot of things, but one of the things you're pushing is that it's a moral concept that I should. That it is immoral to make decisions without skin in the game, that I should insist on having skin in the game. So there's sort of three levels here. One is regulation should ideally put skin in the game rather than take it away. And it often takes away. No, no, no.
- [21:56]
B
Let me say one thing about regulation and hopefully we'll talk about it. The skin in the game to me is a heuristic. The heuristic can be enforced between two individuals engaging in a contact in a contract. Okay? Regulation requires a state. This is the main difference here, you see, is that we're bypassing the state. It's in the contract. I can eat your food if you taste it in front of me. You see, that is a skinny game heuristic and force between two individuals without having recourse to the state. You see?
- [22:25]
A
But you're saying something more than that. You're saying that if. That if I'm the cook, even if my customers don't insist that I eat my own food, I should eat my own food.
- [22:37]
B
That's exactly. So here we have the first level of the contract between individuals. Forget regulation and state. The second level is a moral concept, is I should not make a forecast unless I'm harmed by it. Like people ask me on television or something every time.
- [22:53]
A
This is why they're.
- [22:55]
B
If it's incorrect, if it's incorrect, I should be harmed by it if it's correct. So if I make a forecast, if someone asks me for my opinion, it is immoral for me to say, well, the market is going up or the market is going down or this will happen unless I stand to lose from that advice. Because people take risks based on other people's advice. You see, this is where it's immoral. This is why skin the game is very generalized to daily life. I cannot tell you, well, this is good. Unless I've tasted. Of course, if there's risk, if there's no harm, then who cares? So here we saw two levels.
- [23:32]
A
Yeah, we all understand it's immoral to review a book you haven't read.
- [23:37]
B
Completely, right?
- [23:38]
A
Absolutely. And so to make a prediction about an event that you have no stake in has a similar immorality to.
- [23:47]
B
Depends on how much harm you can cause, you see? So this is. I'm saying in the risk domain, we're saying we're confining it to moral philosophy in the negative tail domain, you see?
- [24:00]
A
Yes.
- [24:01]
B
So if I predict. If I lie to you and predict that tomorrow, I tell you tomorrow is going to be sunny just because, you know, it may uplift your moods or something like that. We're not in a Kantian domain. We're, you know, it's okay because I'm lying to you. I have. And I'm not going to be here to enjoy, you know, the sun or something, but I think that I'm making calculated things that make you feel better. So. So because there's no harm, if there's no harm, you know, that's not my space, you see.
- [24:30]
A
Yeah, I understand.
- [24:32]
B
So. So. So we saw two levels, as we saw the first, the contract with the individual. Two, the contract with yourself, which is the moral domain. Yourself or with humanity, which is the moral domain.
- [24:42]
A
And it seems to me. Go ahead, go ahead.
- [24:45]
B
No, no, go ahead.
- [24:46]
A
So it seems to me. So skin in the game emerges naturally in the dealings between two people. And it seems to me that what. And it should emerge, you're suggesting, morally, if it doesn't emerge through the choices of the interacting individuals, we should be eager to impose it on ourselves as a moral heuristic, a moral code. But it seems to me that one of the challenges of politics, political life, public policy, is that what government's really good at is getting rid of skin in the game. That's kind of their specialty.
- [25:21]
B
Not all governments. I mean, let's look at it. What you and I tend to call governments, because we have the same political colors, is a centralized government. A government can be a local, neighborhood, union. And then what I figured out from the history of countries that have been very successful, like Switzerland or Sweden, place like that, is that people making the decisions are usually embedded in a community, and their skin in the game is typically shame because they're socialized by the community. This kind of Game is shame. Whereas a government official in Washington can make a mistake and it's a spreadsheet looking at them. It's not someone at church on Sunday looking at them and making them feel shame. And that's where there's. That's where the main difference is.
- [26:10]
A
You could argue, though, that there's an offsetting effect at the national level, which is history. Ben Bernanke is going to. It's true that I don't spend a lot of time with him in church or synagogue, but his name will go down in history to some extent for his decisions. Now, the real problem for me is that, as you point out, a lot of times there's opacity, it's opaque. People tell me, well, he did such a great job, you can't even evaluate it. We don't have the data to evaluate whether he's done a good job. So the shame part of, of a blemished historical legacy is very limited in economic policy actions.
- [26:49]
B
Exactly. It's very simple. The law of large numbers works much faster at the micro level than at the macro level. At a macro level, you need thousands of years to figure out to someone a repeat experiment if someone did the right decision. At the micro level, we have thousands of experiments all every day, you see. So it's very different. And this is probably the big distinction. And municipality is a micro concept, whereas a Washington is a macro concept. So my whole idea in fragility and antifragility was to build a system where mistakes can occur but do not threaten the system.
- [27:31]
A
So let me go back to the point about government. It's true at the local, local level there's some natural incentives. But at the national level, say in the United States, a lot of what government does is to remove skin in the game. Bailouts, insurance policies, do overs, ad hoc interventions. Is there any hope that public policy might take your idea and think of it as a way to guide regulation?
- [27:57]
B
I think that the only hope is to rely less on public policy, because systems, when they're localized, tend to enforce kind of game very naturally. You see, you look at, for example, a system where they're small, you can identify cause and effect very easily, and then you can force the cook to eat his own cooking. And you see it in small communities, like in the army, for example. In almost every country I looked at, people who repair helicopters are sort of forced to take rides on them, and people who. And almost every. Everything entailing parachutes, jumping has that kind of skinny game involved where if you fall, someone's Parachute, you may randomly be asked to jump in it, jump using it, or something like this. So you have this. So there's a natural tendency by systems when they're small to produce skin in the game rule and function properly that way. But the thing, it looks like the things get larger, the size effects. Large is ugly, you see, and large doesn't have skin in the game. And large has these warped incentives. And large, of course, likes regulation rather than these simple heuristics. And as a libertarian, I like the government to be the last resource. It doesn't mean no government at all. It could be minimum, but whatever is last, not the first. So when I think of things that should be less of government, you know, these things are related to, of course, law enforcement and things that we have either market failures or systemic inability to protect individual against the largest threat. And there maybe you may need regulation. Like, for example, I'm very concerned about genetically modified food because the person producing them may threaten my backyard, you see. So I want protection and I can't really protect my backyard because I can't fence it against the changes in ecology that can happen from it. So maybe regulation is something to consider in these exceptional cases, but sort of exactly like medicine, emergency room surgery is completely different from cosmetic surgery.
- [30:20]
A
Agreed, but my point only is that if government stopped bailing out banks, there would be a natural skin in the game. And we seem to find it difficult to avoid. There's no we here. The politicians find it difficult. I don't. The idea of it doesn't bother me, but somehow the politicians can't seem to bear to let these people have skin in the game.
- [30:47]
B
Yeah, that's true. This is where we have a problem with metastatic government, is that the larger and more convoluted government is, the more they dislike from the game and the more you have inverse care in the game that people who have everything to benefit from causing you harm, namely through the lobbyists.
- [31:11]
A
I want to go back to the philosophy, to the moral principles you talked about. Can you talk a little bit about the role of asymmetry and the negative asymmetry? Because I want to talk about positive asymmetry, as you do at the end. So talk about what skin in the game has in common with an eye for an eye, with Silver Rule, the golden rule, etc.
- [31:33]
B
I mean, what it has in common is establishing symmetry, the way I approached it in Antifragile. And people are starting to get there nine months after its publication, because it's page 600 that I start talking about skin in the game, the way I looked at it is in terms of fragility. If I had more upside than downside, namely a positive asymmetry, I tend to benefit in the long run from random events. I am what's called anti fragile, so I benefit more than I lose. Now if I have more downside than upside, then I am fragile. And it's very strange how the Think Map, you know, this is one of the wonders of science, is that this asymmetry, you know, of course, is what everything that's fragile on earth shares. And it also comes from disliking random events, disliking volatility, disliking stressors, disliking things. They all is the same equation that explains them all. And the equation that we use in option trading when you short volatility. So basically people are short on option or long an option. So the problem is the class of people who are long on option, namely bankers. If there's volatility, they either break even or make money, they're not harmed. And you know, society is short an option. Basically if the bankers make money, we don't get anything. If they lose money, we end up paying for it via taxes, more regulation, more burden of having people to understand 2500 pages of codes and stuff like that. So this is where the asymmetry is very, very similar to short term option. And of course people not understanding option at the core. There is this misunderstanding of the number that when you have an option because everybody's still economically still in not even 20th, 19th century statistics, where people don't really understand probability distribution and they think that you can figure out performance from short periods of time and that distribution are sort of symmetric, that kind of stuff. So there's a mismatch between compensation, period, however you want, whatever you want to call compensation, and the time it takes because of large numbers for the variable to show its properties.
- [34:00]
A
Yes.
- [34:01]
B
So this is a big problem is that you have an asymmetry and it's undetected. The undetected part is what we're insisting on. When we looked at the paper, we realized that while economists understand the agency problem, they understand moral hazard, but they don't understand moral hazard. Okay, with respect to fat tails, because they never really gave too much thinking to fat tails other than saying, well, it's a peso problem and we can't talk about it because we don't understand it. Thanks, bye. All right. Instead of saying, well, this is a central part of life, what they call the outlier and Then let's deal with it as a central component of life. And if we don't deal with it, it's so large in effect, that is going to end up eating us. So this is sort of like the background. So when I say, you know, that there's option and optionality, it's quite general. You can look at any individual and say, well, this guy is long optionality at the expense of others. Or this guy is short optionality being used by others.
- [35:13]
A
Explain that last case.
- [35:17]
B
That's the taxpayer, the taxpayer or sometimes the investor. You invest in a company, okay, if some lot of the returns are going to go to the industry, are going to go to Wall Street. Because if the fund manager makes money, it's his money. And if he loses money, you pay for the losses. So you get a smaller share of profits than you do of losses. In the long run, you're going to lose money. And this is effectively what people didn't realize about hedge funds is that they think hedge funds are the best thing after sliced bread. Some are very good, but in general, hedge funds have a high level of fees and a lot more optionality to the hedge fund manager, but has been corrected thankfully by people forcing hedge fund managers to have a lot of their money in their funds.
- [36:07]
A
But you also talk about positive asymmetry where people assume risks for others to help altruistically to help them in society. Not just, not so much in the financial sector, but elsewhere, right?
- [36:20]
B
Anywhere in a lot of places, people voluntarily take risk for the sake of others. And this is what economists are fully equipped to understand and honor. When you take the notion of honor, honor is someone who has courage, someone person who he or she has a lot of courage and is using that to save others or help others. So if you realize that people, when you see Solon's statement that you're only happy if you have a glorious death and you see these, that was a mentality throughout the ancient world that you're as good as the risks you're taking for the sake of others, city, anything or even yourself. So in other words, you're not transferring risk to others, you're the one who is taking the risks. And of course the two most famous figures of today are people who died for their ideas, not just to fight for others, but Socrates and a few hundred miles to the southeast, another fellow who you know, died for, you know, who's crucified. So when you think about it, prestige that we have gotten has almost always been proportional to the risks you take for the sake of others. So you have heroes and war and stuff like that. That's the notion of hero. And of course, the lords in medieval England and medieval Europe were people who were forced to protect the peasants. And they have a high rank, but it came with obligations. You took more risk. You were the first person to die. And of course, Hannibal was first in battle. Same with Julius Caesar. Same with all the big warriors, with almost no exception. Now let's talk about our generation. Who do you have? George W. Bush. Right. He escaped war, although his father, up to that generation, they were war heroes.
- [38:28]
A
You see, it was honorable. It was honorable. It was dishonorable to find a way out. It wasn't considered clever or smart. It was dishonorable. Exactly.
- [38:36]
B
So your rank and studied a lot, the mafia and antifragiles by saying, you know what, what's kind of behavior, people, you take risks for the sake of others. That's it. You don't, you're not a risk lover. But you, if someone got to take the risk, you're the one who takes it. And that's sort of like the idea was not to die in a nursing home with tubes coming out of your nose. The idea is to die in battle. Okay, that's, that's, that's what you're, you're made to do, okay. And that was. So prominence came with that kind of lord, you know, lord concept. So. And that I think was prevailing on society. But today, when you know, the gentleman who's the head of the CIA but was a big military person when he was busted, was his name Petraeus? I thought, I mean, I looked at his Wikipedia page and there were all these decorations, hundreds and hundreds of decorations. I saw that the fellow jumped from helicopters at night, climbed walls, the things through grenades, from, you know, through small windows. No, it turned out that the fellow had never been in battle. He had never been in battle. So here we have a generation of people who have never had to take risk for the sake of others. And society cannot function when you have an imbalance between. What I call the first column is people who make others take risk for them. And then you have the right column, people who take risk for the sake of others. You see, it can't function that way. You cannot, you cannot have too many of the Petraeus and George W. Bush who have never taken personal risks but engaged others in war. You can't have too many of these. We need the reverse. And we had plenty of these a generation earlier.
- [40:29]
A
I wonder why that changed.
- [40:31]
B
Technology. The problem is technology, modernity is Causing disruptions in the entire system.
- [40:41]
A
I don't know. I think part of it is how wealthy we are. Staying alive is definitely what we call an economics, a normal good, meaning we want more of it as we get wealthier. So I think we value our lives and our health a little higher than we used to. And so our willingness to sacrifice it is a lot harder.
- [41:00]
B
I don't think. I think it's just probably the culture because the change was very recent. It took place very abruptly.
- [41:09]
A
Right, but why did that culture change?
- [41:13]
B
I mean, we can't look for explanation. The thing that you have to look at the world in which we live today. All right, it's highly technological. It is probably, you know, it's highly technological. We still, we have dangers, but not the same kind we had before. And modernity put the bureaucrat in place of risk taker. Now, risk taking isn't just physical. Risk taking is entrepreneurship. So instead of worshipping entrepreneurs who take risks really for the sake of others, because the probability of success is much lower than that of say, venture capitalist. So they take risk for society and they fail a lot. But collectively we need them because otherwise we can't advance. Instead of having all these people glorified and put on a pedestal. Who do you put on a pedestal? Harvard grads. That's not how society can evolve because it's not the Harvard grads who got us.
- [42:12]
A
Did you say Harvard? Harvard grads? Is that what you said, Nassim?
- [42:15]
B
I use Harvard graduates. I use it as a metaphor. I think that England was built and America were built by adventurers in the economic sense, what Adam Smith called adventures, not by bureaucrats. And then after on the benefits are reaped by the class of bureaucrats who come and try to control the process.
- [42:39]
A
Your remark about entrepreneurs reminds me of when Amazon started and they were losing money every year. And I remember being so grateful to the people who had invested in them because they probably at the time, most people thought they weren't going to make it. But I was building a beautiful library of inexpensive books that arrived on my door very quickly and had a wonderful way to look, shop for them. And they were bearing all the costs and I was getting all the gains.
- [43:02]
B
I mean, look at restaurants. Restaurants, you know, to open a restaurant is like in New York City is an act of suicide, you see. But yet without these people, we wouldn't be where we are. So what I'm saying is that we no longer are giving the respect do to these people who take risk for the sake of others, whether in the military domain or in the economic domain or in other domains as well, in the political domain. I mean, people who have the courage to voice their opinion. But we can protect ourselves, just like the ancient, by finding very simple heuristics. Like Ralph Nader had a heuristic for war. He said, if you're going to vote for war, you should have a member of your family, a descendant, a son or grandson, on the draft. Then you can vote for war. And in a way, it's liberating on both sides. When I manage money and, you know, of course I lose money very frequently, I lose every battle, but ended up winning wars. So you lose money every day. And the clients would call you, you pick up the phone, you're not even. You're not uncomfortable. Because I've lost 10 times more, at least 10 times, up to 50 times more, the share of my net worth that day than he did from the loss. So I wouldn't have to. So it's liberating on both sides. You don't feel guilt if you have skin in a game. If you are sharing the losses, you don't have any guilt. When you don't have that, then people start having these conflicts.
- [44:36]
A
Right? So the investor sleeps well at night knowing that the manager is sleeping with the same portfolio as he is. And you don't have the guilt. It's good.
- [44:45]
B
Exactly. And the manager doesn't have the guilt because he has 50 times the relative exposure. You have half your net worth in your fund. That's the rule of thumb today. Half your net worth in a fund. And these people are very diversified across funds. I mean, the clients.
- [45:02]
A
So I want to get to a paper you wrote with Philip Tetlock on prediction, because I think it's full of some insights and they tie back into these ideas that we've been talking about. You suggested we talk about this as well. So in that paper you talk about a distinction between binary variables and vanilla events. Binary events and vanilla events explain what the difference is.
- [45:25]
B
Okay, if I tell you that the stock market is going up or stock market is going down and tomorrow I'd be right or wrong. That's a binary event. There'll be war or there won't be war. This is a binary event. Most predictions are binary, in a sense, binary. It can take a value of 0 or 1 or so it cannot event can happen or not happen. In real life, we have a third dimension, the depth of the event. So, for example, you can predict the market is going to go down, but it can go down 20% or a tenth of a percent. You see, the problem is that people who forecast are judged on how frequently they're right, when in fact how frequently they write doesn't matter. It's the cumulative that matters, you see. So there's that dimension that makes a huge difference. And of course there are a lot of pathologies in food by randomness. I already identified the problem because one day I was asked by someone, hey, what do you think the market is going up or down? I told them, of course it's going to go up. They say, well, with a high probability. And then the person saw my trading account and realized that I was exposed to the market going down big time.
- [46:38]
A
You were betting when you said exposure?
- [46:40]
B
Yeah, yeah, my exposure, what I call exposure. As you know, it's not bets in the sense of bets are binary. So people may mistake them for something, for binary proposition. My exposure. In other words, I had options that paid off big time. So market went down. So he said, what's going on here? So yeah, but you told me the market is going up. I said, yeah, I think the probability of the market going up is very high. But should it go down, it should go down a lot. You see, in the real world is about expectation, not probability. In other words, probability times how much it goes down, you see. So the rational thing is to be short the market, but it's more likely to go up because if it goes up, it goes up small. If it goes down, it'll go down big. He was involved in trading. He said, oh, you're right, but I didn't think about it this way. Well, you'd better think about it this way because that's how the world works. So when, for example, people, when we're talking about the first war, some people predicted that the first war would happen, but they imagined it would last two weeks.
- [47:45]
A
The World War. World War I.
- [47:47]
B
The First World War. Exactly. So when this event in the fat tailed domain, an event is not defined. So you can't say up or down. You have to say how much up and how much down.
- [47:57]
A
It's very deep. Because I mean, I think I had the same thought about the Civil War. The Civil War I think people thought was going to last two weeks. They knew a war was probably going to happen. If they had a little bit of imagination and thought about the possibility it would last five years and kill billions of people, they might have had some second thoughts.
- [48:14]
B
Exactly. So what happened is that the event, war can kill 5 people or can kill 5 billion people. So in Fat tail domains. The probability matters much less than a payoff if it matters. Which is why we. So when we talk about prediction markets, people think that prediction markets hedge you, they don't hedge you. A prediction market is a binary and an exposure is open ended and life.
- [48:41]
A
Is open ended and life is open ended.
- [48:43]
B
So natural variables are what I call vanilla. And I wrote a book on the mathematical difference 20 years ago. It published 18 years ago called Dynamic Hedging Vanilla and Exotic Options. And I used the word vanilla at the time because I was a trader. That was my language. I was not a scientist yet or I was not involved in science. So when I wrote that paper with Phil, Phil is involved with prediction. And I said, you know, the guy can have a perfect track record. Doesn't mean anything if the payoff is small in a binary space. Things are very different from where they are in a natural or vanilla space. So. And we called of course that natural variable vanilla. And it did effectively change. This paper was great because we sent it to an agency that's involved in forecasting and immediately they realized, hey, you know what? Forecasting is not done properly. You see, we should not ask people if this event will happen. We should have buckets. What's the probability of having earthquake of 3 on the Richter scale or 4 on the Richter scale or 1? You see, you should have different buckets for different sizes of exposure or sizes of the event.
- [50:03]
A
Yeah, obviously it's like you say, if the odds of an earthquake tomorrow. I'm right now in the Bay Area. I'm in, I'm on the campus of Stanford. So the odds of an earthquake tomorrow are zero, close to zero. Very unlikely, right? It's very unlikely. So I can. Let's make a prediction. I'm going to predict no earthquake. If there is an earthquake, what really matters is whether it's a little murmur or whether it's giant. That's what matters.
- [50:27]
B
Exactly, exactly. So like when you look at economic variables, they're fat tailed, meaning the probability doesn't matter much. It's the event associated with the probability that matters more. So you can have a very small probability of a large event. Now there are some pathologies. People didn't understand it when they created prediction markets. I was then opposed to the idea of prediction markets. People didn't understand me. So I went to Phil and said, listen, let's write this paper so we can have grounds to discuss a prediction and separate predictions between binary and vanilla predictions or binary or natural ecological. I call it also prediction. So when it comes Back to skin in the game. If you have skin in the game, it's never binary, you see.
- [51:19]
A
Because what matters is the amount.
- [51:20]
B
What matters is the amount if you're based. If you don't have skin, the game you're going to gain the reputation system. And the reputation system is based on binaries and perception of others is more binary. So this is where we connect the two papers.
- [51:41]
A
I guess you could turn skin in the game into binary with execution in some sense, right? I mean, the reason that it doesn't work very well is that without skin in the game is that if you say, well, if you lose money I'm going to slap you on the wrist and if you win, make money, I'm going to give you an ice cream cone. That's not a very good system. Because it doesn't.
- [52:00]
B
No, exactly. Because if you then you make 5,000 ice cream cones you can set up, if you buy an option that has very, very, very small probability of being under money, you see, then you slap on the wrist once, you see.
- [52:13]
A
Yeah.
- [52:14]
B
It reminds me of your incident where the event happened. It kills all the people who ever took that roller coaster, you see. So this is where we have to be very careful on dealers. Fat tails. The probability in the fat tailed domain, if I fat in the tails of a distribution, the probability of a given event drops, you see, but then the magnitude of the event increases. So I've been very, very annoyed at the interpretation of black swans. People thought that I was saying that black swans are more frequent. My point, no, they're not more frequent, they're less frequent, but when they happen they're deeper. So they have a lower probability than you think. There's a deeper effect which is what makes them vicious.
- [52:55]
A
A dangerous, evil, crazy man with a pistol can do some damage. A dangerous, crazy, evil man with a nuclear bomb is not just dangerous.
- [53:04]
B
Exactly, exactly. The problem is that the nuclear bomb story, you don't have enough of a record to figure out what's the probability of having such a fellow and you never have an idea about the danger till it's too late. That's a problem.
- [53:26]
A
Yes.
- [53:27]
B
Let me tell you why it's important to talk about binary in the context of skin and game. Because people who talk without harm, who talk about events without harm with impunity, write papers and people are not active, don't really understand the real texture of the reality. When you have skin in a game, you're much more, in a way scientific about things, you see. You're much more rigorous about things, you're much more rigorous about risk. And there's a class of people who are trying to promote the idea that people that small probabilities are overpriced by the system. And they find experiments that in fact are binary or by nature or very well is bounded and say, well, people tend to overpay for lottery tickets or overpay for these artificial setups.
- [54:16]
A
They're overly cautious about events that are unlikely.
- [54:19]
B
People are cautious if people that pathology doesn't extend to options, financial options, for example. People don't understand the following is that if you have an option and there is a stock market crash as the one we had in 1987, and the option can cover 30, 40, 50, 100 years of P&L, in some cases 300, USB and L, you see, because it's a very small, tiny payoff. So true. You know, we may not have an idea of that probability, but we definitely shouldn't make statements about things that have a very, very, very, very, very strong payoff like that. You see? And people naturally in natural setting tend to be careful about it. We overpay for insurance, we overpay for lottery tickets. So there's something called the long shot bias. And people think it's a pathology where I think that it's only pathology in things that are modern, you see. And that pathology doesn't extend to vanilla variables.
- [55:30]
A
Let me see if I understand what you're saying. I think you're saying that a lot of laboratory experiments that psychologists and some economists do, people overestimate the probability of very unlikely events. They over the so called overreact. So they'll, they worry a lot about a plane crashing and in fact the odds of a plane's very safe compared to a car. But they overreact. But what you're really saying is that they don't overreact because the costs of a plane crash are very bad.
- [55:57]
B
No, no, no, no, no. For plane crash maybe you may overreact. It's not a big deal. The point is that you cannot generalize from an experiment that is not natural to natural settings. And we identified exactly what the mistake people make. Because you're making a statement derived from a variable that has a bounded payoff and generalizing to things that have open ended payoff.
- [56:22]
A
Yeah, my example, the airplane's a bad example.
- [56:24]
B
Yeah, it's a bad example. And let me tell you say people overplay for, you know, people overpay for lottery tickets, for example, which is true. They overpay for lottery tickets, meaning the.
- [56:36]
A
Expected value is negative.
- [56:38]
B
Exactly. Hence, if people overpay for financial options. Okay, hence let's sell remote probabilities in finance. Well, anybody who has a brain would realize that banks are engaged in the businesses selling small probabilities in finance. And they lost $5 trillion in 2008, which means more money than they ever made in the history of banking. So therefore that statement that people overpay for protection and finance is false.
- [57:12]
A
When you said the number 5 trillion, for a minute I thought you just meant a lot of money, like a zillion. But it actually is close to 5 trillion.
- [57:19]
B
Right? 5 trillion. Yeah. 5 trillion is. I mean that's what they lost before governments bail them out. But $5 trillion is a lot of money. So what I'm saying is that instead of doing experiments, just look at the variables themselves. You see, insurance companies haven't really made money until recently, you see, and then all it takes is sometimes one event to. And insurance companies aren't involved in very complex things, typically except for reinsurance. So look at the data banks. Bets on small probability events by financial firms have proven disastrous in history. And wealth came from bets on open ended remote, probably events, namely entrepreneurship.
- [58:08]
A
So I want to. That's beautiful. I want to close with one last topic. We're going over time here, but I just wanted something that I thought about when I was reading your paper, which is parenting. It seems to me that when our children are younger, we don't want them to have skin in the game, literally. We don't let them get near the stove that's hot because they'll burn their hand. And as they get older, good parenting, it seems to me, which is hard to do, means letting our children have their own skin in the game rather than the skin of the parent. Think that's right?
- [58:41]
B
I think that's right. I think that traditional parenting has some merits in a sense that you protect. There's an expression in Lebanon that the first seven years you play with them, the second seven years you let them get in trouble. And the third seven years you advise them on how he got in trouble.
- [59:04]
A
You what? What's the last one?
- [59:05]
B
You advise them on how he got in trouble.
- [59:07]
A
Oh, you explain it to them.
- [59:08]
B
Yeah, you explain it to them exactly. And then they're 21, you see, that's a Lebanese expression. The first seven years you protect them, you see, because they're fragile. The second seven years are sort of anti fragile. So you need to get in trouble because they'll never learn unless they have skin. The game now, before finishing, I'd like to talk about something. We're talking about skin the game, which is quite, you know, relevant here, and it's to talk about academia, because we're both academics. No, you cannot. Yeah, sorry, we're academics, but. So the only way you can reform economics is by installing some kind of skin again mechanism, because it looks like now the system on its own allows them to be wrong with total immunity.
- [60:05]
A
Agreed. But doesn't that come back to the problem we talked about in the first few minutes about the chair of the Fed? And if you're an economist, one group of economists says in 2008, we need to spend $2 trillion by the government and it doesn't matter what we spend it on. The other group says, no, we should spend zero. Then we spent about a trillion. We spent 800 billion, roughly. And things didn't come out so well, but it's very possible. That's not proof that they were wrong. The people who said spend a lot of money. It could be a thousand reasons. And I think we just have to accept the fact that economists can't have skin in the game and therefore we should discount what they say.
- [60:47]
B
Exactly. The point is we need to lower the dependence on people who don't have Skinny Game.
- [60:56]
A
Yeah. Or ignore them.
- [60:57]
B
That's the problem. Sorry.
- [60:58]
A
Or ignore them.
- [61:00]
B
But you cannot ignore them. You have to build a system because people can take over the prestige. There are a lot of things, a lot of pathologies we can't control that way. The best way to do it is build a society in which mistakes made by economists stay on campus. Exactly. That's the idea. The idea is, if Larry Summers wants to make mistakes, more mistakes, let them make them at Harvard, where we're insulated from it. Okay. Sort of like the ivory tower. Not because they're, you know, it's because we're protected from them, not because they protect themselves from us. We should work both ways. You see, they don't want the real world. We don't. Sorry.
- [61:39]
A
It's a great slogan. What happens on campus stays on campus.
- [61:42]
B
That's exactly it. So keep the mistakes local on campus, and that way everybody will be happy.
- [61:51]
A
My guest today has been Nassim Taleb. Nassim, thanks for being part of Econ Talk.
- [61:55]
B
Thank you very much. Thanks for inviting me again.
- [62:03]
A
This is Econ Talk, part of the Library of Economics and Liberty. For more Econ Talk, go to econtalk.org where you can also comment on today's podcast and find links and readings related to today's conversation. The sound engineer for Econ Talk is Rich Goyet. I'm your host, Russ Roberts. Thanks for listening. Talk to you on Monday. Sa.