Nassim Taleb - "The Black Swan"

ah hello we're going to get started now thank you for coming in as these days get more and more beautiful I'm appreciative that everyone still comes to our programming but you couldn't miss today's right so today's talk is with Nassim Nicholas Taleb he's a wall street tree he was a Wall Street trader for 20 years as many of you know who traded in pun intended his trading to become an academic and has since been writing about the work that he's done and bringing together the research and practice from his two points of view today he'll discuss his best-selling book that many of you probably know about the Black Swan it's on sale outside and will have a book signing after this and he's also going to talk about a new book anti fragility the nests necessity of disorder that is supposed to be published this fall and the title may change because he took a call from someone in my office and they were discussing the title so anyway be beyond the lookout for that perhaps they'll come back and do a signing for that as well published in 2007 the Black Swan focuses on the extreme impact of certain kinds of rare and unpredictable events and humans tendency to find some simplistic explanations for these events retrospectively um this sort of reminds me of when you know the kids are little and say why is the grass green and after you've exhausted every sort of scientific explanation you just say because it is so hopefully his uh has I think his work is more profound than that the book has been crediting credited with predicting the banking and economic crisis of 2008 and I'm sure he will go into that in more detail and a review of the book by The Sunday Times called it one of the most the twelve most influential books since World War two so that is quite an honor and we're very honored to have him here today to speak so after he speaks we'll do our usual Q&A and as always will will preference students so students please think of your questions come down to the mics when you're ready to ask them and we'll have a good talk Thanks either thank you thanks for the introduction hopefully I figure out how this works let me see if it works does it work it does work now okay so too bad I was hoping for a little bit of randomness in this and to define what is fragile two random events I like the idea to illustrate the fragility of technology because we have a trade-off in life you have gains and efficiency that always sort of come that almost always come with losses somewhere in other words an increase in fragility so I'm going to talk about that but first let me ask you what is the opposite of fragile who's who say robust all right okay student holding my glasses see how many people are gonna say robust what else durable all right okay what else surely all right some of the words okay so that's a common mistake because let's look at this package right this package is fragile you're shipping this item to the North Pole where you know all time before the wedding all right if you if the package is fragile you write handles here on it so a fragile package has four upper box we define fragility for upper bound okay unharmed agree if it were robust what would you write on the package so you're shipping now a different kind of wedding gift to the North Pole right why would you write on it on a robust package nothing you would not write I don't care you know you want all right so once we're going to now have a definition of fragility once we write it down mathematically but realize that the opposite of fragile is definitely not robust okay it is something it's opposite to be a package on which we would write please miss handle okay because fragile is what is harmed by this or by disorder whatever by you know mishandling and the anti-fur the opposite would be what is what gains from disorder gains from mishandling gains from okay so there's no word for it and I've tried I'm sure someone's going to come up to me and you know during a wine and and because we have foreign students here and explained that in his native language there is a word for anti fragile there's none okay which I've tried all languages all right not all languages like 68 nineties right so there's no word in for some things that's the opposite of fragile you know in common vocabulary in actions we are aware of that property of course but we have a lot of saying but nmath ology will see they are mythological equivalent but there's no word in common language and it is a problem why because you know I was driven you know largely driven to find a common word for we'll call antifragility but technically I call it a final stochastic loves volatility love stressors love randomness after reading a book on perception of colors that explained that Homer the not have during the days Homer there was no word for blue people were not biologically colorblind they were culturally colorblind and even a lot of population don't have a word for blue but they can match colors properly okay so we are aware of that intuitively but not culturally of that property of antifragility or is gonna see ramification can be quite severe the let me show well the representation and time series okay it's fragile what has small gains okay and large losses so the upper bound is what for a package is zero it's now going to improve at all but it improves moves a little bit that is a fragile you can see it from financial packages that is a robust the robust upper bound on harm lower bound on harm or close to you know on harm and of course the anti fragile is something that gains that has much more upside than the outside now let's go to mythology you know if you recall the first slide can someone recognize the animal hydra hydra in greek mythology you cut one head to grow back so the last thing you need is harm it the robust what would the robust being weak mythology phoenix phoenix is your killer that comes back that's a robust so the let's get deeper into this but let me tell you what i'm going to take you okay i started defining and looking at the world there's no it's okay I'll manage looking at the world in three categories the fragile the robust and the anti fragile the anti fragile is well gains from disorder but you think that if it gains from disorder it's going to be hungry for disorder and if you don't give it this order it's going to suffer okay the robust of course doesn't need the solo doesn't care and the fragile is what breaks under stress so you can have a very simple mapping of things about the hundred categories thanks 100 domains and economics that makes you fragile robust equity is robust for the structure and of course venture capital has this equal log file of stochastic attribute and as with benefits from sharks and uncertainty so I built this triad I call the Triad as a replacement as a replacement to a two-for casting call this is a very non predictive method and let's look at the problem forecasting the central problem and the Black Swan was that you don't the problem that you cannot forecast these events mathematically it is impossible to be precise about small probabilities and perfectly is almost impossible okay it has been a part I mean has been very very rare for people to predict events so you could look at it in records particularly in what I call a fat tail domain but mathematically it's very very hard they are it's what we call the problem of induction and and it applies with full force to rare events you don't have a large sample if you're basing yourself on empirical methods you don't have a large sample ok hundred-year floods happen once every hundred years so if you have 50 years of data you're not going to see it okay and so rare events are not predictable and Berkeley there's a BIA you need a log lot of data for a lot of things we have we don't have enough data and that's assuming you believe that it passed is relevant and mathematically small perturbation we're going to look I'm going to use a very same technique that I will use to define fracture to show you why they're very fragile these small probably a very fragile to model error so that's so let's set it forecasting measuring plus people use in the English language and I'm sure Danny Kahneman would confirm that the use of certain words people start you know imagining that they can measure the value of the future they can forecast they can measure risk simply because you you provide them with a certain framework okay if you tell people instead of use the word measure risk or measure probability of rare events if you tell them that you're trying to your prophesizing for example there would be a lot more careful so there are a lot of psychological issues you know involved in the words we use for in Semitic languages to predict means to is you know you the way you say it in Arabic for example is to prophesize so it's very careful about prediction if you use Arabic but if you're using Wall Street lingo and say I'm measuring the probability of some event you have the illusion of measuring something that doesn't exist okay you have the solution measuring it the way you would measure this table okay so this is a problem prediction and the Triad as a replacement to prediction I don't have to measure a future event I can and we're going to see now there's a technique to figure out fragility to measure fragility it's vastly more measurable with much much smaller error rate then then the probability future events so let me figure out where I am here in these slides and let's get there all right the let me show you a few things for example you know which item here is more fragile now very simple you see gains you're reckless some of you may not recognize one of the items the one on my right all right okay some of you in school all right the the the one on the left visibly is vastly more fragile I have files computer files you know my first you know popular book or non-technical book fooled by randomness I have the files on its get 12 years ago I can't retrieve them okay this book is 505 years old the one on the right is you can read a I don't know you can read the newspaper you can read this so the one on right so you see a gains in technology come with increases and fragility I use a certain method the way III I mean my statements about the crisis were very simple if something is I can't predict what will happen in the future but if someone is using predictive methods and building fragile items based on these predictive methods these items are going to break so I can predict who's going to blow up based on the reliance on prediction banks for example they they they predicted a lot when you have debt you need to be predicted that map's one-to-one with overconfidence about the future small mistakes what you're out so for someone to have a lot of debt it means that you think that I'm going to have make mistakes look at government forecasting for any economic number okay all prices today should have been twenty two dollars a barrel according to the Office of Management and Budget and we trust them now with a forecast of what's going to happen in 2012 2013 2014 okay so the people when they are very predicted they build something very fragile when they're not relying on prediction they have redundancies in the system that makes them a lot more robust and sometimes anti fragile so I I was asked by the economist to predict what was going to happen next 25 years thinking that I was they asked me thinking I was going to overreact but so you can't predict I filled out the she alright I said what are you the author the Black Swan or some imposter or did something happened to you right and I don't know of course I I can is very simple most people predict the future by taking the present as baseline and adding flying you know flying motorcycles all the kind of stuff to the future it's much easier to take the present and and project the future very moving I don't know what what the future will add I know what was future what time will remove okay anything fragile is likely to go bust okay and and effectively the technique works if you take what is it you know I'm going to show you how this technique works alright you recognize these items okay this is a kitchen a pompeii kitchen right glass three thousand-year-old things that are very old have staying power simply because I'm not fragile technology is very fragile technology comes and goes the future will be technological but not with the current technologies and effectively the very simple application of something called the Lindy effect is you take if a technology has hundred years old you can expect it to last an extra hundred years time is a very good detector of fragility if a book has been in print for 15 years you can expect it to be printed extra 15 years in the absence of other information it's just very statistical just as you would predict someone like a pregnancy from his age and no other information and when you see an old person and young person you can expect the young person to outlive the old person for humans anything non-perishable the young will outlive the old anything perishable I mean we humans happen to be personable not our genes but we are perishable but for items that are not perishable like cultural items things that don't decay the opposite holds and for everyday technology lives is likely to live an extra day and you see it in the data you see then again that's the the application of fragility so let me continue with what the things this is the economy of things that are equivalent when I talk about fragility and in a few minutes I will define fragility and we're going to see all maps together what the anti-federal gains from these and the fragile losses from these so the equivalence is uncertainty variability imperfect and incomplete knowledge chance chaos volatility disorder entropy time the unknown randomness turmoil stressors and error if this working when I stand here I don't have to get close to the microphone No okay right through these maybe taught at Princeton in different departments in different buildings okay but they have the same phenomenology although in theory we don't know why these have the same effect what happens when something is anti fragile is craving randomness and you choke it of randomness you don't give it randomness you see what's up top this is a natural system it you know it's always has randomness if you do at the bottom what Alan Greenspan did here try to eliminate the business cycle by stabilizing artificially the system you eliminate randomness and system craze randomness you end up what we have here is you had not have fluctuation but then the busts are going to be horrible people who are familiar with forest fires would see an equivalence there if distinguish every small fire the thing collapses okay well we all we are aware of it in our own domains if you put a child in a sterilized environment for fifteen and a half years no germs right and then he would be very healthy no and then put them in a New York subway what would happen you see what would happen all right so the private system that things that are anti fragile of stressors or harm or whatever you want makes them weaker and this is a common mistake that comes from not understanding antifragility and incidentally that phone call we have decided on the subtitle of the book the necessity of disorder okay the things that you like disorder bailing out companies is not a good thing because what you're doing is delaying a larger collapse later said letting things the natural way fail fast early and at minimum harm to others so let me find now I'm going to we have a very easy way to string all this together with a way with a technique to detect fragility and antifragility okay and that explains why what is fragile doesn't like volatility what is anti fragile loves volatility and why we can Elson option trade for 20 years why we can use all these things derivatives are bad things we've been bad for society but at least here they were useful for something okay to show who is short volatility and who's long volatility so in a Midrash tehilim in sort of the Talmudic not quite a boutique literature there is a story of a king who has to punish his son by crushing him with this big stone but of course the king doesn't have any intention of killing his son so he had a quandary and the quantity was solved by an advisor and what did the advisor recommend what do you think he did exactly small pebbles you break up the stone in the small pebbles and you Pelt the king's son with the pebbles this is a time-consuming you know every takes hours and everybody's happy alright the punishment went as you know dictated by the law and and the king's son was safe right and except that someone had to spend a lot of time cutting stones any so from here we have a very simple lesson which is as follows for the fragile very simple you have a non-linearity this is the horn size of stone event size size of stone and a horn the harm is nonlinear every if you look at the size of the stone harms the the son of the king all right and then none disproportionately non linear in non linear way in other words a hundred pounds will be more than a hundred times as harmful as one pound okay we have a definition of fragility if I mean you can try the experiment if you want take your car and drive it against the wall a hundred times on one miles per hour or okay or do it a thousand times a tenth of a mile per hour and or and then compare to compare try it once at a hundred miles per hour okay so you see what I mean here is that you have an increase in speed and increase in something so you have non-linearity fragility come from non-linearity with the proportionately larger impact so this is exactly like what we used to call the short gamma as option trader is exactly what we call short volatility it has a negative second derivative well with this we could do a lot why is it that everything fragile has to be nonlinear well very simple statistically you have a lot more events of mild impact than events of larger impact if you were linearly fragile to jumping or a tutu - you know shots by jumping you know from some Heights then jumping from 10 meters would be you know as harmful as jumping ten thousand times one millimeter I don't know if I got it right this is thousand times one millimeter Telstar 1 centimeter who cares right ok if it were otherwise just walking to the office would kill you say so harm has to come from the nonlinear fragility is in a non linear and I think fragile to the linear would have been harmed from cumulative shocks like a coffee cup gets tons of sharks over its life that one single I don't know if they want to experiment with here but one single hit you know of some intensity breaks it but a lot of small hits don't cumulatively harm it so with this we can do a lot the fragile is what has this curve you know a very simple lesson concave and convex we'll see how the it's very easy to figure out the concave and convex using this graph ok the concave is the one this is a concave concave is no good convex is good concave it means that disproportionately larger harm convex is proportionately larger benefits and we can use this to map anything the rule is as follows I have a variable I wish I had a no maybe see here and groan no it's okay the the you see the arrow you are there if that the variable moves up by a certain amount you harm more then you would be then you better gain if it moves down by certain amount you are concave and it's fragile its market goes up 10% all right you make a million the market goes down 10% you lose 10 million you are concave you say therefore you are fragile very simple so upside versus downside if you're upside bigger than a downside your anti fragile with respect with source of randomness and vice versa with this we can figure out number one why we had a crisis visibly the system was driven to a nonlinear cusp but we can the best comparison would be to traffic you know if you know what New York City is Manhattan okay all right say you have 90,000 cards in New York travel time half an hour between point A and B point B you add 10,000 cars travel time increases but very little you add 10,000 more cars and then you hear honks and you know the travel time now that doubles so nonlinearities is the reason why traffic has been a problem I was at just I was last week in Sao Paulo coming in from the airport three hours going back it was Good Friday you know at 5:00 a.m. 28 minutes all right so you realized that they are they have too many cars you realize were you're fragile to a number of cars well that's the case was all over up to my systems have you heard of the airport called Heathrow and London okay this was built by someone who forecast number of traffic not realizing that there were very concave to the number of flights and it's so over optimized that the smallest disruption causes 8 10 15 hour delays was people sitting on the floor playing guitars you know it's kind of thing people young people do in the summer when you have you know airport delays so with this we can detect very simple you can detect fragility you can detect model error where is I take any model increase the parameter 1% decrease the parameter of 1% and see if I have more losses than gains and fragile if I have more gains than losses I'm anti fragile from anti fragile in long run I'm going to be ok gain effectively from disorder if I'm fragile and the long run I'm going to be harmed and act don't mope beyond harm I will no longer be around ok part of the gene pool whatever it is so another example let me give you another example what this is too technical I don't know how technical the crowd is that I can get technical with my graph someone asked a question let's look at why sighs let's look at why sighs every lot of people in economics talk about things called econo economies of scale all right you hear about it it exists everywhere except in the data companies get large and they're not more efficient and effectively they attempt to die alright so there is a problem why what because as you get larger you may be more efficient because you have higher other savings and you have four secretaries your this fear that okay but if there's a squeeze ah you're harmed a lot more this explains sort of why we don't have very large animals an elephant is vastly more efficient than a mouse metabolically but we don't have a lot of elephants we love animals of that size on on earth yeah that I've seen right on on on land so let's take the stalled experiment have you heard one there's a story with French to see theory at all have you heard of it I know it's a French man who guys don't care but this is very interesting story soceity journey I had a rogue trader this fellow he was national hero in France he may actually run for president one day right so his national hero his name is Carol Jerome Kerviel very nice fellow DUP where is this guy okay Caravelle it was discovered that Kavya was hiding fifty billion euros plus a lot of money even in today was a euro where it is still a lot of money fifty billion euros like sixty five seventy billion dollars was hiding was you know in the drawer he had bought stocks and they discovered that right it was not very pleasant thing and they decided to liquidate given that the the was in France they thought they didn't weren't aware of the existence of Martin Luther King or at least that he had a holiday right that New York that Monday was closed so they tried to liquidate the fifty billion dollars of course the French way which means twelve people dictating to one poor guy you know one person liquidating and twelve people telling what to do all right but whatever it was they lost they ended up losing four billion dollars four billion euros from the event from liquidation okay so we can play the thought experiment following six all experiment is if instead of having one large bank you had ten small banks each one would have its mini kvl they would still look like him just as massive so it would be pleasant this is technology so instead of having one large Bank you had ten small banks and they would all have the same probability of having a rogue trader so the rogue trader would have 1/10 of the size okay we have ten times such event look at liquid the net effect would be zero losses why because if you're to liquidate and other traders here liquidating 50 billion euros even on a regular day and even if you did it the right way would cost a lot of money these are liquidation costs very non-linear liquidating five billion euros all you need is two traders and four cups of coffee and it's only sip your coffee and your liquidate progressively and when the coffee is finished it should be finished liquidating and guess what zero or minimal costs so you realize that you have this economies of scale stochastic this economies of scale coming from size that when when a large event hits you don't if you I don't know if you have pets as someone here is a lot of people come from countries that have elephants if you have an elephant as a pet okay thing will be okay on a regular day but if you have a shortage of water you have to pay up a lot more per liter of water that if you had a mouse no okay so so during squeezes the costs go up non non proportionally this is one way to figure out it's the same equation we used earlier okay you have nonlinearities very simple once you detect this concavity this concave non-linearity you know you're in trouble but we can apply this to a lot of things that looks like I don't have Alain Caesar with the watch okay so looks like I have an extra five minutes now we're going to move to a big problem the problem we have today in society that you have two classes of people you have the class of people making the upside but having no doubt side from the risks and a class of people who have the downside for the risk of others but no upside from it agree this is we saw from the crisis and this is continuing at no time in history have we had so much what I call a transfers of anti fragility and let me explain anti fragile means you benefit from volatility you have the upside you benefit from volatility if you look at the stock market the stock market or when I calculated about a few months ago had lost over three trillion dollars because over the past 12 years okay because you know company's been issuing at all prices okay higher than here and lower than here as well but mostly higher than here with issuing stocks and retirees and investors lost about three close to three trillion dollars cumulatively farm managers are paid but only when they do well agree it's called the Centers and people have the illusion that such incentives it's part of the capitalistic system okay so you would say that okay since the stock market lost by three trillion let you know after financing that you'd have no incentive now how much were incentives given to not fund managers who company managers negative 400 million no how much 400 billion dollars positive why because it's very simple if we split this room in two we have managers in this room half of you go along the market and half of you go short the market okay and you're paid an incentive at your end half of you will make money those will make money will get a bonus and those who lose money what will they get nothing there's no malice right okay well that's counter to everything since what we call Western civilization has started and let's go to the beginning of Western civilization does anyone recognize gentlemen here sorry Hammurabi all right okay Hammurabi's law go visit the louver and you see the law and I'm sure this year this law had to precede this concept Hammurabi is as follows if the architect builds a house and the house collapse and kills the owner of the house the architect shall be put to death now or not you know recommending such policy now not a little piece before all right but think about it what is the logic behind Hammurabi's law what do you think sorry what the real thing is now if the architect can hide risk in a bit better than anyone else you see in a thing so it's the best and the architect will know more about the risk than any regulator that's central no inspector will ever be able to know what's going on in the foundations of the house better than the architect themself so it gives an incentive to the architect it's the best risk management rule ever scoped skinnin again the best risk measurement rule where we lost that the limited liability company was not what Adam Smith's had in mind and effectively you can see that wealth has not been generated by companies after you know the the owner is no longer the owner but before I don't miss you know figured out they can have this as an agency problem there's moral hazard agency problem and let's see how it plays in finance when I'm short an option at the expense of someone else sorry I'm rushing like this okay very simple let's say that these are all profits okay the first few months are all profits and the rest is a big loss you make make make make make make make and every year you make money you have a bonus that's what bankers had and when there is a loss as what happened in 2008 who eats it sorry shareholders first and then who else you taxpayers Bob Rubin Robert Rubin made 120 million dollars and bonuses from citibank citibank lost more than they ever made you know they were going to be in receivership I know from the inside he kept his home 20 million dollars now you guys are paying retros you know actively for it okay and and these risk or hidden risks bets on local ability events detectable to some anyone who knows about fragility can figure out if something is fragile and someone's making these bets and actually there was an incentive to have these hidden risks particularly if you don't have what I call skin in the game so we have moved from an environment in which skin in the game was the only risk management to an environment where technology is a risk management it doesn't work you have to you cannot be and then I have my rule as you cannot you shall not be long an option hit an option that's anyone else's expense but you know we have that effectively we're moving even into a new kind of world vastly worse at no time in history that we have people who had a high rank in society and did not take personal risks right it's not like the end war was not a video game like in the short time you bushes days all right usually you can generalize this outside finance to everything peak you have declare war first and battle commander in chief pre technology pre modernist so we have this loss of skin in the game is very generalized and the crisis happened I just wrote a paper I wrote two paper on this topic the first one called the Black Swan of Cairo in which I explained how you know stifling volatility in these countries like Egypt okay will cause blow ups and Saudi Arabia now of course were contributing and the second one is on the first the title was wider the crisis happened and that I changed it more to be more prescriptive how to avoid another crisis by saying the problem we had is a combination of scientism bogus scientism and absence of scandal game on the part of all operators now it prefer you we've been teaching methods in finance called portfolio theory how many of you have studied these particular methods okay you all know they don't work no okay they keep teaching them and they keep teaching them and there's the biggest intellectual fraud you can imagine right because these portfolio construction don't work you're endangering people there's of course no FDA but why do people keep teaching them because a professor has the upside of giving you a nice equation and doesn't have any other downside if you blow up very simple okay so the idea to generalize skin in the game you have john ruskin again alright so if you had a legal system via other techniques and and I like what I call heuristics and simple solutions to these classes of problems I think I'm pretty much done but let me finish with a picture of Hammurabi and a lot of things will come up in questions here in the Q&A let me put Tom Wahby all right and finish with Hammurabi and so incidentally I want to mention one thing before I finish since we have some it's the same thing with the electronics when you're very ill you are like longer an option medicine because the harm of the medicine is not going to harm you much let me give you a very simple example now something called numbers needed to treat and empty my initials if you have very high blood pressure and you take medicine the numbers needed to treat are two out of three means out of three patients who take the medicine to will will be will benefit from it but if you have a borderline high blood pressure one in 24 patients will benefit from the drug and of course the the pharma company will always benefit from the drug so the numbers needed to treat so it's very nonlinear and if you rationed drugs to those who are very ill but these people given even more than they want okay you you would do both a favor the one who is very ill and the one who is ill because what I call the enteric genex is harm done by the healer it is exactly your your fragile if you're slightly ill and using a drug and you are like long volatility anti fragile sorry if you're not very ill your fragile alright we give lipitor to people not knowing the side effects so this has this probability distribution I don't know how technical you guys want to be but this is the stick week of looking at convexity effects and s symmetries applies directly to medical treatments particularly if you look at it at a macro scale how much we spend and we have data actually if you have data for occurrences of hospitals that went on strike and focused only on eliminated elective surgery and of course people have longer I don't know whether it is longer because it eliminated elective surgery or because maybe they did better job with emergency you know treatments where both were needed anyway so thank you very much for listening to me and let me put back Hammurabi and their questions please sit down hi my name is Cristian I'm an undergraduate studying financial engineering here at Princeton and your book The Black Swan was one of the principal things that convinced me to study probability so I'm just wondering you know as an undergraduate here how should I supplement my education - to use your terminology avoid becoming a nerd okay the probability and what we call someone focused on a very narrow domain they don't go together very well because probability is very fragile - estimation errors so two things either avoid studying probability or or if you study it don't study the probability mathematically because it looks like for what we use it for in social science it's just an attainment nothing else and it's actually harmful it has yet regenexx it is harmful so if you consider insurance and generalize kind of that something that promotes fragility isn't there also an increasing non-linearity when you take risks so say you invested you're going with the Google X Lunar Prize and you're trying to invest in rockets or going to the moon if you only invest $10 in that then you're not going to learn very much about the can you speak slowly oh yeah so if here so it also seems on the opposite end that there's a non-linearity in the investments that you make in the risk that you take so if you're only risking a small amount of money add such a grand prize then you're not going to get very much out of it but if you're investing ten times as much then you may know that's not the way to look at it because when I look at the portfolio you look at your portfolio as a whole right so you got to split it into different investments you have $100 are you going to split how many investment you're going to add okay in a fat-tailed world a world that has a lot of larger deviations you should have as many small investments as you can for these hundred dollars in a world that doesn't have fat tails you do what the Markowitz portfolio theory then you would invest in a small number of investment for which you think you're comfortable you know about the return so it's a very different methodology but Don linearity is not the point being being a huge fan of your ather isms so one of them one of my favorites being a good listener is someone with skillfully polished in indifference I really liked the by the freak hostess but since now you don't have a Twitter account where do you keep that for so what can we find them I thought I mean you buy two stores it's like no I put aphorisms I wrote aphorism just as a literary exercise about the 500 of them over a certain period of day and I thought it's gonna take me five hundred days it took me about 60 days to write five out of them and a publish them as a book yeah that's right but you can buy the book no after after reading the book you still have a new effort I guess I know I don't know I don't I've stopped or if I if I have a personalized I got published I thought they want to get it for free because people like people used to Twitter wants everything for free I have a question about the implications of your worldview with respect to what it means to let's talk about banks for example so Thanks so if we manage everything for the tail risk the belt the Black Swan the implication would seem to me be that if a bank had ten billion in capital it could only really make ten billion in loans because we have to assume that any one loan could go you know to the to zero and we don't want to do bailouts because you say they're bad so we end up with a we eliminate fractional reserve banking and the implications for the the size of the economy are somewhat significant and I wish you would just comment on it very good he's telling me that if someone used my idea of perfect robustness seeking perfect robustness and you would not let banks lend more than their capital right and that would be very bad for the economy let's go to the banks in between number one the the just reducing the leverage of banks would be a good benefit but let's look at what the banks do all right banking is a very strange business it's involved in lending you agree all right so it's learning a good thing or a bad thing who thinks that lending is a good thing okay you think that lending is a good thing they thought that economic growth came from lending Hussein economic growth came from lending okay who think that they know the the innovation and all that came from thanks to lending to people who were involved in these things okay directors look at the history of lending and history banking you'll realize the following Industrial Revolution takes off landing comes when companies no longer need the money okay where and then for the manager and the lending increases dramatically when the manager of the company is no longer the owner and start you know gambling society's money let's take the numbers here number one if you want you have computer laptop computers up there how much debt does Apple have okay there you go now the the I know if you remember what happened in 2004 those of you still who could read the $2,000 long time ago all right the students all right in 2000 we had a collapse in California how much debt were there zero there was no like really all right so the economy growth doesn't come from that except the portion of it we paid back when you pay back the debt okay so it looks like I mean that I'm not talking about talking speculative that you have different forms of debt you have debt for letters of credit debt for now speculative purpose and then of course speculative debt all right or is that lending companies to take some classes of risk in history you go back to this fellow all right the Babylonians discover deals of that that blew up society what happens is you know creates mad habits and of course almost every civilization we've had has ended up either banning or severely curtailing debt whether he tells talk about Islam Catholicism Catholicism had been of debt most of using Islam I don't know I wonder what happened in Ireland right but it was Aquinas penis if you read a penis is the fat was much stronger than Islam okay so and then it's almost also racial figures at the point so the Smith's that we need that to grow you need that to grow cosmetically and it's a very fragile type of growth and and of course that leads to a culture of debt that is not very good okay so this is actually a question now what bank should do if you had equity rather than that the right way this conversation be irrelevant because the banks would not play such a large role what you don't have debt banks don't play a very large role today we have between three and six times the amount of debt to GDP in Western countries as we did in 1980 so any so we realized why the finance sector is very big and the financial sector that compensation scheme another data piece of data when I wrote the Black Swan I wrote that but 1982 money Sun and banks in America lost more money than ever made in history banking okay and one day one one one quarter of course cosmetic you know hiding of it and sure enough I get a lot of hate mail this is the past this will never happen again and of course the hate mail I got my published in 2007 needs to say that five trillion if you were to mark-to-market properly were lost in 2007 2008 year in that period and so if you want more convincing that what debt has done is mostly allow some class of people become rich not the bank investors mostly a certain number of your real estate speculators like Donald Trump you know remember he went bust he didn't he's rich now many honey is bust manufacturers Hanover for those who read financial history so there's something about that that is very strange well it's a big myth about the role of debt and traditionally governments have borrowed to wage war so and I think there's nothing has changed you know now we're paying for largely between 1 and 2.2 trillion for vast war ok so now you understand why I'm not like that and why we should be very careful and this conversation by the size of the banks wouldn't take place in an environment where we have more equity and healthier economic climate more question I guess but don't you agree that the way you're looking at the world that a bank would only be able to lend I am not really I would only be able to lend the amount of equity it has because it says this is fairy tale just ask this question this is taking it to the extreme because you want to prevent banks from lending you don't want the bank that Bank shouldn't exist just go buy a bond you see the idea is to prevent banks from playing a very large role in society okay so lending speculative lending but the balance sheets don't come from things that are healthy but I'm not against fractional reserve banking you have to look at what the loans are made for okay are these are the loans for speculative reasons or the car loans or loans you know that are very controllable and helped at a small-scale economic activity plus the other thing is there's difference between the risk of banks of a similar question like with insurance companies if you have two types of insurance there's insurance that is from my calls mediocre Stan because the risks are correlated and small insurance for example on health insurance and stuff like that and then their reinsurance that are lumpy the banks the problem of the banks is that they're most of the loans are very correlated and then invited beyond a certain level the get engaged into very correlated loan rather interesting Craddock risks that are had shovel and diversifiable no more questions because there's wine oh yeah all right big follow up questions to stake out kind of your position on debt so yeah these are sort of debt skepticism extend to forms of debt which seek to benefit from anti fragile systems for example like Ventura County the first of all venture capital is never debt all right has never been that but I put in a triad here I put that in two places all right okay I could the fragile is debt is fragile relative to equity okay and this is a relative scale okay it's like a spectrum but if you see the second line the second entry public that is a mother of all fragilities public that is very dangerous because public debt is never repaid if you merge that with my idea of size and you will see that the problem was was with government is that was a scale of the government is because of projects you know the the let's apply the survey of size a hundred million dollar project will have overruns that are about four times on balance those of a ten million if your energy scale ten times ten million projects you say and that is the problem was probably that is that it's much less it much more complicated to errors I'm writing something with the IMF trying to look at concavity of government deficits they're very concave because they can't cave in a sense that if unemployment goes up one percent their deficit worsens a lot more than if it improves if unemployment goes down one percent or some other elements that enter their deficit particularly if their feedback on interest rates of their existing that making things worse so government debt is very bad if I were you know to view the world and it's say the world has never been richer today in history the world as a whole and almost every country today is richer not Greece but almost every country today is richer than it was ten years ago and that was even last year David or not okay in the aggregate for country and for some strange reason okay we're like rich people who need to borrow well when you get rich you should usually have less than whole reason tomorrow and we're borrowing more than ever this is a big problem so if very simple heuristic if you want to have a healthy economy number one prevent governments from running deficits make them shoot for surplus so you can get the you know a small deficit the family and the second one decentralization because municipalities are better at forecasting their expenditure than centralized governments yeah could you summarize your year and a half ago economist article in terms of the key things that would be and would not be broken by 2036 you know I mean I I use the methodology I had used that technique to say that the banks would collapse and Fannie Mae would go bust right I had used that technique before that economist article but in it they asked me what would be broken right I told them what is fragile needs to be broken and of course even even with government bailouts you can't be larger than a market you can bail out bail out a lot eventually something will be collapsed I think major banks today will be already broken banks today we'll all go bust the banks that you name that you walk outside you see a name of bank your odds are will be off the outside the gene pool banks all right predicting what will go in the future the Federal Reserve Bank of the United States will not exist on this present form too fragile izing and itself you know you can't you can fool some people for a while we can't fall in forever all right the the bonus without malice will not survive I don't know if I said that in the article what will survive I mean just using that Linde effect as a heuristic the bicycle will survive you know actually you know and today will produce more bicycles and cars and it's faster growing than cars except China where it was the opposite before the bicycle will survive the car will survive the Warped wall bookshelf will survive tell you want to think about the not because these reading things aren't good but because it's too unstable of technology and what else did I add in it and I said the economic establishment the occurring existing economic establishment quantitative economics particularly we'll look at them the way we look at Miss Bree fortune teller and Lower East Side of New York I will talk about him that was my article he's an economist he looks like an economist Thomas I'm a junior in the economics department yeah very good how do you think I could knock economics and finance should be taught in the classroom and what rolls very good analysis play both in the classroom and in the real world I just wrote an article in the New York time I don't write articles often my articles are typically 300 words to me that you can do them standing up and I just wrote one in New York Times and they said how should we teach the economics is there's a certain class of economics that should not be teaching and if you want to do something else teach gardening instead they're better for society no because they're harmful and how can you identify them second-order effects and let me show you what we did here very simple I have a technique to identify them I don't have a technique to identify it okay now I have the paper here convexity to perturbation of parameters very simple I can tell as a model has second-order effect by taking the parameters and their parameters there are fixed making them stochastic the paper is on a web okay so you can tell what models makes you make you blow up from black swans and which ones are robust all of classical economics survived everything done that has equations in it over the past 50 years almost everything is bust they say that simply take Markowitz for example if you expand the model to make some parameters help to be constant make them stochastic the model reverses conclusions so is it sound well what makes you blacks want prone what you do better without a model so economic thinking is on you know doesn't change it has survived the test of time it's what has been done since Samuelson optimization that that goes bust once you perturb it you expand the model to put to bake parameters or introduce fat tails or make probabilities of the tail a certain and it's very simple a called fragility parameter and one I used it to show why Fukushima computation done for Fukushima was 40 they said this event should happen one in a million years okay if you perturb 8 the parameters a little bit it happens every 30 years okay these are the models that go out the window and you can identify them by perturbation parameter to see if they're nonlinearities we should have the graph here it is very simple if you start applying it econometrics is out the window something survived of course that have quantitative and everything literary in economics take Hayek no equation survives so strange okay and you can show that quantitatively by taking the equation themselves and expanding them by putting what I call a second layer of uncertainty so it looks like we're all so thank you very much for inviting me you

Channel: Woodrow Wilson School of Public and International Affairs

Views: 113,446

Keywords: 04, 10, 2012Nassim, Taleb

Id: 33kET2YPWls

Length: 66min 0sec (3960 seconds)

Published: Fri Apr 13 2012