Behavioral Finance and Investment Strategy

famous early bubble is this one that people have read quite a bit written quite a bit about and it's called tulip mania and back in Holland when they were flush with lots of cash from from their overseas trading and the invention of modern finance we so forth they needed a way to basically to lose it right so so they got into the tulip business and you know think about tulips is that they became very very very expensive in fact a tulip at one point was worth more than a townhouse or a couple townhouses even and I won't bore you with the numbers but what happened is that the value of these tulips plummeted very rapidly it went back down to earth now of course you can buy them at the orchard supply for you know 50 cents apiece okay now there's that's not the only one there's tons of these bubbles and I love studying these bubbles and I'll walk you through a couple of them there was one that happened in England it's called the South Sea Bubble and this was really the first IPO so you know we've been doing IPOs for a long time and we've had I pee Oh bubbles for a long time and this is actually the coin the term bubble and it's inspired the bubble act in England which banned IPOs for 100 years okay that probably wouldn't go over too well nowadays in Silicon Valley and you know you see that the price of the shares in the South Sea company went went way up and then of course went way down and you might be thinking well who would be dumb enough to buy at the peak of the bubble right who would be dumb enough to buy and by the way this company that basically no business model their business was illegal the rest of the world outlawed what they did but that you know they were trading slaves to countries that basically would not allow them to trade slaves and so the price of stock went way up and then of course it plummet who would be dumb enough to buy at the peak of the bubble people dumb people like Sir Isaac Newton for instance who lost his entire fortune in the drop now of course you know why would Sir Isaac Newton buy at the bubble peak of the bubble because he was a big believer in momentum right he actually only or not he came up with some theories he invented momentum right the problem is that Sir Isaac Newton also he forgot about this but he also invented or discovered gravity and so if you'd remembered that part of his theory he would have avoided the rapid decline thereafter okay now fast-forward to recent times maybe some of you were old enough to remember the Japanese real-estate bubble I first started teaching every thought that Japan's gonna take over the world you know they even bought Pebble Beach and they bought you know van Gogh paintings and everything and it was all based on the land value that had inflated in back in Tokyo so back in 1991 the land value in Japan was nearly twenty trillion dollars so so you could for the same amount of money either buy the land beneath the Emperor's Palace which you know find piece of real estate or you could buy Canada now which would you prefer under your Christmas stocking I'd rather have Canada what are you gonna do here you walk around look it's nice and pretty but they don't mean well or anything or hockey there's nothing good there okay so what happened is for Japan to welcome you and the you know Japan is not recovered yet damn and they're still mired in this recession for the last thirty years twenty years okay now more recently maybe some of you this is gonna touch a little too close to home but the dot-com bubble and just to give you a sense of what happened there I'm not gonna go through all the aggregate details and remember this company VA Linux anybody has a stock still okay so you know this thing opened at nine point five billion dollar market cap now what's their business free software right okay so what happened well this is what happened the stock went from two hundred forty two dollars to seven dollars but the thing is they they were smart enough to keep their burn rate low enough so that they're still in business unlike so many of these other coms they are still in business except that they don't really do software anymore they are now called geek net they sell t-shirts which is actually a more viable business model because at least it has revenue $2.99 a shirt so if you want to buy gonna support G Linux boost uh you know the stock price then you can buy some buy some shirts okay and you know you guys all know about Webvan you know maybe you guys got sucked up in web man you know maybe you got sucked up by pets.com right maybe you got sucked up by the housing bubble okay so I remember this slide came out 2006 I remember talking to people and saying don't buy a house don't buy a house they're like wait it doubles every 10 years gotta buy a house and of course that's what happened to houses except in Berkeley where they're even more expensive than they were in 2006 okay so again what about the stocks that were inflated by the housing bubble like Bear Sterns Bear Stearns Lehman Brothers these guys completely crashed after all of the inflation of their asset prices in in the 2000 and of course more recently poor 'old Zynga okay in adult anybody work at Zynga okay all right they're hiding underneath the chair playing with farm farm bill okay so I shouldn't say anything bad about saying it but you know that's that's another one that's kind of collapsed so you know how do we understand these bubbles you know it's great to be a historian and tell some tall tales and fun stories but oh look at those poor Dutch people look at those poor you know Zynga investors but we can also construct theoretical models and so financial economists are really good at building out models let's assume this and then add in some calculus and boom there's our answer in our paper under tenure another way to do it is to use experiments so experimental economics is this brand new field it's not really brand new it's been around for a while within economics where you try to replicate the techniques of the natural sciences right so if you're trying to figure out what it's a rat fat okay you know you stick them in a cage and you force-feed them all sorts of things like you know McDonald's or whatever and see if they get fat right and then you have a bunch of other racks in another cage mice but they're exactly genetically identical and you feed them you know subway and you know see how fat they get right and that is the experimental method and so economics which has been built around kind of looking at the world doing empirical stuff or cooking up all sorts of theories in your office has now moved in the direction of the Natural Sciences and we can do experiments on people to find out what drives their behavior okay and so we actually have a lab here in the building across the way and it's called the Berkeley experimental social science lab or X lab for short and so what we have is these are these different cubicles and people will walk in generally undergrad students do this MBA is they basically screw up every experiment so we ban them but but you know undergraduates will come in and will administer all sorts of experiments on them to find out you know what is making them behave in a particular way so if we can look at something like a bubble and say oh you know we've seen it in practice but maybe it's not really a bubble because if you look ex-post and you saw there's no way tulips could have been worth you know ten thousand dollars a tulip but for markets to be efficient the only thing has to be true is that the prices have to be consistent with the information set that was available to the people at the time so you know lots of times prices go up in prices go down and that doesn't mean that there's a bubble it just means that people get new information and they think everybody's going to buy their pet food online and then of course they realize they're not going to buy their pet food online and so the stock crashes that may not be a bubble if it's consistent with the information that the investors had at the time so the great thing about an experimental setting is that we can control the availability of information to the participants in other words we can create a market where the fundamental value of the asset is known to all of the participants right if we try to test to see whether or not Zynga really was worth four ten dollars a share would have to have some sense of what the fundamental value was at that time and we don't know it so that's the great thing about experiments so if you're going to set up an experiment that would test for the propensity of people to naturally among themselves create bubbles in asset markets what we can do is create an artificial asset with a value that was known to everybody and then allow people to trade them amongst themselves within this artificial market and then at the end as with all these experiments we pay the participants money according to how successful they've been at engaging in the activity so there were traders that make a lot of fictional money in the lab get paid a lot of real money and the investors that lose a lot of fictional money in the lab wind up you know getting paid less money okay so they're incentivized with these compensation schemes we use real money to pay our you know victims or you know participants in in these experiments unlike the the poor mice who don't get paid they just get their necks broken when we're done with them okay unfortunately can't that's why we don't let MBAs into the experiment because a lot of us will want to break their necks when we're done so to keep them safe okay so here's the the experiment that is sort of the canonical experiment that has been replicated numerous times and which we have carried out up here in the ex lab on a bunch of occasions and it's actually the guy who invented this experiment his name is Vernon Smith and he won the Nobel Prize for inventing experimental economics and the idea is we create this asset which has a finite life and this this asset is going to spin off dividends over its lifetime and then it ceases to exist everybody collects their checks and they leave the experiment okay now the one thing that kind of makes it a little bit interesting is that the dividend is a random number the distribution of which is known to all the participants so if you have even the most basic mathematical ability and our house undergraduates presumably have that you can figure out what the average or expected dividend will be each period and in this case it's going to be 24 units or 24 dollars we get $24 each period and then the students or participants are given a bunch of cash and they're given some shares if they want to increase their holdings in the shares they can buy the shares if they want to decrease their holding of shares they can sell the shares and so they're asked to submit bids and asks you know through a computer terminal and then the computer displays you know the prices of the assets them to supply the demand etc and then you know exchange takes place all right and then we do it again with the same people to see if they you know behave any differently so so here's the asset it's got fifteen periods it's got a dividend of twenty for each period the experiment only lasts like an hour so you don't have to do any discounting or present value computations and so the value of the asset according to this very simple calculation is 360 you know lab lab dollars okay and that's what it should be trading for in the first period now after we figure out who owns in the first period the dividends are distributed to whoever's holding the asset and then they're allowed to trade again now because the dividend has been paid out and there's only 14 periods left the second period it's going to be worth 336 lab dollars and so on and so forth right so the fundamental value of this asset is going to be declining until you get to the final period where the asset has no value left okay so it's worthless and so the value of the asset at period 16 should be zero okay so I've got this whole thing set up and now we say trade let's trade and so this is a display this is sort of what they see so you have your Holdings over here okay you can submit your buy order or your sell order and anybody who's done any kind of trading understand sort of how this works and then over here the offers to sell and the offers to buy are displayed and then there's a market clearing price at which shares will change hands okay and you know you're given a couple minutes to think about it and decide what to do etc so what do you suppose happens we suppose happens boom first period well the price it's way below the fundamental value people are looking at those shares moves a little suspicious I don't know what I really want that I mean oh no I mean I got cash I don't really know what it's worth I don't even know if the stock's popular right you remember this morning popular is good right so they don't know what it's popular yet so they're not quite sure it's the the new kid on the block they have to size them up all right and then you know there are some people that look at that and say wow you know that's pretty cheap cheap asset it's worth below its fundamental value so let's buy it we suppose happens price well you know doesn't really go up much the volume goes up and start people start to see who people are get interested in this market then the price goes up a little more next period that goes up and it goes up and it goes up and wow it's now approaching the fundamental value so at this point you'd think everybody's figured it out they've done the math they know what it's worth and so at that point it should the bargains are now extinguished the p/e is up to its historic high and now the disc the party's over but new keeps going right now it's trading at more than its fundamental value okay and now maybe this is just a mistake it's people look at like ah whoa whoa we overshot maybe you know we should get out of this thing before the party you know where it ends before people figure it out but no keep going up right and this happens with the real people with real money in a real experimental setting and then it keeps going up a little more and then it stays up there and this is people buying and selling this thing at a price that's way above its fundamental value and then all of a sudden people start saying hmm what's going on maybe we should cash out our positions locking our gains and then the price starts to decline and then it goes back to the fundamental value as soon as people see that the game is about to end they're like I got to get out of this thing and they cash out okay so this thing right here this is what we call a bubble now of course this thing is also interesting you know why do we see these things trading at discounts but this is what I'm interested in right here at the bubble okay where it's trade yeah well they have all they have to do is do the math so you average the four dividends and you get 24 and then you multiply 24 by 15 and then BOOM and we teach them how to do this okay then did you take introduction to finance dividend discount model you know it's like week 1 okay so you know we're assuming these good I mean they're not all ha Thunder grads I mean we might bring in some English majors right but even they they wouldn't get into Berkeley if they couldn't multiply right now you might be thinking well what if I what if I you know just said all right you're getting $24 a period you got it it's still you still get a bubble okay I mean you can spell it out as much as you want and you're gonna get a bubble now after this sells yeah yes well we need to give them a finite you so if you don't have a finite end period then you don't have the ability to calculate the fundamental value right so we want to make it so that the fundamental value is is knowable right so it's not like trading you know Zynga stock where you don't really know you have to make estimates about earnings and all this other kind of stuff it's like we got a $20 bill how much you want to pay for it okay you want to make it really easy really really easy okay so we get this bubble now you'd think after this some of the people are going ah man was I thinking I can't believe I bought this thing for more than it was worth you know and he's unappealing I can't believe I didn't buy it back when it was cheap so then we say them all right let's do it again like okay okay this time we'll do it this time we'll get it right so we open up the trading and what happens starts off low again nobody really knew what's going on look people are timid again maybe they're still recovering from the bathroom break I don't know but he starts off low and it goes up and then finally like oh I've seen this tune before now I know what's happening what happens it keeps going up right now notice that even though you get a bubble with the second time around it's a little bit different right I mean it starts a little earlier it ends a little earlier it's a smaller and so in some sense you could say people are a little wiser I don't know they've been suckered once they're not going to get suckered again but they do get suckered again but not quite as much so we do it again let's do a third time and then a third time well okay people have learned their lesson so it seems like twice bitten three times shy so the bubble kind of goes away when you do it a third time so what does this say well experienced people learn what's going on they've been you know punished or rewarded based on their experience and they they generally get it after a while if everything's stable now what's really scary about this is that and you can use this we can start manipulating liquidity we can manipulate the order book we can do all sorts of things to see what institutional conditions will make the bubble smaller or weaker and appears at least based on this version of the experiment that the one thing that really matters is experience and so people on Wall Street back in the you know 60s would say we're never going to have another bubble like the 30s as long as I'm still working here and then what happened is those guys retired and then you know you get the next bubble so the idea was that as long as the people live through the last bubble are still around you won't have another bubble you're gonna need a new new batch of suckers to get the next bubble now that's the recently we've seen these bubbles come you know closer and closer back-to-back and you know I think they're it's because you're not repeating exactly the same thing so if we go into this experiment we just changed one thing so it's that 15 periods it's eighteen periods or you change the payoffs to change the dividends or you mix up the people so you say okay you've done it three times that groups done it three times now let's put you guys and you guys together you get a bubble again so as soon as you change one thing and that's maybe why people who lived through the dot-com bubble said man there's no way I'm ever going to get suckered into buying equities again I'm sticking a real estate right so he helped it got busted again but you know something different so you know maybe the people who've got you know who blew it in the dot-com bubble will never blow it another dot-com bubble but they've got sucked into some other bubble the Bitcoin bubble right okay so experience seems seems to matter as you can see it tends to flatten out these these bubbles now what I made yeah oh that's about the volume the volume so it tells you sort of you know how many shares traded okay now you can repeat this experiment with expect I will say I did this with a bunch of financial managers like financial advisors and and you know these are these are actually I'll show you I'll show you those not one shows all its financial advisors I've done this with them these guys are off the charts in terms of the bubbles that they create I mean it is not I mean I have never seen it you know you can do this with undergrads English majors you'll get bubbles but Anne Shiell advisors there's no room on the chart the computer like melted down because the pry it wouldn't support prices that high yeah there's two questions in small deviation one is do you think the Fed is looking at any of this because there seems to be you know number two is it's related to the question of one but in your work post the bubbles does it affect their future buying behavior right because one of the things we see from these bubbles is that the economy tanks even though the dot-com bubble blew up right there's the there's a spillover effect well yeah of course so that you know if it was just like ah that was fun we had a bubble no big deal but of course what it does is it has long lasting effects in particular on institutions that have been you know suffered through the bubble so the balance sheet of the Japanese banks the zombie banks you know is still weighing down the Japanese economy you know people tend to become more timid and take fewer risks so my colleague Rica mal Madeira has done Studies on investor behavior on people who have lived through crises and bubbles they change their investment behavior going forward dramatically right because they take on fewer risks and so you know that's one of the reasons why the Fed is very aware of how recessions can sort of perpetuate themselves because people unwilling to spend or unwilling to take risks and so they this is part of but part of macroeconomics for a long time but it hasn't been grounded in the micro study of asset pricing and and how asset prices work so so I think the Fed is very much aware of this and they they do pay attention to this okay so what I'm interested in talking about today is well what can you do about this right how do you make money from this so if we go back to this picture we look at the prices you know if you followed the warren buffett rule which is by when the stock is below fundamental value and then sell when it's above fundamental value what would you have done here you would have bought right here okay and then you would have held on and then you would have sold the minute this thing penetrated the fundamental barrier right so is it trades for a penny above fundamentals you're out okay now that guy would have done better than the people you know who had bought into the bubble certainly couldn't you do even better how could you do even better selling up here I was holding on after the asset goes past its fundamental value you can do even better than the basic fundamental guys so I have a lot of friends it work in hedge funds and this is what they're trying to do you know they're like fundamentals yeah you know that's important I got to know about that of course but I'm really interested in how other people behave I'm interested in what the market thinks right where's the market going you know what is the investor do one of the other strategies out there that are driving prices okay so in order to understand investor behavior and in aggregate we need to have some some game theory so game theory is another course I teach and it's a lot of fun and usually game theory is applied to things like industrial organization and labor markets and pricing and rarely as game theory applied to finance and so that's sort of what what I'm interested in doing so if you remember or have encountered any game theory there are a couple very classic games they're simple two by two games where you have two players each of whom has two strategies and then you can depict it in a by to box and so if you've heard of game theory at all you've heard of prisoner's dilemma okay now I'm not going to talk about prisoner's dilemma let's talk about these other two games one's called stag and hare which is a cooperation game and the other one is a Hawk and Dove game and then if you under who if you follow me as I work through the these games in their basic form I'll try to explain how they help you to structure your thoughts about interactions in financial markets okay we teach game theory in our strategy classes it's a key part of strategy so so here's a hypothetical suppose you're going out hunting you go single a hunt here why is it Virginia and ask people hunt you like and you know here it's like hunt are you kidding me you know like you know tofu doesn't move why would I want to right so so if you're hunting for for stag stag essentially is a two-man job you need two people to hunt the stag you need essentially you got a corner of the stag you see them actually wandering around campus late at night but you need two people basically you need one guy to kind of shoot the stag and then when the stag starts running off into the woods you gotta have another guy at the other side to shoot him again right so it helped drop dead you need two people you got to trap this thing all right so it requires cooperation now the other alternative is to hunt hair a hair is a one-man job you know you just chase after it and then you get it and then you eat it okay so the question that I mean I've never done it I'm sure it's more complicated than that but it's it's basically a one-man job and so the classic cooperation game that I'm describing looks like this now what is this saying this says that if you have two hunters and they each decide to go after hair in other words they each operate on their own in isolation they will each get two pounds of meat okay if they cooperate and they corner the stag and they eat the stag then they'll each get four pounds of meat because it's a much bigger beast okay now if one of these guys stands in his station doing his part to kill the stag while the other guy at the other end of the meadow is chasing after hare then the stag guy will get nothing because the stag will escape the hare guy will get his two pound of meat okay so which is best for everybody as a whole what's the social optimum here it's the cooperative equilibrium where both of them go after stag they both fill their bellies and enjoy themselves but what's the equilibrium what's the equilibrium so the equilibrium is is where you know the parties are content with what they have and no one has any incentive to to deviate this is something called the Nash a broom is the most basic fundamental concept in in game theory so look this is not an equilibrium because if I go after the hair and get two and you go up to the stag and get zero I'm going to think boy I should have done the stag if he's doing stag I should have done stag then we both would have gotten for the other guy who got zero says he did hair I should have done hair if I had the opportunity to go back and change my choice I'd switched to hair so if you find yourself in this box what's going to happen is you know one of the parties is going to switch when they find out what the other guy did okay and so if both of them are going after hair then that's an equilibrium because neither party has an incentive to switch because we switch out of it then you wind up with nothing stag stag is an equilibrium because once you both have the big you know stag pin down there's no incentive for you to go after hair because then it would reduce your companies your pay so these two are both equilibria the question that is how do you wind up in this equilibrium so imagine the following you know you are going into a hunting lodge in Palo Alto they got all sorts of hunting apps right and and and so he'd help you're looking around say hey anybody want to go hunting with me ok so do you want to go hunting with him ok so the two of you guys are going to go hunting now I want to hear your conversation because if you decide to go stag you're going to have to go to opposite ends of you know opposite ends of Sand Hill Road and you want to see one another so let me hear this conversation okay so then you're both gonna do the stag right ok so you grab your guns and you go in opposite directions now he's over somewhere on the other side of Sand Hill Road you can't see him what's going through your head let's hunt VC nope well what's what's going through your head so what are you gonna do you might switch so in other words your payoff to going after hair is - what's your payoff to going after stag it's going to be it's going to be either 0 or 4 depending on what you think he's going to do so you're going to say hmm what's the probability that he's going to go after stag now if that probability is greater than 0.5 then you'll stick to your end of the bargain but if your estimate of his reliability drops down below 0.5 then you're gonna be like yeah there's a hair I'm gonna go after a hair okay so how would you figure out what the probability is of him actually living up to his end of the bargain ah there's no sin here you're out of cell phone range man that's right no information personality judgment so let's suppose that Gordon appears to be extraordinarily reliable you know you checked his Yelp ratings there they're real good you know in Gordon's Gordon's a great guy really reliable guy so you're thinking okay I can depend on him but that's not all it matters because what also matters is well what does Gordon think of you because your estimate of his probability of sticking around is going to depend on what you think his estimate is of you sticking around so maybe Gordon's a really reliable guy but he doesn't know that you're a reliable guy cuz he looks up your Yelp rating and it's there's nothing there your Klout score is zero okay do you know what that is okay good so you have to think not only about what he's like but you have to think about what he thinks you're like now maybe he knows that you're reliable but he doesn't know that you know he's reliable yeah well yep he has to understand game theory right so so but look here's the important thing you have to think about what he thinks but you also have to think about what he thinks you're thinking which of course means that you have to think about what he thinks you think he thinks you think he thinks you think he thinks you think all the way down ad infinitum now it started as a very simple computational task you know which is bigger - or the weighted average of four and zero but now you can see that this is a very complex task right psychologically and mathematically even and so you know what you do is going to depend a lot on your beliefs not only about him but your beliefs about his beliefs about your beliefs about his beliefs and so on and so forth okay so it's a very very complicated thing to get to one or the other of these solutions and you may wind up you know anywhere really in these four boxes if these individuals are left to decide on their own what to do now in the backyard I want you to think about what is the relevance of this to finance and trading and etc so think about that the second we will get back to it now there's another way that you could think about this that's sort of the classic economic way of thinking about this problem the other way of thinking about it is the way biologists think about it so so I taught a biology course once which was a lot of fun because I didn't actually ever take a class in biology but it didn't matter because biology and economics are really sort of the same thing they're isomorphic mathematically in many ways and and so in biology the way they learn to think about game theory is they think about not like what's going through your head or what's going through your head what they do is say let's imagine we have two organisms who don't think they just do and they do what their genes tell them to do okay and then we throw them together and have them interact and see what happens so suppose we have a petri dish and we have a whole bunch of hard-wired stag hunters and a whole bunch of hardwired hair hunters we throw them into this petri dish interact with one another in a random fashion and each time they interact they get a payoff they either get four or zero or two and if they if they get four well then they have a lot of offspring and these offspring inherit whatever characteristic you know they exhibit so if their hair hunters and they have four offspring well they can tell up to two offspring then those two will be hair hunters if you're a stag hunter and you have four offspring while those four will be stag hundreds so these numbers represent sort of you know your offspring in a fitness model in biology now what happens now we can solve for the equilibria because your payoff now depends not just on your type but what the ecology of types looks like so if I'm a hair hunter I'm going to get a payoff of two regardless of the distribution of the population I just get two two two two two two the stag hunters their payoff is going to depend on how many stag hunters there are so if I'm a stag hunter in a world of hair hunters everyone's running around I go hungry I get zero but if I'm a stag hunter in a world of stag hunters I get four every time I go hunting so what's the equilibrium ajan we just do this we put we create a artificial society I'll show you example this we just allow them to reproduce well in this world here where we start off our starting point is with less than 50 percent stag hunters what will happen is the stag hunters will do worse than the hare hunters and they'll go extinct but if our initial condition is with 51 percent stag hunters the stag hunters will do better than the hare hunters and the hare hunters will go extinct right so the starting point the original distribution of the population will determine the outcome okay now where do we see this right in life well we see it in standards Wars right so if you're trying Network externalities its referred to right so if you're trying to figure out how valuable is it that you're operating in the windows world well it really depends on how many other people are using Windows I'm the only guy using Windows in a world of Unix well my all my documents and everything else have totally worthless if I'm the lone UNIX guy and I don't do UNIX I got utak site and but M is surrounded by a bunch of Windows people well then you know my documents are worthless so in order to we're going to see a tendency within standards to have a single standard I mean you can think about in all sorts of other areas of life you know suppose you have railroad tracks that are four feet wide you like I got the three foot wide train you know you're not going anywhere except Spain I guess right so so we're gonna see a push towards one or the other and that's why of course a lot of companies that are participating in standards Wars try to you know gain market share as fast as they can so that they can it can dominate the space you know well squeeze it so whether you drive on the left side of the road of the right side of the road this is essentially a game of stag and hare if you're a left hip you drive on the left side in a world everybody drives on the right side you're not going to last long you're not gonna have a lot of offspring so historical conditions will dictate you know whether you go all left or all right so you know most of the world is is pretty much on the right side there's a some chunk of the world that was colonized by England which goes on the left side and of course there's India where you just drive whichever way you feel like driving jury's still out right now well interesting actually Samoa Western Samoa just switched from driving on the right side to driving on the left side how does that happen that's crazy right so they did you need almost never see switch Sweden switched back in the 60s it's really really Sweden actually was thinking of switching they said there's a really dramatic transformation so we're gonna phase it in that got shot down like you know midnight everybody had to switch sides okay so alright so let me tell you a second game let me tell you a second game everybody get this game alright it's called you see it everywhere it's called a coordination game or stack and the key thing is you want to do what everybody else is doing so you have to look around and say what does everybody else doing okay that's what I'm doing so everybody else is buying the asset maybe I want to buy the asset okay so herding for instance is going to be it's going to look like one of these coordination games okay now the second game I want to tell you about I'm only gonna tell you about two games is this other one it's called Hawk and Dove and it's going to work very different and so when you get when you learn these games and game theory you can walk around the world and see things like you walk into a restaurant Amanda's game game chicken going on there over at that table and oh there's some prisoner's dilemma happening over there and it kind of gives you the frameworks and structures to to analyze the the world in it's more detailed form so here's how Hawk and Dove works you got these birds okay and they're in dyads you see two birds and the two birds are approaching a pile of food all right now the food is what are the worms I don't know whatever birds eat so they're approaching the pile of worms and they have a choice should they be a hawk which means they come in all aggressive talons out the beak out we're said to be doves and they just go into like no be nice and so here the payoffs to this interaction it's going to work very similarly to the ones I just described if you have a pile of worms with a value V and two doves approach the pile of worms they will split it peacefully like oh you take half I'll take half great and they swap business cards now they bump one another do people still bump I don't remember but you know this presumably is what would happen if two Haas grads we're fighting over the same deal right they would just split it 50/50 now if a hawk approaches a dove Paco's right and the doves like I'm out right he's going to take off and leave the entire pile of worms to the hook but if two Hawks come together okay they're going to ultimately wind up splitting the pie because they're equally matched but in the process they're going to wind up paying a price and that price is W so you can think of W as the size of the pile shrinks while they're fighting so maybe the worms are escaping okay that could be like wastage or it could be wounding so that you know there's there's blood loss so they fight they fight they fight and then you know each one is a 50% chance of winning but they you know they have to go back and lick their wounds okay so now what's better in this game what's the street is there at equilibrium right it's clear I think that the social welfare maximizing equilibrium would be one of these three because there's no wastage but what's the equilibrium here now if W is trivial then it makes sense to be a hawk under all circumstances but if W is large enough you have to think twice before fighting and so what's going to happen is if W is large enough your optimal strategy would be if you see a hawk coming be a dove rather than suffer the consequences of a fight if you are a if you see a dove coming be a hawk and then you get the entirety of the pie now whether this only works when W is really big so I mean think think France in Germany World War one or you know they fight they fight they fight ten million dead people they're like okay let's just you know neither one of us gets Belgium okay or you know Iraq and a run I mean they just fought and fought and fought and fought and you know everyone just went back to the original status quo but there are millions of dead people okay so that would be the Hawk Hawk story that's sufficiently painful then you're going to want to do not what the other guy is doing but what the opposite of what the other guy is doing and we can show this using the evolutionary logic and it would look something like this if you're in a world of doves if you're in a world of doves every dove is getting half the pie from each interaction now if you inject a hawk into this if you take a hawk and you plant them in the middle of the sea of doves okay it's they're just gonna go crazy I mean this is going to be like taking a hardened criminal from San Quentin and sending them to Walnut Creek you know I mean they're just gonna they're going to do really really well okay and and so so they're going to reproduce like mad and and the proportion of hawks in this society will increase okay because they're going to win all these battles now let's go over the other extreme end where you have a war of all against all right where everybody's fighting tooth and nail trying to kill one another and then you inject you know Mahatma Gandhi in there and he's just I'm not fighting I'm not fighting I'm not fighting oh there's some food okay I'm gonna get the food and then boom he's going to do better than these other guys who are knocking each other silly and so what'll happen is in a world of hawks the doves will do well and what will happen is regardless of where you start off you will converge on an equilibrium where hawks and doves do equally well they both get the same payoff now this is what's called a frequency dependent strategy and it leads to here what's called so as it's really evolutionarily stable these strategies when they exist in these proportions theorists also called this a mixed strategy you see mixed strategy in sports so where where do we see these kinds of strategies where the goal is to you know mix up what you're doing well let me let me you wanted one of my favorite examples from nature because you see this in nature all time by the way the proportion of male and female offspring is a mixed strategy that is a mixed strategy equilibrium you have a 50% chance of boys 50% chance of girls approximately that's because if you had a world of boys it would be great to be a girl and if you had a world of girls it would be great to be a boy in terms of you know offspring so does anybody know how salmon reproduce anybody here fish for salmon No okay well you know about dams and what it would it fit what are they what do they do where they go what happens this sounds apart from the running upstream past all the dams in the fishermen once you get to the top it sounds like a Harlequin romance novel that you're describing okay it's not that pretty sorry that that's a after-school special what what happens is after these salmon make it up through the rapids over the waterfalls past the fishermen past the bear you know they think that they have the end in sight where they can get romantic but that's not what happens that's when the hard part begins that's when they start fighting what's going on here I'm supposed to be plugged in let me try that again again that's what they do this is not a Harlequin romance these are male salmon fighting with one another okay why are they fighting with one another what's the point of all the fighting they're fighting for real estate okay why are they fighting for real estate they want to what they want to be daddy that's right so what happens is the female salmon when they get up there they got to lay their eggs now where you going to lay your eggs if you're a female salmon where yeah you want to find some nooks and crannies to protect your eggs from all the Predators so if you're a male salmon and you want to reproduce you're going to locate yourself right by the nooks and crannies but all the other male salmon want to go by the nooks and crannies and so they fight for who gets you know it's almost it's actually like if you ever tried to buy a house in Berkeley it's kind of like that okay it's real estate right so people are these male salmon or fighting over one other now if you know that these male salmon are beating the living crap out of one another for this real estate maybe there's an alternative strategy as a male salmon that you can use to reproduce what might that be we're think about another way if you are a parent of a young male salmon would you want your young male salmon to grow up to be big and strong of course you would except maybe you don't it turns out that if you have male salmon that are big and bruisers and they're fighting one another it might actually pay to be a little tiny sneaky slippery salmon because while these males are fighting and duking it out the little slippery salmon slips in fertilize the eggs it slipped out right and the male salmon they're fighting it what was that oh jeez he's gone oh then they go back to fighting right so in a world of male salmon it actually pays to be a little tiny salmon but if you have too many little tiny salmon then a big salmon if he jumps in he can guard his territory he's got nobody to fight you can just spend its time wording off the little baby salmon so in fact right this is a frequency dependent strategy and if you look at salmon there's actually this dimorphism where you have big salmon little salmon and it's a stable equilibrium where the proportion of small salmon is sort of constant and small salmon don't forget small salmon large time and don't forget large salmon rather you know you have offspring salmon and there's a probability associated with like male and female ok now where else do we see this we see in sports so if you're taking a penalty kick where are you going to aim it left or right this is right well maybe cuz you see what that guy is doing but look the way it works with penalty kicks is but it's when the guy kicks the ball he doesn't know which way the goal he's gonna go and when the goalie jumps left to right he doesn't know which way the ball is going to go so you have to actually decide simultaneously if you want to have any chance of defending the ball so if some kickers like man I'm much better shooting to the left I'm just gonna always shoot to the left you start shooting the left the other guy the goalies gonna be oh that's the left guy and he's gonna run in and defend the left or if you're the goalie and you're like I always go right then the shooters gonna always shoot left so you actually have to pick a a mix it has to be random right just like with football you know you're gonna do it pass or a run you have to keep the other side guessing all the time even if you're better at passing than running you have to run occasionally to get them off-guard okay so this is actually you can solve this mathematically and you know what you have to have is you have to an internal randomizer because if you show any predictability then the other guy will will will be able to defense you need one of these inside your brain somehow maybe Google glass maybe the the football players will be using Google glass in so there was actually an economist and isn't I mean what else what can't you do research on right so you know if you want she always take a look at all the fun research that Hass is sponsoring this is not Hospit these guys went out and let's guy went out and looked at football players you know european football players and he found that the based on the probability of success that these guys had they were playing almost perfectly they're randomizing both the goalies and the shooters were randomizing almost perfectly even though none of them know anything about game theory presumably okay they just fell into the social optimum okay now there are some other sports where it doesn't happen actually so in baseball you know if you throw fast balls every time then you're going to give up a lot of hits so you got to mix in some off-speed pitches so according to this study there's actually too many fast balls and these guys switch back and forth with too much predictability and so you know we actually have I have a lot of students that work with people in Major League Baseball and other we do research in this area we have MBA students now that all want to go into sports management they're using analytics and they came in second place in the sports analytics competition that was sponsored by the NFL last year okay and so another result is that there's too much not enough passing too much running in football okay now where do we see this in real life this is a problem that a lot of you may have encountered I certainly run into this problem if I'm in Berkeley and I need to get down to Palo Alto okay I've got three possible routes I can go across the Bay Bridge San Mateo Bridge or the Dumbarton bridge okay now if you look at if you if on average these three routes take almost Dave but there's no traffic they both take they all three him take exactly the same amount of time really there's no reason to take one or the other except for you know and they all have the same tolls as well so there's really no no reason to pick one or the other but when Google invented you know Google Maps or before that you know you could get real traffic updates from from other sources and you could look at this what would you do if you saw this this map so I'm saying Berkeley I'm like which route should I take I go look at this app and I go oh man I'm going to go take the Dumbarton now does that make sense is that right that's right so if you're the only one with Google Google Maps then you can take Dumbarton and you're probably going to be in the clear what if everybody else is looking at Google Maps then everybody else is going to go in the Dumbarton right so you might actually be better off taking the Bay Bridge even if it's like you know there's an accident on the Bay Bridge maybe you know by if no-one goes there then you'll be the only guy and you get around the exit but then again right those other people who have Google Maps are wondering what you're going to do and they're saying well wait a second if the Dumbarton bridge is green then everybody's going to go to the Dumbarton bridge but wait a second everybody's gonna figure out that everybody's going to Dumbarton so nobody's gonna go to the Dumbarton something I should go on the Dumbarton so you need to actually you can't just look at this this map doesn't help you wait you need to know not only the state of the roads but what is the state of the the cognition of the other people out there how many people are looking at this and then what are they doing what's their strategy right what are they guessing it's back to the stag in here I need to know what does that guy think I'm gonna do what did he think I think he's gonna do but what do you think I think he thinks I think you think everybody thinks everybody else is gonna do and you know this becomes an extraordinarily difficult problem in a world where everybody has access to Google Maps so when it was just a select few it was great but now everybody has it and it would be better if we just said Google just tell me where to go and then Google would assign you go that way you go that way you go that way right and and if we all followed Google's logic then we you know we would maximize traffic flow I mean that's that's the idea right so so that's up so you know the problem with some of these models is that we can mathematically compute what the optimal what what the optimal routing distribution should be and what the optimal mix strategy is but we don't always get to equilibrium so if we have a world with with say predators and prey and they're coexisting we might think that you know if there's too many predators then the prey disappear and if there's a you know too few prey than the Predators disappear and so when the Predators increase their population then the prey will go down and so you won't think that the predator and the prey would think through this problem they would work out the optimal equilibrium you'd have a nice steady ecology with neither too few nor too little predators and prey everything would be good it's kind of like old MacDonald's farm the problem is that there's a delay in the feedback right so if we start with a lot of hares then the Lynx is here who are predators will multiply but they multiply more slowly than the hares and then as the linked population expands that the hare population will crash and then right these guys will crash a little later and so you see that there's this constant oscillation in predator prey models where the population expanding and contracting you know with this rhythmic regularity ok so there's not a steady equilibrium now you also see this on the freeway so you know suppose you're on the left lane and the right lane is zooming past you right what do you do swear yeah you know you know you're going to the right lane but then of course then the right lane becomes the slow lane and then your left lane you go back the left lane but and you never actually see the lanes sort of smoothing out you just constantly see this this oscillation you know there are people that study this and that you can watch it on video ok so now let's get back to to finance and think about you know you have two strategies as an investor and I don't mean like as an ordinary investor I mean sort of as a professional investor one of which is passive no where you you just put all your money in an index and you just say ok you know prices are right I'm just going to work on asset allocation I'm just going to focus on covariances and correlations get mean variance efficient investing that's all I do I don't I don't look at balance sheets or income statements or any of that kind of stuff ok and there's a lot of vestment professionals that advocate that approach first is another approach which is you know I'm going to look at all this information and data and I'm gonna do research and I'm going to look at balance sheets and income statements and I'm going to make phone calls and all this kind of stuff and I know people who are living both worlds now the question is which is better so suppose that investment research is costly you know you guys spend a lot of time deep in the weeds looking through financials doing you know value-added research are you going to waste your time doing it no because on average we know that the average investor does worse than the index investor so let's all put our money in the index now what happens when everybody puts their money in the index when everybody puts their money in the index well nobody's doing any research so you're like Enron that's 2% of the market let's make sure 2% of our money's in Enron in and then nobody ever finds out that Enron is just a big fraud right because no one's actually doing any work so look in a world where everybody's active everybody's doing research and everybody's doing research to uncover bat you know underpriced and overpriced securities best for you just to kick back and say you guys are you know did a great job thank you very much for the research I'm just going to freeride on all of it with my index fund and thanks thanks for everything I'm paying my 1/10 of 1% management fee we're good but if everybody does that then it pays to be an active managed active investor uncover under price overpriced assets so this is frequency dependent this is a hawk and dove game and so in order to know what your what is the optimal strategy you need to know what the other guys are doing what's the distribution of investment philosophies and strategies within this market so that's why professional investors spend at least as much time trying to figure out what the other investors are doing as they do focusing on right the assets that they invest in yes of course it's unknowable in the same way that it's unknowable what the heck Gordon thinks about you over on the other side of the um you know Sand Hill Road or it's impossible to know right what the other drivers are doing so and especially because they don't necessarily want you to know what they're doing that makes it even more difficult have you ever tried to pry out of you know hedge fund manager what they're doing I mean it's it's it's very difficult unless of course they want you to know in which case you can't rely on what they're telling you okay so look in order to know whether or not you're going to flourish you need to know what the ecosystem looks like so you know if you're a dodo bird you're like am I going to flourish in in New Zealand well there's no humans yeah but if there's humans and right you're going to be the original you know McNugget the dodo McNugget okay so we actually have some tools that we can use to explore the interaction of different strategies so this is this is something which is called agent-based modeling and I'm going to show you a very simple version of this so here we have hawks and doves we start off with a population that's equally distributed between hawks and doves and then we watch them go at it and you see that the Hawks pretty much put the doves out of business when fighting is cheap and free so now let's increase the cost of fighting make fighting really painful we'll start the whole thing all over again and now watch what happens we get we get a nice steady interaction right stable equilibrium between hawks and doves then we could throw in another strategy just for fun let's call this retaliate errs so retaliate errs are people who they retaliate against the Hawks but that costs them something now what happens the retaliate errs will initially keep the keep the the Hawks wait let me start with the cost down okay but with retaliate errs what happens initially is the retaliate errs do really well and they keep the the Hawks at bay okay well this time it happened one way but then what happens is the retaliate errs because they're actually incurring lots of costs start losing ground to the doves who don't pay for this and then we could actually inquiry could the Hawks are extinct here but we could introduce a mutation rate and if we introduce a mutation rate then the Hawks will come back to life and put everybody out of business okay so we can play with the mutation rate we can also introduce learning so learning would be imitate the other guy and then we see other dynamics so well if you want to know what an optimal investment strategy is you need to know what the other guys are doing okay so now when we go back to a bubble for instance what's the optimal investment strategy in a bubble well before we get that what about has anybody seen this there's something that happened recently with there was a news report that was false that Obama was that the White House was and there was a big flash crash so look what happens with this flash crash this is United Airlines there was a Yahoo News error which said that United was going through bankruptcy and was just it was a coding error in the news and look what happened to United stock now why did this happen because you have a program trade which is reading the news and it sees bankruptcy United and then in order to be the first one to sell it initiates a sell so it goes down by a little bit now that's not what caused it because then there are other programs that say if the price goes down by 2% sell now then you got a bunch of other guys who are saying if it goes down by 5% so you'll know when to go down by 20% Sal and so what happens each of these programs cause the other programs to kick in and the stock price plummets and then what happens we'll only got some other guys if the stock price go down by 60% by I think oh no one says if it goes up by you know ten percent in five minutes by and then it goes to two before any of you Minh beings even figured anything out right now you're back to square one okay so so what happened here what's what's going on if you this would not happen with a different universe of programs right so it's only purely the universe of programs that caused this to behave this way so before you put in place a program which says you know sell if it's gone down by 20% you need to know well are there other guys who are going to say you know Sal Sal Sal Sal Sal if it goes down by 20% because then you know you're going to lose even more right so you need to know what the other guys are doing so I have this experiment I'm not going to do so let's go back to the dot-com bubble and just say well what did the hedge funds do the hedge funds are supposed to be the smart money it's the idiot retail investor who bought pets.com and the hedge funds the smart guys they're the ones that are supposed to be sitting on the opposite side of the trade saying oh man this thing's overpriced I'm selling turns out the hedge funds we're along the bubble the smart money was long the bubble not just the retail dummies now are they being stupid and getting sucked into the same idiocy of the retail investor or are they savvy saying I'm going to surf on the sea of irrationality out there and get out in time well so these are two of the most famous hedge fund managers in history George Soros and julian robertson and they had when they were they experienced the bubble together and had alternative philosophies julian robertson said the bubble is ridiculous I'm going to short the bubble and and George Soros said now you know I'm riding the bubble I don't care about fundamentals all I care about is where is the market going and what happened well Julian Robertson who is probably the most famous hedge fund manager in history lost more money for his investors in 1999 then he had made throughout his entire career up until that point of course he still has a few billion himself but you know what happened he was wrong he was right tech was overpriced but that didn't prevent it from continuing to go up even further so as a fundamental investor he lost his shirt because he wasn't thinking about the behavior of others Soros is still in business he's still around he's still doing fine so what's happening people tend to write people tend to look around and see what other people are doing it's called hurting okay if the markets going up they plow money into the market the markets going down they pull money out of the market and so as a result investors do very poorly they don't understand their behavior they don't understand why they do what they do right they're not thinking about the ecology of strategies and the way a professional investor thinks about the ecology of strategies and so I'll conclude by pointing out this is a important debate that's arisen recently about the whether or not hedge funds actually add any value and you know if you look at the industry numbers you see that hedge funds have done better than stocks bonds commodities with a lower standard deviation pretty much then most of them they've got the the best Sharpe ratio which says hedge funds are the best asset class that you can invest in essentially but what this ignores is the fact that all right this is all time weighted if you look at dollar weight when did people invest in the hedge funds when the edge funds were going up and up and up and up and up there really weren't all that many people invested until you know 2006 2007 or B's like me I gotta get into hedge funds those hedge funds are doing great so he puts all their money in the hedge funds and then you have 2008 so the net result is regardless of these numbers here more money has been lost by investors and hedge funds that has been made okay and so there's this book that went out a couple couple of two years ago called the hedge fund Mirage and it basically says it if you took all the money that was invested in hedge funds right in the last 20 years and you put it in t-bills you would have done twice as well twice as good okay so now that the problem so the hedge fund industry responded by saying well look that's not our fault you know we do well it's the investors fault for investing at the wrong time right we're not going to say no to money when when everything is overpriced if you give it to us we will take it and so really the problem is one of behaviors so the key question that I keep coming back to in my behavioural finance class and also my game theory classes you know when does it pay to stick with the herd right when does it pay to do what everybody else is doing and when does it pay to do the opposite of what everybody else is doing when does it pay to stick with the herd and when does it pay to deviate from the herd and I think that's a big part of behavioral finance and I really enjoyed talking about it talking to investors talking to investment advisors talking to hedge fund managers talking to corporations you know talking to pension funds about these issues and questions and if anybody's in that industry you know and wants to talk about it with me I'm really glad to talk about it and my email is you know available on the website and I think our time is up but if if any of you guys have questions I don't know we started a few minutes late do we have time for one question or do you want to just call it call it quit what do you think question no I know but you want to bring up a question so there's two herds I see behaving interesting right now one is around Apple and one is around gold have any thoughts about this huh that's a good question so yeah I think I think there's definite Art of it is brought on by the ease of investing in gold right so the creation of gold ETF another very easy you know easy to invest in vehicles I have some other slides on john paulson in particular who I you know admire for his greatest trade ever and he has done really poorly ever since you know he lost a ton of money in financials in 2011 you know he lost its I know forest and help I mean the guy who's done who's a research-based guy like not doing his research on that and then you know gold guys lost a you know almost a billion dollars I think on on gold which is a rounding error for him of course but yeah so the gold um I'm very skeptical Apple well you know I would have told you when it was at 700 that I think yeah there's an Apple bubble going on but right now I really don't know enough about the the fundamentals to know anything but I do know is that cus products like that companies like that have a lot of retail interest you know Peter Lynch once said you know trade what you know and so if you use an Apple you buy an Apple right if you use if you have a if your pacbell customer right back in the day you would buy back Bell and not BellSouth and so you know that there's some value to that where people will inject their information the problem is that then you know these retail investors who understand a product but don't understand the financials can cause a lot of ala Atilla t in the price which cannot be arbitrage dawei by the by the institution's so so i think you know you have to be very careful about some of these you've brought these companies that everyone know thinks they know something about that the prices may not reflect the the fundamentals very accurately yeah last question oh that's a good question so I actually asked a hedge fund friend of mine recently about you know how he says when the bloom is off the Rose that's when you sell it's like can I turn that into a program trade it's like no I'm not telling you so yeah I mean I think that that's a very you know I don't know if I knew the exact way to figure that out then I would be a billionaire but but I do think that when with bubbles when you start running into people in the street taxicab drivers right well I guess they're not in the street during the car but you know taxicab drivers and you know ordinary folks when you're you know when you're your nephew or grandmother or whatever thinks oh I've found the best way to make money that's the time to get out okay so when people start flipping apartments in Las Vegas you don't know anything about never been to Las Vegas that's time to get out of Las Vegas real estate etrade actually had a really funny commercial back in 1999 because there's a quote from this this economist back in the 30s he said when your beautician their hairdresser tells you what stocks to buy that's the time to get out of the market and so in etrade 1999 they actually had a television ad where this businessperson goes in to get his hair done and the guy's like you got to buy this stock he's like great he runs back to his office and buys it on etrade and I was like whoever is that that ad agency has a tremendous sense of humor and they're basically sending us a secret coded message each rate is to get out of the market so yeah I think that's it when people are overconfident and I think that there's no other way that it can go so you go back to real estate 2006 talk to people they say real estate can't go down you know and you ask them what do you think is gonna happen your real estate in the next 10 years 9% a year it's going to go up that's the time to get out okay when everybody agrees on something that's when you've got a problem I think I have time for a 30 second question do do bubble is there an anti bubble does it work to the gas I fear so papal finances sometimes called fear and greed economics of fear and greed and so when fear is ramped to the landscape when everybody in 2009 is thinking I've got to get out of my portfolio because I can't afford to lose any more that's the time to get in so again when everybody is saying we're going to hell in a handbasket that's the time to buy generally but of course you know sometimes things are down for a reason so you really have to know the fundamentals first and then look at what you think the price is relative to the fundamentals so you really can never abandon pure fundamental research as to keep you grounded okay all right thank you so much it's been great and

Channel: Berkeley Haas Alumni Network

Views: 94,656

Rating: 4.9077811 out of 5

Keywords:

Id: ERDEozGpV74

Length: 73min 14sec (4394 seconds)

Published: Tue May 14 2013