A Brief History of the Efficient Market Hypothesis

on behalf of the American Finance Association the University of Chicago graduate school business and the myron scholes forms it's an honor and a pleasure to introduce gene Palmer this talk is being videotaped for the American Finance Association history project so we speak for the ages Gene's going to tell us how the efficient market hypothesis developed so I'd like to say a few words about why it's so important this may not be obvious to young people in the audience and I'm sure gene is going to be too modest to say much about it market efficiency means that asset prices incorporate available information about values why should they reflect information because of competition and free entry if we could easily predict that stock prices will rise or I should say maybe decline tomorrow we'd all try to buy yourself today prices change until they reflect that information now this seems like a pretty simple theory hardly worth all the big fuss maybe you were expecting general relativity with lots of incomprehensible equations the efficient market hypothesis is more like evolutionary evolution and efficient markets are elegant simple and powerful ideas that organized and energized vast empirical projects and that's the true measure of any theory without the efficient market hypothesis empirical finance would just be a collection of Wall Street anecdotes how I got rich stories and technical trading new sheets the efficient market theory and empirical work are also a much deeper intellectual achievement than my little story suggests it took nearly a century to figure out the basic prediction of an efficient market from Bashi Lea's random walk to the consumption Euler equation it took a lot of hard work and great insight to account for risk premiums selection biases reverse causality endogenous variables and to develop the Associated statistical procedures efficient markets empirical work doesn't just check off easy predictions it typically tackles really tough anomalies each one of which looks superficially like a glaring violation of efficiency and each of which is endorsed by a cheering crowd of rich or maybe lucky traders that empirical work consists fundamentally of applying scientific method to financial markets modern medicine doesn't ask old people for the secrets of their health it does double-blind clinical trials and to that we owe a lot of our health modern empirical finance doesn't ask warren buffett to share his pearls of investment wisdom we study a survivor bias free sample of funds sorted on some ex ante characteristic and we correct for exposure to systematic risk to that we owe our wisdom and maybe as a society a lot of wealth as well this point is really important now a period of great financial turbulence it's very easy to look at the latest market gyrations say oh look at that surely markets can't be efficient that's not how we learn anything of lasting usefulness efficient markets taught us to evaluate theories by their rejectable predictions to do the numbers to do real scientific empirical work and not just to read newspapers and tell stories efficient markets are also important to the world at large the assurance that market prices are basically right lies behind many of the enormous changes we've seen in the financial and related worlds from index funds to market to market accounting and to modern risk management well with 40 years hindsight our markets efficient no not always and Jean said so in 1970 for example prices rise on the release of inside information that means that information was not already incorporated in the price this is great news only a theory that can be proved wrong has any content at all in the first place theories that can explain anything are just as useless as prices went down because the gods are angry Jean went on he argued that no markets ever perfectly efficient since no markets ever perfectly competitive or frictionless the empirical question has always been to what degree a given phenomenon approaches this unattainable ideal still the answer today I think is a lot closer to yes than no in the vast majority of empirical investigations it's certainly a lot closer to yes than anyone inspected in the 1960s or that the vast majority of practitioners believe today there are strange fish in the water but even the most troublesome are surprised a small fry however empirical finance really isn't devoted to debating efficient markets anymore and any more than modern biology is debating evolution we've moved on to other things I think of most current research as exploring the amazing variety and the subtle economics of risk premiums focusing on the joint hypothesis part rather than the informational efficiency part of genes 1970s a this is also great news healthy fields settled debates with evidence and move on to new discoveries now don't conclude from that that efficient markets are passe or that finance is boring as if it as evolution lies quietly behind the explosion in modern genetics markets that are broadly efficient in which prices quickly react to most information quietly underlie all the interesting things we do today and this is the best faith that any Theory can hope to that can aspire to now gene we'll talk about the history of efficient markets people expect the wrong things of history as they expect overly complex theory no lone genius ever thought up a hypothesis went out to test it convinced the world with his two point one T statistic and retired theory and empirical work developed together ideas bounced back and forth between many people the list of salient versus unimportant facts shifts evidence argument and alas age gradually change people's minds occasionally one great mind puts it all together this is how efficient markets developed - and as jean has always graciously acknowledged genes - essays describe the ideas but much less of this process it was an amazing adventure and historians of science should really love this story ladies and gentlemen please welcome gene fama to tell us I'm not going to I'm not going to talk that long and I'm going to talk very informally I'm not good at giving formal talks and as John said I'm going to give you more about the history of the development of this hypothesis and the history during my lifetime because it goes back my academic lifetime because it goes back much further from then that and I'm going at a note I got from Bob Lucas and said he couldn't be here you would like to be here but he said don't do the easy thing and tell him about your current research tell him about what was going on when this when we were students and then young faculty members around the University of Chicago and that that's what I intend to do and let me say thanks to the FDA for doing this project and hopefully somebody will look at this thing when you put it up there okay so we going back to basically the middle 1950s is when empirical work on efficient markets starts theoretical work you can trace back to bachcha area there's a website that John sent me where this fellow traces it back to 1538 and then the first person who ever worked on Brownian motion that's not what what I'm all about here so you can go to that website and look at all these obscure things so all it says is ideas have very deep roots and getting them all out as it's a very complicated business but I'm only going to tell you about my history in this area so the second thing I'm going to give you is oh just a little brief perspective on the context in which this work was done basically finance didn't exist in the middle 1950s it really didn't exist until work that started with Markowitz is his doctoral thesis in the economics department here and then that masterful book that he wrote in 1959 which was the first really rigorous definition of risk of securities and risks of of portfolios but basically at the time important historical fact is that basically at the time when research on efficient markets started there was no asset pricing theory around asset pricing didn't exist the first asset pricing models we had were what was the cap M which came in 64 and 65 and then all the stuff in the in the 70s ending up with the breeding Lucas consumption based asset pricing model but none of this was available and if you want to see a good indication of it go back and read the original M&M papers on capital structure and examine how they struggled with the definition of perfect substitutes and they Beit they come down in the end to something they call the risk class which covers everything basically cash flows has to be perfectly correlated for everything in the risk class and that covers every asset pricing model you can ever think of but it's a it's a very narrow definition of a risk class but the point is that the basic problem or the basic link between asset pricing and efficient markets just wasn't wasn't understood so how did it happen well Morris Kendall and Harry Roberts and people like that were basically statisticians and computers came along and this basically expanded their horizons normally even though the computers were were toys not toys physically because they filled whole buildings but would tour is relative to modern computers but they were released from their from their mechanical calculators and able to do calculations that they couldn't imagine doing before and what's the most easily available data securities prices so there were lots of work looking at distribution of securities returns and the behavior of the serial behavior of of security returns and people observed that if you calculate it out of correlations of stock returns they tended to be very close to zero and then they started thinking well what does this mean what this is me in a world where markets are operating properly well then book economists came along and said if you think about how Margaret should work basically the current price should reflect all available information statement of an efficient market so prices should change only based on unexpected new information and that led to the that led to the random walk model of of pricing and all of this all of this was going on at least in my perception of it it was all taking place here and and at at MIT here we had Merton Miller and Harry Roberts and Lestat Elsa and when I was a kid been my Mandal grows to come quite frequently spent several quarters here and there was just lots of talking about these topics and then at MIT there was a Franco Modigliani Paul Samuelson Paul Kutner Sydney Alexander people that were very interested in in this topic and there was lots of back and forth between those two campuses at that time so I explained the the idea that prices adjust to new information that gives rise to this random walk hypothesis the random walk hypothesis well people quickly figured out that that was a little a little bit too strong prices don't have to be independent identically distributed basically they have to be a fair game which means there there are there they're martingales and so Samuelson and Mandal bro both wrote papers showing that okay random walks weren't what you need to really needed a martingale but there was something missing in all of this so people didn't even understand when they were doing the tests that they were saying something about market equilibrium so when I calculate an autocorrelation coefficient if I think about it in terms of the joint hypothesis problem and what what does it mean what's my implicit assumption about market equilibrium it's basically that the security or the portfolio that I'm looking at has a constant expected return so that it's it's-it's returns are unpredictable based on on past returns and then there was lots of working nobody does this anymore but there was I even wrote a paper with the Marshall boom on filter rules there was lots of papers on filter rules if you develop a trading rule and you test it and the trading rules typically involved going long going short and the presumption was that there was persistence in in prices and the market efficiency hypothesis was that if the market was efficient basically buy-and-hold should be the best thing to do and now if you think of that in equilibrium context what it says is expected returns are positive so you don't want to be ever be out of the out of the security now constant expected returns positive expected return all terribly arbitrary assumptions but they were implicit in this in this in this early work and in 1970 I wrote this reason I wanted to untuk to I've written three review papers let's top it believe it or not on this topic and I look back at their money someday they're pretty good actually I don't know how much more I know much more to say in addition to that so maybe I just give them a little flavor and then people can go back and read them so this paper this 1970 paper called efficient markets review of theory and empirical work was in the Journal of Finance basically points out that you're you're always jointly testing market efficiency with some model of market equilibrium which means that market efficiency puts per se is not a testable proposition now that that doesn't bother me because the converse is also true and that bug is nobody models of market equilibrium asset pricing models aren't testable without market efficiency because basically all of them assume market efficiency so these these two things are now inextricably linked in I think everybody realizes that now and as John said much of the interesting stuff that happens over the last thirty years has to do with testing asset pricing models and then deciding whether they were I don't work because the model is better because the markets in inefficient okay now in nineteen might forget the date exactly 68 or 69 we published this paper the adjustment of stock prices to new information from official Jensen enroll which has an interesting history so I'll give you the loo of the history the history is Jim Laurie had gone to Merrill Lynch or he had good friends at Merrill Lynch and they wanted to know what the stock return was from 1926 to 1960 and so they gave him two hundred fifty thousand dollars to develop the crisp tapes but basically only NYSC stocks at that time so he Larry Fisher did did all the work Jim Jim Jim didn't do it Larry was amazingly foresight 'fl in terms of figuring out everything you had to know to construct true returns so they nineteen sixty two or three whatever they published the number in Merrill Lynch takes out a full-page adil ad in a Wall Street Journal saying here's the return on stocks 26 to 260 it's great give me your money so so but at that point Jim said God he got worried he said I'd like to get more money from them but they are going to give us more money unless they think these tapes are useful for something else we have to freak to unhide set of course this is laughable because maybe 30% of all empirical work in finance uses that the the crisp tapes but he was worried that nobody would ever use these things so he said why don't you study stocks with so that that that's that's what we did and this paper gave rise to I mean it would have been done anyway because the the data are there and this is something that had to have happened we just happened to be the first ones to to to do it but basically this launched an industry the event study industry in in finance I always think we've run out of events to study and then another 200 events that he's getting again get published so it isn't it isn't true event studies initially you know that like every other test they're subject to the joint hypothesis problem initially this wasn't a big deal because people were interested in what happened within a few days around the event and if I'm looking at fairly big events mergers whatever if I'm looking only at a few days around the event it doesn't matter a lot what I call the expected return because it's typically tiny relative to the event return but once I start getting interested in long term adjustment of prices to information now the bad model problem of a joint hypothesis problem starts to rear its ugly head again and we've seen a resurgence of work on our work on long term returns has become very popular and the topic is incredibly interesting of course you want to know whether the long term adjustment of prices to information is is efficient but here the join hypothesis problem just becomes devastating because what we've learned is that these tests are just very sensitive to what you assume about market equilibrium so if the questions are very interesting but I'm not sure the the answers can never be terribly convincing about about long-term returns now I should also mention what was that my Jensen's thesis on the on the evaluation of pension fund a mutual fund performance this was just you know we had out like a run of incredible feces at this time and this is obviously one of the best ones but this launched an industry of its own which is performance evaluation now in perspective performance evaluation didn't exist it did not exist prior to this this time all mutual funds were actively managed and they all claimed they'd be whatever you wanted whatever they wanted to beat and there were there was a nobody ever stepped back and evaluated whether they actually did what they claimed to do and Jensen set up this framework in the context of the cap BAM that was his model of market equilibrium and you know I dropped a huge challenge on their heads because it didn't look like they performed very well now Mike was one of an amazing group of students that we've had here over the years so at his time we had Richard role there was Jensen it was Myron Scholes ray ball Russ watts I have to be forgetting people uh Marshall Blum and then after them or they they keep coming coming and coming along but this group had a big impact on research on on market efficiency especially in the in the early day so I get a lot of credit for the stuff that they actually that they actually actually did but one of the offshoots of the crisp seminar was we had we used to have do this crisper seminar biannually which also was the generate interest in the in the crisp tapes but the participants were business people and what one fellow who came all the time was Mac maquon from Wells Fargo he was in charge of their investment business and he was really a attuned to this efficient markets idea and Rex Sinquefield was one of my early students in my Advanced Course in MBA steam even off to American national he was totally convinced in the two of them and their separate institutions brought out the first index funds which was basically along with Jensen's performance evaluation the first applications of efficient markets now in those days that took a long time for ideas to penetrate into the into applications but one of the big finance I I had this in my first talking to fast but my first slide I had this notion that finance is the most successful area of economics and I think that that's clearly true in what sense finance is the most successful area of economics on the academic level because of the joining of theory with empirical work we test our theories and we develop meaningful tests of our theories and when the theories don't work we go on to other theories this is the best sign and the best scientific tradition but there's also been a huge spillover of Finance into applications for example I right before they got the Nobel Prize I always said the options pricing model was clearly the most important paper of the century the last century in the sense that everybody going through a ph.d program or an MBA program and had to had to in finance had to know about it every PhD economist had to know about it in the whole industry was was was was built around it so it's it's hot to imagine anything that had a bigger impact on the world now efficient markets less so because even to this day less than 20% of mutual funds for example are managed in a passive way the rest and and that's way up from ten years from ten years ago so that hasn't really penetrated as much as I would have hoped okay so that that takes us basically through the period when I was interested in direct tests once I got clued into the joint hypothesis problem I moved on to a different different kind of test and that's this market predictions of interest rates inflation and other macroeconomic variables so this is like turning it around saying markets are efficient let's see what they can predict and so the first paper I did in this tradition was interest rates as predictors of inflation so paper the AR in 1975 which basically says that during the period from 53 to 71 I call this the Golden Age of the Fisher effect it's the only period in which it it it works during that period expected real returns on bills were relatively constant and so the variation the bill rate was very basically variation and the expected inflation rate and that when you regress the inflation rate on the bill right you got slopes very close to one and residuals that were serially uncorrelated so at least to that extent it looked like you're looking at the best possible assessment of inflation now that the simple idea in that paper is that if you want to estimate a conditional expected value you put that variable on the left and you put the predetermined variables on the right it's it's that everybody does that now but at that time people were doing it the other way around so they thought that expected inflation caused interest rates so they would regressed interest rates on inflation and of course immediately they have a huge measurement error problem in the in the in the regression now this is this is a characteristic of my work like I came on this idea serendipitously I mean I don't know why I came on it and then I applied it like 15 different times so there there are literally fifteen papers that use the same approach some of them are about the information the term structure about expected spot rates and holding period returns about forward exchange rates as predictors of spot exchange rates and then there are tests of whether how well stock returns forecast real activity and all kinds of things like that so that that took me to that took me another another ten years and then we come to the time when Ken French comes here and he has the office next to me and we start working together and I figure here's a guy that's got incredible energy he can keep my reputation going for another 25 years so I hooked up with him incredibly smart guys same work habits I do and we just work together very well and the end result not the end result was still writing papers but one of the big results was this this paper that we wrote in 1992 called the cross section respect expected stock returns which I never expected to get published because everyone does nothing new in yet there's absolutely nothing doing it all we did was to take previous so-called anomalies and put them all all together in one place so what had happened well the cat bam came along in 1965 people didn't start testing it until the the early nineteen seventy so there were papers by following Macbeth black Jensen Scholes other people had tested the cap amp and the initial tests looked like they gave pretty good support to the model and then kind of in the earlier 1980s we started to come across things that were causing embarrassments for the model but they came up one at a time one of the time these things came up and so people looked at the morn at a time and said I have one thing that's not too bad so all we did was put them all together and say look this thing just doesn't work and then came this the so-called three factor model which was meant basically to capture the the problems that kept them seem to be having with small stocks and with explaining the difference between value and growth stock returns now this this is a good example of the a conundrum posed by the joint hypothesis problem so we developed this model to explain the size and value growth phenomenon but then immediately a controversy arises Lacan oh shucks like for Vishnu say in others say nah this isn't this isn't a risk-return story this is just market inefficiency and basically the joint hypothesis problem says based on that data alone are based on any data I've seen since then you really can't tell one story from another okay I was going to say something about my research philosopher but I'll skip that I'll say it why not what's my research philosophy keep it simple keep it simple do things that people can understand that that's easy for me because I'm a simple-minded guy if I can understand it almost anybody can understand it and my good friend David Booth told me 20 years ago every business is 50% marketing and academic business is not an exception it's at least 50% marketing the way you write has as big an impact on your impact in the world as a whole as what you actually do so saying things clearly and simply it's going to get you much further the words where did market efficiency efficient markets come from you won't see them in my thesis they're not there you won't see them in any other work up into that point I popped them into this little paper that I wrote for Capital ideas for this for the school that that's the first appearance of them and then that paper got picked up by all these professional journals in the term term caught on but that that's where it developed it didn't exist up until it hadn't been used up until that point so coming up with the name gives me a lot more credit what was actually done than I deserve okay we talked about this we talked about that we talked about that here are all the papers I wrote right don't do there 15 of them right did that okay I'm done