In Pursuit of the Perfect Portfolio: Eugene F. Fama

[Music] hello I'm Andrew low and welcome to the project on in pursuit of the perfect portfolio I'm really thrilled and honored today to be talking with Eugene F Fama the 2013 Nobel Prize winner in economics and the founder of many ideas in modern Finance uh not the least of which is of course the efficient markets hypothesis Jean thank you very much for joining my pleasure so I want to start today by talking a little bit about your background all the way going back to uh the old days uh when you were a high school student

you were you were actually a jock aren't you right still am I see faded jock so what what what sports did you play and what were particularly passionate about yeah um I played baseball for two years we lost in the semi-finals of the state championship football we won the state championship when I was a junior um basketball I was never very good at but track would you believe that I was second in the state in the high jump I heard that that's pretty amazing the only guy that beat me was the first American to jump

over seven feet wow and but he didn't take off his sweatsuit before he beat me he left it on and just still be me so with all of that how did you end up getting into economics oh well I thought I was going to be a high school sports coach and a teacher mhm uh and I was majoring in romance languages of all things and but I was getting sick of that really sick of and I took an economics course as a junior a tough tough yeah and I loved it so and I was pretty

good at it so I stuck with it and was it uh Professor Ernst who played a big role in well he he hired me to work with him it was between my Junior and Senior year um he had a stock market forecasting service and um my job was to come up with trading rules to beat the market and did you yeah I always did they always worked really in Sample never out of sample okay but he always had me keep an out of sample portion of the of the data and I just found out because

I I wanted to see if he was still alive he's not but he was actually a premier athlete a golfer really he's in the BC Boston College Hall of Fame which is really good ab as a golfer huh and so what made you decide to go to grad school then oh because I didn't want to work for a living um uh I figured it that becoming an academic was a way to keep my sports life going and set my own hours so I could uh keep keep it all together and it worked out quite well

actually I'll say and so Chicago was Chicago I'll tell you a story about how I came to Chicago I applied because my professor at toughs all of whom were Harvard econ graduates and I said I want to go to business school not to an economics department they said well go to Chicago cuz that's the only one that has a an academic orientation they didn't mention MIT by the way I remember long time uh so I applied here and I never heard anything and comes April I call and the school's very small at that point so

the Dina students answers the phone Dina students doesn't even have a phone anymore h doesn't talk to students but anyway he picked up the phone I'm chatting with him and he said well what I said what happened to my application he said we have no record of your application and he said well what kind of grades do you have and I said pretty much all a and he said well we just happened to have a scholarship for somebody from tough do you want it wow that's amazing that's Serendipity and without that phone call that would

have been very different right because i' had already been accepted to lots of other places yeah um but we came here but Chicago is the place for you yeah okay and so at Chicago who did you end up working with and who are the most formative for you looking back at that time this is long before your time the finance courses were ridiculous I mean didn't pay to take any finance courses so I took all my courses in most of my courses in the economics department and uh when it came time to write a thesis

um you know we were in the early 1960s first computers are coming about so everybody is getting interested in doing empirical work yeah and uh here was Mt Miller and Lester teler and Beno mandelo came around quite frequently uh and that was a way to get these people Lester tell are interested in working with you I see and I had two kids so I wanted to get out as fast as possible so I thought doing something in this area would be would would have that effect and it did yeah so in fact I think you

were among the first to actually use computers to study the stock market I I was using it and there was a guy in the physics department using it at night during the night because very limited capacity yeah and we were the only ones and we'd call IBM and say this compiler is not working it's doing this and they laugh so let me guess after the second time they didn't laugh anymore right were you programming in Fortran at the time yes right okay and allocating 1,000 memory spaces okay that's a joke so your thesis was on

the behavior of stock market prices ultimately published in journal business but you found that the random walk was actually a pretty good description right at that point people hadn't really come around to talking you won't see the word officient markets in that paper you won't see it before that um people are kind of groping for what it mean meant if so first of was statisticians like Harry Roberts and people like that they were interested and then when Lester telser and the people at MIT Paul Samuelson MH coer all those people got interested in it they

started asking well how would we expect these things help prices to behave if markets were operating properly and that was kind of the seeds of the fishing markets hypothesis if you go back to that point yeah we didn't have any models of market equilibrium right so they weren't aware of the joint hypothesis yeah problem they would just propose a test that if the Market's efficient price prices will be a random walk or something like that yeah well in fact I I think the last paragraph of your thesis actually called for more theoretical models try to

explain some of these asset pricing anomalies so it's very interesting and timely that it sort of happened at the same time that now the capm started developing years later yeah exactly two years later so out of that thesis what came from your mind in terms of how it shaped the way you thought about efficient markets was that a particularly formative experience looking at the data because many people think that theory came first but in fact it seemed like in your case it was a data no no there was groping for Theory at that point right

we're groping around for Theory I don't think it got codified until the paper that I had in the Journal of Finance in 1971 y uh where that's where the joint hypothesis problem got was was stated um the the word efficient markets appears in a little um occasional paper that I wrote for the the school here and then it got republished in uh financial analyst Journal okay so that was the first place that I used that and it caught on right so I mean to to say it caught on is a bit of an understatement because

it's kind of a bad choice though why do you say that oh because it gets mixed with eff fishing portfolios you know right it's a it's a I don't know what an alternative yeah well but I mean the idea that prices fully reflect all available information that also came from you that I think had a huge impact and one of the things that I don't think people realize is what the state of the practice of Finance was at the time so how did your Theory go across the practitioners what was what was Wall Street like

at the time Wall Street at the time there were no standards about you know there were very few mutual funds like things around and you were free to say anything you wanted about how what your performance uh looked like right um and this was a direct challenge to say look let's start measuring and when we had the capam came along a couple years later and Mike Jensen wrote his thesis on you know the performance of mutual funds that kind of lit the bomb really right in the sense that now you couldn't get away without without

actually testing how how well you were you were doing or somebody else was going to do it for sure yeah so that that started the whole performance evaluation uh business which goes on to this day I mean right you always think you've you've seen enough mutual fund papers but then they'll be five more next year Well in fact now it's a standard of the industry that you have to measure performance in the way that you outlined originally but at the time I would imagine it was pretty controversial you were not well- loved on Wall Street

at that time no still aren't actually well I don't know we could talk about that we're g to get to that um but it seems like that was a really big change in the way that academics were actually having to have an impact on the industry it was there was no impact before that right right there was no there was nothing M there was nothing to go back and forth with right in uh in asset pricing or portfolio marwitz didn't penetrate yeah when I joined the faculty here nobody teaching Investments was teaching portfolio Theory right

this is 1963 martz's thesis here was 53 but nobody was teaching it right and I I went to meron Miller and said what should I teach he said we hired you to teach the new stuff you're supposed to teach stuff we on teaching he wasn't he was teaching Corporate Finance but uh so I started I just took Mark wiz's book and handed it to the students and said this is what we do yeah so you know you're known for empirical work and we're going to continue on with that in a minute but I actually want

to take a little detour about Theory okay because you actually wrote a textbook called the theory of Finance with mert Miller it's a very popular was a very popular book at the time now there are other books uh but uh still when it first came out there was nothing else like it and I actually looked at it when I was a grat student it's some really hard and deep theory for example you've got Dynamic optimization in that book right and that was a really surprise that you know somebody with that empirical background the reason I'm

laughing is you know how we I wrote that book no we were using it we wrote it in the process of teaching an introductory Finance course to MBA students introductory to mbas that was it must have had some mbas well I'm not sure how many of them ever really got through it but that's what we wrote it for that's how our perceptions of the market was so poor but it actually ended up being a very influential book for graduate students many many years right so getting back to empirical work now so your your theory about

efficient markets obviously launched a whole host of papers ultimately that really changed the industry and the way it worked how was your interaction with the industry at the time did you have any interaction or really was it just the force of the ideas that caused the change and how people started looking at markets I would have a periodic interaction so Paul Samuelson for example was on the board of Tia cred mhm and he had me go and talk to them about this stuff because they were just you know kneep and active stuff they still are

actually but um he wanted me to talk to them about what this new research imployed for how they should do business because he wasn't having any effect on right right and I did and I didn't have any effect on them either and that was kind of the experience uniformly there there after so I I totally lost interest in talking to uh m applied people because you know after a while I came to the realization they don't want to do this because it's gonna kill their fees why would you want to go from 1% down to

fractions of several basis points to as your management fee they're not going to volunteer to do that right yeah so at the same time you also mentioned that the capm had been developed and people are starting to use that for performance evaluation um you obviously played a big role in that because the paper the F McBeth approach to studying these Factor models actually had a very very important role in getting people to understand uh uh can you tell us a little bit about how you came to start working on that oh well so this is

it was published in 73 but obviously you go back way before that and uh black Jensen sches had written a still famous paper on testing the the capam and I officially was here at the time M and he would come in every morning early 7:00 and I'd be working there and we would argue and I kept saying to him Fisher that thing that you're doing in that black Jensen shills paper you're just running a cross-section regression that's all it is they had this complicated portfolio approach um and I said you're just running a regression there

and he said no I'm not yes I am no I'm not so finally I wrote that chapter of the book to prove to him that all it was was a cross-section regression and then I said well why not write the rest of the book so so uh that was how that arose and but that uh that paper uh became kind of the founding paper in cross-section regression approach to testing asset pricing models M it was amazing because that approach is still used today is an incredibly useful way of thinking about selection bias being careful to

make sure that these measurement errors don't get caught up in the way you construct the portfolios well I I think it has applications more broadly in terms of um you know very popular in economics to run panel regressions right but that's a particular weighting of the data yeah uh in which you rate rate observations weight observations equally whereas in the F MC Beth approach you weight periods uh equally and you should ask yourself which way makes more sense right and I think the standard errors that you get out of the F Beth approach are much

more much simpler to interpret yeah then what you get out of uh the panel regression approach when you put through two pass robustness uh right checks on it which are large sample properties whereas the properties in the F best stuff are small sample properties yeah I mean you can certainly interpret it much more easily it's a lot more intuitive simpler right yeah so that actually led to I think your F French three Factor model the idea that actually it's not just the capm it's something else going on is that right well or how did you

come to that um if you go back in time so the capam had a 20-year run MH uh basically and then like all models I mean so-called anomalies start to to come across the first one was rol Bond's thesis on the small stock effect and then we had uh you know leverage and all other kinds of things jary effect yeah lot lots of things came out so we wrote this paper in 90 2 which was the cross-section of expected stock returns in which we just pulled all this stuff to together and I didn't think that

paper was a big deal actually I said there's nothing really new in here the three- Factor model no that one didn't have the three- Factor model that was just saying but the 93 paper did yeah then the 93 paper that was different yeah um the 93 paper was you know when we the 92 paper basically said there are all kinds of anomalies you can't put them aside anymore and more important not more important but I mean in addition yeah the central prediction of the capam just has never worked this the relation between average return and

beta has always been too flat right um so that became known as the beta is dead paper right and the New York Times actually quoted that phrase bet is dead that's that's kind of that's not the right way to characterize it though the right way to characterize it is there are too many other things that help explain average return so that even if you had a strong positive ation between average return and beta you still have this problem that there are lots of other things that seem to be able to capture variation in average returns

that that model doesn't get now you and Ken have had a very long and productive collaboration for many years where's the state-ofthe-art now in the fmer French thinking what's current uh perspective so we based on that paper the 92 cross-section of expected stock returns we came out with this so-called three Factor uh model in which we added every model of asset price every asset pricing model basically says the market portfolio is the core right and you start with that and then you know the capm is the simplest version in which that's the only uh portfolio

you have to consider and then you have Merton's extensions in which you have lots of other portfolios that are possible candidates hopefully attached to State variables um and we kind of framed ours in terms of of that model although that that's really a stretch because we didn't ident ify any right any state variables so I've come to the opinion that it's really what I call an exercise in empirical asset pricing in the sense that none of our theoretical models work yeah the the most fundamental theoretical model is the consumption cap bam right it's awful right

exactly that doesn't explain much of the data you're you're kind uh so that one's kind of dead in the water the capm is kind of dead in the water so I've kind of come to the opinion that maybe what we're doing is finding a set of portfolios that span the mean variance efficient tangency portfolio and for a while maybe that's the best you can do so you use the characteristics of the data to identify what might be an appropriate model we haven't come to the end of that because you want to limit you know you

want some constraints on that on that process you want to boil down to a small number of so-called factors of portfolios that you use to explain returns yeah and we're not anywhere near the end of that okay or even near the end of Ken and I are working on this right now how you test for model compression right um you know reducing the number of variables that you that you that you look at so I think facing that issue straight out is interesting and important but it's it's basically an empirical orientation there been many attempts

to um identify State variables that might explain these things but the Ken said this from the from the beginning now he said these variables are all just variants of price yeah so if a model doesn't work you expect them to pick up the dimensions of average returns that the model mix misses and it's not because they're related to any particular state variables they're all just linear combinations of different state variables and we don't know how to unscramble them you know right but then how do you prevent the uh inevitable concern of data snooping where oh

oh absolutely yeah real sensitive to that so I I preach robustness okay in other words um I want to see the same phenomenon um so we use for example 63 to whatever it was at that point 90 89 90 yeah uh We've obviously extended it up to date I mean on kid's website he publishes the data regularly it's very helpful uh and we went back and collected the data back to 26 y to see if it worked and then we went out a sample internationally to see if it worked so robustness is the key okay

to all of this stuff because you know we how many of us are out there using the same tapes right you're going to find stuff that's in there whether it's real or not right or whether it's just sample specific so you robustness is my my theme really all right so I want to come back to that when we talk about behavioral anomalies before we get to that I have to talk about your 69 paper the F Fisher gens and Roll paper okay that's a paper that on the surface looks pretty straightforward but it was the

first paper that really looked at stock market reaction to events and information right tell us a little bit about how you came to that how did you come to write that work on that problem okay so the these things are really funny because it's total Serendipity right so the crisp tapes had come out shortly before that right and Jim Lori who was he had solicited money from marily lych to put that together yeah and believe it or not he was afraid that nobody would ever use the tapes really really and it's used by everybody now

practitioners andics right so he he said came to me and said can't you do something with this and and I said well what what's on the tape uh I mean I had just finished my thesis where I Ed my own data right and he said well we've got you know prices and everything else he well we have stock splits that's the only thing that was on there so was stock splits so I said okay we'll do a study of stock spits and uh Jensen and Ro were PhD students at the time so I I hand

it off to them to do the dirty work and then that's how that paper developed so the methodology was something that you had designed just at that moment for that uh particular issue because that method ology is now used not only by all academics but it's actually used in industry I don't know if you realize but you know it's used badly in court cases too exactly in court cases used all the time to calculate damages they don't understand the notion of standard errors right well those are details that judges may not get excited about but

uh right but it's an incredibly influential paper thaty came about well I you know I unabashedly say it gave rise to an industry yeah well at least maybe two or three Industries M Miller this is terrible but mer Miller always said when we had tenure decisions in accounting and most of the papers would be event studies okay that's just event studies but you know I I turn that as dick tlor likes to always have these little I call them anecdotes about when the markets doesn't work don't work and I said okay but there are thousands

of these papers where it seems to work very well yeah all event studies th those are the best I think studies of how well markets adjust in New right the vast majority tell you that markets adjust when the information is revealed and the actual event you don't see much it's exactly what you conjecture because of efficient markets right yeah so now you mentioned dick Thor now only turned to behavioral Finance what are what's your thinking because it seems like behavioral Finance has been at odds with efficient markets but on the other hand they they seem

to be almost opposite sides of the same coin so what what are your thoughts because you obviously hired dick here at chicag right right I was yeah involved with that right so I've started to tease him mhm and saying I'm the most important person in behavioral forance because without me they have nobody to pick on exactly and that I I said to you know 20 years ago I wrote a paper market efficiency long-term returns and behavioral Finance where I said look guys you have to grow up you can't just be Chen you can't just be

complaining about market efficiency all your life you have to come up with something that we can test and reject right and it takes a theory to beat a theory you said right and and a test tal Theory uh and I I I reminded him of that a couple of months ago okay because if you read his recent book yeah misbehaving right same thing right and when I say to him well where is the theory that we can reject that we can call behavoral Finance doesn't exist not yet no well I mean not yet give me

a break okay I'm trying to be kind here okay so now let me turn to more recent work that you've been doing in thinking about how to apply these ideas because obviously your work actually has a lot of practical implications many people take efficient markets as well as the F French 3 factor and some cases five Factor model and put it into practice can you talk a little bit about your interest in practice and maybe some of the work that you've been doing for years now with dimensional fund advisors well um dimensional fund advisors was

started by uh David Boo and Rick singfield who were students uh Booth was one of my research assistants way back when as a PhD student and Rex was in one of my classes Rex was the first one to go off to American Standard as a kid and within a couple years he worked his way up to be the in charge of all their investing he he brought out the first actual index uh fund u based on that efficient markets hypothesis um Wells Faro was doing it pretty much at the same same time but they didn't

come out with an actual Index Fund because myin and fiser always wanted to fancy it up a little bit okay so Rex actually did the first one and then Booth came to me and said I'm starting a company um do you want to be involved and I said sure I've never been involved with the business so I'll be involved okay with it and I've been you know working with them ever since initially all they had was a micro cap fun so the 910 desiles of the of the uh equip ENT NC stock so was Tiny

Port tiny stock portfolio and then I had done a lot of stuff in the 70s on using the structure of forward raids to predict um Returns on longer term uh bonds and they came out with products based on that pretty quickly um and then when the farmer French stuff came out they had clients for that stuff Val So-Cal value stuff before that paper was even published um he brought a guy out one of the clients out here to the university and on my computer screen I showed him the results and he said okay I'll take

20 million of that W and I want a big one and a small one uh he said fine okay and then we went over to the med and had a hamburger but now all of their products are kind of centered on that model you know us products stand International mhm uh products and the business is an enormous business now I mean it's grown phenomenally over the last several even just years never mind the decades it grew right through 2008 2009 right it's very kind of a testimony that uh if people buy into efficient markets they

don't bomb out as easily right as uh people who buy into active investing but jean the ironic thing is that this particular set of portfolios that DFA has constructed has actually managed to out perform a lot of its competitors ah but it hasn't managed to outperform the segments targets all right so you still believe that this is an example of efficient markets at work rather than a counter example no no no it's just I always distinguish between asset pricing and efficient markets those are those are the two what I call the the twins of asset

pricing siamese twins of asset pricing you can't separate them right um and the the risk Return part of it is what they're deal dealing with so I think their products are just riskier they have a little tilts to they have tilts towards value they have straight value portfolios small stock portfolios um and big stock portfolios I mean basically they have value tilted portfolios practically of all of All Sorts yeah it's a fantastic success story because obviously we are sitting in the Chicago Booth School of Business so David has been very generous and has done well

as as uh thanks to you and the ideas that you've developed quite an impact right so uh now we get to the most important part of this interview which is the pursuit of the perfect portfolio in your view what is the perfect portfolio what would you recommend and how far away are we from being able to achieve that for the typical individual investor well I don't think there is a perfect portfolio I think you have least in my current view of the world you have a multi-dimensional surface that's characterized by a Continuum of portfolios with

different sorts of Tils okay and the market portfolio is the center of of that Universe um in an aggregate people have to hold the market portfolio that's that's it and that's an efficient portfolio in any model you want to you want to think of and then you can decide to tilt away from that towards other dimensions that we think capture different kinds of of of risk but and that's a personal decision but you're you're uh as David boo says you want to diversification is your buddy right if you decide to tilt away you want to

do it in the most Diversified way you yeah you can now is there a possibility of over diversifying you know people like Warren Buffett have argued that instead of not putting all your eggs in one basket he prefers to put all your eggs in one basket and watch that basket very carefully yeah so is there a chance that we end up you know somehow and spreading our uh portfolio too thin or no I don't think that's possible okay um you know if you ask Warren Buffett what what people should do with their portfolios he says

go passive right right so that's true and I think he provides an interesting case in the sense that everybody points to him right as evidence of a market in efficiency of some sort but there there are two problems with that one is nobody says that if you run companies you can't add value right right nobody says there's no such thing as human capital right right then there that does that part the other is if we have I don't know several hundred thousand businessmen right and we pull out the most successful one what's the probability that

that was luck not skill right even over a long period of time yeah so you got a big statistical issue there because that's the way these people get identified sure I would I would like like to have somebody examine what have the performance what has the performance of his acquired companies look like since he was anointed you know yeah interesting I had a kid come in one time an undergrad that said he had tested that and he promised to show me the results that I never saw again right must not have been very compelling I

don't know he that he found a girlfriend or something yeah well so in the end then it seems like having individuals pick characteristics that are appropriate for their personal circumstances would be the approach that you well you know I think the big problem facing investors is that they don't understand the importance of uncertainty about outcomes yeah so for example I get to do a lot of talking to institutional people and financial advisers institutional people especially tend to change their portfolios based on 3 to five years of past returns right and I show them simulations in

which that's basically noise you know 3 to five years of pass returns there's almost no information in that about expected returns uh and they're kind of shocked by it and but I think that's the the reality is that there's no there's no free lunch out there the the the higher expected Returns on stocks comes about with a large amount of risk and what that means is over long periods of time you can lose MH um um on a purely chance basis um and we don't have enough data to know what the true expected return return

is so we've got data going back to 1926 that leaves us with an estimate of the risk premium over one month bills that could be um the average number is 5 six% a year um but the standard the two standard are bounds around that kind of big very noisy yeah yeah so what about all of the technolog iCal innovations that we've seen over the last few years things like ETFs and online trading Robo advisors any thoughts on where the markets are going what you think in terms of well you know markets have if if I

think about DFA in the beginning um the people went home with small computers every night as a backup right now they have multiple systems I mean all over the world um that's changed in the old days they did a lot of block trading MH gone right because now everybody splits up their orders puts them out on multiple platforms and does their does their trading SO trading has become much more competitive much more fast moving you have to be very careful not to get front runed by anybody that can figure out what you what what what

you might be trying to do um so that's all changed dramatically and of course information is now much more easily available but you know I I frankly don't see the tracks of any of that and the behavior of returns I mean the markets don't look I get this every time I give a talk you get that question a Market's more efficient now because more information is there and it's more quickly available well maybe there's just too much noise in the data but you can't see the tracks of that in the data do you see the

noise increasing over the last few years or is it about the same about the same okay you know we go through periods of higher volatility lower volatility but after the 30s basically the process looks pretty stationary so you you have a really quiet period I mean believe it or not the 40s to about 63 are a really quiet period you have only two returns outside that barely break 10% monthly yeah and then thereafter you get quite a few that break 10% MH but you don't get what you got in the 30s which is a lot

of plus or minus things bigger than 20 you know right right and has the financial crisis changed any of the Dynamics of markets over the last few years it has you can't tell uhhuh okay if it I mean I can't tell yeah so would you say that you've changed at all your investment philosophy over the course of the last 10 or 20 years are you still investing in the same way i' say no okay uh but if you went past that so before 92 before we did the cross-section of expected stock returns I'd have said

um everybody should hold the market portfolio right now I'd say no your taste might cause you to tilt a little more towards smaller value or whatever right but you know it's still I still think the market is the centerpiece right you know and most people should sit there because it's a cheap way to go you know yeah um it's very inexpensive to hold uh a market portfolio from Vanguard or somebody like that yeah uh there are lots of providers that do it at very low cost but you have to be careful because there are some

that do it at high cost right yeah cost will eat you up right yeah and looking forward any thoughts on a farm French seven Factor model coming down the Pik we were we're doing the we came out with a five Factor model and that seems very robust but I don't think it's been fully vetted yet okay and I'm suspicious about the investment Factor M because um there is this phenomenon that's been there forever that all asset pricing models have problems with uh unprofitable stocks that invest a small unprofitable stocks that invest a lot mhm there

were a big problem in our 993 paper when we look backward we thought it was we thought it was MX NASDAQ you know the high-tech companies that were causing the problem but when we brought it back to 26 it was still there right and those guys weren't there at that at that time do it internationally it's there again yeah so that Factor gloms on to that stuff so I'm a little suspicious of it uh the other one is the profitability factor which is almost entirely due to small stocks because you think about the market the

cap weight Market is Big profitable companies right so that is a high profitability portfolio yeah the the Capo version of it sure well Jean that's no exaggeration to say that your theories have really democratized all of Finance I mean the efficient market Market hypothesis allowed all of us to manage our portfolios in a straightforward way so on behalf of all investors thank you very much for giving us the perfect portfolio listen thank you