In Pursuit of the Perfect Portfolio: Harry M. Markowitz

[Music] hello I'm Steve Forrester welcome to our project in pursuit of the perfect portfolio it's my pleasure and my honor to be here with dr. Harry Markowitz 1990 Nobel Prize in Economics winner and the father of modern portfolio theory Harry thank you very much for joining me today thank you for having me I want to start off by taking you back to the early 1950s and tell me about your first investment decision when I wrote my article my 1952 article I had never invested I was a student with no one sale reports to it I

mean it was a student without funds and the first time I had the opportunity to invest was when I had joined the RAND Corporation and Santa Monica they offered TIAA forces Scripps creff stocks versus bonds and at that time the studio time makes them look forward to telling this story people have it wrong have it right but but they don't have the right conclusion at that time I thought if the market goes up the stock market goes up and I'm completely out of it then I look silly and if it goes down and I'm a

hundred percent in it I'll look silly so I went 5050 so it was at that time minimizing maximum regret now the story is sometimes told and I've seen it even the Wall Street Journal that Harry Markowitz even Harry Markowitz when he makes his portfolio decisions he doesn't use mean variance analysis he doesn't use MPT just uses minimum acts regret well that's what I did in 1952 but that's not what I would do today is not what I would recommend or more precisely it's not what I would recommend a 25 year old do know now I

probably put them 100% in stocks a lot has happened since I published that article in 1952 there's an infant structure we have this data that goes back to 1926 is light at least it shows you how I'm the average over the long run if you put a dollar in small caps who would be worth like 13,000 or something like that because it's compounding at 12% you put it in large cavity we worth three or five thousand because it's compounding at ten percent but if you put it in government bucks and be worth 150 bucks as

compared to fifteen thirteen thousand now we have optimizers we didn't have an optimizer in 1957 so on so the the conclusion is false that now Harry Markowitz 2017 would recommend to a twenty five year that they go 50/50 so I want to come back to that later on in our discussion in terms of your views on the on the perfect portfolio so that's a great foreshadowing of what your thoughts are let's go back to when you were when you were working on your dissertation at University of Chicago and and you have this this unique ability

to see things that that other people don't see can you tell us about one of the aha moments when you were in the University of Chicago library reading John Bear Williams investment book in and what what what were your insights that that came to know that my first that was my first aha moment and I've been seeking aha moments ferger's ever since I was at the stage where I was where I had to write a dissertation well back up a little bit with the story I went to my thesis adviser Yasha Marshak to ask for

suggestions he was busy so I waited his auntie room and somebody's another guy was out there and the other guy was a broker waiting for marsac so we talked about why we here and he suggested that I do a dissertation you know in the start without the starting applying mathematical econometric techniques to the stock market so I went in before he did I told myself the guy out there I suggested that I do a dissertation of the stock market and applied to the stock market and Marshak said well Alfred Coles who adopt the Cole's Commission

would like that he was it was actually interested in application of econometric techniques to the stock market he was the first to do experiments with weather forecasters could forecast have found out they couldn't and it's on so Marshak thought it's a good idea but he didn't have he didn't know the literature so he sent me to Marshall ketchup then dane professor Ketchum was the Dean of the business school I now understand and he gave me a reading list which included Graham and Dodd which is still in my shadows of course up and wheeze and burgers

poor investment companies of their portfolios and we could see how things were diversified and classified by industry and then john byrne williams was the financial theorist of the day and Williams said that the value of a stock should be the present value of its future dividends and now of course dividends are uncircumsized figured he met the the present value of the expected value of future dividends later in the book he does say that when things are uncertain you should use the mean value the expected value my reasoning process went well if you're only interested in

the expected value or the average the mean value of a vostok you must be only interested in the mean value of a portfolio oh and if you're only interested in the mean value of the portfolio the way you maximize that is put all your stock with all your eggs in one basket which I knew that wasn't right so I thought well you're trying to avoid risk as well as seek return I drew a graph with risk on will you want to return on one axis Andrew on the other and I thought of the returns on

security is being like random variables I mean that means that the return on the portfolio was a weighted average of the returns on the individual securities where you choose the weights I knew offhand at the time what the expected value of a weighted average weighted sum was but I didn't know what the variance or standard deviation of a weighted sum was I looked I got a book off the library shelf expense keys introduction to mathematical probability my memories of that goodbye remember that whole thing very well and looked up the formula for the variance of

weighted sum and there it was covariances correlations so I had the aha moment that the variability the riskiness of the portfolio attended not only on the volatility of the individual securities but to the extent to which they went up and down together and so there's still a lot to do that that became essentially my 1952 article plus a little geometry of pictures of how efficient sets look and there's still one more to do like trimming out how to compute these things an estimate but that was the aha moment so these ideas that seems so simple

now but but yet you saw something that that no I don't know why Williams says that if you diversify sufficiently you will get the mean but that is only true if risks are independent uncorrelated if risks are correlated you don't get the mean you get there's a rule that says the variance of your portfolio and this is from chapter 5 of my 1959 book the variance of the portfolio doesn't approach 0 it approaches the average covariance so for all the correlations are 0 then the average covariance of 0 and variance will approach 0 but in

general I won't and how somebody could live through the 1990 but 1929 to 1932 crash and assume uncorrelated is difficult to see in retrospect I think his book came out in 1938 so he should have yeah he had gone through you know he had gone through oh he had seen the 29:32 was fresh in his memory the the notion of diversification goes way back the Bible had some refer to the Merchant of Venice Antonio why are you said is your business going bad my my goods are not to one bottom trusted but not and not

my fortune to this year so my field so he knew about diversification but was like it was a good theory that would catch up with intuition and then you brought that theory and here we are now in March of 2017 right and now this is the almost to the 2050s with the Davis over the sixty-fifth of security on our anniversary v is going xt60 look at the anniversary how times long at papers I want to talk about that this is the the journal Finance paper called portfolio selection and what's so striking about it is is

how different it was compared to other papers that appeared in that same issue that looked at inflation and public policy issues and income statements what is it that made your paper stand out and and made it so different at the time and yet it wouldn't be that different from other journal Finance articles today well my emphasis was on portfolio selection you know because considering the portfolio as a whole they were all talking about evaluating securities or industries and the notion that you should have some kind of a theory up about what makes a well diversified

portfolio and what is the trade-off between risk and return that some it's surprising that the human race went so long to leave me to discover that to fill that need and and really providing a process that helped to create a whole investment management portfolio management and let me tell you a little story of you know Peter Bernstein of course his book on there's a capital idea Capital ideas and something against the gods and strange history of residency of the history of risk he he was at a meeting that I was at to the Robert not

has an annual meeting and I went one of his his advisers and Peter Bernstein Peter was there then he's no longer with us and after he has comment about to I'm somebody else's paper he just as a remark from the audience not as part of his paper he said it's common under somebody else's paper you younger people don't know what institutional investing was like before the 1950s we would sit around and have discussions like you see on television about I think this industry or I think there's some company and somehow we would clobber together cobble

together a portfolio and he said now you have a process and it was at that minute a moment when he said now you have a process that I realized what I started of course there's a lot that goes into it besides you know like data and programs and so and so forth that's most about amazing what look for you to oversight over now 1952 was a great year for you you have the Journal of Finance paper but but a lot of people don't realize that that there was another paper that you referred to as Markowitz

1950 to be right that appeared in the journal political economy the utility of wealth and and it's through that paper that you've now been often referred to as is not only the father of Modern Portfolio theory but the grandfather of behavioral finance behavior like economics can you tell us about that paper and the key insights first confirming that you know certifying that I'm the grandfather of behavioral finance Danny Kahneman has a book out called thinking fast on thinking slow and of course if you're Harry Markowitz and you get a copy I think excessive thinking slow

you get you a first thing you looked in the index to see if I refer to there's two references to me one was sort of inconsequential the other was told us the history of how prospect theory came about and the Kahneman says that there was that he and Tversky were struggling with some some experimental results which they just couldn't understand and finally Tversky came and said now I got the answer it's in a then 25-year old paper by Harry Markowitz it said that if you want to explain actually havior do not attach utility to to

wealth attach it to change in wealth and that you know then that plus one and then it was a sort of a stripe of the curve depending on my curve had a pennant a concave portion the left on the losses so that was explaining why there's insurance people insure and then there was a convex portion which explained why which is part of the explanation why people buy bought lottery and then I was again a concave because otherwise you have the st. Petersburg paradox now that kind of this all came about come about I was teaching

taking the Friedman's class in microeconomics uh he I don't remember whether it was a an assignment an optional assignment or a man notorious einman but he assigned a paper by Friedman and Savage I can't remember the name of the paper but it was trying to explain why there was both gambling and insurance and it had a curve which is very much like the market which currently later Markowitz curve in which there was a concave portion a convex portion of concave portion so I looked at this and I thought well you know if you took a

it looked like a two-humped camel walking uphill so if you take a plank and you know put it against these two humps you get a double tangent and the behavior of people depend on where they are with respect to the these two tangency 's like for example if you have two middle class of people who are half way between the lower tangency and the upper tangency there's no fair bet that they would prefer then when were they flip a coin and one becomes poor and one becomes rich and you don't see that in fact and

then the people who are below the lower tangents of poor people they don't buy lottery tickets and then I don't know who's lining up in front of me when I'm trying to buy a Wall Street Journal on 3rd Avenue area and so on I mean it just didn't make any sense and the only point that made any sense was it the inflection point where they were they be you're cautious on the downside and maybe a little bit you know willing to gamble a little on the upside and then I said I didn't call that that's

usually current wealth but I said I call that customary wealth because if you have a recent windfall gain you move into the convex part I don't you're you're a little bit more devil-may-care you're playing with house money so to speak and if you have a recent windfall laws you become more cautious so so the only point that made any sense was this inflection point and as compared to your group your sake Tversky says common Tversky lives become its current wealth but like I talked about as customary wealth because of the you're usually there but if

you had a recent windfall gain or loss so that was that was another aha moment you've had many of them I love them I just adore effect so you followed up your your 1952 portfolio selection paper seven years later with with a book by the same name and it packed a lot of things into it there was a chapter on matrix algebra that I think many famous people including Bill sharp learned is available well of that absolutely so what what had changed and and what was in there in terms of some of the nuggets that

eventually became the capital asset pricing model with Bill sharp under your tutelage was able to follow up on well there's no that's dispelled the notion that Bill sharp got the idea for the capital asset pricing model from me let me tell you the real numbers were down that path let me tell the relationship between made me a mill we think roughly 1962 success as far as Bill and I can reconstructed I was working at a RAND Corporation he was working at the RAND Corporation II I was working on the script programming language or something else

besides portfolio theory and he compared with my door said my name is Bill sharp I'm a student I worked here like you do and I'm a student trying to get a PhD at UCLA UCLA right Fred Weston who was the publisher who is the editor of the Journal of Finance at that time had told bill oh why don't you ask Kerry for suggestion you like his 1952 article what did you ask him for suggestion and in chapter 5 of Marcos 1952 of 1959 I talked about what I called arrived covariances the fact that there are

just too many covariance is to expand correlations who expect anybody to in a team to individually look at them and estimate them so you need something like a factor model and I a point you know I pointed to him about the the idea that there's too many covariances and whether to try building a factor model and see how well it works so that became sharp 1963 a simplified model of portfolio theory which showed how if you're estimating fact you have factors for each security rather than covariance is you know if you have you know one

hundred securities you got a hundred times 100 divided by two roughly covariances that's fifty thousand whatever it is up so whereas you'd have one hundred betas testament so now is your estimation process easier but he published ways of quickly tracing out efficient frontiers and in those days a computing computing wasn't like it is needed today you know you the biggest computers in those days weren't as powerful as your cell phone you know so so he published Sharp's 1963 article was on factor models his 1964 model was on a capital asset pricing model now I published

on cap M but it's like I have a an article called eat market efficiency a surrogate a theoretical distinction and so what and the came out of the FHA well decade ago or so plus or minus the and it said the theoretical distinction was between the proposition the market is efficient in the sense that it has correct probabilistic information everybody has the same beliefs and they're correct oh and everybody has a mean insufficient portfolio you have to distinguish between that efficiency in that sense everybody's processing information correctly and the statement the market portfolio is a

mean variance efficient portfolio in what I show in that paper that FHA paper on a theoretical dissection and so what is not only do you have to distinguish those concepts but the in order to deduce that the market portfolio is any mean variance efficient portfolio and there's this linear relationship between expectorant and beta you have to assume either like sharp and linter that you can borrow all you want at the risk-free rate or you have to assume like Roy did that you can short and use the proceeds so you can give your your broker a

thousand dollars short a million dollars worth of stock one and go long a million and a thousand dollars worth of stock - and it's a feasible solution you know I don't think I'll get away with that I don't think what it tried tried it your broker if you do tell me what your brokers name is so there's a ranked team that has constraints so and it's also true that there is not a linear relationship between expected returns and and beta defined and justified so these folks who find that there's not a linear relationship and then

have all these fantastic explanations but why there isn't a linear relationship my explanation is very simple there is I don't know about you but I can't borrow I wanted to history write that said I should add that I still think kappa mo is very important because it is a null hypothesis against which you can test you know empirical results you could a lot of people do and it's just their interpret you know when they find there isn't a linear relationship and they have all these fantastic explanations like I say as I said I think it

cap em was very important historically and still is as he as the null hypothesis against which all things get measured let me come back to something we talked about earlier let you and sort of that sort of a circle circle back the perfect portfolio okay what what do you see as as is the perfect portfolio for individuals for institutional investors is there a perfect perfect portfolio and and what might be coming out of your work that would suggest one type of portfolio for sure another okay it sort of tied into first lay since I don't

believe that you know because people can't borrow they want to originally right I don't believe that the market you know the market portfolio plus or minus leverage is even efficient so it's certainly not perfectly the difference between a major difference between mark was 1952 and mark was 1959 is mark was 52 proposed mean variance efficiency and it analyzed it subject to one particular constraint set the sum of the X is equal one of the X's are not negative by 1959 I had come to realize that you might want lots of other constraints on the choice

of your portfolio might want on upper bounds on individual securities you might want upper bounds on on how much you have uncertain in industry or you know sets of securities you might have other linear constraints like you want to have income a certain level its compared to and so on so what I provided in Markowitz 1959 was a computer program that found mean variance efficient portfolios subject to any linear equality or inequality constraints of their linear constraints like the sum of the x's equal 1 or inequality like this plus this has got to be less

than or equal of that and these are used in in practice like I have a client who does portfolios of closed-end funds buying when there a deeper discount selling alert less deep discount but it he has a world index on all country world index benchmark and so he constrains the portfolio not to deviate too much in any one country and not to deviate too much from the benchmark in any one region and he doesn't want to turn over constraint to be turn over to db2 to great and so he has a turnover transfer and so

on so talking about the individual well individuals differs if some are somewhat they differ in their tax situations so it should be an after-tax mean variance analysis and that gets a little tricky because there are things with different time horizons you put it into a the 401k plan you can't get it back out without penalty until you're something you know 59 and a half or whatever okay so the the correct part only not the perfect portfolio with the correct portfolio for the individual depend on all the risk preferences you know they're willing to trade off

you know trade off from is for return you don't want them dropping out of the program prematurely so if you if if it looks like the right thing prove them for the long run if they have them 50% in small camp and you know small camp is very volatile or maybe it's 50% of the emerging market which is very volatile and if it has a bad bounce right out of the box you'll decide that's a stupid asset class this is stupid manager program so there are constraints on the choice of portfolio which varies from individual

to individual some individual they have shorter or longer horizons they they are willing to tolerate their they're willing to invest in certain asset classes or not they have different tax situations and so on so there is no perfect portfolio there is a right note we will assume that there is a right portfolio for them and part of the process is to involve the rightful OOP they want part of the process is always to involve the user the investor what is their trade-off between risk and return in the old days you just showed them up a

frontier if he said pick now you do Monte Carlo simulation to show them whether it's a few points from the probability distribution of how much they can spend when they retire and so on and so forth so there is a perfect portfolio there's just the right portfolio for any specific individual is a lot of work to find them so again it's the process that you provided it's certainly going to be diversification right it's going to be an important element to that but not necessarily the market portfolio so typically not in fact you front frontiers you

know they start out with a fairly unda versified portfolio and they pick up securities and they lose securities and people are different you are different you know different people are in different points on the frontier and nobody's holding the market portfolio because the market portfolio probably isn't even an efficient portfolio and it's certainly not an efficient portfolio for everybody and we can't for one lend at the risk-free rate so it would be if you could borrow and lend if you could borrow all you want at the risk-free rate then the only mean variance efficient portfolios

would be the market plus or minus leveraging but it's not true so where do you think the investment management industry is is going to go in the next 65 years right now we've got Robo advisors which in some senses is going back to the beginnings and looking at things in a systematic way where do you think the industry is going well I might mention that I have a number of clients I work on retainers and so somebody's paying me even now the cookie of talking and a couple of my clients are Robo advisors with different

one is acorns where they think takes small change you know they round up the they round up the bill and your tech to your credit card bill and invested and say and so so yeah I'm all for Robo advisors they have to then make some guess as to where you want to be on the frontier or maybe just as fact that you chose them this particular advisor Robo advisor shows what kind of portfolio you must be interested in the I think oh we are now doing a very good job increasingly good job of exploiting my

19:52 idea and not let me tell you the idea that's in volume two of the four volume book I'm currently writing consciously way yeah yeah I got to plug the book it's called risk return analysis and the subtitle is that the theory and practice of rational investing and actually it's not about a rational investing it's a rational decision-making financial planning it's not seven and volume two talks about decision support systems so a 401k advisory service or robots of is a decision support you with the the better one I mean the more flexible ones you get

to interact with the system and pick what risk class you want to be in and what savings Rachel need did have and so on so you so it helps you it helps you make the decision that it takes your decision and implements it so it's a decision support a decision-maker it's a decision support system and the investment decision I think I view any of the way I view it is part of a game which the family is playing it and these are these are financial this is the financial planning game but it has to do

with many things like births and deaths and weddings and people getting sick and flycatchers all life life events it is the events and decisions which have major impact on the supply and demand of Natural Resources and the just analyzing the portfolio selection that decision in isolation is like trying to decide how our bishops should move in a chess game without considering the chess game as a whole I think the way I see the thing going in the next 60 years it looks the way I'm trying to push it in volume two is for the the

man-machine the human computer division of labor to cover more fully the various aspects of financial planning well we wait anxiously for volumes three and four and and on behalf of all investors I want to thank you for what you've contributed to to help help us make better decisions more rational decisions and I really appreciate you taking the time it's been my pleasure both both to to work at this field for less which 65 years we've decided and and your interview was very very pleasant thank you very much thank you [Music]