Five Questions: Factor Investing with Jim O’Shaughnessy

By Jack Forehand (@practicalquant)

Factor investing has grown dramatically in recent years. But it is far from new. As those of us who invest using factors seek to learn more about them and the best way to utilize them in our portfolios, there is no better place to turn than those who have been there since the beginning.

For this week’s interview, I have the privilege of talking to Jim O’Shaughnessy. Jim was a factor investor before factor investing was a thing. He is the Chairman and Co-Chief Investment Officer at O’Shaughnessy Asset Management. His book What Works of Wall Street is considered by many to be the bible of factor investing. In the book, he showed that simple quantitative strategies, when followed with discipline, could produce significant market outperformance over time. Reading it in college was one of the reasons I decided to pursue a career in investing so I feel very fortunate to be able to get his take on some of the issues facing factor investors today.

Just a note before we begin. This interview was transcribed from a phone conversation so please forgive any grammatical errors. I also should probably stop calling this series Five Questions because I once again have failed to limit myself to that. But in my defense, when Jim O’Shaughnessy agrees to answer your questions, you ask as many as he will respond to. It’s just what you do.

We are publishing this interview in two parts. Part II will be up next Monday.   

Jack: Thank you for taking the time to talk to us.

One of the things many investors struggle with is the balance between following the principles they believe in consistently over time, but also understanding that things change and investment strategies sometimes need to evolve. Altering an investment strategy in response to short-term market movements is obviously a bad idea, but on the other hand, having a level of conviction that leads you to never adopt to changing times doesn’t work out very well either.

As you have issued new editions of What Works on Wall Street over the years and your strategies have evolved, you have been able to adhere to your core principles, while also adjusting your approach when the long-term data supports it. For example, for your value model, you started in the first edition using Price/Sales as your primary value metric and by the 4th edition had moved to a value composite. I was wondering if you could talk a little about how you think about the evolution of investment strategies over time and how you go through the process of deciding when a change is and is not necessary.

Jim: Every one of our evolutions of our investment strategies has come from pure research. People always joke, oh, you’re a quant so you golf all day. Absolutely not. What we do all day is we study, not only our existing strategies and factors, but anything that might be new or promising that we have either come up with ourselves or read about in the literature, and every change that has ever been made to one of our investment strategies follows a very highly detailed format. I could for example, say, oh my gosh, I love price to free cash flow. I’ve read all of the great papers about it – and that would be the operative word, I love – that would mean that I was maybe getting a little emotional about that. The way we evolve our strategies is through constant testing, not only of the strategies, but the factors that make up that strategy, whether we should be using a new or different one, etc.

So our process continues, day in and day out, research on all of the above that I’ve mentioned.  An interesting thing that Patrick wrote in the first quarter letter and something that I’ve always gone on about, he put it more eloquently than I do, but you have to have a big research graveyard. And I’ve always said, you know, 90% of our research ends up on the cutting room floor, meaning that we expend the energy, we go to the trouble to know, we take our time, treasure and everything else and invest it in looking at a variety of things. And the fact is that for the most part, most of that returns a null set, right? So, Gosh, you know, we could look at adding or subtracting a factor or we could look at replacing a factor or we could look at a whole bunch of different things.

And when we’re done with the research, we find that it has added no real value to our process and it ends up going in, as Patrick calls it, the graveyard of research. I’ll give you a couple of examples of how things changed through this process. You yourself mentioned that if we just use my book What Works on Wall Street as our guide post here, in the first one I was a little wet behind the ears and called price to sales the king of value factors. Well of course, that was before I took the time to think that really it depends on what kind of timeframe are you looking at to determine who’s going to be the king or the queen. If you were looking at it maybe 10 years earlier, it might’ve been EBITDA to enterprise value.

Looking at it 10 years later, it might be price to free cash flow. And that is where we kind of alighted on this idea of, well, you know, why not all of them? Every factor has its own strengths and own liabilities. Why not come up with a composite that combines all of these various factors, that cover different strengths, different liabilities, but also, for example, on the value composite side, cover more parts of the balance sheet. After doing that research, we found that yes, that insight was correct. We found it far more efficacious to use a composite of value factors because (a) when you’re looking at single factor performance, it’s always going to be a horse race with a different winner depending on what year you end and begin in. And secondly (b), the idea that you’re getting a much better sense for the overall value when you look at a variety of the factors than if you look at just one made both intuitive sense and then obviously our final test, which is the make it or break it one, it met all of our empirical tests as well. Same sort of thing to the idea of changing the original version of the book. I had a value strategy that used dividend yield as a final sorting factor. And then we took a look at using what we call shareholder yield, which is dividend yield plus buyback yield, and found that that was actually a much better predictor of future results then dividend alone. So every time that we evolve one of our strategies, it is based on a very highly articulated and understood a group of premises that everyone on the research team gets, and its numbers not emotions that make the decisions for us.

I’m the final say on whether something gets put into production or not at OSAM and, you know, it’s pretty easy for me because if there isn’t empirical evidence that it is going to improve the results, then it’s going to be a no. And thus, the majority of the research kind of ending up on the cutting room floor.

Jack: So by relying on the long-term data, you can essentially take your emotions out of the equation completely?

Jim: Yup. Essentially. Because, you know, I think that that actually brings up another interesting question, which is why are behavioral problems, as Danny Kahneman calls them, cognitive mirages? In other words, if it’s a visual mirage, your brain knows that it’s a mirage, right?

If you and I were in the desert and dying of thirst, and even if we both saw an oasis a hundred feet ahead of us, we could intellectually understand that we’re probably seeing things. Whereas with behavioral problems, it’s because they thrive on the emotion of the moment, and the emotion of the moment has the capacity to completely override all of our prefrontal cortex abilities. And again, you know, it’s kind of, if you look at evolution and look at who survived and who didn’t, we’re all the children of the people who ran away. If you looked at most human beings, what you’ll find at the center of all of their anxiety is fear, and fear is a powerful motivator and if motivates people to either fight or flee, but it is ever present. And because of the way we evolved, it’s hard to override; you know, it’s a subject of literature, it’s a subject of philosophy. It’s a subject of a lot of things. But a similar interpretation of it is quite simply the reason why looking at things historically is so much easier than looking at them right here and right now is that right here and right now, you and I, both of us are in the present, our emotional facilities within our brain, within our system are fully operational. I used to give the example, so think about the last time you really screwed something up, right? And then you’re thinking about it six weeks later and you look back and what you see is very different than what you felt at the time, what you see is the facts, right? You see the facts devoid of any emotional input, and they’re crystal clear, right? You’re like, god how could I not have seen that? Well, you couldn’t see it because your mind was literally hijacked by the primal emotional operating system within the human being and that fear or that uncertainty blinded you to what in hindsight of course is easy to identify. And it’s interesting to me because even as much as I tell this to people and they truly do intellectually understand what I’m saying, they absolutely do; It’s not a hard concept to grasp at all. It’s a very easy concept to grasp, people go yeah, of course that makes so much sense until you are flat up against a market that’s down a thousand points in the first 10 minutes of trading and all of that goes out the window for most people and they let their emotions take over and that almost always leads to very, very bad results.

People have asked me, so what are you most proud of in your investment career? And I always say that I think my proudest moment is that I didn’t override, I never let my emotions determine what I was going to do. It is really, really, really hard. Even for somebody like me whose got quant in my DNA. It’s really hard to do.

Jack: I want to ask you about price to book because that is an interesting example of the balance between data and common sense. There’s a lot of academic data over a really long period of time that supports price to book, but at the same time, if you look at price to book logically today you would say that it doesn’t make sense in a world where the vast majority of assets that companies have are intangible. But on the other hand, I don’t know if you read it, but Corey Hoffstein did an interesting post a while back where he talked about how long it would take statistically to say for sure price to book is completely broken. And he came up with something like 80 years or something like that. I know you have removed price to book from your value composites and I am wondering how you look at that balance between the long-term data supporting it and logic that is doesn’t make nearly as much sense as it once did in today’s world.

Jim: Sure, sure. So number one, I love Corey’s work. I think he does great research. It’s one of the things that we at OSAM do all the time is we try to read all of the people that we have a lot of respect for their research and take it into consideration. You know, in our instance there were several things that were driving our hesitancy on price to book. The first was that for the fourth edition of the book, I had access to the CRSP data (that’s the Center for Research and Security Prices at the University of Chicago) which goes back to the late 1920s.

And when we ran that data on price to book what we found, if you go to my chapter on price to book in the current edition of the book, you’ll see it, was that price to book between late 1920 and 1963 where we actually began with our earlier versions of What Works didn’t work at all. And one of the reasons it didn’t work at all was because of the Great Depression in the 30s. And then you’ve got to start really doing some sleuthing and figuring out, okay, what’s going on here? Well, if you start reading like really early accounts of why price to book could be a yardstick to use, you’ll often see the caution that price to book is also a good proxy for risk of bankruptcy. The lower the price to book, the more at risk that company is to go bankrupt.

In the 1920s, it was sort of a perfect indicator for a company’s chances of going bankrupt. Because we were not finding very long-term consistent support. So you know, you look at things, other things and you’ll see, if you say, look it all rolling 10 year periods, you’re getting very similar results, at least directionally, right? With price to book that was not the case. In fact, what you saw in that period between the late 1920s and the early 1960s was an inverse relationship with the stocks with the highest price of books doing much better than the stocks with the lowest price to books. On top of that, it comes into the idea of continual research that we do at the factor level. So for example, one of my associates on the research team, Travis Fairchild wrote an outstanding piece from my point of view, looking at exactly what you’re talking about.

He took a very different look, but ultimately obviously also an empirical look at why price to book might have been efficacious in a much simpler economy, right? Where assets were easy to see and easy to value and how that doesn’t really translate well into today’s economy with all of the intangibles, such as brand value, etc. being very, very difficult to account for. We also have a forthcoming paper by our research associate who is known by the pseudonym Jesse Livermore, in which we believe you’re going to see the final nail go into that price to book coffin. And the reason that we think that this is important is despite all of this ongoing research that is very empirically compelling, there is still something like $1 trillion indexed based on price to book.

And so from our point of view, we take being fiduciaries very, very seriously and you know, if we have mounting and compelling empirical evidence that a factor has either broken down because of changes in the economy, or because of other reasons, or because simply empirically it’s giving you wildly different results based on whichever 10 or 20 year period you take a look at, we think that in good conscience, you really have to not only highlight that and publicize it, which we try to do, but not use it. So as you yourself note, we stopped using price to book because of what we view as pretty serious deficiencies in the ratio itself.

Jack: You have been very vocal over the years about the issue of investor behavior. The way we are wired as human beings can be a positive in many aspects of our lives, but it can wreak havoc with our investment returns. In your talk at Google, you talked about the two points of failure that active investors face in that they will tend to abandon a strategy not only when it is down on an absolute basis, but also when it underperforms its benchmark. For those of us who develop factor strategies, this requires a balance between the superior returns that usually come with more focused portfolios and the better investor returns that come with limiting tracking error relative to the benchmark. How do you look at the issue of balancing actual and investor returns when developing an investment strategy?

Jim: Wow. You know, that is such a great question and it’s one that I have been asking myself with a great deal of seriousness over the last 10 to 15 years. When I started doing the research that ultimately led to What Works on Wall Street and some of my other books, I realized with the benefit of some hindsight (of course, hindsight bias is always 20/20) that originally I was sort of designing strategies that fit Jim O’Shaughnessy, who, you know, is a high risk, give me maximum alpha type of investor. And the more I read about the various failures of investors to stick with even simpler strategies, for example, an index strategy, you’d kind of think, wouldn’t that be like the easiest one for someone to stick with? Well, yeah and no. Right? Because there is a single point of failure and that single point of failure for an indexer is when the overall market has declined so much that that indexer can’t take it any longer. They react emotionally and they throw in the towel. They sell their index fund. And so active management like we do has two points of failure, right? The first is the same as the index investor. The overall market is down to a level where the person invested in that strategy for emotional, usually primarily emotional reasons, can’t take it and sells. But also if you’re an active investor, you have the added problem of let’s say you’re following an active strategy like our Market Leaders Value that buys well valued companies with high shareholder yield. And you know, for the previous three years, the Russell 1000 value is up 10% and your portfolio was only up eight and a half. Well you might use that and say, oh, well, you know, that the strategy used to work but it doesn’t anymore so let’s throw in the towel. So the idea is that there are so many points of failure available to both index investors and non index investors that we do try to keep in mind; what would a person with you know, sort of average risk tolerance and sort of average expectations, what could they withstand? Now obviously we’re going to work hand in glove with the advisors who use our strategy because we do believe, as you know, we don’t go direct to the public on the non institutional side. And so we try to work with advisors who really understand this and who are advising an overall strategy that they think is going to work with the client in question. It’s interesting because just when you think you’ve kind of nailed, well you know it could be this draw down or it could be this underperformance or it could be, you know, and you start going through all the things that might potentially get somebody to throw in the towel. Sure enough, human emotion gives you yet another reason that you hadn’t been considering.

And it’s one of the reasons why I spent so much time on trying to get people to understand these biases. I don’t even know if we should call them biases. I think that what they are is, they’re just part of our programming. And honestly, while we strive to get people to understand the downside, it gets back to that question that we were discussing just a moment ago. And that is, you know, when you’re discussing it, guess what? No one’s emotional. And you and I, you could be here in New York and sit down, and I could show you a portfolio that I’m showing you the results of, say 70 years, right? And I’m pointing, I’m saying Jack look, look at this horrible drawdown here. That could happen. You get that that could happen, right? And you say, of course I get that that could happen. Of course, I get that. Well, that’s a number on a piece of paper that means nothing, because you’re not experiencing it, not in real life. You know, when I started the first version of OSAM it was called O’Shaughnessy Capital Management and we used to actually, when people would come in, they’d read What Works, and at this time we were working directly with the public, and they’d come in and say, “Hey, I want, you know, strategy, X”. It was always the strategy with the highest return. Right? “Okay, okay, but can we show you something?” And they’re like, sure. And then we’d show them the worst rolling 10-year results, which almost always is pretty horrible. And we look at them and say, you still want that strategy?

And what’s interesting is that, for the minority that said yes, they ended up sticking with it a lot longer than the people who said “oh hell no way”. But you know it’s one of those things.  You think that, you know, human beings, it’s so fascinating to me because you think you know what a person is going to do, right? The old saw about the best way to predict someone’s behavior today is look at their behavior yesterday. Right? And that’s true right up until the minute it’s not. And I could tell you story after story after story about advisors that we worked with for years and years and years, who made it through a lot of market turmoil, but then cracked during the great financial crisis. And you know, it’s just so fascinating to me because there’s literally no way I’ve come to learn; there’s literally no way you’re going to know who’s going to be able to stay the course. And you know, by the way, I’m not saying this in any kind of manner that in any way belittles people being emotional about the returns of their portfolios. I get why they are emotional. You know, money is kind of the last taboo, right? And that’s because unfortunately it’s come to be seen as shorthand for success or value in a meritocracy, which I think is kind of crazy, but there it is. And it’s not just money, right? It’s your family. It’s your children, it’s your spouse, itis your grandchildren, its everybody. And when you throw all of that together, you’re talking about an incredible potion to let the emotions just run riot. And, you know, it’s something that we keep trying to figure out if there’s a better way or a better mouse trap to quell those emotions. But for the most part, we haven’t found it yet.

Jack: You know, we have struggled with this too. We manage money as well and trying to figure out upfront the perfect thing to ask somebody to figure out what they’re going to do is basically impossible. Because until you put them in that situation, you just don’t know what they’re going to do. No matter how many questions you have, no matter how many simulations you run through, until they actually see it, until they see 2008 with their own eyes, you don’t know.

Jim: And that’s so true. And you know, we can all look to our own personal experience, right. I’m sure that you like me have been in situations where you made an emotional decision. I certainly have. And you know, when you reflect upon it, you think if you’d asked me to bet a lot of money, I would have bet that I wouldn’t have done that and yet I did. So it’s like – one of my heroes is Richard Feynman the professor and one of his great sayings is rule number one is don’t fool yourself because you are the easiest person to fool. And so, you know, when we can’t even predict our own behavior under duress, it’s going to be very difficult as you point out. Simulation. Simulation. Simulation. Well, unless it’s a holodeck simulation, you really feeling it, it’s just not the same. And you know it’s funny because, especially on the quant side, it’s funny because of our beliefs, we, we really try to put people through these steps. I think even more than conventional managers do, and yet it’s one of those situations where you can simulate things until you’re blue in the face. Gorbachev had a great line, which was something along the lines of – I’m going to quote this wrong – but, he was talking to somebody who was holding forth on something and he said, well, your answer is meaningless because it is academic. It does not take into account the full- blooded reality of the here and now. And I always thought, boy, if there’s ever a way that for us to be able to, you know – maybe virtual reality will do it for us – but if there’s ever a way to make people actually feel what they might feel when experiencing a horrible turn the market, I would certainly advocate using it unless it’s of course, you know, too damaging. We don’t want to kill them. We want to make sure that the patient is okay.

But you know, what’s interesting, and I think that we would be remiss if we didn’t add this, there’s also problems on the other side, right? When things are going way too well, too much in your favor. And you know, the old saw that I break out all the time is the four horsemen of the investment apocalypse are fear, greed, hope and ignorance. And fear, greed and hope have wiped out more money than any recession or depression. But we’ve been there, right. And my operating theory about the stock market has been unchanged. And this is a good illustration for another thing that I always advocate that people do and that is to keep a journal of their investment ideas and thoughts and why they’re doing what they’re doing. And it’s something I’ve done since I was 20 years old and it’s not just for investments by the way. My journal wasn’t. But looking back, I was doing that about a couple of weeks ago and looking back at some of my investment stuff because I was trying to think, you know, when did I start thinking about the stock market in this way? And it turns out I have been thinking about the stock market this way for a long time, and the simple way to put it is the stock market is a complex adaptive system, with feedback, right? And, and it works so well because the feedback is going to be interpreted by people very differently. And that’s called heterogeneous interpretations. In other words, you know, you might want to buy Apple stock and I might want to sell Apple stock and neither one of us is wrong, right?

You want to buy it because maybe you’ve got a young child and you’re starting their college fund and maybe I want to sell it because I have a child who is getting married and I want to pay for her wedding. And you know, there’s infinite reasons, right? But what’s fascinating is markets work so well because that feedback into this complex adaptive system with a variety of opinions works very, very well. However, markets get very scary when they’re complex adaptive systems with feedback and the feedback turns into what I call an information cascade. What’s an information cascade? An information cascade is information that becomes so uniform that virtually all of the players paying attention to it interpret it the same way. So that shifts players from being heterogeneous to being homogeneous. In other words, they all have the same thoughts about something. One of the reasons why I have been recently studying memetic behavior, which is basically the idea that a lot of our behavior is led by copying other people and then making up a reason logically in our own brain for why we do something the way we do it. Well, that’s not really true actually. If you look – and it’s got a fancy name now and everyone likes the term meme – and memetic is, you know, the same kind of thing. But this has been around since the Allegory of the Cave with Plato. People often when they don’t know what to do, right, what did they do? They kind of look around. They say, well, what’s everyone else doing? And suddenly they start doing that as well. And that becomes absolutely vital during market bubbles like the .com or the housing CDO crisis and near market bottoms. So if it’s December 1999 or if it’s February 2009, I think that information cascades had occurred causing people to all believe the same thing, all react the same way. And I got to tell you, it’s like a magnetic pull. That’s why I’m looking into, you know, the virus of memes, they can take over people’s behavior.

Just think about the Salem witch trials for example, if we want to kick it out of our current context. Basically, what you see is this information cascade becomes so compelling and trying to resist it, even people who intellectually resist it have a hard time resisting it in practice. And by the way, I am like the perfect example of this. So in April of 1999 I published a paper called the Internet Contrarian in which I was really, really, bearish on Internet stocks, saying this is the biggest bubble any of us had ever seen. And 85% of these companies are going to be carried out of here feet first, and even the winners of the internet are already vastly overpriced. And oh boy this isn’t going to end well. Well, what did I do a few months later? I started an online investment company.

It’s always great when you yourself could be the best example of stuff like that. People might not want to think that they’re capable of it, but here I am. And so, you know, intellectually, boy did I get that valuations were insane and they were crazy and everyone was going to lose their shirts, and what did I do? I memetically just couldn’t stand it any longer and thought, well, I’m jumping in too. By the way, the practical Jim should have been the better guidance for enthusiastic Jim. Because, when Jack Willoughby wrote his Burn Rate article for Barron’s that kind of pulled all those walls down.

Jack: Yes. It’s really crazy because even those of us who have studied this stuff forever, and even someone like you who’s been in the business for a long time, and I’ve been lucky enough to meet Daniel Crosby author of the Behavioral Investor, and even someone like him who has studied this in every way possible and wrote an excellent book about it, all of us are still subject to this, and I think that is probably the biggest takeaway. You have to realize that no matter how much you learn, you learn that all of us are going to be subject to this no matter what.

Jim: You know, one of the first tweet storms I did included that sentiment. It became very popular, which I was surprised by, and one of the things I mentioned was, by the way, this applies to everyone, especially you. I said, by the way, if I didn’t have quantitative rules keeping me safe, I think I would probably be even more subject to these behavioral biases. And also when I’m giving talks to groups of really, really smart people, and I mean PhDs, doctors, engineers etc., I’ll say your intelligence actually is a disadvantage here. Why? Because really smart people are really good at coming up with very compelling narratives. And back to Feynman, who do they fool first? Themselves. And they are very good at convincing other people too.

I can’t tell you the number of pitches that I used to sit in on for things like CDO funds for example. And literally, I had, thank God, inoculated myself against the madness after my own experiences starting an internet company right before the Internet bubble blew up, and I would ask just like basic questions, like isn’t 40 to 1 leverage on an illiquid instrument dangerous? Literally, other people who were around the table looked at me like I had three heads. Like, are you insane? This is foolproof? Famous last words, right. There’s a great quote: if you think that you can make things foolproof, you’re underestimating the ingenuity of the average fool. And so it’s a really fascinating topic because people in general are really good at ascertaining the strengths and weaknesses of other people, right. So, you know, I thought at one time maybe the best way to have people do an investment questionnaire is to have people fill out the questionnaire for the person sitting next to them and then after they’ve done that have them switch papers. You know, I know I’m not going to panic, but this guy next to me, he’s probably going to panic when the market’s down 20%. Then just have them switch the paper and say, no, no, that’s your survey.

Its another one of those really fascinating things when you dive into this stuff and you read the literature as much as I have, one of the other points that is just so common is that we’re pretty well calibrated at finding other people’s flaws. We are horribly calibrated at finding our own.

In part II of the interview, we will cover a lot more ground, including Jim’s views on trend following and whether simple value investing strategies can still work in a world of big data and machine learning. Jim will also respond to some of the arguments others have made about the death of value investing. Stay tuned for Part II of the interview next Monday.

Copyright: Photo: / vska