In this episode of The Acquirer’s Podcast Tobias chats with Dan Rasmussen, who is the Founder and Portfolio Manager of Verdad Advisors. During the interview Dan talks about his replication of private equity investing in the public markets using levered small cap stocks. He also provides some great insights into:
– How Can Investors Implement A Successful Private Equity Strategy Into The Public Markets
– How Smart Investors Use ‘Good’ Leverage To Magnify Returns
– What Are The Real Drivers Of Private Equity Returns
– More Knowledge Doesn’t Make You A Better Forecaster
– There Is No Relationship Between Who The CEO Is, Or How The CEO Gets Paid, And What Happens To The Stock Price
– You Either Take Risk And Get Return, Or You Buy Things That Everyone Likes And You Get Mediocre Returns
– Investing Decisions Should Be Based On Empirical Data – Not Stories And Emotions
Here’s the link to the “unpopularity” paper I describe, actually a “book” at 163 pages, Popularity: A Bridge between Classical and Behavioral Finance by . The relevant part of the discussion about moats starts on page 84.
The Acquirers Podcast
You can find out more about Tobias’ podcast here – The Acquirers Podcast. You can also listen to the podcast on your favorite podcast platforms here:
Tobias Carlisle: Hi, I’m Tobias Carlisle. This is The Acquirers Podcast. My special guest today is Dan Rasmussen of Verdad Advisors. Dan has a PE replication strategy that is quite similar to my own approach to deep value, so I’m very interested to talk to him. No less a personage than Jim Grant of Grant’s Interest Rate observer, describe Dan as a human Ferrari. So, we’re gonna talk to him right after this.
Tobias Carlisle: (intro music)
pod intro voice: Tobias Carlisle is the founder and principal of Acquirers Funds. For regulatory reasons, he will not discuss any of the inquirer’s funds on this podcast. All opinion expressed by podcast participants are solely their own and do not reflect the opinions of Acquirers Funds or affiliates. For more information visit acquirersfunds.com.
Tobias Carlisle: Hi Dan, how are you?
Dan Rasmussen: Great, thanks for having on Toby.
Tobias Carlisle: My absolute pleasure. So, just so we can understand the strategy that you’re approaching right now, can you give a little background to how you came up with the strategy and how you’re implementing it now?
Dan Rasmussen: Yeah, absolutely. The strategy really borrows from private equities. So, if you look at the broad history of private equity, from 1980 to 2006, private equity was the best performing asset class by a wide margin. And I think 80% of private equity funds that were raised during that period, outperformed the public market equivalent. So a tremendous track record of success. According to Cambridge Associates from late 1980’s until 2006, it was 6% net a fee out performance the public market for the private equity index. So what were the private equity firms doing during that period?
Dan Rasmussen: I had a really unique chance to figure this out when I was at Bank Capital, because we were trying to answer this exact question. What had driven our historical success? What do we need to do to continue it? And when we started to look into it. We found that there were some fascinating elements of what had made private equity work, and it’s gonna resonate very closely with your Acquirers multiple as you indicated in the opening. But, what we found is there are really three characteristics that predicted success in private equity. Well, there were two that defined it and one that predicted it.
Dan Rasmussen: So, private equity relative to public markets; private equity firms are buying companies that are small, generally 200 million of market cap versus 30 billion for the S&P 500, that are levered, typically about 65% net debt to enterprise value verse 10% for the Russell 2000. And 3rd, if you divide private equity by purchase multiple; the cheapest 25% of deals turns in at less than seven times EBITA, accounted for 60% of the industry’s profits. And the most expensive 50% of deals done over 10 times of EBITA, accounted for only about 10% of the industry’s profits. So, in aggregate the story of private equity was buying small, cheap stocks with debt.
Dan Rasmussen: And, if you think about why that worked; it’s small value on steroids. It’s small value times leverage and, gee if you buy something cheap and it’s small, so you’ve got a lot of upside, and you lever it right, no surprise it works. And, it looks like private equity was earning about 6% outperformance of the broader market and they were taking about 6% a year out in fees, so the true outperformance of the growth strategy was about 12% per year. And, so, what I set out to do was say, “Gee, I wonder if we could replicate what private equity had done in the 80’s and 90’s by buying these companies that had the same quantitative characteristics in public markets.” They were small, they were cheap, and they were levered. And probably the leverage is the biggest departure point between me and most other value investors, but I’m sure we’ll talk more about that.
Tobias Carlisle: Well, I think I’m agnostic to the leverage in the company, provided that the operating income is there to support it. It’s not excessively leverage, but I have found, and you perhaps know better than I, whether there’s a tipping point where debt is good up to a point and then beyond that point it’s sort of deleterious to your return. So, do you want to talk a little bit about the sensitives of debt, how you sort of assess whether something is sort of a safe investment, and where you think that the limit might be?
Dan Rasmussen: Yeah, and I would say that the world of safe investments is not the world I play in. You know? I am on one end of the risk and volatility spectrum, and happily so. So, if you want safety, go buy bonds. Where our goal is to outperform and my view is that to outperform, you have to take risks. So, setting that aside, leverage. Leverage, more debt is bad. More debt is bad. If you measure debt as debt to assets, debt to EBITA, debt to interest, any absolute value of metric, you’re gonna find that the more levered a company is, the higher the probability of bankruptcy. And bankruptcy is like getting a zero on your math test in eights grade, right?
Dan Rasmussen: One zero will sink your entire semester grades, so you don’t want zeroes. And debt is what creates the possibility of a zero. And so, what you find is that increasing leverage increases the risk of bankruptcy. Now, what you also find is that just like buying a home or any other asset, if you buy something well and let’s say you buy it for $100 and you borrow 90 of those dollars and you sell it for a $110, wow you made 100% profit on a 10% rise in values. So, leverage amplifies the returns. And so, what you find, I think, of leverage, is sort of a trade off; there’s good leverage which is leverage as a percentage of your purchase price right?
Dan Rasmussen: So, if you think you’re making a good investment, you’d ideally want it as levered as possible to magnify gains when it works. But, on the other hand, if you’re wrong, you want less leverage on an absolute basis. And so, that’s why the intersection of leverage and value is so important. If you buy cheap things with debt, you tend to have the advantages of the magnification and yo don’t have too much bankruptcy risk. But, on the other hand if you buy expensive things with debt, you know, God love ya, it ain’t gonna turn out too well.
Tobias Carlisle: You have a nice strategy in that the debt is by virtue of the fact that it’s raise by the company, the target that yo put into the portfolio, it’s non-recourse to you, so it’s not like your carrying debt at a portfolio level. It’s carried up the holding level. So, in that instance, if you’re going to do it, that’s sort of the way to do it so that any individual stock that might fail doesn’t sort of risk the entire portfolio.
Dan Rasmussen: That’s right, and think you know, this is Robert Schiller won the Nobel Prize for the finding that market prices are 20 times more volatile than fundamentals. So, if you think about where you want leverage, you don’t want leverage on the really volatile price movement of a stock. You want leverage on the balance sheet of a company where it’s dependent on that company’s earnings. And what you find is that when you have leverage there, it’s asymmetric. So, if you say, have a margin loan, you have symmetric exposure. If the markets go up 10 then you’re levered 100%, you go up 20. And if it goes down 10, you go down 20.
Dan Rasmussen: If you buy a portfolio of levered companies, that are equivalently levered, you tend to not quite go up, if say 50% levered, or 50% debt, 50% equity, you don’t go up quite 100% when the market… you don’t up 2x when the market goes up, you go up a little less than that. But when the market goes down you don’t go 2x down, you go only little bit worse, because unless a bankruptcy risk of the company is meaningfully changed, the equity won’t reprice to reflect the fact that the company’s leverage is 50%. And that’s sort of the key insights to making this work.
Tobias Carlisle: It’s a fascinating strategy, and it’s one that you were, I don’t want to say a junior, but an associate, or an analyst at Bain Capital; when you were tasked with this what others drivers of outperformance and I think that he levers might have been operational improvement leverage and possibly the purchase price. So, can you just talk a little bit to that study.
Study at Bain Capital
Dan Rasmussen: Sure. Yeah, so we looked at a whole variety of things. The first thing we looked at was every private equity deal we could look at. So, I think we built a data set of 25 hundred deals, 350 billion of invested capital and private equity, and we looked to a predicted success. And there were a lot of people that thought industry was going to predict success, but industry was sort of an irrelevant variable as it turned out; it was all about purchase price. And, even those other things; size and leverage, every private equity is small and every private equity deal is levered. So, if you’re within private equity they don’t really predict anything, because 200 million of market cap is the average PE deal, right? That is an extreme micro cap in public equities.
Dan Rasmussen: I think there have been only a dozen private equity deals that are larger than the large end of the small cap index. So, you know, buy in large what we’re talking, is tiny little things. So, whether its 400 million of market cap, 200 million, 100 million; it’s all small, doesn’t really matter. Leverage levels, again, 65, 70, 55, doesn’t really matter much at all. Because, everything is levered, so controlling for everything else doesn’t make much of a difference. What really mattered is the valuation. And that was so powerfully predictive. And especially I think in a levered environment, for those reasons we were talking about.
Dan Rasmussen: And we looked at the other drivers, which I think you brought up. I think that private equity firms would tell you that there are really maybe two or three core strengths that they have. One is that they do better diligence; so they spend a huge amount of work, they really know the companies. By the way they have access to private information, so they have knowledge advantage over say, public investors. And the depth of… they own 15 companies rather than 50, so of course they know them better. And then second, that they’re able to improve them, that’s why they have control, they own 100% of the company so that they can sit on the board and make them better. And then, third, and sort of related to that; they have the ability to higher or fire the CEO, and so they can replace them with a better person from their staple of operators. And those sort of key drivers allow private equity in their minds, to outperform.
Dan Rasmussen: And what’s I’ve found is, and this goes back to Phillip Tetlock, a student of Daniel Kahneman, right; more knowledge doesn’t make you a better forecaster. So, great you have access to a data room, great you spent a million dollars on McKenzie, you’re no more likely to predict EBITDA growth than anyone else. And anybody who has every looked at EBITDA forecasts from any buy side firm, will tell you the error bars are so huge as to make even the whole endeavor of forecasting EBITDA growth, worthless. And we broadly found that to be true.
Dan Rasmussen: In terms of operational improvements, is every private equity guy trained at an investment bank magically a better CEO than every public company’s CEO, are they magically better board members than every public company board? Are the McKenzie people that KKR hires that much better than McKenzie people that the public company hires? Right? It’s just not plausible. And if private equity guys were really such better managers, shouldn’t Harvard business school and Stanford be teaching the private equity approach to management? But, they don’t, well why is that? Why is that this is so flaunted, yet they don’t teach it? Well, because the private equity approach to management is lever the company up, that’s it. And then, sit on the board and if things go wrong, bring in McKenzie again and again, or maybe switch to BCG because they have industry expertise. And that’s about as far as it goes and it doesn’t really help. And then third; choosing a CEO. Who the CEO is in my mind, doesn’t really matter. And we can talk more about that.
Characteristics of CEO’s
Tobias Carlisle: Lets dive into that. You’ve more recently released a paper describing looking at the characteristics of CEO’s and whether that is in fact, predictive of performance. So, would you like to talk to that a little bit?
Dan Rasmussen: Sure, I think it goes back to the 80’s and a guy Michael Jensen at Harvard Business school. And Jensen had this idea that he noticed that Harvard Business School students, gasp, were not going into corporate management, the horror. So, why weren’t they going into corporate management? And he asked them, they said, “Well, we don’t get paid enough.” And so he said, “Well, it’s such a disappointment because my students are so brilliant, and if they ran public companies, then cash America would be a better place, because they could pass my wisdom on at their companies, et cetera, et cetera. And so what we need to do is pay CEO’s more in order to attract the best and brightest to corporate management. I mean what could go wrong?”
Dan Rasmussen: And so, his solution, and this is such a common hackneyed phrase now, but he wanted to align incentives, right? And now everybody is so into alignment incentives that you find the VP at the private equity firm is trying to give his nanny a bonus for higher performance or something. It’s like, “Oh dear, right? They’ve really drunk the Koolaid on aligned incentives.” And aligned incentives in the equity world, and then it goes back to another idea; the Milton Freedman idea that all that matters is shareholder stock price performance. And so, why not incentive CEO’s to make the stock price go up? That was Jensen’s logic, it makes sense, right? Tie performance to price, tie salary to performance, and they did that with options, and they did that with essentially massive stock grants to CEO’s.
Dan Rasmussen: And what my research looked at is saying a few things. One, is there any relationships between the best and the brightest, Jensen’s students, being CEO’s and share price performance? And what we found is no. MBA’s are not better CEO’s than non MBA’s. Harvard MBA’s are not better CEO’s than non Harvard MBA’s. In fact, if you look at any sort of pedigree related thing, whether they were a banker or consultant, even whether they founded the company; there’s no statistical relationship to the stock price performance. In fact, there’s no statistical relationship between incentive pay and performance either. Right? I mean, and in fact, even if you look at historical performance; so just say, “Okay, well there’s gotta be some great CEO’s, so let’s look at CEO’s whether their three years predicts the next three years. If you’re a great CEO, your greatness should be persistent.” And we found no relationship there.
Dan Rasmussen: And then we said, “Well what about the CEO’s that do a great job at one company, or a really bad job at one company, and get hired to be CEO at another company?” No relationship there either. There is essentially no relationship between who the CEO is, anything about the CEO, or how the CEO gets paid, and what happens to the stock price? And so, now we’re sort of arguing for a null hypothesis, that none of this matters. You can’t prove a null, you can only defend the null, but there’s shockingly no convincing evidence to suggest anything about who the CEO is, from how they’re paid, who they are; that’s any relationship to equity markets. And so, Jensen’s logic, while it sounded good, was entirely wrong.
Tobias Carlisle: Jensen wrote a series of papers where he was suggesting that the importance for buyouts and takeover in the 80’s, was the free clash flow that these companies were generating, which seems like a pretty trite observation to most investors, but naturally groundbreaking in an academic sense. And then resisted fiercely, even though the company throws a free cash flow, so therefore it can support debt. Pretty clear observation. It reminds me of Buffett’s comment that, “Where the CEO with a reputation for brilliance tackles the company with a reputation for poor performance, it’s generally the reputation of the company that persists.”
Dan Rasmussen: Yes, yes.
Tobias Carlisle: So, you were a Harvard undergrad, then Bain. Stanford MBA, and while you’re at Stanford did you start formulating the idea for Verdad and is that… I know that you were at Bridgewater somewhere in there, I’m not entirely sure where Bridgewater fits in.
Dan Rasmussen: I interned there in college, yeah. And a very short period of time, but very influential in my thinking. But, yeah I came up with this idea while I was at Bain Capital and I said, “Look, you know, I wanna go do what Bain Capital, and KKR, and Blackstone did 20 years ago. Not what they’re doing now, I don’t like what they’re doing now. I want to do what they used to do, and just copy it, because it worked.” And I think most people that create good things are just copying better ideas from other smart people. There are no new ideas, only implementing good ones from the past.
Dan Rasmussen: And, so I decided to go to Stanford because there was this guy Charles Lee who was a professor there, who was just absolutely brilliant. He’s a hardcore quant, teaches a class called Alphanomics, just all about quantitative investing. And I said, “Well gee, I have this very simple insight, which is buy cheap, small, and leverage.” And though I’m not a quant, I’m not like a super algorithm guy, those are all quantitative things. Right? Size is quantitative, value is quantitative, and leverage is quantitative, and so what I want to do is figure out how those interact, how those work, and try to find the principles of what works. And then also take all the other knowledge from quantitative investing and then apply it onto those variable and see how we can improve. So that was really my idea in going to Stanford and it was a great decision because Charles Lee is an amazing guy and an amazing mentor.
Tobias Carlisle: Its funny, I came to it in a possibly in a similar way, to you did. I was a junior attorney working on a lot of private equity deals, sort of happened in the early 2000’s. And it’s an enormous amount of effort to take something private, there’s a diligence process, and there’s a lot of paper that’s debt and equity and lots of things to be considered. The company’s taken private, then you’ve got this enormously liquid asset that you can’t shift easily, and you pay a premium when you take it private. And at the same time, I could look at companies that were listed on the stock market that you can buy for virtually no effort, open your brokerage account and up and buy them, and then don’t pay that premium and get at least equivalent returns, and likely better returns for less efforts.
Tobias Carlisle: So, I think that the strategy is, I’m sort of talking my own book a little bit, but I think the strategy’s a really good one and it’s likely to perform very well, but it’s still a value strategy. And so, it’s been a very tough time for value. How do you feel about it relative to something like the value factor or EVE which is probably a reasonably good proxy? Are you gonna track that closely, or do you think that you’re gonna be able to out perform? Or how do you think about it in those terms?
Dan Rasmussen: Yeah. I think there are two dimensions to it. I think one, is we have leverage right? So, in theory if we both own something for the same price, but I’m 50% levered, I should do the whole… I’d say, broadly levered small value performs like small value on steroids. So, so there should be a magnification effect to investing in the levered portion. So, I think that that broadly, theoretically makes sense, now, it’s gonna be painful when things go down, but it’s gonna be really nice on the upside. So, that’s sort of the trade off.
Dan Rasmussen: And I think the other thing that’s worth thinking about, with all value strategies is, and all quant investing is in some sense a ranking, right? YOu’re saying, “Well I like cheap things, and so well the 10 cheapest things should be better than the next 10 cheapest things, should be better than the next 10.” And what’s sort of interesting is, sort of, if you look at the distribution of the markets, most of the really cheap things are also really small, for sort of related and obvious reasons. But, what this means is that if you want to be a very disciplined, extreme value investor… If you say, “Gee, I really love EBIT to EV.” And I do, right, it’s a great signal. … And you wanna own the 50 cheapest EBIT to EV companies, probably the average market cap on those, is like 200 million.
Dan Rasmussen: And so, the other sort of question you have as a sort of factor investor or quantitative investor is, “Am I willing to run a fund that can invest in those things? Am I willing to know, and tie my hand behind my back and know that I can’t ever get bigger than say, 200 million of capacity because then I couldn’t buy the very things that define the value universe.” So, I would say, when I think about what I do and how it compares to the broader universe of small value, I’d say there are really two differentiators; one is the focus on leverage, which hopefully should amplify things. And then the next is sort of, a very conscious commitment to staying very small and focusing on the extremes of the value factor, which I think is also really important to generating Alpha.
Determining Which is Likely to Out Perform
Tobias Carlisle: I think you touched on it a little there, but you also have a very recent paper, which is a fascinating read, about looking at the very cheapest and then seeing if you can determine which, cherry pick out from that group, or determine which of those are likely to out perform the others; which is sort of the Holy Grail, if you can really figure that out, then you’ve got something special. So, can you describe the paper and what were your findings?
Dan Rasmussen: Yeah, so that would be the Holy Grail, if it were possible Toby. So, we’ve gotten really into machine learning. And I like machine learning because it’s sort of Bayesian, it’s probabilistic. And, it’s sort of in my mind, after linear factor models, I think machine learning is sort of the next stage for quants. And so, we thought, well what’s the first way? If you’re looking at quantitative research, you know that the linear factors are the most important, right? They show up in regressions, of course they matter most.
Dan Rasmussen: So, what you want to do is start with those linear regressions and then layer on machine learning on top. So, what we naturally said is, “Okay, well let’s start with the linear model,” and we know what small value is, so does everybody else, “let’s look at the extremes of small value, which are really what drives the Alpha, and then lets look within in that to try to pick out the things that don’t work.” So, what we did is we took the one third of worst outcomes. So, where the linear model said it’s gonna expect a return of 30% and it actually had a return of negative 50. And we tagged those third of worst outcomes going back 25 years, in the U.S. and Europe.
Dan Rasmussen: And then we said to the machine learning algorithm, we gave it everything we could think of, everything; and we said, “Go and tell us why the model’s wrong.” Why does small value produce such high error rates and such high dispersion? Why does some small value stocks do really well, why are others value traps, and why do others go bankrupt? And our hope was that we could more finely tune our factor model, right? And so, we came back and we were delighted. The machine learning said, “Hey we can achieve something close to 50% accuracy at telling you what are the 30% worst outcomes.” I said, “well great, look at how wonderful…We’re so brilliant, this model’s great.”
Dan Rasmussen: And so he said, “Well tell us what the biggest predictors, of your essentially, probability of being wrong, what are the biggest predictors?” The number one was the linear regressions expected return variable, so basically the model was telling us; the higher the risk, the higher the return. And in fact, the higher the realized return. So, if you purely said, “I wanna only buy the things that are most likely to be wrong.” You had the highest returns relative to the stuff the lowest likely at being wrong. So, there was almost like a beautiful proof of market efficiency.
Dan Rasmussen: Now, we did find that in some extremes, probably the 5% of most extreme probability of being wrong, the machine learning model was actually really good at identifying really bad things. And it was also good at sort of identifying things which might be a little bit less risky than their price implied. So, it does improve a factor model, using machine learning does definitely improve the factor model, but for the most part, it’s a very incremental improvement. Okay, it’s machine learning isn’t magic, nothing in investing is magic, but the linear regressions are pretty darn good. And markets are pretty darn efficient, but machine learning can help you fine tune those things. But, by and large, even with the most advanced tools, you’re gonna find that things like price matter most.
Tobias Carlisle: We took a similar approach in quantitative value. We looked at that cheapest decile of EV, EBIT, and then tried to divide it into two halves, because it’s already a fairly small universe of stocks. And so, to get a sufficiently large portfolio, you can’t really divide it much more than in half. We looked at a variety of things; margin strength, and so on. And that’s one of the departures between Wes and I, is that I just prefer cheaper, and Wes prefers cheap and good. I think that, just changing gears slightly, I first read your work without realizing who you were at the time, was when you wrote a critique of Porter’s Five Forces. It’s a wonderful paper. Can you just… What is your critique of Porter’s Five Forces?
Critique of Porter’s Five Forces
Dan Rasmussen: Well, to put it very simply, my critique is that there’s no evidence that it’s right. I think you and I are in this very controversial school of investing that you call evidence based investing, which is the revolutionary idea that maybe your ideas should be supported by evidence. And, unfortunately, Michael Porter’s ideas just aren’t. They sound nice, they sound good, they’re just wrong. And what’s sort of fascinating, what’s sort of most interesting in some sense about Porter’s ideas, so you have to kind of go into the DNA of his ideas. So, Porter studied in the field of industrial organization, and the field of industrial organization in the 60’s and 70’s was very focused on anti-trust and monopolies.
Dan Rasmussen: So, they had this idea that monopolies were evil and bad, because they could screw everyone over, and thus earn really, really high profits. And in the industrial organization field, there was this theory called Structure Conduct Performance; so they said the industry structure determines the firms conduct, which determines performance. And let’s make it even simpler than that. Basically, they said, “The closer you are to a monopoly, the higher your margins will be.” So the higher your market share, the higher your margins will be. So, if you want to study business, what you should be studying is industry structure and figuring out which companies have monopolies. And there’s something very intuitive about that right? You sort of think, “Wow, gee, if I owned all the railroads, I could charge a big price for my trains.” That hasn’t worked well for Amtrak, but again we’ll get back into the evidence a little later.
Dan Rasmussen: So, Porter was sort of a disciple of this school and he came to Harvard Business School in the 80’s and Harvard, business school has a checkered reputation as an intellectual field. When Harvard Business School first opened, people were saying, “What are you gonna do, you gonna bring in cobblers and butchers and chefs, an teach them…” I mean it just seems like not something Harvard should be doing, you know teach them Latin. But, at the time, in the early 80’s, Harvard Business School’s approach to teaching corporate strategy was to teach people SWOT analysis; strengths, weaknesses, opportunities, threats. And I think to anybody with half a brain, that sounds like a dumb idea, or at least in the realm of, “Really, I go to Harvard Business School and you’re teaching me SWOT analysis?”
Dan Rasmussen: There was something lacking, and Porter came in. He said, “Let’s overhaul the strategy curriculum and teach Structure Conduct Performance. And we’ll teach them how to take Structure Conduct Performance and apply that to corporate strategy.” And that was the origin of the five forces. So he said, “The more market power a company has, the higher the market share, the closer it is to a monopoly, the more power, the more force it should be able to apply, relative to its competitor, suppliers, et cetera.” Right? And that was the idea, and it should be no surprise that the idea of an advantaged, powerful establishment dominating everyone else, so appealed to generations of Harvard Business School graduates.
Dan Rasmussen: And this became the dogma of business, and sadly… and I felt Warren Buffett a little bit here, because he adopted this wide moat concept, became a dogma among value investors; who started to say, “Ah, well we don’t just want cheap companies, we want cheap companies that are competitively advantaged, that have these five forces.” So, it caught on like wildfire, and yet at the same time, it’s really interesting, this whole idea of Structure Conduct Performance, was also being applied in the legal system and in academia. Right?
Dan Rasmussen: So in the legal system, this was being used to break up monopolies and break up companies with high market share. And the Supreme Court, in a series of landmark decision in the early 80’s, they basically looked at a bunch of sophisticated econometric analysis, and found that there was no evidence that higher market share led to higher margins, no evidence whatsoever. There never had been. And so, they said, “You know what, you can’t use market share as an indicator of anything anymore. It’s not ipso facto, anti-competitive to have high market share. There’s just no evidence of that.” I mean, you actually have to prove consumer harm, and that was what the Supreme Court said, so basically, in the early 80’s, Supreme Court said, “Structure Conduct Performance, as far as the courts are concerned, is dead. We don’t believe it, we don’t buy it, we’re not gonna apply it. So, keep it out of the legal system.”
Dan Rasmussen: And at the same time, the field of industrial organization was starting to do all these market share margin studies, industry margin studies; and what they ended up finding was that industry had no relationship to conduct or performance. And market share had no relationship to conduct or performance, just the legal scholars are finding. And by the late 90’s the field of industrial organization had basically admitted that industry analysis was dead, traditional industry analysis didn’t matter, there was no evidence for it.
Dan Rasmussen: And this is at the… Porter’s star is rising and rising and rising. I think the economists or Forbes named him global guru. Which, you pretty much know anyone who’s a global guru is wrong. But, it had really, his star had peaked. And what I wrote, a piece is just saying, why are value investors still so devoted to this cant? Which is what it is, it’s cant: it doesn’t work, isn’t true, and isn’t even basically supported by logic. Which, to what you were saying, is it buy cheap and good? Well, no, it’s just buy cheap things, and you try to buy good things, you just move away from buying cheap things, and then it works less well.
Tobias Carlisle: It’s a little reminiscent of another guru, Tom Peters, who had that book, In Search of Excellence. And, he said these are the criteria for excellence, and it included various things like; a high return on invested capital, high sales growth, and so on. And an analyst, Michelle Clayman along and said, “Well, let’s go in search of the un-excelled.” And she found, that the companies that had the characteristics on the other end of the spectrum, which were; very low returns and invested capital, little to no growth in earnings, they outperformed quite substantially, the excellent companies, so called. And that study was updated by a gentleman who worked for her, his name escapes me now, but I put it into deep value.
Tobias Carlisle: Basically, over the full data set, it’s a stunning outperformance for the unexcellent companies, and the reason is very simple; the excellent companies, you pay two times book for them on average in this data set. And the unexcellent companies you get them for .6 times book. In both cases there’s this sort of diminution in that the businesses get worse after you buy them, in the 12 months after you buy them. But, so the driver of the performance is purely that reversion in the price to book, or price to underlying intrinsic value. It’s one of those fascinating things, I’ve been saying it for a long time, but I can’t find many value investors who are prepared to, or who believe it at all, so I’m always very happy to find a fellow traveler.
Dan Rasmussen: No, it’s exactly right. And I always say investing is a game of meta-analysis, not analysis. So, it doesn’t matter if you think a company is good, if you everyone else thinks it’s good. It’s what you think of, relative to what the market thinks. And so, ultimately, a company’s price today depends on everyone else’s projection about what’s going to happen in the future. The sort of, if you wanna be a meta analyst investor, step one: find things that everyone agrees are bad, and then among those figure out which ones you wanna own. And understanding that consensus pessimism, if the future is completely unpredictable, then both extremes of optimism and pessimism are gonna be wrong. So, what you wanna own is the things people are pessimistic about; the unexcellent things.
Dan Rasmussen: And I think that’s just logical, but I think part of it is, you try pitching this to sophisticated institutional investors and they don’t wanna hear… You’re like, oh what do, “I buy really bad companies that everybody else hates, that might get a little better. I don’t have any reason to think they’re gonna get better, I just sort of think, who the hell knows what’s gonna happen in 2020. Maybe newspapers won’t be as bad as everyone thinks they’re gonna be.” And people are like, “Well, over here, this guy who’s really done a deep dive on sales force, and boy is there a bright future for sales force. And they own 20% of their funds sale, they have super high conviction.”
Dan Rasmussen: And I was joking with one of these large institutional investors, I said, “You can either have higher returns or you can buy things that you have really high conviction on. But, the universe of things that are high conviction, high return is a null set. You either take risk and get return, or you buy things that everyone likes and you get mediocre returns.” There’s just no more, sort of, obvious truth in investing in that.
Tobias Carlisle: The Venn diagram doesn’t really overlap very much in that.
Dan Rasmussen: Exactly. We wish it did, but it doesn’t.
Tobias Carlisle: Buffett has sort of been a proposed…as you pointed out, the reason for this is that Buffett, who is the highest profile, most successful value investor, possibly we wouldn’t even know the term if Buffett hadn’t been so high profile, and been so generous with his writing. The problem though, is that if you try to do it yourself, you find it extremely difficult to do that. And I think Michael Mauboussin has this great study where he took buckets…he could rank say the Russell 1000, he could rank them from highest return on invested capital, to the lowest return on invested capital. And then I think he puts them into qintar, so one fifth each. And he tracks them over 10 years to see what they do. And, as you’d expect, they sort of have this mean reverting function where the ones that are the worst, tend to get much better, and the ones that are very best tend to get worse. Reason’s very simple of course, it’s because everybody wants those very high returns and they compete for them, and nobody wants to be in the industry that has very low return and they leave and that sets them up to do better.
Tobias Carlisle: Maubousssin’s looked at the drivers of those returns, and he’s never been able to say prospectively which of the companies, what the drivers of it are. He can come up with conclusions at the very end where he says, “Buyer technology, Pharma did very well, you don’t want to be in retial, you don’t want to be in anything that’s got these very high…. you want higher margins rather than lower margins. But, nothing is sort of particularly predictive.” I find is fascinating that this research is out there, and for whatever reason… sadly I have these arguments on twitter all the time. There’s a recent, recent paper on unpopularity. Did you see the–
Dan Rasmussen: Yes, I loved that. I thought that was so good.
Tobias Carlisle: Capturing that again, where Morning Star has those three categories, where they say there’s a wide moat type of company.
Dan Rasmussen: Same story as the Porter stuff.
Tobias Carlisle: Wide moat, narrow moat, no moat, and I love the Morning Star definitions of these things. I think it’s an excellent description of what you would look for in a moat. I don’t think there’s anything wrong with their method, I think it’s a very good method. It’s just that when you look at the returns to those three categories; no moat outperforms narrow moat, narrow moat outperforms wide moat. And the reasoning is always, “Well it’s because narrow moat is riskier, sorry, no moat is riskier, narrow moat is less risky, wide moat is the least risky.” But then, I always think, if we did a Monte Carlo simulation of it, shouldn’t we get equivalent returns across all three, because we’re getting failures in no moat, and wide moat sort of persists.
Tobias Carlisle: It doesn’t seem to be the case. It’s sort of accepted so well in the literature, and then it manifests in these discounted cash flow models that everybody builds. Because, they say this particular company that I’m looking at has this very high return in invested capital. So, therefore should be able to grow and compound, and it’s gonna grow over this 10 years or so, and then the terminal value is gonna be enormous, because it’s gonna grow in perpetuity; sort of at a slightly higher rate than GDP. Are you a proponent of DCF’s, do you use them in your firm?
Dan Rasmussen: Well, yeah, there’s the old joke that an economist and an engineer are on a desert island and they’re starving, and there’s no food, and a can washes up. And the engineer says to the economist, “Well what are we gonna do?” And the economist says, “Well I have an answer.” And the engineer says, “Well what is it?” And he says, “Well assume a can opener.”
Tobias Carlisle: I love it.
Dan Rasmussen: You know, in theory, discounted cash flow is right. Predict the entire future, discount it back based on it’s riskiness, and then that’s the value of any security. But, what alludes me is why investors think they can predict the future in the first place. If you say, “Well predict how much money is gonna be in your bank account in ten years.” Well you’re the worlds leading expert on you, surely you should be able to develop a very sophisticated excel model, that should give you a pretty precise answer to that. But nobody does it, because we all know it’s impossible. You know, Coca Cola is a much more complex system than you. Why do you think Coca Cola’s balance sheet should be predictable? I mean, it’s just nonsensical, but it starts from the premise that, if you can predict the future yeah it’s right, but you can’t predict the future so it’s wrong. So, it’s like the first premise is flawed.
Dan Rasmussen: And yet again I think business schools are to blame for this. The DCF models, again, it’s a Harvard Business School idea from the 1930’s. For some reason, people wanna plan, and they wanna say, “Well, what should this be worth?” And I think the answer is, no one has any idea. We don’t know what the future is gonna hold, so you might as well buy it cheap.
Tobias Carlisle: I don’t mind the Beau Williams model as a statement of what you are looking for, that is true that something is worth now, the cash flow is discounted from here until kingdom come, that’s absolutely true. The difficulty is in implementing that theory in any practical way, in any individual company. It’s virtually impossible to do that. The other thing, and I’m sure that you have built these incredible complex Excel spreadsheets with multiple tabs, all sort of linking through, projecting out margins and growth and so on, and then discounting that back at whatever is the correct discount rate, I have no idea. But, they all sort of boil down to this; you only really have a handful of inputs.
Tobias Carlisle: It’s the future, it’s the dividend or the cash flow that you expect over the next few years, the discount rate, and the growth. That’s three very simple inputs that are then expanded out, exponentially. This is another thing, I get trolled about all the time; I’m anti-DCF, but only because there are only a handful of inputs. And they say, “it’s silly to use a multiple, it’s silly to use a ration when you can look at these other…” Well, really the only thing that the ratio is missing is the growth rate, and I don’t know.
Dan Rasmussen: Right, well I think there is a two by two. Right? There’s importance and know ability. And I think what investors miss is the know ability portion. So growth rate, really important, unknowable.
Tobias Carlisle: Right.
Dan Rasmussen: If we knew the growth rate, we’d know the value of the company. Well we don’t know the value of the company, it’s unknowable. So you multiple that in the equation and it comes out to zero right? Cuz know ability is zero, and so even though importance might be a hundred, if know ability is zero, the value of that variable is zero. Multiple on the other hand, is both very important and very knowable, so you multiply those together and you get the answer right. I mean, that’s the logic why people miss the know ability portion of it. And that’s the fundamental mistake people make. They think that the future is predictable when it isn’t. And I think there’s no more fatal intellectual flaw, than to believe that you are a fortune teller, or a prophet, or a sooth-sayer; when in reality, you’re just a guy with a spreadsheet.
Chicago School of Business Approach
Tobias Carlisle: So, I’ve seen you describe your approach before, as the Chicago School of Business approach. What do you mean by that and what does that mean in a practical sense?
Dan Rasmussen: Yeah, you know the sort of joke around here, we call ourselves part of the Chicago’s. Well I think the Chicago School, to me, means sort of the disciples of Eugene Fama. And generally, people that think it in terms of evidence. I mean, the Chicago School of Economics was famous for saying, “No, we need to prove this works in the real world. We need some empirical evidence.” And for example, on growth rates, there’s a great, I think it’s a 2004 paper, I think it’s by [foreign 00:42:20] and it’s–
Tobias Carlisle: That’s the contrary in investing.
Dan Rasmussen: No, it’s a different one.
Tobias Carlisle: Not contrary in investing?
Dan Rasmussen: Maybe I got the authors. It’s the Persistence and Predictability of Growth. And, I might have the authors wrong, but it’s The Persistence and Predictability of Growth, and they find that growth is neither persistent, nor predictable. So, if you actually look, and you actually try to predict it, you can’t. And nobody can, and they haven’t been able to. Right? And I think for me, when I say Chicago’s School that’s for me, it’s saying, “Oh, okay. You’re gonna build a DCF model, well before you build it, why don’t you prove that you can forecast the growth rate. Prove that you can forecast the growth rate, then let’s do the DCF model. Don’t just build the DCF model assuming you can break the growth rate.” And I think that, for me, is what the Chicago School is.
Dan Rasmussen: And I think it’s funny right, because people will say, “Why don’t you do more analysis? Why don’t you meet with management? Why don’t you forecast the growth rate?” And I say, “I’d do it if it it worked. You show me evidence that it works and then I’ll do it.” But I’m an evidence based investor, so I don’t do things that I don’t have any evidence or logic to think that they work. And I think the most common conceit, is that some people believe that the more you know about something, the better you can predict the future, about that thing. And I don’t think that’s true, in any meaningful way. I don’t think that just knowing more about something make you better at forecasting it. There are a lot of people that really know a lot about baseball, but it doesn’t mean that they can predict who the winner of the World Series is, anymore than anyone else can, because it’s unpredictable.
Tobias Carlisle: James Montier has a great collection of these studies that show all of the various different ways of predicting the future. He has a great one on horse handicapping, where he gives people, you get a small amount of information about which horse is likely to win. Then the professional handicappers rank the horses, then they give them increasingly, more and more data about these horses; they sort of randomize it so no individual is getting the same data at the same time. And it demonstrates two things; one is that we tend to anchor on the first bit of information that we see more than any other, and the other is that we get increasingly confident with each little bit of data that we receive about the horse that went right. But our accuracy doesn’t improve at all beyond the sort of one, two, or three pieces of information that we receive at the start. It’s kind of a… And then that is replicated over and over again through the literature.
Tobias Carlisle: I think Paul Mele said something like, “When you see this phenomenon replicated that many times, it sort of becomes this golden rule.” And the golden rule is that simple statistical models do better than experts, which you and I sort of seem to embrace. But, I’ve yet to see it really penetrate the investment world, other than the quants, who sort of seem to construct very large portfolios. When you go about constructing a portfolio, how many positions, how diversified, how concentrated, what are you looking to achieve?
Constructing a Portfolio
Dan Rasmussen: Yeah. So, I tend to like the 40 to 50’s stock range. So, I think that, in theory you have to embrace the idea that your models are good, but the world’s unpredictable. So, the R squared of even the best model is not that big. And again, as we talk about the extremes of cheapness, you have really high dispersions. So, you wanna hit enough, you wanna see enough pitches that your statistical insight works out. However, on the other hand, you’re balancing that with of course, you have a ranking system. So, is the stuff that you can buy at three times[inaudible 00:45:47], then three and a half, then four, then five, then six, the more things you own, the less of a hit you’re getting from buying cheapness.
Dan Rasmussen: For example, whatever it is you’re ranking on. And so, my view is that you have this trade off where you want enough names that you have at least some diversification, but not enough that your alpha starts to deteriorate, because you’re looking too much like the index or diluting your factors exposure. So, I think it’s a compromise between the two things. I think there’s no perfect answer, but I think obviously factor investing, in the big realm of things; should a 50’s stock factor investor be better than a 500 stock factor investor be better than a 15 hundred stock factor investor? Of course. Right? Of course, if the factor is right, the guy with 50 is gonna beat 500, beat 15 hundred. Is 70 or 30 the right answer? I don’t know that I know, or anyone really knows.
Tobias Carlisle: It sort of becomes this slightly, it’s almost a nihilistic approach to investment, just to say that the only things I’m gonna use are the the things that I can. You know, I prefer historical earnings to afford projections, so on. So, I want things that have been printed in black and white, that have been recorded presumably because they actually happened. But it’s sort of, the approach becomes this Tetlock. If you embrace Tetlock, if you embrace the behavioral arguments about why everybody else is so bad at investing and then potentially… and you’re clearly a very intelligent guy; you’ve got Harvard undergrad, Stanford MBA–
Dan Rasmussen: Or maybe I’m just a really good actor, and I’ve learned to read my script very well Toby, and tricked you.
Tobias Carlisle: Who really knows you could be picking the stocks, but what do you take from the… Tetlock has the fantastic book, it’s Tetlock Super Forecast? I’m forgetting now.
Dan Rasmussen: Yes, yeah that’s right. Yeah.
Tobias Carlisle: Tetlock is su–
Dan Rasmussen: That’s right.
Tobias Carlisle: …So, what’s the Tetlock story?
Dan Rasmussen: Yeah. So look, I think at it’s heart, forecasting, that is what we do. Right? That is what investors do, that is what investment models do, they forecast prices. They say, “These are the things where the price today is the lowest relative to the forecast price.” Very, very simple. We are all forecasters right? And so, it seems logical to me, that if you’re a professional forecaster, which is what we all are, you should probably study forecasting. Well, there’s only really one scholar, I mean their Mele, and there’s Tetlock basically right? There’s Kahneman’s school of thinking, but there’s basically they all agree on a few sort of, simple findings about forecasting.
Dan Rasmussen: One, is that when you forecast, you want to forecast like an insurance actuary. If you have two alternatives, how long is Toby gonna live, well I’m gonna send in McKenzie, they’re gonna spend three months with him, they’re gonna do a projection of how much he works out, what he eats, we’re gonna forecast into the future. We’re gonna 50 slide power point deck, we’re gonna interview everyone he’s ever known, and we’re gonna come up with the answer to how long he’s gonna live. And the insurance guy just says, “Okay, he’s a male, he’s this age, does he smoke or not? How much doe he drink? What’s his family history of chronic illness?” Plug it in to a model and say “Here’s the answer.” And Tetlock basically says that’s the better way to do it. Because you want to use probabilities, base rates, historical statistics.
Dan Rasmussen: So, you want to orient yourself around what the historic incidence of something has been. That’s the best way to make decisions. You say, “I’m gonna renovate my house. How long is it gonna take to renovate my house?” You know, one way to do it is ask an architect, think about what it is, how long is gonna take. Otherwise just say, “How long have similar renovations in this area or similar size, taken?” That’s the base rate approach, and it tends to work much better because you’re actually basing it on data. And this is why the statistical models work better than the experts, because the statistical models are just saying, “Take all the historical probabilities and assume that.”
Dan Rasmussen: Now, I think there’s a role for, not expertise, but analysis, because someone has to say, “Okay, well what is the base rate?” What variables are important? How do we define that actuarial table? There’s an element of nuance to that, and I think there’s also some element of saying, “Does this fit within the base rate, or is there an acception?” I think Tetlock would even say, “At the end of all that statistical work, you might wanna adjust based on some specific details.” So, if you’re actuarial table says, you’re gonna live til you’re 80, but we know you’re going to do base jumping next week, we might adjust the model just a little bit. Right? And that would be sensible and logical.
Dan Rasmussen: And I think that’s what I take away from Tetlock, which is; you have to use base rates, if you’re not doing base rates, if you’re not using statistical models, you’re not paying attention to what we know about forecasting and good forecasting, and if you’re not paying attention to what we know about good forecasting, why are advertising yourself as a professional forecaster in the first place? It beats me.
Tobias Carlisle: That’s an interesting idea that you bring up there, because there’s a phenomenon that’s sort of, I’d describe it as broken leg theory; where you have some information about whether John goes to… and I think this is from a Paul Mele paper, I think he gives this example where he says, “You have an estimate about whether John goes to the theater on a Friday night. And you might include such things as, it’s an action movie and he likes action movies. It’s raining and he doesn’t like to go out in the rain.” And then you can find the statistical guess as to whether he goes out.
Tobias Carlisle: Let’s say on this particular occasion, he’s got a broken leg. Should you be allowed to update the model, or override the model rather, to reflect the fact that he’s got this broken leg? And the answer tend to be no, for the reason that we find many more broken legs than there actually are. And it’s particularly apt in a deep value world, because every single company I look at has a broken leg, which is why it’s cheap. That’s why you get there in the first place. So, if I was to override the model, I wouldn’t have anything in the portfolio, you know? Because I think the businesses are as junky and as cyclical and everything, as everybody else. I just have to rely on that base rate bailing me out.
Dan Rasmussen: I talk to respective investors a lot, it’s part of my job. And the one question I dread the most, is when they say, “Tell us about some of the companies in your portfolio.” And I’m just like, “Oh no. And I’m like, the more you know about the companies, the less you’re gonna like the strategy. Just hold your nose, blindfold yourself, and buy. Because, do you really wanna know that the largest holding is a Russian steel company, or would you rather not have known that?”
Tobias Carlisle: There’s two reasons to hate it. It’s Russia and steel. That’s why it’s cheap.
Dan Rasmussen: Yeah that’s why it’s the cheapest darn thing in the world. But, it’s funny. I think that’s exactly why, and I think people make decisions so much more on stories and emotions than they make it based on data. And that’s what I think quantitative investors are fighting, right? Because someone’s gonna make a decision to invest in a stock or invest in a fund, based on an emotional or person connection. Not necessarily based on reason or that sort of, what Kahneman calls The Type two, or type logic. And I think that’s often a challenge we face, with being quant investors.
Dan Rasmussen: And that’s why I think work like what you’re doing, is so important, because we need to tell the story of why statistics work. We need to talk about things like the acquirers multiple, and say, “You know, if you actually walk through he intellectual journey, you’ll start to come to an emotional connection with this way of thinking. And if you come to an emotional connection with this way of thinking, then maybe you’ll act based on reason.” But you have to do so much work to overcome people’s bias, because it would be so much easier to walk in and say, “What I buy is great companies, the best companies in the world.” And you know, “Gee, it’s not just important to buy the best companies, you gotta know them better than anyone else. So, we put more work and more money into understanding these great companies and why they’re great, than anyone else. And blah, blah, blah.” And that’s what so many people pitch, and by and large that’s where the money flows to those types of people, and to me it’s just crazy because it’s just not likely to work.
Tobias Carlisle: You have to find the good story to sort of explain the data you’re using, which is what I’ve tried to do in the books. Here’s the story that you can remember, but the real message is the underlying data that tells you how it works. But I’ve seen many examples, and I’m sure you have too, where there’s no limit to the amount of research that a firm can do. Big firms that have multiple analysts and can send them out. And I know that, for example, I won’t mention the name of the firm, but a local firm in Los Angeles had a very big position in Sino-Forest, a $100 million position, established weeks before the fraud was uncovered. And they had done as much as send an analyst to China to go and have look at the forest and they’d been taken out and shown, here is the forest that forms part of the portfolio. The only problem was that that particular forest wasn’t actually in the portfolio.
Dan Rasmussen: It was just a forest. It was a nice forest. I think back to my days in private equity, and we’d go on these factory tours and you’d see all these machines, and you’d be in your suit and go, “Oh that’s a nice machine over there.” You know? “Great factory, really great.” And what do I know? I can’t tell the difference between a good factory and bad factory if my life depended on it, but yet somehow that was part of the diligence process. It’s just so silly. But, you know, the other thing people miss is that the bigger you are, and thus the more you can spend on research, the fewer opportunities you have to invest because there are so many fewer big companies than there are small companies. And the bigger amount of capital you have to deploy, the more constrained your opportunity set is.
Dan Rasmussen: That’s why an individual investor actually has an advantage over a point 72 or a Bridgewater, when it comes to choosing stocks. Now, maybe there are other areas that they are really advantaged in, but if you’re trying to pick stocks, the smaller you are, the more options you have to choose from. So, even if you have slightly worse analysis than some brilliant hedge fund, you are so advantage by being able to invest in small stocks, relative to being only able to choose from the stocks of 10 million dollars of daily volume. I think it’s an interesting sort of truth about markets that research budgets can’t overcome deficits in size.
Tobias Carlisle: Dan, absolutely fascinating discussion. Really enjoyed it. If folks want to get in touch with you, how would they find you?
Dan Rasmussen: So, I’m on Twitter. @VerdadCap is my Twitter handle. I have a website, www.verdadcap.com. We have a weekly research newsletter, which I promise is controversial, occasionally polemical, but hopefully always empirical.
Tobias Carlisle: I’m a subscriber. I think it’s absolutely wonderful. Dan Rassmussen Verdad, thank you very much.
Dan Rasmussen: My pleasure, thanks Toby.
For more articles like this, check out our recent articles here.
Don’t forget to check out our FREE Large Cap 1000 – Stock Screener, here at The Acquirer’s Multiple: