In this episode of The Acquirer’s Podcast Tobias chats with Michael Mauboussin. He’s the Director of Research at Blue Mountain Capital Management. He’s been an adjunct professor at Columbia for more than 25 years, earning multiple awards for doing so. He’s the Chair of the board of trustees of the Santa Fe Institute, and he’s a prolific researcher and writer on all things investment. During the interview Michael provided some great insights into:
- The 3 Most Powerful Ideas From Expectations Investing
- How Should We Calculate The Cost Of Capital?
- The Importance Of Base Rates In Valuation
- It’s Not About Growth Or Value It’s Value Creation – Buying Things For Less Than They’re Worth
- Companies Have Focused Too Much On Growth And Not Enough On Returns
- The Past 10 Years Have Shown That M&A Deals Can Create Value For The Acquirer
- Regression To The Mean Has Such A Powerful Influence In Investing
- You Have To Earn The Right To Use A Multiple
- Why Are Some Companies Able To Resist Mean Reversion?
- The Inside View Vs The Outside View
- The Paradox Of Skill
- If You’re The Stronger Player, You Want To Reduce Variance
Books and Papers Mentioned:
The Outsiders (William N. Thorndike)
Thinking Fast and Slow (Daniel Kahneman)
The Psychology Of Prediction (Daniel Kahneman, Amos Tversky)
Fooled By Randomness (Nassim Nicholas Taleb)
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: When you’re ready sir, let’s get underway.
Michael Mauboussin: I’m ready.
Tobias Carlisle: Hi, I’m Tobias Carlisle. This is the Acquirers Podcast. My very special guest today is Michael Mauboussin. He’s the Director of Research at Blue Mountain Capital Management. He’s been an adjunct professor at Columbia for more than 25 years, earning multiple awards for doing so. He’s the Chair of the board of trustees of the Santa Fe Institute, and he’s a prolific researcher and writer on all things investment. I can’t wait to talk to him. We’re going to talk to him right after this.
Voiceover: Tobias Carlisle is the founder and principal of Acquirers Funds. For regulatory reasons he will not discuss any of the Acquirers funds on this podcast. All opinions 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 Michael, how are you?
Michael Mauboussin: I’m great Toby, how are you?
Tobias Carlisle: So much the better for chatting with you right now. I can’t tell you how much I’ve been looking forward to this conversation. You may be able to see in my bookshelf back there, many of your books are on prominent display. I thought we’d start with the book that you first cowrote with Alfred Rappaport, Expectations Investing. I think the story about how you came to write that is quite interesting, so would you let us know, how did that come about?
Michael Mauboussin: I’ll first say that I was a liberal arts major in college. I studied government. I’d never taken a business class. Well I take that back, I took accounting for non-business majors when I was a senior and got a C plus in the course, out of the gracious heart of the professor. So I really had no business experience, and so I came onto Wall Street in the mid-1980s having no idea what was going on, and really there were so many old wives’ tales and rules of thumb. I guess to some degree we still have those, but that was certainly the main thing I confronted.
Michael Mauboussin: Then in the late 1980s one of my trainee classmates actually handed me a copy of Alfred Rappaport’s first book called Creating Shareholder Value. That was published originally in 1986. For me that was my epiphany. The light bulb went off. There were three things in particular he talked about in that book that have really become the centerpiece of my thinking all along in my professional career. The first is something, Toby, you’ve written a lot about and you know a lot about, is that it’s not accounting numbers, but it’s cash that ultimately drives the number of businesses, and so how do we think about that as value investors in general.
Michael Mauboussin: Second is, and I still tk we do a poor job of this, he argued that competitive strategy and valuation really shouldn’t be separate, they should be joined at the hip. Which is to say, the litmus test of a good strategy is that it creates value, and that you can’t really do a thoughtful valuation without understanding the economics of a business and the industry.
Michael Mauboussin: Then the third and final thing of this chapter, a chapter that is aimed at corporate executives, but it was called Stock Market Signals to Managers, and the argument was, “Hey CEO, what is crucial is what’s priced into your stock in terms of expectations, and for you to deliver excess returns in the stock market, it has to be consistent with a revision in expectations. So this idea of the markets being expectations machines.
Michael Mauboussin: So I immediately started writing about this very enthusiastically as an analyst. By the way, then it was a little bit out there. It was a little bit quasi-academic, but I had a couple of key executives in particular, and investors, who were supportive. One I would mention, by the way, is Bill Stiritz, who was the CEO at the time of Ralston Purina. He’s prominently featured in Will Thorndike’s book, The Outsiders. And getting the sort of imprimatur from someone like Stiritz really was an attaboy for a guy like me who was young at the time.
Michael Mauboussin: Then I collaborated, just I would talk to Rappaport from time to time, and then in the late 1990s he sat down and said, “A lot of these ideas do make sense for investors. Would you like to collaborate on a book where we take those core principles and apply them specifically to investors?” We wrote that in 1999. By the way, it actually launched on September 10th, just think about this, September 10th, 2001. So the day before a national tragedy in the middle of a three-year bear market.
Michael Mauboussin: So the timing on the release was not ideal, but it was just an awesome project. And just being able to work with someone who’s your mentor, and someone for whom I have such deep respect, and now Rappaport, I still … He’s in San Diego, he’s not so far from you, and I still talk to him frequently. And just such a thrill, and just been such a huge influence in my life, not only professionally and even personally. That’s the Expectations Investing story.
Tobias Carlisle: One of the really powerful ideas in Expectations Investing, and it’s built right into the title there, is the idea that you can see implied from the market price the expectations for a business. So we’ve got historical data on earnings and cash flow and growth rates, and then you can take a forward look at what the market price employs, and then your job as an analyst is to determine whether that is a fair expectation or not. I think that one of the nice things about it is, it does a really good job of tying together, that Buffett would never allow them to be separated but they are often separated, the growth and value for investors. So can you just talk a little bit about that process, how you come up with the expectation and the mechanics of that undertaking?
Michael Mauboussin: Yeah. 100%. There are a lot of people who have negative things about discounted, I think everyone agrees that discounted cashflow is intellectually the correct way to value all financial assets, so you get very little pushback on that, but people get the devils in the details about how models can lead to whatever values, assuming whatever assumptions you make.
Michael Mauboussin: So the first beautiful thing about expectations investing is, it actually reverse engineers the process. One thing we know for sure is the price, and then you basically are asking a simple question: what has to happen for this price to make sense? So mechanically what you would do is, can I do a good job, or a reasonable job, of trying to understand the consensus expectations that are built into this price, and value drivers. So sales growth rates, operating profit margins, capital intensity, and understanding what has to happen for that price to make sense. So now you’re making an over-under judgment rather than saying, “I have a laser precision as to what I think the thing is worth.”
Michael Mauboussin: You bring up this value versus growth thing. This is really essential, because the key insight in all this is, we want our businesses in general to create value, so their returns are in excess of their opportunity cost to capital, which is finance 101, or economics 101. So what we know is, if you’re earning excess returns, growth is a wonderful thing, and the faster you grow, the more wealth you will create. You can show that mathematically in a very trivial way. Of course if you’re earning below your cost to capital, growth is bad, because the faster you grow, the more wealth you’ll destroy. So this notion of value and growth, I think, has always been a false dichotomy. In the factor worked usually you’re talking about sort of proxies for these things. But ultimately they have to be connected to one another, and value creation’s the key principle.
Michael Mauboussin: That leads to some very interesting simple heuristics. If you have a very high return on capital, high growth expectation business, very very modest tweaks downward in the growth rate will lead to very sharp declines in the value of the stock, mathematically appropriately so. Likewise, values stocks, which I think is why they’ve done well over long periods of time, tend to be low expectations stocks, so that’s why they tend to do well.
Michael Mauboussin: So to me, and I teach at, as you mentioned the Columbia Business School connection, I teach at the Heilbrunn Center for Graham and Dodd Investing, so very much in this tradition. But I think the name was chosen very carefully, to not be just pure Graham, cigar-butt type of thinking. It’s really, if Graham were around today, how would he think about value investing. I think he would embrace many of these same ideas. So sometimes cheap stocks are cheap for a good reason, and sometimes they’re cheap because the expectations are unduly low, so it’s distinguishing between those things that really seems to be the crucial things.
Tobias Carlisle: There’s a few questions that fall out of that for me. When you’re thinking about cost of capital, are you calculating that in a traditional, efficient markets, looking at the beta of the stock? Or how should we be calculating the cost of capital?
Michael Mauboussin: It’s a great question, and we’ve written a bit about this, and I’m completely familiar, and to some degree sympathetic, to some of the charges against that traditional way of doing it. There are a lot of adjustments you can make, for example using industry betas versus firm-specific betas, and some regression techniques to allow you to get to a slightly better place. But I also say you should triangulate. The bottom line I always say, I say this to my students in particular, you’re a business person, so forget about formulas and filling out formulas. Think about this as a business person. And you have a couple things you can do.
Michael Mauboussin: One is, you have, for example, the credit markets. You have bond yields. That should be a touchstone. You have options markets. Options markets can give you some sense of what’s going on, so you have other comparable things. In other words, there should be ways to triangulate to get into something that’s intelligent. Then you want to just make sure that you’re consistent, so if it’s an expectations game, that consistency becomes very important. So it should pass, I always say, whatever the cost to capital you come up with should make business sense, it should make common sense to some degree, and of course you want to have it tied to some degree to principles and finance. But the notion of starting building with a risk-free rate and adding some sort of excess risk premium, that probably does make sense. But you’re right, we try to have it both ways, which is using a traditional asset pricing model, but also understanding that that’s not the end-all, be-all, and certainly don’t do that by rote. It should be thoughtful, and understanding the adjustments that will get you closer to where you think you should be in terms of the real world.
Tobias Carlisle: I’ve tried to understand, in some of Buffett’s writings, he writes about, he just says if you reinvest over a period of time and you find that you’re trading at a discount to book value, for example, then that’s an obvious example of where the reinvestment is turning a dollar of earnings into 50 cents, or less than a dollar of retained value, and vice versa. Is that an overly simplistic way of going about it, or do you think it’s an effective …?
Michael Mauboussin: I’ve actually opened a number of my reports with that basically dollar-bill test, and that makes some sense. If you’re earning exactly your opportunity cost, and you invest in your business your dollar should be worth a dollar on the marketplace, and so that’s one [inaudible 00:11:07] booking that does make some sense. And of course if you’re earning high returns, if the value of that stream of cashflows then becomes more than a dollar, and as you mentioned below a dollar, it should be less than a dollar. So at least as a simple framework I don’t think that’s a terrible place to start.
Michael Mauboussin: The devil becomes in the details in terms of how the real world works. It’s often that even businesses that are valued … I mean there are a whole slew of issues with book value to begin with, but even simplistically, companies that even are value-neutral, value-destroying, often trade at a premium to book value, I think because they have embedded options for restructuring. So it’s very tricky to map that one-to-one from theory to the real world, but I think as a conceptual way to think about it, it absolutely makes sense, yeah.
Tobias Carlisle: As a practical matter, I really like the expectations, or the implied growth rate style of investing. That’s one of the first things that I always look at when I look at the valuation of a company: what sort of growth rate does this company need to justify this stock price?
Tobias Carlisle: An example that’s in the news at the moment and has been a popular stock for a while is Chipotle, because it was a very high growth rate stock for a while. Then when people got sick they said it was food poisoning, but I think it was norovirus, which is, you can just get very unlucky with norovirus. It takes down cruise ships and things like that. So I thought they it was a little bit unfair, but when I went to look at the valuation it still implied these incredibly high growth rates. I know that Bill Ackman bought it recently and did very well out of it, and I looked at it again today, and the implied growth rate is something like 40% compound over the next decade, which makes it an enormous business, and I think that means that means that it’s almost like a Starbucks-Chipotle on every corner. Do you look at individual stocks, is that the way that you recommend folks go about it?
Michael Mauboussin: Toby, exactly. Again, I always say when you’re doing your reverse engineering or understanding the expectations, you should be completely agnostic. So really you should have no view of the world, no view of the business in particular, you just want to understand what has to happen for this thing to make sense.
Michael Mauboussin: In the example you’re giving, you’re saying these very high growth rates have to transpire to justify today’s stock price. Then you want to ask the next question, which is, what’s the likelihood that that will actually occur. And you can do that one of two ways. One is to do your bottoms-up research, and like you said, if it’s restaurant chain you say how many stores are they going to build, what are the economics of the store, so on and so forth.
Michael Mauboussin: Then the other thing is to appeal to base rates. Just say how many companies of this size have grown at this rate for this period of time, in history? You know what that distribution is going to look like, and you know where this company’s growth rate’s going to fall on that distribution, and you say okay, what kind of bet do I want to make on that. We use a lot of examples of that. When you’re far right tail distribution, so you’re saying that 40% growth rate maybe has been done only 2% or 3% of the time in history, doesn’t mean that it can’t happen in this instance, but it probably suggests to you it certainly wouldn’t be your base case. You would shade the probabilities to reflect that to some degree.
Michael Mauboussin: So base rates, I think, your own bottoms-up work, combined with some of the base rates to try to inform you, and I would just go back to this Graham notion of margin of safety, and if it’s close you shouldn’t be playing. So if you have to go out four digits on your HP-12C to figure out if it’s a good investment, it’s not a good investment, I can tell you robot now. So you really want to find situations where there are a lot of opportunities for upside and ways to make money, presume year-long, and there aren’t that many ways to lose, so this sort of asymmetry.
Michael Mauboussin: And again, it’s a really important thing to say, is that almost always, the tricky thing, when sentiment’s high, when sentiment’s optimistic, that’s when stock prices go up and when sentiment’s bearish. So you’re fighting against a little bit of the broader sentiment and the broader points of view and the consensus view. That’s where it becomes psychologically difficult to do this. But the discipline and the mathematics of it should be quite straightforward, and I think you nailed it exactly how you’d want to think about it.
Tobias Carlisle: When we discussed very briefly yesterday when I said I wanted to discuss a return on investment capital, and you sent me through 152-page book that you produced, which is base rates across any number of industries. Over the last decade it’s no secret that value has had a very rough run, that more traditional value, and to the extent that any value investors, I think, have succeeded over the last decade, it’s because they have been investors who have been seeking those higher growth opportunities.
Tobias Carlisle: And there’s some suggestion that the market has changed, that for whatever reason, software really will eat the world. Software as a service is going to be the dominant business model in many industries, and so that’s going to change the base rates. How do you handle something like that, when you’ve got … I think the research goes back to 1950, or it’s at least 50 years. So you have 50 years of base rates across different sectors and industries, and we’ve got potentially a new industry, I guess, almost, because software is sort of distinct from the underlying industries that it’s trying to disrupt. How do you get a handle on whether those growth rates are sort of a short term thing? Is it because it’s a reasonably new industry? Are growth rates driven by competition? Or has the world shifted? What’s your impression?
Michael Mauboussin: This is obviously really interesting, and a very challenging debate. The first thing I would just say is, and I think you’ve already made this distinction, but just to reemphasize it, often when we talk about value versus growth we’re talking about factors, and those are often valuation based factors, whether it’s price to book or price to earnings or something like that. So I want to step away from that statistical definition of value versus growth, and revert more to what you described a moment ago, which is, value is buying something for less than what it’s worth, and growth might be a willingness to pay a fuller price or something, betting on that. So that’s the first thing. But I think this whole debate about value versus growth is really a fundamental one, and this is where doing bottoms-up research and understanding what’s going on can be beneficial.
Michael Mauboussin: There’s no question we have a trend, which has been going on for decades, of investments being more based on intangibles than tangibles, and I think our traditional factor work probably lended itself more to businesses that were more tangible asset-based. So then you have to say “Okay, if I’m starting about non-rival goods versus rival goods, and the economics of information versus the economics of physical goods, what is different about it? So no laws of economics have been repealed, let’s be super clear about that, but we’ve always known, and we’ve known for a very long time, that the economics of information goods are different than those of physical goods. By the way, there’s a great book written 20 years ago, and notwithstanding some of the examples in the book are dated, but Carl Shapiro and Hal Varian wrote a book called Information Rules. And I think that Information Rules has a lot of good models to understand how we should be thinking about these kinds of things.
Michael Mauboussin: The other issue that you raise, which I think is a very important one, is the strength and limitations of base rates. And like you said, we have data in many cases back to 1950, and you should ask, and it’s a very valid question to ask, is whether those distributions reflect everything that I should think about might happen. The answer would be obviously no, because every time there’s a new extreme on the left or right hand tail, that’s something we’ve never seen before. So I would not say that becomes the gospel from which you can never deviate, but rather to say that is something that should introduce an element of sobriety, to say we shouldn’t expect things to go on for very high rates for very long periods of time. Or at least I don’t want to make that bet explicitly.
Michael Mauboussin: So I think this debate’s a really difficult one. I’ve always tried to reframe, and I’ve been doing this for many many years, is reframe the value versus growth thing, just say it’s value creation, and buying things for less than they’re worth is a value investor. And I think Munger’s got some line like this too, like all intelligent investing is value investing, in the sense you buy something for less than what it’s worth. And there have been a lot of, obviously, amazing franchises that were never super cheap. It would’ve been really hard to by Walmart at any point in the first 15 or 20 years that it was public at a valuation you would deem to be cheap. It’s been very difficult to buy Amazon at any valuation you would deem to be cheap. But those things created enormous amounts of wealth.
Michael Mauboussin: So I’m a little bit agnostic, but I think you’re raising all the key issues, and I would not be tethered to the view that traditional factor value has to come roaring back at some point. That said, we’ve seen these kinds of cycles before, so I would lean toward leaving statistical factor value, but I’m somewhat agnostic on this. I don’t think people should have super strong views one way or another.
Tobias Carlisle: I think the challenge over the last decade has been that many of these companies have, in order to justify their valuation at the time, you have had to imagine that the future growth rate is going to be quite high, and sustainably high, and what has transpired is that they have in fact sustained, in many cases, these very high rates of return, and these very high returns on invested capital. I think Amazon probably is the best example of that, because I think even a decade ago it would’ve been very difficult to imagine just how big Amazon is today, and that’s because you would’ve had to have foreseen something like Amazon web services, which was just nonexistent. So I would never ask you the question, how do you ever find something like an Amazon. But do you think that that’s something we should be considering? Should be moving up our growth rates, or is that the sort of question that everybody asks right at the very top of a bull market?
Michael Mauboussin: Yeah, I would be hesitant to do that. But maybe we can step back and say there’s one other aspect to this that I find really interesting, which is, for most of my career I’ve lamented that companies have focused too much on growth and not enough on returns, a return on capital in particular, so they’ve allocated capital in such a way that the growth becomes the first and foremost thing. I think there’s been a market difference since the financial crisis in particular, so two things going on. One is, I think ROICs in corporate America basically have been going up, and part of this is the digitization that you described, and moving from tangible to intangible assets and so forth. And that means there’s more sensitivity to growth. What I’ve found interesting is, I think that really since the financial crisis, and the very sort of modest recovery in the economy globally, is that growth has become at a premium, and people have been willing to pay for this.
Michael Mauboussin: There’s a weird sort of side story that I’ve always found very very intriguing, and that’s with mergers and acquisitions. M&A, if you go to business school, and even today people say it, although it’s not as accurate, they would say something like, 60 to 70% of M&A deals are value destructive for the acquirer, so we do some cumulative abnormal return around some window. That was true for decades. Essentially very hard to create value as an acquirer. You create value in the aggregate, but for the acquirer. And there was this really weird thing, that for the first seven or eight years coming out of the financial crisis, that the majority of M&A deals created value for the acquirer. And this is completely anomalous if you look at a five decade history. So I don’t know the answer to this, but I do suspect that this notion of growth being amplified because higher returns on capital, that it’s become more valuable. So it’s an interesting dimension, we should demarcate the growth rates and then the returns on capital.
Michael Mauboussin: The high returns on capital, the growth becomes a really big amplifier to wealth creation. That’s an interesting thing to throw into that mix, which is, we’re looking at these distributions of growth rates historically, but if the ROICs are just a shade above the cost of capital, that doesn’t amplify all that much. If I then bump up my ROICs, and [inaudible 00:23:34] for some subset of companies they’re much higher, and then I show that right hand tale growth, then it really amplifies wealth creation. I don’t want to be an apologist for anything that’s gone on, but I think that may be at least a set of considerations to look at, to try to understand what’s happened. I think that gets short shrift.
Michael Mauboussin: And the M&A thing, I think there’s not enough discussion of this. It was weird 10 years following the financial crisis, relative to what we’d known for decades and decades.
Tobias Carlisle: Is it possible that traditional M&A always, you needed to pay a premium price in order to get control, there needed to be a reason to sell, so it always involved some, there was a financial measure and then there was a takeover premium that was built into it, that presumably gave the vendors value, or more than value for them, in order to let it go. So possibly the prices paid were lower. But then there’s also the other side of the equation too, which is that the cost of capital is lower than it has been traditionally.
Michael Mauboussin: That’s true. But there are some counter arguments to that. One is that the observable premiums really aren’t materially lower, but we can observe that from case to case. The cost to capital thing is an interesting one, but there are two things. One is, the cost to capital, if that’s truly lower, that should be reflected in the stock price pre-deal, so before you get bids, so that price is just higher. Now you could say, on the margin is it cheaper, that’s an interesting question.
Michael Mauboussin: The other counter point is, there was a period certainly in 2007, probably in the first part of 2008, where interest rates were relatively low, if you were a borrower you could borrow at very favorable terms, not like today but not that far off from today, and we did not have the same phenomenon. In fact it was actually almost the opposite, that it was very difficult to create value as an acquirer. So I don’t really know the answer to this, but it’s just an interesting thing to think about.
Michael Mauboussin: I always find it interesting that you’ve had something that becomes sort of conventional wisdom on M&A being a value destructor for acquirers, and again it’s been taught in business schools for decade after decade, and for this last 10 years, or let’s say the 10 years or seven, eight years after the financial crisis, for that to be so different, is something to make note of. So I think the discount rate could be part of it, I do think the premium thing, there are M&A waves, and usually when you act at the beginning of a wave versus the end of the wave it’s more favorable, but I’m not sure any of those things fully explain what we’ve seen.
Tobias Carlisle: Just to change tack slightly, one of my favorite charts of yours that I’ve included in a couple of my books, and possibly three, is you rank all of the companies on their return on invested capital, and you have say five buckets with the highest return on invested capital, after backing out the weighted average cost of capital, and you show over about 10 years that there is this mean reversion in the returns. I had an earlier gets on who’s a business professor in South Africa, and he did a similar examination of South African companies, and found almost exactly the same trend, except that he found in South Africa it was more pronounced. It had occurred in about three years, that over a course of three years the top performers were at the mean, and the worst performers were at the mean. Could you just talk a little bit about that chart? And I understand that you’ve updated that over the years, and it seems to have stayed pretty consistent as you’ve done it.
Michael Mauboussin: Right, and it will always be consistent, by the way. You just described, what we’re doing is documenting regression toward the mean and returns on capital. You could do returns on invested capital, ROE, and in our most recent work we were doing cashflow return on investment, but it’s all basically the same story.
Michael Mauboussin: First, by the way, I would draw people’s attention to one of, I think, the more interesting chapters in Danny Kahneman’s book Thinking Fast and Slow. It’s chapter 17, and chapter 17’s all about regression toward the mean. So again, just to be super clear, this is a psychologist writing about regression to the mean in a somewhat different context, but he makes a point in there that’s really important. Which is, any time the correlation between two measurements, so return on capital today versus return on capital next year, any time that correlation is less than 1.0, perfect correlation, you’re going to see regression toward the mean. So that’s the first really important idea just to get out there.
Tobias Carlisle: What does that employ, just before you go on?
Michael Mauboussin: Yeah. Let’s think about a world with … Anywhere from a zero correlation to a one correlation, so zero correlation means there’s no relationship between event one and event two. By the way, it just says you should expect complete reversion to the mean for the next outcome. So let me back up and say, what does regression toward the mean say in plain language? It says outcomes that are far from average will be followed by outcomes with an expected value closer to the average. Outcomes that are far from average will be followed by outcomes with an expected value at closer to the average. Now that correlation coefficient, one versus zero, that tells you the rate at which regression toward the mean happens. So if the correlation I sone, there is no regression to the mean at all, and if the correlation’s zero that means you expect complete regression toward the mean.
Michael Mauboussin: So the first thing to say is … And by the way, this is really interesting. You show this to a business school student, or even a business school professor, and you show that chart of high returns going down, low returns going up, there’s a classic example for that in microeconomics 101, which is competition. Toby opens Toby’s lemonade stand, very profitable, Michael says “I’m going to come in, build mine right nextdoor to you, lower prices 10%.” Still a good turn for me, you have to now lower your prices to match mine, and so on, and we iterate down to our opportunity cost of capital, which is fine. And I certainly don’t want to deny the notion that competition happens. But what I’m saying is, for any possible set of reasons, not just competition, but it could be any exogenous factors, that the correlation of returns goes down, you will get regression toward the mean.
Michael Mauboussin: That’s a very very important message, and again, tying it back to something we talked about a few moments ago, if we’re saying that the value creation is the key principle, and value creation is return on capital, less cost of capital spread, so some return, what we’re saying is, economics are going to drive these returns down over time. So you should be very, very measured about what you’re willing to pay for future value creation. I think that’s the main thing.
Michael Mauboussin: There’s another exercise we do that freaks people out, which is, we actually do regression toward the mean backwards. So for example, you take 2018 numbers, do the same quintile ranking that you just described a moment ago, and then go backward in time, and you see the same exact pattern. So while it’s very intuitive that competition would affect businesses going forward, it makes no sense that they would affect businesses going backward, and of course that’s what you get.
Michael Mauboussin: That’s the first principle that’s really important to understand, is that regression toward the mean is a powerful thing, it’ll happen every time the correlation is less than one, and we can actually measure those correlations and understand the rate of regression to the mean. So it’s two lessons here. One is that it happens, and second is the rate at which it happens.
Michael Mauboussin: What we did more recently, and I don’t know, the professor in South Africa, what he did exactly, with what industries he selected or whatever it is, but I would have an educated guess about that, but what we did then is looked at different sectors. And we could do it, of course, on industry level. And what you find is something that’s incredibly intuitive, which is sectors like consumer staples have a slower rate of regression toward the mean than sectors such as energy. Saying that differently in plain language, and again I think very intuitive for most investors, is all things being equal, two companies, a consumer staple company and an energy company with the same return on capital, and say roughly equivalent cost of capital, you’d be willing to pay a higher multiple for the staple company than you would for the energy company. Why, because the staple company will have a more gentle slope back toward the cost of capital, and the energy company will have a much steeper one.
Michael Mauboussin: What’s also important is, we’re talking about return on capital regression, but you of course can apply this to anything. You can apply this to growth rates in sales, you can do it for margins, you can do it for anything, to understand the rate at which that regression toward the mean is likely to happen. So I think it’s incredibly powerful, the notion that we can start to measure, not just that regression happens, but the rate at which it happens is also incredibly fascinating. It ties back to intuition, that you pay less for financials and energy than you do for staples, for example, and healthcare, and that’s what we see empirically. I think it’s an credibly important idea.
Michael Mauboussin: We didn’t talk much about this, but a lot of people operate with multiples, and this is mostly discretionary investors, operate with multiples, and multiples are good heuristics, but these are nuances in the discussion that get lost, I think, when people just bluntly say this is worth 10 times EBITDA, or 12 times EBITDA, without really understanding what are the underlying economic drivers. So I always say to my students, you have to earn the right to use a multiple, which means that you know what the underlying economic assumptions are that get you there. And that’s not always done, actually.
Tobias Carlisle: I like using multiples for the simple fact that if you’re looking for … It just stops you from overpaying for most things. Certainly there are some businesses that are worth paying up for, but in the vast majority of cases, if you think about it the other way around too, if you’ve got a very high yield, and the growth rate implied by that very high yield is very low or negative, then often I think you’re finding something that’s closer to the bottom of its cycle, because it’s having a very bad run. The market is pricing it as if it continues all the way to zero, and I think the better bet across a portfolio is that there will be some recovery. You’re still going to make mistakes, you’re still going to have some blops, but that’s quantitative factor investing, that’s the nature of it.
Michael Mauboussin: And I 100% agree with that. Again, let’s rephrase it, you’re buying low expectations.
Tobias Carlisle: Right.
Michael Mauboussin: You’re saying the world is pricing it as if it’s going to be worse, it’s going to be a little bit better, there are some reasons for us to believe that, and we do a portfolio of these things, not every one of them’s going to work out, but on balance it’s going to work out very favorably. That’s exactly right.
Tobias Carlisle: In your research, one of the papers that you wrote, possibly more than a decade ago, it was death, taxes, and mean reversion. You talk in there about, some companies are able to resist mean reversion. I don’t know if it was, you explicitly said it, or if I went in and worked it out, but I think it was about 4% of companies in your universe which might have been at the top thousand by market cap in that one. Which is a very small handful, that’s about 40 companies. And I think you examined the factors in that one, you did a DuPont analysis and you looked at the rate at which capital was turned, and the margins and so on. I think that my recollection is that you found it difficult to really identify what the drivers were, but you said it tended to be companies with higher margins. Then you looked at the industries, and I think you said biotech and pharmaceuticals had tended to do a little bit better, and retail had tended to do a little bit worse. Do you have any more color on how companies resist mean reversion, or what you need to look for?
Michael Mauboussin: Yeah, I would love to do more work on this, but I will say this, that I would direct the attention of our listeners and our viewers to work by Michael Raynor and Mumtaz Ahmed. Do you know this stuff?
Tobias Carlisle: No, I don’t.
Michael Mauboussin: There’s a Harvard Business Review article that summarizes their findings, and they wrote a book called The Three Rules. I think it’s chapter two of that book, is from my point of view sort of the money chapter. So they worked with another statistician down at University of Texas named Andy Henderson, and they worked out, what they try to do is figure out how much of corporate performance was a function of common cause variation versus special cause variation. So common cause variation, we’ll just say it’s the randomness in the system, special cause variation we’ll say some sort of skill, what we’re trying to identify.
Michael Mauboussin: I should first say, by the way, if you look at investment management industry, the vast majority of what you see out there is explained by common cause variation, but we also see in money management that certain money managers do generate, they are skillful. The challenge is not that we can’t figure that out ex-post, the challenge is figuring it out ex-ante, can we figure that in advance.
Michael Mauboussin: A very similar story is true for companies, and what they did essentially is, they created a transition matrix. They basically looked at deciles, I think, in ROA, and then studied empirically how those transition matrices worked, and then they simulated the world zillions of times and said, does any company defy the simulation. And again, it’s a very similar thing. As you mentioned, a low single digit percentage of those companies that seem to be, we’ll call them skill, whatever you want to say, that are defying, essentially, that gravity of returns.
Michael Mauboussin: Then they were asking, what are the factors that are behind that, and they come up with a couple things. And it turns out, one of them was better revenues before cost, and better before cheaper. The way I would restate that in Michael Porer language is that it seems to me it’s a differentiation strategy. And going back to the DuPont thing you alluded to, if you said to me how would I characterize differentiation versus low cost producer using DuPont, differentiation would be high margins, relatively modest capital velocity, and low cost producer would be relatively low margins and high capital velocity. So I think that the arrows are pointing toward this differentiation as you pointed out accurately, sort of these high sustaining margins, as sort of the key to all this.
Michael Mauboussin: Again, this ex-post, ex-ante question doesn’t go away, can you identify those kinds of things. So they try to identify or think about or talk about things in the book that would lead you at least to a higher probability of success in identifying these kinds of companies. But they claim that these rules that they’ve identified are quite universal, they have nothing to do with industry, whether a company is acquisitive or non-acquisitive, and so on and so forth.
Michael Mauboussin: We actually replicated some of their findings, and came very close, so part of that feels very buttoned down, but that’s another area of research that I’d love to do more work on. I think it’s really fascinating, and whether we can actually start to take a step toward getting some things that would be predictive to allow us to identify those companies.
Tobias Carlisle: It’s striking to see being proven empirically what Buffett has been saying for a long time, which is that brands are very important, and that high return on invested capital needs to be supported by some sort of competitive advantage, which he calls a moat. Striking to figure that-
Michael Mauboussin: I will say this though Toby, this is where I always feel like … If you had asked me, even before we wrote that piece, but if you had asked me like 20 years ago, 15 years ago, what seems to be the most defensible competitive advantage, it’s partly because I think a lot of our business school case studies are things like Southwest, they’re disruptors, they’re like Southwest Airlines or Dell, and what characterizes those businesses typically is lower margins at higher velocity. So I would’ve been leaning toward the low cost producer model, but you’re exactly right. I think Buffett, that’s the Buffett slash Munger slash Phil Fisher influence, for the better businesses you should pay up a little bit for better business, but it’s a really interesting thing, so I feel like I’ve had to sort of shift my own thinking on that a little bit, and sort of recognize what the data are telling us. More needs to be done, I think, and that’s a really fascinating, exciting area for us to do more work on.
Tobias Carlisle: One of my favorite books of yours is Think Twice, which has, and I saw you point this out, and so then I went back and looked at the cover, and [inaudible 00:39:37] you have Thin Ice hidden in the title cleverly. You talk about inside versus outside views, which I understand to be the difference between the base rate and the case rate. Can you just describe Think Twice, if you would, and what you mean by that?
Michael Mauboussin: Yeah, so Think Twice is a book with the idea that, and this really is, the whole thing is almost a homage to Danny Kahneman, and this is the system one, system two thinking. So system one’s your fast system, it’s experiential, it’s quick, it’s automatic, and then system two is your slower system, analytical, more purposeful, more deliberate, and more costly, candidly. The book is trying to run through different situations where you should be recruiting your system two to think through the problem.
Michael Mauboussin: It’s interesting, the inside-outside view, this is the one idea, I think it’s so powerful, we’ve already talked about it, but just to be more formal, and I think you describe it well with case versus base, but inside view is the natural way for us to think about problems. If I pose a problem to you, whether it’s, how long will it take you to remodel your kitchen, what will it cost, when will you finish your term paper if you’re a university student, or how will this asset class or this stock perform, the classic way that we all do it is, you gather a bunch of information, you combine it with your own experience, your own input, and then you forecast. And for most cases that’s perfectly fine.
Michael Mauboussin: The outside view is a markedly different way of looking at the world. It says we’re going to think about this problem as an instance of a larger reference class base, and we’re going to ask, what happened when other people were in this situation before? It’s a very unnatural way to think, and that’s why the think twice comes in. It’s a very unnatural way to think for two reasons. One is, you have to leave aside all your cherished information. We think what’s going on in our head is a pretty good representation of the world, and you have to discount that, which is not a natural thing to do, and second is, you have to find and appeal to the base rate, which is not often at your fingertips. So there are many things that are unique to you, you may have moved from one city to another, and it’s a new experience for you, but many other people have done something like that before, so you don’t really know what the aggregate experience is like.
Michael Mauboussin: So the argument, and this is a famous Kahneman-Tversky paper, they talked about this back in 1973 on the psychology of prediction, they said look, the way to think about a really good prediction is to combine the inside and the outside view in an intelligent fashion. And just to follow up, going back in our conversation about regression toward the mean, here’s a really simple, I think a really powerful, heuristic. If your activity is all luck, no skill, so think about roulette wheels or lotteries or some things, it’s all outside view, no inside view. It’s all base rates. Does not matter what your prior experience brings to the table. If it’s all skill, no luck, so running a race, or I’m sure you’re a better tennis player than I am, then it’s all inside view, no outside view. So my past tennis record, it doesn’t matter if I’m playing you, because you’re going to beat me every single time that we play. Then of course almost everything in life that’s interesting is going to be between these two polar extremes.
Michael Mauboussin: So this is another way for us to think about, how do we weigh inside versus outside view. It also goes back to our stuff on regression toward the mean. So in a way these are all sort of intimately connected concepts to some degree. So that’s the inside-outside view. Notwithstanding, I think many people are starting to understand more about this, most people don’t employ this regularly or systematically. That’s why the base rate book, you saw that was an accumulation of pieces we had trickled out. We’d put them in one big volume. I would use that as a reference. The first 15 pages or 20 pages, I described this regression toward the mean, so I would say that’s a pretty reasonable thing to read. There are a lot of pictures, it’s not that bad. Then the rest of it is just a repository of specific financial measures for companies. I just think it’s an incredibly powerful idea.
Michael Mauboussin: This is one where you should, like you said, we tend to use our own views of things and embed these expectations, or growth rates, without really thinking, has anybody ever done this before? Is this even plausible? Even if it has been done before, if it’s been done a small percentage of times, you may want to temper the probability you assign to that outcome. Again, that’s being a value investor, in the sense of, you don’t want to overpay for the future. You don’t want to pay for cheery expectations.
Tobias Carlisle: One of the great vignettes in that book is the business magazine covers, where the companies that were spoken about in a glowing way tended to have very positive stock price returns in the previous three years, and they were companies that, there were bearish articles written where they had very negative stock price returns over the precedence three to five years, whatever it was, and then of course as soon as they publish, it flips, which is a very good example of mean reversion, I think, and a little bit reminiscent the Ibbotson, what does he call it, the paper or the book that came out recently, the popularity book?
Michael Mauboussin: Exactly, yeah.
Tobias Carlisle: Did you read that? I have a collection of those.
Michael Mauboussin: Yeah, I did, and it’s great. I would just say that, and Kahneman actually talks about this also in Thinking Fast and Slow, but it’s this idea, if you think about, extreme positive outcomes are almost always lots of skill plus lots of luck. And extreme negative outcomes are almost always not great skill but bad luck. And if you presume that luck is unlikely, by definition, it’s unlikely to be persistently good or bad, it’s in some sort of distribution, then by definition the guys that are on the covers as being bullish are the ones that have enjoyed a lot of good luck. That’s why we say the expected value next outcome is something closer to the average. You expect the luck on average is not going to be as good, and they’re going to mean-revert. And likewise, the bad companies have had bad luck, and they’re going to mean-revert. In a sense, it also all fits in, because you don’t want to extrapolate …
Michael Mauboussin: And again, all streaks, Stephen Jay Gould said a streak is great skill plus great luck, it has to be. Joe DiMaggio, great skill plus great luck. It’s the only way you can get there. If you accept that an outcome is some combination of skill and luck, it has to be the combination of, the right hand draws from both, they get used to the extreme right hand outcome. So yeah, I think that’s a really powerful one. Again, when everybody’s bullish on something, you should always be careful about, how can things go wrong, basically.
Tobias Carlisle: I think that’s a really great segue into my favorite book of yours, which is The Success Equation, which is in the second shelf in the orange red just behind me there. There’s lots of great ideas in it. I love the idea of just trying to understand the role of luck in a lot of the data that you look at too, that the data’s not necessarily a pure reflection of skill. What is the success equation about?
Michael Mauboussin: It’s interesting that one of the latter chapters of Think Twice was about skill and luck. I originally had it as a second chapter, because I thought it was really interesting. And my editor reads the book and she says, “This skill-luck stuff, I’m not sure if anybody’s going to really care about this. You can leave it in, but put it at the end.” So I’m like “Okay, whatever.” So I put it at the end of the book. And then some people who, they’re friends, so they would call me up and they’d go, “I really enjoyed Think Twice,” and they’d go, “But that chapter on skill and luck, that was really cool.” So I was like, okay, that’s interesting. Then I wrote one piece when I was at Lake Mason on luck and skill that just seemed to be, really resonated with people for whatever reason. So I’m like all right, I’ve been emboldened to think about this concept.
Michael Mauboussin: I also have been influenced by a number of things, like many people. Nassim Taleb’s book, Fooled By Randomness, which came out in 2001, right around the same time as Expectations Investing. Michael Lewis’s book Moneyball. So, statistical techniques applied to athletics, but you could see how that moneyball idea is carried around to different domains. So I thought, Fooled By Randomness is a great message, and being aware that there’s a lot more randomness than people think, that’s great, but can we do a little bit better than that? Can we start to really think about how to quantify the contributions of luck in various activities? That was sort of the motivating force, and again, very much guided by the sports guys.
Michael Mauboussin: I love sports, I’ve played a lot of sports myself, and am an enthusiastic fan. I love the world of business, and understanding how businesses work and so forth, and of course investing has really been my professional career. So all three of these things have these very large doses of luck, and so it made a lot of sense for me to try to work on that. So yeah, the motivating point behind The Success Equation is to say, can we untangle this contribution of skill and luck? And what does this mean for how we think about the world, how we evaluate ourselves, how we judge the likelihood of future outcomes? And it’s an incredibly rich … Not to suggest, there’s much more to be done on this, but it was an incredibly rich, I think, experience, just trying to think about that even in a basic way. So that’s what it’s like.
Michael Mauboussin: There are basic things, like skill tends to follow an arc. Almost everywhere you go, athletics is a great example, you get better and better and better, at some point you peak, and then you degrade. Lucky is very weird, because some things are independent draws like coin tosses, but more often luck has to do with social factors, and so these path dependent processes. That applies to big parts of life as well. I think it’s just a fascinating topic. Then rolling up our sleeves, just tying back to all the things we talked about, I think it has a lot to say concretely about regression toward the mean.
Michael Mauboussin: I think most investors, if you go to an investor at a conference and you say “Regression to the mean,” you get a lot of vigorous nods, affirmative nods, but if you actually say, “Do you know what this means,” I think most people do not understand. In any way, I don’t think they understand what it means exactly. So being able to take it to the next level, and to put some flesh on the bones in terms of the ideas, is really important.
Tobias Carlisle: One of the really powerful ideas in the book, that I had observed in lots of different topics without ever realizing what it was, is the paradox of skill. What is the paradox of skill?
Michael Mauboussin: I should say that I think I gave it that name, but I learned it from Stephen Jay Gould, the very eminent evolutionary biologist, and he was talking about this in the context of baseball batting averages. The idea of the paradox of skill is, when outcomes combine skill and luck, which is most things in like, it can be the case that as skill improves, luck becomes more important in defining the outcome. That seems like that’s the paradox, is more skill, but luck is more important.
Michael Mauboussin: The key insight is to understand the distinction between absolute skill and relative skill. So absolute skill, and I think we would agree, is you look around the world, it’s never been higher. Certainly in the world of athletics, and you can do this especially in things measured versus the clock, so swimmers and runners and so forth. We’re leaving aside some results from performance enhanced … We’re at all time records. Business, the world of business, the world of investing. Can you imagine if I put you back in the 1960s with the tools at your fingertips in terms of computing power and data and so forth, you could run laps around your competitors. So absolute skill, we’ll all agree, has never been higher.
Michael Mauboussin: The second point, and this is the point that Gould made, was on relative skill. What’s happened in domain after domain is that relative skill has shrunk. Which is to say, the difference between the very best participant and the average participant is smaller today than it was in years past. So Gould illustrated this with Ted Williams, who was the last player to hit over 400 in major league baseball. 1941, he hit 406. So it turns out that Ted Williams was almost exactly a four-sigma event, four standard deviations. By the way, batting average itself hasn’t changed all that much over the decades. In fact the powers that be in the baseball leagues want to keep it roughly the same for competitive. But we have, pun intended, an arms war between pitchers and hitters. So he’s a four-standard deviation event, that got him to 406. If he were a four-standard deviation hitter last year, which would be awesome obviously, you would hit about 380, which is tremendous, you win the batting title and so forth, but you’re nowhere near breaching the 400 level. And that’s because the standard deviation of batting average, what that sigma means, that standard deviation means, is less today than it was in years past.
Michael Mauboussin: So in domain after domain, what you see is this consistency, this excellence, this consistency of excellence, which means that luck becomes more important. Again, going back to our tennis example, I’m sure that you’re a much better tennis player than I am, and I’m sure you [crosstalk 00:52:32]-
Tobias Carlisle: I confirm that I am not.
Michael Mauboussin: But if we were somehow metaphysically identical tennis players then it really would, the outcome of every match would be some sort of coin toss, and so it would be complete luck. What you see in professional sports, by the way, is grinding toward parity. Every league grinds toward parity. Why, because the quality of the players improves, they’re drawing often from global populations, the training techniques, the coaching techniques, the nutrition and so forth, is becoming uniform, and uniformly excellent. So as a consequence you get more parity.
Michael Mauboussin: In pro sports, it’s really interesting. The one area where you don’t see as much of that is basketball, and basketball is one where superstars really can make a very big difference, and they’re almost irreplaceable. And it turns out most superstars are men that happen to be 6’7” or 6’8” or 6’9”, they’re tall. And there aren’t that many of those people out there.
Michael Mauboussin: So the paradox of skill, clearly relevant in the world of investing. One of the charts that we’ve shown many times, and I expect you’ve seen it as well, is we just look at the standard deviation of alpha, standard deviation of excess returns. And there are some complicating factors, the fact that the sample size has gotten much bigger and so forth, but basically what you see is the standard deviation alpha just comes down. What’s particularly interesting if you look at mutual funds is, it actually comes down, down, down, and then it goes up a lot around the dot coms, and then it goes right back to trend. So the dot com introduced, as individuals rushed into the game, a lot of opportunities for excess returns offer professional investors, and then once essentially those people went away, we reverted right back to the big guys fighting each other.
Michael Mauboussin: The other thing I’ll say, I find this to be a little bit surprising, but a lot of people go, this fact everyone’s going into investing and passive and systematic investing, that’s really great for us active managers, because those guys are just not … And the reason I hesitate about that is because it’s the poker table analogy. What you want when you’re playing poker, if you want to make money, is to have someone who’s across the table who’s weak, so that you can take that person’s money. And if that person decides, “I’m going to go index,” so they essentially leave the table, or essentially they’re not betting, they may be drinking their beer, but they’re not betting any more, and now you’re just competing with the other good card players, your life’s got a lot more difficult, not a lot easier. And I think there’s some of that that’s out there too, so the people who remain are actually the most skillful, not the least skillful, including mom and pop, who used to be a fairly important part of the investing ecosystem. They’re basically gone, and as a consequence the big boys are fighting each other, and that just makes it more competitive.
Tobias Carlisle: I’m so glad you raised poker as an example, because that was one of the examples that I was thinking of, where there was the massive boom in the early 2000s of the online poker, where the kids could play 10 tables at one time. And so they got this enormous amount of experience in a very quick time. So a kid could be in their early 20s and have as much experience as someone who’d been playing for 60, 50 years, someone who’d been playing for a very long period of time. But all of these kids had it at the same time, so all of a sudden the televised games were won by very young players. But then because they’re playing against each other …
Michael Mauboussin: Exactly.
Tobias Carlisle: And this is one of the really smart things, I think, that they did, they actually recognized what was happening, and so they would change the way that they played to this very aggressive loose style which just increases the amount of luck that you have in the game.
Michael Mauboussin: Yeah, you want increased variance. The other thing that’s interesting in all that stuff, just in terms of the spread of excellence, is chess is another fascinating example. So you look at, Magnus Carlsen … All right, Kasparov lost to Deep Blue in ’97, so I don’t know how old Magnus Carlsen was, he was probably a little kid, five or six years old, and so he grew up in a world where he learned not only from a coach, but also by playing online. So there have been some really interesting things, where these guys will track chess moves versus that leading computer programs, and you see that the best players in the world today play more like the programs than their predecessors, which is so cool. These are all ways to get better, and it just makes the world more competitive, but like you said, that’s interesting.
Michael Mauboussin: That’s another idea in The Success Equation which is really interesting, which is, if you’re the stronger player you want to reduce variance. You want to make it as simple as possible, and then overwhelm the competitor with your skill. If you’re the weak player you want to increase variance. You might go down in a ball of flames from time to time, but every now and then you’re going to have a strategy that actually works. So if you’re playing in a football game or whatever it is, you want to run trick plays, you want to run weird stuff that the other team is unlikely to have seen. You’re still going to be the underdog, but it gives you a much better chance of success. So whenever you’re the weak one, increase variance. That’s your poker thing, just increase variance. That’s how you try to win. If I ever had to play heads-up against a poker professional, I would just play with high variance. I would lose most of the time, but every now and then you’d strike gold.
Tobias Carlisle: Thanks so much for spending time with us. Just before we go, if folks are looking to get in contact with you, what’s the best way of doing that? You’re on Twitter …
Michael Mauboussin: Yep, Twitter, @mjmauboussin, DM on Twitter’s great. michaelmauboussin.com, so my website has got a bunch of stuff on the different books and so forth, some of the stuff we talked about today. And yeah, I love to talk to anybody.
Michael Mauboussin: It was so much fun, I really appreciate that we were able to talk about some pretty important and serious concepts. I hope it was fun for our listeners as well.
Tobias Carlisle: I had an absolute ball. Michael Mauboussinsin, thank you so much for spending the time with me today.
Michael Mauboussin: Thank you Toby, appreciate it.
For all the latest news and podcasts, join our free newsletter here.
Don’t forget to check out our FREE Large Cap 1000 – Stock Screener, here at The Acquirer’s Multiple: