Invest Like the Best with Patrick O'Shaughnessy - Michael Mauboussin - Sharpening Investor & Executive Toolkits - [Invest Like the Best, EP.308]

Episode Date: December 20, 2022

My guest this week is Michael Mauboussin. Many of you will know Michael and his work well. He’s Head of Consilient Research at Counterpoint Global, one of the sharpest investment minds I know, and a frequent guest on this show. In this discussion, we go deep into his recent work on market share, returns on capital, and capital allocation - all of which are coming under increasing scrutiny for different reasons. Please enjoy this great conversation with my friend Michael Mauboussin.   Listen to Founders podcast   Founders Episode #136 A Success Story: Estee Lauder    Invest Like the Best with David Senra: Passion & Pain   For the full show notes, transcript, and links to mentioned content, check out the episode page here. -----   This episode is brought to you by Tegus. Tegus streamlines the investment research process so you can get up to speed and find answers to critical questions on companies faster and more efficiently. The Tegus platform surfaces the hard-to-get qualitative insights, gives instant access to critical public financial data through BamSEC, and helps you set up customized expert calls. It’s all done on a single, modern SaaS platform that offers 360-degree insight into any public or private company. I’ve been so impressed by the platform that my firm, Positive Sum, recently made an investment in Tegus. We did so because we feel that Tegus will be the gold standard platform for investing research for decades to come. As a listener, you can take Tegus for a free test drive by visiting tegus.co/patrick.   -----   Today's episode is brought to you by Brex. Brex is the integrated financial platform trusted by the world's most innovative entrepreneurs and fastest-growing companies. With Brex, you can move money fast for instant impact with high-limit corporate cards, payments, venture debt, and spend management software all in one place. Ready to accelerate your business? Learn more at brex.com/best.   -----   Invest Like the Best is a property of Colossus, LLC. For more episodes of Invest Like the Best, visit joincolossus.com/episodes.    Past guests include Tobi Lutke, Kevin Systrom, Mike Krieger, John Collison, Kat Cole, Marc Andreessen, Matthew Ball, Bill Gurley, Anu Hariharan, Ben Thompson, and many more.   Stay up to date on all our podcasts by signing up to Colossus Weekly, our quick dive every Sunday highlighting the top business and investing concepts from our podcasts and the best of what we read that week. Sign up here.   Follow us on Twitter: @patrick_oshag | @JoinColossus   Show Notes [00:03:46] - [First question] - Overview of Michael’s recent research on market share [00:05:48] - Market share dynamics in modern history [00:08:43] - How market share data is useful for investors [00:12:30] - Investing in early breakout companies from low-concentration markets [00:14:34] - Surprises from his recent research project [00:15:29] - Using the value stick for stakeholder satisfaction [00:19:12] - Examples of value creation using the value stick [00:23:33] - Market power in relation to markups and willingness to pay [00:32:00] - Identifying a company’s real ROIC numbers [00:44:00] - How important absolute ROIC is when picking investments [00:47:07] - Research on capital generation and allocation trends [00:54:25] - Characteristics of great capital allocation strategies [00:59:26] - Surprises in the market since his deep-dive research [01:02:54] - Artificial intelligence and other sources of disruptive innovation

This episode is brought to you by Teegas. Over the years of our partnership with Teegas, they have evolved from a pure expert network into a full company intelligence platform. I've been so impressed by the platform that my firm, positive sum, recently made an investment in Teegis. We did so because we feel that Teegis will be the gold standard platform for investing research for decades to come. Teague streamlines the investment research process so you can get up to speed and find

answers to critical questions on companies faster and more efficiently. The Tegas platform surfaces the hard-to-get qualitative issues. insights, gives instant access to critical public financial data through BAM SEC, and helps you set up customized expert calls. It's all done on a single modern SaaS platform that offers 360-degree insight into any public or private company. As a listener, you can take Tegas for a free test drive by visiting tigis.co slash Patrick. Today's episode is sponsored by Brex, the integrated financial platform trusted by the world's most innovative entrepreneurs and fastest growing companies. With Brex, you can move money fast for instant impact with high limit corporate cards, payments, venture

debt, and spend management software all in one place. Ready to accelerate your business, learn more at brex.com slash best. That's bR-e-x.com slash best. Hello and welcome, everyone. I'm Patrick O'Shaughnessy and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money. Invest Like the Best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at join colossus.com. Patrick O'Shaughnessy is the CEO and founding partner of Positive Sum and the CEO of O'Shaunacy Asset Management. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect

the opinion of Positive Sum or O'Shaunacy Asset Management. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of positive sum or Oshonnessy asset management may maintain positions in the securities discussed in this podcast. My guest this week is Michael Mobeson. Many of you will know Michael and his work well. He's the head of conciliant research at Counterpoint Global and one of the sharpest investment minds I know and a frequent guest on the show. In this discussion, we go deep into his recent work on market chair, returns on capital and capital allocation. all of which are coming under increasing scrutiny for different reasons.

Before we transition to the episode, I want to highlight the Founders podcast, which is part of our Colossus Network. David Senra, who hosts founders, has devoted his life to learning from history's greatest entrepreneurs, and every week he distills the lessons of a different founder. If you want an entry point, I highly recommend starting with episode 136 on Estee Lauder. I hosted David on Invest Like the Best This Summer, and it's hard not to walk away insanely energized after listening to any episode with him. You can find a link to founders and that episode in the show notes of this conversation. And you can search all past transcripts on our website, join colossus.com.

Please enjoy this great conversation with my friend Michael Moveson. So Michael, if you can believe it, it's actually been almost two and a half years since you and I last did this. I think this will put you back on top as the all-time guest leader for if that's like the best. Beating out our good friend Bill Gurley, who I think Curley has the title. And it's a good occasion to do it because you have had, from what I can remember, the most prolific research publishing period, I don't know, of your life or something in the last three months, putting out huge, huge reports on big central issues for business and investing. So I'm really excited to do this with you today. We've got a lot to choose from. But I think the one that I'd love to begin with is the research that you've done recently on market share.

And the reason I'm picking that is there was a lot of new stuff in there, especially around markups, which I'll have you explain. But I think it's a great entree into strategy and into some really helpful ideas for business leaders and for investors that are investigating business strategy inside companies they invest in. So maybe you can begin there with your motivations. Why did you choose to do such an exhaustive deep dive into the dynamics of market share? And then we'll start to pick apart some of the lessons that you learned. Well, first of all, Patrick, it's awesome to be with you. It's always a pleasure.

I can't believe it's two and a half years, but it's a thrill to be back. and I'm just a huge fan of the podcast, and so being part of it, it's great for me. Taking a step back, I teach a module and strategy to my students, and we spent a lot of time talking about market share and some of these dynamics, but I always had the sort of nagging question in my head, which was, if I simply showed you the list of competitors in an industry, so the number, and then I showed you their market shares, would you be able to infer anything about the industry or market structure? So almost going backward. So our classic wave doing this is we say, oh, what is this company's competitive advantage? And does that competitive advantage

ultimately manifest and things like profitability and market share? We don't really think about it going the other direction. That was the opening motivation. And as you point out, when you go through the report, it sort of led down this meandering path into some things that I was familiar with, but I hadn't studied in great depth, including how we define and think about market share, how we think about things like market share stability, how we think about market concentration, markups, the rise of superstar firms. This whole area that I think is not only super interesting intellectually and relevant strategically, but also foundational if you're an investor, because this is a lot of the

stuff that investors think about. So that was the motivation was to unpack this area and meander down this path and learn about all these various components. One of the things that stands out from a couple of the charts in the report is some of these industries, let's take Google's share of the search market or something, where on the data in the chart, it just looks like a monopoly. There's another one in there on Microsoft and the word processing industry. Obviously, Microsoft does a lot more than that. But you see some of these modern, usually technology-driven examples of complete dominance in a certain industry, depending on how you've

defined the industry. So maybe to begin, talk about how the dynamics of market share have changed over time? Have they changed? If we went back to the 1950s, would it just be fundamentally different because of the nature of businesses that are market leaders or has something changed over time? The hard one to answer, I think that we've always had these winner take-all type of outcomes. Historically, however, those tend to be industries that are regulated. So governments tend to identify those and regulate them. The most obvious and sort of canonical example would be AT&T in the telephone system, something like that. What I think is probably more recent, you know, recently, and technology enabled this to some degree is, as you point out, these winner take

most or winner take all outcomes. So when we think about market concentration, there are two ways to think about how it happens. One is consolidation. A couple examples come to mind would be the U.S. Devets industry has consolidated massively the last 25 years, and that's literally just companies buying and merging and getting more concentrated that way. You could think about the airline industry recently last 20 years or so. another example.

And then there are these winner-take-most. And you point out, we selected two illustrations. One was Google in the search business. And the other was, this is a much older example, but Microsoft in the word processing business. What's interesting about each of those particular cases, though, is now looking back on it, it seems obvious that Google's got these dominant shares or Microsoft were at the time had these dominant shares.

But when you decompose the process, what you realize is there was just an enormous amount of shuffling happening early on. And when you think about, for example, internet search, it was by no means a foregone conclusion that Google would emerge as the dominant player. So we document this through the measures of market share stability, just showing that stability changes quite dramatically. The other thing I'll mention is sometimes you get these reversals of a concentration. So I thought the automobile industry example in the U.S. was a fascinating case.

If you go back to the 1960s, for instance, basically a handful of companies had three quarters of the market share. So it was a very concentrated, very high. And that concentration has actually come down substantially over the last 50 or 60 years. All these things, as you point out, they're dynamic, they change. I don't know if it's useful to make broad characterizations that it's more of these winter take most markets than there were before. But I do think they are more enabled.

They may be fleeting. In other words, you may not be able to sustain that leadership for a very long time. The question comes into whether regulation starts to impose itself on it. I'll ask some questions on this from both the perspective of the business builder, the entrepreneur, but also the investor. So if I'm an investor and I want to add a tool to my toolkit, is there a tool to be added in the area of market share? Meaning, like, if I approach a new industry for the first time and I have the market share data, is that helpful to an investor, do you think? And if so, in what ways? I do think so. A couple tools I think I would recommend. One is the idea

of market share stability. The underlying premise here is pretty straightforward, which is it's hard to have a sustainable competitive advantage if market shares are shifting around a ton all the time. You want to be king of the hill, and it's hard to be king of the hill if the hill itself is shifting. Interestingly, one of my former colleagues at Columbia Business School, Bruce Greenwald, a really well-known value investor and teacher of value investing, developed a technique to do this. But by the way, these ideas have been around for a very long time. There's a paper written by a professor named Michael Gort back from the 1960s. So again, nothing new under the sun. And what Bruce did, which I always liked, which was you would look at market share over two increments. So let's say

five-year periods apart. And then we'll look at the absolute average change in those market shares. So pluses, whether you lose share or gain share, that becomes a positive number and then take the average of those numbers. So when that number's high, what that's telling you is there's a lot of change going on. If that number's low, it tells you there's a lot of states. It's a sort of stable. And so that's, I think, a really simple litmus test, number one, that all investors should do and can do readily. The second one, Patrick, and I emphasize this in my class a lot, but I think that we don't do enough of this, which is simply looking at entry and exit. There's a very rich literature on entry and exit into industries. I think when people see the data, they're usually surprised at how much entry and exit there is.

This ties into life cycle. So it turns out that early in a lifecycle of an industry, there tends to be a lot of entrance, not. a lot of exit, you at some point peak, and then sort of back when you get more mature, you have more exits than you have entrance and you get a stabilized number of competitors and their market shares solidified to some degree. By the way, auto industry is one of the beautiful case studies in this exact thing. So those would be two tools right off the top, which would say, look at market share stability, look at entry and exit data, and have a sense of it.

Now, there are a couple other things that are important. Number one is how many competitors are there? In the report, we talk about airlines and automobiles where there are a half dozen companies, each with shares of less than 20 percent. No one has sort of a dominant share. Then you have other industries where, I guess, two or three players are the dominant players. You get three-quarters share of the market or more. Those dynamics are important, and that leads to two related concepts that they're really super important.

And every investor and executives should know this, although these can be difficult to discern, but one is this concept called minimum efficient scale. microeconomics. So this is the output you need to generate to reach your long-term average cost. So I'm a potential entrant looking at an industry, and I look at the bigger players and I say, how big do I need to get in order to compete with those guys? What kind of market share do I need? How many units do anything to sell? So that's your minimum efficient scale. So that's a really important number to know. Obviously, entrants have to think a lot about it. The second one related to it is total addressable markets. How big is the market?

So obviously, a large market can accommodate number of companies at scale. You could have four or five big companies. They themselves may not have a specific advantage at scale, even though they're huge. I think we had the stat that something like five or six global auto companies have revenues of $100 billion or more. Or you can have a relatively small market, and then one company is not that big, but they have captured minimum efficient scale, so they become the dominant player. So those would be some of the tools I think that investors can think about and can use to gain

some insights. One of the charts that really stood out was the time series of the industry concentration, I think I'm thinking of the one in internet search, where like you mentioned, in the late 90s or late 2000s, it's extremely low. Like the Herfindal Hirschman Index, which is a proxy for industry concentration, is really, really, really low. There's Ascheves, there's Excite, there's Google, there's tons of others. And then you just see this up into the right line that levels off at 85, 90 percent for Google. It looks like to me, there's lots of information in that chart. And from the investor's perspective, what they should look for is stuff where it is, where there's a breakout, I don't know what else to call it, where there is low concentration,

and then track companies that are breaking out of that. Do you think that that's a repeatable, sustainable investing strategy? Absolutely. I think that's exactly what you want to do. And by the way, that's all the stuff on the main guy on this is a guy named Stephen Klepper, who's died, but he was a Carnegie Mellon's done an enormous amount of work on industry. I mentioned Michael Gort, Clepper and Gort worked together on a bunch of stuff. And this is a very common pattern, which is you have sort of a battle for the industry.

And by the way, a lot of it is actually standard setting and stuff like that. It actually things about what is going to work in the marketplace and price points and all that kind of stuff. And once a company or companies establish themselves as the sort of pole position, the other guys fade away. And that's probably a pretty interesting time to invest. So the answer to that is yes. Now, that said, these are really really. easy to see through the rear for a mirror and really difficult to see through the windshield. I was part of the team when I was a Lake Mason Capital Management. We were working on the

IPO of Google in the summer of 2004. And it's hard to remember this, but everyone now sees everyone should have bought it on the IPO. First of all, the press about the Google IPO was you should avoid it for a bunch of different reasons. And it was not completely clear that there would not be another company that would come along and unseat Google as it unseed all the companies you had mentioned. Again, in retrospect, it seems super easy. easy. But at that time, it was not clear that that was going to be the dominant force. Was there anything in the investigation where you felt something seems like it should matter to investors or to businesses and it just didn't where there just wasn't a dog barking,

even though you might have expected one? One of the things I found to be most surprising was the lack of connection or causal factors between market concentration and profitability. So usually you'd say like a more concentrated industry should be a more profitable industry, and it turns out that isn't really the case. It is the case that market share tends to align with or predict profitability. It's a very conditional statement, but that tends to be more true. So to me, that one made in the biggest eyebrow razor was if I just find industries that are concentrated and assume that those industries are going to be the more profitable, and that relationship

just doesn't hold up. I'd known a little bit about that, but I really went back and looked at that literature and solidified that point. A really good opportunity to talk about this idea of the value stick and really just value creation. Like market share, of course, on its own, doesn't matter at all. I can get complete market share as a nonprofit. If I never wanted to turn a profit or something, it seems like it'd be easier. And market share usually is measured on revenues. So talk about this notion of the value stick and willingness to pay, willingness to sell and value creation. I love this idea. By the way, this is original framework was laid out by Brandon Berger and

Stewart. The paper is from the mid-1990s. So this idea has been around for a long time. There's a relatively new book last year or two by Felix Oberholzer G, who's a professor at Harvard Business School. The book is called Better, Simpler Strategy. I recommend that book. It's an easy read. It's very clear. And he provides vivid illustration to this. So to your point on the value stick, what are we talking about? You imagine just a stick, a vertical stick from top to bottom. The very top, your jot down willingness to pay. We'll come back and redefine the terms in a second or define the terms. Next, below that would be price, below that would be cost, and below that would be willingness to sell. So willingness to pay, what is that? It's pretty common sense notion is

what is the most somebody's willing to pay for a good or service. As consumers, you can totally relate to this. So if you pay less than what you want for an ice cream cone, you go, oh, it's awesome. There's consumer surplus at your side. I would have paid more for that ice cream cone. So the difference between pricing and willingness to pay is consumer surplus. The difference between price and cost for the company is exactly what traditional, and that's value creation for the company. And then the difference being willingness to sell is the price of which suppliers are willing to sell their good or services to the company. Usually the most important supplier is employees. So that's the basic setup. So the aggregate value creation, which includes consumer surplus,

supplier surplus, and the company's value creation is a difference between willingness to pay and willingness to sell. The point that Oberholder G makes, which I think is a very powerful framework, he says, listen, as a company, don't worry about raising your price. Think about increasing willingness to pay. Because if you increase willingness to pay and you don't raise your price, you've increased consumer surplus, your customers are really psyched. On the other hand, if you need to raise your prices, you have a little bit of flexibility because they value your good or service much higher than the price you're charging for today.

You're starting by adding value to the consumer and then figuring out how to split that with the consumer. Likewise, and this is, I think, is much less intuitive and trickier, but how do you lower willingness to sell? How do you make your employees feel fairly compensated at levels that might be lower than competitors? How do you make your suppliers feel happy to sell to you at lower prices than they might otherwise? So he goes through the book, specific techniques to do both of those things. That to me is a fundamental idea. And you're exactly right. My market share by itself, you have to balance market share with this idea of value creation. And that's even when I define total dressable market. I don't define it as how many units I can sell or how much revenue

can I generate. I define it as how many units can I sell while creating value, which is a very different definition. I really like that framework. And I think that it's a really good way to think about businesses in general. And you can think about countercases. There have been some recent criticisms of sort of the, I'm going to call it for a lack of better term, the general electric model where you go out and just say, can we raise prices? Can we cut costs as a way to create value for our company? And you're not thinking as much about willingness to pay and willingness to supply. As a consequence, you sap that consumer surplus, you sap that supplier surplus. And that eventually gets you into trouble. And you think about other companies who really do try to build on willingness

to pay or lower willingness to sell. And we can talk about the components of how to do those things, but to me, that's the broad framework. Maybe dig into just a couple examples on each side, just to nail home the point, because I want to make sure I understand the difference if any, between willingness to pay and just basic customer value creation. Yeah, I think it's related. The canonical example of increasing willingness to pay is capturing network effects. So network effects exist when the value of a good or service increases more people use the good or service.

Again, that's why a lot of these platform businesses, if they take off, they create enormous amounts of value because they do two things simultaneously. One is they benefit from supply side economies of scale old school. You just have more revenues spread over the same cost. That's good. But more importantly, you're increasing demand side economies of scale, which is your increasing willingness to pay more than you're increasing your price. So you're getting more for what you're doing.

So classic network effects. So you think about platform businesses would be, and look, everybody waves their hands, say we're a network effect business or a platform business. Not that many succeed in doing that, but those that do certainly would benefit from that increasing willingness to pay. Another good example is this idea of compliments. So a complementary economic good is something that pairs with what you're doing. So the classic examples are video games and consoles, hot dogs and hot dog buns.

If you are selling something that has a complement, what you'd like is to have your compliment be as cheap as possible and in fact, ultimately free because that makes your good or service more valuable, right? So you say it's a summer picnic and I'm supposed to buy hot dogs and hot dog buns, of course. If the buns are basically free, you're going to buy more hot dogs or vice versa. So complements are another one. So you might think about like in this environment, one great example would be electric vehicles and charging stations.

So it's a chicken and a egg. You don't want to buy an EV until you know that you can charge everywhere you go, but no one wants to build a charging station until they know someone's going to show up and so on and so forth. So this chicken the eight. So that's a classic compliment. So if the government comes in and says, we're going to build charging stations all over the place, that adds enormous amounts of value for the electric vehicle manufacturer

because that compliment just got cheap or cheaper and makes the value of their good or service much higher and the willingness to pay goes up. So that would be the classic example on the willingness to pay side. On the willingness to sell, this is a little bit trickier, but there are a couple things that are interesting. One is as a company that gathers information, this is obviously something that's changed quite dramatically

in the last 25 or 30 years, but companies themselves are gathering much more information about what their customers are doing. If you can take that information and package it and give it to your suppliers and allow them to be more efficient, that could be a case where they can, for example, increase their invested capital turns, lower their margins, still generate good returns for themselves and benefit the company itself. So let me one example, just to make that more vivid, there's a fairly famous case study between Walmart and Procter and Gamble. Walmart, early on, huge reliance on data.

They knew what was going on. your Procter & Gamble, obviously, they're a very big customer is yours. So Walmart says to Procter & Gamble, here's what I'm going to do this. We're going to tell you basically where all your stuff gets sold, when it gets sold, so and so forth. P&G then in turn allows them to take that information and improve their own manufacturing processes, lower their inventories, so they can increase their inventory velocity and so forth, and actually allow them to charge a slightly lower price because they're more efficient on the back end. So there's a win-win.

The data is there, and that allows them to lower, willingness to sell, and then hence that allows Walmart to be more effective. The last big one is employees, and we all love to talk about the importance of culture in organizations, but there's a really interesting thread of research that shows if you truly do create a great culture, that people are willing to come and work for a little bit less than a competitor, for instance. And if you pay at a fair value, so you pay what the competitor would pay, and your employees are willing to take a little bit less, that's employees. surplus and they're totally psyched.

So this is an employer, think about how do I create an environment where people are psyched to show up every day? I'm giving them employees surplus. They don't want to go anywhere because they feel they're paid fairly, but they're really psyched and they would actually work for a little bit less than what they're getting paid. So that to me is another area that, again, it's a little bit subtle. The willingness to sell stuff is all a little bit more subtle, but there are some really clear examples that we can reveal that could be useful. I want to talk about power as it relates to all of this because

seven years ago or something before seven powers and other similar frameworks became so popular amongst the entrepreneurial crowd, usually if you talk to an entrepreneur, they were just trying to create a great good or service and charge fairly for it. And now when you meet startups as I do all the time, oftentimes these young people that are just brand new into starting a company are already thinking about generating power, competitive advantage, these sorts of things. It seeped into the entrepreneurial consciousness. And I think the concept of market power is, is really interesting and important, especially as it relates to this notion of markups that you explore in the paper. Can you explain that section and what you learned exploring the data

behind markups? Absolutely. And I had a conversation recently with Felix Oberholzer G, the professor I just mentioned, and he said, when young people come, entrepreneurs come wanting to start company. They say, how do I monetize my good or service? He says, don't focus on monetization. Focus on increasing willingness to pay because the monetization will follow. So this is a good precursor of this conversation. So what is market power or how do we measure it? In classic economics, the equilibrium is when the price of a good or service equals the marginal cost. So go back to your cost and demand curves that you learned in Econ 101 when the good or services price at that intersection, that's considered to be competitive equilibrium and there's no market power.

If the price of the good or service is above that level, it reveals market power. You're able to price your good or service above the marginal cost. cost. Now, what happens, by the way, when you do that, two things fall out of that. One is consumer surplus actually goes down because all of a sudden there are some people who are willing to pay a little bit of a lower price, but they're out of the game now. But then it introduces essentially economic profit for the company. So just being aware that this idea of market power increasing your prices above the marginal cost, there's a tradeoff that consumer surplus goes down while you're making economic profit. That's why the willingness to pay is such an imperative.

So the question is, how do we measure this? The classic way to do this in economics is a concept called markups. And a markup is, again, this difference between marginal cost and price. Now, I will just say, Patrick, this was something I learned a lot about and I thought it was a fascinating literature, and I want to contrast two schools on this. Obviously, a markup of zero basically means that you're, or a markup of one, pardon me, is you're pricing at your good or service at the marginal cost. If markups are above one, that means you're pricing at a premium.

If you look at the data series from 1955 to now, which you're a marketable, which you you see are two very distinct periods. From 1955 to 1980, essentially markups were relatively low and relatively stable. So a little bit of like an inverse use, a slight arch. In 1980, they're back to sort of 1950s level, but basically low. Then since 1980, we've seen this up into the right hockey stick move in markups. What suggests that companies have more market power today, suggest they're pricing way above the marginal costs and so forth. So that's the first. And there are a number of leading academics. They've written books about this topic. And when you think about people's concerns, for example, they give rise to work on antitrust and so forth. This is going to be the

kind of literature that they prefer to. Now, when you unpack that, there are two really interesting dimensions to it. The first is asking, why are those markups going up? And there are two possible ways to think about it. One is some industries are getting better than other industries. So some industries are still sort of mired, other industries are taking off and they're pulling up the averages. Turns out that's not the case. It's not the between industry differentials that make a difference. Where the difference happens is within industry. Think about this this way. You go from industry to industry and what you're finding is one or two companies are pulling away from the pack, these so-called superstar firms. The between industry effects are modest, but the within industry

effects are really strong. And the last way we can document that, which we did, is by looking at quintiles. So simply, you take the full population, you break it into quintiles, so these buckets of 20% each, and then you look at how they perform over time. Bottom three quintiles, completely boring, no action, not going up very much. And their slopes are positive, but tiny. Fourth, quintile goes up a bit, and then the top quintile is really where all the action is. So that's the basic picture, flat from 50s to 80, then up, no action between industries, lots of actions within industries, and then quintiles, it's all the top quintile, which is consistent with that superstar firm.

By the way to calculate these markups is basically using a production function. I'm not going to bore you with the details, but the basic equation to figure out a markup, and this you guys could do, for example, at Oshonancy asset management, any quant can do this. Production function is input output. The output is going to be sales,

divided by the input, which is going to be cost-a-good sold. You then have to have some sort of output elasticity function. A good one is about 0.8. So basically sales divided by COGS times 0.85 is going to be your markup. Okay, so that's the thing to keep in your mind. So then a couple academics came along and they said, wait a second, time out. Maybe the input is not just COGS. Maybe there's more to our input and we should consider intangible investments. And gee, when have intangible investments really taken off really since

1980? So what they did is they said, we're going to keep sales as our output, but now for our input, we're going to have cost of good sold. But now we're going to add a component to reflect a slice of SG&A, sales general administrative costs that we would deem to be intangible. And by the way, there's a lot going on in this area that's interesting, but there's sort of a standard is called Peters and Taylor framework where they say you should take 30% of non-R&D SG&A plus R&D, and that's going to be your intangible investment. So these academics added that to the cost good sold. So obviously what's happening is also that relationship's changing. Your output has to say change a little bit.

But basically, the markup picture goes away. So it's still positive. The slope goes up a little bit, but it's a gentle rise. It's certainly not alarming your hair wouldn't catch on fire by looking at it. And more importantly, then you break down this between, okay, so there's still no between industry effects. But the industry effects actually go away as well. And when you look at the quintiles, again, the fifth quintiles is the highest, but you don't get the runaway effect. So another way of saying this is the so-called superstar firm.

firms appear to be the firms that are spending exactly the ones that are spending the most on intangible investments. And this goes back to your thing, what's here that we don't see? And I think the intangible discussion is something that is here that we don't see. It's in our accounting numbers, but we probably have it in the wrong category. So when you make that adjustment, you get a very different outcome. So that to me was a really big revelation. Now, all these things are tricky to measure.

Let's go back to the beginning. I mean, measuring market share is difficult. Measuring market concentration is difficult. All these things are difficult to measure. But this is another example where I think a sensible adjustment, I think all of us would agree that intangible investments have become more prominent. I don't think anybody would debate that. We could debate about how we measure that and what the details are, but we don't debate

that it's basically happening. And when you introduce that, it gives you a completely different picture. So to me, that's a completely fascinating discussion. And basically what that's saying is the costs are higher than what we said. so the markups are not as high as people perceive. And then that gets you right back to the beginning, which is let's continue to focus on willingness to pay, price, cost, and willingness to sell

is sort of our core way of thinking about creating value. And again, you think about a lot of the great compounders, people take off these companies that are great compounders of the last 20 or 30 years. In almost every instance, you'll see that they are, they may not be overt about it, but they're playing to this playbook effectively. I mean, return on invested capital. I was tempted to say, well, what metrics cut through all this

where it's just like, okay, you introduce a different accounting methodology and the effect goes away and careers are dedicated to this stuff. But at the end of the day, our job as investors is find great investments where companies produce lots of cash, return in different ways, discount cash flows like the basic stuff. So I'm tempted to say, well, let's just take a look at real cash flow and the relationship between that and the return earned on that on the investments, but then the investments themselves rear their head again, which is how do you measure this stuff? And you published a ton of work recently on ROIC with and without intangibles. This is like the workhorse metric of all investors. We're trying to figure out what returns do companies

earn on their outlays, on the stuff that they invest in. And it just seems like a cloudy picture. So help us with the recent work there. How would you encourage investors to cut through some of this noise and get a pure picture of how good a company's earning returns on their investments. Yeah, it's great. I mean, maybe I'll do this in three parts. The first is just to define terms very quickly. There's a lot in the report, and I don't want to get too deep into the nerd factor. So if anybody wants to really delve into it, they can do that. But at a very high level, return on invested capital, the numerator is a term called net operating profit after taxes. And the way to think about no pat is that it's the cash earnings a company would have, assuming it had no debt or any

excess cash. Basically, the cash earnings of the business without any leverage or excess cash. That becomes very handy because, as you point out, that becomes a foundational number for a calculation of free cash flow. It's obviously the numerator of return on invested capital. It's a number from what you can then subtract a capital charge to come up with economic profit. So no pad is a really handy number. Then investic capital, in a sense, is trying to capture the cumulative amount of investment made in the business. On the left-hand side of the balance sheet, which is the asset side, obviously, we're trying to figure out what is the stuff we've had to buy and put into place to generate the cash flows. And on the right-hand side, which is liabilities

plus equity, it's how do we finance it? So the left is what's the stuff? On the right is, how do we finance it? So that's the basic picture, RIC, is how much earnings do I get for how much stuff I put into the business and to state the obvious hire tends to be better than lower. Importantly, there's a threshold, which is the cost of capital. So in other words, all the capital that you've deployed in your business has an opportunity cost. It could be going off and doing something different and earning a return. So you have to meet that minimum threshold and ideally do better than that for you to

create value. So that's the first section. That's basically what we're talking about. And in this piece, I went out of my way more than I do normally to make sure that people can make the connections between net operating profit after tax and investment and free cash flow. By the way, investment is really the change in invest in capital from year to year. So all those ideas tie together.

And again, people can go through that in some detail. I mentioned one of the thing on free cash flow, Patrick. And I think that in your own work and the quantitative work you've done historically, I think you guys really appreciate this. The classic definition of free cash flow on Wall Street is cash flow from operations minus capex. So a lot of companies talk about this, a lot of investors. But that is a very potentially misleading number.

should be used with great deal caution. The first big reason for that is stock-based compensation is added back to cash flow from operations. I would argue stock-based compensation, by the way, it may be perfectly fine, but it should be appropriately reckoned as a financing activity, not as an operating activity. So that's one big thing. Second is, it's a levered number, right? So not unlevered. That's fine, by the way, we have to make adjustments. And the third is, there's no reckoning for leases in CAPX. And increasingly companies also lease things. And there are companies like, for example, Amazon.com, which they'll say, here's our CAPX, here is our lease number, and then they give the numbers together. And they literally say in their

10K, you should think about these things together because it's the same whether we buy it or or lease it. So that's the first big section. Second big section is just empirical, and I'll try to do this really fast. But we did two things I thought were always fun. One is just literally look at the distribution of returns on capital. And what I want people to envision is sort of a normal distribution in the middle, kind of a normalition. And think about a mean or median around nine, low nine, something like that. And then you have two tails on the left and on the right. So, On the left is low return on capital businesses and on the right, very high return on capital businesses. And those are fairly well populated. So that's the interesting thing. The medians don't

change all that much, even if you bring in those tails, but you get these lumps on the side. The other thing we did, which I thought was really fun empirically, as we said, if you think about strategy, going back to Tynespecta strategy, Michael Porter talked about two specific generic strategies. One was cost leadership and the other was differentiation. And the way we could think about those in terms of breaking down RIC, right? So RIC itself can be broken into two pieces, Nopat margin, which is Nopad divided by sales, times capital velocity,

which is sales divided by Investa Capital. So when you multiply Nopat over sales times sales divided into sales, the sales cancels out. You get Nopet over-invested capital, and that's R.OIC, right? But what's interesting when you do that little simple, it's called a simplified DuPont, is allows us to plot companies on a two-by-two matrix. Now imagine the X-axis is no-pat,

no pet margin. Again, no pet divided by sales. The y-axis is sales divided by investor capital. So this measure of capital velocity, how fast do I turn my capital? And then you multiply it. Every company becomes a dot. And what you see is the companies in the bottom right hand corner are high margin, low capital velocity. Those are differentiators. And think about Tiffany's, for example, I mean, at that. On the upper left-hand corner is low margin, high capital velocity. So you would think about a grocery store, something like that. They're selling a ton of of stuff, making a tiny bit on each, but they have a lot of capital velocity. So we can plot every single company on that, which is super cool. Everyone's going to fall on that matrix. Sorry, so that's

the second thing we did that. And you can then start to categorize company, broad brush as cost leader or a differentiator. Then when you open up your strategic analysis playbook, it guides to how you think about it. The last thing I'll mention on this, and I won't go into any great detail, Patrick, but actually was inspired by one of your podcasts with Will Thorndyke. And Will, by the way, He's awesome. He was describing RIC, and he was like, oh, the way we look at it is without goodwill and intangibles, which is typically recorded on the asset side. And if you say, like, hey, Will, why do you do it like that? He'd probably say something like, gee, we're really on the underlying economics of the business. So we don't want to pollute it with this M&A activity,

which I want to, okay. So I said, like, there are two big adjustments you can contemplate with and without this goodwill adjustment. So either do we leave the goodwill intangibles in there, or do we take them out? And the second, do I make an intangible adjustment, which I'll mention just a second. So we're going to take internally generated and tangible investments and put them on the balance sheet. So when you do that, you turn one on or the author for possible combinations. We did the numbers for Microsoft. And if you basically turn them both off, the ROIC is 94%. And if you consider the fully loaded, it goes as low as 34%. And the traditional calculation's like 49%. Right. So in other words, what I wrote at the piece was like, the answer you're going to get is

essentially based on the question you pose. So just being mindful of that. Now, the third part is building a little bit on this intangible. So exactly as you pointed out, even with this stuff on markups, what we know is increasingly companies are spending on their intangibles. And just to give you some sense of this, our calculations, our estimates are that in 1985, intangible investments were 0.7 times capex. In 2021, we estimate that intangible investments were 1.6 times capex. So like a watershed change and what is my investment career, basically. This is how long I've been run in doing this kind of stuff. So the key is, are there techniques to take, again, this component of sales general administrative SGNA costs

and consider what part of that is an intangible investment? We call it internally generating tangible investment. There are two big problems with this, big challenges. Challenge number one is what percent of SGNA should be considered? An intangible investment I mentioned before this Peters and Taylor. They have this 30 percent of non-R&D, SGNA and all of R&D. But there's actually some very interesting and I think relatively recent academic research where they actually break this down for all the Fama French industries.

So you have customized numbers for each industry, which is helpful. The second big calculation is asset life. So you buy a machine, it's got a five-year life, you throw it on the balance sheet and depreciate over five years. Well, if we're going to create these intangible assets by capitalizing internally generated investments, we also need to assign asset life so we can amortize appropriately. in this paper, they also give asset life. So just to say, without going to details,

we now have a framework that we can do this by industry, understand what component of estrogenation appropriately be deemed to be. And by the way, the other side note is, we typically assume that all of R&D is considered to be an investment and tangible investment. And what these guys claim,

and I think it's right, absolutely not all of R&D is now considered an investment. There's a lot of R&D that's maintenance, classic maintenance. You get an update on your Microsoft operating system or something, some R&D person's doing that likely. They actually go through with both SG&A and R&D and then give you the asset life. So that allows us to now do this calculation for the full population.

So empirically what happens, I guess I'll mention a couple things that are interesting. One is fascinating. Is it the tails, which I mentioned before, right? Remember the sort of normal distribution in the middle and then we have these tails. The tails get pushed in. It regresses all the distributions toward the mean. Now, what is the intuition? One way to start is with a company with a very bad return on capital.

One of the things we looked at in the piece we talked about was Snowflake. I think we'd all agree. It's like a really exciting young company, interesting future and so forth. Well, the RIC in 2021 was negative 416%. What's going on? Well, the answer is they're spending a ton to build their business, and there's very limited to invest the capital. So when you have a lot of numerator that's not good and a little bit of denominator,

it looks really bad. When you make these calculations, what you're doing is you're capitalizing part of the investment. You're saying this expense is actually an investment. So by definition, you're increasing your earnings, in this case, by a lot. And then you're increasing your invested capital, in this case, by a decent amount. But you're adding really high return on capital to low return on capital. And it turns out it brings the number from Snowflake from negative 416% to 3%. Now, 3% is not spectacular in a scheme of life, but completely consistent with a company early in its lifecycle, completely consistent. So you get a very different, and I think a more accurate view of things. So that's the first thing is you're pulling in the tails. The second one

is going back to our thing on the strategy matrix. What happens is when you make these adjustments, it moves companies from the upper left-hand corner to the bottom right-hand corner. So you're increasing margins and decreasing invested capital velocity. Why? You're increasing margins. You're increasing because you're adding earnings, but you're adding investment so that it has that effect. So you're shifting companies from the upper left to the bottom right, broadly speaking. What we did in the report, which I think was really fun, is we just picked a handful, something like a dozen companies, and we showed their journey. And for some companies, they make a huge migration.

Typically younger companies are very intangible, intensive companies. Other companies, more stable companies, more bored companies, less intangible companies. They don't move at all, basically. They move a tiny bit in our framework, but very little. So I just think this sheds a completely new light, and I think a very illuminating light on understanding the underlying economics of these businesses. So that's the RIC. The first is just the mechanics of it. Second is the traditional distributions.

And then the third part is when we introduce intangibles, how we get a very different picture. I personally think a more accurate picture. And the last thing I'll say on this, Patrick, is that this is a little bit of the Wild West. In other words, this notion of really discerning what percent of SG&A is an intangible investment versus maintenance, for example,

discerning asset lives. This is work in progress, and there are actually people working on this in the strategy community, people working on this in the finance community, people working on this in the accounting community. I should appeal to you on this because I do think that Quant, who have started to show intangible investments on their balance sheets for their factor models, increases the strength of the factor models, and particularly the value factor becomes a stronger or better signal when this is

introduced, which is important. So you're trying to map the world in an accurate way. These are all, I think, steps toward the truth. If I go back on my investor hat, and I think back to that conversation with Will, the reason we brought up RIC is this is something I will, I'm looking for in businesses that I make big investments in. So one of the features of the businesses that I back is I want them to earn high returns on capital in the way that I define it, which is a bit unique. I immediately jump to, okay, hire better for investors. And I'm curious how true that has been historically. So the question is, does the market figure it out or not?

Do you have a high return on capital business? It's not going to be like a revelation to the market to some degree. A couple things I'll mention on that. And we will probably use these data to spin out a couple other pieces and we'll replicate this thing we did before, but I'll just mention what the results were before. We said, let's just use a completely naive strategy and just say, I'm going to buy the high return on capital businesses and sell the low return on capital businesses. Or I buy them all and to see how the profile.

And it turns out the sharp ratio is for, for all those portfolios are basically exactly the same. So in other words, adjust it for risk, you get the same returns. What's going on? I mean, the market knows that high return on capital business are better, and so they give them higher multiples. By the way, we open that piece with a very rudimentary chart, which is on the x-axis is just return on capital cost, capital spread, so some measure of your economic profits. And on the y-axis is enterprise value divided by invested capital, which is a souped up amount of price to book. What that relationship should show is if you have higher returns, you get higher price to bulk and lower returns,

lower price. And that's basically the relationship. I mean, the correlation is like a 0.7. It's not perfect, in part because there's no reckoning for growth in that calculation. If you put it in a third dimension for growth, you could probably reduce that R-square quite a bit. But you get the idea. So the answer is that. That's the first thing.

Then the second thing where I think it's more potentially exciting is that companies move around. So you can rank them in quintiles, for instance, and then ask these questions and these transition matrices, just ask a question, hey, if I started in the top quintile, so I'm a high return on capital, what's the probability I'll be in the top quintile five years from now, or three years from hour, 10 years from now, whatever it is? And that's the sustainability component. There we see a lot of action. So if you go from a high return to low return, you get crushed. Your stock price gets crushed. If you go from low return to high

return, you kill it. Your stock price does extremely well. So to me, Patrick, I think the key is to find the deltas versus the absolute numbers. Now, The last thing I'll say is, if you are in the first quintile and you stay in the first quintile for a really, really long time, you tend to do really well. Now, notice in the first exercise I said, you just buy all the first quintiles. Well, what's going to happen is they're going to scatter, right? Some will stay in the first quintile. Some will go to the second.

Some will go to the third, so forth. You're going to have a mix of returns on capital. But if you just bought the first quintile and you could identify a priori which companies would stay in the first quintile, you would do really well. So that's this idea of sustainable competitive advantage that's not priced in. If you can figure out changes, that's good, or if you can figure out sustained high returns. But again, what we know just in terms of base rates is only a small percent of companies remain in the first quintile throughout their lives. Well, I'm working closely with a company that we invested in called TIGIS, which serves professional investors with its products.

And early in that company's history, three of the canonical ideas were Bezos's customer centricity, Thorndyke's outsider's capital allocation profile in seven powers. We started to hit on some of these key things that are really informing great entrepreneurs out there. And I think capital allocation is early on in the company's history, they wouldn't call it capital allocation. Maybe they would call it resource allocation or just decision making because there's not a ton of capital to go around.

As companies mature and they're more successful and they produce cash, capital allocation, so-called, becomes ever more important in what happens with shareholder return. I can't remember when it was, but you wrote the piece that got me back in my quant-focused days focused on this as my research priority. And I think it's an opportunity now having you had just updated the report on capital allocation across the U.S. stock market to really learn from your research. So we'll get to some principles for good capital allocation. And I do think there is good and bad capital allocation pretty objectively.

But I want to start with what companies actually do. and what the trends have been and what they do and how they allocate their capital. So maybe we'll start with just where do they tend to get their capital from? There's a couple different sources. And then where have they been putting it on average through history? What have they been doing with their money and their capital? And how has that changed over time? Thanks for that.

We wrote a piece about this, I think, in 2016. And we recently published a new piece on the same topic. I think it is not just new, but it's new and improved a lot. And we've added some components to it. it and did a lot of work with academic research. I should just say on the capital allocation in particular, we toggle between looking at empirical data and addressing the research has been done in academia. I think those two things are really helpful to frame it. As you said, where does the money come from? The answer is for most companies, corporate America, it's internally generated cash.

It's a very, very strong component of it. Issuance of debt ends up being a source of cash. And then equity actually is net reduction. So companies on balance have been reducing their equity, negative issuance. Now, especially that last one, that's episodic, of course. Sometimes there is issuance and so forth. The numbers last year, for example, hopefully I get this right, the stock-based compensation account was something like negative 250 billion. In other words, companies allocated 250 billion.

And their buybacks was just a shade under 1.1 trillion. That gives you a little bit of order of magnitude. So that's a little bit of where the money comes from. By the way, the fact that most capital is internally generated is sort of a good news, bad news thing. The good news is you don't need to access the capital markets all the time that reduces friction. So those tend to be good things. The bad news is, as you pointed out, companies can take good money and put it into bad projects. There's no capital market check on that and companies can do dumb things and basically shareholders can do very little about it overtly.

Is there good or bad capital allocation? The answer is absolutely. The key to me, you mentioned Will's book The Outsiders, and to me, the North Star of all this is building long-term value per share, building long-term value per share. So that's the metric and standard by which you're going to measure all your investment decisions. And we wrote this in the report like four times, but the answer to every capital allocation question is, it depends. Is an M&A deal good? Is buy-back good? The answer is it depends.

and all those things depend on what's going on, especially the relationship between price and value. If you go back over time, historically, the largest source of capital allocation has been mergers and acquisitions. And again, for many companies, it's the most significant. Now, again, large amount of variance. Some companies do a lot of M&A, others don't. Some do it very sporadically. But M&A in the aggregate tends to be the biggest component. Interestingly, what's new, and I think what's changed, as I mentioned over the last 30 or 40 years,

is the number two item by our calculations is actually investment, SGNA, XR&D. So it's this intangible investment. That's what's novel in this particular report. It's actually very significant. It's a burgeoning area of research in academia, and I think super interesting. And as I mentioned, I think the quantitative community is actually very tuned into this, but this is something that will continue to be a lot of work. That's followed by CAPX.

By the way, buybacks and CAPX in recent years have been roughly neck to neck, and then it's dividends and divestures. and I don't know if I'm missing R&D, I guess, is also in there. The other thing I'll say, Patrick, which I find fascinating is John Graham is a professor of finance at Duke University. Every year, the president of the American Finance Association gives a presidential address in January, and that's published by the Journal Finance in August. The Journal Finance quite generously makes that PDF available. So Google John Graham, Presidential Dress, August 2022.

And what John goes through in that talk is essentially 30 years of, CFO surveys asking CFOs about how they actually make capital allocation decisions. So it's totally cool. So you get this long-term view of how they do things and you can monitor how they change and so forth. One thing is fascinating is there's a very different mindset. And let me just home in, for example, on buybacks versus dividends. We show the report, by the way, under some theoretical conditions that are not true in the real world, but under theoretical conditions, buybacks and dividends are the same. They are completely different in the CFO or CEO's mind when the board's mind. So dividends once initiated are considered to be sacred. And you want to raise them

the fall possible. You'd never want to cut them. Worst case canaries keep them? And they're considered to be tantamount to capbacks. That important. By contrast, buybacks are like, oh, we've paid all our bills. We made all our investments. We got a little money laying around. You know, our capital structure and we'll buy back stocks. It's a residual concept. So even though on some level, they're the same concept, they're thought of completely differently. M&A, by the way, same thing. So the standard deviation of growth rates for CAP-X is totally tiny. Standard deviation of growth rates for dividends, totally tiny. Standard deviation growth rates for M&A, for buybacks. These are more pro-cyclical things. They're huge. So that's another thing that's really important to bear in mind is that when things are

rolling and doing well, certain types of activities will come more into vogue and things are going poorly. Things drop out of vogue. Hopefully those answer those two top questions. By the way, you are right. There's also a life cycle component to all this stuff. Obviously, on one level, companies have to invest the most when they're young because they're essentially they're called pre-production cost. You have to spend money in order to make money. So you tend to have low returns on capital because you're loading up the capital. You sort of use the term resource, and I think that's the better way to think about this.

This report dwells exclusively on capital allocation. You could argue that people and stuff show up in the intangible, but basically capital allocation is how companies spend money. But you're exactly right. I mean, if you're a company, you have to also think about. about your human resource allocation and so forth. And those are super important things that are not captured in these data. So just to be clear that we're not pretending this is everything.

So resource allocations are the more encompassing framework. But this would be a subcomponent, albeit obviously a very important one. Our estimates last year, this is 2021, is that total spending was like $7.1 trillion. So there's a lot of money basically being spent. If you think about the characteristics of the great capital allocators, lots of whom Will has outlined in the outsiders and now is doing with us via 50X as well. Some common themes start to pop out. Maybe the summary of all those themes is you already said it, which is, it depends.

Like the great capital allocators look at the situation. They don't just operate via policy. They look fresh at every new environment, every new opportunity and make a decision based on the best information they have. There's not a lot of inertia, I guess, to their decision making. How would you sum up some of the things that you've uncovered, maybe a small handful of things that you've uncovered that indicate great capital allocation or that you could feel comfortable saying to kind of everybody, yeah, you should kind of think this way when allocating capital? No, I think you actually described it pretty well. We opened the report with a piece of research I found to be absolutely fascinating and it very much speaks to this exact point.

So these were academics that asked the question. They actually examined how companies spent capital, specifically CAPEX, for their business. business units. So imagine a company with multiple business units. They look at their CAPEX year after year and the question is how much they move it around. So zero meant they spent exactly what they did or the same exact percent breakdown from year to year. One meant they completely started anew and they moved money around and shuffling completely. So not surprisingly, what the professors found was this inverted you and an optimal level because the world changes and markets mature and so on so on. Sometimes some businesses

deserve more capital than others and so on and so forth. So there's an optimal level. What they also found was 98% of companies were below the optimal level. 98%. So in other words, companies are, as you pointed out, they're very conservative. There's a lot of inertia. They do what they did last year, give or take a little bit. So that's one thing that comes right out of this, which is you could call it zero-based resource allocation, but basically this idea that every year can we truly start with a clean

sheet of paper and ask what are the resources, both financial and human resources, this business needs to maximize its potential. That's the basic question. And again, that's usually not how people think about things. Another one is thinking about strategies versus projects. It's not uncommon within corporations that small-level projects get handled at the unit level. They get a little bit bigger. It maybe goes to some CFO, bigger it goes to the board and so forth.

It's like a gating thing. You really want to have everything to think about things strategically versus just projects. So you might imagine you're a company that has fulfillment centers. You might imagine there's some little project within that fulfillment center that looks great on an IRR basis. But the overall strategy, you have too many distribution centers. So you end up shutting down the distribution center. So it's a good project within a bad strategy.

So being able to understand and think about things strategically is really important. And then the other one I would mention is just literally a willingness to act. One of the chapters in Will's book is about Bill Steeritz at Ralston Purina. and Bill was described as a bit of a card player. He's holding cards and he thinks about how do I improve my lot in buying and selling businesses. By the way, Billy Bean, probably Oakland Hays does just do. Like every moment, they have a sense of what their players are worth

and what other players are worth and at what point should they trade them to build value. And a willingness to act is really important. So that flexibility is also key. When is the right time to buy back stock, when's the right time to issue stock, sell a business, buy a business, so and so forth. I'll tell you one funny anecdote about this, that every year my course at Columbia Business School, I bring in an executive to talk about competitive strategy. And I brought an executive.

They were a consumer products company. They had their core business phenomenal. And they had the side business. It was a consumer products company, but totally different characteristics, much more cyclical and so forth. The students were pounding on this guy. Like, why do you own this thing? Like, this is not your core, so and so forth. And by the way, the story was the start of this business was like 20 miles down the road.

It was a family business. this they kind of struggles. These guys said, they screw, we'll just throw them a little lifelines. They end up buying it. And then they end up building making a bunch of acquisitions around. Right. So the CFO is trying to like dance around this thing.

And at the end, the class is over. In the end, he comes up to me and he goes, okay, let me just tell you the real story. The real story is that the CEO lives across the street from me. And we coach our kids baseball teams together. He's part of the community. So I know intellectually that we should not own this business, but it's really hard for me to go to the CEO and ultimately to the board to say,

let's sell this thing because my neighbor and my co-coach is going to be gone. And he's like, it's not that bad. So these are the real human elements that come into this stuff. And I'll just circle back on John Graham. Last thing I'll say that he mentions the true characteristics, just observing these CFOs through these surveys. They tend to be very conservative. They use hurdle rates way above their cost of capital.

They tend to be very conservative for their capital structures, for example. They tend to be very miscalibrated. So they don't really understand the future. and they don't understand how much they don't understand the future. Those are all dimensions that companies could improve on. And I think those are the things we'd look for for the companies that are great at capital allocation. If you think about the, let's call it six, seven years from the time you first did this deep dive to today, what has most surprised you about how markets and companies have unfolded?

Even just between the last time that you and I talked and now has been one of the craziest market shakeouts ever. You know, when we talked to was still in the early phases of, of 2020 and the run-up in valuations of some of the internet stocks, the work-from-home stocks, et cetera, and that's all mostly crashing back down. What does it feel like for you between the two times that you've written this thing on capital allocation? Has anything major changed in your mind? Or are a lot of these things fundamentally human and basic to business that if we did

this again in 20 years, it's going to be similar? The big thing that's probably changed is the backdrop in which we do business. We spoke, I think, in the summer, around the summer of 2020, and that was when there's a lot of equity issuance, hot IPO market. The spec market, I think, may have peaked in the first quarter of 2021, but that was sort of rolling big time. As you pointed out, the businesses that were doing well were the digital type of businesses, the value stocks were doing very poorly. To me, what the biggest change has been, what's happened in the last 12 months.

One is real rates on the 10-year have gone from negative 100 basis points to positive, whatever it is, 120 basis points. So when you have a 200 plus basis points swing in real rates, I don't care what asset class you're in. There's no place to hide. So equities, credit, you're seeing that. By the way, interestingly, if you look at Aswat the Motor and a wonderful professor of finance at New York University, every month he publishes an estimate of the equity risk premium, which allows you to estimate the expected return for the equity markets. His January 1, 2020, reading was 5.3 quarters, 5.75%. The expected inflation component is off the 10,000, year. It's like 250 basis points. So you're talking about 3%-ish real, a little over 3% real,

which is half of what we've earned historically. The most recent readings now are in the high eights. The inflation numbers have actually been very stable. So now we're much more in the range. To me, what's interesting is we've gotten very frothy, expected returns certainly a year ago had gotten very low. And in the last 12 months, high yield spreads are up 150 basis points. Triple B spreads are up 70 basis points, equity risk premium is up. So as painful as the asset class adjustments have been, the expected returns are much higher. So then you say, as a company, what do we do about that? And the answer is these sharp changes tend to create dislocations and potentially opportunities for companies to do things with capital allocation.

Now, the other thing I'll just say, I think Patrick, we may have talked about this briefly even in 2020. COVID was really hard for management teams and investors, of course, but management teams. because you had this screeching to a halt anything related to services. You had a huge amount of stimulus. So if you're Amazon.com, you say to yourself, there's huge demand for what we do. They spend a huge amount of money to try to accommodate that. First, I'm sure, just to capture the sales, but also almost societally, they're an important good. And then that reverses.

So now all of a sudden you've built all this capacity, put in all this investment. All of a sudden, the demand doesn't perpetuate. It's been very difficult, I think, for management teams to it. adjust. But these all end up presenting opportunities from a capital allocation point of view, whether that's M&A, share buybacks, and so forth. That to me is the biggest change. And again, typically presents not only challenges, but also opportunities. In that market share piece where we started, you invoked Clayton Christensen a little bit in this notion of disruptive innovation and how that plays such a key part in changing dynamics of market share or the creation of new industries where those dynamics

then start to shake out. As we sit here today, it appears like the one that is most relevant right now is artificial intelligence, which you and I've talked about many times for many, many years, but really has seemed to hit, at least on the consumer side, like a tipping point, whether it's just of awareness or of capability or of whatever. But all of a sudden, you're hearing people talk about Google's moat in the face of artificial intelligence, which is maybe the most unassailable business note that exists today. How are you thinking about the playing field right now for sources of destructive innovation? What has your attention? What are you watching? I think you've nailed it. I mean, the one area is this area of artificial intelligence. I have to say I'm certainly far from

being an expert. And there are other people know much more about this than I do. That said, my oldest son is a data scientist, works in artificial intelligence type of areas. I try to pick his brain all the time about these types of issues. I think the one thing that's been a little bit surprising is the performance improvements we've seen. For example, Open AIs, GPT to GPT3, we'll see GPT4 at some point. I think those performance improvements have been greater than people had anticipated. That, I think, will completely be disruptive. But there's two big questions. One is, every time new technologies emerge, people always freak out that there are going to be tons of jobs lost. And historically, that's not been the case. So the question is, what happens to the job situation? Does that disrupt the

labor market even further. And then the second question, who wins in this? Who creates value ultimately? And it's not always, at least in my mind, crystal clear at the outset. I also think you have to follow still amazingly interesting developments in health care. That's another fascinating area. So there's certainly a lot of really exciting stuff going on that will certainly change the world we live in the next five or ten years and just trying to think about it and keep up with it as well as possible. Well, I would encourage everyone to go read these three reports that have been out in the last couple months. Actually, it's four, but three themes on return on capital, capital allocation and market share. I just feel like the theme of the market is, in my mind, return to normal,

return to talking about these concepts and caring about the basics of the businesses and where they earn their excess return. Do they earn an excess return? I don't really know of anyone that can teach that better than you. So I'm glad that you've had this explosion of research. I'm thankful for your time, as always, today. And encourage everyone, go out and check out these three reports. Michael, as always, a total blast, 47,000 notes and books. I need to go read and chase down as every time we talk. And I know you didn't ask me, but the kindest thing anyone's ever done is let me go on their podcast again.

Thanks, Patrick. Have a great day. See, Michael. If you enjoy this episode, check out JoinColossus.com. There you'll find every episode of this podcast complete with transcripts, show notes, and resources to keep learning. You can also sign up for our newsletter, Colossus Weekly, where we condense episodes to the big ideas, quotations, and more, as well as share the best content we find on the internet every week.