Introduction

Patrick: [00:01:21] My guest this week is Pat Dorsey who is the longtime director of research at Morningstar where he specialized in what we call economic moats, sources of sustained competitive advantage that allow a few companies to deliver huge returns over time. Several years ago he left Morningstar to form his own asset management firm, Dorsey Asset Management, and build a portfolio of companies with wide moats like those he studied at Morningstar.

And while moats are critical, equally important is how companies allocate the capital generated, or made possible, by the existence of the moat.

In the vast majority of the conversations you hear I'm meeting the guest for the first time. I mention this to encourage you to connect me with anyone whose story or way of looking at the world might resonate. Always feel free to contact me with ideas.

Pat and I begin our discussion with the key differences between the sell-side and the buy-side, and then discuss all aspects of moats and capital allocation. Please enjoy our conversation.

Understanding Moats

Patrick: [00:02:21] Pat, well I thought a fun place to start would be the transition from the sell-side to the buy-side. I'm always intrigued by people that cross that threshold and so I'd love to hear what was the most surprising and/or difficult thing during that transition phase.

Pat: [00:02:38] There are two big differences in going from the sell-side, which I promise not to take umbrage at given that we were an independent firm and not commission based at Morningstar, but that's all right. I mean we were selling equity research. On that side of the street, you're paid to say yes. On the buy-side you're paid to say no.

That's the simplest answer I can give because if you're selling research everybody wants an idea. Everybody wants an idea. What you got? What you got? What you got? What you got? On the buy-side, you're paid to not lose money. You're paid to do a small number of smart things which means you have to say no to a lot of stuff.

I think the second big difference is that on the sales side there's no sizing. Everything is binary. I like it. I don't like it. It's a buy. It's a sell. But it's much harder on the buy-side because is it a 5% buy? Is it a 10% position buy? Is it a buy all 5% now? Is it a buy half and wait? It's a much more nuanced process than just, "Do I like it or not?"

Patrick: [00:03:40] We're going to spend most of our time talking about your process for evaluating businesses which was rooted in your research on the sell-side. You developed Morningstar's, or led the development of Morningstar's moat ranking, or I don't know if there's a specific name, but effectively a look at how defensible an economic moat is for stocks around the world. The framework that we'll follow today is first to talk about what the hell a moat is.

Everyone talks about moats but it's kind of hard to dimensionalize or understand what that means. Then two, capital allocation and how the market misprices moats. We'll begin with moats themselves, a term popularized I think by Buffet. I'll let you begin by just describing what that might mean and then we'll get into how to identify different kinds of moats and what they might mean.

Pat: [00:04:30] The simplest way of thinking about it is it's a structural characteristic of a business, and structural meaning inherent to the business. You cannot separate this characteristic from the business itself. That insulates the business from the competition. That's the simplest way of thinking about it.

You can't think of Tiffany without the Tiffany brand. The two are one and the same. You can't think of SAP without thinking about switching costs. It's really hard to rip the stuff out once you have it, and so that structural attribute is what gives the business leeway to reinvest capital at high rates of return not without competition, but making it a bit more difficult for the competition to do what economic theory says it should, and which generally does happen for the vast majority of businesses, which is drive down above average returns on capital.

Patrick: [00:05:14] Return on invested capital is an interesting way of thinking about moat, but the reinvestment potential is a key attribute as well. Michael Mauboussin and others have done really interesting research on the mean-reverting nature of returns on invested capital. Now, they do mean revert but not necessarily instantly, or it could take up to 10 years. Could you give a sense for what part of the stock market universe might have in percentage terms or something like that, might have one of these kind of moats that we talk about?

Pat: [00:05:46] Oh, sure. That's a great question because I think it's something that's utterly undiscussed in classical finance textbooks, which is that not all businesses are created equal. It's also not something you really get from the sell-side because the sell-side is full of specialists who are paid to say, "Of the really awful companies I cover this is the least awful." That was, in fact, actually my personal frustration with Michael Porter "Five Forces" work that led me to develop the moat work at Morningstar, which is that Porter's role was writing for business owners and he was a consultant.

You don't gain business as a consultant by narrowing your market and saying, "If you run an auto parts company you're hosed. Just do something different." You want auto parts customers. The idea there was maximize even if you have an awful business. Whereas as an investor, well you can just skip the business. You don't have to invest.

It's a critical, I think, thing to remember, that business quality, economic moats, are not normally distributed across the market. There are some parts of the business, consumer packaged goods, enterprise software, that just have the competitive advantage all over the place. There's plenty of others, life insurance, commodity chemicals, pick your area, that just stink, just really hard businesses.

Patrick: [00:06:57] Let's talk about your framework for identifying and measuring moats. Four key categories, and maybe we can start with intangibles. What intangibles might represent a moat? And throughout all these questions I'm also curious to hear your opinion on things that might signal or look like it might mean a moat but in fact, is sort of like fool's gold.

Pat: [00:07:17] Yeah, absolutely.

Patrick: [00:07:18] We'll start with intangibles. How might companies, this is relevant for business owners too I think, how might companies build, or as an investor, identify a moat and intangibles.

Pat: [00:07:27] Yeah. The intangible asset is just what it sounds like. It's not on the balance sheet but it gives pricing power. You can think of a brand. You can think of a patent. You can think of a license. You have a landfill or a gravel pit, or to make a particular type of aircraft part. All of which are intangibles but which give the owner of that intangible pricing power.

I think the key thing to be wary of, to address your fool's gold question, is companies love to talk about their brand. They love to say how well known they are. But if that brand doesn't translate into pricing power it's frankly not worth what the company spends to maintain it. You always have to be thinking, what advantage does this brand give you? Does it allow you to raise prices? Does it allow you to price ahead of the competition?

If it's a license you need to think about the regulatory regime. Because what the regulatory gives he can take away as well. You've seen people write up really beautiful thesis on companies because they have this license to do X, Y, and Z, without ignoring the fact that it's perhaps in a country that doesn't have the best rule of law and there's six more licenses that are about to get handed out to the boss's brands.

Power of Brands

Patrick: [00:08:32] Let's talk about brands specifically and the different types of brand, and what might be more or less durable. We'll come from the positive angle. When you're looking for a company whose moat is dependent on its brand what are the markers that you're looking for?

Pat: [00:08:46] I think the thing to ask yourself is always, "What does this brand do for the consumer?" Does it create positional value for me? When I wear a Rolex watch, which I don't, but if I did it tells the time just the same as a Timex. But it says, "Hi, I have a lot of money," and that gives positional value to people, right?

That's the value of that brand. There are other brands that confer trust. A Morningstar Mutual Fund rating, the opinion of a Gartner IT analyst. A Moody's Bond rating. These are all things that confer trust because enough people have used that rating that it has value greater than what it cost to produce it.

It doesn't take Moody's any more to produce a bond rating than it would you or I to look at a bond and rate the damn thing, but yet they can price a lot higher than we can because there's social proof involved. Nobody will accept the Pat & Pat bond rating, but they'll accept the Moody's Bond rating and enable to access capital markets with it.

Patrick: [00:09:44] It's a really interesting idea where in all those examples I think a key part of that is history, and ubiquity. Even beyond Rolex other watch ads will show like intergenerational pictures and-

Pat: [00:09:55] Oh, that Patek Philippe ad, yeah. That's great.

Patrick: [00:09:58] These brands have, for the most part I think everyone you listed, been around a long time and been doing something very consistently. That trust, or even positional value is based on everyone knowing about it and it having been around for a long time. How often do you see upstarts in those kind of well-established spaces?

Obviously Moody's and others may have been the first rating agency, so there's a first-mover advantage that can grow into the brand we're talking about, but what about the capacity, or from your investor's seat, the potential for a new upstart to come in and take part of Rolex?

Pat: [00:10:32] This is a fabulous question because it is so unbelievably relevant right now, especially to anyone who has historically invested in consumer product, consumer packaged goods, where the brand does not really confirm positional value. Nobody thinks you're a better person because you use Tide, but what it does is it reduces your search cost.

You see the orange box and you grab it. Those brands, I would argue, that type of brand rather, is much more subject to disruption now than it has been at any point in the past 30 or 40 years because those brands were built by control of mass media.

If you're Proctor & Gamble, if you're Unilever, you could afford to advertise on broadcast TV. Nobody else could. You had a privileged position right there. Now you look at what's happening with Dollar Shave, with Chobani, because you can reach millions if not hundreds of millions of consumers with YouTube, with Google AdWords, with viral videos, with tons and tons of digital media, the barrier to creating mass recognition is vastly lower.

I think those brands, I would argue, are much less durable today. Much is maybe a little overstating it but I think you need to be very cognizant of that if you own one of the Buffetist inevitables. How inevitable is it truly?

Conversely the ability of things to go viral and scale quickly, I think it's fair to say Amazon is one of the more trusted consumer brands right now. It didn't exist 20 years ago. I mean this is not Rolex. This is not Cartier. It's not been around 200 years, but because from day one Amazon realized that offline is actually a better business model than online because it has the ability to confer trust.

Offline I give you a dollar. You give me a good, and we're finished. We have no relationship anymore. Online there's all this trust involved in terms of the user experience being what I think it's going to be, or what I've been promised that it will be. Amazon figured that out from day one.

Patrick: [00:12:25] Let's move to switching cost which is the second of four key categories. Tell us a bit about what you look for there or look to avoid when thinking about moat and durability of a moat.

Pat: [00:12:35] Yeah. The durability really is the key question here because if a business has high switching costs, or captive consumers is another way to think about it, it has the incentive sometimes to abuse its position. Those of us in the financial community generally loathe the word Bloomberg when it is mentioned given the pricing power that it tends to exercise in often very rude ways I think is the best way to put it.

I would argue that's an example of perhaps close to abusive pricing power, where perhaps you're not maximizing your market. You're maximizing price. What you want when you have a high switching cost business is one that is reinvesting in improving the product and delivering value so that the consumer feels like, "Yeah, I just paid 3% more but I'm getting 4% more," in terms of the value of the product.

The switching cost businesses that treat the product as a cash cow and don't reinvest, and just milk it, Advent to use another financial example, would be a perfect example of this, of a product that looks like it was made in about 1960, but we have no choice but to use it.

Patrick: [00:13:38] What is a good example or a favorite example of a firm that's done managed switching cost the right ways, in the way you just described?

Pat: [00:13:44] Oh, boy. I think a good place to look for this is on vertical market software, so software that's become the language of an industry, the plumbing of an industry. You could argue that Microsoft has managed it pretty well, although the usability of Windows and of Office has often been exceeded by other products; the G-Suite for example. A lot of people actually like G-Suite better.

The ability of Office to link in and hook into lots of other things that people use has kept it much stronger than you might have thought. I mean the demise of Office has been called I don't know how many times. What would be another good example? AVEVA, which is a company that we own. This is a company that makes essentially very high-end AutoCAD software for designing process plants.

Think a chemical plant. Think of an oil rig with a lot of pipes in it. Very, very complex processes. Their customers are in a highly cyclical business. If you're an construction contractor, you're KBR for example, you're project-based. When capex is high you're loving life. When capex is low you're hurting.

What they've done is made it so that those businesses can flex down their spend but not too much. AVEVA doesn't have to watch all of its business go away, but the business is able to flex down a little bit because it is a cyclical business.

Patrick: [00:14:55] The third which is one of my favorites is this notion of network effects. This is the talk of the town that a lot of... I know you own quite a bit of Facebook so maybe a prime example of a network effect. But that's obvious right? Network effect being the more people that come involved the more valuable the network becomes.

But dig into that in a little more detail. What, outside of the obvious examples, are some good things to think about when looking for network effects, and how businesses themselves should be thinking about developing them?

Pat: [00:15:23] Well, I think the key thing there is to always think as my users grow, does the fact that there are more users on the network add value to users number 10, 20, and 30, the folks who were early on for example? Restaurants are the canonical opposite example. The more people who go to a restaurant it actually becomes less valuable to you, because the more crowded the food may get worse.

But of course the more places I can use my Visa or MasterCard the more valuable that network becomes to anybody who uses it. I think one place people often get confused with networks is what are called radial versus interactive networks.

Interactive network is Facebook, basically tons and tons of nodes connecting with each other. A radial network has basically lanes. There's more hub and spoke. Western Union would be a good example. Western Union likes to tout having the most locations globally to send money back and forth, but most of the money is sent along specific lanes. It's sent from Chicago to Mexico City, or it's sent from Chicago to Bangladesh. It is now sent from Bangladesh to Mexico City, okay?

Those individual lanes are vulnerable to a lower-cost competitor. The fact that it's a red herring. It's saying that, "Oh, we have the most locations" because that's not a relevant number to the moat. The moat is much less durable than say a Visa or a MasterCard where it's more like a web than a radial network or hub and spoke.

Patrick: [00:16:42] Really interesting. Yeah. I never thought about the distinction between those two, but I guess you want it to be as disperse as possible like a Facebook, where you remove any one of those vertices in the network and it doesn't really matter. Whereas I guess if you pull Chicago, Bangladesh out of Western Union you've got a big issue.

Pat: [00:17:00] Yeah. Exactly. Exactly. Or it's more from a competitive standpoint than anything else. Because if you wanted to attack Facebook you have to get all these users on one place or another. But if I wanted to basically just attract Mexican immigrants in Chicago who want to send money back home to small towns in Mexico I can attack Western Union with that without touching its Asia Franchise, without touching any of these other franchises. I don't have to go after the whole thing. I can just go after a piece of it. You can find focused competitors doing that.

Patrick: [00:17:28] The last one is cost advantages, economies of scale, et cetera. Maybe sum up the key... In today's environment, obviously scale economies is pretty straightforward, but in terms of change what new ways are cost advantages being used-

Pat: [00:17:41] Yeah. No that's a great question.

Patrick: [00:17:42] ... by markets?

Pat: [00:17:43] I think we are trained to think about cost advantages in terms of manufacturing and B school 101, like reduce the marginal unit cost because you can produce at a lower marginal cost. But what's becoming more and more common with value being produced by intangible assets, value being produced by R&D, value being produced by scope, is the value of scale to lend durability to a business.

Think about Google and Facebook's investments in artificial intelligence right now. It's a hot buzzword but it is a good example because machine learning benefits from scale. The more data you feed through it the better it gets.

The more data you feed through what you've got that's now better, the better that gets. It's not scale advantage in terms of it reduces cost. It doesn't reduce cost at all but it improves the product. It improves the output for the consumer. I would argue that's a scale advantage in the same way that having the biggest factory to produce widgets is.

Another thing I think people often miss is scope advantages, that you've just been doing something longer than anybody else. If you've been focused on a very small part of the market for 30 years and you've nailed all these small little nuances to running dollar stores. In the US the dollar store has come down to basically a duopoly at this point.

Retail is a tough business, so it's certainly not as durable as a moat as having the AI scale that Google and Facebook do. But I think, and I'll admit to this, I think it's easy to undervalue experience, the ability to serve a niche profitably for a very long period of time, and to have developed just lots and lots of little trade secrets that make it hard for the competition to replicate.

Moat Ranking Strategies

Patrick: [00:19:29] These four dimensions are all interesting, especially when they start to interact to create maybe very defensible moats. But now we need to move to the investor side of the equation. First question will be, how do you take those four components and as objectively as possible come up with some ranking or system? I'm a quant so I can't help but think about it this way, but is there a method by which we could take these four...

And I realize that the determination of the four might be very subjective, but let's say you've got some opinion of the four for company X, Y, Z. How do you start to build this into an investing framework that can be used in an actual strategy?

Pat: [00:20:07] Yep. Well, I think the first thing is four is not better than one. Realize that if you've got one incredible brand, Tiffany's brand... One incredible moat, Tiffany's brand, or Facebook's network effect, whether you have the others or not doesn't really mater that much. You don't need to have... Four is not better than one.

Four moats is not better than one moat. What matters is your ability to maximize that moat, to reinvest back in that moat. That's why we have gravitated towards businesses that are less mature and can reinvest, because I think where moats become Hungarian lollapaloozas is when each dollar of cash flow can be reinvested in an incrementally higher rate of return.

That doesn't happen. If you have a static moat that's great. Coca-Cola, right? The moat is the moat, but it's not really getting bigger, and the reinvestment potential isn't very high. All the cash needs to get paid out to shareholders who have dividends or repurchases. But if you have a business where each year the moat gets a little bit better, and so the returns on capital get a little bit higher each year and there's reinvestment potential. The dollar that's reinvested at 18% this year gets reinvested at 20% next year, gets reinvested at 22% the year after that.

Patrick: [00:21:14] How do you think about this from a mispricing standpoint? There's only alpha if you've done a better job of evaluating the value of say, a moat, than the market has. Some of this comes in timing. I think it's hard to argue markets haven't focused more and more on the shorter and shorter term. It's where a lot of the brainpower and machine power is being focused is next quarter's earnings.

Tying into this idea of a longer time horizon as an advantage how do you think about, okay let's say... Facebook. Let's stick with Facebook since you own it and it's obviously well known. It has an enormous network effect, but I don't think anyone doesn't think that. Then the alpha becomes, well, its network effect is actually more powerful than the market has discounted. How do you think about identifying those gaps specifically?

Pat: [00:22:02] That's actually a perfect example, because on Facebook's Q2 17 earnings call that was just last week, probably half the analyst questions were on Messenger, which is not monetized now and will probably not contribute materially to earnings for several years. No one asked a whole lot of questions about the core platform that's growing 50% per year, and cranking out a 47% operating margin, and how long that can continue.

That's what you need to be focused on is what's the return on ad spend for an advertiser right now? In the case of Facebook the alpha comes not from saying, "Oh, there's a network effect here." The alpha comes from looking at it from the advertiser's side, not the consumer side, and getting icky.

Advertising is icky, right? Google doesn't like to admit it's an advertising company. Facebook doesn't like to admit it's an advertising company. "We connect humanity." You're a big ad company. But once you sit in the advertiser's seat and say, "What does this platform do for you that other platforms cannot do? How does it enable you to reach people in ways that other platforms do not enable you to reach people? What is the return on spend on the dollars you give to it," the answers that come back are incredible.

Well, that's when things get interesting. Network effects are great but if they're not monetized they don't mean doodly to an investor. I think focusing on not the ethereal, "Ooh, there's a network effect," but how do you make money off that? How do you monetize it? I mean Yelp is a great example. Yelp has under-monetized. Or TripAdvisor has arguably under-monetized their own moats and the review databases that they have. The network effect, it's not as valuable if you can't monetize.

Patrick: [00:23:36] How do you think about price; traditional, statistical, cheapness? So Google and Facebook, two examples we've talked about, by traditional measures pick your fundamental, price to whatever, look very expensive relative to... Maybe not Google as much but ...

Pat: [00:23:49] Yeah. Not Google as much.

Patrick: [00:23:51] ... certainly Facebook looks statistically very expensive. How do you square that multiple? How do you come to a market consensus multiple and incorporate current price or valuation in your review for potential future return?

Pat: [00:24:05] Yeah. That's a great question and I think the first thing is to take one of the first words in your question, statistically, and eliminate it from your jargon. Eliminate it from your analysis, because the businesses with moats that compound best over time have futures that are vastly different than their past. Statistical, quote-unquote, valuation just can't see that.

It can say where is it today relative to where it has been? When MasterCard came public it was the first of the payment networks to go public. It was printing like a 13%, 15% operating margin. Okay. That's pretty good, right? It just doesn't look that good as a business.

But if you then think about the qualitative characteristics, that it was owned by a whole bunch of banks before it became public which means it was probably not managed optimally, and then if you just run through the very simple mental exercise of okay, MasterCard gets one dollar of incremental revenue, what is the cost it needs to put in to generate that dollar?

Not much. The operating leverage thesis is just not rocket science to come up with, but statistical analysis isn't going to show you that. Facebook, same thing. Two years ago they printed a 31%, a mid 30s operating margin, low 30s, because they just bought WhatsApp and a lot of that got allocated to R&D. But that doesn't have anything to do with the incremental economics of the business, the fact that they pay nothing to generate their content and the average advertiser pays them a whole bunch to get access to the users. What it involves, I think, is carefully, logically, and creatively thinking through the core economics of the business.

Patrick: [00:25:38] Let's say there's a moat. What should Amazon do for example? What should be its future strategy?

Pat: [00:25:42] Well, I mean basically strategy and capital allocation should both serve the moat, both serve to strengthen the moat. As you point out Amazon is the canonical example of this because everything has been focused on making the customer's life easier. If the customers life is easier they'll use it more. To go back to Facebook it's how do I keep users engaged?

If users aren't engaged, if they don't come back to the platform every day, there's less reason for the other users to come back to the platform every day. You have to keep that flywheel going. In their case it's adding things like live video and whatever else it might be. If your moat is a brand then you have to ask yourself... Again, break it down into is it a positional brand?

Is it a low search cost? Is it one that Garner trusts? If it's one, just to give the example one that Gartner trusts, you have to ensure that that trust is based on social proof. That's why you see Gartner analysts always out there in the media saying, "Oh, we think the market share for this thing will be this or that."

And why? Because that gets the name recognition of Gartner out there. It says that if I am a CTO and I go to my boss and say, "We should spend, on this software product versus that one," why? Well Gartner thinks we're the better one. That has some resonance to him or her.

Patrick: [00:26:49] Before we get deeper into capital allocation I've got a fun question which is, we established the four types of, or four dimensions of the moat. We'll go sub-moat, so a brand could be the answer here versus intangible. If you were starting a business, and we don't know what the business is, but I bestowed you ahead of time with the knowledge that you're going to get...

One of these aspects of brand is going to be fantastic. What would you pick? What would be the most intriguing dimension of moat if you could guarantee you had it ahead of time?

Pat: [00:27:22] Well, if you can guarantee that you have it then it's a network effect because it feeds on itself. It grows... A network effect business grows itself with very little marketing dollars. The fact that you said guarantee takes me straight to that position.

Patrick: [00:27:34] What about...

Pat: [00:27:35] But without the guarantee-

Patrick: [00:27:37] Yeah...

Pat: [00:27:38] Yeah. Without the guarantee, I think it's high customer switching costs. When you have a product that becomes integral to a customer's life, they literally cannot do their job or function without whatever product you might have, that is an insanely powerful place to be and it just gives you enormous pricing power over time. Sometimes those businesses do have challenges growing because basically, you become integral to the lives of all the customers you're going to get.

Then it becomes a return of capital story. You see this in vertical market software ("VMS") companies. They sometimes have a hard time growing. That's where, to your question, capital allocation becomes a bigger and bigger question, because what you don't want is a VMS company that spits off cash going out and buying another VMS company.

I mean if you're Mark Leonard at Constellation, and that's part of your strategy well then great, but if you're just dominant in one business and you go out and buy another one, two plus two does not equal four. Two plus two probably equals two and a half.

Patrick: [00:28:31] My guess, I was going through the list before our conversation, and I would have guessed that you would have said positional value brand, something like the Rolex, because of how... I'm always interested in things that don't change, so things that reliably in 10 years will be the same. Sort of like the opposite of a technology company or something like that. It seems like there's such low turnover in the big positional brands.

Pat: [00:28:56] You're totally right about that. I think those are ones where, as an investor if you can essentially look at those... I would say those are the brands that lend themselves best to your statistical valuation analysis which you were just saying. When Tiffany gets statistically cheap, when Richemont joins Cartier and a whole bunch of other big brands it gets statistically cheap, you should take a hard look.

If the moat has not been impaired you should take an even harder look. I mean the only caveat there is you do have occasional spurts in demand. There was about 10 years ago a huge spurt in demand for luxury products coming out of China partially because people were getting more money, but partially as bribes.

With the big corruption crackdown a few years ago you saw a structural slowdown in growth. You would want to make sure that your statistical analysis incorporated a longer view of growth, which is probably like 4% to 6% volume growth, and not the one that was bumped up for a few years by the China experience. Because if you use that as your arbiter of statistical cheapness, you probably would be misled into paying too high a price.

Thoughtful Capital Allocation

Patrick: [00:30:01] I love your line that capital allocation is effectively the link between business value and shareholder value. We've spent the conversation thus far really talking about business value, things that create good operating companies. But it's a really important point that two things can screw that up: The price you pay, so you can overpay for a great business and then it's not a good investment.

But maybe more interesting is that if you have a terrible capital allocation plan or managers that are awful in allocating capital once it's been generated, that can also destroy, from an investor's seat, the value of a fantastic operating business. Talk about your view on what makes for a good capital allocator and what you look for in investments.

Pat: [00:30:40] Thoughtfulness I think is probably, if I had to summarize it in one word. We frequently talk to companies and say, "Okay, let's talk about buybacks... "Oh, well our goal is to do about a third buybacks, and a third dividends, and then a third accretive M&A." I mean that's the dumbest assed answer I've ever heard.

I mean that's not a thoughtful answer. That's a pull it out of your ass answer, you know? If you have tons and tons of opportunity in a given year cut the dividend and buy a company. I mean great. The market will probably kill your stock but you'll generate a lot of value. It's thoughtfulness.

It's about thinking through what makes the most sense for generating future returns for your business, whether that's buying back stock, or reinvesting in projects. I think you see this in the UK, Australia, some other parts of Europe. You have what we call the dividend fetish, where basically paying out dividends is the "sine qua non" of a great business, which is just asinine.

Of course in the US you have this buyback fetish, which is of course equally moronic because when your buybacks are just sopping up options dilution it doesn't add a damn bit of value to me. I think it's that capital allocation needs to be thought through just as carefully as business strategy in the same way that you would want a competent CEO to say, "Okay, how are we building the moat," and lay that out very clearly.

Say, "Okay, how does your capital allocation strategy help me as your shareholder make money?" You want a thoughtful answer to that. I know that's loosey-goosey, but because there is no good answer. For some businesses it could be pay out the dividend. For some business it could be just buy back shares like a banshee over five years' time.

For some businesses it could be "acquire like mad". But the key is don't have it be a stock off the shelf answer. Have it be an answer that's specific to that business, and that does not do things just because other people do them.

Patrick: [00:32:26] One of the things we studied quantitatively, which obviously this is a very qualitative area but-

Pat: [00:32:32] But it can be studied qualitatively, sure.

Patrick: [00:32:34] ... quantitatively one of the interesting findings is looking at lumpiness. You mentioned that idea of a third, a third, a third. What you find is that companies that do that kind of thing, just have a steady-state capital allocation, it doesn't mean they're going to be terrible under-performers, but the best-performing stocks tend to have a little bit more lumpy type results.

Take buybacks as the example. Companies that do 1% every year, 2% every year, don't really outperform, but ones that randomly gobble up 10% of their shares in a very short window tend to outperform by quite a wide margins.

Pat: [00:33:06] Yeah. That's cool. You kind of run a high level of variance and repurchases?

Patrick: [00:33:11] Yeah, exactly. Or variants of capital allocation policy, if there's a pie that you can do these five or six things. If that moves around a little bit more, I guess you'd have to have success at moving it around so it's not easy, but that that might make more sense than the just like, "It's fixed. Just do this same thing-"

Pat: [00:33:30] Yeah. No.

Patrick: [00:33:30] "... year in year out."

Pat: [00:33:31] Totally agree. Totally agree. That makes intuitive sense because acquisitions don't come along every day, and your stock's not cheap every day.

Patrick: [00:33:38] I think maybe one rule of thumb would be, you said thoughtfulness, I would think about it in terms of what's the highest return? It sounds obvious just as investors look for companies that are going to deliver the highest return. The CEO should be doing the same thing. Sometimes that might mean buy a company, sometimes it might mean send everything back because you're mature.

Pat: [00:33:55] Yeah. Highest return and setting a high and reasonable hurdle rate. We see this especially with the European companies we look at given how low interest rates are in Europe, is that companies think that beating their WACC in year three, when WACC is seven, is like a good outcome. That's our hurdle.

God. Oh my god. Just a few months ago I sat down with a company in London and they're one of these, "Yeah, our goal is to beat our WACC in year three," kind of things. I was just like, "What? Why is that your hurdle rate?" I just got this blank look from the woman I was talking to, like, "Well, what other hurdle rate would we use?" An absolute number. Something higher.

As an investor I can buy anything in the world. I can allocate capital anywhere. Why would I allocate my capital to you, sir, or in this case it was a ma'am, to then reallocate my capital with a 7% hurdle rate? That makes no sense whatsoever. That's another one I would say is that in addition to thoughtfulness, absolutely hurdles.

Relative hurdles, like meeting our WACC, it's squishy. The goal in investing is to make money, compound capital, whether you're a CFO or a portfolio manager. If you don't have an absolute bar that, "We think 10% is good. We think 12% is good. We think 20% is achievable," whatever, "Things need to clear that hurdle for us to allocate capital," you can talk yourself into some pretty dumb stuff.

Patrick: [00:35:17] You really do need independent thinking, not just buying back shares. Because other people are... Independent thought seems like another key attribute of a thoughtful capital allocator.

Pat: [00:35:26] Yeah. Yeah. A willingness to do stuff that looks optically stupid, but then later on turns out to be the smartest thing going. Just to go back to Amazon, do you remember when they locked Amazon Prime? I mean people thought that was... "Free shipping? What?" People went crazy, right? That's a good example.

I never thought about this. That's actually probably a good one, not a bad test that when you have a decision by a company where there's a giant hew and cry, and everyone just says, "Oh, that's the dumbest thing ever," it's probably worth a hard look.

Identifying the Best Investment Opportunities

Patrick: [00:35:59] I'd love to try to codify your process a little bit. We talked a lot about the key things you look at, but like anyone... I think it's all equities right, your portfolio? And fairly concentrated, so 10 to 15 stocks, so certainly no closet index. From start to finish what are the stages? You've got some investible universe. You've got LP dollars. You need to start winnowing that universe down. What are the stages of that?

How do you think about it? As many stages as we can draw. I'm fascinated by this, for more discretionary, traditional asset managers relative to quant where obviously it's exactly codified. How does that... Starting at 3,000 companies, or 5,000 companies you can buy, what happens next?

Pat: [00:36:44] I think the first thing for us is throw out all the stuff that we're not interested in. These would be commodity chemicals, life insurance, businesses that just have structurally awful economics. Those tend to comprise a higher proportion of the equity universe outside of the US than in, so it actually kind of narrows the world pretty fast.

Then you start fishing in the pond. That's step one. Fish where the fish are. Don't fish where there are no fish. Then the next step is really thinking about businesses, just sitting down and thinking about businesses that have structural tailwinds behind them. In the case of Facebook, digital advertising, your advertising going from offline to online, eCommerce in the case of Amazon.

When companies have the wind at their back, and their end market is growing quickly, you can... I don't want to say cover up a lot of sins but just really it sure makes their life easier. Thing one, find businesses with good structural economics which tend to, as we said earlier, cluster in specific industries.

Patrick: [00:37:40] Could you list maybe, I don't know, five examples of industries like that?

Pat: [00:37:44] Oh, yeah. Okay, so the payment networks right? Obviously that's an easy one, you know MasterCard, Visa. The processors, that's, I wouldn't say a commodity industry, but there's a not a lot that differentiates one merchant acquirer from another merchant acquirer outside from scale. To some extent they all kind of do the same thing. They have similar tech stacks I would argue.

The vertical market software, software that becomes the language, the plumbing you might say of an industry. Exchanges. Exchanges tend to be beautiful businesses. Futures exchanges tend to have much better margins than cash equity exchanges because futures are non-fungible. You have to open and close the position in the same exchange as opposed to, you can buy some NYSE sell it on ARCA.

Those tend to have very good structural economics. Obviously as we've talked about the positional goods, luxury brands, consumer products, those tend to have very good economics but have challenges when it comes to reinvestment. That's where I think capital allocation becomes more of a challenge.

Patrick: [00:38:36] You used the term good economics. Can you be a little more specific on what that means? Is it margins? Is it...

Pat: [00:38:41] Well, yeah. Yeah. The end result, I mean good economics at the end of the day means good or high returns on capital, which you can achieve with low margins and high asset velocity, or low asset velocity and high margins. But when I think about good economics it's not just ROIC. It's the ability to grow the pie over time.

Because a high ROIC that can't grow, I then have a high capital allocation risk. If you have a business that has a high return on capital and just spits off cash, well either they're going to pay it to me as a dividend which means I need to figure out what to do with it so why would I study this business? Or they're going to do something with it, in which case now I have capital allocation risk, which if the business has a great capital allocator well fine. That's good.

That's a whole separate set of questions that you need to answer. Whereas if the business has high returns on capital, the ability to reinvest capital... And the ability to reinvest capital usually comes from having small market share and being able to take market share in a market that is not growing. Or the ability to just keep your market share but grow along with the market that's growing very quickly.

eCommerce grows at 15%. Even if you're not Amazon you can do pretty well in eCommerce just because there's a structural shift in dollars from A to B. I don't want to say it doesn't take the capital allocation question off the table, but it means that you usually can just stay focused as an analyst on what are they doing internally?

How are they reinvesting those dollars? Because there are plenty of companies... Would Jeff Bezos be a great acquirer? I don't know. If you put Jeff Bezos in charge of a company that did not grow and said, "You now need to..."

Here's a great example. What if you exchanged Jeff Bezos and Mark Leonard? Could Jeff Bezos run Constellation? Because it's a very different business. I mean capital allocation is capital allocation. Capital allocation of buying businesses, nurturing independent teams that then send cash back up to the headquarters, which is what Constellation does, versus I would say more ruthless culture in Amazon that is just laser-focused on one thing, delight the consumer... Damn the torpedoes.

I'm not sure those are interchangeable. But that's all internal reinvestment right? Deciding between project A, project B internally, and some don't work, like the Kindle Fire. Some work beautifully like Amazon Prime. I guess I would argue that's a different skillset.

Patrick: [00:41:02] This reminds me of a fun question, kind of a version of which I asked Will. The question is, if you could take any one business, so you get to choose your business today, but you have to swap out its current leadership, its current CEO, and put someone else in, what company would you choose and who would you swap in?

Pat: [00:41:20] Oh, man. That's really fun. Let's see, and it would be... It's more fun if you would go kind of off the run.

Patrick: [00:41:43] Yeah. No rules other than that.

Pat: [00:41:43] No. No. I know. That's why you want to go off the run. I'm trying to think of mismanaged businesses. There's Yahoo! but honestly it was so horrible by the time she (Marissa Mayer) took over that I'm not sure you could have plunked anybody else there and they would have done much of a better job honestly.

Man. It's just funny because we tend to throw these things out very quickly in our process, ones where you know you can't change management. We don't do activism, and so we just keep on going. Ooh, Blackbaud. There's a good one for you.

Blackbaud dominates the market for fundraising software. I you're a YMCA, or you're the city opera, or whatever. Current management A, is vastly over paid, and B has made some absolutely stupendously bad capital allocation decisions we think in terms of getting outside their moat. I think if you just...

Because it's a very sticky product. I mean you think about most nonprofits, they don't have a giant IT department. So once the software comes in you're just going to keep using it. It's called Raiser's Edge. R-A-I-S-E-R. The customer work we indicated that the platform was almost MS-DOS like for a long time, very low reinvestment. But the customers, this is all they had. Salesforce didn't really work for them. I mean, boy you could have maximized that.

Patrick: [00:42:56] Who would you pick to do it?

Pat: [00:42:58] Oh, boy. Who would I pick to do it? I'm not sure Mark Leonard would take the job. He's got too good of a job.

Patrick: [00:43:03] That's all right. You've got the power of God in this question.

Pat: [00:43:05] Look, I think for any VMS company that's under-managed you just drop Mark Leonard in. I mean he's got, I would say, a more nuanced understanding of what drives a good VMS business than anyone I've ever met. I had the great good fortune to meet him one on one at their offices one time.

Then I saw him speak to a class of Columbia students about a year and a half ago. Those were both master classes. There's some people you meet with in any industry, I mean for Mark Leonard it's VMS software, who you're getting chapter and verse. There's one, the CEO of one company we own currently. I still remember he visited us not long ago and he talked about as a salesperson at a company, and in this case it was software, it could be anything.

He said, "Your goal as a salesperson is to get the hall pass. Your goal is to show up and be able to walk around that building and talk to anybody you want. You want to be so familiar with the security guard, with the secretary, with everybody. You want the hall pass." I was like, "That's a great way to think about a sales culture."

I mean that's a great way to think about it when you're talking to companies and assessing whether they can sell well or not, how are people getting out on the street? Do they have a hall pass? It was just kind of a cool way of framing it.

Patrick: [00:44:17] Very cool. Coming back to your process.

Pat: [00:44:18] Yeah, anyway. Fish in ponds where the fish are. Mechanically what we'll typically do is we'll do a quick idea, which is just sort of a paragraph or so and, what's the moat? What's the opportunity? What's the idea here? Then I'll decide to do what's called a first pass memo on, which is basically two thirds to a day's worth of work, which is not answering the question of should we invest. It's answering could we invest?

It's, does this business have the characteristics we want in terms of reinvestment potential, structural advantages, a management that does not misallocate capital? That's honestly, I think, one of the hardest jobs of any portfolio manager is to allocate the 100, or 200, or whatever hours of research time you have per week.

Because that's a finite asset. We do probably seven to eight, nah, seven or so, six to seven first pass memos per month. Maybe one or two of those will get approved for further work. The approval process for us is very team-based. We're not a top-down shop.

I have a very diverse team. They come from different backgrounds which I think is a topic for another day, which I think is very valuable in our profession because we all know that any successful investment shop has a process and philosophy, a way of looking at the world. But that can lead you into groupthink.

That can lead you into, "Our way is the best way." You can kind of reject outside input. I think one way to mitigate the risk of that is building diversity. Mauboussin talks about this a lot, building diversity into your organization, diverse viewpoints. Those decisions on what to move forward with are very, very much team-based.

I mean you have to have somebody that makes the final say obviously. You can't invest by committee. I get the input from everybody, and then we go to what's called an in-depth stage where we're going to do customer calls. We're going to go to trade shows, get to know the industry backward and forward. Then the key we found that really adds value to our process, and everyone can cook the pie differently, is to give ourselves multiple opportunities to say no.

Because the most finite asset is not capital. It's time. You always want to be maximizing the ROI, and the research time.

Because I think if you just send an analyst off and say, "Go get to know this industry, and this company, and give me a 20 page in-depth memo in X amount of weeks," if in three days they find out something that kills the idea they're going to waste the next month. But if the boss said, "Go write that memo," they're not going to come back and say, "Boss, this is kind of a really bad idea because I just found out X, Y, and Z."

But you have to give people that opportunity in your process. You have to have those windows where you say, "What do we know now that's different than what we knew a week ago, and should we keep going or not?" Because the most finite asset is not capital. It's time. You always want to be maximizing the ROI, and the research time.

Patrick: [00:46:53] Unfortunately the flight schedules at LaGuardia conspire to shorten our window a bit today so we're a bit short on time, but for our last question, one of my favorites that I ask everybody, is for the kindest thing that anyone's ever done for you?

Pat: [00:47:07] The kindest thing? That's a really cool one. The kindest thing anyone has ever done for me... There's a professor names Scott Hendricks at Kansas State University. He's used one of my books in their undergrad investing classes for a while. When I host the K State students when they come to Chicago every year or so, he was actually one of the very first undergrad programs to go see Buffett, those Sunday things. He actually was one of the progenitors of that.

To this day I think, I'm pretty sure that K State goes every year. Even the top MBA programs are on an every two-year rotation. But I got a call one Wednesday and he said, "We're going Saturday. You want to be an adjunct professor a day?" Because technically only professors get to go. You have to be faculty to go.

He figured he'd kind of shoehorn me in because I'd written the books or whatever. I was like, "Do you want my right leg or my left leg? Yeah, of course." He didn't have to do that. He got no benefit out of doing that. I couldn't ask any questions. The students got to ask the questions, but yeah. It was like, here's a chance to hang out with the man for two hours. He got no benefit out of that.

But I'm glad you asked that question because I think it's actually really important in our profession, because the goal I think for all investors should be that the next generation is smarter than you are. I think that the importance of making time for the ones who are coming up... I think about that a lot actually because I came into this profession a weird way through Morningstar, and I never really had a mentor.

I actually didn't have the moxie to reach out to people I really respected, and so I respect the hell out of these 25 year olds who just email me blindly. Some of the emails are misspelled. That's a problem, but I think kindness matters, and the importance of making time for the ones who are going to be hopefully replacing you and better than you in 20 years, 30 years time is important.

Patrick: [00:49:05] Great place to end. This has been really fun. I hope we get to do a round two with a little bit more time next time.

Pat: [00:49:09] Yep.

Patrick: [00:49:09] Thank you very much.

Pat: [00:49:10] No, thank you.

Created by @pbrychcy1