The Exchanges

Q So you don't know if the underlying economics of the businesses do better if the CEO is better. Ok. When you go through this work, okay, now we're constructing the portfolio. How do you think about what the portfolios should look like?

A So I think that you basically want to have the best on this metric. So you want to be the cheapest, highest free cash flow yield, and you also in Europe and the US want highest quality. So ideally, if you can have a higher credit quality as measured by lower debt to assets, lower debt to EBITDA, higher free cash flow to debt, all those things are going to be beneficial. So We're very focused, first and foremost, on making sure the portfolio is maximizing those quantitative criteria. And then second, within that, you don't want correlated risk, so you want to diversify your bets, because you could be really wrong about the auto cycle, and you better hope that you are right about that packaged food company in London, because you've got to have those big winners that make up for what are inevitably going to be big losers in a deep value levered portfolio. So we were looking for dispersion, An important thing about having very high dispersion is having diversification. So we're trying to, let's say something is ranked number 70th, but it's the only technology company in our top hundred. We're going to put that in the portfolio, even though it's a little lower ranked, maybe it has a little bit worse score than something else because it's going to add diversification and it's an uncorrelated bet to the rest of our portfolio. So I think that's how we think about it. And then I think t…

AI assessment note: “We're very focused, first and foremost, on making sure the portfolio is maximizing those quantitative criteria.”

Q to Strangers, and he talks about this concept of default to truth, and how some people, but in the minority, don't do that, and effectively means they're natural cynics. That certainly comes through in both your work, and particularly having seen you speak on occasion, Take me back to, I don't know, your childhood or whatever. Where did this come from, this kind of innate skepticism and cynicism in you?

A I hope it doesn't veer into cynicism. I think Gladwell points out that that would be corrosive to social trust, which obviously I think can be bad, but I grew up in a family of four kids, and my dad is an antitrust lawyer, and we'd have family dinner every night, and my dad would put us each on the witness stand. How did that science test go? Oh, I gotta be, why do you gotta be? I didn't study. Why didn't you study hard enough? Do you think if you studied more, you would have done better? And I think this constant Socratic method, more than a cynical or skeptical perspective, it's just asking the second and third order questions and asking for proof. You said you did okay on that. Show me how you actually did. And I think that's a legal mindset. And I think my father's sort of legal mind shaped me so much, both in sort of saying, hey, trust, but verify. And also I think something that is maybe unusual is in applying the Concepts and principles that I read to what I do. That there should actually be a flow through from reading, logic, coming up with a, a rule or a guide for behavior, and then acting in accordance with that. And I find that that's actually relatively rare in investing. I think a lot of people are very story driven, and they're very idiosyncratic driven, so they're bottoms up, quote unquote. And they don't say, well, I'm gonna do X, but would X if it were applied …

AI assessment note: “my dad is an antitrust lawyer, and we'd have family dinner every night”

Q Yeah, interesting. So at what point in time in your tenure at Bain did you decide that the empirical data you were seeing almost required you to do something else?

A It was right after doing that, after doing that study, and I think my first instinct was to say, well, why doesn't Bain do this? And I think the reason that Bain and most private equity firms can't do it is competition. Okay, great. You think it's a good idea to buy things at six times EBITDA. Okay, where are you gonna buy them? Not in the U.S., right? Too much competition. Now, I think in Europe or in Asia, I think there are still cheap LBO opportunities, right? So a firm like Bain or Blackstone, they're still seeing opportunities to do really attractive buyouts ex-U.S. But within the U.S., you can't. The opportunity's been gone, and now I think private markets are much more expensive than public markets, and you can see this with the venture-backed stuff like WeWork. Private markets say it's worth seventy-five billion. Public markets say it's worthless. You can see it in the deal multiples, where if S&P 500 is at 12 to 13 times EBITDA, private equity claims it's at 12 to 13 times EBITDA, but, you know, those EBITDA numbers are not accounting for all the proforma Adjustments, which people say is like, 25% of EBITDA, so you do that math, and they're trading, basically buying micro caps at 16 times EBITDA, when the S&P 500 is trading at 12 or 13. That's just nuts. It ain't gonna work. It defies logic. And it wasn't as bad then as it was now, but it was roughly that bad. And so w…

AI assessment note: “It was right after doing that, after doing that study”

Q solid, economic backdrop. Should that economic environment change in any way, how do you think this plays out? Because on the one hand, right, The softening economics would tell you businesses will fall, you'll have defaults. On the other hand, there's so much money on the sidelines in private equity and private credit to boost this. So how do you think about what that looks like in the out years?

A It's really tough to say. I'll give you two sort of extreme scenarios. On the one hand, I would call a normal business environment, which is there's a recession. What happens in a recession? Things that are over levered go bankrupt. And so you see a default wave. And so much of private equity today of LBOs is single B or triple C type quality, that I would say 25 to 30% of LBOs would go bankrupt in a normal business cycle, normal recession with normal levels of delinquency. So that's sort of normal scenario. But it seems like in this world we live in, nothing is ever normal, right? Especially the post-financial crisis, it seems like we live in an upside-down world. So what's the sort of upside-down world possibility? And I'd say it's roughly looks like what happened with energy private equity. Energy private equity is fascinating, because 2013, 14, and 15 vintages of energy private equity funds, when oil prices drop 70%, they've started to rebound a little bit, a huge drop in oil prices, S&P small cap energy down 70 or 80%, peak to trough, 90 plus percent of 2013, 14, and 15 energy private equity funds are marked above one, ok? So you look at that, and you say, that may be the other possibility, which is that The chickens never seem to come home to roost. There's this extend and pretend zombie type behavior where the money that you put into private equity never blows up, but yo…

AI assessment note: “I'll give you two sort of extreme scenarios. On the one hand...”

Q How do you think about incorporating management teams?

A I did this big study on that because I got on that question a lot. So we basically said, okay, what we want to look at is CEO quality. So management teams, let's just narrow it to the CEO. And we can think about quality in two ways. And you'll detect a hint of sarcasm to this, but we'll define the first way as do they have a great pedigree? Okay. So they went to a top business school or any business school. They went to a top Ivy League institution. Second is their own track record. So they were either CEO of another company and then got hired to run this company. So we can look at their track record of the prior company and then this company. And then we can also look at just, hey, how do they do the last three years? Does that predict the next three years at the same company for people that have longer tenures? And what we found were all these tests had no statistical significance. Turns out that going to business school empirically does not mean that your stock price does better. Turns out working at an investment bank or top tier consulting firm has no impact on the stock price of the company that you run. Turns out that your track record at a prior company has no impact. Also turns out that your tenure at that company itself has no predictive power. So every one of these tests failed. And so what I look at that and say is, there might well be great CEOs. Of course there ar…

AI assessment note: “what we found were all these tests had no statistical significance.”

Q So when you looked at the empirical data, there's a key signal you saw in price. The private equity firms will talk about improved operations. What did you see in the data in terms of the ability to improve the operations, whether it's top line or bottom line?

A Yeah, so I think top line, it's the same thing. You can't predict revenue growth, and it's really hard to control it. Wish we could, and if it were possible, there'd be a class at Harvard Business School, how to drive revenue growth, okay? It's just impossible and unpredictable, and maybe you get things right, maybe you get things wrong, but it's really hard to predict revenue growth. Unless, I'd say, the one exception to that is this sort of compounding Businesses where you can see a clear return on investment. You know, okay, I increased spending by a hundred million. That should lead to an incremental. That's a logical way. I think you can predict revenue, but without investments, sort of same store sales type growth, really, really hard to predict. Bottom line, I think private equity is better. I think the empirical evidence would show that private equity firms are good at cost cutting. They are good at cutting the fat. They are good at streamlining operations, and we saw that a lot in the data. However, These days, if you're buying a company, the majority of private equity acquisitions are from other private equity firms. So now you have to make a case that your private equity firm is better at cost cutting than the previous private equity firm, or those idiots that sold it to you didn't. In contrast, I think a lot of what you see is that to a man with a hammer, everything…

AI assessment note: “the empirical evidence would show that private equity firms are good at cost cutting”

Q Some of this research, as you're learning it, you're putting it out in weekly emails. They're always sort of interesting and data driven. How does that integrate into your process?

A So I think it's in a number of ways. One is that We have sort of a written culture at Verdad. We want to write up our results and share them internally, and then we want to share them with our investors. And you could say, well, why would you share your research with others? Your stuff's just going to get arbed away. And we think, actually, it's not going to get arbed away. It's not going to get arbed away for a few reasons. One, stuff we're trading is so illiquid that even if the guys at AQR or DFA read it, what are they going to do? How are you going to move a ten billion dollar portfolio? I mean, it's just not a concern. It's not going to have any effect on Arbitrage. And second, we probably believe that the deepest arbitrage in doing deep value, in doing illiquid deep value, is basically taking on volatility pain. That buying the stuff is wildly volatile. If you add up the years when the market is down, and then the years that small value underperforms the rest of the market, that's 50% of years. The easiest way to win in small value is just last enough years that you're actually hitting that 50% Of years that are winners. And so many people just get flushed out of small value because they don't have a long enough time horizon. So what we're trying to do is to our investors is to say, look, you need to be making your investment decisions on very long-term empirical data. An…

AI assessment note: “We want to write up our results and share them internally, and then we want to share them with our investors.”

Q And then how do you trade off the concept of transactions cost, right? In theory, these things are bouncing around, and if they bounce up and down and up and down, you may be able to capture a little bit of buy low, sell high.

A Yeah. I think it's patience, and I think you just sit on the bid, and you just say, hey, we're, these are illiquid stocks. We're providing a little liquidity as a buyer, but any excessive provision of liquidity is going to move the stock price in our names. Even if you try to buy, in our stocks, you try to buy a million dollars worth of them, you're going to move the price. Almost. So buy 50,000 a day for two weeks. I mean, it's painful. But our logic, again, is that if we are doing things that other people are unwilling to do or unable to do, unwilling because it's tedious and boring, unable because of capacity issues, and we can do that, that should be where alpha is. I mean, that should be what you're compensated for. There's a fundamental logic. If you're buying really cheap, really out of favor, really illiquid things, you're seeing multiple ways to win. That's the logic.

AI assessment note: “I think it's patience, and I think you just sit on the bid”

Q So when you put together the desire for, say, industry diversification, these liquidity issues, how do you weight positions in the portfolio?

A I think that this is one of those things which I don't think a computer that we've designed can answer yet. I mean, I think there's an art to it, and I think there are a few things that play into it. One of the things that plays into it, obviously, is how highly ranked the stock is. How good of an example is it according to your model? The next thing is liquidity, and then I'd say the third thing is how well you know the stock. Let me say, give an example, something that jumps from being ranked 500 in your ranking to number one in your ranking system. You say, oh, I've never seen this company before, and it looks really interesting, and it's an added diversification because we've never seen a company in this industry before. And you say, well, how did it end up jumping from 500 to number one? Well, there's only one way, which is the stock price went down, 80 or 90%, right? And so then you step back and you say, why did the stock price go down? And what does that sharp drop in stock price tell me is likely to happen next quarter? And the answer is something really bad and not pretty, and you might not want to find out. And so what we also then think is, as the portfolio's in motion, right, you're constantly rebalancing, you should be really careful about new things. Really careful. Because those new things, the price movement that made them interesting to you, is also signaling …

AI assessment note: “I think there's an art to it, and I think there are a few things”

Q And then the last lever is the financing environment. So the private company has this optionality to refinance their debt. They have too much debt. With all this money on the sidelines, they could put more equity in and extend the option on the ownership of the business. How do you think about that optionality in terms of the success historically of private equity investments?

A I think leverage is a double-edged sword. On the one hand, If you finance something with 90% debt to enterprise value, and you increase enterprise value, then levering at 90% was an amazingly good idea. On the other hand, Debt imposes certain handcuffs on a business, and not only is it most handcuffs, it increases bankruptcy risk. If anything goes wrong on the top line or the bottom line, you're gonna be in trouble. Now, the environment that we've been in, and it's been really interesting since the financial crisis, has been one of very, very benign default environment. There's been very few bankruptcies outside of energy and retail in a few troubled areas, very, very little bankruptcy, falling rates, and also in private equity world, the massive Influx of private credit money, and private credit is a really frightening development. So private credit, basically the banks, starting after the 98 and oh one recession, started to really get out of risky corporate lending, because they basically found it was unprofitable, and they always realized it was unprofitable at the worst times, right? The recession comes and all that stuff, you underwrote an eight percent yield, all of a sudden, so much of it defaults, you end up with a four percent return when you could have lent to double B bond at Five percent. And you're saying, oh my god, why did I go through all the headache of doing t…

AI assessment note: “I think leverage is a double-edged sword.”

Q firms moving to Ridgeline gain a decided advantage. That's why customers call it miraculous, game-changing, and an awakening. If that's not how you would describe your investment management tech, request a demo at ridgeline.ai. And now back to the show. So in that first machine learning model, what are the feature set of the model that you found is predictive of the companies more or less likely to go bankrupt?

A So it's all logical, and good machine learning is first of all going to find things that regressions find, and then it's going to fine tune. So if your machine learning is finding something that standard regression models didn't find, it's probably not a good model. The big things are obvious in the data set, So I'd say the first thing, the biggest predictors of debt pay down are, did they pay down debt last year? What's the free cash flow? What's the free cash flow relative to debt? I mean, these things are not totally surprising, but it's honing the probabilities and looking at things like, oh, gee, but if they did a massive impairment charge, that might be a really worrisome signal. It's basically all the nuances that the machine learning helps you identify. And then the second model, which looks at errors in our existing model, so we were really excited about this, because we got the first results back, and it said it could predict errors with 50% accuracy. So if you think 30% of our universe was tagged as an error, and it's predicted with 50% accuracy, that's a big lift. So we said, okay, great, like, we're geniuses. And then we ran a study of, well, what are the returns based on that Probability being wrong score. And we basically found it was a linear correlation. So the higher the risk you took of your model being wrong, the higher the returns were. So we basically just…

AI assessment note: “biggest predictors of debt pay down are, did they pay down debt last year? What's the free cash flow?”

Q So it sounds like as you go through your process at Verdad, you start with this, this filter of certain criteria of, you know, say low EBITDA multiple and leverage. And then you're trying to figure out which of those won't go bankrupt. How many companies does that leave you to then filter through?

A So we divide it regionally, North America, Europe, and Japan. So in each of them, you want to start with Basically the traditional value screens that should be relatively obvious to anyone who's ever read Fama in French or familiar with the academic literature, right? You want to essentially rank by multiple valuation metrics like EBITDA to EV, EBIT to EV, free cash flow yield, price to book. The things that sort of score well on those blended value factors are going to be the things that are most attractive. I think within that we specialize in the levered ones that are, have the high probability for deleveraging, which is very related to free cashflow yield. These are not dissonant concepts, but we're going to focus on the ones that are levered because when they get multiple expansion and they're levered, that's where you really make a ton of money. So we start with that universe and then we apply two machine learning algorithm tools that we developed. So the first looks at the probability of debt pay down and basically says, okay, we took 60 years of US company financials. We said based on year negative one, how likely was this company to pay down debt in year zero? And basically score every company on that and eliminate the ones that look like they're not going to pay down debt, which are essentially bankruptcy risks. The second thing we did is we took those models, we ran …

AI assessment note: “So we divide it regionally, North America, Europe, and Japan.”