This is Part II of my interview with Tobias Carlisle of the Acquirer’s funds, which just released its first ETF, the Acquirer’s ETF (ZIG) earlier this year. In Part I, we talked about some of the arguments against value investing and how they can be refuted. You can read that here.
In Part II, we talk more about the intricacies of building a value portfolio. As a reminder, this interview was taken from a phone call transcript, so please forgive any grammatical errors.
Jack: I wanted to ask you about sector constraints because you take a nonstandard view among quant investors by allowing your portfolio to take overweight positions in certain sectors when the fundamentals dictate it. Many quants tend to view sector overweights as uncompensated risk, but you argue that there is a benefit in moving a greater percentage of your portfolio into areas of the market that are cheap. I wonder how you look at that balance between seeking greater returns by concentrating in specific sectors vs. the tracking error it can generate?
Toby: I think it’s a matter of taste how you construct a portfolio. I think that if your objective is to not have a great deal of tracking error, which is better for raising assets and better for short term performance, then you should constrain by sector or industry. I think if your objective is long-term, outperformance, then you need to be prepared to accept a lot of tracking error. I’m not trying to build in tracking error. That’s a side effect of what I’m trying to do, which is I’m simply trying to buy what I think are the cheapest companies in the market at any time. And what tends to happen is industries get cheap together because there are things that are external to the industry that treat all of the companies inside it as being no longer worth what they were worth a few years before. Financials are a good example of that. I think that the 2007-2009 crash is vivid in many investor’s minds. And so financials at the moment are the cheapest sector and they look to be unusually cheap relative to history. I think that means that you should take a bigger bet on financials than you would otherwise. And if you’re sector constrained, you’re not going to be able to take full advantage of that, whereas I am. But if I do that by not being sector constrained, it means that my performance in the short term is going to be tied quite closely to the fortunes of that industry. But I hope that over the longer term, over a three to five-year period, that will turn and I’ll be there to capture more of it than anybody else. So in short I think it’s a matter of taste for the construction of the portfolio. But mine is long-term outperformance and I don’t mind tracking error in the short term.
Jack: One of the more interesting debates going on in value investing right now is the one regarding the Price/Book ratio. It is an interesting case study in the balance between long-term data and the application of logic and human decision making in an investment process. On one hand, the academic research supporting Price/Book goes back to 1926 and it has produced significant excess return. On the other, in a world where intangible assets dominate tangible ones, it is easy to argue that Price/Book is no longer applicable. I know you don’t use Price/Book, but I was wondering what your thoughts are on how applicable it is today and how to you look at that balance between long-term data and the fact that things change that make the long-term data less applicable?
Toby: I gotta say, I don’t think that the
reason that price-to-book doesn’t work anymore is because there’s so many
intangible assets or there’s more intangible assets in the world. I think that
the reason that price-to-book underperforms is because the change in accounting
rules means that we haven’t been able to deal with companies like McDonald’s, which
have brought back so much stock that they now have negative equity and
price-to-book doesn’t properly classify that as being undervalued because it
doesn’t have a place for negative equity other than more debt than assets. So I
think we’ve got this kind of metaphysical problem and I think that to the
extent that price-to-book was ever a good way of measuring it, it is not as
good as it was in the past.
And I think a better way of measuring value is if you use flow measures, and the criticism of flow measures, like income or cash flow or whatever it might be, versus something like an asset measure like Price/Book is that they are much more volatile. So book is pretty stable from quarter to quarter and year to year. And so when the price in the market dips below book that might indicate that you have a value stock. And when it’s at a premium to that, it might indicate that you don’t. The flow measures I think will give a much noisier answer, but I think that they are still iterating towards a more correct answer. So I have a preference for flow measures and I think that where a company is generating a lot of cash flow and it’s cheap relative to that price, what that indicates is that it’s doing pretty well on a competitive basis and it’s still probably a pretty good bet regardless of where it is relative to its book value.
Jack: That makes sense. One of the interesting things I found when I interviewed Jim O’Shaughnessy, when he had gone back for one of the editions of his book and looked at Price/Book, he found that in the Great Depression it was basically a proxy for bankruptcy risk. So stocks with low Price/Books had very high bankruptcy risk. And so it didn’t work at all in that period either.
Toby: There’s also some research, the Loughran research on EV/EBITDA. He said that the problem with price to book value is for one thing it tends to buy, and you’ll notice if you look at low Price/Book companies, what you think is that you are buying something that is cheap relative to the assets. What you’re really buying is something with enormous amount of debt and a little residue of assets. And then you’re buying it at a big discount to that little residue. And so I think what happens is when that works, it works really well. That’s a company that’s paid off its debt. And when they don’t, it’s because the debt has overwhelmed the company. In that Loughran research they say that Price/Book never really worked in any but the smallest 4%-6% of the market.
And then you’ve got other problems with measuring in that part of the market because you’ve got wide bid/ask spreads. It’s not clear that you can trade at any of those prices, so the back test may not match reality, whereas something like EV/EBITDA, which they look at in that paper, they said that scales really well. Even in the S&P 500, it’s more likely that the cheapest 10% will outperform the most expensive 10%, so it scales through the whole of the universe. So that’s why I prefer flow measures to book value measures for that reason.
Jack: That makes complete sense. Another interesting debate that many value practitioners have is the debate over which metric to use or whether to just use a bunch of them in a value composite. On one hand, if you think you have identified the best metric to use, using one can make the most sense, but on the other, if you think the best metric is difficult to identify in advance, then using a composite can be the most sensible approach. I know you have found an individual metric that you think is superior with the Acquirer’s Multiple, but others like Jim O’Shaughnessy have migrated to a composite based approach over time. I was wondering your thoughts are on using an individual value metric vs. a value composite?
Toby: Well, the way that I’m using the Acquirer’s Multiple, I use that for the industrial companies. So I don’t use that for Financials. For Financials, I use a much more traditional book value measure. And then I want them to be doing all the other things that I want other cheap stocks to be doing like buying back stock and generating free cash flow. Negative financing income basically: paying down debt or buying back stock or paying a dividend. And I think that indicates some sort of financial health measure.
So I use a slightly different beginning
way of measuring financials versus industrials. But then I think that the next
step is the same for both and you want to make sure that the cash flows do
match the reported income. So that’s a second measure that I’m looking at to
make sure that the cash flows match the accounting measures of income. And I
think Jim’s point is a good one and Jim’s experience is one that I have learned
from because the first edition of What Works on Wall Street had price to sales,
and I think that the argument there was compelling because it was undiluted by
management’s changes to the income statement, and the bottom line is a very
noisy, messy figure because it’s been manipulated. The top line is harder to
But it also misses the fact that some companies have a higher gross profit margin than others. And companies with higher margins should be better targets then companies that have a lower gross margin, even if they are equivalent on a price to sales. So I think that it’s necessary to have other measures in there that confirm what the main measure is saying. And so that’s what I have done. I want to make sure the cash flows match. I want to make sure they’re doing things that a good management of a cheap company would do, which is buying back stock and paying down debt. And I think that when those things come together, I think that basically what I’ve built is a composite, even though the composite that I have built is the way I think of value investor would go through an evaluation.
Jack: So for non-financials, you are using the Acquirer’s Multiple as a starting point and then you are going through and doing a variety of checks to make sure the other data backs up that initial opinion?
Toby: Yes. Which goes to answer one of
your other questions. I’ve been running this model for about a decade. And over
the course of that decade, I’ve had this iterative process where it spits out a
bad name, which I don’t like for whatever reason, and then I go back to my
process. And when I said I don’t like the name, I don’t like the name because I
don’t think it’s undervalued, not because I care about how it’s gonna perform.
So what I have been trying to build is a model that acts like a value investor rather than a model that acts like a Chicago School of Business quant investor. So the model that I have built over 10 years is sort of going back and forth, looking at the outcome and going back and adding another part of the process. So now I’m comfortable that basically the model behaves the way that if I was discretionary, I would still be doing the same thing.
Jack: Yeah. You know that’s interesting. That’s kind of the same way we tried to do it when we built Validea. We wanted to use quant models to analyze stocks as closely as we could to the way a human would look at it, but without the human emotion and biases.
Toby: I think that’s the best approach. It makes it easier for the manager to stick with the process as well because the outputs you’re getting match the way that you think about investing. The first time you run a quant model, the output is madness. There’s some stuff in there that you just wouldn’t buy. It takes a long time to get from that point to where you’re comfortable with the output of the model.
Jack: Thank you again for taking the time to talk to us today. If investors want to find out more about you and Acquirer’s Funds, where are the best places to go?
My email is firstname.lastname@example.org and I’m on Twitter @greenbackd.