My thoughts:
https://actiquant.com/2019/05/07/warren-buffett-says-book-value-lost-its-relevance-i-object/
Buffett was talking about the metric he’d touted quite often, which was change in book value (as a sign of growth), not just book value (as a sign of value). That’s where he switched tracks. I don’t think he’ll abandon ROE anytime soon.
I have a very different view of book value than you do, Marc. I’ve written about it here: https://seekingalpha.com/article/4143240-equity-versus-debt-showdown. The most pertinent sentence in that overly long article is probably this one: “Simply put, the relationship of the market value and the book value of equity tells you almost nothing of any use because it ignores almost everything of importance about the company and treats the entirety of a company’s debt as a problem rather than a possible solution.”
Another article that’s worth reading (probably worth more than mine) is Travis Fairchild’s article on book value here: https://www.osam.com/Commentary/negative-equity-veiled-value-and-the-erosion-of-price-to-book. I love his opening paragraph: “The price-to-book ratio has a problem. More and more U.S. companies report negative book value, the result of accounting rules and structural changes in the market. This creates broad confusion and problems for the famous value factor, and indexes or strategies which rely on it as a measure of cheapness. Negative equity companies are often written off as distressed, but after reporting negative equity, most of them survive for years and have, as a group, outperformed the market 57% of the time. There are currently 118 companies in the U.S. market with negative equity. These companies have had negative equity for an average of over three and a half years, and 25% have had negative equity for over five years. One example is Domino’s Pizza which has had negative equity since their 2004 IPO but has outperformed the S&P 500 by a cumulative 1,442%. McDonalds, H&R Block, Yum Brands, HP, Motorola, Denny’s, AutoZone, and Wayfair are also on the list of those with negative book value.”
The basic problem with book value is that by subtracting a company’s liabilities from its assets, it treats those liabilities as detracting from a company’s prospects (which is the basis of its valuation). Nothing could be further from the truth. Both debt and non-debt liabilities can power a company’s growth and are often necessary to its survival. Book value measures a subset of a company’s assets, but it’s the more expensive subset, as the cost of equity is by definition higher than the cost of debt. And investing in companies with a lot of debt on their books can provide an investor with cheap leverage.
Your article says nothing about debt or liabilities. Why not? How can book value be considered apart from those things? After all, you can’t actually calculate book value without subtracting liabilities from assets, so shouldn’t they be a big part of the discussion?
As for ROE, the lower the book value, the higher the ROE. But the lower the book value, the more “expensive” the stock is. Does this make any sense? P/B should be tossed in the trash, and the sooner, the better.
(Well, except for financial and real estate companies, for which debt and liabilities play an entirely different function. For those companies, price to tangible book value is indeed still a very important measure of value.)
Of course I can talk about book value without talking about debt and other liabilities, just as I can talk about football without talking about baseball. IOW, I can talk about book value instead of talking about capital structure, just as I can talk about football without addressing the broader topic of team sports.
As to capital structure in general, you can’t talk about cost of capital without also talking about risk since the two are inextricably intertwined. Yes, everybody has always understood that a company could juice its returns by loading up on debt. But the risk of debt-induced corporate failure is extremely real, as I’ve seen way too often. We’ve had a good-long debt-friendly cycle, but fortunately for economy, most asset buyers know bette than to be prisoners of the moment (the most dramatic exception having been in realestate leading up to 2008).
As to the presence of negative equity companies, then don’t rely on PB for them. That’s why we have other metrics.We have negative earnings companies but those do not invalidate p/e. We have negative cash flow companies but that does not invalidate cash-flow-based measures. Etc., etc., etc. It’s the responsibility of the strategist to understand what can and cannot be gleaned from each statistic and act appropriately. To me, as I’ve stated often here, this means thinking in terms of a portfolio of metrics to diversify away the risk of oddities that impact particular metrics here and there.
If you want to toss BV in the trash, you can, of course, do so. But that means you also toss ROE, ROI and ROA in the trash. Again, you can do that. My suggestion, though, is to go a lot slower when it comes to tossing things in the trash. Time and changing circumstances have an incredible way of teaming up to embarrass those who are in a rush to abandon the body of knowledge acquired over many tears by many smart people.
Marc or Yuval,
Would counting debt be double-counting sometimes?
I mean if debt is used to buy an asset doesn’t that asset get counted in in the “book value?” Then if you count the debt that bought that asset would it be double counting?
I understand that things like Enterprise Value can be better than book value so I may agree with Yuval on the larger point. Enterprise Value does include debt and I understand why this is the case. And it looks like Marc uses enterprise value himself at times. Also I do not use book value so I am not disagreeing with Yuval larger point.
This is just a question—probably regarding accounting about which I know little. I do know my accountant must be doing something wrong, I just can’t figure out what it is—which is the limit of my understanding.
I must be missing something about book value and why one would add debt to that. Use a different factor, I do understand some of those arguments.
-Jim
Marc,
Just to clarify, I’m not advocating throwing book value in the trash, just the price-to-book-value ratio, while making an exception for financial and real estate firms. And I’m not advocating throwing ROE or other ratios based on shareholder’s equity in the trash either. I also don’t agree with your football/baseball analogy because you can calculate a football score without looking at baseball scores, but you can’t calculate book value without subtracting debt. For me, talking about book value without talking about debt is like talking about football without talking about field goals. (Or, to take up your suggestion, talking about capital structure without talking about risk.) I completely agree with you that debt is risky. Lastly, the difference between negative equity firms and negative earnings or negative cash flow firms is that negative equity firms tend to outperform positive equity firms, while you couldn’t say the same for negative earnings or negative cash flow firms.
Jim,
When you take on new debt and purchase an asset with it, the debt goes into the liabilities column and the asset goes into the assets column. Book value is obtained by subtracting liabilities from assets, so it remains unchanged. That’s assuming, of course, that you don’t depreciate the asset you’ve just purchased.
- Yuval
Thank you!
Yuval, I am having a hard time really believing that, over a long time frame (not just what is covered in the P123 database period) that “negative equity firms tend to outperform positive equity firms”. Is there data on this, say from the early 1900s to today? or even the 1950s? I am skeptical…but always willing to see something new.
See footnote 1 on this page: https://www.osam.com/Commentary/negative-equity-veiled-value-and-the-erosion-of-price-to-book -
“Negative book value companies have outperformed in 57% of rolling three-year periods from 1993 to 2017.”
Also see this paper from 2010:
https://aaapubs.org/doi/abs/10.2308/acch-50094?journalCode=acch
I can’t access it myself, but the abstract is telling.
OK. I will read the articles. Thank you.
Synopsis from the American Accounting Association posting:
“We document a phenomenon that, along with the increasing trend of negative earnings, the frequency and the magnitude of negative book value of equity have also grown substantially over time. Although negative-book-value firms are commonly perceived as financially distressed, we find that the majority of these firms survive for a long time, and that many continue to report negative book value for several years. Over the most recent decade of our 30-year test period, 1976–2005, we find that based on per-dollar of assets, the market, on average, prices negative-book-value firms higher than positive-book-value firms. In addition, we discover that the correlation between market value and book value for negative-book-value firms is negative. Searching for explanations of these phenomena, we examine R&D expenditures of negative-book-value firms. Our results indicate that R&D, especially R&D accumulated over time, not only contributes to the increasing trend of negative-book-value incidences, but also plays an important role in the market’s valuation of firms that concurrently report negative earnings and negative book value.”
Synopsis from the OSAM blog:
“The price-to-book ratio has a problem. More and more U.S. companies report negative book value, the result of accounting rules and structural changes in the market. This creates broad confusion and problems for the famous value factor, and indexes or strategies which rely on it as a measure of cheapness. Negative equity companies are often written off as distressed, but after reporting negative equity, most of them survive for years and have, as a group, outperformed the market 57% of the time.1 There are currently 118 companies in the U.S. market with negative equity. These companies have had negative equity for an average of over three and a half years, and 25% have had negative equity for over five years. One example is Domino’s Pizza which has had negative equity since their 2004 IPO but has outperformed the S&P 500 by a cumulative 1,442%.2 McDonalds, H&R Block, Yum Brands, HP, Motorola, Denny’s, AutoZone, and Wayfair are also on the list of those with negative book value.”
Another thing to consider is that when a company buys back stock trading above p/B = 1, then the book will get reduced. Actually the company may be getting more valuable, if purchases were before intrinsic value, but P/B will increase and shareholders equity will decrease (can even go negative). In such scenario’s I agree with Yuval that P/B becomes meaningless over time.
This negative book value thing causes problems only to the "physics envy"crowd, not to real-world investors.
No datapoint is perfect. Every data point is subject to whacky things. That’s why we have a multitude of valuation metrics. They all tell different aspects of the story and the reason we look at more than one is the same reason we own more than one stock – to diversify against the risk of idiosyncratic problems with a single metric. That’s why the biggest thing to fear is the unwittingly mis-specified model.
Book value can be negative. Earnings can be negative. Cash Flow can be negative. Free cash flow can be negative. Sales can’t be negative (absent some rare and unusual accounting protocols that net things out against gross inflows before reporting on the sales line), but sales can be wildly distorted up or down. Growth can be negative.
Yet all are regularly used in valuation. And anyone with a computer and a desire to propound an eye-catching conclusion can concoct a study associating supposedly negative metrics with superior equity returns. Sometimes, they’re picking up risk-on attitudes on the part of the market. Other times, they’re bumping into stocks whose merits are better explained by metrics other than the one they think they’re debunking.
Maybe next week I’ll post the factor investing presentation I prepared for an AAII group in Pittsburgh next week. (Any p123 users in Pittsburgh? Come on over and say hello.)
BTW, as to buybacks, here’s the deal (sans propaganda).
If a company doesn’t think it has enough opportunities to profitably reinvest its profits (i.e by earning more than the cost of capital which presents a whole different hornet’s nest since cost of equity is incalculable except in grad school and there, you use whatever method will induce the professor to give you a good grade so in the real world, its just a number one pulls of of one’s you-know-what), it should give excess back to shareholders. They don’t do it as a matter of honor as MBAs. They do it because if they don’t, activist shareholders and/or raiders will start sniffing around. Excess cash on the balance sheet is like blood in the water. Both attract sharks.
Dividends are the traditional method by which excess profits are given to shareholders. But from the management perspective, share buybacks have a particular appeal.
One, the stock often bounces because commentators over a generation convinced many in the investment community that this is better than a dividend because its better to think in terms of tax on a capital gain that may or may not ever be realized in the future as opposed to tax on income (albeit at a favored rate) that you actually get cash in hand; so managements love to be loved and, hence, love buybacks.
Also, for management, buybacks are way better than dividends because this gives management the freedom to be completely random and erratic in what they’ll give and when they’ll give it as opposed to dividends, which tend be regular and, hence, a sort-of commitment. Put another way, buybacks are a guy thing; a series of one-night stands as opposed to the dreaded commitment!
(There is another consideration nowadays. With debt so cheap, it’s tempting to leverage up the balance sheet by borrowing to repurchase shares; like an LBO without actually doing an LBO – but with all the dangers if something goes wrong down the road.)
So do buybacks above or below book value enhance or detract from intrinsic value? Propaganda aside: none of the above. First off, nobody knows what intrinsic value is because its incalculable by humans who can’t see into the future and, hence can’t precisely value the biggest component of intrinsic value.
Beyond that, it’s like the question of dilution or non-dilution of acquisitions. One can always concoct a day-of-transaction answer (and we see the answer WB is concerned about people honing in at BRK regarding its buybacks). Ultimately, though, it takes years to see if a buyback (or an acquisition) was a wise decision, and the answer is based on what happens going forward. This is why the better analysts out there try to project forward, rather than getting obsessed with the day one number crunching.
Here’s the thing. What we’re all talking about is accounting book value. This is an accrual accounting measure, not an economic one. Financial managers have little to no incentive to make it reflect its economic value.
The whole corpus of literature on value investing are variations on a theme of adjusting accrual metrics to economic ones.
P/b doesn’t work as well as it used to because of information/computing-power diffusion. Top line metrics are increasingly losing their predictive powers. However, even if true, something might still be gained from thorough analyses of financial disclosures (even if that something may be initially hidden).
I have a question for Marc. Am I correct in positing the following? Or is this incorrect?
In addition to what you just wrote, it seems to me that there are two other good reasons for a company to buy back its shares. First, it reduces its cost of capital by reducing its equity (and, incidentally therefore, its book value). Second, it increases ownership and control and is the first step in going private.
Private companies go public in order to raise money. But going public means that the company holds no shares in itself, and control of the company is delegated to shareholders (some of whom work for the company, of course). Once the company has a lot of money to spare, it can then go private if it wants to by buying back all the shares it has issued, and buying back some of its shares can be a first step along that road. Assuming it’s not buying back shares from people who work for (and run) the company but is instead buying them from outside investors, it is increasing its own control and ownership.
- Yuval
Yes and no to the first question. In a textbook sense, they could be said to be trying to reduce overall cost of capital by swapping more-expensive equity for less-expensive debt. But remember that outside the world of textbooks, nobody really knows what cost of equity is. What they’re really doing is recognizing that the cost of debt today is dirt cheap, and that such cheap costs, once they go away, may not be seen again for a generation. So they tale advantage of it to stockpile while they can. Of course at some point, they need to put the newly-acquired cash to work, but with the cost of new debt so low, the hurdle rates for reinvestment ideas is likewise pretty low. (Is there a chance companies may be making the kinds of investments today they’ll be regretting and writing off five years from now? Yes. This could be a big topic of conversation in the mid 2020s.) When the cost of anything is cheap, the quality of buyers declines. What’s really going on is a sort of interest-rate arbitrage; not cost of debt vs costs of other kinds of capital, but cost of debt today vs cost of debt down the road (like an airline locking inlow fuel costs).
As to the second question, it’s definitely possible. But to build that into an investment theme, we’d have to support it with other notions consistent with an eventual transition to private ownership. Many companies today buy back stock without any thought at all to going private.
With regards to buybacks… buying back stock is a way of enabling the share price to move higher without having an excessive market cap, important for companies like Apple that have difficulty expanding their market capitalization beyond what it already is. So I would say that share buyback may be in the self-interests of company insiders especially if they hold stock options that are essentially leveraged instruments that depend on maintaining a certain minimum share price. This is my cynical side expressing itself 
Steve - that is an excellent point!
Jim O’Shaughnessy just gave a long interview to Validea in which he advocates throwing price-to-book in the trash. The link is http://blog.validea.com/five-questions-factor-investing-with-jim-oshaughnessy/.
“Ror the fourth edition of the book, I had access to the CRSP data (that’s the Center for Research and Security Prices at the University of Chicago) which goes back to the late 1920s. And when we ran that data on price to book what we found, if you go to my chapter on price to book in the current edition of the book, you’ll see it, was that price to book between late 1920 and 1963 where we actually began with our earlier versions of What Works didn’t work at all. . . . So you know, you look at things, other things and you’ll see, if you say, look it all rolling 10 year periods, you’re getting very similar results, at least directionally, right? With price to book that was not the case. In fact, what you saw in that period between the late 1920s and the early 1960s was an inverse relationship with the stocks with the highest price to books doing much better than the stocks with the lowest price to books. . . . We also have a forthcoming paper by our research associate who is known by the pseudonym Jesse Livermore [this is the genius behind the Philosophical Economics blog], in which we believe you’re going to see the final nail go into that price to book coffin.”
Good link!
I probably operate similar to Jim O’Shaughnessy (or try to)—using rank performance as he does in his books. So for practical purpose, I may tend to agree.
But didn’t Piotroski used price-to-book in his original paper?
Did price-to-book work when used with a Piotroski score? Even if it did work I am not sure it still does.
Can Marc make it work, still, with his buy/sell rules?
More generally, interaction are important in multiple regressions. Are interactions important for rank performance too?
Marc certainly thinks interactions are important in the buy/sell rules and perhaps within the ranking system itself. I think interactions are important in both.
With interactions in mind, how much work needs to be done to be able to say a factor is NEVER useful?
-Jim
I want to amend my earlier statements: I don’t think the price-to-book ratio should be just thrown in the trash because it’s still quite useful for evaluating companies in the financial sector.