A blog reader emailed me this question about why I appraise stocks using a pure enterprise value approach – as if debt and equity had the same “cost of capital” – instead of using a Weighted Average Cost of Capital (WACC) approach:
“…debt and equity have different costs. In businesses with a (large) amount of the capital provided by debt at low rates, this would distort the business value. In essence I am asking why do you not determine the value of the business using a WACC, similar to how Professor Greenwald proposes in Value Investing: From Graham to Buffett and Beyond. The Earnings Power Value model seems theoretically correct, but of course determining WACC is complicated and subject to changes in the future. Nevertheless, your approach of capitalizing MSC at 5% is basically capitalizing the entire business value, including the amount financed by debt, at what is presumably your cost of equity for a business with MSC's ROIC and growth characteristics. Perhaps I am coming at this from a different angle than you, but it seems a little inconsistent from the way I am thinking about it, and for businesses with more debt this would lead to bigger distortions. AutoNation would be a good example of a business with meaningful…debt that this approach would distort the valuation on.”
When I’m doing my appraisal of the stock – this is my judgment on what the stock is worth not whether or not I’d buy the stock knowing it’s worth this amount – I’m judging the business as a business rather than the business as a corporation with a certain capital allocator at the helm. Capital allocation makes a huge difference in the long-term returns of stocks. You can find proof of that by reading “The Outsiders”. Financial engineering makes a difference in the long-term returns of a stock. You can read any book about John Malone or Warren Buffett to see that point illustrated.
But, for me…
My appraisal of Berkshire Hathaway is my appraisal of the business independent of Warren Buffett. Now, knowing Warren Buffett controls Berkshire Hathaway would make me more likely to buy the stock and to hold the stock. So, it’s an investment consideration. But, it’s not an appraisal consideration for me. When I appraise Berkshire Hathaway, I appraise the businesses without considering who is allocating capital. Otherwise, I’d value Berkshire at one price today and a different price if Warren died tomorrow. I don’t think that’s a logical way to appraise an asset. Although I do think that buying an asset that’s managed by the right person is a good way to invest.
A good example of this is DreamWorks Animation (now part of Comcast). Quan and I valued DreamWorks Animation at a level that was sometimes more than double the stock’s price.
There was a point where we could have bought the stock at probably 45% of what we thought the business was worth.
However, we asked each other: “If you bought this stock today and then Jeffrey Katzenberg died tomorrow, would you hold the stock?” Both of us said no.
That didn’t change our appraisal of the stock.
It did change whether we’d buy the stock or not.
We thought that – with or without Katzenberg running the company – the business of DreamWorks Animation (its distribution deals, the library it had, the characters it had the rights to, the systems it had in place, etc.) meant it was worth more than the stock market was then valuing it at. We just weren’t comfortable investing in a movie studio – even at a discount to our appraisal of that movie studio as a movie studio – without being sure we liked the guy running the place. What I mean is that if I was a private buyer and I was offered DreamWorks stock at $18 or whatever the low on that stock was at one point – I would’ve said “does it come with Katzenberg or not?” And if the answer was “No, he’s retiring”, my answer would be “Thanks. But we’ll pass on this deal”. I don’t want to own a movie studio without knowing who is going to run it.
I can value a movie studio without knowing anything about the head of that studio. But, I can’t invest in it.
When I appraise the business – I just appraise the business like I’m being asked by a private buyer and a private seller to arbitrate the case of how much cash should pass between the buyer and seller for this asset.
So, questions of how a business is organized as a corporation – where it is incorporated, how much debt it uses, how aggressively it avoids taxes, etc. – can be part of my thought process when it comes to a “go or no go” call on investing in the stock. But, these considerations don’t change how much I appraise the stock for. I would never arbitrate a dispute between a buyer and seller differently because of those considerations.
Honestly, I always value a stock – that is, a single piece of equity in the company – by first valuing the entire enterprise as if it was being sold to a 100% private buyer. Although I’m looking at a public company – I always think of it as if it’s about to become a private company.
I always use capitalization independent measures of value (an enterprise value based approach) when valuing a business. I understand the logic of valuing a business controlled by a certain capital allocator - Warren Buffett, John Malone, Robert Keane (head of Cimpress / "Vistaprint") etc. – using a mix where debt and equity are valued differently. However, when valuing a business, I am trying to appraise the day-to-day operations of the business in the sense of what is inherent to the operation – not the current corporate structure, current capital allocation policies, etc. which can all change if the company's management changes.
Sometimes, there are long-term financing advantages in place at a company. For example, among U.S. supermarket: Village Supermarkets (VLGEA) has long-term leases (often running 20-40 years) that allow it to occupy good locations in New Jersey at reasonable rental rates and Kroger (KR) has effectively financed the supermarkets it owns outright (about half of everything it occupies) using long-term fixed rate bonds that mostly pay low after-tax interest rates. When considering whether or not to invest in these businesses, I certainly consider the fact these leases don’t run just 5 years and these bonds aren’t due in just 5 years.
But, normally, I try to value businesses using multiples of EBIT, EBITDA, etc. that are independent of how the business is capitalized (whether it is using debt or equity) and even often how the business is taxed (if it is incorporated in the U.S. or Swizterland for example).
When writing Singular Diligence, Quan and I disagreed a bit about this. And so, sometimes we applied a higher EBIT multiple for European companies than we did for U.S. companies. I am not sure I agree with that kind of thinking. Over time, corporate tax rates in a European country could rise and corporate tax rates in the U.S., Japan, etc. could fall. It is probably unwise to assume a 100% probability of the same corporate tax rate everywhere in the world. However, I have always also felt that it's actually wrong to assume a 100% probability that the current corporate tax rate in a country will remain the current corporate tax rate.
Within a year, the U.S. corporate tax rate could be about 40% of what it was 50 years ago. I’m not sure that – 25 years ago – your ability to guess whether the U.S. corporate tax rate would stay the same, double, or halve over the next 25 years would have been very good.
Taxes from country to country – and state to state within the U.S. – vary quite a lot. Countries and states can change their tax rates. And corporations can change where they do business to avoid taxes.
Often, the difference a change in tax rates would make is not taken into account by investors.
For example, Village Supermarket (VLGEA) hasn’t really paid less than 41% in taxes at any point in the last 15 years (there’s one exception that’s too complicated to get into here), and would – under some proposed tax plans in Congress – have made about $1.90 to $2 in EPS last year instead of $1.60 if rates had been different.
If you just take the most recent tax rate as being the tax a company will always pay – you’d be changing your appraisal of Village by something like $6 a share depending on whether the federal tax rate was what it is now or what it might be soon. It’s only a $23 stock. So, a $6 adjustment in your appraisal price is about 25% of the market price.
The stock’s EV/EBIT of less than 7 looks low. The stock’s P/E over 14 doesn’t looks so low. If I could only use one measure: I’d always favor EV/EBIT over P/E. So, I’d say that stock looked cheap to me. Other folks – looking at the P/E of 14 – don’t see anything noteworthy there.
It sounds like a small point when I warn someone that Village might be worth $5 to $6 more per share if corporate tax rates were cut or Apple might be worth $15 less a share if top management decided to be less aggressive about avoiding taxes.
But, this becomes an issue in some cases that the market doesn’t pay enough attention to. I can see some industries right now (like ad agencies) where global peers inside and outside the U.S. have tracked each other nearly perfectly in terms of stock price movements this year but which – if the U.S. cuts its corporate tax rates – will see very different movements in their P/E ratios in the years ahead. The U.S. headquartered companies are going to grow their after-tax earnings a lot faster than the non-U.S. headquartered companies. And this isn’t because they are doing business in different countries – it’s just because of where they’re headquarters is.
Now, you can certainly use an approach where you value debt and equity differently and still be aware of these things. But, that’s a lot to think about. And you’re likely to default to just assuming that whatever the market now values stocks at is what tells you the correct cost of equity, whatever tax rates now are goes into your model, etc.
I’d rather move up the income statement and think in terms of how much pre-tax income (in cash form) the company is producing versus the amount of total capital it’s using (not what is equity and what is debt in that mix).
I find the approach that works best for me is to try to value whatever company I am looking at – Village Supermarket, Apple, etc. – as if I’m appraising the business independent of the corporate tax policy, corporate debt policy, etc.
Then I just take the amount of debt the company has and give the bondholders the first portion of EV (up to the face value of the debt) and the appraised business value that’s left over is what I assign to the equity holders.
This gets into the issue of what I'm trying to accomplish by using an appraisal method of Enterprise Value (not just market cap) that uses EBIT (earnings before interest and taxes).
A business model does not have an inherent and immutable interest rate it pays, it does not have an inherent and immutable tax rate that it pays, and it does not have an inherent and immutable mix of debt and equity. These things will change over time. They are really "corporate" issues - that is, issues of financial engineering - rather than business issues.
When I appraise a company I generally want to use a "highest and best use" type approach like one would use with real estate. So, if a company is - like Village Supermarket (VLGEA) is - paying a 41% tax rate (between U.S. Federal and New Jersey state taxes) I don't want to make the mistake of assuming that's entirely a result of business decisions rather than financial engineering (corporate) decisions. The same is true - in the opposite direction - of something like Cimpress. Cimpress would be difficult to engineer in a way where it would pay less in taxes. The same is true for many of the big tech and drug stocks in the U.S. It would be hard to engineer these companies in a way where they would pay less in taxes than they already are. We need to dock them for that relative to company’s that can be financially engineered to pay less in taxes, carry more debt, etc. than they are now.
I want to be careful not to overvalue Apple and undervalue Village. The EV/EBIT ratio is helpful in avoiding this. The P/E ratio is not.
We could assume that we should treat things like debt-to-equity ratios, tax rates, interest rates, etc. as givens that won't change.
Or: we could assume that these things should tend to be leveled off much the way they would be in a 100% buyer's mind.
I think that’s the best approach. Instead of always thinking about what tax rates might change, how the cost of debt and equity might change, etc. we try to think in terms of what a private buyer would pay for 100% of the business if he could organize the business under any corporate umbrella he wanted to.
Let's use Village Supermarket as an example. Imagine it was for sale. Now, Village is - I think - the second biggest member of Wakefern. There are limitations on a non-Wakefern member being an owner. So, you can't really own supermarkets inside and outside the Wakefern (Shop-Rite) system. This means Kroger can never acquire Village. However, there is nothing stopping someone like a private equity owner from buying both the #1 and #2 biggest Wakefern members, firing the family members who work at those companies, leveraging up the combined (and now more synergistic) Shop-Rite operator and thereby paying less in corporate expenses, paying less in taxes, and tying up less capital relative to sales than is currently done at Village. In fact, if I had hundreds of millions or $1 billion on hand that is exactly how I would consider buying 100% of Village in a negotiated transaction. And ultimately it is this figure - what a company would be worth to a private control buyer - that I want to nail down. It’s important that - if I'm going to come to a conclusion that's independent of the stock market - I think in terms of a transaction for the whole company instead of asking myself what a passive, minority shareholder would pay for a share of the stock.
I don't want to ask: what would the stock market pay for Village?
I want to ask: what would a knowledgeable, private owner/operator of supermarkets pay for Village.
I find treating debt and equity the same useful when making that calculation.
Using Cimpress as an example takes us the other way. Cimpress may be valued more highly in the stock market than it would be valued by a private owner, because the person who controls Cimpress is running it in a way to maximize how much he can report in "adjusted" earnings and how little the company pays in taxes and so on. This isn’t necessarily the wrong way to run the business to maximize intrinsic value over time. It might be the right way. But, what I’m saying is that – if I was analyzing the acquisition of Cimpress as a 100% control buyer intending to take the company private, I’d have a hard time figuring out what I could do to end up with more after-tax cash in my pocket than the company is producing now. Likewise, I’d have a hard time figuring out how I could keep less of my own capital in the business (Cimpress has debt, capital leases, etc.) when I owned it than shareholders are now keeping in Cimpress.
If you have the time, go look at what Cimpress has paid over the last 15 years or so in taxes and what Omnicom has paid over the last 15 years or so in taxes. Then try to figure out how much of each company’s earnings have come from high tax countries like the U.S., lower tax countries, etc. I can’t come up with a business explanation for the tax differences. I can, however come up with theories on how you could do that through corporate level decisions.
That’s what financial engineering looks like. Based on that, I don’t think it would be a good idea to award as high a P/E ratio to Cimpress as you would to Omnicom. You can level this by looking at measures like EV/EBITDA and EV/EBIT.
Overall: I really want to discourage investors from ever using the P/E multiple in place of the EV/EBIT multiple.
In some cases, like Berkshire Hathaway (BRK.B) and Village Supermarket (VLGEA) there are steps the companies could take immediately to report higher earnings after-taxes than they now report. The stock market often does not see this or does not care about this. However, a potential buyer of the entire business always thinks in these terms. He thinks about how he could use debt instead of equity, how he could pay less in taxes, how he could combine one company with another, what he could do with excess cash on the balance sheet, etc.
The concept of weighted average cost of capital (WACC) is popular with academics. Bruce Greenwald is an academic. And so he used WACC in the book he wrote on value investing. If I ever wrote a book on value investing, I'd never once mention WACC.
And I don't think Buffett or Munger would either.
Now, that isn't to say that "cost of capital" doesn't matter. But, the way WACC is used by academics is problematic. Here's why. One, it's unnecessarily complicated. There is no need for the board of directors, the CEO, etc. to know what the cost of equity capital is. Their job is to maximize the return on equity. This means they may want to take on debt to the extent they can safely increase the after-tax earning power of the equity. It also means they may want to buy back stock if it increases the after-tax earning power of the equity. And, they may even want to issue stock if it increases the after-tax earning power of the equity. So, they need to think in terms of earning power. But, it's not actually necessary for any capital allocator to know what the market is going to value their equity at. John Malone is a good example. John Malone does not need to think in terms of Discovery Communications having a higher cost of capital when it uses equity, because investors are likely - since Discovery owns cable channels - to assign a lower earnings multiple to Discovery's equity than to the equity in other Malone holdings.
First of all, they could be wrong. In fact, I suspect John Malone and passive, minority investors differ in their appraisal of Discovery Communications equity.
And honestly, it is John Malone's appraisal of Discovery Communications stock that probably matters more than passive, minority investors' appraisal of the stock. That's because the value in Discovery Communications can always come from a transaction done outside of the stock market (taking it private, combining it with another company for cash, combining it with another company for stock, etc.).
Or, we could take the Disney and Fox negotiations. Murdoch has to consider whether he wants to be paid in stock or cash. He might also prefer one buyer over another, because he prefers one stock over another. Murdoch doesn’t need to think in terms of WACC to make this decision. He just needs to think of the price he is being offered in terms of the intrinsic value of Disney shares rather than the market price of Disney shares. If you’re locking yourself into a stock for any reason, the market price of that stock isn’t what matters to you. What matters to you is your appraisal of the intrinsic value of that stock.
I just don't see how thinking in terms of WACC makes sense when you are talking about equity. Now, what the cost of a company's liabilities are does matter. And we can get into that discussion in a second. But, the cost of equity is not something you need to figure out. It's an unnecessary complication in determining value. It's overly academic.
Very often, when you believe one company’s equity is worth a low multiple of book value and another company’s equity is worth a high multiple of book value – what you’re really saying is that one company (the high book value company) has ample access to stable, long-term, and low-cost funding as part of its day-to-day business. Examples would be: Berkshire Hathaway (BRK.B), Frost (CFR), Progressive (PGR), Omnicom (OMC), and Dun & Bradstreet (DNB).
In all those cases though, the low-cost funding comes from the basic business model. It is not the result of commodity type liabilities like issuing bonds. For example, Omnicom – and actually Berkshire long ago in its history (about 1989 I think) – did engage in some financial engineering to get low cost funding where they issued zero-coupon bonds. In a sense, issuing zero coupon bonds – the Omnicom zeroes were convertible into the stock – was a long-term speculation on interest rates (and because the bonds were convertible, the stock’s P/E ratio). Basically, Omnicom tried to lower its cost of capital (its WACC) through a pure financing play of trying to take the other side of a speculation from the market.
Omnicom was – this was in 2003 just after the millennium bubble years – betting that long-term interest rates were low and the stock’s own P/E was high. The imputed interest – interest you owe each year but don’t actually pay yet – also has tax implications that a financial engineer might be interested in.
Basically, these are bets. To the extent Berkshire and Omnicom were able to lower their cost of capital via these deals it’s really just because they were smarter than the people on the other side of the table from them. You too can find ways to bet P/E multiples won’t expand from a certain point, interest rates won’t rise from a certain point, etc. If WACC got lowered by these deals, that’s because management was smart and opportunistic not because it was an inherent characteristic of the business model.
I’m only interested in things that lower WACC that are a constant feature of the business itself.
I have written a ton about a company's cost of liabilities. But, I've done it specifically about what I consider day-to-day cost of liabilities concerns rather than financial engineering concerns. If I was a control investor - or at least an "influence" investor - like John Malone often is, then I might worry about financial engineering. But, I can't buy Village Supermarket with the idea of getting them to make changes to their cash, debt, owned buildings, leased buildings, etc. in order to reduce the amount of equity capital in the business while also reducing the amount of taxes paid. John Malone can do that in some cases.
So, what companies have a low cost of liabilities?
One, insurers have a low cost of liabilities if they are able to operate at a combined ratio below 100. This is what Berkshire Hathaway's insurers do. Berkshire owns a ton of niche insurers that have very low combined ratios (but aren't big enough to get individually discussed in Buffett's annual letters). Berkshire also owns GEICO. Progressive is a good comparable for GEICO. Progressive has been able to achieve high returns on equity (and grow intrinsic value per share) because it has a lower cost of liabilities than a life insurer like MetLife. In the long-run, MetLife is going to have trouble compounding shareholder value the way Progressive does because MetLife is paying more for its liabilities than Progressive is.
Other companies I have written about also have ample day-to-day business access to liabilities that have a cost of 0%. These include database / subscription type businesses like John Wiley, Dun & Bradstreet, IMS Health, etc. These companies can operate with negative book value if they want to. Basically, they can use all their profits to buy back their stock at above book value. And - over time - they've tended to outperform companies with business models that require them to pay more for liabilities.
Who pays more for their liabilities?
Any company that needs to borrow long-term to finance some kind of specialized capital like a steel plant, a race track, a copper mine, a stretch of railroad, a stretch of cable, a cruise ship, an airplane, etc. runs into this problem.
Some of these businesses can be good enough - because they may have monopoly like characteristics - to offset a high cost of liabilities. Examples would be: Hilton Food, Ball (BLL), U.S. Lime (USLM), etc. The business model requires capital. But, once you’ve built what you needed the financing for – you usually have zero local competition.
A low cost of liabilities is a feature in many - probably most - of the stocks I’ve bought recently.
Frost has a low cost of liabilities. In general, Frost can fund about 90% of itself at a cost that is no more than the Fed Funds Rate. Here, I am including both the actual interest cost Frost pays and the net non-interest cost it pays. What I mean is that - in a normal year - the Federal Reserve and Frost are paying similar amounts for their funding. That's really the key element of my analysis of Frost.
I get emails all the time about banks that trade at lower price-to-book ratios than Frost. However, the price-to-book ratio only works well when you are comparing two firms with the same cost of liabilities.
So, say MetLife trades at 1.2 times tangible book and Progressive trades at 3.8 times tangible book. Is MetLife stock really cheaper than Progressive stock?
Well, Progressive should trade at a very high price-to-book ratio because it normally has a negative cost of liabilities. The company is "paid" to make use of other people's money.
Many bank investors think Frost is expensive at 2.5 times tangible book. However, I disagree. The book value you are judging the expensiveness or cheapness of is only the equity portion of the company. A bank's value resides not in its equity but in its deposits. So, you can only decide whether Frost is expensive or not once you've done your best to determine how much or how little it pays for its liabilities.
Ad agencies pay nothing for their liabilities. This has been critical to their outperformance of other public companies over time.
So, if you want to talk about WACC in the sense of comparing the fundamental elements of a business model between industries, firms in an industry, etc. - I'm right there with you.
In the long-run, I want to own ad agencies rather than railroads. People forget this, but Omnicom stock has (even now when it's cheap) outperformed Union Pacific stock (even now when it's expensive) over these past 30 years - and not by a small margin. Over the last 30 years, the economics of railroads have improved a lot more than the economics of ad agencies have improved. But, advertising is still such a better business because it pays 0% on its liabilities (client money) while Union Pacific pays anywhere between 2% and 8% (pre-tax) on the bonds it uses to finance itself. In addition, Union Pacific's financing needs constantly increase in nominal dollars (because assets must always be replaced in real dollars) while Omnicom's financing needs constantly decrease in nominal dollars (it gets more float as client billings rise each year).
So, I do think of a company's cost of liabilities - whether it needs capital on top of the equity it has outstanding and how much that will cost. This is a key part of assessing the quality of the basic business model.
A great example of this is NACCO (NC).
NACCO has one consolidated mine. Other than that…
NACCO's customers finance the mines NACCO operates to supply those customers with coal. These economic liabilities - they don't show up as accounting liabilities, because under GAAP you don't consolidate an enterprise you are incapable of financing yourself - are non-recourse to NACCO but the economic benefits (the free cash flow) is paid out 100% to NACCO each and every year (it's not held at the unconsolidated subsidiary level where the liabilities are).
If NACCO was the company funding the mining operations – as it does with the one consolidated mine it has – there’s no way I would buy the stock at any price. So, the way the company’s business model works – in the sense of what liabilities cost the company – determined whether or not I’d invest. The company’s business model – not corporate level financial engineering – also gives it a lower tax rate (since it operates the mines that take the coal out of the ground – it gets the depletion benefit for tax purposes even though it isn’t financing those mines).
But, these liabilities come from the day-to-day operations of the business.
So, when Bruce Greenwald talks about something like the negative working capital cycle at Dell, I'm in agreement with his thinking. But, when he talks about using WACC to value a business - I'm not on board with his thinking.
The simple answer is that I never use WACC.
The more complicated answer is that I appraise a business in terms of what it would be worth to a 100% buyer who put those assets to their highest and best use.
And then the honest answer is: I focus heavily on stocks with access to low cost, no cost, or negative cost liabilities. These are the best businesses in the world. But, all my thinking goes into how the business model provides these valuable liabilities. None of my thinking goes into financial engineering.
One reason for this is that a business model seldom changes while financial engineering changes dramatically all the time at many public companies.