In yesterday's article "How Warren Buffett Made His First $100,000" I told you what stocks Warren Buffett bought when he was in his early 20s. Today, I'm going to talk about how he thought about those stocks.
There's a general perception that Warren Buffett switched from investing in cheap, low-quality businesses when he was very young to investing in less cheap, higher quality businesses when he was older. There is some truth to this.
But only some.
We know that Warren Buffett still buys some lower quality companies. Just in very small amounts. In fact, we know that he bought Korean companies in the early 2000s. This is what he told Harvard Business School students about those Korean stocks:
· Citigroup sent him a manual on Korean stocks.
· Within 5 or 6 hours… (he) found 20 stocks selling at 2 or 3 (times) earnings with strong balance sheets.
· Korea rebuilt itself in a big way post 1998. Companies overbuilt their balance sheets.
· Daehan Flour Mills – 15,000 won/year earning power. Stock was selling at 2 and change times earnings.
· …Strategy was to buy the securities of 20 companies thereby spreading… risk. Some of the companies will be run by crooks.
· …Doesn't know any of the companies and… (has) never been to Korea before…
· The investment presented an opportunity to make 150% at which point the stocks would still be selling at 7 or 8 (times earnings)…
· …Put $100 million into this by doing very little work.
So even in his later years, Buffett didn't limit himself to only buying wonderful companies at fair prices. Sometimes he still bought less than wonderful companies at very wonderful prices.
Likewise, 60 years ago, when Buffett was just getting started in investing – he put 75% of his personal portfolio in GEICO. GEICO was not a Ben Graham-type stock. It was selling for about 8 times earnings. That's cheap. But Buffett bought GEICO as a growth at a reas! onable price investment – not as a Ben Graham liquidation value investment.
So how is it possible for Buffett to move between these two extremes?
Sometimes, he buys very good businesses – like Coca-Cola (KO) or IBM (IBM) or Wells Fargo (WFC) – that only seem cheap if you believe in their franchises. These are far from Ben Graham bargains.
And then other times, Buffett buys companies like Daehan Flour Mills. Or he buys into a liquidation like Comdisco. Or an arbitrage position like Dow Jones.
How does Buffett choose between:
· A wonderful business at a fair price
· A fair business at a wonderful price
· A business that is liquidating
· An arbitrage opportunity?
Very few successful investors buy stocks that fall into all these categories. Ben Graham did arbitrage, liquidations, and fair businesses at wonderful prices. But he never bought wonderful businesses at fair prices.
Phil Fisher bought wonderful businesses at fair prices. But he never bought fair businesses at wonderful prices, or liquidations, or arbitrage.
Is Buffett just combining Ben Graham and Phil Fisher?
No.
Buffett invested in GEICO – in fact he put 75% of his net worth into GEICO – while he was still taking Ben Graham's class. GEICO is a great example of Warren's departure from the Ben Graham approach. Buffett was departing from Graham's approach from the moment he set foot in Graham's class.
How?
He was focused on his return on investment. He was focused on compounding his wealth. Graham wasn't. Buffett was. That was the difference.
And so Buffett immediately started buying the same stocks as Ben Graham – but he focused on just the very best ideas in Graham's portfolio. A great idea for Ben Graham would – at most – account for about 10% of his common stock portfolio. A great idea for Warren Buffett could be – like GEICO was – 75% of his portfolio.
When Buffett started his partnersh! ip, he ha! d a 25% position size cap. But he removed that to allow for a 40% investment in American Express (AXP). Buffett made many investments of 10% to 20% of the partnership's portfolio over the years. For Ben Graham, 10% to 20% was a really big position. It wasn't the kind of thing you bought every year.
So a huge difference between Ben Graham and Warren Buffett was focus. Buffett was always focused on his best ideas. This is part of what makes Warren Buffett similar to Phil Fisher. And very different from almost all other investors.
The other part of Warren Buffett's approach that separates him from most investors is that he's wedded to a very specific idea – return on investment – rather than a very specific style of investing.
The only way Buffett can sort through a range of different ideas including good companies, mediocre companies, liquidations, and arbitrage – is by looking at his return on investment.
I wrote about this back in 2011 in an article entitled: "Warren Buffett: Mid-Continent Tab Card Company."
That article was based on Alice Schroeder's description of Warren Buffett's investment in Mid-Continent Tab Card Company.
And it's a good article to read if you want to know how Warren Buffett thinks about stocks. Because it includes such heretical ideas as: "…growth had the potential to be either an added kicker or the most serious risk to his investment" and "you build the margin of safety into each step. You don't just slap a 40% discount on the intrinsic value estimate you get at the end."
But the most important statement in that article was:
"Buffett doesn't seem to make actual estimates. Alice Schroeder says she never saw anything about future earnings estimates in his files. He didn't project the future earnings the way stock analysts do."
How is that possible?
How can you sort through a variety of different investment options without using any explicit future estimates?
Y! ou have t! o think in terms of return on investment.
In fact, the reader who asked me the question that prompted the Mid-Continent Tab Card Company article actually got very close to identifying how Warren Buffett thinks about stocks:
"You wrote that Buffett just looked at the initial return (>15%) he was getting and the business's own ROC. When you aid 'initial' do you mean the 1st year? I think that sort of makes sense because his return of the subsequent years would be taken (from) the firm's own ROC and sales growth. Is that how you see it?"
Now, what did that reader get wrong? He came very, very close to describing how Buffett looks at a business. But he just missed.
What variable isn't being considered there?
Is it really true that: "his return of the subsequent years would be taken (from) the firm's own ROC and sales growth"?
Let's say a company has zero leverage. And its return on assets has been 10% a year for each of the last 100 years. You can bet on that 10% a year. Okay. Now, let's say it is growing sales by 10% a year.
How much is the business worth?
And how much should an investor expect to make in that stock if he pays exactly tangible book value?
Can the investor expect to earn 20% a year or 10% a year?
Or something in between?
Now, if you expect to hold the stock for a short-period of time your return will largely be based on what the market is willing to pay for each dollar of earnings the stock has in the future. So, you can certainly make over 100% a year if you buy a stock at 10 times earnings and sell it at 20 times earnings exactly one year from today.
I'm not talking about that. Don't worry about the resale value right now. Just look at the question of what the owner of a business can expect to make if the following facts are true:
· Total Assets: $100
· Annual Earnings: $10
· Future Annual Sales Growth: 10%
Do you think you can answer that ques! tion?
!
A lot of people think they can answer that question. But Warren Buffett would say you can't answer that question.
Not until you consider two possible future scenarios. Ten years from today, that same business could look like:
· Total Assets: $260
· Annual Earnings: $26
· Future Sales Growth: ?
Or it could look like:
· Total Assets: $100
· Annual Earnings: $26
· Future Sales Growth: ?
Or it could look like anything in between. In fact, I'm simplifying. If you look at their 10-year records, quite a few businesses grew assets faster than earnings. So, the range of possible outcomes in terms of the ratio of change in earnings to change in assets is even wider than I just presented.
If we look at two businesses each earning 10% on their assets, each unleveraged, and each growing at 10% a year – we can imagine one future where assets have grown by $160 over 10 years. And we can imagine another future where assets haven't grown at all over 10 years.
Which is the better future for an owner?
Obviously, the future with sales growth that far exceeds asset growth.
That would allow the company to buy back stock, pay dividends, etc.
So we can think of the combination of a company's return on assets and its change in assets and sales as being like the total return on a stock. The total return on a stock includes both price appreciation and dividends.
The total return on a business includes both the return on assets (from this year) and the growth in sales. But it does not include sales growth apart from asset growth. Rather, to the extent that assets and sales grow together – growth is simply the reinvestment of more assets at the same rate of return.
In other words, a business with a 10% ROA and 0% sales growth and a business with a 10% ROA and 10% sales growth could be more comparable than they appear. If the company with no sales growth pays out 10% of its assets in dividends ! each year! , why is it worth less than the business with a 10% ROA and 10% sales growth?
In the no-growth company, I get 10% of my initial investment returned to me. In the growth company, I get 10% of my initial investment reinvested for me. If the rate of return on that reinvested cash is the same rate of return I can provide for myself on the cash paid out in dividends – why does it matter which company I choose?
Doesn't an owner earn the same amount in both businesses?
Now, I think there are qualitative reasons – basically safety issues – that would encourage me to prefer the growing business. Usually, companies try to grow. If a company isn't growing, it could be a sign of something serious.
So a lack of growth is sometimes a symptom of a greater disease. But growth is not always good.
In more cases than people think, growth is actually a pretty neutral consideration in evaluating a stock.
There is an exception. At unusually high rates of growth – growth is almost universally good. This is a complex issue. But I can simplify it. Very few businesses that grow very fast do so by tying up lots of assets relative to the return they earn on those assets. Therefore, it is unnecessary to insist on high returns on capital when looking at very high growth companies. You'll get the high returns on capital – at least during the company's fast growth stage – whether you ask for them or not.
What do I mean when I say growth is often a pretty neutral consideration?
Let's use live examples.
Here is Hewlett-Packard (HPQ)…
10-Year Average Return on Assets: 3.2%
10-Year Annual Sales Growth: 10.7%
10-Year Annual Asset Growth: 14.5%
And here is Value Line (VALU)…
10-Year Average Return on Assets: 76.2%
10-Year Annual Sales Growth: (8.2%)
10-Year Annual Asset Growth: (11.1%)
Whose assets would you pay more for?
I have a problem with an 8% a year decline in sales. And worry tha! t the fut! ure looks really, really grim for Value Line.
But it's hard to say Hewlett-Packard has gained anything through growing these last 10 years. The company has retained a lot of earnings. And it retained those earnings even while return on assets was low.
The 10-year total return in Value Line shares has been (0.9%) a year over the last 10 years. The 10-year year total return in Hewlett-Packard has been a positive 4% a year.
So it sounds like Hewlett-Packard has done much better. But all of that is attributable to investor perceptions of their industry. If you look at HP's industry, total returns – from 2002 to 2012 – in the stocks of computer makers were around 14% a year. Meanwhile, publishers – like Value Line – returned negative one percent a year. So, Value Line's underperformance relative to Hewlett-Packard is probably better explained by the miserable future prospects for publishers compared to the much more moderate future prospects for computer companies.
Why does this matter in a discussion of Warren Buffett?
Because it illustrates the one future projection I do think Buffett makes. I think he looks out about 10 years and asks himself whether the company's moat will be intact, its growth prospects will still be decent, etc.
In other words: will this stock deserve to sell at a fairly high P/E ratio 10 years from today?
Warren Buffett doesn't want to buy a stock that is going to have its P/E ratio contract over 10 years.
To put the risk of P/E ratio contraction in perspective, consider that Value Line traded at over 5 times sales and nearly 25 times earnings just 10 years ago. Whatever the company's future holds, it's unlikely we'll see the stock at those kinds of multiples any time soon. Publishers just don't deserve those kinds of P/E ratios any more.
So, how much the market will value a dollar of earning power at in the future matters. And that is one place where projecting the future is probably part of Buf! fett's ! approach.. This is mostly a tool for avoiding certain companies rather than selecting certain companies.
For example, Buffett was willing to buy newspaper stocks in the 1970s but not the 2000s. The reason for that was that in the 1970s he thought he saw at least a decade of clear sailing for newspapers. In the 2000s, he didn't.
Today, I think Buffett sees at least a decade of clear sailing for the railroads and for IBM. In both cases, his perception of their future prospects was almost certainly the last puzzle piece to fall into place. It wasn't an issue of IBM (IBM) getting to be cheap enough. It was an issue of Warren Buffett being confident enough to invest in IBM.
By the way, let's look at IBM's past record:
10-Year Average Return on Assets: 10.3%
10-Year Annual Sales Growth: 2.8%
10-Year Annual Asset Growth: 1.9%
As you can see, IBM isn't much of a growth company. But that doesn't mean the shares can't be growth shares. IBM has improved margins and bought back stock. That has led to a 20% annual increase in earnings per share compared to just a 3% annual increase in total revenue.
So can we answer the question of why Warren Buffett is interested in companies like IBM and Norfolk Southern (NSC) rather than Hewlett-Packard and Value Line?
Well, Value Line is obviously too small an investment for Buffett. But we're using it as a stand in for all the publishers Buffett once loved but now shuns.
Buffett is a return on investment investor. He isn't exactly a growth investor or a value investor – if by growth we mean total revenue growth and if by value we mean the company's value as of today.
Buffett wants to compound his money at the fastest rate possible. So he looks at how much of the company's sales, assets, etc. he is getting. Basically, he looks at a price ratio. And then he looks into the company's return on its own sales, assets, etc. When you take those two numbers together you get something ! very close! to a rate of return.
The last part you need to consider is the change in assets versus the change in sales (and earnings). Does the company need to grow assets faster than earnings?
Or – like See's Candy – can it grow sales a little faster than assets?
Let's take a look at Norfolk Southern as a good example of the kind of railroad Buffett would own – if he didn't own all of Burlington Northern.
Norfolk Southern
10-Year Average Return on Assets: 4.9%
10-Year Annual Sales Growth: 6.0%
10-Year Annual Asset Growth: 3.6%
Now, how much earning power do you get when you invest in Norfolk Southern?
Total Assets are $28.54 billion. And the market cap is $21.28 billion. So, $28.54 billion / $21.28 billion = $1.34 in assets for every $1 you pay for the stock today.
Now, Norfolk Southern's return on assets has averaged a little less than 5% over the last decade. But I think that – like he does with IBM – Buffett believes the current returns on assets of the railroads are sustainable. So, we are talking something in the 5% to 7% range for a railroad like Norfolk Southern.
On top of this, he sees that the railroads have grown sales faster than assets. Now, we could do an elaborate projection of future margins, returns on assets, etc. to try to figure out what the railroads of the future will look like.
Or, we could just assume that over the last 10 years, Norfolk Southern has grown sales about 2.5% a year faster than it has grown assets. And Norfolk Southern can earn 5% to 7% on its assets. As a result, an investor in Norfolk Southern will see his wealth grow by about 7.5% to 9.5% of the company's assets he owns. This doesn't sound like much. But, railroads use leverage. And they often have price-to-book ratios lower than their leverage ratios. As a result, investors can often buy more than $1 in railroad assets for every $1 they spend on a railroad stock.
Let's say you get about $1.33 in railroad asset! s for eve! ry $1 you spend on a railroad stock. This is roughly the situation at Norfolk Southern today. In that case, your expected return as an investor would be what you expect the company to return on its assets (7.5% to 9.5%) times the amount of the company's assets you own per dollar you spent buying the stock (1.33). In this case, an investor's expected return in Norfolk Southern would be around 10% to 12.5%.
Is this exactly how Warren Buffett's thinks about stocks?
No. I don't think there's any way to boil Buffett's thinking down into a single formula.
But I do believe that the formula that best approximates Buffett's thinking would include the value of assets he gets when he buys the stock and the return those assets will earn as they move through time (hopefully growing slower than sales).
I don't think there's any way to incorporate growth into a "Buffett formula" and stay close to his actual thinking unless you are netting growth in sales against growth in assets.
Most formulaic approaches to Buffett's thinking depend too much on the stock's P/E ratio and its growth rate. And too little on the company's return on capital.
Buffett doesn't just see return on capital as a marker of quality.
He thinks of it as part of his own return in the stock.
How Warren Buffett Made His First $100,000
Warren Buffett: Mid-Continent Tab Card Company
Read Geoff's Other Articles
Ask Geoff a Question
Check out the Buffett/Munger: Bargain Newsletter
Check out the Ben Graham: Net-Net Newsletter
“That opening sentence has got to be clarified before you can have an intelligent conversation about it,” said Milevsky (left), speaking by phone from his office in Toronto.
