Get Premium to unlock powerful stock data

Warren Buffett: Make The Past Your Crystal Ball

Author's Avatar
Geoff Gannon
Jan 23, 2011
Article's Main Image
Recently, a reader sent me this question: “When valuing a company, how many years of financial data do you normally use?”

The answer is that I go as far back as the data goes.

If Wal-Mart (WMT) posts 38 years of financial data on its website – which it does – I’d go back 38 years if I was thinking about buying Wal-Mart.

I put all the data in a spreadsheet. And I look at it. That’s the first step.

Now, on my blog, I stick to using 10-year averages simply because I want to keep things standardized. I figure 10 years sounds doable to most people. And websites like GuruFocus show you the past 10 years. No website I know goes back 38 years. You have to go through the annual reports yourself for that.

The SEC reports at EDGAR go back to 1994-1996 depending on the company.

From everything I’ve heard Warren Buffett also likes to look deep into the past. He asks for all the past financial data a business has before buying it. Yet he doesn’t ask for any future estimates.

I think 30 year old financial data is at least as useful as future estimates. Estimates aren’t real. I only look at estimates to learn what management is thinking. I will decide against buying a company because the estimates make it clear the business is headed in a direction I don’t want – which usually means a new and allegedly exciting one.

The ideal business is one that doesn’t change. Absence of change makes it much easier to evaluate a stock. If the key drivers of profitability have been the same for the last 10, 15, or 30 years, they’re likely going to be the same for the next 10 or 15 or 30 years.

This is what Warren Buffett told a group of Wharton students about change:

“Some businesses will change very quickly. We are looking for ones that don’t….We are looking for the absence of change. Fruit of the Loom and Hanes together have 80% of boys underwear in the U.S. I guess we will keep wearing underwear. Bill Gates welcomes the absence of change, he just doesn’t get it in his business. Microsoft and eBay have some moat. If you can identify change, that is great, but it is a lot riskier and so is the chance of our strategy not working. So we look for absence of change. We don’t like to lose money. Capitalism is pretty brutal. We look for mundane products that everyone needs. Patents are the worst way to ensure demand.”

In fact, capitalism is so brutal, you can actually see change in the past record of most successful businesses.

Why don’t I want to buy Apple today based on its earnings?

Because if you look at Apple’s long-term record – not it’s 10-year record – but it’s truly long-term record, all you see is change. I’m not imagining the threat that Apple will no longer be cool and cutting edge and – you know – profitable. You only have to go back 10 years to see a time when it wasn’t.

And the reason most people give for why Apple was once all these wonderful things, then wasn’t, then was again is Steve Jobs. So we’re talking about a competitive advantage being one man.

If you’re buying Apple based on today’s earnings you’re saying that yes Apple will need to change faster and better than everyone else just to stay in place – because these aren’t razor blades or underwear or cans of Coke they’re selling – but that’s okay because you’re confident Apple will change faster and better than their best competitors.

Now that’s fine if you believe that. If you have facts and reasoning to support that assertion about Apple’s ability to constantly change faster and better than its competitors in the future, then you can buy Apple based on today’s earnings. Warren Buffett isn’t going to stop you. Ben Graham isn’t going to stop you. Even Ben Graham believed that there was more than one way to invest. He just didn’t have much faith in the other approaches.

So you can bet on change. And you can invest in Apple. And you can legitimately call it an investment, if – and only if – you can legitimately claim the same kind of confidence in Apple’s future ability to change as you can claim in the ability of businesses like razors and underwear and cola to resist change.

For Warren Buffett , the competitive inertia of a business at rest is a lot more reliable than the competitive inertia of a business in motion.

Now, that doesn’t mean you can invest solely on a business’s past record. But you can and should start with the past record.

The past record of consumer electronics is a record of change. So you don’t need to go any further than that. You know right off the bat that in a business like consumer electronics you’ll need some crystal ball other than the past record if you want to be confident even in a business’s current level of earnings being maintained.

No one’s talking about growth here. I’m not saying it’s unclear whether Apple will grow its earnings. I’m saying it’s unclear whether Apple will be earning as much in 2020 as it did in 2010.

So it’s easy to see when the past isn’t going to work well as a crystal ball because the past record itself is spotty. Sometimes the past record looks pretty solid. But it actually won’t work as a crystal ball, because of changes that have only recently started to hammer the business.

I’m not going to buy a newspaper no matter what the 30-year data says. I’m not going to buy a broadcast network like CBS (CBS) based on past data. I’m not going to buy any country’s telecom monopoly based on past data. Because those businesses are now subject to change in a way that invalidates the past. Since the past is a bad guide to the future in those businesses, I’m not going to use the past record as a crystal ball.

A good example of this is Barnes & Noble (BKS, Financial). I was interested in the stock up to the point where they were making too radical a departure from the past for me to be able to analyze the business.

Now, Barnes & Noble was – and still is, it’s only up $1.50 from where I bought it – a very cheap stock based on its past record. If you had confidence in the past being a useful crystal ball, you could invest in Barnes & Noble. I mean, this was a company selling for only about 4 times its average free cash flow of the last 10 years. It was the best bricks and mortar bookseller. And it’s next biggest competitor – Borders (BGP) – was lagging far, far behind.

So, to the extent the future was only going to be half as good as the past, you still had a good investment here. The future had to be a lot worse than half as bad as the past to lose money buying Barnes & Noble stock at $15 or $16 a share.

But then came the build up to this past Christmas and the heavy investment in the Nook Color.

The Nook is a big plus now for most investors in Barnes & Noble. It’s the one bright spot analysts who cover the stock cite.

For me, it’s exactly the reverse.

It’s the reason I sold the stock.

If Barnes & Noble had partnered with someone like HP (HPQ) the way Ron Burkle wanted, I might not have sold my shares.

Instead, Barnes & Noble pursued investment in the Nook very aggressively. Which is wonderful for the company’s employees and maybe its customers and all that.


It’s not a big plus for investors.

At least not investors like me.

I can’t count hardware sales as being valuable, because Barnes & Noble doesn’t have competitive advantages in the Nook that will stand up against investment from competitors. I didn’t go into Barnes & Noble wanting to own something like Apple (AAPL, Financial).

I can only evaluate two booksellers and two e-reader makers: Barnes & Noble and Amazon (AMZN). They are the only two companies who have competitive advantages in the e-reader market by virtue of their powerful bookselling operations.

But there’s no moat around the devices themselves.

There’s no moat around a better mousetrap. And to the extent Barnes & Noble and Amazon compete by putting out better versions of the Nook and the Kindle – there’s nothing sustainable there. You have to ante up again next Christmas. I can’t evaluate those sales the way I can evaluate sales of books (whether print or digital, online or offline).

So, I think Warren Buffett and I start from the same place in that we want to see past financials that tell us something about future profits. For me, past financials at Apple don’t matter. Apple re-invents itself every few years. I don’t want that. I don’t want to invest in something that’s making a new bet – the odds of which I can’t evaluate – every 2 to 3 years. I want to find something that’s placing the same wager every year.

Now, if Apple were selling at a discount to its current assets, that wouldn’t matter. Then you wouldn’t be looking at the past to justify the future. You’d be looking at the assets. That’s what Ben Graham did. And it worked for him.

But buying a stock based on its earnings is different. You need some confidence that the past is a kind of crystal ball into the future. Otherwise, who cares what it earned last year? The whole point of the P/E ratio is that somehow last year’s earnings can predict earnings in 2011 and 2012 and 2013 and 2014. Not perfectly. But approximately.

When you make the same bet every year it becomes like insurance, or roulette, or whatever analogy you want to use. And then all you need is an edge on the odds. As long as the company is making the same competitive bet year after year, all you need to do is buy a high return on capital business at a high free cash flow yield.

That’s it.

The problem with past financial data at most companies is that It’s like having historical loss data in insurance. It’s not going to do me any good to know how likely an insurer was to pay out on a fire insurance policy if it was writing them solely on small businesses in Utah from 2000 to 2010 and now plans to start writing them on small businesses in Ghana from 2011 forward.

I don’t know anything about losses from fire in Utah or Ghana. But I don’t feel comfortable assuming your past experience in one of those places is going to accurately predict your future experience in the other place.

A lot of businesses are that way. I think Apple is that way. I think it’s a great company. I think it’s also pretty close to impossible to value on an earnings power basis. What is Apple going to earn 5 or 10 or 15 years from now?

I don’t even know what Apple’s going to make in 5 or 10 or 15 years.

Now, something like Google (GOOG, Financial) today or the Washington Post (WPO, Financial) in the 1970s – that’s different.

Google is an advertiser supported media company that thinks it’s a tech company. To the extent I can know how much Google will invest inside it’s moat – search advertising – and how much it will invest outside it’s moat, I can try to value the company. The big question there is really just how much money Google will blow on things other than search.

It’s the same question with Microsoft (MSFT, Financial).

Google and Microsoft are – from an investor’s perspective – very similar value propositions. They have an easy to value profitable core business attached to the risk of future impossible to value non-core businesses. If you could take over Google and control its cash flows so it basically just bought back stock and paid dividends, you could value it quite easily.

Now, you might think that sounds silly. But Warren Buffett actually did exactly that at the Washington Post. He didn’t take over the company. But he was able to strongly influence the head of the company, Katherine Graham, in such a way that for years all the Post did was buy back stock. If Warren Buffett hadn’t been there, the Washington Post would have certainly bought some cable properties and other overpriced assets he didn’t like. Most media companies did back then.

I’m always trying to find something that’s cheap on the basis of its past. But I also want to know that the past is a decent guide to the future.

A lot of people think Warren Buffett predicts the future. That he projects earnings. That he does a discounted cash flow calculation.

He doesn’t.

Warren Buffett’s biographer, Alice Schroeder, went through his files when she was writing The Snowball and she found no evidence he ever did a discounted cash flow calculation or really any kind of future earnings estimates.

He just did calculations using past data. He looked at data from the company and its competitors over a lot of years. And then he studied it. He didn’t do a lot of calculations. What he was really doing is describing the company. He was laying out its past in the language of business: accounting.

Then he got comfortable with that past. And he looked at the initial return he was getting and the growth he thought he could get. And that was it.

But when I say the growth he thought he could get – there’s no evidence he quantified any of this. There’s really no evidence Warren Buffett has ever sat down and said it looks like Coca-Cola will grow 15% a year or 20% a year or 2% a year.

What Warren Buffett does is make sure he’s getting a decent return on his initial investment. In the earlier days that was 15%. That means he didn’t want to pay more than 7 times earnings (1/0.15 = 6.67x). Now, whether that was 7 times forward earnings or trailing earnings or “normal” earnings might vary from situation to situation depending on what kind of business it was. But basically he wanted to start off with a 15% return and then grow from there.

As long as he got 15% to start, the rate of growth wasn’t terribly important. I mean, it mattered. But it wasn’t a deal breaker. If you start with 15% and you know it can grow from there, you know you aren’t going to do worse than 15%.

The way Warren Buffett looked at an investment really depended on 3 things:

1. The initial free cash flow – or owner earnings – yield

2. The return on capital

3. Whether the business was growing or decaying

Ideally, in the early days, he wanted a business with a 15% return on capital selling for no more than 7 times its normal earnings and growing. The business had to be reliable in the sense that there wasn’t a risk of catastrophic loss. Exactly what he would use – whether it was free cash flow or earnings or whatever – probably varied depending on the nature of the business. And the return on capital measurement he used must have varied too. Banks earn almost nothing on assets, but can earn a decent return on equity. He must have used return on equity for banks and insurers. But how did he decide what an appropriate amount of leverage was?

The truth is that Warren Buffett never had a one size fits all mechanical approach like Greenblatt’s magic formula or Piotroski’s F-Score. He had a way of looking at investments and judging what was a good return. He had to tailor it slightly to each situation. So, whether the perfect “price” to use is the stock price or the enterprise value, whether the perfect return on capital to use is the ROA or the ROE, these are things that might change depending on whether it was a financial company, a utility, a railroad, or an advertising agency.

But Warren Buffett ’s basic approach seems to have been that he:

1. Checked for a risk of catastrophic loss

2. Studied all the past financials of the business and its competitors

3. Checked for a moat around the business

4. Asked what the initial return on his capital was

5. Asked what the return on tangible capital inside the business was

6. Checked to make sure the business would grow instead of decay

That’s it.

Or at least that’s an approximation of “it”.

A better approximation than thinking Warren Buffett actually estimates future earnings.

There’s no evidence Buffett looks at future earnings.

Exhibit A is this video of Alice Schroeder discussing Warren Buffett’s1959 investment in Mid-Continent Tab Card Company.

In case you skip the video – you shouldn’t, it’s terrific – I’ll leave you with the thoughts of one blogger who wrote this reaction after seeing the video:

“What was interesting to me from this case study was that (Buffett) did not seem to bother himself with projecting revenue and profits out five years, did not grind over whether to use a 10% or 12% discount rate, nor worry about which terminal multiple to use. There certainly was no investment banker’s “book.” Instead, he approached the investment wanting a 15% “equity coupon.” From there, he had to decide for himself whether it was a cinch to get such a return while relying solely on analyzing historical profit and loss statements.”

This was a company growing at 70% a year. And yet Warren Buffett didn’t make any projections of future growth. He just looked at the initial return he was getting and the business’s own return on capital.

Both were adequate.

So Buffett bought the stock.

That’s Warren Buffett’s crystal ball into the future.

He just looks at the past.

And Buffett's been pretty open about this. Although he's constantly said the value of a business is equal to the discounted value of its future cash flows, he's never said he does such a calculation.

For example, in the same 1992 letter to shareholders that people cite where Warren Buffett mentions John Burr Williams's The Theory of Investment Value definition of value as being equal to discounted cash flows, Buffett said this about his own approach:

“Though the mathematical calculations required to evaluate equities are not difficult, an analyst - even one who is experienced and intelligent - can easily go wrong in estimating future ‘coupons.’ At Berkshire, we attempt to deal with this problem in two ways. First, we try to stick to businesses we believe we understand. That means they must be relatively simple and stable in character. If a business is complex or subject to constant change, we're not smart enough to predict future cash flows. Incidentally, that shortcoming doesn't bother us. What counts for most people in investing is not how much they know, but rather how realistically they define what they don't know. An investor needs to do very few things right as long as he or she avoids big mistake.”

The problem with estimating future earnings - or doing a discounted cash flow calculation - is that it's very easy to make very big mistakes.

Warren Buffett 's lesson is to avoid big mistakes.

Focus on situations where you can use the past as your crystal ball.

Follow Geoff at Gannon On Investing

Also check out:
Not a Premium Member of GuruFocus? Sign up for a free 7-day trial here.
4 / 5 (48 votes)
Load More

Please Login to leave a comment