Get Premium to unlock powerful stock data

Warren Buffett's (Modern Day) Margin of Safety

Author's Avatar
Geoff Gannon
Feb 14, 2012
Article's Main Image
Someone who reads my articles sent me this email:

Dear Mr. Gannon,

…in your analysis of Mr. Buffett, I often read you say that his margin of safety isn't always achieved by simply coming up with an intrinsic value for a business and then slapping a discount on it. However, why does he often make reference to what he thinks a business is worth and then compare it to what he bought the stock for (e.g. Washington Post was selling for $100m when he thought it was worth $400-500m, Wells Fargo was worth more than double what Berkshire paid, PetroChina was selling for $30bn when he thought it was worth north of $100bn etc). This seems quite different from the day one return of 15% requirement mentioned in the Mid Continent Tab Card Company example. Do you think this difference arises from the fact that he has a different valuation procedure when investing in stocks to whole businesses (although he says they are essentially the same thing)?

Kind Regards,


No. I think Warren Buffett usually thinks in terms of the return he can get on a stock. He has mentioned valuations to students before. He doesn't do it very often. Honestly, he doesn't give precise figures of either returns on investment or valuations very often. But he has many times – even with Bank of America (BAC) recently – talked about the kind of returns he can get in that stock versus everything else. I don't think he comes up with an intrinsic value estimate for something like Bank of America. I think he looks at the earning power. Also, if you look at some of the intrinsic value estimates Buffett has given in the past for companies, they seem to be closely connected to demonstrated earning power.

Someone mentioned a bank Warren Buffett bought for his partnership. If you read what he wrote about that bank it's clear his intrinsic value estimate is merely a rounded off percentage return. So, if he thought stocks should return 8% a year, he might say that a company earning – on average – about $5 a share is worth $60. In this case, it's reasonable to think he is just multiplying a company's earnings by the inverse of what a decent return would be in securities generally. Since 1 divided by 8% is 12.5. And 8% is a reasonable return. He may round that off to 12 and slap a 12 P/E on a stock for that reason.

I mention this because when Buffett uses the word "conservatively" in a valuation – it tends to be a valuation similar to what I just described. It almost seems as if his idea of conservative is a sort of "no-growth" valuation. This makes sense in context – especially at Berkshire – because Buffett's habit has been pretty much to never buy companies he expects to stagnate or shrink.

He only buys growing companies.

In that case, as long as growth is expected and the return on capital is solid, he just needs the price to average earnings ratio to be reasonable. The one thing I would say there is no evidence for Buffett doing is a discounted cash flow analysis. Whatever valuations he has done – I don’t think the idea of projecting earnings into the future and then discounting them back is part of it.

He has – however – mentioned future valuations in static terms. In other words, he's said that if we had to hold the stock it might increase in value from say $60 a share to $140 a share ten years from now. What there is not much support for is a Ben Graham type analysis. He doesn't seem to talk about a stock apart from its normal earning power.

And since Buffett now has held several stocks for a very, very long time the idea of worrying too much about the valuation at the time you buy seems silly. He has to know the price is reasonable and the return on his investment will be good. But no one time discount to current intrinsic value is going to still be supplying great returns in 15 or 30 years. And he has now owned some stocks for that long.

Let's use 20 years – about the length of time he’s owned Coca-Cola (KO) and Wells Fargo (WFC) – as an example.

If Warren Buffett buys a stock at 33 cents on the dollar in terms of price to intrinsic value and then the stock grows intrinsic value by 6% a year for the next 20 years at which point the stock trades at its new intrinsic value his compound return would be 12% a year. Buying a big cap stock at 1/3 of intrinsic value is a pretty extreme discount. And yet a return of 12% a year – while still quite good – is possible in some stocks that are already trading at a fair P/E as long as they are growing.

My point is just that over a period of 20 years or so, you're not going to get more than maybe 6% a year in your return coming from closing the initial valuation gap. In this sense, I mean a multiple. So, if you pay 5 times earnings for something that is worth 15 times earnings or 1 times book value for something that is worth 3 times book value – that kind of valuation gap will only provide about 6% a year over 20 years if the company itself is compounding book value, EPS, etc. at a leisurely 6% pace. You see how paying a bit more for something that could compound at 10% a year starts to make sense when you have a holding period of 20 years. Thinking you will have 10% growth and then ending up with 3% growth would be a big deal. But this kind of mistake isn't that hard to make outside of companies with good moats, indispensable products, etc.

So my point is just that having a clear idea of the intrinsic valuation as of this moment is more important when buying stocks you intend to hold for 1, 3, 5, or maybe 10 years. Many investors – Ben Graham included – had turnover in their portfolios that meant the average stock might be held for anywhere from 2 to 5 years. In those situations, the discount to intrinsic value – in other words the low price to NCAV, book value, earnings, sales, etc. – can have a very large influence on your annual return over the full holding period. But as your holding period gets longer – and Buffett's holding periods have gotten very, very long – you need to spend a lot of time thinking about the return on your investment and the return on what the company retains and so on.

IBM (IBM) is an interesting example. I'm not sure Buffett would buy IBM if it weren't for the share buybacks. The company has worked to avoid dilution and reduce the share count over time. He singled this out as a reason for the purchase when he spoke on CNBC. There are other reasons for buying IBM. But I honestly don't believe Buffett would be buying IBM if instead of buying back shares over time IBM was treating its shares the way many tech companies do.

Buying back shares gives Buffett clarity on long-term returns. Basically, some of IBM’s earnings can be used to buy more IBM stock. When IBM trades at reasonable prices, the return on these earnings used for buybacks is predictable and acceptable. Combined with decent sales growth in the single digits and strong pricing you have a recipe for good long-term returns in the stock without having to buy it at a super low price to current earnings or expecting a lot of growth. This has been his habit with companies like Wells Fargo, Coke, etc.

Sometimes the prices appeared cheap at the time he bought them. Sometimes they didn't. But in every case he was assuming a certain level of long-term profitable growth. In this sense, you could say he is more of a growth at a reasonable price investor. I don't think Buffett – at this point – spends a lot of time trying to figure out exactly what a company is worth today. I think he starts by finding the perfect company and then waiting for an acceptable price.

And, no, I don't think he has a different approach based on stocks versus owned businesses. There is some difference. I think – historically – he was willing to pay a bit more for a low to no-growth business if he could buy the whole thing because he could reinvest the cash elsewhere. At times, he has made comments that hint at this.

For example, he said he bought Dairy Queen as a business but wasn't interested in buying it as a stock. You could assume he just was talking size (he didn't want to waste time buying small part of a company that wouldn't move the needle at Berkshire rather than the whole thing).

But I think there may have been more to it than that. A high return on capital business is attractive as part of Berkshire in a way it would not be as a stock. With very high returns on capital, a wide moat, etc. a company's biggest issue will be constantly lower returns on all new investment. They will try to grow, make acquisitions, etc. You can see this as part of the genius of Teledyne, Berkshire, etc.

An oasis of extraordinary economics is worth more inside Berkshire than outside Berkshire, because it won't waste capital. I would say franchising businesses almost always work this way. They throw off more capital than can be reinvested prudently. At some point – this is true of a lot of specialty retail, restaurants, etc. but it's especially true of franchises – you hit super high average returns on capital. Most public companies will continue to invest far beyond that point and bring down the average return to near normal.

Considering how low the return on all investment past the peak average return on capital point was – these companies basically invested chunks of capital year after year at pathetic rates of return just to keep posting growth. Very moderate location growth combined with dividends, stock buybacks, etc. often works better.

You can study a bunch of publicly traded restaurant stocks to see this.

Restaurants are an extreme example.

But there are other businesses like that. Some great businesses have almost no good choices for additional investment within their own circle of competence. The best owner for these kinds of businesses would be someone like Berkshire, Teledyne, Biglari Holdings (BH), etc. Or a private individual who does the same thing by investing in different industries, regions, etc.

So, yes, there is a difference between stocks and wholly owned companies.

Wholly owned companies can't do super stupid things outside of their core business. They can't try to buy up big competitors. They probably won't try to grow aggressively when it isn't economical. Stocks – as public companies – will be eager to maintain at least a modicum of growth even when it makes no economic sense. And some industries have a growth fetish. A tech company for instance will always seek growth even when it can't find anything related to its core competency that offers hope for growth. This is dangerous. And wholly owned companies avoid this problem for Buffett.

On the other hand, negotiated sales are never super cheap.

Wholly owned businesses won't make as much sense as stocks in a market bottom. Today, they are fine. Stocks are not – generally – super cheap right now. But in the 1970s, stocks made more sense. You could buy pieces of businesses for far less – on a pro rata basis – than you could ever negotiate a sale of the whole business for.

That's the only differences I see. By the way, Buffett has – in the past – mentioned pretty similar hurdles for investments in each area. He has sometimes said he "pays 10 times earnings" for a whole business. This was said to a seller – not shareholder – and is obviously not some ironclad rule. In a shareholder letter from the early 2000s – Buffett mentioned the idea of wanting to see 10% plus returns in common stocks. When he didn't see those as reasonably likely, he wouldn't buy the stocks. He said he was usually – but not always – able to find such stocks. Those numbers are pretty close to each other. The 10 times earnings number must have been pre-tax. And the 10% reference to returns in common stocks would be allowing for some growth. You have to squint pretty hard to see how these numbers are that different from one another. I'd say he's willing to evaluate both stocks and private businesses in the same rough range of return possibilities. There may be some differences in how he sees them. But he values them in an awfully similar way.

Talk to Geoff About Warren Buffett’s (Modern Day) Margin of Safety [email protected]

Check out the Ben Graham Net-Net Newsletter

Also check out:
Not a Premium Member of GuruFocus? Sign up for a free 7-day trial here.
4 / 5 (51 votes)

Please Login to leave a comment