Someone who reads my blog sent me this email:
I just came across your blog last night and i just finished reading the article how to value a business. You wrote... that means I appraise a stock’s earning power value as being worth (Cash - Debt) + (10 * EBIT) or (Cash - Debt) + (15 * Free Cash Flow)…my question is where do you come up with these figures 10 (for EBIT) or 15 (for FCF)? Does this apply to all businesses of any sort?
No. Those multiples doesn’t apply to all businesses.
I was just talking about businesses that are normally unlevered. So we're not talking about banks, insurers, railroads, utilities, etc. Those businesses are a little more complicated than a grocery store, a software company, an advertising agency, etc. So, let's put them aside.
The figures of 10 times EBIT and 15 times free cash flow are estimates of what the largest public companies normally trade for. If you go back to 1871 as Professor Shiller did and look at the S&P 500, it trades around 15 times the average (inflation adjusted) earnings of the last 10 years.
So, I'm talking about 10 times "normal" EBIT and 15 times "normal" free cash flow. Normal - like Warren Buffett said in this year’s letter to shareholders - means a general business climate better than 2010 but worse than 2005 or 2006.
In other words, the economy is still running a bit below normal right now. It was running a bit above normal in 2000. You kind of smooth that out as best you can. A good way to do that is look at average past earnings. Or, to use the free cash flow margin or return on capital over the last 10 years and apply it to today's current sales or capital base. Something like that. You don't just want to take last year's earnings. And you certainly don’t want to project next year’s earnings. Because that means you might be multiplying a peak year by 10 or 15 - which would be very bad if the peak is say 30% higher than "normal".
So, where do I get 10 times EBIT and 15 times free cash flow?
Well, Shiller found that 15 times past inflation adjusted earnings is normal for the market as a whole. Basically, Grantham and Hussman get the same answer in different ways.
That makes sense, because 15 times earnings is 6.67% (1/15 = 6.67%). I'd guess that's close to what the real return in stocks has been - so with 3% inflation and the market at 15 times earnings, you'd expect about 9.7% a year in total returns. Let's call it 10%. That's been possible because prices were at a certain level and inflation was at a certain level. If prices were normally higher at all times, returns would have to be lower.
Normal for the stock market has been a Shiller P/E of 15 times.
Why free cash flow instead of net income?
We don't have free cash flow data going more than a couple decades back. But I just want to make sure investors who read what I'm writing know to look for cash earnings instead of the kind of earnings railroads and cruise lines report. The depreciation expense at a railroad, utility, etc. is lower than their actual cash expenses. So, the "right" P/E ratio for a railroad might be closer to 10 times earnings than 15 times earnings. The right P/E ratio for an advertising agency might be more like 15 or 17 times earnings. The difference could be that big with some companies.
The amount of earnings turned into cash varies a lot by industry. So, I don't want anybody turning down an advertising agency with a P/E of 13 because that's higher than a railroad with a P/E of 10. In reality, the ad agency might be a little cheaper. Don't focus on free cash flow year-by-year. But over 10 years or something, free cash flow is what you want to look at. That's the only part of a company's earnings that can eventually be paid out as a dividend.
At the very least, you want to look at the historical relationship between net income and free cash flow. If a stock’s report net income always exceeds its free cash flow – that stock’s reported earnings deserve a lower multiple than if the reverse was true. Earnings that can’t be paid out in cash today are worth less than earnings that can be paid out in cash today.
The EBIT estimate is just a tax rate calculation.
Usually, at some point, you’ll want to evaluate a business separate from its debt structure. What is the underlying business worth? That's especially true when the amount of debt or net cash rises or falls over 10 years. Well, countries like the U.S. tax corporations around 35%. Or, I should say, that's what many companies I see actually pay. Some big multinationals pay less in U.S. taxes – but a big part of that is their refusal to bring earnings back to the United States. If you just look at small local companies in the U.S. they really do pay about 35% even if General Electric (GE) and Pfizer (PFE) don’t.
For example, Village Supermarket (VLGEA) runs grocery stores in New Jersey. It doesn’t get any earnings from overseas. And it does pay taxes in New Jersey. So, about 41% of its earnings go to taxes each year.
As you can see, 35% is just an estimate. But, it’s the best estimate we have when deciding how much of a public U.S. company’s EBIT will go towards paying the government.
Of course, a company’s tax rate will change depending on its mix of debt and equity. But, we don’t have to worry about that if we use enterprise value instead of market cap to judge the cheapness of a business.
Enterprise value looks at a company as if it had neither debt nor extra cash. In that case, the tax would be around 35%, because the company couldn't deduct any interest charges.
The formula is: Pre-Tax Earnings * (1-Tax Rate) = After-Tax Earnings.
In the U.S. that means: EBIT * (1-0.35).
And 15 * 0.65 = 9.75.
In other words, 15 times free cash flow is usually the same as just under 10 times EBIT.
The reason the two numbers don't match is that the company is using debt or paying more or less in taxes than is normal. A new owner could add debt or remove cash or do anything he wanted. So, he might not look at 15 times free cash flow the way you and I would as passive outside investors. He might look at 10 times EBIT. Normally, buyers of whole businesses look at EBIT or EBITDA. They look at some operating earnings or cash flow number. Not reported earnings.
I should point out that none of this means you should pay 10 times EBIT or 15 times free cash flow. It just means that you will likely do about as well or as badly as stocks generally do when you pay those kind of prices. Those are the kind of prices the S&P 500 often traded at in the past. That's what companies like JNJ and Microsoft often trade for.
Not always. Microsoft is cheaper than that right now. That’s a reminder that there’s no perfect multiple – either of free cash flow or EBIT – to use when valuing a business. What multiple you pay for Microsoft ultimately depends on what you think of Microsoft’s future. Right now, Mr. Market is thinking gloomy thoughts about Microsoft. That’s why the multiple on Microsoft is so low right now. Whether it’s the right multiple or the wrong multiple depends on how quickly Microsoft’s competitive moat disappears. If the moat is going to remain for many years, the stock is clearly cheap today. If the moat is going to vanish any day now, the stock might actually be too expensive.
That’s the way it works with every company. It’s not really about growth – or not only about growth – the way many models make it seem. A stock’s multiple has a lot to do with its competitive position. How certain are future earnings? Coke doesn’t trade for a high P/E ratio in normal times because it has great growth prospects. Coke usually trades for a high P/E ratio because it has a great brand. It has a wide moat.
When you’re looking at the overall stock market, none of this matters. The S&P 500 doesn’t have a moat. The S&P 500 isn’t above average or below average. It is average. So you don’t have to worry about those things when looking at the overall market. Something like the Shiller P/E ratio makes a lot of sense when looking at the S&P 500. It makes less sense when looking at Microsoft or Coke or Facebook.
Things are different when you’re looking at specific stocks.
You would need to make sure the business was of high quality and you would want to pay less for your stock than most people pay for the average stock. After all, why pick stocks if you only get the same prices everyone else pays through their mutual funds and index funds.
Ben Graham wanted a 33% margin of safety. In other words, he never wanted to pay more than $2 for something that was worth $3. Applying this rule to stocks generally, that would mean paying less than 7 times EBIT and less than 10 times free cash flow.
Again, we mean "normal" EBIT and free cash flow here. If the company's having an unusually good or bad year, you might need to adjust the number up or down a bit so it looks more like the company's normal performance over the last decade or so.
Unfortunately, stock prices aren't real low right now, so it's kind of hard to find large companies trading for less than 7 times normal EBIT and 10 times normal free cash flow.
Finally, I should say a bit about what “normal” means.
It’s easier to recognize “abnormal” earnings than normal earnings.
But there is one rule you can always use. To be normal, last year’s earnings absolutely must meet at least one of these three conditions:
1. Last year’s EBIT is less than 10-year average operating margin times current sales.
2. Last year’s EBIT is less than 10-year average return on invested tangible assets times this year’s invested tangible assets.
3. Last year’s EBIT is less than 10-year average EBIT.
Meeting one or – in theory – all of these rules doesn’t necessarily ensure that the P/E ratio you see for the stock reflects that stock’s normal earnings. But you have to be able to use one of these three measures to justify last year’s earnings or you need to throw out last year’s earnings for being abnormal.
Normal means the margins are normal, or the return on capital is normal, or the overall profit level is normal.
By far the most controversial measure is #3. Be careful using it. The only reason to use a long-term trailing average of actual EBIT or free cash flow instead of applying normal margins or returns on capital to today’s sales or assets is that the actual volume of business in the industry bounces around a lot.
As long as you know the company has maintained or grown its market share over the last decade, it may be helpful to look at a long-term average free cash flow or EBIT measure because you can’t predict what normal sales for the industry actually are.
Obviously, it’s better to avoid companies like that.
The more stable market share, sales, margins, and returns on capital are – the easier it is to evaluate a business.
Taking Microsoft as an example, it’s very hard for us to know exactly what Microsoft’s future will look like. But, it’s actually pretty easy to see what the company’s normal earnings are. The problem is not the business cycle. It’s technological change. If we knew technology wasn’t going to change, we could definitively say that Microsoft is cheap right now.
That’s different from a company like Posco (PKX). We aren’t as worried about technological change in that case. But we are very worried about what “normal” looks like in the steel business. It bounces around a lot. You can see that in Posco’s bouncy margins and bouncy capital spending. We need some way to smooth out that bounciness.
So we look at things like margins, returns on capital, and long-term average earnings to get some idea of what a smooth, “normal” earnings number would look like.
We can then compare that number to Posco’s enterprise value.
If we’re secure in our knowledge of what normal looks like – and with Posco you’ll never be completely secure – the right price is probably about 15 times free cash flow.
But there’s a lot of uncertainty built into steel. And even if we were pretty sure of our estimate, Ben Graham would say there’s no margin of safety unless we paid 10 times free cash flow instead of 15 times free cash flow.
Obviously, Ben Graham didn’t talk about free cash flow – the cash flow statement wasn’t standardized until his career was over. But the idea of paying no more than two-thirds of what a stock is worth applies equally to stocks bought for their earnings and cash flow as it does to stocks bought for their current assets.