Always Use Normal Numbers

Author's Avatar
Dec 14, 2012
You always have to use normal numbers. Don't just use trailing earnings numbers of any kind. Know what goes into those numbers.

For example, Valley National (VLY, Financial) is a bank that depends a lot on its net interest margin. This, in turn, depends very heavily on the interest rate on long-term high quality credits. So, when long-term interest rates are very low - as they are now - Valley's earning power may be understated. You need to look at deposits - which is what they have available to lend out and make money - and apply a normal idea of the spread they can earn on those deposits they lend out. That may be very different from what they are earning now.

Likewise, Carnival (CCL, Financial) has a large fuel cost component to its business now. At 2012 level oil prices the company has margins, ROE, etc., at the level it does today. Those margins, ROE, etc., would become amazingly wide if oil prices were at the level they were at in the 1990s.

Honestly, I do not apply a simple EV/EBITDA valuation when I look at a stock. When I look at a bank, I look at deposits and how cheap they are to gather (in terms of fees taken in versus expenses the bank has to pay). This is probably what Warren Buffett was doing when he looked at Wells Fargo (WFC, Financial) both in the early 1990s and again at the time of the financial panic. He did not just look at a P/E ratio. And Wells - like Valley - was not actually super cheap (far from the cheapest bank) in terms of price to book. It was however cheap on a normal earning power basis if you looked at deposits (which they gather very cheaply) and imagine what they could be loaned out at over an entire banking cycle.

For many companies, you can look at EV/revenue. In fact, this is one of my favorite measures. But it is one of my favorite measures for consumer facing, more stable businesses. If you were offering to sell me a local TV station somewhere in the U.S., the first thing I would want to know is what price you were asking for in terms of enterprise value to revenue. Why? Because I don't think different owners will get much more or less revenue out of owning a TV station. I do think they may run it more or less efficiently in terms of costs. And I know they can choose totally different capital structures for the business. So, I would always start the valuation of a TV station on an EV/revenue basis.

What about building materials?

That's very, very hard right now. You can't value a building materials company on revenues right now. Vulcan had $3.6 billion in revenue in 2008. That was the peak. Often when you see a P/E quoted it is on this year's earnings - or actually estimates for next year's earnings - which just happen to be a peak.

One of the first things I do with any company is look for their past peak in earnings.

Vulcan's peak operating income was $714 million in 2007. Roughly speaking, a P/E of 15 would be $7.14 billion for the entire company. That's because corporate taxes in the U.S. are 35%. So 10 times pre-tax earnings is equivalent to 15 times after-tax earnings for companies that don't use debt.

There was consolidation in the building materials industry. And capacity is not being used fully. So, I think you would need to do a lot of research to figure out what earning power would be in a normal building cycle.

But that is always the idea you need to focus on. What is the normal earning power of this business? Be conservative. Look at it from a few different angles.

For example, Western Union (WU, Financial) will have lower earnings next year than this year. The company expects a 10% to 15% decline in earnings due to price cuts that will start to be offset within the next 6 months to 18 months by more volume.

Instead, let's assume they are too optimistic. Let's assume a 20% decline due to price cuts. And no offsetting volume increase. Let's just assume the normal level of earning power is 20% below this past year's earnings.

Operating earnings were $1.385 billion last year. Let's lop 20% off that number. We get $1.108 billion. Now, let's apply a 35% corporate tax rate. In reality, much of Western Union's business is outside the U.S. so it is never taxed as high as 35% - but we want to err on the side of conservatism. That would leave $720 million in net income. There are currently 615 million shares outstanding, so our estimate of after-tax earning power is $1.17 a share. This is potentially conservative for three reasons:

1) We made a much more pessimistic assumption about near-term earnings and price elasticity than management did (20% drop with no increase due to lower prices versus idea of 15% drop that they recover from over time)

2) We assumed a 35% tax rate which is more than WU ever pays

3) We used net income rather than free cash flow. WU does a much better job of converting net income into free cash flow than most companies. So, it should trade at a higher P/E ratio if it is to trade at the same price to free cash flow as most stocks.

So there we have it. Our estimate - which we believe to be conservative - of $1.17 a share in "owner earnings." The stock price is $13.11. So $1.17 / $13.11 = 9%. That is a 9% owner earnings yield. Western Union does not require much capital to expand its business (just signing bonuses) so any growth beyond its present earnings would basically fall straight to our total returns. In other words, we would expect to earn 9% plus growth in the stock. If WU grew earnings by 3% a year we would expect to earn 12% a year in the stock.

Let's check this estimate. Is there another more conservative way of estimating earning power at WU?

The first thing I would do is check operating income in the past. And margins in the past. Have they been lower than they are now? Not since the company was spun off. So our estimate of about $1.1 billion in pre-tax earning power is actually lower than what the company has experienced in the recent past. This includes the financial panic. So it seems we made a decently conservative estimate of earning power when we came up with $1.17 a share.

It's conservative because it assumes lower results than the company now has or ever has had in the recent past. It assumes lower margins than they now have. It assumes a worse near-term result than management expects.

It also assumes a higher tax rate than they will have. And it understates owner earnings because we know the ratio of owner earnings (free cash flow) to reported earnings is abnormally high at Western Union due to the economics of the business (payment processing).

Now we still might pass on Western Union because we don't think 9% is a high enough earnings yield. Or we think the long-term future is gloomy. That's fine.

But I think we did our best to come up with a reasonable but conservative estimate of present day normal earning power. If the situation changes in the future, it changes. We have to deal with that possibility separately. But I think we estimated what "normal" looks like at Western Union pretty well right there.

That's the approach I take with any stock. I look for the conservative side of reasonable. I look at the worst year they had in the last 5, 10 or 15 years. I look at margins today and in the past. I look at the median of everything - returns on equity, returns on sales, etc.

And I ask how reasonable is it that a company can fall to levels it saw in the past. And most importantly not what the highest peak and lowest valley is but what the middle path looks like.

I think doing that with building materials companies right now is way too hard. I think the valley we've seen these last couple years is way lower than normal. But I also think everything we saw from 2003 through 2007 was way too high a peak.

That means you have to look for other measures of normal than just the recent past. I think that's tough. And very speculative.

If you look at something like Valley National I think it's a different story. Right now, they are earning 68 cents a share. If interest rates stay at rock bottom (as the Fed has promised they will for the next few years at least while unemployment is high) then you basically know what you are getting. You are getting about a 7% earning yield. At the moment, that entire 7% earning yield is being sent to you as a 7% dividend yield.

What if interest rates start to rise again. Now, personally, I believe Valley's earnings in the boom were not as exaggerated as some banks because they did not expand nearly as much (in fact they pretty much didn't grow in the 2000s) and they did not make a lot of bad loans. The only influence of the financial panic you see in their results is the macro economy's decline which leads to more bad loans and lower demand for loans, the low interest rates caused by the Fed's response to the panic, and some investment securities they owned. Putting those things aside, the influence of the boom and bust is not terribly obvious when you look at Valley.

So let's shake the idea of today's interest rates from our minds. We don't want to be over focused on the present. We want to avoid recency bias here. We want to imagine the full scope of what the future could look like. It could be a future with higher interest rates, higher inflation, etc. It could look like any period from the 20th century. It doesn't have to look like the last few years.

In fact, I'm sure it won't. I have no idea what the economy's future will look like. But on average, the U.S. economy of the next 10 to 30 years will not look like the U.S. economy of the last 3 or 4 years.

It will average out to a more varied experience. And the value of a stock is linked to its long-term earning power. A stock should be valued as if it is an equity coupon that will be paid out over 30 years or more.

Valley has absorbed other banks (with the help of the FDIC). It is a bigger bank now. It has more revenue. Conservatively, I think we can say it will one day - who knows when - return to its past highest peak of net income. That was $164 million in 2006.

Valley has 193 million shares outstanding. That gives us $164 million divided by 193 million which is 85 cents a share. I think that's a reasonable but conservative estimate of Valley's earning power. We can do a lot of other calculations. But I don't think you'll find one as simple and common sense and yet conservative as that. This is a stock that should have a "normal" earning power of at least 85 cents a share.

It now trades at $9.26 a share. So that would be a 9% "normal" owner earnings yield. If we think VLY's normal earning power is 85 cents a share than we are buying a 9% equity coupon when we buy VLY shares today.

It may seem strange to treat earnings as 85 cents rather than 68 cents. It may seem delusional. It's not. It's a more honest approach to valuing the business the way an owner would. Don't get tricked by recent headlines. Look at the full sweep of past experience and imagine just how different the future can be from today.

You'll often find that today is not as normal as you might imagine. And therefore your estimate of earning power is not the same as what the company is reporting in EPS right now.

That's okay. Do the work. Trust yourself. Not the headlines. But always be both conservative and reasonable.

Finally, notice that nothing I did was precise. I didn’t try to be exact. That’s not the point of normalizing numbers. As a buyer of stocks, you just need to draw a line in the sand where you can say the upside from that point is a lot more reasonable looking than the downside.

In other words, it is reasonable to believe that Valley’s earning power – under normal conditions – would be 85 cents a share or higher. And Western Union’s earning power – under normal conditions – would be $1.17 a share or higher.

Just look for the conservative bound of the reasonable range. Plant your flag there. And decide whether the current stock price is a good bargain compared to that arbitrary but honest point.

Ask Geoff a Question