Someone who reads my blog emailed me this question:
“I'd love to hear how you think about return on invested capital. It seems every investor has their own definition.
How do you calculate it, what do you consider bad, average, good, exceptional, etc.?”
I use the pre-tax return on net tangible assets. When I’m looking at a non-financial company, I remove cash and other securities from total assets. I also remove intangibles such as goodwill and other intangible assets. The number that is important to me isn’t the return on net tangible assets. It’s the flipped figure. Let me explain. Let’s say a business can generate a 30% pre-tax return on net tangible assets. At a 35% corporate tax rate – like the U.S. has – that’s about a 20% after-tax return on net tangible assets. I don’t spend much time thinking in those terms. I don’t look at the company as earning 20 cents per dollar of net tangible assets. Instead, I think of the company as needing to retain one dollar of earnings to grow earnings by 20 cents – or, if you’d like to put it in even simpler terms, the company must retain 5 cents of earnings to grow EPS by one cent. This is the most useful number for me to consider, because I already know what the company is earning now. Return on capital matters to me as a long-term investor, because it determines the value of the growth I’m going to get.
Let’s look at an example like Bank of Hawaii (BOH, Financial). Bank of Hawaii is a financial stock. So, we don’t have to break out cash or anything like that. We can just talk in terms of return on equity. In a normal interest rate environment (something like a 3% Fed Funds Rate), Bank of Hawaii could earn an after-tax ROE in the 20% to 30% range. But how fast can it grow? Historically, the bank grew deposits a little faster than 5% a year. The bank has as high a level of tangible equity relative to total assets as it is ever likely to want to have. So, it doesn’t need to retain more earnings than usual. It can pay out anything it doesn’t have to retain to grow. Bank of Hawaii needs to retain 5 cents of EPS for every 1 cent of EPS growth. BOH is now making about $4.20 a share in earnings per share. It’ll make a lot more in a normal interest rate environment. So, it can grow EPS faster than it grows deposits while the Fed Funds Rate is being hiked. But, let’s pretend $4.20 was a normal level of EPS given the deposits the bank has. In that case, BOH growing at 5% a year for the next 10 years would be able to grow EPS from $4.20 to $6.84 ($4.20 * 1.05^10 = $6.84). So, the bank could be earning $6.84 a share in 2026. How much would it need to retain to do that? Over the next 10 years, the bank would grow EPS by $2.64 ($6.84 - $4.20 = $2.64). And if the incremental after-tax return on equity was 20% a year (which is very doable for Bank of Hawaii in a normal interest rate environment) then the bank would need to retain $13.20 a share in cumulative EPS over the next 10 years. It could pay the rest out in dividends and share buybacks. I don’t have a calculator in front of me, so I can’t figure out the exact cumulative EPS that Bank of Hawaii would have over those 10 years. But I can estimate it’s probably not much less than $55 a share over 10 years. In other words, it’s not that different than if the stock had an EPS of $5.50 a year for 10 years. In reality, the stock’s EPS would start at $4.20 today and end up at $6.84 in 2026. But, we can approximate this situation by acting like the bank was going to earn $5.50 a year for all 10 years. So, you have about $55 in cumulative earnings expected over 10 years and you have the need to retain about $13 over those 10 years to support this growth in EPS. You can check this by doing $13/$55 = 24%. And, yes, Bank of Hawaii could earn 20% to 25% after-tax if the Fed Funds Rate was about 3% or so. So, we have a cumulative payout potential for the stock of about $55-$13 = $42 over ten years. The bank can pay out about $4.20 as if it grows 5% a year. When Quan and I wrote about Bank of Hawaii for the newsletter, what we said is that Bank of Hawaii would pay out no less than 80% of its EPS in stock buybacks and dividends and it would pay out no more than 100% of its EPS. Before the financial crisis, it paid out a little more than 100% of earnings. After the financial crisis – while it was deleveraging – it paid out closer to 80% of EPS. But even during that time of de-leveraging it didn’t pay out less than 80% of EPS on average. So, it’s easy to estimate that if the bank is already earning $4.20 a share and it isn’t going to pay out less than 80% of earnings in any year – you’re never going to get less than about $3.35 a share (roughly $4.20 * 0.8) in a combination of stock buybacks and dividends. The bank pays out just under 50 cents a quarter in dividends. So, let’s say $2 a year in dividends. So, share buybacks should be about $1.35 a share at a minimum in any given year. They should also grow in line with the growth in EPS. So, if we expect EPS to grow 5% a year over the next 10 years – we’d also expect the amount spent on share buybacks (per share) to start out at $1.35 a share and grow by 5% a year over the next 10 years. This means the company should be – in 10 years – spending about $3.25 a year in dividends in 2026 ($2 * 1.05^10) and about $2.20 a year in share buybacks. The stock now trades for $85 a share. So, you’d expect the stock to appreciate in price to keep the dividend yield in 2026 from being too high or the stock buyback rate to cause EPS growth that is too fast.
So, what we are talking about here is a return on retained earnings. That is how I think about return on capital. Because Bank of Hawaii can’t spend very much on its actual business. But then, it can grow without spending very much on its business. Some companies can grow without having to retain any earnings at all. The example I always give is Omnicom. Because of the way payment cycles work for ad agencies – they get paid by clients a couple weeks on average before they pay for the services they purchase on behalf of clients – an ad agency’s free cash flow increases as its billings increase. The same thing is true for companies that collect money up front and then provide services later. So, if John Wiley (JW.A) is growing 3% a year, it can do that while paying out all of the earnings from its academic journal business in dividends and stock buybacks. It has no need to retain earnings. In that sense, growth is free.
So, that’s how I prefer to think about return on capital. I don’t look at how much capital a business has and how much earnings it has and decide that a 50% return on capital is better than a 15% return on capital. That’s not actually true. Let’s think about pre-tax return on capital. I own George Risk (RSKIA, Financial). I’d say that George Risk has probably gotten close to a 50% pre-tax return on capital in some years. But, George Risk doesn’t really grow. Now, let’s compare that to a company like LifeTime Fitness (now private). LifeTime Fitness can keep building new gyms across the country for a long time to come. What if LifeTime Fitness could get a 15% pre-tax return on capital. That would be a 10% after-tax return on capital. I can’t make 10% a year in the S&P 500. And I can’t take out a mortgage when investing in the S&P 500. But, LifeTime Fitness certainly can borrow about 50% of the cost of the land and building needed to create a new gym. Let’s be conservative and say LifeTime Fitness can earn a 10% after-tax return on capital when it builds a gym and it can get a long-term mortgage at 7.5%. Now, a mortgage at 7.5% would be 5% after-tax (because you deduct the interest cost and pay a 35% corporate tax rate). So, LifeTime Fitness would be financing a new gym in two parts. One part would be with equity. And it would be earning a 10% return on equity. The other part would be with debt. Basically, the company could earn 10% to 15% a year on a new gym for its shareholders. And it could do this while being pretty conservative. A mortgage with a loan to value of 50% on the kind of locations where LifeTime Fitness puts its gyms is not aggressive. And the earnings from the gym itself should do a good job of covering the interest cost. It should also be possible to have a very long-term mortgage even at a rate that isn’t remotely close to the return you are getting on your capital. In the example I gave, a 7.5% pre-tax yield isn’t low. And yet it’s far below the nearly 15% pre-tax return we’d expect LifeTime could make on a new location.
So, George Risk can’t create value for me despite its nearly 50% pre-tax return on net tangible assets. Meanwhile, LifeTime Fitness could have created value for me despite its measly 15% pre-tax return on unleveraged net tangible assets. Why? Two reasons. One, it could reinvest at a 10% unleveraged after-tax return. I don’t expect the stock market to do anywhere near 10% a year. In fact, I expect it to do more like 6% a year. So, a 10% unleveraged return on retained earnings is attractive compared to owning the S&P 500 at today’s high price for the index. And, two, LifeTime will use leverage when it builds its gyms whereas I will not use leverage when I buy a stock market index. This adds an element of risk. But, people won’t lend to me on the same terms to buy stocks that they will lend against land occupied by a brand new, large gym. You can’t get low cost, long-term financing to buy stocks. Warren Buffett (Trades, Portfolio) can. He can get no cost float. But even if we wanted to use leverage – the leverage we’d use would be very dangerous. It would be something like margin debt which is an extremely unstable source of funding compared to a long-term fixed mortgage. If land falls 50% in value but the gym keeps producing enough income to cover its interest payments – nothing changes about the mortgage debt. It doesn’t become due sooner. It doesn’t become more expensive. So, it’s easy to finance this kind of growth with less than 100% equity.
In principle, this is the way I think of return on capital. I think in terms of unleveraged return on net tangible assets. And I think in terms of the cost of growth. To me, Omnicom doesn’t have an infinite ROE – what it has is free growth. The question then is just whether Omnicom will grow 2% a year, 4% a year, or 6% a year. But, I know that growth will be free. What matters to me is the return on retained earnings.
For purposes of standardization, I always use the unleveraged return on net tangible assets. So, I take total assets. And then I remove cash and cash equivalents, goodwill, and other intangibles. I then divide EBIT by those net tangible assets.
In reality, the figure shown on the books may not be an accurate measure of economic value. EBIT isn’t free cash flow. And, of course, the assets have a fair market value different from what is shown on the balance sheet. For example, LifeTime Fitness’s actual return on net tangible assets is lower than what I described here. That’s because when LifeTime was a public company, the fair market value of the land it controlled was higher than the book value of that land. I’ve mentioned this before with other companies. For example, Copart (CPRT) carries land at less than its fair market value. This doesn’t much matter. The company gets a higher return out of using the land for its operations than another owner would get from putting the land to a different use. So, you don’t want to “count” the land as an asset in addition to the company’s operations. That would be double counting. But, you should be aware that it might cost Copart more to buy the same sort of land today. So, Copart’s actual return on earnings it retains today might be lower than the ROC of its older sties.
This just means you must use your best estimates. There is no perfect number. But, I think what you want to start with is the concept of the “cost of growth”. I think in terms of how much earnings a company would have to retain to support EPS growth of 1 cent. So, if you have a 20% unleveraged return on net tangible assets – you need to retain 5 cents of shareholder money to produce one extra cent of earnings.
An acceptable return on net tangible assets is 15% pre-tax. A good return on net tangible assets would be something like 22% (about 15% after-tax). Anything in the 30% pre-tax range or higher is more than you will ever need. In the long-run, no company is going to keep retaining all its earnings while earning an after-tax ROE of 20% a year. Companies just can’t grow that fast for that long. So, a lot of a high ROE company’s earnings are just going to be paid out in dividends and buybacks. What matters most is the difference between a company eking out a 10% after-tax ROE and a company doing a 20% ROE without the use of leverage. The first company is nothing special. The second company is really attractive if it can grow.
Â
