I recently received a couple questions about how to screen for “compounders” and whether it is better to look for a high growth rate or high return on capital.
First, we need to remember one basic principle. A company is limited to growing at a rate similar to its cash return on assets (really its incremental return on its net tangible assets) or to using outside financing. It is the incremental (cash) return on its own assets that matters. The reinvestment of these cash earnings relative to assets is what fuels the company’s growth rate. If the company runs out of this fuel, it will need to borrow money, issue shares or stop growing. There is no way around this reality.
That is something we can know rationally. Now, we can move on to something you would only be able to discover experimentally.
It is worth noting two statistical (screening) tendencies here.
One, screening for growth stocks normally also turns up high return on net tangible asset stocks because of this relationship between growth and return on net tangible assets. Today’s return on net tangible assets is what fuels tomorrow’s growth. So if you look at companies that have grown the most, they tend to be companies that had the most fuel available to grow.
Two, companies that increase assets tend to see returns on those assets decline, while companies that decrease assets tend to see returns on those assets increase. For a normal company, returns on assets are often lowest right after growth in assets had been highest.
It is possible to find many, many companies with high returns on net tangible assets while having almost no growth. I will use an example: Dun & Bradstreet Corp. (DNB, Financial). The company has a 15-year return on capital (calculated using the magic formula) of somewhere in the 100% to 1,000% range every year, yet a revenue growth rate of just 2%. Note, though, revenue per share growth was over 7% for those 15 years, which actually makes Dun & Bradstreet not an especially slow-growth stock despite being a no-growth business. In this case, we can assume the business - in real terms - has truly not grown at all in 15 years. There is basically no reinvestment opportunity at all. So the corporate top line has grown just 2% a year and the sales per share have grown a little over 7% a year.
The stock itself has outperformed the S&P 500 over the last 15 years (12% a year versus 4% a year). So it is possible for a no-growth business with very high profitability to beat the market. It is unusual, though. I consider Dun & Bradstreet the most unusual example I could possibly come up with. The business has literally no growth. Returns on capital are close to infinite. The capital allocation reflects this by simply returning all capital. This capital allocation approach is so rare, it is often better to look for companies with high returns on capital and some place to put that capital. Most boards would not just plow all their free cash flow into buying back their own stock like Dun & Bradstreet does. They would try to go out and acquire stuff.
Can we find a true growth stock that is not a true "quality" stock?
That is harder to do. The best I can come up with is Micron Technology Inc. (MU, Financial). The company has grown sales by 15% a year for 30 years. That certainly qualifies it as a growth stock.
Regardless, some factors suggest Micron may be "high growth" but not "high quality." For example, I said revenue at the company grew 15% a year over 30 years. However, sales per share grew only 13% a year over 30 years. The stock's return has been 12% a year. Certain assets, like inventories, have grown much faster than revenues and profits. Inventory has grown 19% a year. Property, plant and equipment has compounded at 20% a year. We can use Joel Greenblatt (Trades, Portfolio)'s magic formula approach to calculate return on capital here. Over 30 years, the arithmetic mean is 15% and the median is 5%. Especially in a series like this, the arithmetic mean far overstates the actual return on capital trend over 30 years. It is probably much closer to 5% than 15%.
If Micron's internal rate of return is too low to support the level of growth the company achieved over the last 30 years - where did it find a source of funding for that growth?
Between 1999 and 2001, Micron issued $1.5 billion worth of stock. Some of this stock was issued at prices higher than where the stock trades today (17 years later). So it took advantage of a bubble to shift money from mutual funds into semiconductor property, plant and equipment and inventory. The company briefly paid dividends for a few years in the 1990s, but has not paid them since. In addition, it had not bought back more stock than it issued until the last couple of years (really 2015 to 2017).
It also made use of debt at various times. Net debt has grown about 15% a year over the last 30 years.
Now, the end result of all of this - if you held Micron stock for the full 30 years - is a return of about 12% a year. So you would get a growth stock return over the long run. And that is despite me saying the business is not high quality and should not be capable of self-funding growth of 12% a year indefinitely.
Does this mean I was wrong?
At first, it seems so. But let’s dig deeper into Micron’s past to see if we can prove my larger point that only when return on capital is high can a company safely keep growing at a brisk pace.
Micron did survive the entire 30-year period despite growing debt the way it did. The Altman Z-Score would predict bankruptcy was imminent in 2008-09 (it was 0.4 to 0.5 for about two years with a Piotroski F-Score of 2), yet Micron recovered.
There is also the possibility we are - right now - at about the top of the cycle in Micron's industry. This would skew all my calculations severely. For example, if you sold Micron stock 12 months ago after holding it for 30 years, you would have a 9% annual return, not a 12% return like you would today. If you sold it 18 months ago, you would have gotten a 7% return over 30 years instead of a 12% return today. However, this is mostly an issue with the stock performance rather than the business performance. Yes, if I did the same calculation a year ago or so, you would get about 1% less in your compounded growth calculation for sales and other metrics over the last 30 years. That is far smaller than the huge difference in your long-term return as a shareholder, depending on whether you sold the stock two years ago or still hold it now.
By some measures, the business has done even better than the stock. EBITDA has grown by probably 13% a year over the last 30 years. That could certainly support a stock return of 12% a year over 30 years with no problem. However, as I pointed out, some asset items like property, plant and equipment have grown very close to 20% a year.
That is a disturbing trend. It is also inconsistent with the point I made at the start of this article: you need a high return on capital to maintain a high growth rate.
So what is the story with Micron? Is this a cyclical blip?
Does it only appear that Micron has compounded value at growth stock-level rates (12% a year over 30 years) despite having poor returns on capital because we are measuring during a great period for the industry?
Micron had a better stretch from 1993 to 1997 (in terms of return on capital) than it has had from 2013 to 2017. So this high five-year average ROIC by Micron's standards (12% by the Greenblatt method) is not a once in 30 years occurrence. It is more like a once in 15 years occurrence. In the mid-1990s, Micron hit a five-year average return on capital of about 30%. That would self-fund growth of 20% without any problem. But that period happened once and lasted five years. Today, it is having a five-year period that would fund growth of 8% a year without any problems.
So it makes sense Micron could grow a lot in the 1990s (20% a year without any problems), and it makes some sense Micron could grow at around 8% a year today without any problems. But it has grown faster than that.
For example, Micron's return on capital so far this millennium (2001 to 2017) has been so poor I would predict it could only sustainably grow at low single-digit annual rates. In fact, it has grown sales per share at just 1% a year from 2000 to 2017.
Now, it might be unfair to measure from the top in 2000 to today.
If we used the previous peak in revenue per share (1996) and measure to today, we would get 3% annual growth in revenue per share. If another cycle peak is happening now, that is a good measure of Micron's normal growth rate over 20 years.
Wait? How did Micron grow 12% a year over 30 years but only 3% a year over the last two-thirds of that period? What happened?
The truth is, I think Micron's growth fits with my points about return on capital. From 1988 to 1997, my way of thinking about return on capital would lead you to the conclusion that - provided there was enough growth in Micron's area of the economy - the company could grow about 25% a year without issuing stock or borrowing. If there was a little borrowing in there or taxes were lower than I expected or something, you would quickly get to an eight-fold increase in the company's size from 1988 to 1997.
Then, based on the return on capital after that point (1998 to 2017), I would say Micron would be completely dead in the water. It could self-fund growth of only about 3% a year (after some taxes) and the stock should badly underperform the S&P 500 no matter how fast the area of the economy it was growing in.
If you look at history - even here - the growth record matches up with the return on capital record in a way that says return on capital fuels growth. The stock returned about 35% a year for 10 years, while return on capital suggests after-tax ROE should be 25% or better (that is the late 1980s through 1997). And then the stock returned about 2% a year over the next 20 years while return on capital suggests after-tax ROE should be about 2.5% a year.
The company has tended to put everything it can back into the business. In the 1980s and 1990s, it could put something like 25% or more of its book value every year back into the business. Since 1997, it has really only been able to put something like 3% of its book value back into the business every year. In both cases, Micron was trying to grow as fast as it could. Only, the economics of the business (the return on capital) meant the fastest you could grow was over 25% in the 1980s and 1990s and then well under 5% in the 2000s.
The lesson of Dun & Bradstreet versus Micron is a particularly important one. There certainly was not more growth in demand for Dun & Bradstreet's services from 2002 to 2017 than there was for Micron's products. In fact, Micron's corporate revenue grew much faster than Dun & Bradstreet's corporate revenue (13% versus 2%). The difference, however, in revenue per share growth (9% versus 7%) was much smaller because Micron issued shares while Dun & Bradstreet bought them back. The latter also - starting in 2007 - paid out 4% of its sales in dividends. Micron did not pay any dividends.
So it is no surprise Dun & Bradstreet's stock grew 8% a year in price over the last 15 years while Micron grew only 5% a year. That does not include dividends (which the former did pay and the latter did not). Additionaly, it does not make any adjustment for the fact I am measuring a period that ends with Dun & Bradstreet stock down 18% over the last 12 months while Micron is up 105%.
This is the most extreme comparison I could come up with of a company that has all the growth potential but no return on capital to back it up (Micron), versus a company that has all the return on capital imaginable but no growth potential (Dun & Bradstreet).
The only time Micron could be a better investment than a company like Dun & Bradstreet was when the company was also a high return on capital stock. In its best period, the company had something like 25% unleveraged returns on equity for about 10 years. That is a high return on capital.
I have said before that once you find a stock that has about a 30% return on capital (so about a 20% after-tax unleveraged return on equity in the U.S.), you do not need to worry about ROC beyond that. Once you have that, worry about whether it can grow. You do not need more than 20% a year ever.
But - and here is the problem for Micron - I have also said that once you find a stock with a sub-15% return on capital (so worse than a 10% after-tax unleveraged return on equity in the U.S.), you do not need to worry about growth. The only thing you need to worry about is return on capital because no amount of growth at less than a 10% unleveraged return on equity is going to be worth passing up no-growth but high-quality companies like Dun & Bradstreet or average growth and average quality companies like the S&P 500 as a whole. Think of it this way: how is reinvesting at a 9% or worse annual rate of return any better than just paying you a dividend?
In practice, long-term growth companies have to be long-term, high-quality companies. The return on capital is what fuels growth. A business can only run on fumes for so long before its growth rate plummets toward the rate of return it gets on its own business.
Return on capital is always a limiting factor for a growth stock, as we saw here with Micron.
So what tends to happen is a stock like Micron will show up on both "growth" and "quality" screens in the 1990s and then drop off both screens in the 2000s.
Disclosures: None.
