Consistent Corporate Track Records Few and Far Between

Typically when we run stock screens, we present the rationale behind the screen along with the results. For example, why it makes sense to look for companies with high returns on capital, and which companies have demonstrated that ability. This time, though, we'd like to focus not on the outputs of the black box, but crack open the lid of the black box and peek inside.

Here's what I mean. Most everyone knows that it's rare for a company to achieve high returns on capital for many consecutive years. But how rare is it? How many companies are able to do it for two years running? Three years? Ten years? And a related question: If a company posts a high return on capital in a given year, how long might we expect that performance to continue? These are the kinds of questions we'll examine below.

Consistent Growth: Attrition Rates
First of all, let's see what history tells us about a company's ability to maintain steady growth. In the table, we track companies through time, seeing how many of them are able to maintain growth--any level of growth above zero--in sales, earnings per share, and free cash flow. Generating these numbers is easy using  Morningstar.com's Premium Screener.

As you run your eye down a column, you can see how many companies (and what percentage) remain after another year rolls by. For example, our database contains 1,920 firms with sales 10 years ago. Of those, 1,513 managed to increase sales in the first year. Of that number, 1,280 managed to boost sales again the next year. By the 10th year, 316 firms remained out of our original group. In percentage terms, that's 16% of our original pool that managed to increase sales each year. Or, as the pessimists might point out, 84% of companies failed to do so. We repeat a similar procedure for growth in earnings per share and free cash flow, in which persistence of growth is even more negligible. Only 3.2% of firms boosted earnings each year, and 0.2% boosted free cash flow each year.

The Difficulty of Sustaining Growth
Years in a Row	Firms With Rising Sales	Percent Remaining	Firms With Rising EPS	Percent Remaining	Firms With Rising Free Cash Flow	Percent Remaining
Start	1,920	100%	1,904	100%	1,898	100%
1	1,513	78.8%	1,218	64.0%	1,050	55.3%
2	1,280	66.7%	798	41.9%	458	24.1%
3	1,038	54.1%	469	24.6%	176	9.3%
4	852	44.4%	311	16.3%	73	3.8%
5	704	36.7%	202	10.6%	40	2.1%
6	496	25.8%	111	5.8%	18	0.9%
7	384	20.0%	91	4.8%	11	0.6%
8	341	17.8%	72	3.8%	7	0.4%
9	316	16.5%	61	3.2%	3	0.2%

So what? Why might these numbers be important? First of all, they help refine one's sense of skepticism. When forecasting, it's easy to assume a smooth upward trajectory in sales, earnings, or free cash flow. But it's important to remember that reality is rarely so kind. Such records are exceedingly rare. (I should note that such records are even rarer than the table suggests. In our Premium Screener, I can only consider companies that still exist; the ones that failed, were delisted, or got bought out are no longer in the database. The data exhibit survivorship bias. The universe we're looking at here is composed only of winners--only firms that have managed to survive until now--so actual history would likely look even worse than the table suggests.)

Another benefit of such data is how they help inform anyone who constructs stock screens. For example, let's say I screen for firms with steadily rising free cash flows over the past five years. I would get a list of very fine companies. But what do the numbers in the table suggest? They suggest that the vast majority of companies that pass such a screen will fail to repeat their performance. If you had the run this screen in year five of the table, 40 companies would have passed the screen. Of those, only three remain today.

Since it's such a small group, you may be curious which three stocks managed to boost free cash flow each year. They are  Brown & Brown (BRO),  Danaher (DHR), and  UnitedHealth Group (UNH). I'd encourage you to read the Analyst Reports on each firm--they're truly fascinating companies, even if history suggests they'll have a difficult time continuing their winning streaks.

If you want to run these screens yourself, here they are:  consistent sales,  consistent earnings, and  consistent free cash flow.

Abnormal Growth and Profitability: Attrition Rate
Another way to examine persistence of a good record is to ask this: What percentage of firms are able to maintain consistently high levels of operating performance? For example, sales growth of at least 20% or returns on equity of at least 25%?

It's easy to use our Premium Screener to find out. We start with all the firms in our database for which we have revenues 10 years ago. That's 1,908 companies. We then see how many posted 20% sales growth in year 1. That's 540. Out of those 540, how many achieved 20% growth the next year? 281. And so on. Rolling the clock forward another year knocks about half the companies out of contention each time. After 10 years, we're left with just two firms. For the curious, they are  Apollo Group  and R&G Financial .

Sustainability of Great Records
Years in a Row	Firms with Consecutive 20% Sales Growth	Percent Remaining	Firms with Consecutive 25% ROE	Percent Remaining
Start	1,908	100%	1,918	100%
1	540	28.3%	231	12.0%
2	281	14.7%	141	7.4%
3	142	7.4%	92	4.8%
4	72	3.8%	67	3.5%
5	37	1.9%	53	2.8%
6	17	0.9%	37	1.9%
7	6	0.3%	31	1.6%
8	4	0.2%	27	1.4%
9	2	0.1%	23	1.2%

It's a little easier to maintain high returns on equity. At a level of 25% for ROE, 23 companies managed to clear the hurdle each and every year. That's 10% of the 231 companies with 25% ROEs in year one. Even so, you can see from the decay rate why, when Morningstar values companies, we always assume that abnormal returns on capital disappear eventually.

What we're really illustrating here is basic statistics. When you see an outlier--like a company with 20% sales growth or a 25% ROE--the reason it's an outlier is some mixture of luck and skill. The luck disappears quickly by definition, and even the skill will tend to be competed away over time.

Again, a good lesson lies in this data. Let's say I want to find companies with high revenue growth and run a screen for companies with 20% growth in each of the past five years. If I had run such a screen five years ago, 37 companies would have popped up on my screen, ignoring the survivorship issue I mentioned earlier. If we play the tape forward, how many of those would have maintained that pace of sales growth? Just two. And it's perhaps little surprise that the median five-year stock return of those 37 firms is slightly negative, right in line with the overall market.

Regression to the Mean
Though a little more involved, we can also use the Premium Screener to get a feel for how fast stellar performers, as a group, fade to the middle of the pack. Let's consider sales growth. First, we look back 10 years and rank companies from highest to lowest sales growth in 1996. We then divide this group into quintiles, and take the median sales growth of each quintile. We then track each quintile through time. As you'd expect, the median sales growth of the highest quintile falls over time, and that of the lowest quintile rises.

Rather than display a lot of raw data, what we've done in the table is show the difference between the top and bottom quintiles, and how that gap changes over time. In 1996, the gap between the median sales growth of the top and bottom quintiles was 56.5%. (The top quintile median was 46.47% and the bottom quintile median was negative 10.07%. The difference of 56.5% is in the table.) By 2003, the gap had disappeared. So at least in this particular period, it took about eight years for the two groups to become indistinguishable.

The Difference Between Top and Bottom Quintiles over Time
Year	Sales Growth	Return on Equity
1996	56.5%	38.9%
1997	18.7%	23.5%
1998	12.9%	20.8%
1999	7.0%	15.2%
2000	5.1%	14.5%
2001	4.8%	9.4%
2002	2.0%	8.7%
2003	-0.7%	7.6%
2004	-1.1%	11.4%

For return on equity, on the other hand, the two groups maintained their respective identities. The original gap did shrink considerably, but seemed to stabilize by about the fifth year. This jibes with our earlier table that suggested it's more likely for a company to maintain high ROEs than high sales growth.

(For those of you interested in running similar studies, here's how, using sales growth as an example. First, screen for companies with sales growth available 10 years ago. That's your universe. Then rank highest to lowest by sales growth in year 9/10. Divide the group into quintiles. Take the sales growth figure that corresponds to each quintile cutoff and create a new screen for companies in each quintile, which gives you five sub-universes. Then, take each sub-universe and rank by sales growth in each subsequent year, taking the median each time. To save time, use the custom view of the  Premium Stock Screener to show sales growth each year--that makes the ranking within each quintile easy.)

Conclusion
Outstanding performers in one period tend, on the average, to become mediocre performers in another period. This fact doesn't mean, of course, that corporate performance as a whole tends toward mediocrity. Previously mediocre firms bubble up to become stars (think  Apple Computer (AAPL)), and brand-new companies constantly emerge from the brains of entrepreneurs (think  Google (GOOG)). It's akin to a horse race. At any given instant, some horses are leading and some lagging, but the order is constantly shifting.

But the overall lesson is clear. When forecasting future sales, earnings, or returns on capital, keep in mind that time is the great equalizer. You need a very compelling reason to believe that a company can remain above average for long stretches. For us at Morningstar, it's the economic moat that serves as our touchstone. Unless a firm has a demonstrable competitive advantage--a moat--we'll forecast it to become mediocre very quickly.