Growth investing is often seen as dicey at times when markets struggle. But this time around, things have been seemed upside down in terms of what works and what doesn't. Indeed, we see some of that in the growth arena, where it is possible to develop a successful strategy by focusing on a big picture, as opposed to using the sort of here-and-now approaches that are typically preferred.
What usually works
In theory, it's long-term growth that counts; the proven, sustainable track record as opposed to immediate experience which may or may not be a flash in the pan. But in practice, investors have often placed more faith in what they see today than in longer-term trends that may or may not lie beneath the surface. That's especially so in the current generation, the one that came of age during the earnings-surprise/estimate-revision era.
We haven't gone so far down the path of immediacy as to render historical growth rates irrelevant, but we have lately been looking at the past through a narrow window. This can be seen in the 3/31/01 - 7/28/08 performance record of a growth rating I created in Portfolio123.com based on the following factors:
- EPS year over year growth in the latest quarter
- EPS year over year growth for the trailing 12 months
- Sales year over year growth in the latest quarter
- Sales year over year growth for the trailing 12 months
- Improvement in EPS growth rate from the trailing 12 months to the most recent quarter
- Improvement in Sales growth rate from the trailing 12 months to the most recent quarter
For each factor, companies are evaluated relative to the entire universe (an approach that adds a top-down flavor to the model by favoring stronger businesses and steering us away from struggling areas) and relative to their respective industries (something that opens the door to companies in cold businesses if they sufficiently outperform peers).
Figure 1 shows the results of a backtest. The right-most green vertical bar represents the annualized performance of the best stocks (the ones whose overall ratings were in the top 10 percent), the left-most red bar represents the S&P 500, the second bar from the left (blue) depicts the annualized performance of stocks whose ratings were in the bottom 10 percent, and the other bars depict the intermediate groupings. (Note: The results for this test and for all others below exclude stocks that trade on the OTC and any others with market capitalizations below $250 million or share prices below 3. Four-week re-balancing is used throughout.)
click to enlarge images
Figure 1 - Model based on recent growth (backtested 3/31/01 through 12/29/07)
This isn't by any means a perfect rating system. There's not as much differentiation as I'd like to see among some of the middle groups. But highly ranked companies have conspicuously outperformed low ranked firms.
Especially noteworthy is how Figure 1 compares to Figure 2, which depicts the results of a Portfolio123.com backtest of a longer-view growth model based on the following:
- Five-year rate of EPS growth
- Five-year rate of Sales growth
- Improvement in EPS growth rate from the 10-year span to the 5-year period
- Improvement in Sales growth rate from the 10-year span to the 5-year period
Here, too, each factor for each company was evaluated relative to the universe as a whole and relative to the industry in which the company operates.
Figure 2 - Model based on five-year growth rates (backtested 3/31/01 through 12/29/07)
However one might feel about the shorter-term model depicted in Figure 1, that approach obviously shines relative to the latter model.
What's Working Now
When we look at how these two models fared so far in 2008, it seems as if we've entered a new dimension of reality, a bizzaro world.
Figure 3 depicts the 12/29/07 - 7/28/08 performance of the model based on trailing 12 month and latest-quarter growth rates.
Figure 3 - Model based on recent growth (backtested 12/29/07 through 7/28/08)
The best group still outperforms the worst group. But one could have done even better by investing an any one of seven intermediate groups. That's definitely not what we want to see when we backtest a model. Moreover, there's nothing substantive to suggest that the seventh group, the best performer in Figure 3, owes its stature to anything substantive or sustainable.
Now look at Figure 4. Consistent with the bizzaro flavor of today's market environment, the five-year model that fared so poorly in the 3/31/01 - 12/29/07 backtest turned in a reasonably good showing in so far in 2008.
Figure 4 - Model based on five-year growth rates (backtested 12/29/07 through 7/28/08)
It's still not perfect. The eighth group outperformed the tenth (highest rated) group. On the whole, though, an investing strategy built around the tenth group seemed to have a good chance of relative success.
What's Going On
Is it possible that investors picked the present, a period characterized by volatility in commodities, weakness in economic activity, disasters in the housing and financial sectors, and potentially higher inflation and interest rates, to stop worrying about the here-and-now and to do what theoreticians wish they'd have been doing all along?
Was that a conscious choice? Probably not.
Nevertheless there are some defensive qualities that can be found in a five-year growth strategy. The longer the time horizon we use to assess growth, the harder it will be for companies whose claim to fame lies in their presence within industries that enjoyed a recent boom to make the grade. And by not looking at the latest quarter or trailing 12 months, we bypass the most intense among the boom-bust situations.
What we have, here, is a situation where consistency wins out not so much because that quality is being cherished for its own sake, but because those are the firms that get left over after we filter out the ones the Street most fears, those that boomed and then busted. In other words, it looks as if consistency is being favored by default.
To explore this further, I tweaked the five-year growth rating and added two factors relating to consistency (both of which were evaluated relative to the entire universe and relative to each company's respective industry):
- Sustainable growth rate (retention rate multiplied by return on equity)
- EPS Stability (the standard deviation of the last 20 EPS results divided by the average)
The 12/29/07 - 7/28/07 performance of this version of the model is shown in Figure 5.
Figure 5 - Modified model based on five-year growth rates and consistency factors (backtested 12/29/07 through 7/28/08)
It, too, is imperfect, but modestly better than the original 5-year growth approach.
Going from rankings to stock selections is often challenging. Here, our testing universe included about 3,300 issues, meaning the top 10 percent offered about 330 choices. Additional testing showed that narrowing further was not fruitful. In other words, while the top 330 or so stocks tended to outperform the market, the same could not necessarily be said about the top 15-25 issues.
Even though this growth strategy was not able to stand completely on its own, it seems quite capable of being a difference maker when added to other approaches. One example is a simple screen based on the more widely-used type of here-and-now growth strategy.
After narrowing the universe to eliminate OTC stocks, stocks priced below 3 and companies with market capitalizations below $250 million, I created the following simple screen:
- EPS growth rate in latest quarter is above industry average
- EPS growth rate for trailing 12 months is above industry average
- average analyst recommendation score less than or equal to 2.00 (1.00 = best, 5.00 = worst)
Figure 6 shows the results of a 12/29/07 - 7/28/08 Portfolio123.com backtest of this screen.
Figure 6 - Backtest (12/29/07 - 7/28/08) of near-term growth screen
Based on what we saw above regarding a strategy based on near-term growth, it should come as no surprise to see that this screen turned in a pedestrian backtest performance during the recent market environment.
But we achieve a substantial change for the better by combining that screen with the five-year growth rating (including the consistency factors). Figure 7 shows what happens when we choose the top 20 issues, not from the full 3,300-stock universe but only from among the 250-350 that make this screen.
Figure 7 - Backtest (12/29/07 - 7/28/08) of near-term growth screen combined with five-year growth rank
The combined rank-screen model significantly outperformed the Russell 2000, and did so with a list that is capped at 20 names, a grouping that an investor could actually trade.
Table 1 lists the 20 stocks that pass the models of this writing:
A bear-market strategy
Consistent with the poor performance shown by the five-year rank during normal times, this model is strictly for bear markets.
Figure 8 shows that it held up relatively well during the nasty 2002-03 period.
Figure 8 - Backtest (4/1/02 - 3/31/03) of near-term growth screen combined with five-year growth rank
A bull-market test conducted on Portfolio123.com between 4/1/03 and 1/31/05 confirmed the defensive nature of this strategy. During that interval, the Russell 2000 appreciated 73.2 percent. This model marched in lockstep, gaining 75.6 percent during the period. Interestingly, during months when the market was up, the model underperformed by an average of 0.45 percent. In months when the market fell, the model outperformed by an average of 1.66 percent.
Opinions differ as to whether the worst is now over for the market. If more bad times lie ahead, this approach is an example of a strategy that could work well for growth-oriented investors. If, on the other hand, we've had our major bottom and the time for defensive investing is over, we have solace in the backtests which show the downside as being performance in line with the Russell 2000. As a penalty for being wrong, that one seems pretty manageable.
Disclosure: Prism Financial, Inc., owner of Portfolio123.com, contributors, editors and employees may at times have positions in the securities mentioned and may make purchases or sales of these securities while this report is in circulation.