Most investors have probably never heard of the R-Squared Growth Rate. And even fewer know what it means. (Ten years ago, you could count me as one of them.)
But, now I do. I want you to know too. So let's get right to it.
The R-Squared Growth Rate is a measure of how close the actual earnings come to the earnings growth on a regression basis.
In other words, how closely do the earnings conform to the regression line? (Don't worry, I'll expound on this in a bit.)
R-Squared Growth Rate Explained
The range for an R-Squared value is between 0 and 1. (Or, if you express it as a percentage, between 0% and 100%.) The higher the value, the closer the data points conform to the regression line. The lower the value, the worse it conforms to the regression line. (In this example, the 'data points' are EPS growth numbers.)
A value of 1 means the data is a perfect fit - very rare to see. A value of zero (0) is the worst, meaning the date is scattered everywhere. In other words, if the data points are all over the place, it shows there's no rhyme or reason for how the data is coming in - extreme unreliability.
If, on the other hand, the data points are all plotting close to the regression line, that shows there's less deviation from the regression of the growth rate. And the less deviation there is, the more reliable (you would think) those numbers would be.
Investors will use this item to get a sense of the stock's ability to produce trendline EPS results. Of course there are no assurances that future data points won't veer off course. But knowing how closely matched the data points have been in the past is good to know.
What's interesting is that the distribution of the R-Squared values for the stocks in the Universe is an inverted bell curve (or well curve) distribution. (See below.)
Bell Curve and Well Curve Illustration
Usually, with a normal distribution (bell curve), the majority of the data will be in the middle of the range, with the smaller amounts of data falling on either side of the middle to form a symmetrical bell curve.
A well curve (abnormal distribution) has the majority of the data falling on either side with the smaller percentage of data in the middle.
For example: nearly 25% of the stocks had a value of .33 to .66. But roughly 38% of the stocks had lower values. And roughly 38% had higher values.
This distribution was the exact opposite of a normal distribution, so I decided to test it.
Before I did (and before I saw the distribution), I had at first thought that a ratio of 1 would be best and 0 the worst. But in my testing along with other items, these proved to be less reliable.
What I did find, however, was a range that produced the overwhelmingly best results. And that is: above the median with a 50% to 66% fit with the growth rate regression. And that's how we're applying it in this screen, which is aptly called the R-Squared EPS Growth screen.
So the screen I'm running today looks at the following parameters:
• Price greater than or equal to 5
• Volume (Avg. 20 Day Shares) greater than or equal to 100,000
• Zacks Rank less than or equal to 2
The Zacks Rank #1 and #2, Strong Buys and Buys, have handily beaten the market over the last 26 years.
• R-Squared EPS Growth: In (range) between .50 and .66
Above the median and one of the best tested ranges, i.e., EPS growth rates that show a 50% to 66% fit with the growth rate regression.
• PEG ratio less than or equal to 1
If a stock is trading at a multiple higher than its growth rate, it will have a PEG ratio over 1. (If it's two times its growth rate, it'll be 2.) If it's lower than its growth rate, it'll be lower than 1, and potentially be considered undervalued.
• P/E Using 12 Month EPS: In (range) 5 and 15
A top performing value range for this item.
• % Change Price over 4 weeks greater than -5%
Stock cannot have dropped more than -5% over the last 4 weeks.
Over the last 10 years, using a 4-week rebalancing period (and using 4 different start dates to verify the screen's robustness), this strategy produced an average annual return of 15.3% vs. the S&P's 7.4%, all while assuming less risk and less volatility than the market. (See below.)
R-Squared EPS Growth Screen (10 Yr. Study, 20042014, 4-week rebalance)
Courtesy of the Research Wizard
What's even more impressive is that this screen gained an average total compounded return of 14.2% during the devastating bear market in 2008, while the S&P 500 lost -37%.
It should also be noted that while there's no specific limitation placed on the number of stocks to come through, it'll typically generate, on average, 3-5 stocks each period.
Here are 3 stocks from this week's screen (for Tuesday, 7/29/14):
(ARW - Analyst Report) Arrow Electronics
(AVT - Analyst Report) Avnet, Inc.
(PRU - Analyst Report) Prudential Financial
Get the rest of the stocks on this list and start using the R-Squared EPS Growth Rate in your own stock picking.
Plus, imagine your friends' faces when they ask you what your secret is and you tell them the R-Squared Growth Rate. After some blank stares and some head nodding (the kind when someone nods up and down but has no idea what you're talking about), they'll think you're a genius.
Sign up now for your free trial to the Research Wizard and become a better stock picker today!
Click here for your free trial to the Research Wizard.
Want more articles from this author? Scroll up to the top of this article and click the FOLLOW AUTHOR button to get an email each time a new article is published.
Disclosure: Officers, directors and/or employees of Zacks Investment Research may own or have sold short securities and/or hold long and/or short positions in options that are mentioned in this material. An affiliated investment advisory firm may own or have sold short securities and/or hold long and/or short positions in options that are mentioned in this material.
Disclosure: Performance information for Zacks portfolios and strategies are available at: http://www.zacks.com/performance.