Determining What’s Reasonable and What’s Not: An Update

[AF: In this update, I’ve added a substantial amount of new material, indicated in blue italics, on how to determine if our SSG Forecast High PEs and High Prices are reasonable.  My original post was published on July 14, 2008 and focused on how to determine if our SSG Sales and EPS estimates were reasonable.  As always, let me know what you think.]

Every SSG must make several judgments about the future, the most important of which is estimating the growth of earnings per share for the next 5 years. Estimating Sales growth is also important, especially when it is used to derive EPS growth as part of the BI/NAIC Preferred Procedure.

How do we determine what’s reasonable and what’s not?  The answer is easy to express (by comparisons) but much more complex to explain.  Consider this concrete example:

Is it reasonable to estimate 24.0% Sales growth for the next 5 years for Cognizant Technology Solutions (CTSH)?  Is 21.4% EPS growth reasonable?  How do we decide?

In the August 2008 issue of Better Investing magazine, guru Cy Lynch SSGed CTSH using S&P data.  Cy found that Cognizant had Sales growth of 48.4% per year during the last 10 years, “the highest rate in the IT services and consulting industry.”  His SSG also showed that CTSH averaged an even higher EPS growth rate, a whopping 49.6% annually, during the same 10 year period.

Moreover, over the last 5 years, CTSH’s annual growth rates actually increased to 55.3% Sales and 50.7% EPS.  So what do you think is a reasonable estimate for CTSH’s future growth in the next 5 years?

Here is Cy’s answer: “…applying arbitrary limits to projected growth in your stock studies under the guise of being conservative can lead to missed opportunities. I find estimated revenue growth of 24 percent to be reasonable.” [emphasis added]

DISCUSSION

(A) Determining if our SSG Sales and EPS Estimates are Reasonable: 

– Cy asserts that a 24.0% estimate for Sales growth is reasonable, but marshals no evidence to support his claim.  He does say that Sales grew by an “exceptional 48.4 percent over the last 10 years,” which on its face would seem to support a substantially higher estimate.  However, Cy makes no effort to explain why he is estimating so much less than the company’s historical average or why he chose 24.0%.

– For CTSH, Morningstar is estimating 32.10% Sales growth for the next 5 years and Zacks is estimating 55.66%.  Cy’s 24.0% estimate is conservative by comparison even though it is greater than the 20.0% maximum advocated by some other BI gurus.

– When I looked on July 12th, the analysts were estimating long-term EPS at around 28-29% with FactSet CallStreet low at 27.00% and Zacks high at 29.86%.  S&P was 28.00%, FirstCall was 28.49%, Value Line was 28.50%, and Reuters via Morningstar was 29.00%.  There were six estimates at FactSet that ranged from 32.0% to 20.0% low while the seven estimates at Zacks ranged from 49.0% high to 20.0% low. 

– When I looked again almost one year later on June 29, CTSH had fallen in price by -22% and the analysts had lowered their long-term EPS estimates to an average of 21.78% (down from 28.48%) with First Call low at 18.17% (down from 28.49%) and Value Line high at 28.50% (unchanged).

– Thus, 20.0% EPS is the very lowest estimate by any of these analysts when Cy wrote and and his 21.4% estimate seems conservative by comparison, especially compared to the 28-29.0% consensus analyst estimate.

– Cy derived his 21.4% estimated EPS from BI’s Preferred Procedure (PP)  which involves making four estimates for the next 5 years: Sales Growth, Pre-Tax Profit Margin, Tax Rate, and Shares Outstanding.  To get his 21.4% estimate, Cy estimated 24% Sales Growth, 19.2% PTPM, 25.0% Tax, and 300M Shares.  

– Certain tax breaks that CTSH receives from the Indian government are ending and Cy’s 25% Tax Rate estimate is close to Value Line’s 24% estimate in the next 3-5 years and identical to Morningstar’s estimated long-term tax rate of 25%.

– Cy wrote that his 300M Shares Outstanding was based on VL’s estimate, but VL was actually estimating 288M Shares Outstanding as per its 5-23-08 report. If Cy had used that number, his EPS estimate would have increased from 21.4 to 22.4%.

– Cy also used a Net Profit Margin of 14.3%, but did not explain how he converted that to get the Pre-Tax Profit Margin for his PP.  One way is to use this formula [NPM/ (1-Tax Rate)] which works out to 19.1% PTPM.  I also worked backwards from all his other factors to get 19.2% which is a slight decrease from 2007’s 19.4% and 5-year average of 20.2%.  PTPM has been trending down for the last 5 years.

(B) Determining if our SSG Forecast High PEs and High Prices are Reasonable:

– Our starting point is: High PE x High EPS = High Price.  When PE trends are not obvious or I know next to nothing about the stock, I most often begin by choosing a Forecast High PE using Toolkit’s Alt-M command.  Alt-M is an undocumented feature that averages the five lowest High PEs (and Low PEs) in the last 10 years.  It is usually the most conservative option but, when PEs are trending down, is not very useful.

– Some SSGers like to limit their Forecast High PEs to no more than 1.5 times their EPS growth estimate which is the same as using a PEG of 1.5  while others like to use 2.0 and still others prefer 2.1776 <grin>.  To me, all such limits are arbitrary and unreasonable as no one number fits all stocks.  Worse, there is no authority for a 2.0, 2.5, or 2 point anything High PEG.

– What I do is compare Forecast High Prices for the next 5 years to Value Line’s Estimated High Price for the next 3-5 years.  I use a two-part test: (A) Does the SSG High Price substantially exceed or fall below VL’s estimate and, if yes, (B) Do I know of a good reason that supports the SSG’s High Price.  If my answer is no to (B), I most likely would revise my SSG judgments.

– You can define “substantially” any way you like as there is no right or wrong for matters of judgment.  There is, of course, good and bad judgment.  I consider substantial to be 15-20% per share greater (or less) than VL.  I consider bad judgment to be ignoring VL.  And, good reasons can come from research, reading all I can including, for sure, the company’s latest annual report.

– One nice thing about VL’s High Price estimate is that it is expressed as a range from a low of “X” to a high of “Y”.  (VL’s low end is not comparable to the SSG’s Low Price).  For me, VL’s estimated High Price range defines what’s reasonable, and what’s unreasonable is anything that substantially exceeds or falls below VL without a good reason.

– For example, VL was estimating a $60-90 High Price for CTSH when Cy did his SSG.  I’m comfortable being at the high end of VL’s estimate because I know the company’s growth has been like a rocket ship.  If I didn’t know much about the company, then I’d prefer to be at the low end of VL’s estimate.

– Cy’s SSG estimated 21.4% EPS growth for CTSH and a 29.0 Forecast High PE that resulted in a $94.50 Forecast High Price, only $4.50 per share or 5.00% more than VL which, by comparison, seems reasonable to me. 

– VL makes a High Price estimate for only about 1700 stocks in its Large Cap edition.  Its estimate is for the next 3-5 years which means, as I understand VL, for the next 3 and next 4 and 5 years.  That is, the same estimate is applicable to a three or five year period.  VL’s Small and Mid-Cap edition contains no High Price estimate.

–  Keep in mind that VL provides benchmark guidance and is not an iron law.  What’s most important, I think, is not to overlook or ignore Value Line’s guidance.

CONCLUSIONS:

 – I have two, related quarrels with Cy’s analysis:

(1) his way low Sales and EPS estimates both of which were substantially lower compared to CTSH’s historical growth and to what the analysts were estimating.  Because of the way the NAIC/BI Preferred Procedure is structured, a Sales estimate that is way low results in an EPS estimate that is way, way (too) low; and

(2) I am not content with his unsupported assertion that a 24.0% Sales estimate is “reasonable”: simply claiming that something is appropriate does not make it so. 

– I no longer use the PP because it involves too many estimates and too much guesswork for me.  I prefer to survey the analysts, ascertain their long-term EPS consensus estimate, and thus learn what is high and what is low.  That way, I have some basis to know what EPS estimate is reasonably in the ball park and what is not.  See:  Pondering the Preferred Procedure.

– Make no mistake, I emphatically agree with Cy that arbitrary limits on our SSG estimates are serious mistakes, but I also insist on having some evidence to support each judgment and any claim of reasonableness.

– Here’s a concluding example that tries to put all this together in the context of assessing the Take Stock analysis of Starbucks (SBUX):

(1) On July 1, 2009, Take Stock was estimating EPS growth for Starbucks at -19.10% for the next 5 years (that’s a negative 19.1%).  That low-ball estimate works out to a TS Forecast High Price of $6.90 for the next 5 years with SBUX currently selling at $13.89.

(2) I checked the long-term EPS estimates for SBUX by six different analysts and they averaged 15.97% with S&P high at 19.00% and Value Line low at 10.5%.  By comparison, -19.1% EPS growth for the next 5 years seems patently unreasonable;

(3) In contrast to Take Stock’s $6.90 Forecast High Price for SBUX, Value Line is estimating a High Price of $25-35.  That means TS’s forecast is an whopping 72% UNDER the low end of VL’s estimate which again seems blatantly unreasonable by comparison.

(4) Take Stock’s Forecast Low Price ($10.01) is higher that TS’s Forecast High Price ($6.90).  As a result, TS’s High Price is actually low while its Low Price is actually high and that seems way, way wacky to me.

– Armin

[Ellis Traub, the developer and defender of Take Stock, and frequent poster to the BI Discussion List, is invited to comment on what he thinks is reasonable and what’s not.]