January 2001
Some Estimates are Better than Others
| "If I have ever made any valuable discoveries, it has been owing more to patient attention, than to any other talent." |
| -- Isaac Newton |
ESPITE WHAT THEY MAY SAY, investors are becoming more and more shortsighted. For a prime example, look no further than companies' quarterly earnings releases.
Every three months, stocks dramatically rise or fall as a result of making or missing analysts' earnings estimates. We're not talking significant variations here, just pennies will do. It's not uncommon for a stock to move 25% either way based on these short-term results.
|
|
    Source: Baseline/First Call In each of the examples above, analysts consensus earnings estimates are compared to actual results that have been lagged by a year. In the first example (Cisco Systems), estimates have been quite accurate. In the second (Ford) and third (Target) examples, estimates have run above and below actual earnings, respectively. In the final example (Pharmacia) estimates have been all over the place, with no discernable pattern. |
The most volatile period comes just before the quarter ends when companies "pre-announce" earnings that will fall short of estimates. Again, a few cents per share are enough to cost millions in market cap.
Given emphasis on matching analysts' predictions, we wondered just how accurate these estimates are. After all, if estimates - the benchmark - aren't accurate, why penalize companies that fail to live up to them?
Checking the Scorecard
Baseline provides a neat graphical way to do this. By lagging actual annual earnings by one year, you can compare them directly with analysts' projections. If estimates are accurate, the points will coincide. If not, estimates will run above or below actual results. In the worst case, estimates will be above in some periods and below in others, essentially showing no definitive pattern.The latter situation is the worst. Even if estimates consistently run above or below actual results, you can still form a reliable trading rule. But if there's no specific pattern, estimates are essentially worthless to a short-term trader; there's no pattern to work with.
The S&P 500 has numerous examples of each pattern as illustrated by the nearby price charts.
Next we wanted to see if there is a pattern to the patterns. In other words, are estimates more accurate in some sectors than others? To check, we used the Baseline data comparing consensus estimates and actual results over the past five years.
Each stock in the 500 was categorized 1-4 with 1s showing consistent agreement between estimates and actual results, 2's estimates consistently exceeding actual results, 3's estimates consistently falling below actual results, and 4s showing no discernable pattern.
|
|
 Source: Baseline/First Call We divided estimates into four categories and then compared them by sector. Category 1 estimates are highly correlated with actual earnings. Category 2 consistently exceed actual results while Category 3 consistently underestimate them. The final group, Category 4, show no definitive pattern one way or the other. |
We were most interested in those sectors and stocks that had either a high degree of agreement between estimates and actual results or where estimates were consistently below results. Short-term traders can make money with them while the other two categories don't offer similar opportunity.
A Few Surprises
Now estimates are just that, estimates. You naturally wouldn't think they'd always be extremely accurate. In fact, you'd think cyclical sectors (e.g. Consumer Cyclicals) or volatile ones (Technology) would be the least consistent. On the other hand, you'd probably also expect relatively steady earners like Consumer Staples would be the most likely to be accurate.Well you'd be surprised. As illustrated by the accompanying chart, the sectors with the most stocks falling into the favorable categories are Financials (90%), Consumer Cyclicals (82%), Cap Goods (71%), and Healthcare. Those with the fewest were Energy (13%), Basic Materials (14%), and Transports (33%)
Financials, Consumer Cyclicals, and Cap Goods are all sensitive to interest rates and the economy. Given that, it's surprising to see them consistently meeting or exceeding estimates. Energy and Basic Materials are fairly stable, so you'd think their estimates would be a lot more consistent than they evidently are. Oh, and by the way, defensive Consumer Staples and volatile Technology finished in a dead heat (60% and 57%, respectively).
The Bottom Line
As usual, easy generalizations just don't work. Given the range of results, consistency does seem to be sector specific, yet price volatility and earnings stability don't offer any viable indications or reliability.
Perhaps it all boils down to the quality of analysts covering each particular sector. Or maybe it depends on management's level of guidance -- maybe it's better in some sectors than others.
Whatever the explanation, the one thing to take away from this is the unreliability of earnings estimates. While they may move markets in the short-term, they aren't very accurate in the long-term. Short-term traders may be able to exploit them in some sectors, but real investors -- long-term investors -- should probably focus on other things like fundamentals and their own research.
If a simple measure fails these tests, maybe it can be modified to overcome its weaknesses. Consider, for example, the Price/Earnings ratio (P/E). What could be simpler? Once you decide what earnings you're using -- trailing or projected -- you just divide them into the current market price.
But is this an adequate means of comparison between Dow Chemical and Cisco Systems? Sure you can compare the ratios, but when one is growing earnings at 40% a year while the other grows in single digits, can the P/E really reflect this difference?
P/Es don't just vary from company to company, they also vary between sectors. The P/E for Technology companies will almost always be higher than that of those in Basic Materials. Earnings growth also shows the same variance.
Now you've got a means of comparing sector to sector and company to company. Unlike the simple P/E, the PEG is adjusted for sector and company specific differences.
In light of this comparability, the PEG can now be used to devise a trading rule. Value investors typically look for companies with PEGs of 1 or less. Growth-at-a-reasonable-price (GARP) investors will accept PEGs up to 2. PEGs at any level probably won't hold pure growth investors back, so we won't waste time with them.
But whether you're comparing companies or evaluating just one, you need to consider why the PEG is high or low. By itself, the actual value of the ratio isn't as informative as the reasons behind it.
For example, consider why a PEG would be low. First, as obvious as it sounds, it's important to realize there are three parts to the PEG: the price, the earnings, and the projected earnings growth rate. The first two components create the numerator, the P/E, while the final element is the denominator.
The ratio can be low if the P/E is low or the earnings growth rate is high. Obviously it's a positive sign if the earnings growth rate is high, but a low P/E can go either way.
The P/E will be down if the stock price is low or the earnings are high. Both of these are good, right? Well, not necessarily.
First off, low priced stocks may be that way for good reason. There may be some company specific problems such as management turnover or legal issues that have (and may continue) to depress the share price. Secondly, think about cyclical companies - their earnings usually peak just as the cycle is about to turn down. In other words, their P/Es will be at their lowest at precisely the worst time to buy their stocks.
In each of these instances, P/E and PEG ratios will be low, yet the stock may not be a good buy. While use of the PEG ratio is an improvement over the simple P/E - especially when comparing stocks across different sectors -- the ratio alone is not the be-all and end-all of individual stock selection.
You've still got to understand the ratio's components and what it's conveying. You've got to understand why it's giving you the values it's giving you. In short, although it's an improvement over other traditional metrics, you shouldn't hang your investing hat solely on the PEG.
Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials: 
"If you don't know where you are going, you will probably end up somewhere else." -- Laurence J. Peter
OU CAN EVALUATE STOCKS ON A NUMBER OF METRICS, but the trick is to make sure they make sense. Are measures comparable across sectors or stocks? Are they meaningful to each? Do you even understand what you're measuring?
It's simple math again -- just divide the P/E by the earnings growth rate.
PEG's Point
You can overcome these difficulties if you adjust the P/E ratio to reflect earnings growth. This is the whole purpose of the P/E to Growth (PEG) ratio. Just as you have to decide what earnings value you'll use when calculating the P/E, you also have to decide if you'll use the earnings growth rate for the next 12 months or the long-term consensus rate. Once you've done this, it's simple math again -- just divide the P/E by the earnings growth rate.
Hanging on a PEG
Cyclical companies' P/Es will be at their lowest at precisely the worst time to buy their stocks.
Homepage
Return to Top
