22 September 2005

Is the PEG Ratio worthwhile or worthless?

Dunno.

That's why this post is titled in such a manner.

I've done a bit of homework, pulled some nice research. Oh yes, we'll get to that shortly. I still just don't know. If anyone reading this has an opinion, please share in the comments.

I did a search on PEG ratios on a9.com (just for the 1.57% savings at Amazon) and found the old site
of Richard Miller, the purveyor of triplescreenmethod.com. Apparently, this fellow has a commercial concern based on this method of investing. With that grain of salt, let's take a look:

IBD and Zacks approaches to stock qualification identify fundamentally sound companies, but that's not enough. There's another piece of information needed: investors need to determine whether a stock is overpriced or not. Great fundamentals just aren't enough to produce winning trades. Frequently, these otherwise well-qualified stocks become overvalued as others see the potential, buying pressure builds, and price eventually exceeds the support of its earnings growth. PEG ratios (price/earnings/ earnings growth) provide one that estimate of value.

Earnings should be important to traders as well as investors, since future price is related over the intermediate-to-longer time frame to just three factors: earnings, earnings growth, and the stability of both. Peter Lynch in "One Up on Wall Street" said: "The P/E ratio of any company that is fairly priced will equal its (earnings) growth rate." That is, a PEG=1 marks a fully valued stock. The Motley Fool calls the PEG ratio their "Fool's" Ratio, and defines an undervalued company as one with a Fool Ratio less than 0.50. The importance of PEG ratios is well recognized. You probably have two questions: How does one calculate a PEG ratio? And how useful are such ratios in predicting price?

Mitch Zacks, in "Ahead of the Market," points out that analysts do only one thing well: project the next two year's earnings estimates from discounted cash flow modeling. Historically, they haven't rated stocks effectively, and they haven't provided good long-term earnings-growth estimates either. Though my PEG-ratio calculations differ from those made with historical earnings and five-year growth projections, I'm specifically targeting the value expected over the next two years, and the earnings projections over that timeframe are all that's needed.


I painstakingly plugged in data on all the stocks from the WershovenistPig Stock Watch List, using this so-called triple screen method. Assuming I didn't screw the math up, here are the results as of today, after the markets closed:

Ticker/PEG This Year/PEG Next Year/PEG Average
APC - .15/.59/.37
ARXT - na/1.38/na
ATYT - na/na/na
BDE - na/.81/na
BJ - .99/1.16/1.08
BYD - .29/.94/.62
CBH - 1.56/1.09/1.33
CTRN - na/1.44/na
CUB - na/.11/na
CVTX - na/na/na
DSW - na/.82/na
DWRI - .73/.30/.51
EDO - 2.58/1.31/1.95
FDO - na/1.65/na
HET - 2.06/.86/1.46
IGT - na/5.58/na
MRH - na/.02/na
TGT - .69/1.11/.90
UNT - .16/.54/.35
WMT - 1.72/1.06/1.39

An "na" appeared if the stock is too new to have prior year earnings, or if the stock's earnings declined.

So according to my calculations, the above stocks with the most promising PEG's are MRH, CUB, UNT, APC, and DWRI.

The least promising are EDO, FDO, HET, CTRN, and WMT.

Well, that really cleared things up for me. I'm really seeing the differences now.

On that note, let's look at some arguments from Professional Investor
against the PEG:

The PEG ratio has become increasingly popular over the past few years, particularly among growth investors, who want some way of relating the high P/E ratios they commonly have to contend with to a company's growth prospects. It is calculated by dividing the P/E ratio, usually the forecast P/E for next year, by the expected earnings growth rate. The convention is that a PEG ratio of one (for example, a P/E of 15 and a 15% forecast growth rate) is fair value; a PEG significantly above one means the company is overvalued, and if it is significantly below one then it merits further investigation.

Unfortunately, academics have documented major problems with using past earnings growth, and predictions of future growth, as a basis for anything. Back in 1962, Little of Oxford University showed that earnings growth is not a permanent attribute of a stock: past earnings growth does not carry over into the future any more than would be expected by chance alone. In 2003, Chan, Karceski and Lakonishok came to similar conclusions using large amounts of US data. As regards forecasting earnings growth, Ben Graham memorably described analysts' earnings predictions as 'somewhat less reliable than the simple tossing of a coin'. This is not hyperbole: it has been confirmed by a lot of recent research. For example, Keil, Smith and Smith in 2004
showed that analysts' forecasts would have been better in two-thirds of cases if they had simply said that company earnings would grow at the same rate as the average for all companies, rather than trying to guess it for individual companies.

Since the PEG ratio is calculated by dividing one of these predictions by the other, and the evidence is that analysts cannot predict either figure better than by chance alone, its usefulness should be in some doubt. But leaving aside these well-documented problems, here we concentrate on one particular aspect of assessing growth
stocks: the insistence on a history of growing earnings. Does a history of growing earnings mean good subsequent returns?
...
Many investors, and not only growth investors, believe that a company with a long history of growing earnings should tend to have good prospects for the future. This belief is false: knowing about past earnings growth tells us nothing significant, good or bad, about future returns. Where such companies do score over the average of companies in the market, however, is in the softer attributes of returns variability, and the chances of going into administration.


Alfred Rappaport and Michael J. Mauboussin also think the PEG is a load of hooey: I've distilled their effusively windy report
just for you:

[A] frequently used rule of thumb in the investment community is that stocks are attractive when they have P/E multiples less than the company's projected three- to five-year EPS growth rate. This so-called PEG ratio is expressed as the ratio of the P/E divided by the growth rate. The basic idea is that the lower the PEG ratio, the less you are paying for a company's future earnings. This rule of thumb is no more defensible than the simple P/E. After all, the numerator of the PEG ratio is the P/E itself. Further, we see no economically sound relationship between the P/E multiple and earnings growth projected over an arbitrarily short period.

I think I'll end up using the PEG as a minor consideration, one of many factors in determining whether to purchase a stock.

1 comment:

John Coumarianos said...

Earnings growth by itself means nothing if the business is not earning more than its cost of capital. The PEG ratio doesn't tell you whether it is or not, so it's of very limited use.

Earnings growth is great, but not if it's coming at the expense of dumping massive amounts of new capital into the business. The PEG ratio has no way of telling you at what expense growth is being achieved. ROA, ROE, and return on invested capital (ROIC) are better indicators of whether a business is earning more than its cost of capital or whether it's truly turning an economic profit. In other words, you've got to relate the income statement to the balance sheet at some point (which means going beyond the PEG ratio) to determine the price you're ultimately paying for the growth and whether it's worth it or not.