Occasionally, a new method of gauging investment prospects becomes popular. One of the most recent methods to come along is the PEG ratio — the relationship between price/earnings multiples and earnings growth. For a while, it seemed to be the No. 1 stock-picking method.

The idea behind the PEG ratio is to identify stocks trading at a P/E multiple less than their earnings growth rate. Such stocks, it is argued, should be worth buying. But some U.S. analysts have come to believe that the PEG ratio works better with stock indices than with individual stocks.

In the Canadian context, the PEG ratio is a lesser tool in the investment analysis armory. That reflects the widely varying correlation between earnings and stock prices in Canada.

This is easy to explain: the Canadian market has a huge cyclical resources sector.

At times, the S&P/TSX composite index — which has been calculated since 1956 — has no earnings, therefore, it has a negative correlation with price changes. Occasionally, there has been a high correlation between index earnings growth and price growth.

In Canada, the relationship between index PEG ratios and future earnings growth has been loose. Identifying cyclic lows in P/E multiples provides clearer examples of buying points. But it is not an everyday method of analysis, as the PEG ratio is.

This is the current conundrum: at yearend 2006, S&P/TSX composite index earnings had gained 51% a year compounded over the preceding four years, and the index traded at almost 16 times earnings. (Earnings of the S&P/TSX equity index — sans income trusts — had gained almost 52% a year and it traded at more than 16 times earnings.)

With a P/E multiple of 16 and recent earnings growth of 51%, it looks like a no-brainer: the market is a “buy.”

But is it really? Go back 10 years and see what happened. At yearend 1996, preceding four-year earnings had gained at a rate of 84% and the index P/E multiple stood at 24. The market looked like a sure “buy.”

But that proved to be not the case — earnings dropped in the two following years. And, over the four years of 1997-2000, earnings gained only 12% a year.

Let’s go back 10 more years, to 1986, when earnings had grown 15% a year and the P/E multiple was 17. Stocks looked a little expensive, but they weren’t compared with what happened the following year — earnings gained 33%. Over the four years subsequent to 1986, however, compound earnings growth was a slim 3.5%.

Going back another 20 years, to 1966, stocks seemed expensive in relation to recent earnings growth. In the following four years, their compound yearly return was very low, at 3.3%.

The PEG ratio worked in 1976, however, suggesting the market was a “buy.” It was; these were the early years of the great 1974-2000 bull market.

Making such comparisons on a sectoral basis is not possible because the TSX’s current global investment classification system started in 2002 and earlier earnings figures are not available.

That makes interpretation of current PEG ratios for the S&P/TSX composite index’s 10 sectors problematical. For example, should we really believe that the energy sector is a bargain when its P/E multiple of almost 12 is far below its recent earnings growth rate? And does the same hold true for the financial services sector?

The challenge is to determine whether energy earnings growth in the next few years will be anything like the 34% annual compounded growth rate since 2002, and whether the financial services sector can come close to its past 23.6% earnings growth rate.

There are similar challenges for the materials, consumer discretionary and telecom services sectors.

For the cyclical resources industries, a better guide may be relative P/E multiples. Energy stocks were a “buy” in 2003 when the sector subindex’s P/E multiple was in single digits. Industrial metals shares under the old TSX index system reached lows when their P/E multiples were below 20, and sometimes below 10.

Bank stocks in the previous index system were buys when their P/E multiples reached cyclical lows, but the lows followed trends. P/E multiples dropped from the mid-1960s to the 1982 market bottom, and then started rising so the lows were successively higher.

Thus far, in the short history of the new financial services sector, the lowest P/E multiple has been 13.4, but that occurred in late 2004 — which was almost two years after the index reached its low. IE

@page_break@