What will the s&p/ TSX composite index earn this year? How about dividend growth from the financial sector? And how high will the energy sector’s earnings rise?

I have some answers to such questions — as do a lot of people. The problem is there are always unknowns that wreck estimates — even the well-considered estimates made by powerful minds in the investment research community.

So I employ a back-of-the-envelope estimating method that avoids strenuous mental exercise but seems to work as well as any — a simple trend projection of recent results. Applied to earnings and dividends for the 10 sectors of the S&P/TSX composite index, it gives a picture of the possible outcome in 2006.

Mind you, as in any analysis, some results will appear unreal and improbable. In trend projection, a single large shift in recent data will skew results. Just about every forecast depends on past actual results — and that is the virtue of trend projection: it uses only known data. Securities analysts, of course, delve deeply into the unknown, studying every bit of corporate income and expenses to make assumptions and estimates on each before arriving at a conclusion.

Return on equity offers another shortcut to earnings estimates, useful mainly for individual companies. If, say, a company’s ROE has been no higher than 15% and no less than 10% over the past decade, you have a fair idea of what it can earn in the next year — probably between 10% and 15% of its latest book value.

Thanks to the wonders of computer spreadsheets (thank you, Microsoft Corp.!), projecting the earnings trend of an index or a stock is easy to do. Key in several years or quarters of data, apply a trend formula to several periods and you get a range of probabilities.

As with every other method, the further out you go, the less accurate the results become. That is because the long term is shorter than you think.

Growth is a sprint rather than a marathon run. A company or industry that has been having good results for two or three years is much less likely to extend this to a fourth or fifth year.

This is what destroys the certainties promised by all methods of earnings estimating — those done in great detail with great labour as well as the simple methods given here.

The table shows one set of projected earnings for the composite index and its 10 component sectors. The figures are projections four quarters ahead from a base of the past six quarters of actual results.

Because there are recent blips in data for several sectors — the problem mentioned above — estimates for these sectors are averages of projections based on several periods of data (see table footnote).

The projection for S&P/TSX composite earnings is $655, or 15% higher than the January earnings total of $568. But this gain depends on the three dominant sectors of the Canadian market: energy, financials and materials. Together, they account for 74% of the index.

Earnings trends in these three sectors diverge. Financial-sector earnings growth has been so slow, its earnings may not change much this year. Materials earnings may have peaked, judging by their trajectory over the past five or six quarters, but the trend points higher.

Energy sector earnings rise in one quarter, drop in the next — that has been the recent pattern, at least. Trend analysis smooths the data and says we should expect energy earnings to be 28% higher at yearend.

Projections of dividend growth show greater variance. The trend says information technology dividends will drop sharply. Financial-sector dividends, which have grown steadily, should rise 18%; dividends from the materials sector should gain 31%. Overall, the projection for composite index dividends is a 17% rise.

Finally, a description of the source data: in December, income trusts were added to the S&P/TSX indices. This ballooned dividends and yields and altered earnings on the indices.

To bypass this apples-and-oranges problem, the index manager continues some indices in their original form, with no income trusts included. Thus, the S&P/TSX “equity index” is the composite index as previously formulated. The 10 sector subindices continue in their original form as equity-only subindices. These are the data used in this study.

To illustrate what has happened: the S&P/TSX composite index dividend (including income trusts) at the end of January 2006 was $226.95, but for the equity index was $180.66 — a difference of 26%. The difference in earnings was only 2%.

@page_break@The biggest spread is in dividends on the energy subindex. With income trusts included, the subindex dividend was $60.32 in January, 132% of the $25.96 dividend on the equities-only version of the subindex. IE