IN the euphoria of the final year of Peps we are likely to see millions salted away before April 1999 in tracker funds.

This week it emerged that a new monthly record of #1700m was poured into Peps in April. And now the Government plans to give only trackers the CAT seal of approval.

The financial pages provide evidence that trackers are one of the flavours of the month for Pep investors and glowing advertisements attest to their merits. They do indeed have their good points, but some research findings suggest caution. Quite simply, investing at present in a tracker fund that mirrors the FT All-Share Index might not be a successful medium-term strategy.

Between 1919 and 1993 there were 30 years when the yield on the All-Share Index was 5% or more at the end of the year and in every single case there was a positive (after inflation) rate of return on equities over the following five years averaging 14% per annum, less tax. In contrast, there were 12 years when the dividend yield at December 31 was 4% or under and following 10 of these years the rate of return on equities, again in real terms, was negative over the subsequent five years.

It is, of course, a fairly basic concept that yield bears in inverse relationship to share prices, thus when the yield is higher than average, as it assuredly is at 5% or more, share prices are, by definition, depressed; always a good time to buy.

These findings are obviously a summary of the past, and by taking the year-end data, arbitrary. Furthermore, no claim is made that this study represents a high degree of sophistication but dividend yields as predictors of share price changes do have actuarial legitimacy.

In a paper to an international actuarial colloquium in Rome in 1993, Professor A D Wilkie of Watsons the consulting actuaries gave an account of his investigation into the question: Can the dividend yield on a share price index be used as a predictor of future changes in that price index?

The study covered UK experience for a period of nearly 70 years and showed a very strong correlation between the performance both of a share price index and of a rolled-up index at the beginning of any period, with the correlation coefficients reaching a peak between six and seven years ahead.

Tracker funds have been marketed quite vigorously during the past year or so by many well-known providers as a relatively cheap way of participating in the stock market.

Part of their appeal has been that they do not required individual share selection, which takes up much fund management research time and money in trying to identify hidden value in individual shares. Instead, tracker funds are invested in the components of the index itself.

Retail funds in the UK must contain a minimum of 20 stocks with no more than 5% of the total fund being invested in any one stock but this could change.

In the US, investors can buy a single share in the index itself. These are known as Spiders (an acronym for Standard & Poors Depository Receipts) and it is forecast that UK investors will soon be able to participate in similar investment vehicles.

The research evidence does show a certain consistency and makes a fair case for only investing in an All-Share Index tracker fund for the medium term when the yield is above average and selling when it is below; it hardly provides a ringing endorsement for doing the reverse.

The gross yield on the FT All-Share Index is currently 2.78% (subject to daily variation), and well below average, so buying into a PEP fund based on it with a view to holding for the medium term, might not be too clever.

In any gold rush, it is well to remember that the surest money is made by those who sell the shovels, but I have yet to read an advertisement for a financial product which is expressed in quite these terms.

n Robert Calder is a private investor and a member of the UK Shareholders Association.