Investment Insider: Investors, don't be too quick to wish upon a star

ONE year's top-performing investment fund can all too often be the following year's loser. Looking at 1994's "winners" among unit trusts, almost half ended 1995 in the bottom quarter of their performance league tables.Sadly, that experience is the rule rather than the exception.

It should not, however, come as a surprise. More than likely, for an investment fund to beat its competitors, the investment manager must have been incredibly lucky or taken bigger risks in search of bigger rewards. If it was down to luck, stock-market history proves that it is rarely repeated; if it was the result of the manager taking above-average risks, it is more than likely that a price will have to be paid on the downside as well as the upside.

The word "risk" to a fund manager means volatility - not whether you lose money or how much you might lose, but how you lose it. In the mumbo- jumbo language of the investment professional, volatility is measured by "standard deviation" - the degree to which a unit trust fluctuates around its trend line, or average.

For example, two unit trusts can have an identical performance - say up 50 per cent over three years. But one may have experienced average fluctuations (or deviations) of 5 per cent a month, while the other trust may have fluctuated by only 2 per cent a month. This may not seem very important: an extra 3 per cent of volatility a month, to end up with the same result after three years, is unlikely to cause many sleepless nights.

But to the investment professional, what he would label "the risk-adjusted return" would strongly favour the second trust.

That is important, because in a further attempt to confuse you with statistics, unit trusts are now being rated in risk-adjusted form, the so-called "star ratings" which divide the total investment return by the standard deviation and recast the performance league tables. The trust with 5 per cent monthly fluctuations would have risk-adjusted performance of 10 (50 divided by five), the second trust with 2 per cent monthly fluctuations would have risk-adjusted performance of 25 (50 divided by 2) - twice as "good" according to the statisticians.

As a way of picking unit trusts, risk-adjusted performance tables and the "star ratings" which go with them are no more predictive of the future than the conventional tables. The trouble is, these star ratings are being used in adverts. Micropal in the UK and Morningstar in the US go to considerable lengths to point out that their star-ratings are not intended as "predictive" devices. The trouble is the marketing departments of unit trust companies tend to overlook this when designing the ads!

So is a five-star fund better than a one-star fund? Probably it is, but you need to do far more than pick a unit trust simply on past performance, risk-adjusted or not.

q Peter Jeffreys is managing director of Fund Research Limited, an independent company engaged in qualitative appraisal and rating of investment funds.