Investment insider: Top of the form one year, dunce the next

COUNTLESS books have been written on investment performance. But few conclude - let alone offer statistical evidence - that the past performance of individual unit trusts, PEPs and the like is any guide to the future.

This does nothing to deter the annual ritual of performance awards for last year's top-performing unit trust, which precedes a flood of advertisements announcing the "Fund of the Year", the "Fund Manager of the Year", the "Fund Management Group of the Year", and endless derivatives thereof.

The case for investing in the latest "Fund of the Year" is, however, weak. Only three out of 21 winners for 1994 came close to repeating their success in 1995. These were Morgan Grenfell European Growth (first among its peer group in 1994 and second in 1995), Hill Samuel US Smaller Companies (first in 1994 and sixth last year) and BG Convertible (top in 1994, second in 1995). Of the other 18 winners in 1994 only five achieved above-average performance in 1995, 13 were below average, and nine of these - almost half the total - finished in the bottom quarter.

But if last year's winners are a poor guide, what about last year's losers? Specialist trusts apart, just over half the losers failed to beat the average fund of their type in 1995.

Furthermore, success does not improve with time. Studies suggest that unit trusts have no better than a one-in-five chance of repeating top-quarter performance over one five-year period. More sophisticated statistical techniques - adjusting for the volatility of the investment, for example - improves these odds slightly. But it still leaves them short of what could be called statistically significant.

Does this mean that studying past performance is a waste of time and that using a pin is a better method?

Despite the statistics, the answer to both questions is no. But it does demand that you should look at performance in a different way to that which is presented to you. Never look at a single period of time; a better starting point is consistency of performance over separate investment periods.

Second, look behind the numbers; try to find out whether fund managers achieved their results by luck or judgement - next week I will tell you how the experts do it. Find out if the fund manager who achieved the success is still in the job. All too often, the star manager is spirited away - either by his existing employer to manage a multi-billion pound pension fund - or by another employer who entices him away with a golden hello, or the promise of nirvana in some other guise.

Many management companies argue that the investment process is more important than the individual. In some respects they are right but, all too often, when there is a change of fund manager, performance tails off. Try to find out if the fund manager's boss has changed. Chief investment officers are hot numbers and are just as vulnerable to financial inducements as their minions.

Finally, don't worry if you cannot undertake this research yourself; there are many independent financial advisers who can - indeed it is part of their due diligence or "best advice" process to do so.

Next week, I will look at the subject of risk, not in the mumbo-jumbo language of the investment industry (standard deviations, alphas, betas and R-squared) but in language you will understand, in the hope that the industry will listen to what you want.

I will also discuss some of the techniques you can use to look behind those past performance numbers.

o Peter Jeffreys is managing director of Fund Research, a leading investment analyst.