IMA accused of 'flawed' past performance report

Research published by the Investment Management Association has been criticized as 'flawed' because it suggests consumers can use past performance when selecting an appropriate investment as an indicator of potential future performance.

Stuart Fowler, investment industry consultant of Investment by Design, has accused the Investment Management Association of using data which does not reflect the typical analysis of advisers because the persistency of past performance is based on raw returns rather than risk-adjusted data, and does not include tracker funds.

Based on this second round of research produced by Charles River Associates, the IMA suggests funds from the UK All Companies and UK Equity Income funds across a 20-year period offer higher than random persistency over the long term.

Analysis of all funds over the last 20 years suggests funds which sit in the top quartile have a better than average chance of future top quartile performance, says the IMA.

Likewise, performance persisted over short periods for all four sectors tested - UK All Companies, UK Equity Income, UK Smaller Companies and UK Equity & Bond Income, says the IMA.

At best, however, there is a 45% persistency that funds being in the top quartile over a period will still be in the top quartile for the next period.

Fowler suggests CRA's raw analysis of around 940 funds between 1981 and 2001 has focused on the absolute returns of funds in some cases and taken some charges out in others.

"The argument [the IMA gives], that it replicates how investors actually use past performance, is appealing but plain wrong," says Fowler.

"Firstly investors and professional advisers do not pick 50 or more funds at once, and secondly, they are highly likely today to be benefiting from some kind of risk-adjusted selection process, such as the rating agencies' stars and rankings or the selection process of an IFA's 'best advice' panel. Indeed, the FSA's suitability regulations require some form of risk matching by advisers. For both advisers and self-directed investors, risk differentiation is commonly based on standard deviation or the Sharpe ratio, neither of which Charles River Associates even mention," he adds.

In a study of UK investment funds between 1981 and 2001, the IMA commissioned research to analyse raw returns, as consumers are more likely to look at returns on this basis.

CRA evidence highlights these two of the largest UK sectors showed performance and past performance "broadly persisted" in the 21-year period, says Richard Saunders, chief executive of the IMA.

"The IMA has always believed that past performance data can help investors and their advisers. This research, far more thorough than any previous study, shows conclusively that this belief is justified by the facts.

"These numbers do not show that picking last year's winner guarantees you out-performance next year. But they do suggest that, on average, past performance relative to peer group has a tendency to carry forward into the future, for both strong and weak performance. Thus, although past performance should never be the sole reason for choosing an investment fund, it cannot be ignored."

According to the IMA, if performance were based purely on random selection, at least 25% of funds could expect to achieve top quartile performance in that time, however, evidence shows persistent performance figures were in fact consistently higher.

The table below, compiled by CRA, shows the probability of a top quartile fund remaining in the top quartile, after initial and annual charges.

But Fowler still disputes these findings again, arguing an investor might need to invest in around 50 funds at a time in order to achieve these findings, rather than the one to three funds they might hold at any one time.

"Charles River should have calculated how many past top quartile funds you need to own in order to be reasonably sure of a better than random outcome. They needed to show whether holding the tiny number of funds most investors actually hold chances any different from random. You will not find any of this basic distribution analysis in the paper," continues Fowler.

To reader the IMA's full report - Performance persistence in UK equity funds - An empirical analysis" - click thru the right-hand link.

• Tweet  
• Facebook  
• LinkedIn  
• Google plus  
• Send to