Do Fund Managers Earn Their Fees?

Charlie Nelson
director foreseechange
March 2003

The answer is no – but they could do better.

In 1999, Barrie Dunstan reported a survey that I had undertaken (when I was at ACNielsen) concerning the asset mix preferences of superannuation fund members (“Survey unearths aggressive stance on investment choice”, Australian Financial Review, April 12 1999).

In general terms, the survey found that members wanted about 60% in growth assets and 40% in defensive assets.  However, the average member of a super fund was much more oriented towards property as an asset class than fund managers.  Super fund members wanted a roughly 50/50 split between shares and property while fund managers typically favoured a 5:1 ratio of shares to property.

Barrie Dunstan found that assuming benchmark returns for the previous five years to the end of December, the fund members preferred portfolio would have returned 8.37% if the property allocation went to direct property and 9.32% if the property allocation went to listed property trusts.  This compares with returns of the average asset mix used by professional fund managers of 8.8% from the general managed funds and 9.2% from the slightly more aggressive growth funds.

In other words, Barrie Dunstan concluded, on average the sample of members would have matched, or almost matched, the best available median results from a managed pooled fund, and done so with an asset mix with which they felt comfortable.

Since 1999, however, property has performed better than shares and the fund members would be ahead if fund managers used the allocation preferred by members rather than their “expert” mix weighted towards shares rather than property.

The survey has been repeated in 2000 and the asset mix preferred by fund managers varied little over several waves – the proportion allocated to shares fell only slightly immediately after the April 2000 “tech wreck” and recovered to earlier levels after three months.

The average allocation of fund members compared with that of a typical balanced fund is shown in the chart below. 

To compare the performance of these mixes, we have assumed that all funds allocated are to Australian assets only and that the property component is put in listed property trusts.  The former assumption is made because international shares have the additional volatility caused by exchange rate variations and because the survey did not ask about international shares.  This assumption makes little difference to the conclusion reached.

Based on 13 years performance of asset classes (financial years 1989/90 to 2001/02) the average annual return of the two mixes would have been: 

• Fund Members 10.4%;
• Fund Managers 10.1%.

Thus, fund members have every reason to question the performance of fund managers – they have subtracted, rather than added, value over the 13 year period.

Furthermore, unless there is a substantially recovery in the Australian share market before 30 June 2003, fund members would have outperformed fund managers by an even wider margin and will have even more reason for dissatisfaction.

Over the 13 year period, the mix with the best return was 100% in listed property trusts (average 12.4% per year) – the very asset class that fund members wanted more of relative to fund managers.

It is possible to do even better with dynamic mixes.  For example, if we could predict which asset class would do the best in the next year and allocated 100% of our funds to that class, the 13 year return would have been 17.2% per year on average.

While it is difficult to predict this accurately, if we had put 100% of our funds this year into the asset class that performed best last year, the 13 year return would have been 13.2% per year on average.  This is better than the best fixed allocation.

If we took a contrarian view and allocated 100% of our funds this year to the asset class that performed worst last year, the 13 year return would have been 10.5% per year on average – still better than the typical balanced fund.

While I have conveniently ignored switching costs, which would be very low for the professionals, it is clear that fund managers could do much better.  They should have had a higher proportion of funds in property and they should have had a proportion of the funds allocated on a dynamic basis.  This would have enabled them to achieve better returns in the long term and less years of negative returns.

Fund members preferences will be measured again in 2003, along with risk/return tradeoff measures.  To receive this information, subscribe to Prophets Profit.  

Back