The industry is constantly debating how to measure performance, and too often the arguments thrown up are designed to show one or other category of funds in a favourable light.

It is generally accepted that returns of diversified options should be quoted net of tax. Tax applies at the fund (option) level and has the same percentage impact on the returns of every member invested in that fund. (We’re talking here about members in the taxable, or accumulation division of the fund, since pension members’ investment earnings are tax-free.) Tax is an unavoidable cost to the member for having their money invested, and they bear that cost indirectly and automatically because tax is accounted for when unit prices are calculated and declared.

The same logic applies to quoting returns net of investment fees. Again, we are talking about the indirect cost borne by the member for having their money invested. Clearly, for a member, it would make no sense to compare an option’s pre-fee returns, because fees are an integral part of the end result.

Where the debate goes off the rails is the suggestion by some parties (including Cooper) that returns should be reported after deducting non-investment fees – administration and adviser commissions – as well as investment fees. The argument is that members want to know what return they’re getting after all fees and tax. The problems with this measure are:

1. Once you start deducting non-investment fees, you are no longer measuring an option’s investment performance. These other fees pay for other services – administration, marketing, education, member communications, governance etc – that vary across funds and have nothing to do with how well the fund is investing its members’ money.
2. The services paid for by these non-investment fees are impossible to value objectively. For example, one member may be a heavy user of educational material and frequently use the fund’s online calculators etc, while another may never use these services. Clearly, they are going to put different values on these services and the price they pay for them by way of administration fees. These are value judgements, not objective measures.
3. It is difficult to produce any fair or relevant measure where non-investment fees are charged as a fixed dollar amount, eg a member fee of $1.50 per week. This is typically the case with not-for-profit funds. We’ll demonstrate this point by way of example later, but first we’ll comment on the current state of play.

Table 1 summarises what happens in the industry today – how not-for-profit funds (NFP) and retail funds report their returns.

To demonstrate what difference their practices make, we’ve assumed a common $50,000 account balance, a gross return of 10% for the year in question and a typical effective tax rate of 6% for a growth portfolio.

Currently, almost all not-for-profit funds publish their returns net of investment fees and tax only. They do not deduct the dollar-based member fee (administration fee) because in percentage terms it is different for each account balance. And, of course, they don’t deduct adviser commissions because they don’t pay any.

Retail funds generally abide by the standards set by their representative body, the Financial Services Council. FSC Standard No 6 requires them to deduct all fees when reporting returns, including (the maximum) administration fees and ongoing adviser commissions. So even though their investment fees are, on average, lower than those of not-for-profit funds, the net returns they declare are significantly lower (1.27% lower in this example) because of the non-investment fees (administration fee and adviser commission).

[Note: At Chant West, we want to compare the performance of all investment options on a consistent, ‘apples with apples’ basis. For this reason, we have arranged for retail funds to provide us with returns on the same basis as not-for-profit funds, ie after deducting investment fees and tax only.]

Table 2 takes a look into the future. From 1 July 2012, under the Government’s Future of Financial Advice reforms, retail funds will be required to separate their product fees from adviser fees. This means that adviser commissions will no longer be deducted from returns, so by adding that element back in, the returns of investment options in retail funds will look better.

At the same time, a fair comparison would require not-for-profit funds to deduct their member fees (administration fees). These are generally a fixed dollar amount, so as a percentage they will vary depending on the account balance. We have assumed a fee of $100 per annum.

The end result is that the performance of not-for-profit and retail funds will converge, simply because of a change in reporting. In our example, instead of the not-for-profit fund ending up 1.27% ahead, the performance gap shrinks to 0.52%. This is before any differences in investment returns are taken into account. Of course, historically, we know that NFPs have outperformed retail funds by between 1% and 1.25% per annum (ie before taking into account non-investment fees).

We made the point earlier that a fixed dollar-based fee (eg $1.50 per week member fee) represents a different percentage figure when applied to different account balances.

Chart 1 illustrates that clearly. The standard template figure of $50,000 is something of an arbitrary amount. In fact, the average industry fund balance is about half that. As the chart shows, at a relatively small account balance of, say, $10,000, a member fee of $100 per annum starts to look quite substantial (1%). Indeed, it turns the performance advantage around so that, on a ‘net of all fees and tax’ basis, not-for-profit funds actually produce lower returns than retail funds (0.23% lower compared with a 0.52% advantage for a $50,000 account balance).

Chart 1 is getting closer to the real world, while Table 3 takes us closer still. Here, we have taken six large, not-for-profit funds, assumed a 10% return net of investment fees and tax and a $100 per annum member fee, and applied those assumptions to the funds’ actual average account balances.

The results are easily open to misinterpretation, and that is the danger of reporting returns in this way. Taking the figures at face value, a member could be forgiven for thinking that REST’s investment performance is substandard at 8.8%. Yet our model assumes the same return and the same fees for all six funds. The only difference is that REST’s average account balance is so low. That affects the published returns, but it doesn’t affect any individual member’s returns – and that’s the point.

The only person who receives the published return under this method is someone whose account balance is exactly the average for that fund. A higher or lower account balance, or even a difference in cashflow during the year, will produce a different result. As a method of comparing investment performance, this approach is deeply flawed.

Let’s assume that our member is currently in a different fund, maybe changing jobs, and is considering REST as an option under choice of fund. If they were relying on published performance data, and that data had to be compiled in the same way as our table (as it would under the Cooper proposals) they are unlikely to choose REST.

But knowing what we do about REST, which in reality has been one of the best-performing funds for many years, it would probably be a very good fund to choose. But how would our member know that? The point is that performance figures calculated net of all fees can be misleading and can lead to poor decision-making if they are not based on a common account balance. That is why our preference is to show performance net of investment fees and tax only. That is what we believe members should be focusing on. It is certainly what trustees and investment committees are focusing on.

If our view doesn’t prevail, and non-investment fees are to be taken into account, then the only sensible way to do so is to apply the calculation to an assumed account balance of, say, $50,000. That provides some basis for fair comparison between funds.

Some proponents of the ‘what the member actually gets’ approach also argue that, for those funds that still carry investment reserves and declare crediting rates rather than earning rates, it is the crediting rate that should be used in performance reporting. We disagree.

Chart 2 shows how one fund’s default option crediting rate was 0.5% per annum greater than its earning rate over the seven years to June 2010. If we had reported the crediting rate, it would have ranked 25th in our Performance Survey ie about median performance. By reporting the fund’s earning rate, it ranked 39th, just above the lower quartile.

The practice of reserving is used to smooth out returns, but in the process it results in a loss of transparency. It prevents members from seeing what the fund has actually earned, so they have no way of judging whether its performance is good or bad. And, while it smooths out returns, it also shields members from the reality of where their money is invested and how it reacts to movements in investment markets. For all these reasons, we believe that true earning rates, not crediting rates, should be used in performance measurement.