·       Investment returns net of expenses  for all funds averaged 6.68 per cent per year 

The paper shows significant differences by type of fund, ranging from corporate funds with an average net return of 6.96 per cent to retail funds with an average net return of 4.51 per cent per annum.  There is evidence that larger funds have higher returns than smaller funds.

·       Estimated annual expenses for all funds were 1.28 per cent per year on equal weighted basis, and 1.07 per cent per year on an asset weighted basis

There is evidence of scale economies in superannuation:  funds with over $ 500 million in assets generated an estimated 1.08 per cent in annual expenses, against an estimated 1.37 per cent per annum for funds with less than $10 million in assets.  There is no evidence that superannuation expense ratios have decreased in recent years.  There are substantial differences in estimated expenses by industry sector, with public sector funds the lowest on an asset weighted basis at 0.58 per cent per annum and retail funds the highest at 1.32 per cent per annum.

·       There is substantial variation in volatility between fund types

The paper revealed a substantial difference between fund types, with retail funds returning the lowest volatility and corporate funds the highest volatility.

·        Retail funds have produced the lowest net returns in each of the past seven years

Given their apparent lower investment risks, retail funds would have been expected to under-perform in the good investment years from 1996 to 2000, and outperform other sectors in the less rewarding 2001 and 2002 years.  The fact that retail funds have underperformed in each year is an area for future APRA research.

·       There is no evidence that funds which incur higher expenses generate better investment outcomes for superannuation fund members

The data, in fact, indicate that the higher the estimated expenses incurred, the lower the expected investment return.

In preparing the paper, an in-house APRA team headed by Dr Neil Esho compared total returns, estimated expenses and volatility of returns across corporate, industry, public sector and retail funds of varying sizes.

Mr Charles Littrell, APRAs Executive General Manager  Policy, Research and Consulting, said that APRA will use the paper and other relevant data in its supervision of superannuation funds. 

By constructing more finely detailed investment benchmarks based on sectors and fund sizes rather than against the entire industry, we can more easily identify unusual performance from individual funds that would warrant more attention, he said.

APRA has released for consultation expanded data collection forms intended to commence from June 2004.  The regulator will use the expanded data set to further refine its supervisory analysis and oversight of the Australian superannuation industry, particularly at the individual fund level.

There are some significant data areas relevant to superannuation investment performance, which are not currently collected by APRA. 

Mr Littrell added: Notably, we lack sufficient data on the composition of investment portfolios by asset class and do not yet collect sufficient data on expense composition or any data on entry or exit expenses.  The expanded data set will assist us in achieving these goals.