Year to 31-Mar	Return
2002	-2.49%
2003	-11.21%
2004	20.08%
2005	4.56%
2006	11.70%

The next line of the table makes that claim of 6.99% again, but if you compound the 5-year returns together, you in fact get a total of 21.42% over 5 years, which is a compound average return (CAR) of 3.96% per year. The figure used in the report was (as the small print discloses) an internal rate of return (IRR) or "dollar-weighted" figure which was higher because there was more money in the scheme in the later, positive years, than there was in the earlier, negative years of the scheme. You can't depend on that pattern repeating itself, but even if exactly the same sequence of annual returns occurred, for those in the scheme from the beginning, the weight of the money already invested would drag the IRR back towards the compound average. The CAR is the fair measure to use.

In any event, 5 years is not a long time in market terms. We have no reason to believe that the gross average performance of all funds will be any worse or better than the underlying markets as a whole, but the net performance is what matters, and what the report completely avoids mentioning is the annual impact of fees and expenses, and how much lower that impact would be if people were allowed to invest their own money directly in the markets or in index funds.

Expenses running at 2.5% per year

In Jun-04, the MPFA published the Code on Disclosure for MPF Investment Funds, but most of its requirements did not kick in until recently. Fund expense ratios (FERs) were calculated for any financial year which commences after 31-Dec-04. Some fund providers, such as HSBC, use a June year-end, so the first figures, for the year to Jun-06, have only recently appeared.  The Code requires an "On-going Cost Illustration" in a prescribed format which shows the FER for the latest year, and the dollar cost of such expenses after 5 years for each $1,000 invested, assuming a gross return of 5% per year.

Some, but not all, trustees publish these figures on their web sites. The table below shows, for each trustee (in alphabetic order), the lowest and highest FERs of the funds in their disclosure, and the projected 5-year cost for each $1,000 invested:

			5-year cost on $1000
Trustee	Min FER	Max FER	Min $	Max $
Bank Consortium Trust	1.79%	2.04%	100	114
BOCI-Prudential	1.61%	1.75%	90	98
Fidelity	2.24%	2.53%	123	139
HSBC/Hang Seng Bank	2.00%	2.81%	111	154
ING Pension Trust	2.05%	2.81%	114	154
Manulife	2.32%	2.33%	128	129
Principal Class A	2.23%	2.28%	123	126
Sun Life	2.07%	2.11%	115	117

Note: BCT maximum FER is adjusted for refund of a maintenance fee

Under the Code, those expense ratios are actually understated, because they exclude transaction costs on the underlying investments, such as stamp duty and brokerage costs on buying or selling equities. If these aren't expenses, then what are they? The more actively a fund turns over its portfolio, the higher those costs will be, eating into the returns on the funds relative to the markets in which they invest. You can probably add another 0.2% to those FERs to take account of underlying transaction costs, so you'll be looking at a mid-point of about 2.5% per year, which after 5 years will cost you about $138 for every $1,000 invested.

That's just 5 years! If the MPF continues, over a lifetime of contributions, the performance of all funds will be crushed by the expenses.

The lack of competition on MPF fees should be no surprise, because it is the employers who choose the fund providers, not the employees, and the only way to change providers is to change jobs, in which case you can move the money to a fund provider of your choice, or to your new employer's fund provider, but in the process you may incur redemption charges.

To stimulate competition, the MPFA should publish on its web site the expense ratios and complete historic performance figures of every fund in the MPF schemes, in an easily searchable database with comparison tables by fund type, so that the public can more readily find out which are the highest-cost and lowest-cost providers for each type of fund. At the moment, you can hunt all over the internet and still not find that information. All the MPFA has produced is clumsy Excel spreadsheets of unit prices, one for each of the last 12 months, but no performance or expense data. The MPF, as long as it exists, is always going to be expensive to administer, and those costs are going to be built into the funds, but competition would help limit the damage to employee's wealth.

Incidentally, the transparency in the mutual fund industry (outside of the MPF) is even worse, as the SFC's Code on Unit Trusts and Mutual Funds contains no disclosure requirement on expense ratios in mutual fund advertisements reports or fact sheets. We call on the SFC to address this problem, and then to publish on its web site a database of mutual fund returns and expense ratios. Indeed it would be better if the SFC and MPF worked together on this so that people can compare MPF funds with non-MPF funds.

Past investment performance may not be a guide to the future, but past expense ratios are.