Long read: Does the Two-Pot System mean it’s time for capped, asset-based administration fees?A deep dive into the impact of the proposed Two-Pot System and Reg 38 preservation on pension-fund administration fees, and why capped asset-based fees may be the most equitable way to recover future retirement-fund admin fees.Article by Old Mutual Corporate Consultants | Date: 18 October 2023 | Read time: 9 mins

Old Mutual Corporate Consultants have reviewed the impact of the expected Two-Pot System and Reg 38 preservation on pension-fund administration fees. The thinking is that capped asset-based fees may be the most equitable way to recover retirement-fund administration fees when the new system takes effect. Here’s why.

Retail vs Institutional – a history

Apart from the obvious difference that retail products are selected by individuals and institutional funds are selected by employers for their employees, Retail Preservation Funds and Retirement Annuities (Retail Funds) and Occupational Retirement Funds (Institutional Funds) differ in two main ways:

1. Retail Funds typically have minimum investment amounts (either lump sum or ongoing), whereas Institutional Funds don’t. This is broadly because of income tax laws, which require that, if an Institutional Fund is made available to qualifying employees, all qualifying employees must belong to the Fund no matter how small their contribution amount is, or the tax advantages of the Fund are withdrawn.

Since all members must be accommodated at all wealth levels, no minima can apply at an asset level.

2. Retail Funds have typically charged higher asset-management fees than Institutional Funds. There are a few reasons for this:

• There’s an onus on any administrator to KYC (know your client), and in the case of retail, in contrast to Institutional Funds, Retail Funds must do so individually without any support from a trustworthy aggregator like an employer.

• Because the asset managers typically owned the administrators in the retail space in the past, cross-subsidisation of the cost of administration from the asset-based asset-management fees was easier.

The Collective Investment Schemes Control Act requires that all investors in a unit-trust class be charged the same fee as a percentage of assets. This meant that the percentages were set at a level that made those investors near the minima of asset value commercially viable, or at least tolerable within the bounds of acceptable cross-subsidisation. Nevertheless, the asset-based fee lends itself to the reduction in yield (RiY) concept. A fee of 0.3% per annum is equivalent to a reduction in return (yield) of 0.3% per annum.

• Nowadays, there are more and more asset managers that can’t or will not run administration management committees, or mancos, of their own. Mancos that also provide a platform (or LISP) typically charge an access fee to make use of an external asset manager.

• Clients are more-or-less free to move their investments between providers via section 14 of the Pension Funds Act, which acts as both an incentive to keep retail administration fees broadly in line and comparable with competitors.

It also forces administration teams to be both able and resourced to give effect to transfers at an individual level, which tends to carry a cost that is smoothed out over time by being included in an ongoing asset-based charge.

Section 14 transfers are administratively quite intensive for all parties involved, including the transferor Fund (who can no longer extract fees from the ex-client), the transferee Fund (who have to KYC, too), the FSCA who must approve the transfer, as well as the person trying to move their money (with or without assistance from a financial planner). In short, both direct and indirect costs are carried.

Debit orders, varying bank and beneficiary details must also be handled individually on a case-by-case basis by the administration team.

The above has resulted in retail administration fees tending to be priced between 0.3% and 1% per annum of assets per annum. Depending on asset size and market segment, this is sometimes included as an average in the unit price or in some cases charged explicitly on a sliding scale.

A few mancos apply caps to the rand amount of the fee, especially those that identify High Net Worth Individuals (HNWIs) as their target market – both because their assets are larger and because the administration of larger, stickier assets (i.e., less likely to move their money or withdraw significant proportions) is much more profitable.

The upside of the above differences has been that Institutional Fund administrators have typically favoured “rand per member per month”, or “percentage of salary” or “percentage of contributions” pricing models for the Default packages at least, while in contrast Retail Funds have tended to be bound to “percentage of assets” pricing. (I use Default here in reference to Regulation 37 fund Default portfolios/strategies.)

OMCC has considered the features of the Two-Pot System that may force convergence or harmonisation of the pricing models.

The relationship with the new Two-pot System

In the Two-Pot System as currently proposed, retirement-fund members will no longer be able to cash out all their savings when they change jobs. They may indeed be able to cash out a third of their accumulated contributions annually through the Savings Pot (subject to certain conditions) but as we move past the implementation date of the new system on 1 March 2024, they will certainly have at least one Retirement Pot (under the new system) somewhere, provided they have changed jobs at least once.

The design of Regulation 38 (before Two-Pot) tries to assist and protect members by encouraging them to preserve their benefit in the existing fund or “port” their money between providers when changing jobs, theoretically to gain the benefits of aggregation mentioned above as well as the simplicity of a single savings vehicle.

However, there are other considerations:

1. People changing jobs seldom have the energy or will to complete the paperwork required to port their money to their new fund (assuming they have one), which involves several steps and close observation.

2. In the gig economy, contracting or temporary employment among a variety of employers may involve both Retail and Institutional Retirement Funds, perhaps simultaneously.

3. Temporary unemployment may follow resignation, dismissal or retrenchment, meaning the current administrator must continue to administer the members’ assets, perhaps until retirement.

4. Pensionable salaries are still very common, and seldom 100% of the total cost to company. Combined with low contribution rates, many (potentially a majority) of employed members of Institutional Funds have very small shares of funds at the time that they cease being employed. These amounts may be smaller than the minima that Retail Funds consider viable, yet must be preserved conscientiously in terms of Regulation 38.

The solution until now, as mentioned above, was for members of Institutional Funds to simply take the savings, pay the tax and deploy the capital elsewhere. Under Two-Pot this is specifically prohibited by design, and the result (at least under the Two-Pot proposal in its current form) will be that several pots of savings at several administrators (from the investor’s perspective) are inevitable over time.

What is the cost of institutional administration, anyway?

OMCC inspected the public financial statements for two large umbrella funds to estimate the actual cost of administration, and how the pricing compares to retail asset-based fees.

This analysis is of the largest umbrellas, which have presumably achieved economies of scale driven by retirement reform. The market cost of administration seems to be around 0.3% of assets per annum.

However, this will be somewhat sensitive to asset values and market returns and whether our economy is shrinking or growing, particularly employment. It is also equivalent to a rand amount in the region of R75 (as at 2021) per member per month as the democratic average of administration cost across all the members.

Note that Cost of Admin considers the total cost of the administration for all members (including active, deferred, retired, preservation, deceased, new and former as well as S37D events) – not what was actually charged to any particular member group.

Nevertheless, since retirement funds are by their nature not supposed to be profitable, the cost of administration will ultimately be borne by the members of the fund, where the employer doesn’t directly fund administration costs.

Given the wide range of asset values, an average rand amount divided into those different shares of fund results in vastly different RiYs.

For a member with R10 000 saved, the reduction is in excess of their expected real return, meaning their savings would be entirely consumed in administration fees. On the other hand, members who have accumulated R1 million experience a very low RiY of under one basis point. The member with R10 000 would potentially be better off with an asset-based admin fee, while the member with R1 million has the advantage of the rand fee.

In a fund where contributions are expected in future, an initially high RiY is bearable provided they accumulate to higher amounts.

What happens if you apply rand costs to zero future contributions?

Retail Retirement Funds already have administration as a percentage of assets, and it is included in the total investment charge, and termed reduction in yield. This is ultimately instructive as the Retail Retirement Funds (and unit trust discretionary investments) have to deal with frequent individual member interaction.

Under the Two-Pot System, potentially every member will demand service from the administrators on relatively small cash considerations, similar to the Retail Fund experience.

Frequent ad hoc withdrawals by hundreds of thousands of members will indeed carry a cost that is problematic to keep low and even more problematic to apply as a withdrawal cost since it could amount to a large percentage of the member’s total eligible (accessible) withdrawal amount.

Institutional Funds will not be able to rely on future contributions for these paid-up or deferred members. Section 37C will still apply to these small amounts, however, and the cost per section 37C is not charged against the individual member directly. Some cross-subsidisation is necessary here too.

The result is that the “percentage of salary” and “percentage of contributions” fee models would require revision at the point of becoming a deferred member.

Currently, administrators often apply a reduced rand-based administration fee for deferred or paid-up members. However even a nominal rand-based fee of R20 per month would result in a RiY of 3% p.a. on a fund of R8 000.

These members have the option of withdrawing their money before the Two-Pot System comes into use. In addition, sliding scales are quite complicated to explain clearly to investors. OMCC believes that a capped asset-based fee may be the simpler, fairer alternative.

Retirement-fund fees in new retirement normal

Recognising that Institutional Retirement Funds are going to be met with similar problems as Retail Retirement Funds in future, we believe that retirement funds may be forced to apply some level of asset-based fee in any event, even on Default portfolios.

This would potentially help to strike a balance between reduction in yield on low asset amounts and in some way help to cover the total costs of administration, which will have many more paid-up members with small amounts in the future, without increasing the complexity of administration pricing.

OMCC believes that the asset-based fee should be capped at some rand amount so that some of the benefits of institutional schemes remain. However, it’s clear that compulsory preservation will have the effect of converging retail and institutional pricing structures.

A criticism of asset-based admin fees in the past has always been that the size of assets has little to do with administration. However, in a world of preservation without access and with more frequent access to savings, the cross-subsidisation has to come from somewhere.

OMCC believes that capped asset-based fees may be an imperfect remedy to this social-justice necessity.

Learn more about Old Mutual Corporate Consultants or get in touch here.

Related articles