Benchmarking investment performance – the unintended consequences of Government proposals
- On 29/10/2020
- investments, performance
In the 2020 Budget earlier this month, the major showcased reform was in superannuation, namely the Your Future, Your Super Package.¹ This proposed package would implement some of the recommendations made by the Productivity Commission (PC) in its major three-stage review of superannuation over the four years to November 2019.
One of the proposed changes is to monitor the investment performance of superannuation funds relative to benchmarks based on listed assets and to penalise products (and trustees) which under-perform. While the aim of raising overall performance is sound, the implementation is flawed and may lead to unintended consequences unless significant changes are made.
Holding funds to account for underperformance
The stated government objectives are to:
- Maximise the savings for retirement of all members by protecting them from underperforming superannuation products.
- Ensure funds are maximising net (after fees) investment returns.
- Hold funds accountable for the outcomes they deliver to members.
- Make members aware when their superannuation product is underperforming.
The proposed system begins with the default MySuper products which will be measured annually from July 2021. Those funds deemed to be underperforming against a new benchmark by more than 50 bps p.a. over eight years will need to advise their members by October each year. From July 2022, those MySuper options which have underperformed two years running will be unable to take on new members for at least 12 months until their performance has improved on the formula adopted.
In July 2022, the system will also be extended to other Trustee-Directed Products, using the new acronym of TDP, and covering all superannuation investment strategies (options) which are multi-sector. This will cover options such as Conservative or High Growth, but not options covering a single class such as Australian Equities or Property.
At this stage, retirement products are not included but we assume the system with be extended to them in future years. We expect this will be part of the Retirement Income Covenant legislation which has been delayed pending release of the report of the Retirement Income Review.
The PC undertook extensive research on past performance of the superannuation industry and noted the wide variation in performance between funds and between sectors. It noted that underperforming funds were a drag on retirement savings and members of those funds “effectively lost the SG lottery”. Everyone agrees that underperforming funds/products needed to improve or be removed from the marketplace, though there are a wide range of ideas on how to weed them out.
The PC report focused on underperformance and made no comment on the extraordinary outperformance of many Australian superannuation funds over extended periods of time, let alone the asset allocation decisions which were integral to achieving this performance. Some funds have outperformed CPI by 4% to 5% over periods of 30 or more years. Had the PC set a benchmark of (say) CPI + 3.5% over rolling 10-year periods, it would have set a reasonable bar without needing a new metric for funds. Further, it would not have interfered in any of the current asset allocation decisions made by successful funds.
To measure underperformance, the PC undertook considerable research on past results and concluded the benchmark for measuring performance should consider the counterfactual position of investing entirely in listed markets via indices (in other words, eliminating the impact of any skill or of accessing a wider opportunity set). Applying the fund’s strategic allocation to asset classes, a passive performance could be derived from which fees and taxes are deducted. This would be the outcome if a fund simply undertook passive management.
Notwithstanding that funds (encouraged by APRA) measure their MySuper products against a target of CPI + X% over rolling ten-year periods, the PC chose a different measure. It also used a measurement period of eight years without providing any evidence as to why this is a superior period over the industry norm of 10 years. We have been unable to find any empirical evidence for preferring such a time horizon.
The PC determined that any fund more than 50 basis points a year (0.5%) below their own constructed benchmark would be deemed to be underperforming. The threshold of 50 bps is very tight and may result in artificially high levels of apparent underperformance in some investment periods. For example, some infrastructure assets will generate poor returns following the current pandemic, but these assets have mostly performed well over long periods and are not easily traded, making funds likely to hold them awaiting a return to better days.
The Your Future, Your Super package accepts the PC’s formula and APRA will now conduct the measurement. Based on current performance figures, from October next year, several funds will need to advise members that they have underperformed. The funds will have a serious dilemma. The following year, they will need to improve performance to avoid a consecutive year of underperformance. Yet, seven of the eight years measured that year will be the same, giving them unavoidable past performance, so they must hope that the new year is much better than the year dropped out.
Most of these funds will be unable to turn their eight-year performance around enough in one year, so they will need to set up different structures to accommodate new members. This will be very messy for all concerned. The government appears to hope that these funds will exit the industry.
Quirks of new measurement structure
Most funds will want to keep their current targets based on CPI + X% over rolling 10-year periods. This is an established metric, and it is easy for members to understand. Further, the target can be reached by different strategic asset allocations.
If the funds stick to their current investment goals and objectives, they will then have two different formulae to measure, and to explain to members.
There are many ways that funds will deviate from the new benchmark or encounter issues:
- Most market indices are cap-weighted whereas funds should seek industries which will grow in future rather than those that grew in the past. As an example, if the fund did not want to be heavily invested in banks and mining companies, it would need to vary from the Australian index (ASX 300). Similarly if a fund wishes to avoid areas of the market which it considers to be over-valued, then it needs to be able to stay the course if they miss out temporarily from these shares becoming even more over-valued, as happened with some technology shares in the late 1990s before the tech-wreck of 2000-2003.
- Funds can use derivatives to change their exposure – and this will alter their returns.
- Funds can seek franking credits to maximise after-tax returns. They will participate in off-market buy-backs as these provide strong after-tax returns.
- Funds investing in infrastructure will be measured against a benchmark which could be quite different from their holdings. The chosen benchmark for Australian infrastructure is the FTSE Developed Core Infrastructure Index hedged to AUD. The structure is different from local holdings and from the infrastructure asset class, as the most highly prized infrastructure assets are often tightly held outside the listed environment. Further, the benchmark fee is 0.26% which, following RG 97, is about one-third of the fee for unlisted infrastructure stocks.
- Lifestage products have multiple asset allocations. Each will be measured separately – and theoretically a fund could find itself underperforming in one area. Sorry, you can’t join our default for people under age 30 this year, but why not join the 30 to 45 group instead!
- Some funds will revert to bland indexed investments thus avoiding the chance of underperformance – but forgoing the opportunities for higher performance from unlisted investments.
- While we know that past performance is no guide to the future (and ASIC requires funds to state this prominently), funds cannot change the first six years on the forthcoming eight-year test. Even if a fund totally revamps its strategic asset allocation, moves some classes to passive and brings some funds in-house to cut costs, it will still have this past performance within its measured returns.
- Some funds have had poor returns (after fees) and they will have no option but to wind up. They will not be able to recover in two years when 75% of their return will still be poor. This applies even though new members will not receive the past performance. They could start again but are more likely to merge with a high-performing fund (even a very small one) and SFT into that option to preserve the good performance.
Finally, we know that the strategic asset allocation is one of the largest contributors to investment performance. Yet, this is not being measured! It is theoretically possible that a fund investing in volatile assets will provide a sound return and deliver on its member target yet fail the benchmark test in some periods. Conversely, a fund could invest entirely in cash and not be at risk of measured underperformance. Yet, it would deliver a poor retirement outcome. This is an extreme example of some unintended consequences, and the reality is that few members would choose this option, but it shows how the new process could distort behaviour.
Impact of new measures
Things could have been even worse – the government could have adopted the PC’s flawed “best in show” concept of allocating new members to superannuation funds using this flawed methodology to determine which products are the best. Even so, the proposals have potential unintended consequences which are clearly identifiable and with high impact.
The over-arching effect of the proposed measures would likely be to pressure funds to forgo opportunities for long-term outperformance to mitigate the risks of underperformance against a nominated benchmark.
It is likely that many funds will become passive on Australian shares, to a greater extent than would otherwise be optimal. Surviving funds will be large and many won’t be able to deviate much from the index anyway. They will still want exposure to franking credits, so will keep a sizeable portion of Australian shares. One of the outcomes of this is that a much higher percentage of the Australian share market will become passive. In time, prices will be set by the trading activities of foreign investors, the retail investors and SMSFs!
More investments will be brought in-house as funds seek to reduce costs. The allocation to higher-performing unlisted assets might be reduced to minimise apparent failure against the benchmark. Technically, these assets have much higher tracking errors, so they will have different results to the benchmark both better and worse in different periods. The shift away from higher performing infrastructure would be a poor development for members and the overall economy alike.
Some funds could fail (even slightly) on investment performance yet do well in other areas such as retirement or life insurance. Unlike the APRA heat-map which covers several areas, this is a blunt single-issue measure and there does not appear to be any leeway for tolerance.
Many retail products have slowly improved their MySuper products by cutting fees and reviewing asset allocation. Some of these products will not survive despite their efforts to improve value as they will be weighed down by past poor performance.
The new system will lead to changed behaviour. We hope funds stay the course and continue to seek alpha in unlisted asset classes. However, they will have to watch the benchmarks carefully and might use derivatives to protect against any major deviation from the fund’s own assets.
Impacts will not include identifying and addressing situations where funds are under-performing due to issues with their strategic asset allocations. This means that the most likely source of material under-performance would, at best, remain unresolved. Indeed, it may even be worsened by disincentivising allocations to the unlisted assets which have contributed so much to the world-class returns of successful Australian superannuation funds and other leading long-term investors such as the Future Fund.