Superannuation: Performing for all members?

My thanks to the McKell Institute for this opportunity to speak today on our superannuation Inquiry. This is our first (non-media) speaking engagement following the release of our draft report last week. And it is timely as we now enter an important phase of consultation on our draft report – its findings and recommendations.

Whilst the Institute may be relatively young as a think tank – especially here in Victoria – it has already made many thoughtful contributions to public policy thinking in the arenas that matter – health, education, housing, infrastructure and, not least, productivity.

And it’s an honour to follow on from the eminent speakers the Institute has previously hosted at this forum.

Of recent note for the Commission, a speech given by Tim Pallas, the Victorian Treasurer, on federalism. His thoughtful yet frank insights afforded a contemporary perspective on federal state relations whilst completing our recent Inquiry on the GST distribution.

I would like to think the Productivity Commission shares some historical common ground with the Institute’s namesake, Sir William McKell. The historical records plainly reveal that Sir William was driven to provide a better standard of living for ordinary Australians. It is through this lens of the wellbeing of all Australians (not just the vocal few) that we have approached our current inquiry on the super system’s performance – by looking at how the members themselves are faring, and how we can make outcomes better for Australians with super when they retire.

There is perhaps one thing that all participants in our super inquiry can agree upon: there is much content in our draft super report – one commentator suggested more than a Netflix Original. So today, in the time we have, I will focus on its main findings and recommendations. You have probably read and seen much already in the media as the anticipated debate unfolds, much of which has been accurate, but there a few things that haven’t got as much airtime as others and which I’ll draw attention to today.

I’ll briefly canter through:

• our approach to the Inquiry, with some important social context on the system today – it’s a story of an evolving social context;
• the outcomes members are getting – it’s not a world of averages; and
• our proposed package of improvements.

Our draft report is the culmination of much endeavour and consultation by the Commission. It is now more than two years since the Government tasked us to assess the performance of our $2.6 trillion super system. We knew this would be a gargantuan public policy assignment. Its performance matters today for the retirement savings and thus wellbeing of 14.8 million Australians and their families. And those savings represent a fifth of the wealth of all Australian households. Only the family home is of greater wealth value to Australians today.

Some of you may also recall the Commission’s endeavour in retirement incomes commenced even earlier in 2015 with two Commission-initiated reports – Post Retirement Super and Housing Decisions of Older Australians. Indeed these reports had a hand in the Commission being tasked with our current Inquiry.

For our current Inquiry, the Government afforded us the rare opportunity in public policy today to hasten slowly. And for that we are fortunate. It allowed the Commission’s endeavour to span a three stage investigation. In stage 1 we developed criteria for assessing the performance (efficiency and competitiveness) of the super system. Then in stage 2 we developed alternative models for allocating default fund members to products. This was designed to be a ‘top drawer’ report as an input to our current stage 3 inquiry.

We released a draft report for stage 2 last year, and the Government subsequently agreed for us to complete stage 2 alongside the final stage 3 Inquiry.

Since we began the three‑stage process in early 2016, we have received well over 300 submissions, met with over 120 parties in both Australia and overseas, held five technical roundtables and conducted a round of public hearings. And we will continue to consult widely as we finalise our report by the end of this year.

Much of the analysis in our draft report is new and novel. No-one has assessed the performance (through the lens of efficiency and competitiveness) of a superannuation or pension system in its entirety before, so we had to venture into new territory. Now many may emit a yawn in thinking of our stage 1 study – developing an assessment framework. It consisted of 5 system level objectives, 22 assessment criteria, and 79 points of evidence to gather and analyse. But this stage was critical to the draft report we released last week. For it gave us the latitude to undertake in-depth and novel analysis to inform our findings. For good public policy can rarely emerge unless it is coaxed through a comprehensive and robustly harvested evidence base. So we worked out in stage 1 what we had to harvest and how. And we also identified where we had to extract or develop new data to inform that evidence base and analysis. And in a calmer environ than we find ourselves in today.

One of the key innovations was to construct investment benchmark portfolios. To allow us to compare performance, agnostic of asset allocation. The benchmark portfolios are weighted averages of financial market index returns, with the weights determined by the asset allocation of the fund, segment or product we are benchmarking. We then made adjustments for investment fees, admin fees and taxes that reflect the experience of super funds.

So our benchmarks are a measure of whether funds are adding value against the market’s performance. And by adjusting for differences in asset allocation, we are able to do what no-one has managed to do before – compare apples and zebras. Now, such benchmark analysis is not new and novel for the super funds. Indeed to undertake basic performance attribution should be bread and butter. But what is new and novel is to do it across an entire super system (by segment, fund and product). And the sweat that accompanied our endeavour was a function of the very poor data we had to work with.

Importantly, we are using our benchmark to identify where there is long-term underperformance in the system. We define underperformance as being 25 basis points or more below the benchmark – a margin allowing us to err on the side of affording funds the ‘benefit of the doubt’ and to reflect the modicum of uncertainty in the benchmarks arising from data issues, all of which we document in our draft report.

Another novel feature of our report is its cameo analysis – we used these to show how the performance and problems of the super system can compound over time to add to or erode members’ retirement balances. Most of the cameos assume a full-time worker starting their first job today at age 21 and retiring at 67, in the year 2064. But we also undertook a cameo for a 55 year old today, also retiring at age 67. The cameos reveal both the downside of today’s problems and, of equal import, the upside to the problems being removed. And some of the cameos show how regressive some of the system’s failings are for particular cohorts of disadvantaged workers – those with disrupted participation in the workforce.

We conducted econometric analysis to measure the impacts of product proliferation on costs and fees, and used stochastic analysis to look at how these fees affect members’ retirement balances. We also used stochastic analysis to assess the value of life cycle products to members. Today these products represent 30 per cent of MySuper products.

And we have advanced our econometric work on economies of scale, to be released as a post-draft report supplementary paper in August.

So before being able to assess the system’s performance we needed to harvest data. To fill the many data gaps we identified in Stage 1, the Commission undertook 4 surveys. Two member surveys and two fund surveys.

The first member survey was an experimental choice survey of members on how members might behave in choosing a super fund when assisted by a shortlist of good products. This was a key input to our assessment of alternative default models.

The second member survey was broader – to gather evidence about members’ understanding of super and their experiences with the system. This provided insights for example on what members value and the incidence of unintended multiple accounts.

On the fund surveys, the first was to gather data on fund activities and outputs – largely filling gaps in what the regulators collect or where the reporting to the regulators is of poor quality. Alas the overall quality of responses to the survey was disappointing. Many participants skipped questions or provided incomplete data. And responses were particularly poor for questions relating to net returns and fees. This has prevented us from doing some important analysis – international performance comparison and better understanding the systemic difference between the performance of industry and retail segments.

The content of the responses (and the gaps) when viewed across all funds proved evidence in and of itself.

We wrote to all fund CEOs on the day of release of our draft report providing a further opportunity to provide the Commission with the key data needed to complete its analysis.

So what outcomes are the 14.8 million Australians with super (along with their families) getting?

Our draft report presents a pretty mixed report card. The good news is that most members are doing reasonably well. The not so good news is that our $2.6 trillion super system has become an unlucky lottery for many members.

And why an unlucky lottery? The system suffers two fundamental flaws that set the odds against many members:

• Members accumulating unintended multiple accounts, and paying billions of dollars each year in unnecessary fees and insurance premiums.
• Entrenched underperformance, not just in the choice segment but also the default (today’s MySuper) segment.

We found that the odds of being a fund member with these two problems is both too high and highly regressive in its impact – causing greater harm to young people, workers on low incomes and workers in an out of the workforce. These are two awkward truths that need to be addressed. And perhaps it seems well known in the industry. As ex RBA (Reserve Bank of Australia) Governor Bernie Fraser observed last week, ‘The problems have been there for yonks but there has been a hell of a lot of inertia’.

Turning first to unintended multiple accounts: We found that one in three member accounts in the system are unintended multiples – that’s about 10 million accounts. That’s because the super system continues to staple the member’s account to the employer and not the worker. So every time a worker changes job, especially for the two-thirds of members who default when they change job, they typically end up with another super account.

To date, the onus has been on members to proactively consolidate their existing accounts, and we know that many members have failed to do this or left it so late that their balances have been seriously depleted. As we can see from this chart, it’s not just young people either – people in their 40s are most affected. But it would appear that most of these unintended accounts are first opened when people are younger or in the formative stages of their careers.

Our estimates suggest these unintended multiples collectively cost the members who hold them $1.9 billion a year in excess insurance premiums and $690 million in excess administration fees — or $2.6 billion in aggregate each year.

And over a working life, an unintended multiple account can leave a typical member with $51,000 or 6 per cent less to retire with. So a case of less member accounts can mean much more in retirement.

The second big flaw is entrenched underperformance. Our analysis reveals that many members are ending up in underperforming funds. For funds as a whole – so taking account of both default and choice members – we found that over 1 in 4 funds underperformed over the 12 years to 2016. We were constrained by data here, and hence only able to look at those funds that have MySuper products (for technical reasons that are documented in the draft report).

What’s important about this chart is it’s a tailored benchmark – tailored to the asset allocation of each individual fund. Poor performers cannot hide behind differences in their asset allocation from that of other funds in the market.

The 20 underperforming funds represent about one-third of the nearly 15 million member accounts in our dataset. About half of the underperformers are retail funds, and a third industry funds.

But again, it’s not all bad news. About 10 million member accounts (67 per cent of our sample) are in funds that beat their benchmark over the 12-year period. As you would have seen from the media coverage, we also found that not-for-profit funds outperform retail funds on average.

But, it’s the distribution that really matters when it comes to the outcomes for an individual member. Or in simple terms – not all members experience the average. Indeed millions do not given the size of our super system.

This cameo shows just how much the dispersion in member outcomes really matters. Being in a bottom-quartile fund can leave a typical member with $635,000 less (or 53 per cent) when they retire, compared to being in a top-quartile fund.

Looking specifically at the default segment, we took the current set of MySuper products (and their predecessors) and tracked as many as we could back over the past 10 years to see how they have performed over this time.

The top 10 did well – a median return of 5.7 per cent a year for their members. And importantly they hold just over half of default member accounts (some 6.1 million or 55 per cent) and just under half of default assets ($225 billion).

But 26 – or about 1 in 3 – underperformed a benchmark reflecting the average asset allocation of MySuper products. They generated a median return of just 3.9 per cent a year for their members. They account for 1.7 million member accounts.

Of these underperforming MySuper products:

• 12 are in retail funds, 10 industry, 3 corporate and 1 public sector
• 8 are life‑cycle products (which make up nearly a third of all MySuper accounts).

Averages and medians conceal a lot of variation – what matters is where the individual members are. The take-out message from this chart is that there is too much variation in the long-term outcomes that default members are getting.

The large differences in investment performance for MySuper products have significant implications for members. We estimated that being in an underperforming MySuper product can leave a typical new workforce entrant $375,000 or 36 per cent worse off by retirement.

And if all members in the underperforming MySuper products had been in the median top 10 MySuper product, they would have collectively been better off by over $1.3 billion annually (or about $770 each annually, on average).

Turning to the choice segment. Looking at the choice segment was challenging – there are over 40 000 products with much variation between them, and data on many of these products are hard to come by, especially over a 12-year time period.

But we managed to capture 362 of the larger investment options in the system, covering about 13 per cent of assets in the choice segment.

And we found that 1 in 2 underperformed a benchmark tailored to their own asset allocation. About three-quarters of these are offered by retail funds, and a fifth are offered by industry funds. And given the self-reporting (positive) bias here – one could argue this is about as good as it gets for the choice segment even if we were to capture more of this segment through an expanded sample.

While we don’t know how many member accounts this represents, there are 11 million members in the choice segment of the system. So, clearly, many members in choice – as well as default – could be doing a lot better.

We also found problems elsewhere in the super system. And given time today I’ll focus less on the detail and more on the common theme. And that is all these problems have regressive impacts – hurting most those members with lower balances, many of whom happen to be younger members, those on lower incomes through their working lives and those who move in and out of the workforce.

Three of the major ones are:

• Fees erode balances. Australians pay around $30 billion each year in super fees (and that doesn’t include insurance premiums). We found that products with higher fees tend to deliver lower net returns over time, even after we have netted off the fees. And just 0.5 per cent in higher fees can leave a member $100,000 worse off by retirement.
• There are over 40 000 products in accumulation, and comparing them is diabolically hard, even for experts. Members are lost in the weeds of the unhealthy troika of product proliferation, poor disclosure and questionable advice. And the irony of the system is that, if anything, products are most complex during accumulation and most simple in retirement – when the reverse constellation is needed for most members. Moreover, there is a positive relationship between the number of options offered by a fund and the average ratio of their fees to net assets. That delta can result in super balances at retirement being between $140,000 and $230,000 lower.
• Insurance. Many members are being defaulted into insurance products they don’t know about (1 in 4 from our member survey) or that materially erode their retirement balances (by up to $50,000). Many young people (especially under 25s) are paying for insurance they simply don’t need. And many end up with ‘zombie’ policies they can’t even claim on. Income protection is the chief culprit here.

And so the super system needs to change and this imperative will only grow in terms of member harm in our ever evolving workforce.

Compared to the environment that gave rise to Australia’s super system back in the late 1980s, contribution levels are much higher, women are more likely to work and life expectancies are higher – put bluntly, much more is at stake today in financial terms.

And things will only get worse unless we change the system. The labour market itself is changing, with multiple job holding becoming more common, and people changing industry and occupation more than in the past. We are already starting to see the gig economy and technologies such as automation breaking down some of the industry and occupational boundaries we once had.

Retirement itself is also evolving. People are working for longer and are retiring with bigger super balances. Retirement is getting more complex too, with rising longevity, wealth tied up in housing and different kinds of risks that need to be guarded against. Retirement will not get any easier for Australians to navigate.

We’ve some ideas on modernising the super system so super is no longer the unlucky lottery it has become for many.

To recap, the odds are not great for members today. 1 in 3 accounts is an unintended multiple. 1 in 4 funds persistently underperform, and even in the better‑performing default segment, 1 in 3 MySuper products underperforms. In choice the odds worsen, with 1 in 2 products underperforming, notwithstanding our small sample size.

Policy initiatives to date have chipped away at some of the problems, but have clearly not gone far enough. Our draft report calls for some fundamental changes to the super system. We want to head-high tackle the twin problems of a member being defaulted into an underperforming fund and accumulating unintended multiple accounts. Fixing these problems would benefit members to the tune of $3.9 billion each year.

To do this, we want to make default the exemplar, and all about the member. So behavioural economics has inspired and informed our thinking. As noted by Professors Nicholas Barr and Peter Diamond in a submission to our inquiry, some people will have bad outcomes if forced to make choices from a large number of funds, meaning the normal forces of supply and demand cannot be relied on. But the use of a competitive process to afford funds access to default members (competition for the market) can harness the benefits of a robust competitive process for members — and based on criteria that matter most to them — in driving better performance and better member outcomes (while also weeding out underperformance).

We have a package of improvements – including 22 draft recommendations – with three elements at its core.

First and foremost, members only default once. And their account is attached to them and moves with them when they change jobs – their super follows them. They will only have a different product if they choose to switch. We are recommending the ATO (Australian Tax Office) put in place new processes to facilitate this, as well as taking stronger action to mop up the existing stock of unintended multiple accounts.

Second, we make funds compete for the new job entrant component of the default market every four years to appear on a ‘best in show’ list of the top performing funds. This list would include up to 10 top performers, with simple and comparable metrics to help members choose. This will make member engagement easier, especially for new workforce entrants. And comparing fund performance at last becomes a possibility.

The ‘best in show’ shortlist will support member engagement, but it does not solely rely on member engagement to work. Those new workforce entrants who fail to make a choice will simply be defaulted into one of the ‘best in show’ funds, determined by sequential allocation. And we learned from our member choice survey that number is likely to be low – some 5 per cent of members.

Funds will need to compete to be on this ‘best in show’ shortlist, and in doing so extend their best in show offer for new workforce entrants to their existing default members. This would see the benefits immediately spilling over to many existing default members. Funds will be selected every 4 years by an independent expert panel set up for the task. And this independent panel should be comprised of experts who can make decisions about what benefits the members.

To guard against poor decisions, the panel’s processes should be transparent, by publishing selection criteria and weights, as well as the reasons for decisions.

The third core element of our package – one that has got much less attention than the other two – is an elevated threshold for MySuper. This is essential to make MySuper a safe list of good funds for members – which it is not today. Raising the bar will mean that underperforming funds must either lift their game, exit with APRA (Australian Prudential Regulation Authority) shepherding their members to another fund or merge. And existing members will be the largest beneficiaries of this.

We are aware that there is already a scale test in MySuper, but it clearly hasn’t worked, with 112 funds in the super system still having less than $1 billion in assets. And these funds have 2 million member accounts. The Government is trying to legislate an outcomes test — a step in the right direction, but we think the outcomes test needs to be bolstered too.

So what will this all look like for members? We are also recommending that the ATO set up a ‘centralised online service’ – a one-stop-shop for members to choose where their super goes when they start a new job, to move to a new fund or consolidate accounts. And to review their current super product against the ‘best in show’ shortlist each time they log on.

We recommend this electronic system be compulsory for all employers and employees. And its design should be based on thorough consumer testing, so we are not being too prescriptive about what it should look like.

Essentially, new workforce entrants are nudged to select a product from the ‘best in show’ shortlist. But they can also see the list of all MySuper products and choose from them, or nominate any other product if they want something else. All MySuper products, whether on the best in show list or not, will have a simple and comparable one page dashboard that will be easily accessible through the online service. Nobody will be forced to choose or switch products at any time.

Existing members can log on at any time, to compare their current product with the best in show list, and to switch if they want. If they have just started another job or re-entered the workforce, their last active super product will become their default product – a very different situation to what happens today.

We see all members benefiting from our proposed changes to default allocation, not just new workforce entrants – though it is new entrants that will immediately no longer be facing the twin risks of being defaulted into multiple accounts and underperforming products.

Under our proposed system, member engagement will be greater, and fewer members would default each year.

Competition would be harnessed and its benefits unlocked – not the unhealthy competition we see today (especially in the choice segment) that favours the funds, but competition that delivers for members. These benefits would include stronger performance, lower fees, and greater innovation.

Now, this is not all about new workforce entrants, and nor is it about shortlisted funds gaining control of all the default contributions in the super system. On the contrary, existing members will remain in their current fund unless they choose to switch. Those that do nothing will benefit from greater and healthy competition in the system, and the requirement that shortlisted funds extend the benefits to their existing members. As a result, most existing default flows will remain where they are unless a fund loses MySuper authorisation under the much needed higher threshold.

But we do expect to see much more voluntary switching than in the past, as we are making engagement a lot safer and simpler. We have undertaken transition modelling. And modelling that APRA has found to be sound and consistent with their role as shepherd being manageable.

A key outcome we envisage from our policy changes is the exit of persistent underperformers, with the tail of persistently underperforming funds departing in an orderly manner and their members shepherded into better performing funds. Ideally the shepherding would be by the existing Trustee board, through a merger they see as in the best interests of their members. But ultimately the regulator stands ready to shepherd when needed.

This won’t lead to a ‘super’ oligopoly as some have suggested. The ‘best in show’ only applies initially to new job entrants — some half a million new members each year with about $1 billion in assets (albeit growing quickly from that starting base). The sheer size of the system — $2.6 trillion today and projected by others to grow to $4.3 trillion by 2032 — clearly has the capacity to support many, many more than just our 10 best in show funds. And the distribution of fund performance over the past 10 years suggests that many good funds will be nipping at the heels of the shortlisted funds every 4 years when the list is revisited.

Importantly, the benefits are not just about default – choice members would see benefits too, and not just from funds lifting their game generally. The ‘best in show’ shortlist would serve to improve financial advice – by becoming a benchmark for advisers to use in recommending products, and for their customers to put pressure on advisers to explain why any product advice diverges from the list. It will also help regulators to enforce the Future of Financial Advice laws. And ASIC (Australian Securities and Investments Commission) has provided us with advice to that effect.

And just how big are the benefits of change? We ran some cameos to look at how much members would be better off if the super system was no longer an unlucky lottery.

By fixing the twin problems of unintended multiple accounts and entrenched underperformance there are huge benefits for many members.

• Even a typical 55 year old worker today would be over $60,000 better off in retirement.
• And for today’s new job entrant, being defaulted once into a top performing fund would see them over $400,000 better off when they retire in 2064.

Specifically, these numbers are for someone being defaulted into a single top performing MySuper product rather than two underperforming products.

Our package of improvements has many other elements too. We are recommending:

• Clearer, simpler and more widely applied product dashboards to help members shop around. ASIC should play a ‘make it happen’ role here.
• Making insurance opt-in for members aged under 25 and with inactive accounts, as well as a process to strengthen the industry’s insurance code and make it binding and enforceable.
• Bringing governance up to best practice, by requiring all boards to maintain and use skills matrixes, setting stronger disclosure requirements around outsourcing, and greater regulator scrutiny of mergers. Mergers are not occurring as often as they should. In the report we highlight as best practice the example of Westscheme’s merger with AustralianSuper. This was a great example of trustees taking the initiative of ensuring members were transferred to a larger fund when alternatively faced with the prospect of growing net cash outflows.
• And we think that regulators need to be member champions, by confidently and effectively policing trustee conduct, enforcing MySuper authorisation and collecting and using more comprehensive and member-relevant data.
• We have not (as some have suggested) made any recommendations relating to funds having compulsory independent directors. We have recommended (draft recommendation 5) what we think should be the de minimus in fund governance to be ‘in show’ at all. Its focus is on the calibre and skills of the trustee board – which we view as mattering most. We have not made any recommendations relating to funds having compulsory independent directors, though we do view a ‘critical mass’ (at least one third) of independent directors as best practice – especially in a world of related party conflicts. And those conflicts exist across the system.

You can read more about these in the draft report.

So what’s next? Public hearings are scheduled for 20 to 22 June, to be held in Sydney, Melbourne and any other capital city where we get sufficient interest. You can register on our website but do it soon, as registrations close 5 business days prior to each hearing. So that means you have until next Wednesday if you want to appear in Sydney.

Responses to our ‘top up’ (or second-chance) funds survey are due by 27 June.

Submissions on our draft report are due on 13 July. And we have opened up a channel for brief comments – a way of making it simpler and easier for members to give us feedback on their experience, their needs and their thoughts on our ideas.

We’ll be putting out three further supplementary papers, on economies of scale (are they being realised and passed through), the fiscal impacts of insurance, and the results of our ‘top up’ funds survey.

And we might also hold two technical roundtables — one on economies of scale, and one on our benchmarking analysis.

We don’t have a specific date for sending our final inquiry report to Government just yet. But it will be before the end of the year.

On a final note, this is a draft report, with draft findings and draft recommendations. And we are open to feedback on their merit. But for feedback to help shape our final report it needs to be accompanied with evidence. And especially evidence on what it means for members.