Too much of a good thing. Is compulsory super asking us to save too much?

Superannuation is once again a key political battleground. One Nation is proposing to let renters and mortgage holders redirect one-quarter of their compulsory contributions into their take-home pay, while the Coalition mulls whether to allow superannuation as collateral for housing loans. The Albanese government remains committed to the current system.

Vested interest and ideology run through this debate. But is there an economic case to suggest current settings lead to oversaving?

Compulsory superannuation was introduced to help people save for retirement, reduce financial stress among older Australians, and shift the cost of old age from the taxes of the young to the savings of the old. On those measures it’s working well. But by tying up 12 per cent of an individual’s earnings until retirement, the current system may be creating 2 new problems.

The first is a lifecycle consumption-smoothing problem. Compulsory superannuation increasingly transfers money from when Australians are young and relatively poor to when they are old and relatively rich. Lifecycle income tax calculator estimates suggest that – if people were unable to borrow – the average Australian could now have a retirement income that is 50 per cent higher than the cash they had access to during their working age.

The second is an intra-generational (within-generation) equity problem. Superannuation is lightly taxed because it is meant to fund retirement and reduce Age Pension costs. But in an increasingly self-funded system, retirees are incentivised to hold onto more than they need in case they live longer than expected. That means much of it ends up being inherited. As balances grow, what was meant to be a retirement income system risks becoming, in part, a tax-advantaged inheritance scheme that disproportionately benefits the children of the already well off.

How did we get here?

To judge whether the current system is working, it helps to know how we got here. And to do that, you need to go back to 1900, when NSW introduced Australia’s first age pension. A national scheme followed in 1909, with Australia joining a host of countries (notably Germany and New Zealand in 1889 and 1898, respectively) introducing public pension systems.

The motivation for these new pensions was to provide an income floor for individuals once they reached retirement age – when society no longer expected them to work. They were introduced partly in response to the hardship older generations had faced during the Long Depression of the late 1800s.

Even at the time there were questions about how to fund the new system. Should it be a universal entitlement, or based on contributions paid in? Should the funds for the scheme be collected and held in an investment vehicle (pre-funded) or paid from current tax revenue (pay-as-you-go)?

In the 1950s, an influential paper from economist Paul Samuelson highlighted that pay-as-you-go funding was the more efficient method due to the strong population and economic growth of the day. He also argued that it would be more intergenerationally equitable as it would share the benefits of economic growth with older generations by having richer, younger generations support them in their retirement.

This argument supported the tax funded, pay-as-you-go approach already taken in countries like Australia, and suggested that the Age Pension would be a stable and important part of the system going forward.

However, by the 1980s 2 key things had begun to change.

1. Undersaving. There was a growing concern that individuals were ‘myopic’ in their savings decisions, and so were undersaving for their own retirement. Adding to this, the age pension system created ‘moral hazard’ because it incentivised individuals to work and save less in the knowledge that the taxpayer would pick up the bill for their pension.

2. Slowing growth. Both productivity and population growth had begun to slow. This changed the calculus of Samuelson’s result. With risk adjusted returns on superannuation funds now expected to be larger than overall economic growth, the case for pre-funding payments to retiring Australians became stronger both from an efficiency and inter-generational equity point of view.

These 2 concerns helped drive reform of retirement income policy in Australia, which culminated in the introduction of compulsory superannuation in 1992. A number of other arguments were used at the time, including managing fiscal expenses, capital deepening, concern about levels of national savings, and reducing dependence on ‘benefits’. However, they were all derivative of these 2 fundamental concerns.

Over the last 3 decades compulsory superannuation has largely succeeded in addressing both issues. It links what you receive in retirement to what you put in, reducing the moral hazard associated with the Age Pension and forcing ‘myopic’ people to save more than they otherwise would have.

Contributions to superannuation also occur during an individual’s working life and are placed into a fund. Combined with the increased means testing of the Age Pension, this helped shift Australia towards a more pre-funded model of retirement savings – each generation paid for more of their own retirement costs, rather than relying on the support of younger generations.

That shift has helped make Australia a world leader in managing the costs of population ageing, with one of the lowest and slowest growing rates of public pension expenditure in the OECD.

But, with the compulsory contribution rate rising to 12 per cent of income, there is now a question of whether the superannuation system, in solving an undersaving problem, has created an oversaving one.

When does superannuation become “oversaving”?

Saving for retirement comes with trade-offs. Every dollar you put away today will likely be worth more in the future. But the more you put away, the less you have to spend now.

Retirement also isn’t the only thing people save for. You might save to buy a house, to start a family, or to build a buffer in case you lose your job. More money saved for retirement means less money for these other purposes.

To help think through these trade-offs, we built a lifecycle income tax calculator that includes the full set of federal tax and transfer policies (including a public calculator). We use data from a sample of 3,000 individuals from the nationally representative Household, Income and Labour Dynamics in Australia (HILDA) Survey. We then predict household equivalised income paths over the lifecycle of these individuals under different compulsory superannuation contribution rates – focusing here on 9 per cent and 12 per cent. This modelling is based on several assumptions, including no behavioural change, that are laid out in Box 1.

The chart that follows highlights the level of retirement income that different compulsory savings regimes generate, relative to each household’s working age average – their replacement rate. The modelling suggests that under a 12 per cent contribution regime average replacement rates exceed working age income for most individuals. In other words, an average household will have more cash income each year in retirement than they did when they were working. Under a 9 per cent scheme, average replacement rates fall – but at each income decile the average individual does end up with more cash on hand in retirement than they had during their working life.

Notably, the largest difference between the 9 per cent and 12 per cent schemes occurs at the top of the distribution, with only a smaller difference for lower income households. This is because lower income households are more heavily supported by the Age Pension – which is a flat rate payment. As many of these households live below the poverty line during their working lives, this shows the importance of the Age Pension system for alleviating poverty among older Australians.

The implication is that from a lifecycle perspective, superannuation is effectively transferring money from when people are, on average, younger and poorer to when they are older and richer.

Is this a bad thing? That’s a value judgement, but there is clearly an arguable economic case that a 12 per cent contribution rate leads to oversaving if future returns resemble past returns.

It seems strange, after all, to compel people to save so much that they have more money in retirement than while working. While it’s true that some health-related costs may rise in retirement, many of these are covered by government transfers such as Medicare.

Most people also appear to prefer to smooth their consumption relatively evenly over their lifecycle. International research suggests that for most households their consumption remains similar when they retire, but their spending falls. This is because they have fewer work-related expenses (work clothes, transport, coffee and lunch), and an increase in the time available for home production and to shop around.

Of course, this isn’t an argument against saving more voluntarily, or against compulsory retirement saving altogether, just that we perhaps shouldn’t compel people to save so much.

It’s also true that a 12 per cent contribution rate will not lead to every individual being better off in retirement than during their working years. But even in many of these cases it’s not clear that mandating retirement saving at current rates makes these people better off.

Take the example of a person who spends 2 years out of the workforce between 31 and 32 – because they were retraining, caring for a family member, or could not find work. We model this as a sustained negative income shock in the figure below. Assuming this person does not save or borrow, a 12 per cent contribution rate means they will have nearly $92,000 pa to spend in retirement but will live on the JobSeeker Payment ($21,000 p.a.) for 2 years in their 30s.

By tying up more of their savings in superannuation, the higher contribution rate worsens the hit to their available resources during their working life. People can save outside super too, but compulsory contributions reduce their capacity to do so. And because people generally cannot access their superannuation early or use it to secure a loan, it forces them to prioritise saving for retirement over support during the shock.

One solution would be to allow individuals to access their superannuation early. Doing so would let the individual in our example stay on the same spending path for nearly 2-years after job loss. The trade-off is that their available cash at retirement falls to $75,000 per annum. But this is still 5.8 per cent higher than their average lifetime income. This shows that such access can help to cushion the blow of job loss, even among relatively young Australians.

However, the design of such schemes would need careful consideration. Previous e61 research examining Australia’s COVID-19 early release of super scheme found that while some of the funds withdrawn went toward debt-payments, much was spent on immediate, non-durable goods and ATM cash withdrawals. This highlights the trade-offs that will inevitably be involved when allowing mandatory retirement savings to be used for other purposes.

There are 2 further reasons why the comparisons above could overstate concerns about oversaving, with implications for the policy conclusions. The first is that our replacement rate calculations compare an individual’s working-age and retirement income in real terms. That ignores the potentially reasonable expectation that retirement incomes will rise in line with living standards in the broader economy.

The second is that the degree to which historic rates of return may be stronger than future returns. We use historical nominal rates of return of roughly 8 per cent for superannuation as our base. If they were to fall, then that calculus would change and oversaving rates would decrease.

If nominal returns fell to 6.5 per cent – about what most mortgage holders now ‘earn’ in their offset account – most individuals would still have about the same cash income in retirement as while working, and many would have more.

At a more pessimistic 5 per cent nominal return – below the current 10-year Australian Government Bond yield – individuals in the top half of the distribution would have less cash income in retirement, assuming no offsetting increase in private savings. Lower-income individuals would be less affected because they already rely more heavily on the Age Pension.

Superannuation’s oversaving problem is made worse because people face “longevity risk”

A key benefit of superannuation – having individuals fund part of their own retirement to reduce reliance on the Age Pension – is also one of its key shortcomings.

This is because individuals face a big risk in retirement – they don’t know how long they will live. At 67 there is a 50 per cent chance you live until 90, and a 10 per cent chance you live until at least 100. Faced with this ‘longevity risk’, many people will save more than they need to avoid running out of resources towards the end of their life.

A public pension system is closer to a traditional insurance model. Money is pooled across individuals and then paid out to those who suffer the ‘shock’ of living longer than expected.

Requiring individuals to self-insure compounds the oversaving problem during their working lives. Not only are people likely saving more than they need for a retirement of average length, but they are also incentivised to hold onto more of it as a means of self-insurance.

The result of these 2 forces is that people are increasingly dying with large superannuation balances. And what was designed as a system to support people in their retirement is slowly turning in part into an inheritance scheme. Treasury’s 2019 Retirement Income Review predicted that by 2059, 1 in every 3 dollars paid out of the superannuation system will be paid out as an inheritance, up from 1 in 5 dollars in 2019.

In solving an inter-generational equity problem are we creating an intra-generational one?

By shifting Australia to a pre-funded retirement income system, superannuation has helped reduce transfers from younger Australians to older Australians, and in turn alleviated concerns around whether it is equitable for younger generations to fund the retirement of older generations.

But in doing so it may be exacerbating an intra-generational (within-generation) equity issue. The oversaving discussed above may lead some people to leave larger, superannuation-funded inheritances than they would have chosen without compulsory savings. And those inheritances will pass opportunities unevenly to future generations.

The issue is not so much that people are passing money down to their children. It is that they are doing so through a scheme whose light tax treatment (relative to other capital income) was justified on the grounds that it would fund retirement and reduce spending on the Age Pension.

These tax concessions are expected to grow further with a 12 per cent contribution rate. The 2026 Intergenerational Report expects the forgone tax revenue will reach 2.7 per cent of GDP by 2065-66, overtaking Age Pension expenditure, which is projected to fall to 1.8 per cent of GDP.

For wealthy individuals with large balances, superannuation effectively risks turning (in part) into a tax advantaged inheritance vehicle, further exacerbating the intra-generational inequality created by those who stand to inherit and those who do not.

An important caveat here is that the taxation of capital income at full rates would heavily reduce the ability for individuals to accumulate income for their retirement. In the below example, an individual with twice the income of the average full-time worker will end up with a superannuation account that is 60 per cent smaller with the same contributions, due to their higher marginal tax rate reducing the accumulation of superannuation funds.

Policy options

Options for reform of the superannuation system can be broadly broken into 2 areas. The first is altering how superannuation is paid out in the retirement phase, to address oversaving caused by individuals’ self-insuring.

Here a more efficient mechanism would be to move to an annuity-based system. Annuity payments could be linked to lifetime contributions, meaning that if individuals end up living longer than expected they would not experience undue hardship.

This would reduce the cost of the scheme as people would need to save less and remove one source of inherited income from the equation – unspent superannuation balances – mitigating concerns around intra-generational equity.

The second area of reform is altering how superannuation is paid in and accessed during an individual’s working life. One option here would be to reduce the compulsory contributions rate. This would lift take-home pay earlier in life, potentially helping younger Australians at times when they are liquidity constrained and money is tight – such as when saving up for a first home. But it would come at the cost of reduced retirement savings and increased pension expenditures.

An alternative would be to allow some superannuation to be used directly for other purposes that people save for, such as housing equity or for help during major shocks. This would reduce the risk that withdrawals are used for purposes that policy makers do not approve of but would still come with similar downsides of reduced retirement savings.

Retirement income policy involves hard trade-offs, between moral hazard and risk pooling, between maintaining a comfortable retirement and support through other lifecycle shocks. Vested interest and ideology also run through the debate. But there is a reasonable economic case that current settings could lead to oversaving. That is a case for revisiting how much we ask people to save and how, not for abandoning a system that has served us well in other ways.

This paper uses unit record data from Household, Income and Labour Dynamics in Australia Survey [HILDA] conducted by the Australian Government Department of Social Services (DSS). The findings and views reported in this paper, however, are those of the author[s] and should not be attributed to the Australian Government, DSS, or any of DSS’ contractors or partners. DOI: 10.26193/GMYVPW