Why superannuation concessions are no budget goldmine

by Andrew Podger

The idea that superannuation tax expenditures amount to around $30 billion dollars a year, vastly favouring the rich and presenting easy pickings to fix the budget has become widely accepted. The facts, however, are very different. The tax treatment of superannuation is excessive, and does favour higher income groups inappropriately, but the total amount involved is a small fraction of $30 billion – almost certainly well under $10 billion a year.

Despite being one of the original proponents of regular reporting of tax expenditures with the annual budget papers, I have become increasingly uneasy about their interpretation and many of the proposals to rein them in, including by commentators and academics who should know better.

With three colleagues I co-authored the Department of Social Security submission to the 1981-82 inquiry into tax expenditures by the House of Representatives standing committee on expenditures. We supported the idea of an annual statement and set out in the submission estimates of the tax expenditures that fell within the functional category of social security and welfare that was then used in the budget papers for outlays.

In recommending an annual tax expenditures statement, the committee highlighted the Canadian advice we had drawn attention to in the DSS submission:

“Since the treatment of benchmark tax structure is an abstraction there will always be room for legitimate disagreement about its nature and, thus, about whether certain tax provisions are properly characterised as tax expenditures.”

This issue remains central to calculating superannuation tax expenditures. The DSS approach was based on the concept of a comprehensive income tax, the concept subsequently – and still – used by Treasury despite their criticisms of us in 1981. This was the then conventional concept based on the idea that “income is income is income” and should be taxed. The concept can be seen in such influential reports around that time as Asprey (1975) and Campbell (1981).

Consistent with this concept, contributions to superannuation, and the income earned on those contributions, should both be included in individuals’ taxable incomes and taxed at the relevant marginal rate; the resulting balances should then be tax free. Compared with this “TTE” benchmark (tax tax-exempt), we calculated the tax expenditure in 1981-82 at around $2 billion. This same concept now forms the basis of the Treasury’s current estimate of around $30 billion.

When the Hawke government reviewed retirement income policies on taking office in 1983, Treasury’s starting point was to suggest reducing tax expenditures by taxing contributions and fund earnings (as well as tightening the then income test on age pensions). By then I had joined Finance and its brilliant secretary, Ian Castles, who questioned the wisdom of pursuing the comprehensive income tax concept for retirement savings.


Castles highlighted the strong disincentive to save that a TTE regime would impose and recommended instead that contributions and fund earnings be exempt and benefits be taxed in full. This “EET” (exempt exempt tax) – or comprehensive expenditure tax – approach is now widely acknowledged internationally as the ideal, encouraging people to spread their lifetime earnings and consumption.

Castles was keen to shift the debate to focus on ensuring superannuation was used for genuine retirement purposes and promoted benefits in the form of pensions or annuities rather than lump sums, firstly by removing the then tax concession on lump sums (we calculated tax rates that would be equivalent to the tax on an annuity) and secondly by issuing indexed bonds to allow the market to trade the risk of inflation and offer indexed annuities as were commonly available in the public sector’s then unfunded schemes. Despite strong Treasury opposition, the government accepted both Finance proposals.

With the welcome advent of compulsory superannuation under the Hawke/Keating government, initially through the accord with the ACTU and later through legislation, Keating took advantage of the shift to funded superannuation with its actual contributions and earnings, imposing a 15 per cent upfront tax, but allowing these to offset future tax liabilities when benefits were paid. It was a quick revenue grab aimed to meet short-term budgetary pressures. The Keating regime might be described as “TTT” rather than EET or TTE, the taxes at each stage being less than a full income tax might involve. Costello took a big step further away from the expenditure tax approach exempting all benefits from taxed contributions (TTE). This retained nearly all the short-term revenues Keating grabbed but added to future revenue problems as the superannuation system matures and our population ages.

After more than 20 years of taxing contributions and earnings, it is now too hard to go back to EET. Apart from the administrative problems (for ATO, funds and individuals), the revenue implications of a shift to EET would be extremely adverse for a decade or more.

Treasury continues to present superannuation tax expenditures against the comprehensive income tax (TTE) benchmark, identifying a total now in excess of $30 billion a year. Last year, but only as a once-off, it presented another version of the tax expenditures against what it termed a comprehensive expenditure tax, but based on TEE (taxing contributions only) not EET (taxing benefits only). This revealed about $12 billion in tax expenditures. But the tax rate under TEE is only equal to the rate under EET if the tax scale is proportional not progressive. A full tax on contributions would be at individuals’ marginal tax rates as they would also have other significant taxable earnings, but a full tax on benefits would generally be nearer to individuals’ much lower average tax as most would not have substantial other taxable income and their benefits would be lower than their pre-retirement incomes. This suggests that, even with the impact of earnings on contributions, tax expenditures based on an EET benchmark would be lower than $12 billion, probably much lower.

Because our superannuation system is not yet mature – with benefits still to grow a great deal from current levels of contributions – it is very difficult to calculate the concessions meaningfully against an EET benchmark, and arguably unhelpful if we cannot move now to an EET regime anyway.


An alternative is to try to identify a TTE regime that might have a similar impact to EET for most people, and then to calculate the additional revenues involved and the distributional impact. Very broadly, the Henry report approach may come close as it involves applying people’s marginal tax rates to contributions less some standard amount, which could be similar to the average tax that would apply to benefits under an EET regime. With a standard offset of 20 per cent as Henry suggested, the majority of contributors would still pay 15 per cent tax as most taxpayers face a marginal tax rate of around 35 per cent. But the contributions tax would be higher at higher incomes and lower at lower incomes. Henry also suggested a lower tax on fund earnings but applied this in both the accumulation and pension phases. That might also get closer to an EET outcome.

Swan and Shorten did take a first step in this direction by imposing a 30 per cent tax on contributions by those earning over $300,000 (corresponding broadly to their marginal tax rate less 20 per cent). Reducing the threshold for the 30 per cent tax to $180,000 would be consistent with the Henry model, as would reducing the tax on (or exempting from tax) contributions by those earning under $37,000. These measures would clearly make the system fairer, but the additional revenues involved would be small relative to either of Treasury’s tax expenditure estimates particularly if the tax on earnings was reduced from the current 15 per cent.

Some commentators, such as the Grattan Institute and the Labor opposition focus more on containing the contributions and fund earnings that attract the current 15 per cent tax. These proposals step around identifying the benchmark tax regime and simply try to reduce claimed excessive concessions for high income groups. But if EET (or the equivalent TTE) is in fact the ideal benchmark, why deny self-funding middle and high income earners having it apply to them?

The Grattan Institute proposal of a $11,000 contributions cap seems particularly extreme. It is less than anyone earning over $90,000 will be compelled to contribute when the legislated 12 per cent mandated contributions come into force, it is way under the 15 per cent or more that people not eligible for an age pension should contribute to ensure an adequate retirement income, and it would greatly hinder people with interrupted careers from catching up their superannuation savings.

Rather than the crude approach of tough caps, the focus should be on a tax regime that most closely corresponds to the ideal EET approach. Then the retirement incomes agenda should turn to complementary measures such as those recommended by the Murray report to encourage people to take their benefits in the form of regular and secure income streams, to promote longer periods of employment and saving, and to better align superannuation and age pension arrangements.

This article was first published in the AFR on 4 January 2016.

Leave a Reply