Jim Chalmers’ ‘truly bizarre’ $64bn tax expenditure

Treasurer Jim Chalmers. Picture: NCA NewsWire / Gary Ramage

One of the roles I try to fulfil in writing regular columns on economics is to relieve readers of the need to plough through turgid, often impenetrable government reports. Rather than waste your time on badly written, excessively long documents, my aim is to convey the main points as well as assess the credibility of the findings and recommendations.

Because superannuation tax concessions have been in the news lately, I thought it would be useful to outline the government report that sets out the estimates of the costs of tax concessions. Recall here that the point is regularly made that the annual superannuation tax concessions will soon cost more than $50bn per year, which in turn will exceed the annual government outlays on the Age Pension. It’s code for the government’s aim to wind back tax concessions for current and retired superannuation members.

It used to be called the Tax Expenditure Statement, but now goes by the name Tax Benchmarks and Variations Statement. This statement is a requirement of the Charter of Budget Honesty Act 1998.

The theory of tax expenditures is relatively straightforward. Instead of a government spending taxpayers’ dollars to obtain certain outcomes, it is possible for a government to use tax concessions to obtain the same or similar outcomes. We should therefore know the cost of these concessions in the same way we know the cost of government spending.

While the theory may sound self-evident, in practice, what is and is not a tax concession is not at all obvious. Deliberate design features of the tax system – say, the progressive income tax scales – shouldn’t be regarded as tax concessions and are not included in the Statement. (Just imagine the cost of not taxing everyone at the top tax rate of 47 per cent.)

The point is the inclusion of certain items and the benchmarks used to estimate the cost of tax concessions are essentially arbitrary exercises. This is particularly the case with superannuation tax concessions and the contentious decision to use income tax as the benchmark. No sensible policy analyst thinks taxing saving, particularly compulsory, locked-in saving, at full income tax rates makes any sense. But this is precisely what Treasury assumes.

In fact, there is an even more fundamental problem with the way Treasury goes about delineating and estimating tax concessions and that is the implicit assumption 100 per cent of income belongs to the government and any amount individuals or entities are able to retain is a concession. No doubt there are some far-left governments that might take this view, but hopefully no Australian governments will ever get to that point.

As a first observation I can tell you reading the phone book is more interesting than the statement on tax expenditures. There is page after page of weirdly selected tax concessions but only asterisks in the section of the tables that would normally have numbers. Why bother to include these tables if they are so immaterial that there are no numbers attached to the supposed tax concessions?

Believe me when I say there are some fantastical inclusions, such as: the Medicare levy exemption for blind pensioners; exemptions for valour or brave conduct decorations; exemptions for Australian Defence Force personnel on overseas duty; even exemptions for winners of the Prime Minister’s awards.

I could go on, but no one in their right mind thinks these concessional tax arrangements should be included in any list of real tax expenditures.

Let me get to the large items in the statement where all the fuss emanates. By far the biggest item is the main residence exemption, adding $64bn per year because the government doesn’t impose a capital gains tax on the sale of a main residence and the capital gains tax, in turn, involves a “discount”. This is truly bizarre stuff.

No mainstream political party has ever proposed imposing a capital gains tax on owner-occupied dwellings or to remove (at least fully) the discount on the tax. Does Treasury really believe homeowners should pay full income tax rates on the full capital gain when they sell their properties? Surely the flipside is that they would be able to claim tax deductions for the costs of owning their homes, including mortgage payments, which would be enormously costly to the budget.

When it comes to superannuation tax concessions, the two main items are the concessional taxation on fund earnings and on contributions. But there are some important points to make. The reliability of these estimates is only low to medium. Moreover, there is an important difference between revenue forgone and the revenue that might be gained were the concessions removed.

Even Treasury concedes “revenue forgone estimates (the figures quoted in the press) are not estimates of the revenue increase if a tax benchmark variation were to be removed. Care should be exercised when comparing tax benchmark variation estimates with direct expenditure estimates”. The fact is that were the concessions to be removed or altered, people’s behaviour would significantly change.

Treasury does concede there would be a substantial loss to the super pool over time. It is entirely possible that, with less generous concessions or caps, people would take their money out of super and invest it in other tax-effective ways. So when you next read that the super tax concessions are going to cost over $50bn per year, don’t believe it. It is based on a completely inappropriate benchmark and does not capture the smaller amount of revenue the government would receive should a foolish decision be made to alter the tax/regulatory arrangements.

In fact, the real problem is that we have a highly defective way of taxing superannuation. Most countries exempt both contributions and fund earnings because this allows members to maximise their final payouts. Withdrawals are then taxed at normal income tax rates. It’s the EET model – exemptions on contributions, exemptions on earnings and full tax on retirement incomes.

These countries do not count exemptions from tax on contributions and earnings as part of their tax expenditures – they would be vast if they did so. They are simply seen as undisputed parts of the saving and taxation landscape. Our model of ttt (small t is for concessional tax treatment) is one of the worst because it reduces the scope for members to become self-sufficient upon retirement.

There are also serious doubts about the way in which Treasury calculates the burden of superannuation taxes. Henry Ergas, of this newspaper, and Jonathan Pincus have demonstrated that, far from providing significant tax breaks, including for high-income earners, our unique system of taxation leads to extremely high effective rates, often over 50 per cent, because cumulative taxes are paid for long periods.

If the Treasurer were really interested in reform, he might like to consider the overseas model of EET. This could be achieved by grandfathering current accounts then everyone starting again with new accounts with the new tax arrangements. Making super voluntary also would have benefits.