Australia, like most other ‘advanced’ economies, has a ‘progressive’ personal income tax system – that is, one in which people on higher incomes pay not merely more income tax, but a higher proportion of their income in income tax, than people on lower incomes.  In the 2013-14 financial year, the 2.9% of Australian taxpayers whose incomes are sufficient for them to be subject to the top marginal rate pay 30% of the total amount of personal income tax collected by the Australian Government, which is almost twice as much as their share of total taxable income [1].

Taken at face value, Australia’s personal income tax system is actually more progressive than that of many other countries – particularly by comparison with others whose overall tax systems are broadly similar to ours, and which seek to raise similar proportions of national income in order to finance spending by governments. Australia currently imposes the English-speaking world’s second highest top income tax rate on incomes above the English-speaking world’s second-lowest top income tax threshold [2].

There is very little empirical evidence to support the assertion that Australia’s relatively high top marginal income tax rate has any adverse impact on incentives to work, especially for high-income professional men. There is far more evidence suggesting that high effective marginal rates arising from the interactions between the income tax system and the withdrawal rates for various benefits have an adverse impact on the incentives to work facing women considering re-entering the workforce after a period of caring for children, or elderly relatives.

However, there is also a considerable body of evidence suggesting that Australia’s relatively high top marginal income tax rate has a significant impact on how high income households arrange their affairs (for example, conducting business or professional activities through companies and trusts), and how they allocate their savings and investments. That’s why people in the top tax bracket – who, as noted above, account for less than 3% of all taxpayers – account for more than 15% of the value of tax deductions claimed for the ‘cost of managing tax affairs’. It’s a large part of why they account for 36% of all taxable income accrued through non-farming trusts and partnerships. It’s a major reason why they account for almost 21% of the value of non-compulsory superannuation contributions. And it’s a major reason why 23% of them – as opposed to fewer than 8% of taxpayers with taxable incomes of less than $80,000 – have negatively geared property investments.

At the other end of the scale, Australia also has a relatively high tax-free threshold, by international standards. Since 2012-13, the tax-free threshold in Australia has been $18,200. This is slightly lower than in the UK, where the tax-free threshold of ₤10,000 is equivalent to about A$19,250 (although people on incomes of less than ₤10,000 still have to pay Britain’s equivalent of the Medicare levy). But it is much higher than in the US, where the tax-free threshold for single taxpayers is US$9,300 and for married couples filing jointly is US$12,600 (equivalent to about A$12,250 and A$16,600 respectively), or Canada where it is (effectively) C$11,474 (equivalent to just under A$12,100). Since April 2013, New Zealand has not had a tax-free threshold at all.

The principal purpose of a tax-free threshold is to ensure that people on very low incomes don’t pay any tax at all. However, it also has the (unintended) effect of leaving a chunk of income earned by people on much higher incomes completely free of tax – with the result that those people end up  having to pay higher marginal tax rates on their last dollars of income in order that they pay a given amount of tax in total.

A tax reform option which might therefore be worthy of serious consideration is the possibility of reducing or eliminating the tax-free threshold for high-income taxpayers in exchange for a reduction in the marginal rates they pay on their last dollars – with perhaps the beneficial impact of reducing the incentive which those high marginal rates on their last dollars creates to engage in tax minimization strategies.

I asked NATSEM to model an option like this, using their STINMOD+ microsimulation model [3]. Specifically, they modelled the effects of reducing the tax-free threshold for taxpayers with taxable incomes in excess of $80,000 (the threshold at which the second top marginal rate of 39% becomes payable) by 12.2 cents for each additional dollar of income, until taxable income reaches $180,000 (the top income tax threshold), at which the tax-free threshold will have been reduced to $6,000 (the level which it was before 2012-13). The NATSEM modelling suggests that the additional revenue thereby used could be applied to raising the threshold at which the second top marginal rate becomes payable from the present $80,000 to $130,000.

NATSEM’s modelling indicates that such a change would have:

• no effect on people whose taxable incomes are less than $80,000 (since the thresholds and rates which they face are unaffected);
• a slightly positive impact on people with taxable incomes of between $80,000 and $177,066, for whom the increase in the threshold at which the 39% rate becomes payable more than offsets the drop in the amount of income which attracts a zero tax rate, by an average of $460 per annum; and
• a negative impact on people earning more than $177,066, for whom the increase in the second top income tax rate threshold does not fully compensate for the drop in the amount of income on which no tax is payable, averaging $2,220 per annum.

The second group – the people who are on net better off – account for about 16% of those with appositive taxable income; while the third group – who are on net worse off – account for around 3½% of those with a positive taxable income.

It would be quite possible to ‘tweak’ this proposal in other ways – for example, by extending the tapering of the tax-free threshold down to zero, and further raising the threshold at which the second top rate becomes payable; or alternatively by leaving the top two thresholds unchanged, and lowering the top two rates (which would be my preference, because of the favourable effects which that might have on the incentive to engage in tax-minimization strategies).

Either way, the modelling highlights the myriad possibilities for improving the structure of Australia’s personal income tax system in ways that reduce its unwelcome disincentive effects whilst maintaining fairness.

Saul Eslake is a Vice-Chancellor’s Fellow at the University of Tasmania and an independent consulting economist. He was formerly Chief Economist at the Australia and New Zealand Banking Group (ANZ) and of Bank of America Merrill Lynch Australia. He thanks Dr Jinjing LI of NATSEM at the University of Canberra for his assistance.

[1] These figures are for the 2013-14 income year.

[2] The present top marginal rate (including the 2% Medicare levy and the 2% ‘temporary deficit repair levy’ imposed between 1 July 2014 and 30 June 2017) is payable on taxable incomes in excess of A$180,000. The highest top marginal rate in the English-speaking world is paid by residents of New York City, who pay 52.2% (in federal, state and city income taxes) on taxable incomes in excess of US$500,000 (roughly A$658,000 at current exchange rates). New Zealand’s top marginal income tax rate, of 33%, is payable on taxable incomes above NZ$70,000 (about A$64,200 at current exchange rates). The choice of ‘English speaking countries’ (Australia, New Zealand, the USA, Canada excluding Quebec, and the UK) is for simplicity and because of broadly comparable tax systems, and does not reflect any belief that the taxation systems of these countries are in any way ‘superior’ to those of other countries. 

[3] Microsimulation models apply the Commonwealth Government tax and transfer rules to data at the individual and household level. This means complex tax/transfer policies can be modelled, incorporating complex interactions between different policies. The same kind of models are used extensively by the Commonwealth Government in Australia, and government worldwide. STINMOD+ primarily focuses on the ‘day after’ impact of a policy. The analysis in this case does not include the second round impact such as behaviour shifts.