Australia's Future Tax System

Final Report: Detailed Analysis

Chapter A: Personal taxation

A1. Personal income tax

A1–1 The structure of personal income tax

Key points

The personal income tax system should continue to be progressive, but it should operate in a simpler and more transparent way. The centrepiece of the system should be a high tax-free threshold with a constant marginal rate for most people.

The personal income tax system should support workforce participation by limiting high effective tax rates, especially for those people who are likely to be most responsive to financial incentives to work.

The primary unit in the personal tax system should continue to be the individual, and subsidies for dependants through the tax system should be restricted.

Income support and supplementary payments should be exempt to simplify tax and transfer interactions.

Where possible, tax offsets that are structural in nature should be incorporated into the personal income tax rates scale, along with the Medicare levy. Tax offsets that provide a concession for a particular group should be removed or delivered as a direct payment or service.

The taxation of personal income is the most important means of raising revenue in developed countries. However, personal income taxes discourage workforce participation and savings, both of which are important for economic growth.

The main purpose of the personal income tax system is to allow governments to raise revenue to pay for public goods (like education, health care and law enforcement), and to provide income support for those less able or available to support themselves. Underpinning the design of the income tax system is the desire to provide a balance between ensuring that those people with more capacity to pay contribute more (vertical equity) and that those with a similar capacity to pay bear the same burden (horizontal equity).

The practice of taxing those with greater capacity to pay reflects the view that an extra dollar of income is generally of more value for a person with a lower income than for a person with a higher income. That is, people on lower incomes benefit more from a lower average tax rate than people on higher incomes lose from paying a higher average tax rate.

An individual's capacity to pay is difficult to define. In the absence of existing wealth, there is an argument for redistributive tax policy to be based on an individual's potential earnings capacity. However, as information on potential earnings capacity is not readily available, observable proxies are required.

Actual labour income is often used as a proxy for potential earnings capacity. But wage and income differentials may reflect a number of other factors, including choices about how much to work or study. Taxing wages or income therefore biases decisions to undertake paid work and may also affect decisions about undertaking education and training. Savings income is also relevant to a person's capacity to pay, particularly for pre-existing savings or taxes on economic rent. But taxing the normal return to saving is likely to bias savings, labour supply (for those that save part of their wages) and consumption decisions.

As income taxes can lead to a decline in overall economic output, there can be a trade-off between equity and efficiency when designing the personal income tax system.

A hybrid personal income tax base

Australia's personal income tax system should continue to represent a hybrid personal income tax — with income from long-term, lifetime, savings taxed at a lower rate than other income or exempt from income tax. In particular, the main forms of lifetime savings for most Australians, superannuation and owner-occupied housing, should continue to be taxed at a lower rate or exempt from income tax — consistent with an expenditure tax benchmark that exempts the returns to saving (see Section A2–2). Comprehensive income taxation, under which all savings income is taxed in the same way as labour income, is not an appropriate policy goal or benchmark.

The essential reason for exempting lifetime savings or taxing them at a lower rate is that income taxation creates a bias against savings. The income taxation of savings therefore discriminates against taxpayers who save. They pay a higher lifetime tax bill than people with similar earnings who choose to save less. As savings can be thought of as deferred consumption, the longer the person saves and reinvests, the greater the implicit tax on future consumption (see Chart A1–6). For a person who works today and saves, taxing savings also reduces the benefit from working.

Chart A1–6: Tax wedge on future consumption

Chart A1–6: Tax wedge on future consumption

Assumptions: Pre-tax interest rate of 6 per cent per annum and a tax rate of 30 per cent.

Source: Treasury estimates.

The increasing implicit tax on future consumption provides an argument to tax longer-term lifetime savings at a lower rate. An individual can undertake lifetime saving through a variety of savings vehicles, but there are asset types that are more conducive or related to lifetime savings: namely superannuation and owner-occupied housing. It is possible to convert savings in these assets into present consumption by borrowing against them, directly or in effect. Further, the family home yields a stream of income (imputed rent) that is also a form of current consumption. While these features could diminish their status as lifetime savings vehicles, in practice these assets will in net terms remain major forms of lifetime savings for most Australians, and provide for a major part of their retirement income.

An exemption from income tax or applying relatively low rates of tax to superannuation and owner-occupied housing is common practice around the world and has been a longstanding feature of the Australian tax system. The family home has not been subject to income tax in Australia since the earlier part of last century. Imputed rental income and capital gains from owner-occupied housing are generally exempt in the OECD countries, with a few exceptions.

While owner-occupied housing represents more than lifetime or retirement savings, other factors support its continued exemption. Given there is little community acceptance for applying income tax to the family home, any attempt to subject it to taxation is unlikely to be sustainable. Australia's current approach avoids the worst of the biases found in some other countries, where limited taxation of income or gains combined with full tax deductibility of mortgage expenses encourages people to over-invest in housing and take on too much household debt.

Retirement savings are also generally lightly taxed around the world. Many OECD countries tax retirement benefits at a person's marginal tax rate, and exempt contributions and earnings. In Australia, retirement savings are also taxed lightly but in a different manner — as both contributions and earnings are taxed at low rates while superannuation benefits are generally tax-exempt when paid after the age of 60.


Superannuation and owner-occupied housing should continue to be taxed at relatively low rates or be exempt from income tax, consistent with an expenditure tax benchmark.

Other savings income should continue to be subject to income tax.

Personal income tax rates

Options for achieving a progressive personal income tax rates scale

A progressive income tax is characterised by average rates that rise with income, in line with the idea that reductions in income caused by taxation reduce the wellbeing of low-income earners more than high-income earners. This means that higher-income people bear a greater than proportional share of the tax burden.

Progressivity can be achieved either through a flat tax rate with a tax-free threshold, a rising personal income tax rates scale, or a combination of both. Progressivity does not necessarily require increasing effective marginal tax rates, as illustrated in Chart A1–7.

Chart A1–7: Increasing average tax rate for a single person without children, 2009–10

Chart A1–7: Increasing average tax rate for a single person without children, 2009–10

Source: Treasury estimates.

Imposing higher average tax rates on those with greater capacity to pay is typically better targeted if it is done through a progressive income tax system rather than through avenues that indirectly target income, such as carve-outs from the GST base or higher taxes on 'luxury' goods like cars and wine. Higher wage earners tend to vary their labour supply less than lower wage earners in the face of taxation (Breunig et al. 2008), so differential tax rates can also be less distortionary than flat rates. The overall progressivity of the tax system is reduced by other flat rate taxes, which makes progressivity in the personal income tax more important.

The redistributive goals of progressive taxation need to be weighed against the effects that progressive taxes have on incentives to invest in education, training and skills and to engage in entrepreneurial activity. Even with strong preferences for redistribution, steeply rising marginal rates at the top of the income distribution will be counter-productive — it only makes sense to tax people to the extent that they are still willing to work or engage in entrepreneurial activity.2 A recent OECD report found that 'high top statutory income taxes reduce the post-tax income of a successful entrepreneur relative to an unsuccessful one and can reduce entrepreneurial activity and TFP (total factor productivity) growth' (Johansson et al. 2008). Increases in top marginal tax rates must therefore balance the desire for progressivity with the impact this may have on economic growth.

Progressive taxes can make income splitting more attractive, and give people incentives to manipulate the timing of large income amounts, as different patterns of income receipt will result in different tax liabilities. These effects can be mitigated by provisions that deal with alienation of income (see Section A1–2 and Section A1–3), and by adopting approaches similar to accrual accounting that lessen the tax impact of timing differences in the receipt of income.


Personal income tax should be progressive, both through its own rates scale and also in combination with transfer payments.

Australia's progressive personal income tax system

The progressivity of a tax system can be assessed in various ways. A relatively straightforward approach is to compare the marginal and average tax rates inherent in the personal income tax rates scale at a particular point in the income distribution. The OECD commonly makes such an assessment at different points, illustrating how point measures of progressivity are sensitive to where they are evaluated.3 For example, for a single person in 2008, the Australian tax system was the 7th most progressive in the OECD if they were earning 67 per cent of the average wage, 20th most progressive at 100 per cent, and 11th most progressive at 167 per cent. Using this same measure, since 2000 the tax system has become slightly more progressive at 67 per cent of the average wage, and slightly less progressive at 100 per cent and 167 per cent.

Complementary measures enable progressivity to be evaluated across the whole system, not only at specific income levels. As well as hypothetical calculations, actual outcomes can be assessed using empirical data.

Administrative data show that the tax-free threshold and rising marginal rates of the existing personal income tax system deliver progressive outcomes in Australia. Chart A1–8 compares how taxable income and tax paid are spread across the population, after ranking everyone according to their taxable income.4 Taxable income is unequally distributed across the population, with the top 20 per cent of taxfilers receiving 49 per cent of all taxable income. However, the income tax burden is even more concentrated, with the top 20 per cent of taxfilers paying 67 per cent of all personal income tax. Similarly, the bottom 20 per cent of taxfilers receive 3 per cent of all taxable income, but pay only 0.1 per cent of all personal income tax. This means that post-tax outcomes are more evenly distributed than pre-tax outcomes.

Chart A1–8: Distribution of taxable income and tax paid, 2007–08

Chart A1–8: Distribution of taxable income and tax paid, 2007–08

Excludes people with negative or zero taxable income.

Source: Australian government administrative data.

However, the progressivity of personal income tax also depends on how comprehensively income is assessed. The provision of tax offsets, concessions and exemptions affects the personal income tax rates scale that different individuals face, and hence changes the progressivity of the system. The more exemptions in the tax base, the weaker is taxable income as a guide to a person's actual income. Compromises to the tax base include income received in forms that are taxed more lightly; for example, from salary sacrificing into superannuation and from splitting income with others to avoid higher tax rates.

Alongside the tax system, transfer payments are another key mechanism for delivering progressivity. In contrast to many other countries, which have social insurance systems that pay benefits based on a person's previous earnings, Australia has a targeted transfer system focused on poverty alleviation. This delivers strongly progressive outcomes — Chart A1–9 compares the share of transfers paid to the richest half of the population in the OECD countries in 2005.

Chart A1–9: Share of transfers paid to the richest half of the population, 2005

Chart A1–9: Share of transfers paid to the richest half of the population, 2005

Source: Whiteford 2009.


Overall, Australia has a progressive personal income tax system. The personal income tax and transfer system taken together is among the most progressive in the OECD.

Setting tax rates to support workforce participation

Rates of tax are one of the key factors in determining incentives to work and save. For many taxpayers, tax rates are a more visible part of the tax system than other key design elements, such as the way that income is defined for tax purposes.

As a matter of principle, taxes should interfere as little as possible with work incentives, as this leaves society as well off as possible. In practice, people can avoid taxes by earning less, and this is more costly to society than if the person was willing to work more and pay more tax. For example, a taxpayer can decide to work fewer hours than they otherwise might, or a person who receives income support can elect not to work at all to prevent withdrawal of their payment. People may respond to taxes in ways other than simply adjusting their hours of work. They may alter their education or entrepreneurial plans, or the form in which they receive income.

A large body of literature has explored how best to set tax rates to meet a government's needs for revenue while minimising the disincentive effects of taxes and taking account of societal preferences for redistribution; for example, Diamond (1998), Saez (2001), Moffitt (2008), Brewer, Saez and Shephard (2008). One of the key findings is that it can be more efficient to impose higher tax rates where fewer people are subject to them, such as at very low and very high incomes. Another insight from this literature is that it can be more efficient to impose higher rates on people whose behaviour is relatively unresponsive to tax rates, such as prime-aged men and women who are not caring for dependent children.

People respond to tax and to financial incentives delivered through both the tax and the transfer systems. For people who are able and expected to work full time, the progression from unemployment to self-support through work can involve a high effective tax rate. Relatively high effective tax rates on low earnings, such as earnings from part-time work, can encourage people to choose full time work to get a lower overall rate and a higher disposable income. By contrast, people who have limited capacity or limited availability for work may only ever seek part-time work. This could be due to caring responsibilities, disability or impairment, or age. People with such restrictions who work part-time may respond to a high effective tax rate by withdrawing from work altogether. A more efficient arrangement in those circumstances is to impose a lower effective tax rate on modest earnings. This could be delivered through the tax rate only, the income support withdrawal rate only, or a combination of the two. Greater certainty and transparency result from varying only the withdrawal rates rather than using the tax system as well.

If more people currently outside the workforce worked part or full time, this could help meet the challenges of an ageing population. This would be likely to require more employment services and other support, alongside financial incentives, for people who are sick or disabled, their carers, aged people and those who are engaged in home duties or the care of children (Abhayaratna et al. 2006).

Arrangements to support the employment of people who are sick or disabled are discussed in Section F The transfer system. The workforce participation of those who are engaged in home duties and the care of children is also discussed in that section. A key element of the personal income tax system that supports workforce participation is the unit of assessment.

Tax rates and withdrawal rates can have the same economic effect

In considering the incentive effects of the system, it is important to consider the combined impact of the personal tax system and the withdrawal rates applying to means tested benefits. This is because withdrawal rates can have the same economic impact as tax rates — the effect on a person's disposable income is the same whether part of a payment is withdrawn or an additional amount of tax is collected. For example, the pension assets test acts like a tax on savings, and can affect savings decisions in the same way.

While in general tax rates and withdrawal rates should have the same impact on decisions, the impact is not identical where they use different income bases. In addition, timing differences can also alter the effective tax rate at a point in time by comparison with the final effective tax rate after a tax assessment. The practice of taxing and making payments to people at the same time ('churn') can be criticised on the grounds of administrative cost, but has the advantage of allowing governments to target taxes and transfers with much greater precision than would be possible if it simply reduced tax liabilities. Taxing and making payments at the same time allows the tax and transfer systems to reflect work responsiveness, the presence of children, and other characteristics.

The impact of taxes and withdrawal rates may also differ because an individual may react differently to having their earned income taxed compared to having a transfer payment reduced, even though the effect on their net disposable income is the same.

These different characteristics of tax and withdrawal rates, and of the tax and transfer systems, suggest that decisions about imposing tax rates through the tax and transfer systems should consider the relative strengths of the two systems.


The tax system should limit the extent to which people face high effective tax rates, particularly for those who are most likely to reduce their work effort as a result.

Effective tax rates should be tailored to individual circumstances to support workforce participation for those who are able to work and choose to do so. Tailoring should be achieved through adjusting withdrawal rates on transfer payments rather than through tax mechanisms.

Supporting work in an ageing population

Levying taxes efficiently is likely to become increasingly important as the population ages and there are fewer working people as a proportion of the population. Over the next four decades, the retirement age population is expected to grow faster than the working age population. By 2049, over one fifth of the population is projected to be aged 65 or more, compared to around 13 per cent in 2009. As Chart A1–10 illustrates, a corresponding reduction is expected in the relative size of the working age population, and this suggests that economic growth will slow.

Chart A1–10: Historic and projected labour force participation rates

Chart A1–10: Historic and projected labour force participation rates

Source: Treasury projections.

Australia's current workforce participation rates are high compared to those in the past, at 65.4 per cent in 2008 compared to 60.8 per cent in 1979 (ABS 2009f). Maintaining high levels in the future will require a tax and transfer system that supports work. The Council of Australian Governments (COAG) made the following comment on this issue:

… with an ageing population, there will be relatively fewer Australians of working age. To avoid putting too great a burden on those already in work, more Australians need to realise their potential by entering or rejoining the workforce (COAG 2006).

Key groups where Australian participation rates are relatively low compared to other OECD countries include prime aged men, women of child-bearing age, and older men and women.5 This suggests that incentives for existing workers to remain in work are critical. In addition, increases in participation by those not currently working should be supported and encouraged, whether they are not currently working because of illness or disability, caring responsibilities, age, home duties or the care of children.

High effective tax rates reduce incentives to work and save

The personal income tax rates scale is often the most visible component of the effective tax rates that people face. However, other parts of the tax system can raise effective tax rates above the marginal rates in the tax scale. The Medicare levy collects 1.5 per cent of income, and is phased in at a 10 per cent rate over an income range that is not announced until the end of the tax year. Means tested offsets, such as the senior Australians tax offset and the low income tax offset, also increase effective tax rates when they are being withdrawn.

Average tax rates in Australia are in the bottom third of the OECD for single people at 67 per cent of the average wage and at 100 per cent, and are still below the OECD average at 167 per cent. The top marginal rate is in the bottom half of those in OECD countries, and the corresponding threshold is set slightly above the OECD average in terms of multiples of average earnings (OECD 2009d).6

Other elements of the tax system can result in very high effective tax rates at particular points (such as thresholds for HELP repayments and the Medicare levy surcharge). Crossing these thresholds results in a higher rate of tax being levied on every dollar of income, not only on income over the threshold. This means that people's disposable income can fall even though their private income has increased. Work by Chapman and Leigh identified a statistically significant degree of 'bunching' of incomes slightly below the HECS thresholds, suggesting that these very high effective tax rates do have an impact on behaviour (Chapman & Leigh 2006).

The transfer system overlays additional financial incentives on the tax system, because payment withdrawal rates interact with taxes. Recent studies of effective marginal tax rates (EMTRs) in Australia suggest that around 90 per cent of working age Australians face EMTRs below 40 per cent (Harding et al. 2006; Kalb 2007). (EMTRs measure the proportion of an extra dollar of income that is lost due to taxes and transfer withdrawals.) Due to the widening of eligibility for means tested family assistance, the proportion of working age Australians facing EMTRs over 50 per cent increased between 1996–97 and 2006–07, from 4.8 per cent to 7.1 per cent. However, the proportion facing EMTRs over 80 per cent declined over this period.

EMTRs do not give a complete picture of the incentive effects of the tax and transfer systems. It is difficult to fully capture these incentives, which may also be affected by factors such as child care costs, public housing rent-setting, and child support liabilities or receipts. More broadly, consumption and payroll taxes also affect the returns to work, and even corporate income taxes may be borne at least in part by workers (see Section B1 Company and other investment taxes).

In certain situations, EMTRs may not be an appropriate measure of the returns to work. For example, a person out of work may be less influenced by the effective tax rate on a small increase in earnings than by the effective tax rate when they move from not working to working — a much larger increase in private income. Effective tax rates on these larger increases in private income are often called participation tax rates (PTRs).

Research looking at the labour market transitions of Australian families over time found that PTRs have a moderate negative effect on the probability that women will enter employment, and a very large negative effect on the probability that an unemployed person will find work (Dockery et al. 2007). This may be a particular concern for jobless couple families with children, who can face high PTRs when one member takes up work. Among couple families with children under 15 where the woman is not working, around 19 per cent of the men are also out of work (ABS 2009g). This is in sharp contrast to males in couples generally, who have markedly higher rates of employment.

Chart A1–11: Participation tax rates

Chart A1-11: Participation tax rates

Source: Treasury estimates.

Chart A1–11 shows that, for an adult in a jobless couple family with two children, more than 58 per cent of their pay will be lost to tax and payment withdrawal if one member takes a job at the minimum wage. However, these high tax rates allow tax rates to be lower elsewhere in the system, which means that the overall effect on incentives is unclear.

To gain a clearer picture of the incentive effects of the tax and transfer systems, measures of effective tax rates need to be combined with empirical research on the responsiveness of the people who face them. The Australian Fair Pay Commission recently commissioned research into how much certain groups know about the impacts of the tax and transfer systems, and their motivation to work. This work established that people have limited theoretical understanding of how transfers are affected by changes in income, but also that once people are in receipt of a transfer payment, they may protect their entitlements by avoiding work that would move them off benefits.

Dandie and Mercante (2007) reviewed the literature on the responsiveness of various Australian groups, and found that partnered men, single men and single women without children are generally less responsive to changes in wages than partnered women. Lone parents tend to be more responsive than partnered women. Responsiveness varies according to factors such as level of education (higher responsiveness for those with lower education levels), whether the individual works part-time or full-time (higher responsiveness for part-time workers), and income level (generally higher responsiveness for those with lower incomes).


Effective tax rates can be high for some people, including for those likely to reduce their level of work as a result.

Reform directions — improve simplicity and incentives with a high tax-free threshold and a constant marginal rate for most people

Recommendation 2

Progressivity in the tax and transfer systems should be delivered through the personal income tax rates scale and transfer payments. A high tax-free threshold with a constant marginal rate for most people should be introduced to provide greater transparency and simplicity.

The personal income tax rates scale is a key contributor to progressivity in the tax and transfer systems.

A new personal income tax rates scale would have a high tax-free threshold and a constant marginal rate for most people. This could take the form of a constant rate of tax for most taxpayers, with a higher rate for those on very high incomes. An indicative approach to implementing the personal income tax rates scale for Australian residents is shown in Table A1–1. The indicative scales shown in the table result in lower personal taxes for people with low incomes, and give rise to broadly comparable average tax rates for those with taxable incomes up to $100,000.

Table A1–1: Indicative personal income tax rates scale

Taxable income ($) Rate (%)
0 – 25,000 0
25,001 – 180,000 35
180,001 + 45

This approach sets the tax-free threshold at $25,000, where income support recipients would either have exhausted their payments or have substantial private income. This would mean that more than 1.2 million additional people would no longer pay tax — over 10 per cent of current taxpayers. Many of these would not have to file a tax return (although some would continue to do so to claim withheld amounts or imputation credits). Setting the tax-free threshold at this level would remove the need for the low income tax offset and limit the need for the senior Australians tax offset.

Above the tax-free threshold, a constant rate of 35 per cent would apply for most taxpayers. In the example provided in Table A1–1, over 97 per cent of people over the tax-free threshold would be subject to the 35 per cent rate of tax. A constant rate of tax of this kind has the advantage of transparency for most working people. Combined with a tax exemption for transfer payments, it would be much easier for people to understand their marginal rate of tax.

A higher rate of tax could be applied to those on around three times average wages — $180,000 in this example. A top marginal rate that began at this multiple of average wages would be slightly above the OECD average, although internationally there is a high degree of variation in the level at which top marginal rates apply. For example, in the United States, the top marginal rate applies from nine times average wages, while in the United Kingdom it is 1.2 times average wages. Countries with top personal marginal tax rates that apply from around three times average wages include Canada, France, Italy and Korea.

This indicative personal income tax rates scale broadly reflects the aspirational tax cuts proposed by the Government for introduction in 2013–14. Introducing a rates scale of this kind would have a number of advantages. It would provide a higher level of transparency to individual taxpayers, as the great majority would have a single marginal rate. It would also improve the relationship between the tax and transfer systems: allied with a tax exemption for transfer payments, more people would be in only one system at any given time.

A tax scale of this kind could be implemented gradually, taking into account existing settings on marginal tax rates, offsets and the definition of income.

Incentives to work

A personal income tax system that provides more support to workforce participation should be delivered in a transparent way.

Currently, most taxpayers have more tax withheld throughout the year than is necessary, because part of the effective tax-free threshold is given through the low income tax offset (LITO) and is only available after the taxpayer files their tax return. By incorporating LITO into the explicit tax scale, people would receive better financial returns to work throughout the year, strengthening participation incentives.

As well as incorporating the existing LITO into the tax scale, a substantial increase in the tax-free threshold would increase the attractiveness of work to low-income earners (including secondary earners), who are typically more responsive to effective tax rates.

Reconfiguring the dependency offsets would better target support to those unable or not expected to work, which would improve participation incentives for those secondary earners not in these categories.

These changes build upon those proposed for the transfer system. Together, these reforms would better support employment and position Australia to meet the coming demographic challenges.

Taxing people as individuals

In designing a personal tax system based on anything other than a strictly flat rate of tax, a fundamental choice has to be made about the unit of assessment — that is, whether people are taxed as individuals or as part of a couple or family. This choice involves judgments about how people in couples operate in society compared to single people and about the needs of other family members, particularly children.

The key consideration in determining the unit of assessment is how it gives effect to contemporary social norms about individuals and couples. The judgment implied in the choice of unit is whether horizontal equity is concerned with treating individuals or couples in like circumstances alike. It has particular practical implications for workforce participation.

Specifically, the unit of assessment determines the marginal tax rate that each person in a couple faces. There are advantages in having each partner face different marginal tax rates, according to their earnings and other characteristics. For example, in a couple where one partner is the primary earner and the other earns less, perhaps working part-time and caring for children, imposing the same marginal tax rate on both may cause the secondary earner to reduce their work effort. By contrast, a lower marginal rate for that person may encourage and support work.

This observation is supported by an extensive body of research on how responsive people are to financial incentives in determining how much to work and earn. Research shows that, in couples, women are typically more responsive to tax rates than men, and lone parents are often found to be more responsive still. Table A1–2 presents a summary of findings from Australian studies. A progressive individual tax system, with resulting lower tax rates for typically female secondary earners, is therefore more efficient than family taxation. In a similar vein, it is efficient for withdrawal rates on income support payments to take account of the fact that different groups have different levels of responsiveness to financial incentives.

Table A1–2: Uncompensated wage elasticities for Australia by population group(a)

Population group Range Mean
Married men –0.19 to 0.26 0.00
Married women –0.19 to 1.3 0.30
Single men 0.28 0.28
Single women 0.34 0.34
Lone parents –0.15 to 1.48 0.52
  1. Table summarises the range of wage elasticity estimates from Australian studies.

Source: Dandie, S and Mercante, J (2007) p. 37.

Related to this, a progressive income tax levied on an individual basis corrects in part for the bias towards unpaid home production. For example, a couple where both partners are working has access to two tax-free thresholds, while a single-income family with more opportunity for home production only has access to one tax-free threshold.

There are other considerations in determining the appropriate unit of taxation. Stability over time is a factor: families change over time, as people partner and separate, and society's conception of what constitutes a couple or family also changes. The robustness of the unit of assessment is also a consideration. Given the changes over time in how couples and families are defined, there can be a level of uncertainty about whether a person is single or partnered. How much this matters depends partly on how much more favourable it is for a person to be assessed as partnered. The robustness of the unit is also relevant to administration and compliance costs. For example, it is more difficult to implement pre-filling of tax returns where tax liabilities depend upon partnering status.

For these reasons, the tax system should be based on an individual unit of assessment. However, a progressive income tax levied on an individual basis is not without difficulty. Income splitting becomes attractive: larger differences in marginal rates between partners create larger incentives to hold income-yielding assets in the name of the person who is taxed more lightly, or to split income. These effects can be mitigated by provisions that deal with the alienation of income (see Section A1–2 and Section A1–3).

The effect of taking the individual as the unit of assessment is that there is no recognition of the differences in capacity to pay that arise from a taxpayer's responsibility to support adult dependants. The presence of adult dependants can arise in a range of circumstances — notably distinguished by whether or not the dependant is able to work and derive their own income. These considerations are generally best addressed through the transfer system, while some horizontal equity benefits can be provided by dependant offsets.


The personal income tax system should generally tax people as individuals.

Retain the individual as the primary unit of assessment

The Australian tax system has always been based on individual assessment. This is one of the most important ways in which the personal tax system supports participation, by allowing different marginal tax rates to apply to each person in a couple. However, the system does include some elements that take account of the presence of a partner or children, and their circumstances.

Dependency offsets are examples of tax provisions that take account of partner or family circumstances. The senior Australians tax offset allows any unused value to be claimed by a person's partner, if they have one. The Medicare levy low income phase-in arrangements take account of family size and structure. The Medicare levy surcharge's thresholds are based on family size and structure. The spouse superannuation contributions tax offset is available for contributions on behalf of a spouse, while the medical expenses tax offset allows claims for family members as well as for the taxpayer themselves. These provisions are discussed in more detail in Annex A1.

These provisions depart from the principle that tax should be levied on each individual separately. They are a source of complexity in the system, often because the tax system does not routinely collect spouse information — a factor that can make compliance activity difficult. These provisions also tend not to provide responsive assistance in those cases where they are intended to support the costs of living. In many cases, more targeted support is available through the transfer system or other spending programs. They can also have a negative impact on participation incentives, where they affect dependants who could otherwise work.

There could be a case for optional couple assessment for people of retirement age or of late retirement age, on the grounds that these people are not expected to work and the great majority do not. Joint assessment would provide the same benefits to couples who have not shared their assets equally or where one member receives a superannuation pension from a defined benefit scheme, as for couples who have split their assets equally. Such a proposal would, however, introduce significant complexity into the tax system, by requiring the Australian Taxation Office (ATO) to assess relationships and changes in relationship status, as is currently required in the transfer system. While it is more complex, where participation incentives are not important, relaxation of the individual unit of assessment can assist other policy objectives.


The current tax system is generally based on taxing people as individuals. However, some provisions take account of couple or family circumstances.

Individual assessment supports workforce participation by secondary earners, by allowing different effective tax rates for each person in a couple.

Where participation incentives are not important, relaxation of the individual unit of assessment can assist other policy objectives.

Recommendation 3

The primary unit in the personal tax system should continue to be the individual, and subsidies for dependants through the tax system should be restricted (see Recommendation 6a). However, there could be a case for optional couple assessment for people of late retirement age.

The taxation of transfer payments

Many Australians receive transfer payments, often at the same time as they pay tax. Income support and supplementary payments replace or supplement wages and salary for their recipients.

Commonwealth transfer payments are cash payments provided by the Australian government to individuals and families, including Age pensioners, veterans, people with a disability, carers, unemployed people, and people affected by natural disaster. Transfers play a vital role in the government's redistributive policies and take a variety of forms, from income support and supplementary payments to cash payments for families with children.

Reflecting their poverty alleviation objectives and redistributive goals, cash transfers are typically targeted at low-income individuals and are designed to help recipients pay for daily living expenses and otherwise support themselves. Income support payments in particular assist poor households or those likely to fall into poverty without the transfer.

The rate of income support includes the base payment and any supplementary payments, such as Rent Assistance, Telephone Allowance and Pharmaceutical Allowance. These components should be treated on a consistent basis for tax purposes. The same treatment should also apply to government scholarships.

Family assistance has some different characteristics. It is not wage-like in its nature, as it is paid in addition to wages or income support for costs associated with children. Its tax status need not be the same as the tax status of income support.

Transfer payments have a mix of tax treatments

The current system exempts some transfer payments from income tax but taxes others.

Most income support payments are taxable, including Newstart Allowance, the Age Pension and Parenting Payment. The pensioner and beneficiary tax offsets remove the tax liability of recipients who receive the maximum rate of income support payment for the full year. A tax exemption applies to Disability Support Pension (if the recipient is under Age Pension age), Wife Pension (if both spouses are under Age Pension age), and Carer Payment (if the carer and person being cared for are under Age Pension age). The combined effect of the pensioner or beneficiary tax offset plus the low income tax offset and the Medicare levy low income phase-in is that there is only minimal difference in final outcomes. A mix of taxable and non-taxable payments with these additional provisions is a complex and non-transparent way of delivering effectively the same outcome.

A key difference between taxable and non-taxable income support payments in the current system is the outcome for people whose circumstances change significantly in the course of a tax year. If a person receives income support for part of a year and has a well-paid job for the other part of the year, a taxable income support payment is partially or fully clawed back through tax, while a non-taxable income support payment is not. This is because the taxable income support payment is added to the income from work. The impact of this is greater where the variation in income level is high, and was important when seasonal work was a larger component of the work available.

Supplementary payments are, in effect, part of the rate of income support. These payments include Rent Assistance, Telephone Allowance and Pharmaceutical Allowance. While most income support payments are taxable, supplementary payments are mostly non-taxable. Even parts of a payment can have a mixed treatment — the Pension Supplement (a single payment for pensioners) has both taxable and non-taxable elements.

Family assistance payments, including Family Tax Benefit Part A, Child Care Benefit, the Child Care Rebate and the Baby Bonus, are not taxable. These payments are unlike other payments in that Family Tax Benefit Part A, Child Care Benefit and the Child Care Rebate are assessed annually on a variant of taxable income, while the Baby Bonus takes the same assessment concept but applies it to the half year preceding the baby's birth. These payments address the direct costs of children and so should not be taxable.


Income support and supplementary payments have a variety of tax treatments. Some payments have taxable and non-taxable components.

Recipients of taxable and non-taxable income support payments have similar levels of disposable income, once the effects of offsets are taken into account.

Family assistance payments are tax-exempt and address the direct costs of children.

Reform directions — consistent tax treatment of income support and supplementary payments

Recommendation 4

Income support and supplementary payments should be tax-exempt.

  1. Family assistance should remain exempt from tax because it addresses direct costs associated with children.
  2. Government payments that are similar in nature to income support, such as scholarships, should be exempt from tax to align their treatment with that of income support.

A key reform direction is to provide a consistent tax treatment for pensions, allowances and supplementary payments. All income support and supplementary payments should be exempt from tax. One of the main objectives of cash transfer payments is to increase poor households' real income, and taxing transfer payments can interfere with this objective. Taxing transfer payments also complicates individuals' interaction with the tax and transfer systems.

It should be noted that exempting payments from tax has a different impact from taxing most income support and providing an offset to the tax liability for maximum-rate full-year recipients. These two treatments have different effects because the assessment period for tax and for transfers is different; if they were the same, changes in the timing of income receipt or eligibility for transfer payments would not lead to varying tax outcomes. The income on which payments are assessed is also different, meaning that a person's income is counted in different ways in the two systems.

Making all income support and transfer payments non-taxable may result in some income support recipients, particularly pensioners, facing high EMTRs on income from certain sources where the means test withdrawal rates overlap with tax rates on income above the tax-free threshold. The resulting EMTRs would be higher when income support is non-taxable because withdrawal and tax rates would be additive, rather than offsetting each other as is the case when income support and transfer payments are taxable.

For tax-exempt income support payments, the effective tax rate is the sum of the withdrawal rate and the tax rate the recipient is paying on other income. This can result in very high effective tax rates: for example, a 65 per cent withdrawal rate and a 35 per cent tax rate combine to give an effective tax rate of 100 per cent. To reduce these very high effective tax rates, income support payments can be withdrawn at a slower rate at income levels where it is likely that the person is also paying tax, to maintain a desired overall effective tax rate. The fact that the tax and transfer systems have different periods of assessment would make this difficult to achieve precisely, but it would have participation benefits.

There may need to be some offsetting adjustments in the relevant means test to minimise the impact of this overlap for people with taxable income above the tax-free threshold.

Family assistance payments should remain exempt from tax. These payments are not wage-like in nature but are intended to address private and domestic costs associated with children.

The government will need to review the level of the tax-free threshold periodically to maintain a relationship between the tax and transfer systems that has the simple and transparent character of this proposal.

Tax offsets

Tax offsets provide a mechanism for delivering lower net taxes to taxpayers with particular characteristics or types of income. However, the design differences in the large number of tax offsets add significant complexity to the tax system. People's interactions with the tax system would be greatly simplified by rationalising the number of offsets. This would also provide a simpler and more transparent marginal tax rate structure.

Tax offsets reduce the transparency of the tax system

Tax offsets7 are used in the tax system for a number of purposes:

  • to provide concessional tax treatment for some forms of income over others — for example, employment termination payment tax offsets;
  • to provide concessional tax treatment of income received by particular groups of taxpayers relative to others — for example, the senior Australians tax offset (SATO); and
  • to reduce interactions between taxable transfer payments and the tax system — through the beneficiary tax offset (BTO) and the pensioner tax offset (PTO).

At present there are more than 40 tax offsets, with different design features and impacts on people. Non-refundable tax offsets reduce the amount of tax that is payable on an individual's taxable income by the dollar value of the offset. The full amount of these offsets can only be utilised where there is sufficient tax liability — if the offset is larger than the person's tax liability no refund is available. Refundable tax offsets provide the full amount of the offset to the individual regardless of tax liability (that is, they can reduce tax to zero and create a refund).

The combination of all offsets, with their different interactions and eligibility criteria, contributes significantly to complexity in the tax system. Rationalising offsets could make the system simpler and reduce compliance costs. Many of the objectives of the current offsets could be (or have already been) achieved more effectively if delivered through transfer payments, other government spending, or through direct remuneration.

Some tax offsets are structural — that is, they alter the personal income tax rates scale for the majority or a large number of taxpayers. For example, in 2009–10 the low income tax offset (LITO) of $1,350 increases the effective tax-free threshold to $15,000 and changes the effective marginal tax rates for people with incomes between $30,000 and $63,750.

Most other offsets provide concessional treatment to a smaller group of people in specific circumstances. As concessional tax offsets are usually delivered on assessment, they generally do not deliver assistance to taxpayers at the time that the relevant expenses are incurred, are not transparent because they are not directly related to the incurring of the expenses, and they are generally targeted only at those people who have a tax liability.

The objectives of the current set of offsets could be achieved more simply and effectively if they were rationalised in the following way:

  • Structural offsets (such as the LITO, SATO, PTO and BTO) should largely be removed. A higher tax-free threshold and adjustments to the personal income tax rates scale would facilitate this.
  • Concessional offsets, which have in many cases been replaced by direct transfer payments or other government spending, should in most cases, be removed from the tax system. Exceptions should apply where the dependant is unable to work due to disability or carer responsibilities, or either the taxpayer or dependant has reached Age Pension age.


Structural tax offsets alter the personal income tax rates scale for a large number of taxpayers and create complex interactions between the tax and transfer systems. The assistance provided by structural tax offsets would be more simply and transparently delivered through explicit marginal tax rates.

Concessional tax offsets provide a mechanism for delivering lower net tax rates to taxpayers with particular characteristics. However, assistance provided in this way is not transparent, timely or well targeted.

Medicare levy complicates personal income tax

While the Medicare levy is designed to help fund Medicare expenditure, it only partially funds Medicare, which in turn constitutes only a fraction of total government health spending. Of the $71.2 billion spent on health by Australian, State and local governments in 2007–08, only $7.4 billion was funded by the Medicare levy. In June 2009 the National Health and Hospitals Reform Commission (NHHRC) recommended that the levy be increased by 0.75 percentage points to finance its proposed Denticare scheme (NHHRC 2009).

The Medicare levy raises the marginal tax rate for most Australian residents by 1.5 percentage points. However, the levy does not apply to all taxpayers and it interacts with the marginal tax rates in complex ways, creating high effective tax rates at some income levels.

A complex set of low-income phase-in arrangements operates to provide an exemption from the Medicare levy for people without a tax liability (treating couples and singles differently). The complexity of these arrangements and the income levels at which they are phased in make it difficult to avoid stacking of tax rates and withdrawal rates.

In addition, many people are exempt from paying the levy based on their personal circumstances. For example, members of the Australian Defence Force and non-residents are exempt. As a result of the phase-in and exemption arrangements, in 2007–08 only 75 per cent of the 11.4 million taxpayers with a gross tax liability paid the levy.

The levy may send a misleading message to taxpayers about the cost of health spending. This may encourage inconsistent demands for more public funding of health care combined with an expectation that this can be absorbed without higher rates of tax.

The Medicare levy should be removed and incorporated into the personal income tax rates scale. This would simplify the tax system and remove potentially misleading messages to taxpayers about the cost of health spending.

However, to increase the transparency of the costs of health, a share of revenue raised from personal income tax could be allocated to health expenditure. This allocation could be made whether or not the funds were hypothecated formally to health. Total government health spending accounted for around 56 per cent of personal income tax revenue in 2007–08 (based on tax revenue of $126.1 billion), increasing to 62 per cent in 2008–09 (based on estimated tax revenue of $125.8 billion). This could be applied as a proportion of the net tax payable by an individual. This option would be simpler and raise revenue on the more efficient personal income tax base.

Medicare levy surcharge and private health insurance

To increase the take-up of private health insurance, the Medicare levy surcharge requires individuals with an income for surcharge purposes over $73,000 and families with a combined income for surcharge purposes over $146,000 (increased by $1,500 for each dependent child after the first) in 2009–10 to pay an additional 1 per cent tax on their taxable income (including reportable fringe benefits) if they do not have complying health insurance for themselves and all their dependants. The singles threshold is indexed to AWOTE and increased in $1,000 increments (rounded down). The threshold for families is double the singles threshold. While the surcharge is designed to be entirely avoidable (by purchasing the required insurance cover), it was levied on 725,000 individuals and raised revenue of around $450 million in 2007–08.

As it currently operates, the Medicare levy surcharge is not ideal. Although levied on individuals, it is calculated on a family basis (by considering the presence of a spouse and the number, age and study status of any children). This means that the surcharge is levied on a high-income individual with insurance whose spouse does not have insurance. This complicates the system and makes compliance more difficult. It also creates spikes in EMTRs as it applies to every dollar of taxable income, including reportable fringe benefits, once the relevant income threshold is exceeded (rather than only the amount in excess of the threshold). The name of the surcharge is also misleading as it is not related to the Medicare levy and does not reflect its link with private health insurance. As a result, it should be relabelled to reflect this link.

The surcharge is closely linked with the private health insurance offset as part of a package of government polices aimed at increasing the take up of private health insurance. The offset is also problematic as it can be claimed using multiple mechanisms, including through the tax system, making the system unnecessarily complex and costly.


Tax arrangements relating to private health insurance, including the Medicare levy surcharge and the private health insurance tax offset, are unnecessarily complex.

Reform directions

Recommendation 5

The Medicare levy and structural tax offsets — the low income, senior Australians, pensioner and beneficiary tax offsets — should be removed as separate components of the system and incorporated into the personal income tax rates scale. If a health levy is to be retained, it could be applied as a proportion of the net tax payable by an individual.

Recommendation 6

To remove complexity and ensure government assistance is properly targeted, concessional offsets should be removed, rationalised, or replaced by outlays.

  1. The existing dependency offsets should be replaced with a single dependant tax offset where one of the following circumstances apply:
    • the dependant is unable to work due to disability or carer responsibilities; or
    • either the taxpayer or dependant has reached Age Pension age.
  2. The zone tax offset should be reviewed. If it is to be retained, it should be based on contemporary measures of remoteness.
  3. The mature age worker, employment termination payment, overseas civilian, entrepreneurs' and notional tax offsets should be removed (see Annex A1). The education tax refund should be replaced as part of the single family payment, but as a back-to-school (lump-sum) amount.
  4. The overseas forces tax offset should be replaced by adjusting remuneration to maintain net incomes.
  5. Averaging tax offsets for primary producers, the offset for 'special professionals' and the lump sum payment in arrears tax offset should be retained to minimise the extent to which the timing of such income influences tax liability (see Annex A1).

Recommendation 7

Consistent with the recommendations of the National Health and Hospitals Reform Commission:

  1. The medical expenses tax offset should be removed following a review of the scope and structure of health safety net arrangements.
  2. The Medicare levy surcharge and assistance for private health insurance should be reviewed as part of the package of tax and non-tax policies relating to private health insurance. The Medicare levy surcharge lump sum payment in arrears tax offset should be retained if the Medicare levy surcharge is retained (see Annex A1). Assistance, if retained, for private health insurance should be provided exclusively as a direct premium reduction.
Structural offsets
The low income tax offset (LITO)

The LITO is the mechanism that changes the tax-free threshold for the largest number of people, 6.8 million in 2007–08. This number will increase as the LITO rises as a result of legislated tax cuts. The full amount of LITO is available to individuals with taxable income up to $30,000, and the amount of LITO available is then reduced at the rate of four cents in the dollar. In 2009–10, the LITO has the same effect as increasing the tax-free threshold to $15,000 for those with incomes up to $30,000, increasing the 15 per cent marginal tax rate to 19 per cent (plus 1.5 per cent Medicare levy) for individuals with incomes between $30,001 and $35,000, and increasing the 30 per cent marginal tax rate to 34 per cent (plus 1.5 per cent Medicare levy) for those with income between $35,001 and $63,750 (the point at which the LITO is completely withdrawn).

The LITO not only increases the tax-free threshold, but also increases marginal tax rates at higher incomes (see Chart A1–12). As a result, many taxpayers with taxable income up to $80,000 face an effective marginal tax rate different to that set out in the personal income tax rates scale. For example, the only taxpayers who face a 30 per cent effective marginal tax rate (excluding the Medicare levy) in 2009–10 were those who earn between $63,751 and $80,000.

Chart A1–12: Impact of the LITO on personal income tax rates 2009–10

Chart A1–12: Impact of the LITO on personal income tax rates 2009–10

Note: Does not include Medicare levy.

Source: Treasury estimates.

The LITO should be incorporated into the personal income tax rates scale, both for reasons of transparency and to retain the progressivity of the personal income tax rates scale. This change would also make the benefit of the LITO fully available through the pay as you go withholding rates scale, rather than half through withholding and half on assessment, as is currently the case.

Senior Australians tax offset (SATO)

The SATO increases the effective tax-free threshold for people of Age Pension or Veterans Service Pension age. In 2006–07, approximately 623,000 people claimed the offset at a cost of $1.1 billion. In 2010–11, the SATO, when combined with the LITO, will provide an effective tax-free threshold of $30,685 for singles and $26,680 for each partner in a couple. The SATO phases out (at the rate of 12.5 cents per dollar) for income above these thresholds and will be completely phased out once income reaches $48,525 (singles) and $39,496 (for each member of a couple not separated by illness). SATO amounts ($2,230 for singles and $1,602 for each member of a couple not separated by illness) are not indexed. Unused amounts of SATO can be transferred between partners up to the point where the maximum combined offset amount has been used.

The SATO should be removed, in conjunction with the Review's recommendations that the tax-free threshold be raised substantially and that pensions and benefits be made tax-exempt. As a transitional mechanism, it should be replaced with an offset that takes into account the new personal income tax rates and thresholds and delivers a similar effective tax-free threshold. In light of the increase in the number of Australians accessing tax-free superannuation benefits, and a higher tax-free threshold, the tax concession provided by the new offset should be reduced over time.

Pensioner tax offset (PTO)

The PTO was introduced to ensure that pensioners and some allowees on maximum rates of payment do not incur a tax liability. The PTO is available to recipients of specified payments made under the Social Security Act 1991 and Veterans' Entitlements Act 1986. In 2008–09, the PTO was $2,240 for singles, $2,086 for partners in a couple who had to live apart due to illness or because one partner was in a nursing home, and $1,699 for each partner in a couple not separated by illness. The PTO, when combined with the LITO, provides an effective tax-free threshold of $25,298 for singles and $21,691 for each partner in a couple not separated by illness. In 2006–07, approximately 294,000 people claimed the offset at a cost of $459 million.

As the LITO has increased, there has been no offsetting downward adjustment to the PTO or SATO. This has pushed up the effective tax-free threshold delivered by the combination of PTO, SATO and LITO. As a result, most pensioners, whether full- or part-rate, no longer pay any tax on their combined pension and private incomes.

The PTO should be removed, in light of the recommendations to exempt all transfer payments from tax and increase the tax-free threshold. Pensioners of Age Pension age would have access to the transitional offset outlined in the SATO section above.

Beneficiary tax offset (BTO)

The BTO ensures that recipients of prescribed government payments such as allowances, drought assistance payments and wage supplements following disasters do not pay tax on the benefit or allowance component of their income. In 2006–07, approximately 279,000 people claimed the offset at a cost of $130 million.

The BTO should be removed, in light of the recommendations to exempt all transfer payments from tax and increase the tax-free threshold.

Concessional tax offsets
Dependency tax offsets

Concessional tax arrangements for dependants have been a feature of the tax system for a long time and were generally introduced at a time when spouses (and other dependants) typically depended on a main breadwinner and full-time work was the norm. There are five dependency tax offsets, which provide different levels of tax concessions to taxpayers depending on whether they support:

  • a spouse who has a very low income;
  • an invalid relative who has a very low income;
  • a child who is not employed, but is undertaking (unpaid) work as a housekeeper in the taxpayer's house;
  • the engagement of a housekeeper to care for a child under 21 years, invalid relative or spouse receiving Disability Support Pension; or
  • a dependent parent or parent-in-law who has a very low income.

The multiple dependency offsets complicate the tax system and are withdrawn from a low level of dependant's income. This can affect participation incentives and is generally only appropriate where there is less concern about the impact on participation of the dependant; for example, for dependants unable to participate due to invalidity, or for people over Age Pension age.

The dependency offsets should not be provided where the dependant is able to seek work, because in this situation the offset acts as a work disincentive. The offsets should be more narrowly focused on taxpayers supporting either a dependant who is unable to work due to disability or carer responsibilities or where the taxpayer or dependant has reached Age Pension age.

Other concessional offsets

Several other tax offsets are designed to influence behaviour. In some cases these would no longer be necessary as a result of recommended changes to the personal income tax rates scale or because assistance is already provided through the transfer system. For example, the mature age worker tax offset was intended to provide people over 55 years with an incentive for continued workforce participation. However, the recommended increase in the tax-free threshold provides a more transparent and effective participation incentive for these people.

As a general principle, offsets should be limited to circumstances where the assistance cannot be provided effectively through the transfer system or other government spending.

Further detail on existing concessional offsets and recommended reforms is presented in Annex A1.

Private health insurance tax offset

The Australian government currently subsidises private health insurance premiums based on a person's age through three mechanisms: a direct premium reduction, a reimbursement from Medicare Australia, or a tax offset. Most people claim the rebate as a direct premium reduction, with around 96 per cent of private health insurance subsidies claimed through premium reductions and Medicare Australia and around 4 per cent claimed through the tax system. In 2008–09, total expenditure on the private health insurance rebate was around $4 billion.

Providing multiple ways to claim assistance for private health insurance, including through the tax system, is unnecessarily complex and costly. If government wishes to subsidise private health insurance, assistance should only be provided as a direct premium reduction. This provides timely assistance, as it reduces the cost of insurance at the time it is paid, is simple to administer and is the most common way of claiming assistance.

Whether or not this subsidy is means tested requires a balancing of equity and complexity considerations. Means testing would help ensure this assistance is directed to those who need it most. On the other hand, it would require people to estimate their annual income when they receive assistance and then reconcile the assistance they receive against their actual income at the end of the income year. This would create a risk that inaccurate estimates of income would create debts. Means testing arrangements would also increase administrative complexity for policy holders, insurance providers and the ATO.

The National Health and Hospitals Reform Commission's (NHHRC) final report recommended that the Australian government 'commits to explore the design, benefits, risks and feasibility around the potential implementation of health and hospital plans to the governance of the Australian health system' (NHHRC 2009). This would include examining the potential role of private health insurance alongside health and hospital plans including examining any changes to the Australian government's 'regulatory, policy or financial support for private health insurance' (NHHRC 2009). As a result, tax arrangements for private health insurance, including the Medicare levy surcharge and the private health insurance tax offset, need to be assessed in light of an overall review of this sector. In keeping with this, the Review has only assessed the operation of these mechanisms in relation to the tax and transfer systems. It has not assessed the role, purpose and funding of private health insurance.

2 Theory suggests the optimal top marginal rate is the revenue maximising one, which would be zero if it began at the income level where the highest income taxpayer was no longer willing or able to earn more. See Brewer, M, Saez, E and Shephard, A (2008), Means-testing and tax rates on earnings, Institute for Fiscal Studies.

3 The measure is calculated as (1 — marginal tax rate)/(1 — average tax rate), and is reported in the annual OECD publication Taxing Wages. A variant of this was used in Arnold (2008).

4 People with negative or zero taxable income have been excluded.

5 In 2005, Australia was ranked 25th for prime aged men (aged 25 to 54 years), 23rd for women of child-bearing age (aged 25 to 44 years), and 13th and for older men and women (aged 55 to 64 years).

6 Average rates include employee social security contributions.

7 Known as tax rebates in the Income Tax Assessment Act 1936.