Australia's Future Tax System

Final Report: Detailed Analysis

Chapter A: Personal taxation

A1. Personal income tax

A1–3 Taxation of income from savings

Key points

The income from the savings of Australian residents, other than savings invested in owner-occupied housing and superannuation, should continue to be a significant part of the personal income tax base.

The income tax treatment of these household savings would be improved by applying a 40 per cent discount to most interest income, net residential rental property income, capital gains and certain interest expenses. Doing so would provide a more consistent tax outcome for income from bank deposits and bonds, shares, and rental properties, and provide a means of adjusting for the effect of inflation.

A more consistent treatment of household savings would encourage households to seek the best pre-tax return on their savings and to invest their savings in assets that best suit their circumstances and risk-preferences. It would also largely remove the current bias towards negatively geared investment in rental properties and shares and so reduce a major distortion in the rental property market.

While a discount would provide a more consistent tax outcome for savings, its introduction would also substantially change incentives in some key markets, particularly for rental housing. Given the current problems in the rental housing market, the discount for net residential rental income should only be adopted following reforms to housing supply and housing assistance.

While a move to a broad 40 per cent discount would involve further boundaries in the income tax system, at least in the short- to medium-run, some areas of the current arrangements can be simplified. In particular, capital gains tax should be simplified by excluding some low-revenue generating assets, rationalising existing concessions, removing grandfathering rules, and considering a principles-based rewrite of the rules.

A consistent treatment of savings

Challenges in the taxation of savings

Productivity is reduced if tax-induced distortions lead to a misallocation of resources, with savings directed towards less productive investment opportunities. By favouring one form of savings income over another, the tax system alters the allocation, ownership and the management of the nation's capital. This can have adverse implications for the efficiency and stability of capital markets and the way risk is distributed between individuals.

Internationally, the tax treatment of gains and losses from saving typically varies depending on the asset type, the financing arrangement and the entity or entities involved. As well as reflecting discrete policy decisions to favour particular types of saving over others, these differences arise because of the inherent difficulties in measuring economic income.

In particular, there are difficulties with measuring changes in asset values over time, which have led to changes in value only being taken into account when an asset is sold and a gain realised, and in accounting for inflation.

Realisation-based taxes distort asset allocation

Income can be measured as current consumption plus changes in wealth. Despite this, income taxes in general, and particularly for individuals, are based on the realisation principle. That is, income is recognised as taxable when it is realised through a taxable event, such as the sale of an asset, rather than as the change in value of assets or wealth over time.

Lock-in allows tax to be deferred and can disrupt the operation of markets

Taxing capital gains on a realisation basis lowers the effective tax rate on accrued capital gains by providing a tax deferral advantage — that is, the payment of tax is deferred until the gain is realised. This encourages investors to hold on to assets with accrued capital gains.

This lock-in effect can impede the efficient functioning of the capital market and distort ownership patterns as investors are discouraged from switching assets when they would pay tax on a realised gain. The lock-in effect can also destabilise the stock market and real property market as shares and property are sold when prices decline (to realise losses) and are held onto when prices rise (to defer the realisation of the gains).

Chart A1–16: Benefit from taxation on a realisation basis

Chart A1–16: Benefit from taxation on a realisation basis

Assumptions: $100 initial investment; nominal return of 6 per cent; 30 per cent tax rate on nominal income.

Source: Treasury estimates.

Chart A1–16 compares the consumption possibilities from investing $100 today in an asset according to whether it is exempt from tax, generating a capital gain with tax deferred until sale, or generating a capital gain with income taxed as it accrues (similar to interest from a bank account). The benefit from being able to defer tax under the capital gains tax provisions increases over time and provides a tax advantage over other assets, such as bank accounts.

Realisation-based methods lead to arbitrage opportunities and other problems

The adverse impact of a realisation-based capital gains tax is broader than the lock-in effect. Taxation based on the realisation principle also introduces tax arbitrage opportunities. Under a realisation-based tax, there is an incentive for an investor to hold gains and realise losses, thereby using the realisation event for tax arbitrage. Such possibilities then require limits in the tax system, such as limitations on loss utilisation even where a taxpayer incurs a true economic loss.

The realisation principle for capital gains may also create additional complexity and compliance costs. Under a realisation-based tax, taxpayers are required to keep records for long periods, and are also likely to have less frequent exposure to the relevant tax rules. Separating capital gains from other forms of income also creates uncertainty, and arbitrage opportunities, over how particular forms of income should be classified for tax purposes.

But there are practical impediments to accrual income taxation

While there could be benefits from moving towards taxing on an accruals or accruals-equivalent basis, there would also be a number of practical problems in doing so for individuals across-the-board.

The first is the need to accurately measure changes in asset values. While there are practical difficulties associated with accounting on an accruals basis for business profits and other income, unrealised capital gains (other than for assets where a market price is readily observable) are even more difficult to measure. The act of measurement could also affect market pricing.

These practical difficulties are likely to give rise to their own compliance costs and differential tax treatments of assets, depending on how easily accruing income can be measured for different assets. Different tax treatments, with only some assets taxed on an accruals or accruals-equivalent basis, would also give rise to their own tax arbitrage and minimisation arrangements.

Further, where unrealised gains accrue a taxpayer may not have the cash at hand to pay the tax liability, and borrowing against or selling down assets to meet the tax liability would not be costless. Volatility in asset prices combined with lags in tax liabilities falling due may exacerbate these concerns.

While some of the problems of valuation and liquidity arising from accruals taxation could be addressed by using an accruals-equivalent approach (for example, deeming a rate of return based on the value of an asset), this would have other disadvantages. For example, taxing savings on a deemed return or presumptive tax basis would forgo tax on any above-normal returns or economic rent.

In the longer-run, improvements in technology and changes in the operation of capital markets may mean that some of these practical impediments become less significant.


Savings should be taxed as consistently as possible to minimise tax arbitrage opportunities and to avoid biasing household and investor decisions about what assets best suit their needs and preferences.

Taxing inflationary gains erodes consumption power

The current tax system is based on nominal income. That is, the income tax base includes compensation for inflation as well as real gains. The inflationary component compensates investors for the reduction in their purchasing power arising from inflation, allowing them to purchase the same quantity of goods and services in future periods. By taxing the inflation component, an individual's consumption power is eroded.

Chart A1–17: Real effective tax rate on the return to savings under different
inflation rates

Chart A1–17: Real effective tax rate on the return to savings under different inflation rates

Assumptions: Real return of 3.5 per cent; 30 per cent tax rate on nominal income.

Source: Treasury estimates.

For example, if an individual purchases an asset for $100 and sells it a year later for $106 — earning a 6 per cent return — the full return ($6) is subject to tax. If inflation is also 6 per cent, the individual would have had no increase in consumption power — a real return of zero. That is, the same bundle of goods that cost $100 last year would cost $106 this year. By being taxed on the inflationary return the individual is no longer able to consume the same bundle of goods.

Taxing the inflation component increases the effective tax rate on savings above the statutory tax rate, which may reduce incentives to save. For a given real return, the effective tax rate increases as the inflation component increases (see Chart A1–17). The impact from taxing nominal gains may also be exacerbated under a progressive income tax where the average tax rate increases as taxable income increases.

The impact of inflation is less of an issue for capital gain assets where taxation is deferred until realisation. In this case, the real post-tax return increases the longer an asset is held. In contrast, for an interest generating asset the real after tax return does not vary with the holding period. Consequently, the argument for accounting for inflation for capital gain assets is not as strong as that for other assets (Brinner 1976).

While comprehensive adjustments can in theory be made to measure real rather than nominal income, in practice such adjustments can be very complex. A number of jurisdictions that typically face higher rates of inflation than Australia make or have made adjustments on a comprehensive basis for some items of capital income.

As price stability has been a key objective of effective Australian monetary policy settings that target a low rate of consumer price inflation, the biases caused by inflation expectations on the taxation of nominal income in Australia have been reduced.


Inflation exacerbates the biases in the current income tax treatment of savings, leading to an increase in the effective tax rate on the nominal return to savings.

Income tax applies inconsistently to different types of savings

While Australians save in a variety of ways, most household savings is concentrated in property and superannuation — both of which are either exempt or lightly taxed. According to the Australian Bureau of Statistics, the principal assets of Australian households are their own home (44 per cent of household assets), other property including rental property (16 per cent), superannuation (13 per cent), shares and interests in trusts (12 per cent), personal use assets (11 per cent) and bank accounts and bonds (4 per cent) (ABS 2007).

There are considerable differences in the distribution of the income from these different saving forms between households (see Chart A1–18). Taxable income from savings is typically skewed towards high income taxpayers. Interest income, however, tends to be more evenly distributed over the taxable income scale. Dividends and capital gains are the least evenly distributed.

In 2007–08, the bottom 20 per cent of taxpayers earned around 9 per cent of gross interest income but only 4 per cent of dividend income and around 5 per cent of net capital gains. In contrast, the top 10 per cent of taxpayers received around 27 per cent of gross interest income but over 60 per cent of net capital gains and dividends.

Chart A1–18: Distribution of savings income items, 2007–08

Chart A1–18: Distribution of savings income items, 2007–08

Source: Australian Government administrative data, includes taxfilers without a tax liability.

Tax outcomes depend on the form of saving

The tax treatment of the assets that Australian households typically invest in varies considerably. These differences arise from a long history of discrete and ad hoc government decisions as well as difficulties in properly measuring income from savings.

Before the Asprey Report in 1975, the tax laws recognised many items that fall within an economist's definition of nominal capital income: profits from a business, interest, rent, dividends and other periodic receipts. These were generally included in the calculation of taxable income and taxed at the same progressive rates as labour income.

Items that were not recognised, or were only brought into the tax base to a limited extent, included capital gains, superannuation earnings, retirement lump sum benefits, imputed rent from owner-occupied housing and consumer durables, bequests and gifts received. Of these untaxed or lightly taxed items, capital gains have been generally brought into the tax base while superannuation is now taxed as earnings accumulate in the fund. The introduction of dividend imputation was a major change in the taxation of dividends.

The different tax treatments of these assets can be expressed as effective marginal tax rates (see Chart A1–19). The estimated tax rates quantify the effect of the tax system on an investment in a specified asset that earns a normal risk-free rate of return. A zero effective tax rate represents an expenditure or consumption tax treatment; a rate equal to the statutory tax rate represents a real income tax outcome.

Chart A1–19: Real effective marginal tax rates on savings depend on asset class

Chart A1–19: Real effective marginal tax rates on savings depend on asset class

Notes: Real effective marginal tax rates show the tax levied on the normal real return to saving, and reflect the tax treatment of the income from which savings are made (where it deviates from tax payable if that income had been immediately consumed), earnings on those savings, and the final use of the accumulated savings. A zero effective tax rate corresponds to an expenditure tax benchmark, with the investment funded out of post-tax wages, and earnings and the subsequent realisation of the investment untaxed. The negative rate for superannuation reflects the reduction in tax otherwise payable on wages by making contributions out of pre-tax income. The estimates do not model interactions with the transfer system.

Assumptions: 6 per cent nominal return; 2.5 per cent inflation; for rental property, 50 per cent of the return is attributable to capital gain and 50 per cent to rental income and the rental property is held for 7 years then sold; shares are held for 7 years then sold; superannuation is held for 25 years and the individual is eligible for a tax-free payout at the end of the period.

Source: Treasury estimates.

For interest bearing deposits, the effective tax rate exceeds the taxpayer's marginal statutory rate because the entire return, including that part representing compensation for inflation, is included in taxable income as it accrues annually. Income from listed shares in companies with domestic investments benefits from imputation credits for dividends and a discount for realised capital gains. Income from foreign shares does not benefit from imputation and so has a higher effective tax rate than income from domestic shares. Rental properties benefit from the capital gains tax discount, though net rents are taxed at the full marginal tax rate.

Savings placed in lifetime savings such as superannuation and owner-occupied housing are more preferentially taxed. Owner-occupied housing is outside of the income tax base and faces a zero effective tax rate. Superannuation is advantaged because contributions into a superannuation fund are generally made out of pre-tax income (unlike for a bank account, where deposits are made out of post-tax income), though they are subject to a 15 per cent contributions tax. Earnings in the fund are also taxed at a 15 per cent statutory rate and are eligible for a one-third capital gains tax discount and for refundable imputation credits.

For superannuation, the access to the effective partial deduction for saving (or co-contributions from the Australian government, or both), the very low rate of tax on earnings and the exemption from income tax of retirement benefits, means that for many individuals saving in a superannuation fund is treated more generously than it would be under an expenditure tax.

There is considerable evidence that tax differences have large effects on which assets a household's savings are invested in. Based on an examination of the literature and OECD data, the OECD concluded that while low-income individuals respond to tax incentives with more saving, for high-income individuals in particular savings are diverted from taxable to tax-preferred savings (OECD 2007a).


The tax outcomes for different types of savings vary considerably and have evolved in an ad hoc manner. How households allocate their savings between different assets or savings vehicles is likely to be significantly affected.

Different tax treatment of financing gives rise to arbitrage opportunities

Investments in assets by individuals face different effective tax rates depending on the financing choices of the saver. When equity financed, rental properties yield a positive effective tax rate. When negatively geared, asymmetries in the treatment of expenses and receipts give rise to a more favourable treatment (see Chart A1–20). This asymmetry ranks amongst the greatest tax induced biases to the savings choices of households.

Chart A1–20: Real effective marginal tax rates on rental properties, by gearing ratio (current approach)

Chart A1–20: Real effective marginal tax rates on rental properties, by gearing ratio (current approach)

Assumptions: 6 per cent nominal return; 2.5 per cent inflation; for rental property, 50 per cent of the return is attributable to capital gain and 50 per cent to rental income and the rental property is held for seven years then sold; tax on debt provider disregarded.

Source: Treasury estimates.

For example, assume that the full amount required to purchase an investment property for $400,000 is borrowed. The return, part of which is a capital gain, is just enough to cover costs (including interest repayments). In the absence of tax the investment will break even. The same outcome would occur under an accrual-based tax without discounts, as all income and all expenditure would be pooled together and taxed at the same rate.

Under the current system, however, the same investment receives a tax advantage that allows it to do better than break even after tax. All expenses (less rents received) can be pooled and offset against other income — in full and at the individual's marginal tax rate. But any capital gain would not be taxed until realised, and if the asset is held for at least 12 months, only half the gain would be subject to tax. The same results apply for other types of geared investment that yield capital gains; in particular shares, where margin lending arrangements are used to negatively gear share investments.

The realisation principle also leads to an adverse selection bias. That is, there is an incentive to realise capital losses immediately, while deferring the realisation of accrued capital gains, which would be taxed at a discount. For this reason, the tax law quarantines capital losses, which can only be offset against other capital gains, not against other income.

Negative gearing of rental properties has become more prevalent

Households held around $700 billion of residential investment property assets in 2005–06 (ABS 2007). This represented around 14 per cent of total household assets, a proportion that has increased over the last decade.

Currently, around 70 per cent of individual investors in rental properties are in a net loss position. This figure has increased from 58 per cent in 2000–01 (see Chart A1–21). The increase largely reflects increases in interest deductions, reflecting rising levels of gearing rather than higher interest costs. Rental deduction claims have also increased relative to gross rent.

Chart A1–21: Selected taxation statistics — rental income and deductions

Chart A1–21: Selected taxation statistics — rental income and deductions

Source: ATO, Taxation statistics (various years).

The biases arising from the current income tax treatment of rental properties may amplify volatility in the housing market. (See Section E4.)


Current income tax arrangements for savings lead to significant arbitrage opportunities. The different treatment of capital gains as against other savings income and related expenses is an important driver of these opportunities. This creates significant distortions in how rental properties, in particular, are financed and for the rental property market.

Reform directions

Recommendation 14

Provide a 40 per cent savings income discount to individuals for non-business related:

  1. net interest income;
  2. net residential rental income (including related interest expenses);
  3. capital gains (and losses); and
  4. interest expenses related to listed shares held by individuals as non-business investments.

In conjunction with introducing the discount further consideration should be given to how the boundaries between discounted and non-discounted amounts are best drawn to achieve certainty, reduce compliance costs, and prevent labour and other income being converted into discounted income. Further consideration should also be given to addressing existing tax law boundaries related to the treatment of individuals owning shares in order to address uncertainties about when the shares are held on capital account (and subject to capital gains tax) and on revenue account (and taxed as ordinary income).

Recommendation 15

When the 40 per cent savings income discount is introduced a smooth transition should be provided to minimise any disruption that may arise. The transition to a savings income discount for net residential rental income should only be adopted following reforms to the supply of housing (Section E4 Housing affordability) and reforms to housing assistance (Section F5 Housing assistance).

Recommendation 16

As part of the consideration of alternative company income tax arrangements and dividend imputation (see Recommendation 26 and Recommendation 37), consideration should be given to extending the discount to other savings income.

Towards better taxation of savings

The reform direction for savings income taxation aims to provide a more consistent treatment of savings income, to reduce opportunities for tax arbitrage and to reduce incentives for investors to take on too much debt, while broadly compensating for the effects of inflation, particularly for interest income.

To give effect to this reform direction the Review has considered two primary methods of reducing the taxation of income from savings: discounting savings income (like the current arrangements for capital gains) or taxing savings income at a relatively low flat rate (like the current arrangements for superannuation). Both can be seen as representing a form of dual income tax, as indeed can current arrangements though in a more ad hoc way.

While both alternatives have the potential to represent a good fit for Australia's future tax system, a discount approach is the Review's overall preference as it assists in upholding the current progressivity of the income tax system. Deciding between the two reform paths depends on the trade-off between equity concerns of moving away from progressive marginal tax rates and the potential integrity and simplicity benefits of adopting a flat rate. Both options provide a pragmatic approach to dealing with inflation.

The proportional inclusion achieved under a discount would continue to tax other income from savings at progressive marginal tax rates, which may be desirable from an equity perspective. Proportional taxation of a notional real return to saving may also be efficient. To the extent that savings by high income earners are relatively unresponsive to post-tax returns it may be efficient to tax the returns from savings by higher income earners at higher rates and use the revenue to reduce taxes elsewhere.

However, while a proportional inclusion approach may assist in making the tax system more progressive, the degree to which it would do so is less clear. Progressive rates create opportunities for tax arbitrage, as individuals seek to exploit differences in marginal tax rates, or retain income in companies. For example, under the current tax system individuals can reduce the tax paid on the returns to saving by streaming the income to a family member facing a lower tax rate using a discretionary trust.

A flat tax rate on other forms of savings would also reduce incentives and opportunities for tax arbitrage; for example, from realising income in periods where a person's marginal tax rate is low. Furthermore, a flat tax rate would reduce incentives for investors in high income tax brackets to allocate their savings towards tax favoured assets or to try to evade tax by investing offshore and not reporting income received.

A flat tax rate would also reduce the lock-in effects of a realisation-based capital gains tax relative to progressive taxation. Under progressive taxation taxpayers can be pushed into a higher tax bracket when gains accumulated over a long time are realised. In addition, unlike a proportional inclusion, gains and losses would be taxed at the same low flat rate, reducing disincentives towards risk taking and entrepreneurship. Even so, a proportional inclusion approach would still reduce these drawbacks of progressive taxation, because the differences in marginal tax rates between tax brackets would be smaller.

While a flat tax on the income from savings has many desirable features, the transition to a flat tax rate would raise a number of challenges. A flat tax rate, even at a low rate, is likely to result in an increase in the tax rate faced by some low-income earners. In the long-run, however, flat rate taxation of savings income may be more easily integrated with other potentially desirable directions for the future taxation of capital income, including deeming and accrual taxation of capital gains, and integration of personal income tax with a business level expenditure tax (see Section B1 and Section B2) (Sørensen & Johnson 2010).

A savings income discount would tax savings more consistently

Individuals should be provided with a 40 per cent discount for the returns and expenses from certain forms of taxable savings. This would include interest income from certain interest-bearing assets, including deposit accounts, net rental income from residential properties, including discounting interest expenses, and (as now, but with a reduced discount) capital gains. The discount would also apply to such income earned through trusts and partnerships. The discount would not generally apply to dividends and business income.

It is not recommended that the discount be applied to dividend and certain other business and savings income (such as related party interest and commercial property rentals) while dividend imputation is retained and given the potential for returns to labour to be converted into discounted income. These issues are discussed further below.

A 40 per cent discount represents a more realistic inflation adjustment than the 50 per cent discount currently provided for certain capital gains given the recent history of real risk-free returns and the Reserve Bank of Australia's objective of medium term price stability — with the goal of keeping consumer price inflation between 2 and 3 per cent, on average, over the cycle. Moving to a 40 per cent discount on capital gains would also reduce the arbitrage opportunities currently available while limiting the transitional costs involved with the abolition of the existing capital gains discount.

Certain investment products (such as income bonds, funeral policies, fixed-term annuities and scholarship plans) are currently taxed like bank accounts in some, but not all, ways. Consideration should therefore be given to how these investments are to be treated in light of the general savings income discount.

The savings income discount would reduce the large differences in effective tax rates across different savings vehicles (Chart A1–22). For an individual on the top marginal tax rate, the real effective tax rate on interest income would fall from around 80 per cent to 50 per cent. The treatment of owner-occupied housing and superannuation would remain significantly different, reflecting their lifetime savings characteristics, but the degree of difference would be reduced.

Chart A1–22: Real effective marginal tax rates for selected asset classes

Panel A: Current approach

Chart A1-22: Real effective marginal tax rates for selected asset classes - Panel A: Current approach

Panel B: Recommended approach

Chart A1-22: Real effective marginal tax rates for selected asset classes - Panel B: Recommended approach

Notes: The real effective marginal tax rate on saving is defined as the difference between pre-tax and post-tax return from a marginal investment as a proportion of the pre-tax return (net inflation). A zero effective tax rate reflects a prepaid expenditure tax benchmark, where saving is undertaken out of post-tax labour income and the return to saving is exempt from income tax. Negative rates for superannuation reflect the reduction in tax from either making contributions out of pre-tax income (current approach) or by assessing the recommended refundable tax offset for contributions.

Assumptions: 6 per cent nominal return; 2.5 per cent inflation; for rental property, 50 per cent of the return is attributable to capital gain and 50 per cent to rental income and the rental property is held for 7 years then sold; shares are held for 7 years then sold; superannuation is held for 25 years and the individual is eligible for a tax-free payout at the end of the period. Does not account for interactions with the transfer system.

Source: Treasury estimates.

As previously discussed, the current system for taxing assets that yield capital gains, in particular shares and rental properties, allows for interest to be deductible at the full marginal tax rate, while only half the capital gain is subject to tax. This encourages households to take on too much debt and risk when undertaking these investments.

This bias can encourage surges of debt-funded investor activity in anticipation of concessionally taxed capital gains, potentially adding to the volatility of capital markets. The savings income discount would reduce, but not completely eliminate this bias (see Chart A1-23). Under the savings income discount, income from shares would take discounted (capital gains) and undiscounted forms (dividends). Interest expenses in relation to investments in listed companies should be discounted given the difficulties in assigning debt to particular investments and the significant tax benefits that would otherwise still remain for margin lending.

Chart A1–23: Real effective marginal tax rates on rental property, by gearing ratio

Panel A: Current approach

Chart A1-23: Real effective marginal tax rates on rental property, by gearing ratio - Panel A: Current approach

Panel B: Recommended approach

Chart A1-23: Real effective marginal tax rates on rental property, by gearing ratio - Panel B: Recommended approach

Assumptions: Individual on 46.5 per cent marginal tax rate; 6 per cent nominal return; 2.5 per cent inflation; for rental property, 50 per cent of the return is attributable to capital gain and 50 per cent to rental income and the rental property is held for seven years then sold; tax on debt provider disregarded.

Source: Treasury estimates.

Under the savings income discount, there would also be a generally better outcome for rental property investors that finance out of equity (see Chart A1–23). The more neutral treatment would reduce the crowding out (by those undertaking negative gearing) of other potential investors in rental housing, and improve the long-term stability of the housing market. In the medium to long-run, there would be a shift in how rental property investments are financed. Applying the savings income discount to rental properties would also have the benefit of improving the overall operation and stability of the housing market.

The current system favours returns from capital gains compared to rental returns. Moving to a lower rate of tax on net rental income may also encourage more capital-intensive use of residential land, with increasing investment in higher density, higher rental income yielding developments and less reliance on capital gains from land.

However, there are currently constraints to the supply of housing that need to be taken into account. Amendments to the taxation of rental housing should only be adopted following reforms to the supply of housing, such as the approvals processes around the planning system and land supply (see Section E4). In addition, the tax benefit available to negatively geared properties may place downward pressure on rents though it is poorly targeted to this purpose. As such, steps to reduce the existing tax distortion should only be undertaken following reforms to housing assistance (see Section F5).

Boundary issues need to be considered

Despite achieving more consistent tax outcomes for savings, further consideration would need to be given to a number of boundary issues before implementation. Some existing distinctions in the tax system would become more important.

To prevent the labour income of owner-managers from benefiting from the discount, consideration would need to be given to how best to define eligible interest income. Interest income from deposits with deposit-taking institutions, government and widely marketed bonds should be eligible for the discount. But interest income from transactions involving related parties or associates would need to be excluded or otherwise limited, otherwise returns to labour could be converted into interest payments.

In addition, the interaction of the boundary between eligible and ineligible interest and the boundary between business and non-business income would need some consideration, particularly where the eligibility of interest income or deductions for the discount may depend on the behaviour of the taxpayer. For example, consideration would need to be given to the treatment of interest expenses associated with borrowing to purchase units in a unit trust or company that may carry on a business or may invest in rental properties, debt or listed shares.

The distinction between residential and non-residential properties would become more important. The status of properties on the borderline between residential and commercial property, such as serviced apartments, would need to be clarified. However, although not straightforward, this is an existing challenge in relation to income tax and the GST.

In addition, it would be appropriate to give further consideration to addressing the existing boundaries relating to the tax treatment of income from shares. In particular, whether gains and losses are treated on the capital or revenue account is affected by whether the taxpayer is engaged in passive investment or active trading. Such a distinction can be difficult to apply in practice, because the differences between these are often a matter of degree. Under the savings income discount, there would be a greater incentive for taxpayers to classify their share ownership as a passive investment when they make gains and to classify their ownership as active trading when they make losses so that they can offset (undiscounted) losses against other revenue income.

Transitional issues

The recommended discount would reduce the rate of tax on the returns to existing assets that yield eligible interest and rental income. On the other hand, the reduction in the discount on capital gains would negatively affect individuals with significant unrealised capital gains. For geared investors in rental properties and shares, the application of the discount to net rental income and interest expenses would also have implications for their preferred level of gearing.

Transitional relief should be provided to minimise the disruption that may arise when the savings discount is introduced (see Recommendation 15). Options include a phasing in of the new rate of discount (the best approach) or introducing grandfathering provisions for existing assets. Grandfathering provisions, such as those in capital gains tax, tend to be long lived and are among the most complex provisions in the tax law, and should be avoided where possible.

For example, the discount for savings income and related expenses could be gradually increased from 0 per cent to 40 per cent over five years. A five-year transitional period could also apply for capital gains, with the current 50 per cent discount declining each year by 2 percentage points.

For ungeared capital gains assets, such a phase-in would reduce the likelihood of market disruption caused by the incentive to bring forward the realisation of capital gains between the time of announcement and the enactment of legislation. A phased reduction in the capital gains tax discount over five years would offset the natural decline in the effective tax rate arising from the deferral benefits of a realisation-based capital gains tax (see Chart A1–24).

Chart A1–24: Real effective marginal tax rates on capital gains from a five year transition

Chart A1–24: Real effective marginal tax rates on capital gains from a five year transition

Assumptions: Recommended approach introduced one year after capital gain asset is acquired; individual on 46.5 per cent marginal tax rate; 6 per cent nominal return; 2.5 per cent inflation.

Source: Treasury estimates.

For highly geared investors in rental properties, such transitional arrangements would achieve a smooth transition to an outcome that still provides some tax benefits relative to other investments, though significantly less than before. A smooth transition would limit any short-term disruptions in the supply of rental properties if some investors were to respond to the changed tax arrangements by selling out rather than adjusting their level of gearing.

As well as providing long-term benefits, reforms to address supply side constraints in residential housing markets would also assist with managing any transition and so would become more necessary. Reforms to State land taxes and stamp duties would be of some value in this regard, while there are also potential gains from improving the supply of housing and its responsiveness through other policy reforms, such as to planning and land release (see Section E4).

Treatment of dividends and other business and savings income

The Review has considered whether the savings discount should be extended to dividends, business income, other interest income and rental income from non-residential properties and other assets. Such an extension is not recommended for now, but could be reconsidered in the context of a long-term move away from dividend imputation. However, such consideration would need to account for the deferral benefits afforded by the difference between the company income tax rate and the personal income tax rates.

For larger, more internationally orientated companies, not providing the discount for dividends could partially offset the portfolio bias for domestic savers to hold domestic shares rather than debt and foreign shares. This bias arises to the extent that an imputation credit to resident shareholders is a refund for company income tax that they have not entirely borne given that Australia is an open economy (see Section B2).

Further, if the savings income discount was made available for dividends, there would be an incentive for owner managers to convert their labour income into profits. They could then effectively pay themselves a wage through a dividend of which only 60 per cent would be subject to tax, thereby undermining the tax base for income from work.

The labour to capital conversion problem also arises with non-commercial or non-arm's length loans and with rental income from non-arm's length commercial property. For example, if a discount applied to interest income there would be an incentive to convert business profits that represent the non-wage labour income of owner-managers into interest income from a loan provided by the owner at an artificially high rate of interest.

Nevertheless, excluding dividends and certain other types of income has some downsides. In particular, it would lead to a difference in the tax treatment between debt and equity for domestic savers. This would reduce the cost of debt finance, creating an incentive for domestic companies to finance new investment with debt to the extent that the financing choices of domestic companies reflect the availability of domestic capital. Excluding these items also gives rise to some of the boundary issues identified above.

For the longer term, however, a continuing trend of increased openness in the Australian economy suggests consideration may need to be given to moving away from dividend imputation as a means of integrating personal and company income tax (see Section B2 The treatment of business entities and their owners, Recommendation 37).

Longer-term options for dividends and business income

Under most alternatives to dividend imputation, a typical feature of the taxation of dividends is to provide double taxation relief, either through a discount or a low flat rate. As part of a move to such an alternative, a discount could directly apply to dividends from listed shares where the conversion of labour income into profits is less of an issue.

For unlisted businesses, however, providing relief to all business income would, as discussed above, be problematic. Such income often includes a mix of returns to the labour of the owner-managers as well as the capital employed in the business. On the other hand, if income from unlisted businesses is taxed in full, the savings income component could be over-taxed compared to income from listed companies, discouraging small business and entrepreneurial activity.

To address the difference between saving through widely-held listed shares and through a closely-held business, the savings income discount could be extended to business income through a business allowance. Internationally, business allowance systems are already used where there are dual income taxes, such as in the Scandinavian countries, to separate capital income from other income.

Business allowance systems split the net business income of sole traders, partnerships and trusts into labour and savings or capital components, with the discount applying to the capital component. A similar approach could apply to dividends received from unlisted companies, non-commercial loans and non-arm's length commercial property arrangements. Under the allowance system, owners of unlisted businesses (shareholders in unlisted companies, trust beneficiaries, partners or a sole proprietor) would receive an allowance for a deemed return on their equity (savings) in the business.

Extending the discount, or applying a flat tax rate, to all savings income would mean that many of the boundary issues previously discussed, and the differential treatment of debt and equity, would be of less concern. An allowance arrangement would however give rise to some complexity of its own, though allowance-like arrangements may be an appropriate way of dealing with non-arm's length interest payments.

As part of the longer-term consideration of alternative company tax arrangements and dividend imputation, consideration should therefore be given to extending the discount (or possibly a flat rate of tax on savings income) to all savings income.

Treatment of earnings from life insurance policies

Currently, life insurance providers are taxed at the company income tax rate on investment earnings from assets that support ordinary life insurance policies. Reversionary bonuses (or accumulated earnings) paid to policyholders when an insured event occurs, or when the policy is cancelled or matures after it has been held for more than 10 years, are tax free. Therefore, the policyholder is effectively taxed at the company tax rate on the earnings.

Where the policy is cancelled or matures after it has been held for eight years or less, the accumulated earnings paid to policyholders are taxed at marginal tax rates. A proportion of the accumulated earnings is taxed if the policy is cancelled or matures after it has been held for nine or 10 years. To the extent that the accumulated earnings are taxed, policyholders are entitled to a tax offset to prevent double taxation. Currently, the tax offset is 30 per cent of the taxable component of the earnings, a proxy for the company tax rate.

Accumulated earnings paid to policyholders should not benefit from the savings income discount. Life insurance policyholders would benefit from the recommended reduction in the company income tax rate to 25 per cent (see Recommendation 27), increasing the potential tax deferral advantages of life insurance. In addition, life insurance providers invest in assets that produce income, particularly dividends, that would not attract the discount.

Simplifying the taxation of capital gains

The regime for taxing business and savings income includes complex provisions that reflect the complexity of commercial activity, the increasing sophistication of financial instruments, and the wide variety of saving structures and intermediaries. There will be a continuing need to re-assess such provisions with a view to improving certainty, reducing administration and compliance costs, and dealing with design or integrity failings.

Particular emphasis should be placed on simplifying the rules directly affecting large numbers of individuals who are not equipped to deal with tax complexity. The capital gains tax regime is the primary example of such complexity, and should be simplified to reduce administration and compliance costs for individuals and small business in particular.

Capital gains tax is complex

A number of submissions to the Review have highlighted the complexity of the current capital gains tax regime. Principal drivers of the high administration and compliance costs include the complexity of the legislation, the frequency of changes to the legislation, the number of rules and exceptions, and record keeping requirements.

For individuals, shares and real estate give rise to the majority of taxable capital gains (see Table A1–7). Collectables and personal use assets generate little capital gains tax revenue.

Table A1–7: Total current year capital gains of taxable individuals by source (2006–07)(a)(b)

Source of gain(c) Number of individuals
reporting gains
Value of capital gain ($million)
Shares 436,395 20,415
Real estate 152,056 15,061
Other assets(d) 214,918 11,196
Collectables 1,120 67
Total number of individuals reporting gains 668,415(e) 46,739
  1. Refers to individual taxpayers with net tax payable greater than $0 who completed a schedule.
  2. Includes data processed up to 31 October 2008.
  3. Sources include both active and non-active assets.
  4. Includes other capital gains tax assets and any other capital gains tax events.
  5. This is not the sum of figures in this column as individuals may report capital gains from more than one source, so that the total of individuals reporting gains from different sources will exceed the total number of individuals reporting gains.

Note: The figures in this table are derived from the capital gains tax schedule, which individuals who lodge a paper return are not required to complete. Therefore these figures cannot be directly compared to the statistics reported on net capital gains in Taxation statistics.

Source: ATO (2009).

The complexity of capital gains tax is compounded by the various exemptions and the grandfathering of previous provisions. For example, there are various concessions for small business, in addition to the general 50 per cent discount for individuals, while capital gains made on assets acquired before the introduction of the capital gains tax regime are generally exempt and the pre–1999 indexation arrangements remain available for assets acquired before indexation was abolished. A number of submissions also noted that the mechanical and prescriptive nature of the capital gains tax legislation adds significantly to administration and compliance costs.

Small business capital gains tax concessions

There are currently four separate small business capital gains tax concessions available to qualifying businesses or their owners: an exemption for capital gains made on active assets held for at least 15 years (generally available only to an individual aged 55 or over who retires); a retirement exemption for capital gains made on active assets up to a lifetime limit of $500,000 per individual; a further 50 per cent discount for the sale of active business assets; and a small business roll-over, which allows deferral of a capital gain made on an active asset if within two years the proceeds are reinvested in another business asset.

There are currently two initial criteria a taxpayer must satisfy to be eligible for the small business CGT concessions: they must either be conducting a business with an aggregated turnover of less than $2 million (the small business entity test) or they must have net assets of $6 million or less (the maximum net asset value test).

The concessions are a significant area of complexity within the capital gains tax rules. In a survey of tax practitioners on the drivers of capital gains tax compliance costs, Evans (2004) found that the small business concessions ranked prominently (6 out of 18) in the list of factors. Despite attempts to simplify the concessions, taxpayers are required to navigate a legislative maze of gateway and threshold conditions and then additional conditions that relate to each of the specific concessions.

Evans also found that the concessions have become more complex over time. They have frequently been amended to extend their reach and to ensure that the concessions do not provide opportunities for tax avoidance. The outcome is provisions so complex that specialist professional advice is typically required to access them. Despite this complexity, but perhaps reflecting their value, the concessions are widely used (see Table A1–8).

Table A1–8: Number of claimants of the small business capital gains tax concessions, 2006–07

Concession Companies Individuals
15 year exemption 207 764
Retirement exemption 1,264 10,057
Active asset reduction 2,746 24,220
Rollover 519 4,676

Source: ATO (2009).


The current capital gains tax rules are particularly complex, with that complexity compounded by various exemptions and the grandfathering of previous provisions.

Simplifying capital gains tax

Recommendation 17

The capital gains tax regime should be simplified by:

  1. increasing the exemption threshold for collectables and exempting all personal use assets;
  2. rationalising and streamlining the current small business capital gains tax concessions by:
    • removing the active asset 50 per cent reduction and 15–year exemption concessions;
    • increasing the lifetime limit of the retirement exemption by permanently aligning it with the capital gains tax cap for contributions to a superannuation fund; and
    • allowing taxpayers who sell a share in a company or an interest in a trust to access the concessions via the turnover test.
  3. removing current grandfathering provisions relating to assets acquired before the commencement of capital gains tax, with a market value cost base provided for those assets when the exemption is removed, or before the end of previous indexation arrangements. A relatively long lead-time should be provided before these removals take effect; and
  4. rewriting the capital gains tax legislation using a principles-based approach that better integrates it with the rest of the income tax system.

The capital gains tax system should be simplified by rationalising existing concessions, exempting certain assets, simplifying the legislative provisions, and removing some grandfathering arrangements. While the Review has considered the potential benefits of an annual exemption, it is not clear that there would be net benefits from such an approach.

For any simplification of capital gains tax to substantially reduce overall complexity and compliance costs, trade-offs that favour simplicity over equity and efficiency would be required, as well as an acceptance that there would be losers as well as winners in respect of future tax liabilities. As a possible exception, in the medium to long term the greater use of real time reporting of taxpayer information from share registries could also significantly reduce the need for shareholders to retain records and calculate their own capital gains.

Rationalising and streamlining small business concessions

The small business capital gains tax concessions are a significant contributor to complexity and compliance costs. As discussed above in relation to income from work, taken together with the general 50 per cent capital gains tax discount the concessions also result in a highly favourable tax outcome for those small businesses benefiting from them. This outcome may skew the allocation of resources in the economy to less productive uses and detract from equity goals.

Rationalising the provisions — by removing the active asset 50 per cent reduction and the 15 year exemption — and streamlining the remaining concessions would reduce compliance costs as well as improve efficiency and equity around the treatment of earned income (see Recommendation 17b). The small business roll-over provision would be retained, as it has an efficiency benefit of reducing lock-in effects that prevent assets and businesses being reallocated and organised most productively.

Two of the existing concessions — the 15 year exemption and the retirement exemption — raise issues related to the self-employed and superannuation arrangements, and the principle that lifetime savings should face little or no income tax. Many self-employed people effectively use the accumulation of value in their business as a lifetime savings vehicle for their retirement. But the system should be simplified by providing the retirement exemption only.

Access to the retirement exemption should be increased and better aligned to concessions available within the superannuation system. The current lifetime contribution limit should be increased and permanently aligned to the cap for contributions to a superannuation fund derived from the disposal of small business assets. This would increase the current lifetime limit from $500,000 to $1.1 million (in 2009–10), and ensure the limit is indexed annually.

The small business capital gains tax concessions could also be rationalised and simplified by allowing taxpayers who sell a share in a company or an interest in a trust that is a small business entity to access the concessions using the turnover test. Under the current arrangements the concessions can only be accessed under the maximum net asset value test. Under the recommendation, owners of businesses who already access the other small business concessions will not need to determine eligibility under the maximum net asset value test, and instead rely on the same test used to access the other small business concessions.

Exempting certain assets can reduce compliance costs

The capital gains tax regime could also be simplified by exempting all personal use assets and increasing the exemption threshold for collectables. Currently collectables with a cost base of $500 or less and personal use assets that cost $10,000 or less are disregarded for capital gains tax purposes.

All personal use assets should be exempt from capital gains tax, reducing compliance costs at minimal cost to revenue. Increasing the threshold for collectables would also reduce compliance costs, without establishing a tax bias to invest in high-value collectables (such as works of art).

Rewrite the capital gains tax rules

The complexity and uncertainty of the capital gains tax regime could also be reduced by adopting a principles-based approach to simplifying the legislative provisions. The current legislative provisions are prescriptive and mechanical. For example, under the current rules there are 53 separate capital gains tax events, five elements that make up an asset's cost base and five elements of an asset's reduced cost base, and an anti-overlap rule preventing double taxation through capital gains tax.

A principles-based approach could be used to reduce complexity while also increasing comprehension and awareness of the regime. Subject to consideration of the feasibility and net benefits of alternative approaches, a principles-based approach could build on the existing core capital gains tax concepts (such as events, cost base and capital proceeds) to minimise any impacts or disruption to other parts of the tax law that also use these concepts. Any redrafting should focus on the relationship of the capital gains tax regime to the rest of the income tax system.

In 2000, the Board of Taxation commissioned draft legislation on a more principled expression of the capital gains tax law (as part of a wider project known as the 'tax value method'). That redraft reduced 126 pages of capital gains tax law to only 28, without significant policy change. Though the government of the time decided against proceeding with the more ambitious project of which it was a part (and which the Review is not proposing be relaunched), the redraft highlighted the potential for significant legislative simplification.

More substantial simplification of the legislation would inevitably involve some policy change, with some taxpayers made worse off and others better off. Furthermore, while a principles-based approach has the potential to simplify the law, much of the complexity in the capital gains tax regime is due to concessions and the need to address integrity concerns. Whether such concessions or integrity provisions are worth retaining would need to be re-assessed.

Remove grandfathering provisions

The abolition of grandfathering for pre-capital gains tax assets and for pre–1999 indexation arrangements would reduce the complexity of the capital gains tax regime. Evans (1998) suggested that up to 20 per cent of the capital gains tax legislation is attributable to the decision to grandfather old provisions. Removing the grandfathering provisions would also improve the efficiency and equity of the system. For example, grandfathering increases the lock–in effect of a realisation-based capital gains tax, which can lead to inefficient resource allocation.

'Grandfathering' (that is, preserving the treatment of pre-existing arrangements when rules are changed) often occurs in response to concerns about the equity and efficiency implications of a change in policy settings. However, it can add to the complexity of the tax system, particularly where its effects are long lived.

Capital gains tax only applies to gains made on assets acquired after 19 September 1985. While grandfathering reduced the impact of change for existing investors, more complex legislation is needed to maintain the exemption and prevent avoidance. For those holding grandfathered assets there would be compliance costs associated with ending grandfathering, but as the number of such assets declines over time the case for ending grandfathering becomes more compelling.

The indexation rules also contribute to the regime's complexity, although to a lesser extent than grandfathering. Indexation was phased out from September 1999 but remains available for assets acquired before then. It is likely that only a small number of taxpayers are currently better off under indexation relative to the outcome they would receive under the discount method with respect to capital gains unless they can offset such gains with relatively large capital losses.

Concerns over removing the pre-capital gains tax exemption could be offset to some extent by providing a market value cost base for remaining pre-capital gains tax assets at the time the exemption is removed. This would ensure that only capital gains that accrue going forward are taxable. Providing a relatively long lead time for such a change would also provide an opportunity for taxpayers holding pre-capital gains tax assets to dispose of those assets without capital gains tax consequences.

Taxing savings on an individual basis

One practical difficulty with taking the individual as the unit of assessment for tax purposes is the alienation of income from saving, where an individual can attribute their income to another person or legal entity. A particular issue is the difficulty in drawing a distinction between gifts to others and the assignment of income from assets to others.

A person can transfer ownership of an asset to another person. This can be done formally, or can happen naturally such as when a couple has a joint bank account or owns assets jointly. Income tax systems typically permit these gifts, with the future income from the gifted asset included in the taxable income of the other person. Such gifts, however, raise a question of whether gifts or other wealth transfers should be taxed (see Section A3).

Alternatively, a person can retain ownership of an asset but assign (pre-tax) income from the asset to another person or entity, for example, by legally assigning the right to any interest or dividend to another person, or settling the asset on a trust that then distributes the income to the other person. Income tax systems may not recognise these assignments, particularly where it is only the current income that is applied for the benefit of another.

However, there is no clear line between the gift of an asset and an assignment of the income from an asset. The value of an asset can be seen as the net present value of the expected, risk adjusted, future income stream from that asset. The assignment of part of the future income of an asset is simply the giving of another type of asset, that of the right to income for a defined period.

The relative ease with which savings income can be split between individuals, particularly within a household, may have implications for attempts to improve workforce participation by keeping the marginal personal income tax rates facing secondary earners low (see Section A1–1). Including the household's savings income in the secondary earner's income may increase their marginal tax rates, reducing incentives to work or to work more.

Attempts to limit assignments of savings or business income, from either a person's labour or savings can be constrained by practical considerations such as the difficulty of properly targeting any measures. For example, in the case of a family trust used in connection with a family business, the underlying ownership claims to the assets of the business and the contributions of unpaid labour by the different family members may be diffuse and complex.

Current rules only partly limit the alienation of savings income

For savings income, specific rules apply to limit the transfer of income from property, but transfers of property itself are generally effective in assigning future income to the recipient of the property (though on transfer capital gains tax may apply to any gains in the value of the asset that have arisen up to that point in time). Entities such as companies and trusts can also be used to split income from assets, while the underlying ownership or control of the assets remains unchanged.

A number of submissions to the Review raised concerns about the use of discretionary trusts, in particular, to split income. Trusts have the advantages of preserving tax preferences such as capital gains tax discounts and foreign tax credits. Companies allow deferral of any taxation above the company income tax rate, and the potential to smooth an individual's taxable income over time. Trusts and companies are often used together to obtain the particular tax benefits of each.

There are also instances where a (low-tax) beneficiary of a trust is taxed on trust income (for example, as they are considered to be presently entitled to the income) but the actual income is effectively provided to another. The different components of income associated with the same asset may also be allocated for tax purposes to the beneficiary best suited from a tax perspective to receive them. Those beneficiaries to whom the different types of income are allocated may change over time.


Current rules limit, but do not eliminate, the scope for the alienation or assignment of an individual's income to other persons or legal entities.

Options to further limit the alienation of savings income

The Review has considered options to further limit the potential to alienate savings income, particularly through the use of trusts. However, given the potential downsides of these options for the overall progressivity of the system or for other taxpayers, their adoption has not been recommended. The case for change is also weakened by the difficulty in drawing a line between allowing gifts and common ownership of assets within households and preventing income splitting.

Applying a flat rate of tax to savings income, from the first dollar, would remove all income tax advantages from income splitting. However, a discount for net interest, rental income and capital gains would also flatten the income tax rates to some extent as they apply to nominal savings income. This would reduce the benefits from income splitting or deferring tax through sheltering income in a company.

For trusts, attributing income to the settlor of assets on the trust when they retain control would directly target the alienation of income. Foreign trusts are already subject to such rules for income tax purposes (the transferor trust provisions), and the transfer system adopts a similar approach to private trusts and companies. However, the administration and compliance costs associated with the general adoption of this approach in the income tax system is likely to be significant and enforcement would be difficult.

Another option considered would be to tax trusts, or discretionary trusts, as companies. Taxing trusts as companies does not directly address the problem of the alienation of income. Use of a trust would potentially confer tax deferral advantages (where the company income tax rate is less than the marginal tax rate of shareholders), while income could still be split between the various beneficiaries of the trust who could in turn benefit from refunds of any excess imputation credits.

Taxing trusts as companies could, however, indirectly limit income splitting by imposing tax penalties on the use of trusts, such as the non-flow through of capital gains discounts. But taxing trusts as companies would be poorly targeted, disadvantaging trusts not used for income splitting. Previous consideration of this option following the Review of Business Taxation also pointed to the practical difficulties involved (Board of Taxation 2002).

While the current income tax structure is broadly retained, the use of trusts for tax avoidance or evasion is, however, likely to remain an area of concern that may require targeted responses. The Australian Government has recently announced the introduction of tax file number reporting and associated withholding requirements for closely-held and family trusts. An updating and rewriting of the current trust income tax rules (Division 6) also has the potential to consider any abuses of current trust tax arrangements (Section B2 The treatment of business entities and their owners, Recommendation 36).