This chapter describes the approaches used to measure tax expenditures and provides guidance for interpreting the estimates reported in the TES.
There are three main methods used to measure tax expenditures.
- The revenue forgone approach — this approach measures the difference in tax paid by taxpayers who receive a particular concession relative to similar taxpayers who do not receive that concession. It compares the current/prospective treatment to the benchmark treatment, assuming taxpayer behaviour is unchanged.
- The revenue gain approach — this approach measures how much revenue could increase if a particular tax concession was removed. Accurate estimation of this cost would require estimates of the secondary behavioural effects associated with such a change.
- The outlay equivalence approach — this approach estimates how much direct expenditure would be needed to provide a benefit equivalent to the tax expenditure. This approach measures the direct expenditure required, in before-tax dollars, to achieve the same after-tax dollar benefit as the tax expenditure where the direct expenditure receives the tax treatment appropriate for that type of income in the hands of the recipient.
The three methods can yield significantly different estimates of the value of a tax expenditure.
Consistent with most tax expenditures statements published in OECD countries, Australia uses the revenue forgone approach.1 This is the most reliable method for estimating the level of assistance the tax system provides to taxpayers.
Some caution should be exercised when using the estimates in this statement for wider purposes, such as estimating the budgetary impact of tax concessions. This section provides an overview of the main issues relating to interpretation of the estimates.
Revenue forgone and revenue gain
Estimates calculated by the revenue forgone approach identify the financial benefit of tax concessions to taxpayers receiving those concessions relative to taxpayers that do not. It does not necessarily follow that there would be an equivalent increase to government revenue from the abolition of a tax expenditure. This is because of behavioural responses by the recipients of tax expenditures and overlaps in the coverage of different tax expenditures.
The introduction of a tax expenditure tends to increase concessionally taxed activity. Accordingly, the same activity would be expected to contract should the related tax expenditure be abolished, with consequential implications for potential revenue flows. Other responses may follow. For example:
- the removal of one concession may result in increased use of other concessionally taxed activities, lowering tax revenue elsewhere;
- under a progressive income tax system, the removal of a tax expenditure may result in some taxpayers facing a higher average tax rate, increasing tax revenue; and
- as tax concessions may alter resource allocation and direct scarce resources from one activity to another, removal of those concessions may affect economic efficiency and the overall level of economic activity. This change in activity could affect tax revenues.
In most cases, the net effect of these influences on revenue is unclear. Furthermore, in cases where the level of activity is highly sensitive to a concession, the increase in revenue from removing the tax expenditure could be very small. In these cases, reporting tax expenditure estimates as the cost to revenue (that is, using the revenue gain approach) would give the impression that the tax expenditure has little material effect when actually the recipients derive quite large financial benefits.
Comparison with budget estimates
Tax expenditure estimates may differ from budget estimates because tax expenditures are estimated relative to designated benchmarks. For example, CGT rollover relief is provided for superannuation entities that merge to meet the requirements of the superannuation safety arrangements (E18). This rollover relief gives rise to a tax expenditure when assessed against the benchmark because it provides a tax deferral benefit. A tax expenditure arises because CGT is not collected at the time a ‘CGT event’ occurs (that is, the disposal of assets by one entity and acquisition of those assets by the merged entity). By contrast, this measure was estimated to have no impact on budget revenue because the transactions concerned would not have occurred in the absence of the superannuation safety arrangements and therefore providing rollover relief did not reduce revenue relative to the revenue included in the budget forward estimates (the benchmark for costing budget measures).
Tax expenditure aggregates
Unless otherwise indicated, tax expenditure estimates are calculated on an individual basis and do not take account of potential overlaps with other tax expenditures. While aggregate tax expenditures can provide a guide to trends in tax expenditures over time, overlaps between the coverage of different tax expenditures and likely behavioural responses to their removal mean that such aggregates are not a reliable indicator of the overall budgetary impact of tax concessions.
Estimates and projections
Tax expenditure estimates are separated into estimates (for historical years) and projections (for future years). The estimates for 2006-07 are preliminary and subject to revision upon receipt of further tax data.
The tax expenditure estimates are prepared on the same revenue recognition basis as the budget estimates. From the 2006-07 Budget, the basis for reporting revenue in the budget changed. The changes are outlined in the box below and apply to estimates in TES from 2006-07. All estimates relating to periods prior to 2006-07 are reported in TES on the Tax Liability Method basis.
Revised revenue recognition methodology
Accrual accounting was introduced by the Australian Government in the 1999-2000 Budget. The Australian Accounting Standards and Government Finance Statistics standards for accrual accounting require that taxation revenue be recognised in the reporting period in which the taxpayer earns the income that is subsequently subject to taxation — this is known as the Economic Transactions Method (ETM). But the standards also permit government reporting using an alternative approach when the ETM approach would generate unreliable measures of taxation revenues.
Because ETM is an unreliable measure for several significant revenue heads — and these account for the majority of total revenue — all taxation revenue was recognised in all accrual budget-related documentation from the 1999-2000 Budget to the Mid-Year Economic and Fiscal Outlook 2005-06 using the Tax Liability Method (TLM). Under TLM, taxation revenue is accounted for at the time a taxpayer makes a self-assessment or when an assessment of a taxation liability is raised by the relevant authority.
Commencing with the 2006-07 Budget, the Australian Government adopted ETM revenue recognition for all revenue heads where the measurement issues are not material, but retained TLM revenue recognition where ETM measurement issues may be material. The taxation revenues that continue to be recognised on a TLM basis are:
Treatment of imputation
The value of some concessions reported in this statement is partially offset as a result of the imputation system. For example, concessions that reduce company tax may be clawed-back through the subsequent taxation of dividends in the hands of shareholders. The estimates in this statement generally make no allowance for this clawback owing to the practical difficulties of doing so.
Capital gains tax estimates
Under the CGT benchmark, nominal capital gains are fully taxable upon realisation (for further details see Chapter 4). The most significant tax expenditure against this benchmark is the 50 per cent discount for capital gains realised by individuals and trusts (E9) which affects most capital gains realised by these entities.
Individuals and trusts may also be eligible for other CGT concessions. The revenue forgone methodology that is generally used in this statement implies that estimates for these other CGT concessions should be calculated against the benchmark of full taxation of nominal capital gains.
To avoid double counting, the values of tax expenditures for other CGT concessions are reduced by the CGT discount component and the discount component of these other concessions is included in the tax expenditure for the CGT discount (E9). This modification to the tax expenditure methodology provides more realistic estimates of the value of the benefits taxpayers receive from capital gains concessions in aggregate, though it has the effect of understating the value of individual CGT tax expenditures other than the discount.
The estimates of the tax expenditures in the forward projections are not necessarily indicative of the cost of the superannuation concessions over the long term. In this context, the current tax concessions will help to reduce budgetary expenses in future years, particularly age pension payments, through encouraging private provision for retirement.
Further, the estimates cannot be interpreted as a time series of the ongoing revenue savings that could be obtained if the superannuation concessions were eliminated. This is because the increase in tax revenue arising from the elimination of the tax expenditure with respect to a particular year would cause the superannuation tax base to be smaller for the next year. For example, if contributions and fund earnings in 2004-05 had been taxed according to the superannuation benchmark, superannuation fund assets and fund earnings in 2005-06 would be lower than if the concessional tax treatment had applied in the previous year.
In addition, changes to the taxation of superannuation could be expected to have behavioural impacts, to the extent that people may alter their saving behaviour as a result. The estimated cost of the superannuation tax expenditures assumes no behavioural change involving either the portfolio composition of savings or the saving rate more generally.
This edition of the Tax Expenditures Statement reports the components of the concessional taxation of funded superannuation (C1 in the 2006 TES) as individual tax expenditures for the first time. Previously, these components were reported in a separate appendix. The superannuation estimates reported in this statement are comparable with those published in the 2006 Tax Expenditures Statement, the sum of estimates C1 and C2 in the 2006 Tax Expenditures Statement should be compared to the sum of C4-C14 and C19. This can also be compared to the comprehensive estimate of the superannuation tax expenditure in the 2002 Tax Expenditures Statement and earlier.
Each year there are also variations arising from the revision of earnings and contributions estimates. In particular, taxable earnings of superannuation funds are not readily predictable. A major reason is that it lies within the discretion of a fund manager to decide when any accrued capital gains of a fund are realised. In addition, the earnings series is intrinsically volatile, reflecting fluctuations in interest rates, dividends and asset prices. Fund earnings have been ‘smoothed out’ for the forward projections.
1 The approaches adopted by various OECD countries to measure tax expenditures are reported in Tax Expenditures — Shedding Light on Government Spending through the Tax System, Lessons from Developed and Transition Economies, The World Bank, Washington DC (2003).