Executive Summary
Introduction

Tax Expenditures in Australia

'Tax Expenditure' or 'Direct Spending': Why Does it Matter?

Conclusions
Appendix 1. Significant Tax Expenditures Introduced from 1996
Endnotes

List of Figures

Figure 1. Tax Expenditures as a Per Cent of GDP, 19832001
Figure 2. Tax Expenditures Compared with Total (Direct and Tax) Expenditures, 19832001
Figure 3. 'Social' Share of Major Tax Expenditures, 199192 to 200405

List of Tables

Table 1. Tax Expenditures Share by Function, 200001
Table 2. Tax Expenditures as Share of Budgetary Assistance by Function, 200001
Table 3. Major Tax Expenditure Items, Actual and Projected, Selected years: 199192 to 200405
Table 4. Number of Tax Expenditures, 199697 to 200405
Table 5. Major Social Tax Expenditure Items, Actual and Projected, 19922005, $m
Table 6. Major Business Tax Expenditure Items, Actual and Projected, 19922005, $m

Acronyms

ACOSS Australian Council of Social Service
ANAO Australian National Audit Office
ANTS A New Tax System
ASFA Association of Superannuation Funds of Australia
ATO Australian Taxation Offices
CDEP Community Development Employment Project
CGT Capital Gains Tax
DWT Dividend Withholding Tax
GDP Gross Domestic Product
GST Goods and Services tax
NCOA National Commission of Audit
OECD Organisation for Economic Cooperation and Development
TES Tax Expenditures Statement

Executive Summary

A tax expenditure is a provision of Australian tax law that provides preferential treatment to certain classes of taxpayers or to particular types of activity. Examples are superannuation tax concessions and excise exemption for alternative fuels. Tax expenditures take various forms including deductions, rebates, reduced rates and deferred liabilities. In 200102, the cost of Commonwealth tax expenditures was estimated at nearly $30 billion, the equivalent of 20 per cent of outlays or four per cent of Gross Domestic Product (GDP).

This paper examines trends in tax expenditures in Australia, and the implications for the efficiency of public spending and for parliamentary and public scrutiny.

There has been an underlying upward trend in the real cost and number of tax expenditures since the mid 1990s. Social tax expenditures have been growing faster than tax expenditures affecting business. Nevertheless, tax expenditures remain a greater proportion of budgetary assistance for industry than is the case for the social security, health and welfare functions of the Commonwealth.

The significance from a public policy perspective is that even though tax expenditures are effectively equivalent to direct government spending, and in some cases are alternatives, they largely escape the detailed parliamentary scrutiny and accountability processes associated with budgetary outlays. This is notwithstanding the 1996 National Commission of Audit recommendations that tax expenditures be subjected to the same degree of scrutiny as direct budget spending.

Indeed the only real progress that has been made in achieving greater scrutiny of tax expenditures dates back to the mid 1980s when the government of the day commenced the publication of Treasury's annual Tax Expenditures Statement,⁽¹⁾ a document the Treasurer releases typically in January. Unfortunately, these statements do not encompass some important Commonwealth tax expenditures, notably those relating to the goods and services tax (GST) concessions (for example, the zero-rating of GST on many food items).

Nor, to date, has there has been any public evidence that the Commonwealth Government has met undertakings to integrate tax expenditure analysis into its fiscal management practices.

Ways to improve the transparency and parliamentary scrutiny of tax expenditures could include:

The paper suggests that reform measures along these lines might be considered and developed further in the context of a parliamentary inquiry into tax expenditures policies.

Introduction

What are Tax Expenditures?

A tax expenditure is a departure from the generally accepted ('benchmark') tax structure that produces a favourable treatment of particular types of activities or taxpayers.⁽²⁾ In the United States the Office of Management and Budget defines tax expenditures as revenue losses attributable to provisions of the federal tax laws 'which allow a special exclusion, exemption, or deduction from gross income or which provide a special credit, a preferential rate of tax, or a deferral of liability'.⁽³⁾

In Australia, the Commonwealth Treasury considers a tax expenditure as 'A tax concession that is designed to provide a benefit to a specified activity or class of taxpayer'.⁽⁴⁾

Emphasising the characteristic that tax preferences are similar to direct expenditures in that they are different ways of achieving the same kinds of social and economic policy goals, some commentators identify a provision as a tax expenditure only if its objective could reasonably be achieved by an expenditure program.⁽⁵⁾ In other words, 'a tax expenditure is a spending program and must therefore be analysed in spending terms', using analytical techniques similar to those used to evaluate direct expenditure programs⁽⁶⁾.

Tax expenditures can take various forms and may either reduce tax liabilities or simply defer them. Tax exemptions, deductions or allowances, rebates or 'offsets', rate reliefs and provisions for income averaging or income or tax deferral are different ways of delivering a tax expenditure.⁽⁷⁾ The schedule below identifies some major forms of tax expenditures with some illustrative examples.

Schedule 1. Major forms of tax expenditure

 Measuring Tax Expenditures

A number of conceptual and measurement issues arise from the need to identify and measure tax expenditures.⁽⁸⁾ An important area of debate has been over the norm or 'benchmark' income tax system. The benchmark is the basic structure of the income tax system against which tax expenditures are measured.⁽⁹⁾ The cost of a tax expenditure depends on how the tax benchmark is definedfor example, if the family unit is considered to be the benchmark income tax unit, family payments will not be counted as tax expenditures. The benchmark may differ between countries and over time.

Differences may include:

An Organisation for Economic Cooperation and Development (OECD) survey of tax expenditures and benchmark tax structures found that the norm includes the rate structure, accounting conventions, the deductibility of compulsory payments, provisions to facilitate administration and those relating to international fiscal obligations.⁽¹¹⁾ Discussing such conceptual issues, the Treasury's view is that:

The decision as to an appropriate benchmark for determining tax expenditures is a matter for judgment: benchmarks may vary across countries and within countries over time. The principal criterion of benchmark design is that it should represent the neutral taxation treatment of similarly placed activities or classes of taxpayer (That is, neutral taxation treatment neither favours nor disadvantages similarly placed activities or classes of taxpayer).⁽¹²⁾

The usefulness of the tax expenditures concept as a tool in budgetary analysis has been contested by advocates of targeted tax benefits for favoured industries or constituents, who challenge the labelling of tax concessions as 'spending' because it reduces their attractiveness. Opponents of tax expenditure reporting in the United States have also criticised the effects of arbitrary measurement of some tax expenditure items on democratic decision-making, because of federal budget rules requiring that new direct spending and revenue legislation be deficit neutral.⁽¹³⁾

The undoubted conceptual and practical difficulties of estimating tax expenditures have not prevented their widespread adoption. As United States economic adviser and academic Professor Martin Feldstein observed two decades ago 'The fact that experts disagree about which provisions should be considered tax expenditures does not reduce the usefulness of any estimates of particular tax subsidies'.⁽¹⁴⁾

Surveys of legislators in the United States show a wide acceptance of the tax expenditures concept and it is now well recognised that tax benefits have an economic impact similar to a budgeted appropriation.⁽¹⁵⁾

Although tax expenditure analysis overseas initially focused on the income tax system, tax expenditures estimates have also been compiled for other taxes, including sales taxes and excise taxes, as well payroll taxes and land taxes such as those levied by the Australian states. That said, tax expenditures reporting so far has been most utilised within a revenue policy framework, for understanding the nature of potential tax policy problems and the impact of existing tax policies, rather than as a tool of budgetary decision-making and determining the allocation of public resources.⁽¹⁶⁾ This also appears to be the case for tax expenditure reporting in Australia.

Why are Tax Expenditures Important?

The most important reason for examining tax expenditures is their size. Treasury has identified 218 tax expenditures for the year 200102, with an estimated cost to revenue of $29.2 billion a year. While there are significant conceptual problems in aggregating individual tax expenditure items,⁽¹⁷⁾ the magnitude of tax expenditures can be gauged by the fact that the amount is equivalent to twice the size of the Defence budget and roughly equals Commonwealth spending on health. Tax expenditures account for one in five dollars of Commonwealth spending. At the state government level, tax expenditures exceed $9.5 billion annually.

There are several consequences of tax expenditures that are of interest to public policy.

One is that overuse of the tax system for non-revenue purposes may undermine tax system integrity. Tax expenditures also have important potential implications for budget control and fiscal management because they are often openended and their financial cost difficult to forecast.

Other consequences of tax expenditures arise from their allocative effects. Assistance provided through tax concessions is closely akin to a direct budget subsidy. United States academic Professor Paul McDaniel, one of the architects of tax expenditures reporting in the US, highlighted what he saw as the 'secret double life' of the tax system:

The basic function of a tax system is to collect those revenues which are to be allocated to the public sector and expended for publicly determined purposes. A tax system may also be used by government as a tool to implement policies of government that require the expenditure of funds. Thus when government desires to provide a financial incentive for individuals or businesses to engage in a particular course of action or to share costs in hardship situations, it may employ either direct spending programs or special provisions in the tax system.⁽¹⁸⁾

Tax expenditures may work against other policy objectives or their results may be quite different from those of related direct spending programs.⁽¹⁹⁾ For example, tax expenditures targeting business or industry policy objectives may reduce the effectiveness of direct spending programs promoting environmental goals.

Critics of tax expenditures also emphasise that the distributive consequences of tax reliefs and tax incentives are often very regressive compared to direct spending programs.⁽²⁰⁾ It has been argued that the design of tax expenditures entrenches the economic disadvantage and dependent status of women because the main beneficiaries of major tax concessions such as for superannuation and capital gainshigh-income earners with continuous labour force participation, and those with significant capital gainsare predominantly male.⁽²¹⁾ Much social security and health spending in Australia has been provided through direct outlays, often on a meanstested basis. Recent new tax expenditures such as the baby bonus, the 30 per cent private health insurance rebate, and higher tax thresholds for retirees have been criticised as a 'new welfare state', characterised by 'a steady growth in handouts to people who are far from poor'.⁽²²⁾

While such criticisms may stem from inappropriate design as much as intrinsic problems in using the tax system to deliver benefits, such characteristics may be more common in tax expenditure programs because they are subjected to a lesser degree of scrutiny and evaluation than direct outlays. Unlike direct subsidies, tax expenditures traditionally have been subject to little parliamentary or public scrutiny. As the OECD pointed out: 'the concept of a tax expenditure was developed because accounting for the costs and benefits of tax measures is often less rigorous than for direct expenditures'.⁽²³⁾ Treasury notes that 'unless both direct and tax expenditures are considered, the apparent size of government could be reduced simply by pursuing the objectives of expenditure programs through tax expenditures'.⁽²⁴⁾

For this reason tax expenditures have been dubbed 'the twilight zone of public expenditures'.⁽²⁵⁾ It is therefore noteworthy that a Parliamentary enquiry recently initiated by the Senate Economics References Committee is examining the structure and distributive effects of the Australian taxation system. One of its terms of reference is to inquire into 'the use and efficiency of tax and expenditure incentives to influence social and economic conduct, for instance participation in the workforce'.⁽²⁶⁾

Some United States public choice theorists consider that tax expenditures are less susceptible to 'capture' by special interests than direct spending programs, arguing that Congressional tax committees are more broadly based and less 'captured' by special interest groups than program committees.⁽²⁷⁾ However, others argue that the lack of parliamentary scrutiny reduces tax expenditure transparency, making them particularly susceptible to lobbying by interest groups.⁽²⁸⁾ The measuring and reporting of tax expenditures is thus said to 'serve to limit the ability of the politically powerful to obtain tax relief at the expense of the less powerful'.⁽²⁹⁾

Purpose and Structure of the Paper

Given the importance of tax expenditures, the purpose of this paper is to assess:

The remainder of this paper is organised as follows:

The next section examines the pattern and trends in tax expenditures, raising issues about the adequacy of information and reporting of tax expenditures and whether tax expenditures have been underestimated.

The centrepiece of the paper is the section, 'Tax Expenditure' or 'Direct SpendingWhy Does it Matter?'. This section lists and describes the various criteria that might be used to evaluate individual tax expenditures. It also examines the report of the National Commission of Audit and issues of budgetary control. It covers matters such as preserving the integrity of the tax system and transparency, political accountability and political decisionmaking.

The concluding section brings things together and proposes a series of measures that could address the problems identified in this paper.

Tax Expenditures in Australia

Emergence of the Treasury's Tax Expenditures Statement (TES)

Public and parliamentary debate over Commonwealth taxation and Budget policy in the early 1980s was accompanied by several changes in the way tax expenditures are reported and examined.

In 1982, the House of Representatives Standing Committee on Expenditure expressed concern at the information base of tax concessions and their budgetary cost, and strongly recommended improvement.⁽³⁰⁾ The Treasurer supported this recommendation in a statement on 27 March 1985. Although since 198081 the Commonwealth Treasury had listed major Commonwealth income tax expenditures in an Appendix to Statement No. 4 of Budget Paper No. 1, these were without comprehensive costings.⁽³¹⁾ However, since 1986, the Commonwealth Treasury has published annual estimates of the cost of tax expenditures in Tax Expenditures Statements (TES).

The TES is now published as one of the commitments under the Charter of Budget Honesty Act 1998.⁽³²⁾ The Commonwealth Government's other commitments under this Act are discussed later in this paper.

Most states also now publish tax expenditure information relating to state taxes in their annual budgets. State tax expenditures are discussed below after the examination of trends and patterns in Commonwealth tax expenditures.

Purposes of the TES

Treasury considers that the primary purpose of the TES is 'to provide estimates of the value of concessions received by individuals and businesses as a result of tax expenditures'. This in turn permits review and scrutiny as to 'whether objectives are met at reasonable cost' compared with that for direct expenditures, and 'facilitates a comparable degree of scrutiny for tax expenditures as occurs for direct expenditures'.⁽³³⁾

Another purpose of the TES is to allow comparison of the assistance that the Commonwealth Government provides to different sectors:

An examination of direct expenditures alone can present a misleading picture, particularly since the benefits derived from tax expenditures are for some sectors, greater than those derived from direct expenditures.⁽³⁴⁾

Issues in the Calculation of Tax Expenditures

The Treasury TES excludes many tax items which arguably might be considered as tax expenditures. Some of the exclusions are too difficult to estimate or are apparently not costed in the TES due to lack of relevant information, for example:

Some of these tax expenditures are measures for which revenue cost estimates could have been expected to form the basis of decision-making at the time the concession was introduced.

Other tax provisions can be viewed as preferential but are counted as part of the benchmark and so are not identified as tax expenditures. The more debateable exclusions are:

Unofficial estimates suggest the cost of these excluded items would exceed several billion dollars annually. For example, dividends, interest and royalties paid to non-residents are generally subject to a withholding tax that is the final tax liability for these payments. The interest withholding tax is imposed at a flat rate of 10 per cent on gross interest and royalties paid to non-residents, with a 15 per cent rate generally applying to dividends.⁽³⁵⁾ The revenue cost of this concession compared to a standard rate of 30 per cent is around $5.2 billion.⁽³⁶⁾

Likewise, the Senate Community Affairs References Committee's 1997 Report on Housing Assistance presented estimates of the revenue loss from negative gearing tax concessions of between $800 million and $900 million for the year 199495.⁽³⁷⁾ The revenue loss increased to about $1.4 billion when the loss to partnerships, trusts and companies for similar types of investment was included. Assuming the cost of this tax expenditure has risen at least as fast as GDP since 199495, the concession would have cost around $2.1 billion in 200102.

The general benchmark for the personal income tax is taken in the TES to be the individual. For this reason, tax credits allowed to taxpayers with responsibility for financial support of dependent children are counted in the TES as tax expenditures. Treasury's practice of counting tax rules facilitating income splitting as part of the benchmark thus raises the broader issue of whether the tax preferences that result from judicial decisions that are not then overturned by statute should be identified as a tax expenditure.⁽³⁸⁾

Debate over what the benchmark includes also gives rise to debate over whether the exemption from the one per cent Medicare Levy surcharge for those with private hospital insurance should be identified as a positive tax expenditure or as a tax penalty (negative tax expenditure).⁽³⁹⁾

Aspects of the measurement of some tax expenditures are also questionable with the effect of possibly underestimating the cost to Commonwealth revenue for tax deferral provisions such as accelerated depreciation. The TES treatment of deferral tax expenditures assumes a zero discount rate for such provisions. This understates the cost of such tax expenditures although Treasury considers that this treatment makes measurement comparable and complementary to other Budget estimates.⁽⁴⁰⁾ By deferring the collection of revenue, the provision means the Commonwealth net debt is greater by the amount of the interest cost of deferred taxation. Attributing a positive discount rate to the deferral of tax revenue due to accelerated depreciation would suggest a total budgetary cost of accelerated depreciation of around $3.8 billion for the twelve-year period the provision is included in TES.⁽⁴¹⁾

A further debateable omission from the 2001 TES are tax expenditures relating to the GST. According to Treasury:

Only tax expenditures that relate to Commonwealth taxes are reported in Table 5.1. As the GST is imposed and collected by the Commonwealth on behalf of the States, and the proceeds of the GST are not reported as Commonwealth revenue, this Statement does not cover GST.⁽⁴²⁾

However, neither the Australian Bureau of Statistics nor the Commonwealth Auditor-General treat the GST as a state tax under existing statistical or accounting conventions.⁽⁴³⁾ Tax expenditures relating to the GST, such as exemptions for certain foods, rent, and health or education services, and input taxing of the financial services and building industries, might justifiably be reported as a tax expenditure. The magnitude of these exclusions is likely to be large. For example, exempting food from the GST was estimated to cost $3.1 billion in 19992000.⁽⁴⁴⁾

Likewise, the TES no longer reports customs duty concessions as tax expenditures, although this would be consistent with the treatment of excise duties.

In addition, since the 2001 TES and the 200203 Budget were prepared, a number of major new tax expenditures have been announced. These include additional concessions for superannuation (estimated cost around $142 million for 200405),⁽⁴⁵⁾ the first child tax rebate ($390 million in 200405),⁽⁴⁶⁾ and the national excise scheme for low alcohol beer ($69 million in 200304).

There are other aspects of the TES data which make their interpretation difficult, especially in examining trends over time. Aggregating estimates of individual tax expenditures is particularly problematic. These issues are discussed in more detail in various editions of the TES but include:

It is also worth noting that the progressive personal income tax scale is considered part of the 'benchmark' for income taxation. This means that reductions in personal tax rates reduce the measured costs of some tax expenditures such as tax deductions or exemptions. On the other hand, the benchmark does not include any particular personal income tax scale so that, for example, cuts in tax rates that favour particular income brackets, or increases in tax rates in other brackets, are not counted as tax expenditures.

The data used below have attempted as far as possible to adjust for the effects of revisions and changing coverage of the TES, but for the reasons discussed above, care must be taken when interpreting these aggregates, particularly when making comparisons over time. In particular, these aggregates include only those tax expenditures costed in the TES, thus excluding uncosted tax expenditures and those that Treasury does not identify as tax expenditures.

Patterns and Trends in Commonwealth Tax Expenditures

Commonwealth measured tax expenditures totalled $29.2 billion or about 4.5 per cent of GDP in Australia in 200102. By comparison, Commonwealth outlays are around 23.3 per cent of GDP.⁽⁴⁷⁾

The considerable magnitude of tax expenditures highlights how misleading international comparisons of public spending may be if account is not taken of the different policy 'tools' used in various countries.⁽⁴⁸⁾ Government spending could, in principle, be reduced simply by redesigning programmes as tax reliefs or tax incentives.⁽⁴⁹⁾

The Commonwealth's tax expenditures are unevenly distributed across functions. Table 1 shows that a large proportion of tax expenditures are categorised under the 'social security, health and welfare' function, although this may be overstated due to the possible underestimation of tax expenditures falling under other functions, as noted earlier.

Table 1. Tax Expenditures Share by Function, 200001

Source: Treasury, 2001 Tax Expenditures Statement, Australian Government Publishing Service, Canberra, 2001.

It is also apparent that tax expenditures are extensively used as a delivery mechanism for some government functions and much less so for other functions. Table 2 shows that, for example, much of the Commonwealth's budgetary assistance to the mining and manufacturing sector is through tax expenditures, whereas most social security, health and welfare spending is through direct spending programs.

Table 2. Tax Expenditures as Share of Budgetary Assistance by Function, 200001

Source: Treasury, 2001 Tax Expenditures Statement, Australian Government Publishing Service, Canberra, 2001.

However, the functional classification of tax expenditures is based on the legal, not economic incidence of the tax concession. So, for example, tax expenditures for private health insurance or superannuation may provide industry assistance through permitting higher premiums or asset management fees, notwithstanding the fact that individuals access the concessions. While recognising that the question of who finally bears the tax burden (or tax relief) may be unresolved, this problem arises similarly in allocating direct subsidies. According to the OECD, even when there is a clear duality of beneficiaries most countries allocate tax expenditures 'to the taxpayer who immediately and directly benefits from them'.⁽⁵⁰⁾

It is also apparent that the same major tax expenditure items have remained important over the last two decades. For example, as can be seen from Table 3, concessional treatment of superannuation, rebates for recipients of taxable pensions and so on, as well as exemptions from interest withholding tax, which were identified as major items in the early 1990s, remained prominent even a decade later.

Table 3. Major Tax Expenditure Items, Actual and Projected, Selected years:
199192 to 200405

Note: Grey shading indicates not applicable

Sources: Treasury, Tax Expenditures Statements, Australian Government Publishing Service, Canberra, various years.

 Although the tax reform process resulted in a winding back of tax expenditures during the late 1980s and early 1990s, there has been an increasing use of tax expenditures as a tool of government fiscal policy since then.⁽⁵¹⁾ A listing of significant tax expenditure measures introduced since 1996 is provided at Appendix 1.

Figure 1 shows trends in measured Commonwealth tax expenditures, including the trend when the