Audit Report No. 32 2007-2008
Chapter 4 Preparation of the Tax Expenditures Statement
Background
4.1
Alongside social welfare programs, tax expenditures are two of the
oldest forms of financial assistance provided by the Commonwealth government.
Notwithstanding this, there is no clear definition of a tax expenditure.
Instead, measures constituting a tax measure can change over time and between
jurisdictions. The Taxation Expenditures Statement 2006 defines a tax
expenditure as:
A tax concession that provides a benefit to a specified
activity or class of taxpayer… A tax expenditure can be provided in many forms,
including a tax exemption, tax deduction, tax offset, concessional tax rate or
deferral of tax liability.[1]
4.2
To apply this concept, The Department of the Treasury (Treasury) selects
a ‘normal’, non-discriminatory or benchmark tax system. The benchmark taxation
system should not favour or disadvantage similarly placed activities or classes
of taxpayer,[2] and is the tax system
that would exist if the tax expenditure to be investigated did not exist.
4.3
Once the benchmark is selected, Treasury compares tax expenditures to
the benchmark. Tax concessions that are consistent with the benchmark are not
considered tax expenditures and are not reported, while those that do not match
the benchmark are treated as tax expenditures and are reported.
Reporting on tax expenditures
4.4
Between 1968-69 and 1982-83, some information on tax expenditures was
printed in the Budget papers. From 1987, Treasury (using estimates and advice
from the Australian Taxation Office (ATO)) commenced publication of a separate Taxation
Expenditures Statement (TES). This practice has continued annually since,
with the exception of 1999-2000. The aim of the TES is to:
n allow tax
expenditures to receive a similar degree of scrutiny to direct expenditures;
n allow for a more
comprehensive assessment of government activity; and
n contribute to the
design of the tax system, by promoting and informing public debate on all
elements of the tax system.[3]
4.5
The reporting of tax expenditures also allows private investors to make
decisions with better information on taxation and government assistance in the
Australian economy. This assists in the efficient action of global markets.
4.6
The Charter of Budget Honesty Act 1998 introduced two
requirements for annual reporting on tax expenditures:
n an annual budget
economic and fiscal outlook containing estimated tax expenditures for the
budget year and the three years to follow; and
n a Mid-Year Economic
and Fiscal Outlook (MYEFO) report containing a detailed statement of tax
expenditures to provide information to allow mid-year comparison with the
budget papers.
4.7
The Tax Expenditures Statement 2006 was tabled in December 2006.
It contained details of 272 Commonwealth Government tax expenditures, with an
aggregate value of over $41 billion. This constituted 17.6 per cent of the
year’s estimated value of all government receipts, excluding the Goods and
Services Tax.
4.8
The value of tax expenditures is expected to increase in later years.
Figure 1 illustrates the actual and anticipated growth of tax expenditures
from 2002-03 to 2009-10.
Figure 1 Estimated
Tax Expenditures: 2002-03 to 2009-10
Source ANAO
Audit Report No. 32 2007-08 – Preparation of the Tax Expenditures Statement, p.
29
4.9
Examining the last 10 TESs, an average of 17 new expenditures have been
found each year, with an average of 8 expenditures being withdrawn. It appears
that tax expenditures are withdrawn as a result of sunsetting and policy
changes. The growth in tax expenditures can be attributed to missed tax
expenditures being identified as well as new expenditures being introduced.
The Audit
Audit objectives
4.10
The audit objective was to assess the completeness and reliability of
the estimates reported in the Taxation Expenditures Statement 2006.
Further, the audit examined suggestions for greater transparency in the
reporting of tax expenditures.
Audit conclusion
4.11
The audit report made the following conclusion:
The purpose of the Charter of Budget Honesty Act was to
establish an integrated fiscal framework to provide for greater discipline,
transparency and accountability in fiscal policy. A key element of this
integrated framework was that the MYEFO report was to include detailed
estimates of both tax expenditures and outlays, thereby promoting the scrutiny
of both forms of expenditure. However, due to methodological challenges,
Treasury has not yet found a way to integrate the reporting of outlays and tax
expenditures, with the result that the detailed estimates of tax expenditures
are reported in a separate TES document. Treasury has advised ANAO that it is
not possible to include the full detailed tax expenditure estimates in the
MYEFO release without significant changes to the focus of the MYEFO document
and without delaying the release of MYEFO itself.
Treasury’s view is that the best focus for controlling tax
expenditures is at the policy development stage by ensuring that the Budget
processes require that the cost of any new tax concession proposal (and any
savings offsets) are examined in the same way as occurs for outlays. However,
past practices in this area have been inconsistent. This has been compounded by
shortcomings in the post-implementation measurement, monitoring and reporting
(through the TES) of tax expenditures. In particular:
n the benchmarks used
in preparing the TES are selected by Treasury based on judgements with the
result that benchmarks may vary over time and can be arbitrary; whilst the
Charter of Budget Honesty Act requires the TES to be based on external
reporting standards,[4] neither the Australian
accounting standards or the economic reporting standard issued by the
Australian Bureau of Statistics have been developed to account explicitly for
tax expenditures. In particular, as few tax expenditures arise from direct
transactions and other events of the kind commonly recorded in accounting
systems, neither AAS31 nor Government Finance Statistics (GFS) is designed to
capture all the notional transactions involved in the majority of tax
expenditures. The external reporting standards also do not address the
selection of tax benchmarks;
n there are unreported
categories of tax expenditures. Each TES from TES 1995–96 onwards has
identified, on average, ten tax expenditures arising from tax concessions or
relief already in place but previously unreported. In this respect, during the
course of the audit, Treasury took or foreshadowed action to improve the
coverage of the TES by reporting tax expenditures in relation to Customs Duty
and Goods and Services Tax, as well as expanding the reporting of
superannuation tax expenditures; and
n TES 2006 included
quantified estimates for less than 60 per cent of those tax expenditures that
were reported and, of these, two thirds were not based on reliable estimates.
Modelling of the effect of tax expenditures and estimation of their cost has
been made more difficult by the trend of reducing the compliance burden on
taxpayers, which results in less information being collected from which
estimates can be made. This situation also impedes analysis of whether
individual tax expenditures are achieving their objectives.
Against this background ongoing review of tax expenditures
would be beneficial given the lack of regular, risk-based reviews and
evaluations of tax expenditures as to whether they are achieving their
objectives and, if so, at what cost. Such a review, and ongoing scrutiny of tax
expenditures, would benefit from:
n the development of
standards to govern the integrated reporting of outlays and tax expenditures
under the Charter of Budget Honesty, drawing on international developments in
this area. This should contribute to the development of a more comprehensive
picture of total Commonwealth expenditure, irrespective of the manner in which
it is delivered and provide more rigour over the selection of tax expenditure
benchmarks;
n the identification of
opportunities to better integrate the consideration of outlays and tax
expenditures in the annual Budget process so that the cost of any new tax
concession, and any potential offsetting savings, is fully considered; and
n improvements to the
reliability of those tax expenditure estimates that are published, recognising
that there is a balance to be struck between more reliable estimates and
increasing the demands on taxpayers to provide additional information (the
compliance burden).
Over the last 35 years there have been a number of Government
and Parliamentary reviews of tax expenditures. However, few of the
recommendations of these reviews have been adopted. As a result, each
successive review reported similar shortcomings and made similar
recommendations. ANAO notes that the Government has recently announced[5]
that, before the 2008–09 MYEFO is released, it will undertake a
program-by-program review of government spending and tax concessions with the
objective of increasing efficiency, transparency and accountability.
4.12
To conclude, the ANAO found significant shortcomings in the completeness
and reliability of the estimates reported in the Tax Expenditures Statement
2006. These shortcomings arose from the variation and arbitrary nature of
benchmarks, the discovery of existing but previously unreported tax
concessions, and insufficient data resulting in unreliable estimates.
ANAO recommendations
4.13
The ANAO made the following recommendations:
Table 1.1 ANAO recommendations, Audit Report no. 32,
2007-2008
1. |
ANAO recommends that the Department of the Treasury:
(a)
Develop an approach for the conduct of an ongoing prioritised
review of the existing program of tax expenditures; and
(b)
Publish for each tax expenditure information on the timing and
outcome of the review.
Agency response: Treasury agreed to part (a) and agreed
with qualification to part (b). |
2. |
ANAO recommends that the Department of Treasury examine
and advise Ministers on options to better integrate the consideration of
outlays and tax expenditures in the annual Budget process.
Agency response: Treasury agreed. |
3. |
ANAO recommends that the Department of the Treasury
develop standards to govern the integrated reporting of outlays and tax
expenditures under the Charter of Budget Honesty, drawing on international developments
in this area.
Agency response: Treasury agreed with qualification. |
4. |
ANAO recommends that the Department of the Treasury
promote more comprehensive reporting on taxation expenditures by:
(a)
liaising with Commonwealth entities that collect revenue to
identify all entities that also administer forms of relief from Commonwealth
taxes, including tax expenditures; and
(b)
developing arrangements, as part of the preparation of the
annual Taxation Expenditure Statement, to obtain relevant data from entities outside
the Treasury portfolio.
Agency Response: Treasury agreed. The ATO agreed with part (b). |
5. |
ANAO recommends that the Department of the Treasury and
the Australian National Audit Office identify opportunities to develop
estimates of large or otherwise significant tax expenditures using the
revenue gain method.
Agency response: Treasury agreed. |
6. |
ANAO recommends that the Department of the Treasury:
(a)
develop an approach to prioritise improvements to the
reliability of published tax expenditure estimates;
(b)
examine options for disclosing in the Taxation Expenditures
Statement information on the reliability of individual tax expenditure
statements;
(c)
work with the Australian Taxation Office to develop reliable
models to estimate the revenue forgone for existing tax expenditures that are
large or otherwise significant; and
(d)
when developing advice for Ministers on policies that are
expected to result in a tax expenditure, assess options for the reliable
measurement of the effect of the proposed measure.
Agency response: Treasury and the ATO both agreed with parts (a), (b) and (c) and agreed with qualification to part (d). |
The Committee’s review
4.14
The Committee held a public hearing on Wednesday 17 September 2008, with the following witnesses:
n Australian National
Audit Office (ANAO); and
n the Department of the
Treasury (Treasury).
4.15
The Committee took evidence on the following issues:
n modelling;
§
revenue gained and revenue foregone methods;
§
behavioural responses; and
§
benchmarking;
n reporting;
§
compliance with the Charter of Budget Honesty Act 1998;
§
the Canadian model of reporting; and
n data.
Modelling
Revenue gained and revenue forgone methods
4.16
The Committee noted one of the key findings of the audit was the
differing methods of economic modelling used in preparing the Budget Papers and
the TES. The Budget Papers are prepared using the revenue gain method, while
the TES is prepared using the revenue foregone method.[6]
4.17
The revenue gain method endeavours to account for potential changes in
taxpayer behaviour including ‘second order’ effects such as interactions with
other tax policies. This produces a more accurate estimate of tax expenditures
allowing for more reliable comparison with outlay measure estimates. It must be
noted, however, that the revenue gain method requires more data of higher
quality to produce more thorough modelling.
4.18
The revenue foregone method compares the revenue raised under current
law with the revenue that would have been raised if the tax expenditure
provision did not exist. This method relies on the assumption that the tax law
remains consistent and that behaviour of taxpayers is unchanged. These
assumptions are a weakness of the method. Its major advantage is that it
requires the least amount of data to produce an estimate of a tax expenditure.
4.19
The ANAO noted the impact of the Budget Papers being prepared using the
revenue gain method and the TES using the revenue foregone method:
There is a discrepancy between these two models and it can be
quite substantial or it can be minor. The issue at hand for the audit was that
we suggested firstly that there be some attention paid to applying the model
that is used in the budget process to large tax expenditures so that Parliament
could get a gauge on some of those and their ongoing revenue effects. The
revenue foregone method that is adopted in the TES will not show you those
effects over time.[7]
4.20
Treasury advised that the revenue foregone method was the method used
for the TES as it was the method that best suited the purposes of the TES, and
that it was also the method used by most OECD countries.[8]
Treasury further illustrated the point:
On the outlay side of the budget you do not have forward
estimates which incorporate behavioural response. If you are looking at
spending on aged pension you never ask yourself the question, ‘What would
behaviour look like if we abolished the aged pension?’ You have an accounting
treatment which just says, ‘This is how much we’re spending.’ The revenue
foregone approach parallels that in saying, ‘On the basis of existing
behaviour, this is how much tax we’re not collecting.’[9]
4.21
The Committee is not convinced with Treasury’s reasons for using the
revenue foregone method, as it does not indicate ongoing revenue effects over
time. The discrepancies between the Budget Papers and the TES greatly weaken
the credibility of the revenue foregone method.
4.22
While the Committee understands the difficulty in applying the revenue
gained method in all cases, it is important for the Parliament and the
Australian people to be able to see the ongoing revenue effects of large tax
expenditures over time. Accordingly, the Committee recommends:
Recommendation 7 |
|
That Treasury publish a paper for inclusion in the Tax
Expenditures Statement calculating the twenty largest tax expenditures using
both the revenue foregone and revenue gained methods to allow comparison with
the Budget Papers.
|
Behavioural responses
4.23
The Committee asked Treasury to explain the audit finding of the audit
that the Budget Papers and TES sometimes contained drastically varying figures
for the same tax expenditure.
4.24
Treasury replied that when tax expenditures are considered in the budget
process, attempts are made to consider the behavioural response of taxpayers
and industry. Further, when the TES is prepared, the method undertaken is the
revenue foregone method, which does not incorporate behavioural responses.[10]
4.25
The Committee notes that incorporating the first round (that is, the
immediate) behavioural effects of a tax concession would provide a more
accurate indication of costs to the government.
4.26
The Committee asked how Treasury measured behavioural responses.
Treasury replied that they used a variety of sources including academic studies
indicating how a similar measure had been introduced overseas, and models
internal and external to Treasury. Treasury advised they endeavoured to be as
rigorous as possible as behavioural change had the potential to have large
ramifications.[11]
4.27
The Committee asked whether second and third round (that is, more
indirect macroeconomic ramifications of policy change) effects were taken into
account. Using the example of First Home Saver Accounts, Treasury replied:
…we do stay clear of thinking that the effect would lead to a
change in interest rates or any other general macroeconomic effects. There have
been occasions when the second round has been formally included in Treasury
costings of significant government packages.
…In these cases, the normal guidelines of not including
second round behaviour were waived. Because of the scale of those packages, it
was thought necessary to take a broader economic view.[12]
4.28
Treasury’s acknowledgement of the impact of first round behavioural
responses indicates the weakness of the revenue foregone method. Further, it
illustrates the importance of obtaining better quality data to enable more use
of the revenue gain method to improve the quality of the reporting of tax
expenditures.
Benchmarking
4.29
The Committee noted Treasury and Customs had agreed to adopt a zero rate
for the customs duty tax benchmark, and asked whether this benchmark concealed
tax concessions.
4.30
Treasury replied that the benchmark was set to zero to allow for
comparison with domestically produced goods, arguing that any duty imposed on
imported goods was equivalent to taxes imposed on domestic goods such as
excise. Further, Treasury argued that the zero benchmark allowed measurement of
the benefit tariffs provide to the taxpayer, and that the tariff was shown as a
negative tax expenditure.[13]
Reporting
Compliance with the Charter of Budget Honesty Act 1998
4.31
The Committee noted the discrepancy between the reporting requirements
of the Charter of Budget Honesty Act 1998 and Treasury’s reporting
practices. The Act requires detailed statements on tax expenditures to be
included in the Mid-Year Economic and Fiscal Outlook (MYEFO).[14]
The Committee asked whether the discrepancies between the TES and the Budget
Papers would be lessened if Treasury reported as required by the Act.
4.32
Treasury replied that it was difficult to report on tax expenditures in
MYEFO as a lot of detail was required and Treasury prepared revenue estimates
and costings at the same time MYEFO was prepared. Treasury believed adding tax
expenditure reporting to MYEFO would:
…severely complicate the MYEFO process. It would also
severely complicate the delivery of the important information on the economy,
on expenditure, on revenue and on measures taken since the budget that that
document contains. There is an efficiency issue and there is a resource issue
in terms of the Treasury.[15]
4.33
The Committee asked the ANAO for their opinion. ANAO noted they had not
suggested that tax expenditure information be included within MYEFO, but that
it was a requirement of the Act that was not currently occurring. The ANAO
report noted:
The different methodologies adopted in the Budget papers and
in the TES impedes analysis of the actual cost of new tax expenditures in terms
of what was expected when they were introduced.[16]
4.34
The ANAO continued:
One of the things we did suggest was that Treasury should
publish its own estimate of how reliable that tax expenditure is. Our
perspective was that there is plenty of room in the existing tabulation in the
TES for including one extra item of information which we think would be quite
valuable to tell the reader that this tax expenditure of $X million is, in
Treasury’s view, this reliable rather than that reliable.[17]
4.35
While the Committee understands the challenges facing Treasury in
assembling MYEFO, it is the Committee’s opinion that Treasury should not be
operating at odds with the Charter of Budget Honesty Act 1998. The
Committee will pursue this matter with Treasury and the ANAO.
The Canadian model of reporting
4.36
The Committee notes the tax expenditures reporting model currently used
in Canada, which records measures that are undeniably tax expenditures (as is
done in Australia), and then divides other forms of tax relief into three
‘memorandum items’:
n relief that is part
of the benchmark system;
n relief where data is
insufficient to separate the tax expenditure component from the tax concession
component; and
n relief that could be
categorised either as a tax expenditure or a tax concession.[18]
4.37
This practice of reporting allows the quantum of tax concessions to be
readily available for both public and Parliamentary scrutiny.
4.38
The Committee asked Treasury about the steps it had taken to emulate the
Canadian model. Treasury replied they did not believe their method differed
greatly, as the TES outlined the benchmarks and structural features of the tax
system in a method similar to the Canadian use of ‘memorandum items’.[19]
4.39
Treasury’s method does differ greatly in one area, the Canadian model of
dividing other forms of tax relief into ‘memorandum items’, rather than
aggregating figures, provides a more complete picture. The system employed by
Treasury prevents reporting of tax measures that can be viewed as preferential
but are included as part of the benchmark. These measures, including many
related to capital gains are unofficially calculated to cost several billion
dollars annually.[20]
4.40
The Committee believes the entire quantum of tax expenditures needs to
be reported to improve public and Parliamentary scrutiny. While it is difficult
to categorise all forms of tax relief with precision, the Canadian model errs
on the side of disclosure, rather than precision, allowing the reader a broader
view of a complex part of the tax system.
Recommendation 8 |
|
That Treasury further investigate the merits of the Canadian
model of taxation expenditure reporting, publishing its findings in the paper
proposed in Recommendation 7. |
Data
4.41
In its opening statement, Treasury indicated that there was scope to
improve the reporting of tax expenditures through use of data held by other
agencies. The Committee asked how Treasury had gone about improving and encouraging
data collection. Treasury replied they would seek access to data from other
agencies.[21]
4.42
The Committee noted there were 98 tax expenditures found to be
unquantified, and asked Treasury as to their plans to quantify them.
4.43
Treasury noted data quality was an issue that affected the quality of
the estimate, and that they were looking beyond tax data to improve the quality
of their estimates.[22]
4.44
The Committee is encouraged to hear Treasury’s willingness to seek data
from other agencies to improve its reporting of tax expenditures. The Committee
wishes to see genuine progress being made in improving the reporting of tax
expenditures, and recommends:
Recommendation 9 |
|
That Treasury include information in the Budget Papers on
the extent to which tax expenditure reporting has improved through the
receipt of reliable data from other agencies. |
Conclusion
4.45
The Committee understands the burdens placed on Treasury in its
operations in the current fiscal climate, but believes improvements can still
be made to the reporting of tax expenditures at little cost to Treasury. There
still remains an abundance of data held by other agencies that Treasury is yet
to access. Securing access to this data will improve the reporting of tax
expenditures in the TES.
4.46
The Committee also remains concerned at the method used for calculating
tax expenditures. The Committee is of the belief that there should be little
difference between the figures prepared for the Budget Papers and those
prepared for the TES. These discrepancies occur in large part due to the use of
the revenue foregone method in the preparation of the TES.
4.47
Of even more concern is the lack of data available to illustrate the
quantum of larger tax expenditures over time. Calculating these tax
expenditures using the revenue gained method will provide more clarity for both
Parliament and the public.
4.48
It is important that the practices of Treasury and its obligations under
the Charter of Budget Honest Act 1998 should be compatible. Under the
current situation, Treasury is not meeting its responsibilities under the Act.
The Committee understands the reasons for this, and believes amendment of the
Act to better reflect practice would be the appropriate course of action.
4.49
The Committee notes the review of the Australian taxation system
currently being undertaken by Treasury, and anticipates extensive coverage of
tax expenditures in the review.
4.50
The Committee is pleased with Treasury’s enthusiasm in improving the
quality of data available to it, and understands its reluctance to undertake
practices that may result in an increased cost to Treasury. However, the
Committee is of the opinion that full implementation of its recommendations
would improve the operation of the Charter of Budget Honesty Act 1998,
the quality of the TES and its value as a document for the reporting of tax
expenditures.