Kirsty Laurie and Jason McDonald1
This paper provides a summary of trends in government spending. It reveals strong growth in government spending and the size of government, particularly over the past four years. It also discusses the distribution and sustainability of spending and notes the importance of high quality spending and flexibility in resource allocation in responding to future pressures.
Governments intervene in the economy to achieve a number of policy outcomes, including addressing market failures or improving social equity by redistributing resources. Direct expenditure is one mechanism that governments can use to implement their policies. Spending often has advantages over alternative policy mechanisms, such as tax expenditures, regulation, guarantees and loans. In particular, spending tends to be more transparent, better allowing the community to hold government accountable for their decisions. For example, it is often difficult to determine the burden, distribution and sustainability of regulation as the economic effects are difficult to measure. The greater accountability on spending also means that it is often the most effective means of achieving government policy objectives.
This paper seeks to analyse recent government spending trends by assessing the size, distribution and sustainability of Australian Government spending.
Size of government spending
As spending needs to be financed through revenue, spending has associated costs caused by taxation distorting resource allocation and reducing economic growth. The higher the tax rates, the higher the distortion, so all other things being the same, higher government spending will reduce economic growth. But, importantly, spending may be either welfare reducing or enhancing, depending on whether the benefits from the spending are greater than the costs of taxation needed to finance it. High levels of good quality spending may involve benefits greater than the costs of taxation. The budget task is to identify and reduce spending that is of low value or that reduces welfare, allowing for either lower taxes or for spending which is of higher social value and adds, overall, to wellbeing. Some spending may address social needs that do not increase measured economic growth, but do improve societal wellbeing. The size of government is therefore a decision of social choice involving trade-offs between economic growth and other social objectives. The overall objective should be to increase general wellbeing.
Spending growth in nominal and real terms
Spending as a proportion of GDP is a measure of the level of direct government involvement in overall economic activity. Measuring spending as a proportion of GDP has at least two benefits. First, it provides a comparable base for analysing spending through time. Unlike nominal dollars, spending as a proportion of GDP provides a meaningful comparison between years of relative resource use. Second, spending as a proportion of GDP shows the relative extent of government intervention in the economy and therefore assists in analysis of social choice. GDP represents the resources available and spending represents the share of those resources allocated by government through the budget.
Our analysis reveals an increase in levels of spending over the past 35 years. In particular, over the past decade the total dollar value of Australian Government spending (including GST payments to the State and Territory governments)2 has grown by 54 per cent since 2000-013 from $176.9 billion to an estimated $272.2 billion in 2007-08 (see Chart 1). Based on the Pre-election Economic and Fiscal Outlook 2007, spending is projected to grow further to $314.3 billion by 2010-11 (an increase of 78 per cent since 2000-01). This equates to 5.9 per cent growth per annum.
Chart 1 below shows that, as a proportion of GDP, spending is estimated to fall from around 26 per cent in 2000-01 to around 24 per cent in 2007-08.
Chart 1: Nominal payments and payments as a per cent of GDP
Source: Commonwealth of Australia 2007a and 2007c.
However, Australia’s recent terms of trade increases have led to a significant rise in nominal GDP, reducing the effectiveness of the spending to GDP ratio as a measure of government resource use. The terms of trade effect on nominal GDP is masking a significant increase in real spending — a measure which more effectively represents the government’s call on real resources. Using the consumer price index (CPI) to convert spending into 2006-07 dollars, real government spending has grown significantly over the past decade, from $174.7 billion in 1997-98 to $264.1 billion in 2007-08, and is projected to grow to $282.1 billion by 2010-11 (see Chart 2).
Source: Commonwealth of Australia 2007a and 2007c.
Excluding GST payments to the State and Territory governments, real government spending has grown faster in the period from 2004-05 to 2007-08 than in any other four-year period since the 1990s recession.
Chart 3 below shows the growth in real government spending since 1972-73, with the shaded area representing those years in which it was generally recognised, based on a range of indicators, that the economy was in a recession. The recent growth in spending stands out, along with the growth in spending under the Whitlam Government in 1974-75 and the increased spending following the recessions in 1982-83 and 1990-91.
The recent growth in spending is particularly noteworthy given Australia has experienced 17 consecutive years of real GDP growth. The economy is currently operating at close to its limits of capacity. Unemployment has fallen to 4.1 per cent, a 33 year low, and capacity utilisation is at a record high of 84.2 per cent. By way of comparison, during the recession in the early 1990s the unemployment rate peaked at 10.9 per cent and capacity utilisation fell to 75.7 per cent. In the current environment the costs of the government drawing on the economy's resources are clearly higher compared to earlier periods since resources must be redirected from other economic activities instead of from idle capital or unemployed labour.4
Chart 3: Growth in real government payments and growth in real GDP
Growth in policy decisions
The growth in real spending in recent years reflects both an increase in the number of policy measures and the cost of these measures.
The number of decisions (including tax and savings measures) announced in the Budget or Budget updates for particular years has more than doubled over the past decade from 359 in 1997-98 to 825 in 2007-08 (see Chart 4).
Chart 4: Number of measures
Source: Australian Treasury.
Most of the new spending decisions have been for small amounts. Around 90 per cent of decisions taken each year have totalled less than $100 million over the forward estimates. However, the number of decisions valued between $100 million and $249 million has grown from 16 in 1997-98 to 49 in 2007-08, and the number of decisions worth over $1 billion dollars has risen from one in 1997-98 to nine in 2007-08 (see Chart 5). Of note, the number of decisions valued between $500 million and $999 million has not shown as much variation, despite the growth in the total number of decisions.
In addition, there has been a reduction in the number and proportion of savings measures included in Budget reports since the 1997-98 Budget.5 In the 1997-98 Budget, close to a third of all measures had a savings component, whereas more recently, savings measures have averaged around 1.5 per cent of total measures.
Chart 5: Number of measures $100 million or more
Source: Australian Treasury.
Chart 6 shows the number of savings measures as a proportion of the total number of measures and also shows savings measures as a proportion of the gross value of all measures in each year.6 In both cases, the proportions clearly decrease over the period, especially following the spike in 2002-03.7
Chart 6: Savings measures as a proportion of the total number and gross value of measures (including MYEFO measures)
Source: Australian Treasury.
As a result of the increase in spending measures and the fall in the number and value of savings measures, there has been a growth in total spending over time. The effect of the accumulation of policy decisions on the growth in ‘base’ spending (that is, the level of spending which would have occurred if decisions since the 1999-00 mid-year update had not been taken) can be seen in Chart 7.
Spending can also be delivered through the tax system as tax concessions (‘tax expenditures’). Since 1997-98 there has also been growth in real tax expenditures. Tax expenditures have grown by 51 per cent in real terms since 1997-98, from $33.1 billion in 1997-98 to $49.9 billion in 2007-08, and are expected to grow further by 2010-11 (see Chart 8). These expenditures have similar impacts on the economy as conventional spending.
Chart 7: Accumulation of policy decisions (expenses)8
Source: Australian Treasury.
(a) Note these expenditures show the impact on taxpayers of deviations from the tax treatment that would normally apply. Tax expenditures are estimated on an assumption of no behavioural change and are therefore not necessarily reliable indicators of the budgetary impact of removing particular tax concessions. Consequently, tax expenditures are not additive to direct spending.
Source: Commonwealth of Australia 2008, 2006c, 2005c, 2005d, 2004c, 2003c, 2001c.
Distribution of government spending
Governments have a unique role in redistributing resources in the economy due to their ability to compulsorily acquire resources through taxation and regulation. Analysis of functional spending can therefore assist the community in determining whether government priorities reflect community priorities.
Distributional analysis is also important for assessing the quality of spending decisions. Not all spending has the same impact on the economy. Governments can either spend on current consumption (such as goods and services) or invest for future consumption (for example, financing superannuation). They can spend in ways that either improve aggregate economic supply (such as public economic infrastructure) or reduce it (for example, some forms of industry protection); and can increase aggregate demand (for example, through own purpose government consumption) or not.
Spending by function
In the 2007-08 Budget, estimated expenses excluding GST for 2007-08 were $236 billion. The allocation of spending is detailed in the chart below, including 41 per cent on social security and welfare, and 18 per cent on health.
Chart 9: 2007-08 Budget expenses — functional splits
Source: Commonwealth of Australia 2007a.
Although whole-of-government functional historical data published in Statement 6 of Budget Paper No. 1 are not updated to reflect function classification or accounting changes, broad conclusions can be reached based on these data.
Charts 10 and 11 show that since 1972-73 spending on social security in real terms has been growing substantially faster than other areas of Australian Government expenditure, followed by health. Other purposes financing (mainly transfers to the State and Territory governments) has fallen relative to other forms of spending, although this does not include GST payments.
Charts 10 and 11: Real spending on functions
Source: Australian Treasury based on Budget papers since 1973-74.
Social security spending can be broken down into sub-functions. Again, the data are not strictly consistent across time and are subject to the same caveats but the risk of non-comparability is lessened by focusing on more recent years. In real terms, assistance to the aged and families with dependent children has been rising strongly and assistance to the unemployed falling (see Chart 12). Program level data on growth in spending support this finding, with strong real growth in the Maternity Allowance, Community Care, the Carer Payment and the Carer Allowance since 2003-04.
Chart 12: Social security and welfare real spending
Source: Commonwealth of Australia 2007b, 2006b, 2005b, 2004b, 2003b, 2002b, 2001b.
Transfers provided through social security often affect the economy indirectly by changing individuals’ decisions to supply labour or save their incomes. But when governments directly spend on goods or services they also directly influence resource use in the economy. For example, government spending on industry assistance and development has had an average annual compound growth rate of 6 per cent from the on-set of the commodity boom in 2003-04 to 2006-07.9 Where such spending directly addresses market failures, it may be improving economic supply. However, where it does not, it distorts the allocation of resources. This places more pressure on aggregate prices to redirect resources in the economy, particularly in an economy close to full employment.
Government own purpose consumption spending
Government own purpose consumption is spending on goods and services by the government, as opposed to capital purchases or transfers to individuals and other entities.10 This is likely to be a better measure of the government’s measured stimulus of the economy than total spending since such spending feeds directly into aggregate demand. In contrast, transfer payments (such as unemployment benefits) do not feed directly into the calculation of aggregate demand. They must first feed through private decision makers who can choose whether to save or consume.
Chart 13 shows that government real own purpose consumption has grown from $53.3 billion in 1997-98 to $72.5 billion in 2005-06. As a proportion of total government expenses, government own consumption has grown slightly from 28 per cent in 1997-98 to 29 per cent in 2005-06.
Chart 13: Australian Government real own purpose consumption and own purpose consumption as a proportion of total Australian Government expenses
Source: Commonwealth of Australia 2007a and ABS 2006b.
As a percentage of GDP, government own purpose consumption has fluctuated over the period, averaging 7.2 per cent of GDP — higher than the proportion of GDP in 1997-98.
In the 1990s, government own purpose consumption grew slower than GDP. However, in recent years, government own purpose spending growth is beginning to match or exceed GDP growth.
Number of public servants
Spending on public servant salaries is one component of government own purpose consumption spending.
While the number of public servants has grown since 1997-98 (when the average (full-time equivalent) staffing level (ASL) was 163,297), there was a change to data collection in 1998-99 which makes comparison with earlier periods difficult. However, even when considered over a slightly shorter period, there is still an upwards trend for average staffing levels. In fact, employment in the public service has been growing faster than employment in the rest of the economy. Chart 14 illustrates that total ASL has increased by 29 per cent since 1998-99 from 189,137 to an expected 243,859 in 2007-08. This equates to average annual compound growth of 2.9 per cent per annum compared to average annual compound growth in full-time equivalent employment of 2.1 per cent.11
This growth is noteworthy given that the proportion of public servants with at least a bachelors degree is almost twice that in the private sector (40 per cent compared with 23 per cent).12 Therefore growth in the public sector is likely to be reducing the supply of highly educated labour for the rest of the economy.
Chart 14: Total average staffing levels in the general government sector
Sustainability of government spending
Finally, we look at the sustainability of government spending through time.
Under the Charter of Budget Honesty Act 1998, the Government is required to release a report at least every five years that examines the long-term sustainability of current government policies over the following 40 years.13 The second Intergenerational Report (IGR) was released by the former Government in April 2007. It found that the fiscal sustainability of the Australian Government had improved since the first Intergenerational Report was released in 2002 but that demographic and other factors would continue to provide challenges for economic growth and long-term fiscal sustainability. Spending pressures according to government policies at the time resulted in a projected ‘fiscal gap’ between revenues (assumed to be a constant share of GDP) and projected spending in 2046-47 of around 3.5 per cent of GDP.
The improved outcome for the fiscal gap (see Chart 15) was a result of lower growth in projected spending per person (mainly in health)14 and higher projected nominal GDP per person compared to the first IGR. This latter effect was primarily due to the rise in the terms of trade since the release of the first IGR.15 Consequently, the fiscal gap may have been larger in the absence of the strong terms of trade experienced over the past few years. This highlights that spending projections can be unexpectedly influenced by external events and that continuing reforms are necessary to improve productivity and participation to ensure that the government’s finances can be made more sustainable.
Chart 15: Fiscal Gap
Source: Commonwealth of Australia 2007d.
From the 2004-05 Budget to the 2007 Pre-election Economic and Fiscal Outlook, parameter and other variations have added $391 billion to the budget surplus over the period 2004-05 to 2010-11, while new spending decisions (including income tax cuts) have reduced the surplus by $314 billion over the same period. Revenue variations contributed $334 billion to the budget surplus. Effectively, the additional revenue from the commodity boom has been spent, or provided as tax cuts (see Chart 16).
The terms of trade have led to a significant increase in the GDP deflator since 2003-04. Normally the GDP deflator and CPI move together, but these indices have diverged significantly over recent years. This has resulted in spending as a proportion of nominal GDP falling, even though the government’s claim on the quantity of national output has been rising. The increasing GDP deflator has effectively been masking changes in the underlying size of government. This means that more of the goods and services produced in the economy are affected by government spending decisions, with the growth in government being financed by higher relative prices for the goods and services the country sells overseas.
Chart 16: Cumulative impact on underlying cash balance of
policy decisions and parameter variations
Source: Australian Treasury.
The masking effect of the increase in nominal GDP can be seen by comparing spending as a proportion of nominal GDP and as a proportion of ‘adjusted GDP’, which is constructed by inflating real GDP by the consumer price index (rather than by the GDP deflator). Chart 17 shows that if it were not for the significant growth of the GDP deflator over recent years, payments would have increased as a proportion of GDP over the past five years.
The policy question is whether government spending growth should be moving in line with the growth in nominal GDP. Some elements of recent government spending growth have related to the redistribution of revenues from the increase in the terms of trade through spending on transfers. However, analysing spending as a proportion of real GDP may be a better indicator of the sustainability of government finances and the impact on the long-run improvements to wellbeing. Real GDP growth is directly influenced by domestic policy choices affecting productivity and participation growth. Government spending financed from nominal GDP growth that does not improve the prospects for future real GDP growth may not be as sustainable.
Chart 17: Effect of significant growth in the GDP deflator on (nominal)
payments as a proportion of GDP
Source: Australian Treasury.
Despite cyclical savings in spending, total government spending has grown significantly over the past decade and in particular since 2004-05. Much of this growth may reflect the strong fiscal outlook. However, even with a strong fiscal outlook it is important to have high quality spending. This can help the sustainability of the government’s finances by focussing on measures that enhance Australia’s productive capacity. It is also important because most spending measures are ongoing and they reduce the flexibility of the government to respond to future pressures.
The budget process provides robust analysis of new spending proposals but this is only a small fraction of overall spending. Effective ongoing review arrangements are important for ensuring that the overall stock of programs (including tax expenditures) remains aligned with government priorities. An effective budgetary framework also ensures that changing priorities are addressed through the reallocation of resources (not only through incremental increases in resourcing), that programs are managed efficiently and effectively, and that there is maximum scope for flexibility to respond to future pressures, including the emerging fiscal pressures from demographic change.
Spending trends excluding GST
All charts on government spending in this article include payments of goods and services tax (GST) to the State and Territory governments, consistent with the practice adopted by the Government. This results in a jump in spending in 2000-01 from the introduction of the GST. However, even if GST payments are not included, our analysis reveals strong growth in real payments in the past decade. This is shown in the following table:
Table: Real Australian Government payments excluding GST
Source: Australian Treasury.
Data on nominal spending in this paper have been deflated by the consumer price index (CPI) in order to examine changes in ‘real’ (rather than nominal) spending. A CPI deflator has been used, rather than a GDP deflator, to convert nominal spending into 2006-07 dollars as price impacts on government expenditure depend mainly on consumer prices and nominal wages.
Comparison with other statistics on the number of public servants
Note that the discussion on the number of public servants in this article has used Government Finance Statistics data from the Budget papers. Chart 18 below highlights that the story of the growth in the number of public servants differs depending on which statistics are used. For example, the ABS data show a decline in the number of public servants since 1997-98. The Australian Public Service Commission statistics (APSC) and Budget statistics, on the other hand, show an increase.
Chart 18: Comparison of public servant numbers16
Source: APSC 2007, Commonwealth of Australia 2007a, 2006a, 2005a, 2004a, 2003a, 2002a, 2001a, 2000a, 1999a, 1998a and ABS 2007a.
These different outcomes result from different classifications and methodologies used in each of these sources. The budget statistics provide an estimate of the average staffing level in the Australian Government general government sector while the APSC statistics include only those covered by the Public Service Act 1999 and do not include permanent defence force members. The ABS statistics also exclude defence force members and exclude employees based overseas but are more comprehensive as they include all entities that report to Parliament, including those covered by the Commonwealth Authorities and Companies Act 1997. In addition, the APSC and ABS statistics use a head count approach, which weights part-time employees equally with full-time workers, while the budget statistics use a full-time equivalent.
Which data are best depends on the issue at hand. In seeking to understand the extent to which the ‘bureaucracy’ serving the government has grown, the number of full-time equivalent public servants is of most relevance. We therefore consider the budget numbers to be the best source.
ABS 2007a, cat. no. 6248.0.55.001, Wage and Salary Earners, Public Sector, Australia, Tables 2 and 5.
ABS 2007b, cat. no. 6302.0, Average Weekly Earnings, Australia, Table 1.
ABS 2006a, cat. no. 6227.0, Education and Work, Australia, Table 12.
ABS 2006b, cat. no. 5204.0, Australian System of National Accounts, Table 45.
Australian Public Service Commission 2007, State of the Public Service Report 2006-07, http://www.apsc.gov.au/.
Commonwealth of Australia 2008, Tax Expenditures Statement 2007, http://www.treasury.gov.au.
Commonwealth of Australia 2007a, Budget Paper No. 1, Budget Strategy and Outlook 2007-08, Canberra.
Commonwealth of Australia 2007b, Final Budget Outcome 2006-07, Canberra.
Commonwealth of Australia 2007c, Pre-election Economic and Fiscal Outlook 2007, Canberra.
Commonwealth of Australia 2007d, Intergenerational Report 2007, Canberra.
Commonwealth of Australia 2006a, Budget Paper No. 1, Budget Strategy and Outlook 2006-07, Canberra.
Commonwealth of Australia 2006b, Final Budget Outcome 2005-06, Canberra.
Commonwealth of Australia 2006c, Tax Expenditures Statement 2006, http://www.treasury.gov.au.
Commonwealth of Australia 2005a, Budget Paper No. 1, Budget Strategy and Outlook 2005-06, Canberra.
Commonwealth of Australia 2005b, Final Budget Outcome 2004-05, Canberra.
Commonwealth of Australia 2005c, Tax Expenditures Statement 2005, http://www.treasury.gov.au.
Commonwealth of Australia 2005d, Tax Expenditures Statement 2004, http://www.treasury.gov.au.
Commonwealth of Australia 2004a, Budget Paper No. 1, Budget Strategy and Outlook 2004-05, Canberra.
Commonwealth of Australia 2004b, Final Budget Outcome 2003-04, Canberra.
Commonwealth of Australia 2004c, Tax Expenditures Statement 2003, http://www.treasury.gov.au.
Commonwealth of Australia 2003a, Budget Paper No. 1, Budget Strategy and Outlook 2003-04, Canberra.
Commonwealth of Australia 2003b, Final Budget Outcome 2002-03, Canberra.
Commonwealth of Australia 2003c, Tax Expenditures Statement 2002, http://www.treasury.gov.au.
Commonwealth of Australia 2002a, Budget Paper No. 1, Budget Strategy and Outlook 2002-03, Canberra.
Commonwealth of Australia 2002b, Final Budget Outcome 2001-02, Canberra.
Commonwealth of Australia 2001a, Budget Paper No. 1, Budget Strategy and Outlook 2001-02, Canberra.
Commonwealth of Australia 2001b, Final Budget Outcome 2000-01, Canberra.
Commonwealth of Australia 2001c, Tax Expenditures Statement 2001, http://www.treasury.gov.au.
Commonwealth of Australia 2000a, Budget Paper No. 1, Budget Strategy and Outlook 2000-01, Canberra.
Commonwealth of Australia 2000b, Final Budget Outcome 1999-00, Canberra.
Commonwealth of Australia 1999a, Budget Paper No. 1, Budget Strategy and Outlook 1999-00, Canberra.
Commonwealth of Australia 1999b, Final Budget Outcome 1998-99, Canberra.
Commonwealth of Australia 1998a, Budget Paper No. 1, Budget Strategy and Outlook 1998-99, Canberra.
Commonwealth of Australia 1998b, Final Budget Outcome 1997-98, Canberra.
Commonwealth of Australia 1998c, Charter of Budget Honesty Act 1998, No. 2, 1998, http://www.comlaw.gov.au.
National Australia Bank 2007, Quarterly Business Survey, December Quarter 2007, http://www.nab.com.au.
1 The authors are from Budget Policy Division, the Australian Treasury. This article has benefited from comments and suggestions provided by Gordon de Brouwer, David Gruen, David Martine, Tony McDonald, Adam McKissack, Penny Sirault, Anupam Sharma, David Tune, Megan Thomas and Lukas Weber. The views in this article are those of the authors and not necessarily those of the Australian Treasury.
2 Unless otherwise stated, GST payments to the State and Territory governments have been included in the analysis from 2000-01. These payments replaced the Financial Assistance Grants that are included in the data prior to 2000-01 as well as a number of State and Territory taxes (which are not included).
3 2000-01 is the first year in which GST payments are included in the data.
4 ABS 2008 and National Australia Bank 2007.
5 A measure is included as a saving in this analysis if it reduced an entitlement for any entity, or if the measure were made to better target policy outcomes. Efficiency improvements in government departments, compliance measures or indirect savings have been excluded.
6 Some of the savings measures identified were included as a part of a package of measures. As there was no breakdown of the financial impact of these measures, the value of these savings may be underestimated as the value reflects the net value of the new spending proposals.
7 The spike in 2002-03 is a result of the 2002-03 Budget and Mid-year Economic and Fiscal Outlook (MYEFO) having the smallest net expenditure of any year from 1996-97. Furthermore, a number of savings were achieved in that year, the most significant of which was the reform to the Pharmaceutical Benefits Scheme, realising savings of almost $1.2 billion.
8 This chart shows the accumulation of expense decisions in nominal terms. However, the upward trend remains in real terms, with the base spending growing from $182.7 billion in 2000-01 to $198.9 billion in 2010-11 (in 2006-07 dollars). A base year of 1999-2000 has been chosen due to the introduction of accrual accounting in 1999-2000.
9 Industry assistance and development is estimated to grow by 14 per cent in real terms from 2006-07 to 2007-08 based on estimates in the Mid-Year Economic and Fiscal Outlook 2007-08.
10 Government own consumption is defined as the net expenditure on goods and services by public authorities (other than those classified as public corporations) which does not result in the creation of fixed assets or inventories or in the acquisition of land and existing buildings or second-hand assets.
11 Australian Treasury and ABS 2008.
12 ABS 2006a.
13 Commonwealth of Australia 1998c.
14 Note that the fall in projected spending per person on health includes the impact of some changes to projection methodologies.
15 Commonwealth of Australia 2007d.
16 Note there are some timing differences between the data from different sources.