This section is divided into two subsections. The first highlights the relevance of tax expenditures in a budgeting and fiscal management context. The second establishes a framework for analysing tax expenditures in that context. The framework comprises six components that were developed from the OECD’s analysis of the four case studies in the Annex.
Use of tax expenditures in budgeting
2.1.1. The use of tax expenditures in central governments
Tax expenditures are often implemented to create incentives for activities considered helpful for growth, wellbeing or social objectives. Some common examples include encouraging savings for retirement by offering tax breaks on retirement savings accounts, encouraging charitable activities by offering tax breaks on donations, or encouraging investment into certain sectors, regions, or products by offering favourable tax conditions for such investments, or to reduce the investment cost of research and development. These examples are frequently included in budget proposals for consideration in the government’s annual budget. There is some evidence that such tax expenditures can be successful in achieving their policy objectives – for example, analysis of firm-level records in 20 OECD Member countries found that R&D tax incentives are effective in boosting business R&D levels, although effectiveness varied significantly across firms and countries (CEPR, 2020[2]). Further research found that R&D tax incentives are less effective for large, R&D intensive firms, which make up a large proportion of R&D tax reliefs (Appelt et al., 2025[3]), and their effectiveness depends heavily on the design, with stable, predictable tax incentives complemented by direct support measures working best (Appelt et al., 2016[4]).
In some contexts, tax expenditures can reduce the government’s administrative costs relative to an initiative that requires direct expenditure. For example, a programme that provides a small benefit to many people is likely to be expensive to administer through direct expenditure but may involve a lower administrative cost for the government through a tax expenditure. This is a relevant consideration when the programmes involve activities that are clearly defined and easily verified (Toder, 2000[5]).
Within the four OECD Member countries in this paper there has been a steady increase in the use of tax expenditures. As shown in Figure 1, growth in the cost of tax expenditures (measured as revenue foregone) has kept pace with GDP in Korea and the four case study countries, and in several cases this growth has increased at a higher rate than GDP growth. While Figure 1 expresses the data as a proportion of GDP, it indicates that as of 2022, the revenue foregone as tax expenditure is equivalent to around a third of the government’s direct expenditure in the Netherlands, around a fifth in the United Kingdom, around 15% in Canada, and around 8% in Ireland and Korea.1 From a budget perspective, tax expenditures therefore represent a significant budgetary cost. While the figures vary, there are strong correlations between which tax expenditures are most dominant. In all four case study countries, at least one tax benefit related to pensions appears in the top ten largest tax expenditures, while in several countries multiple appear (i.e. both pension contributions and employee contributions to superannuation schemes in Ireland). Generally, tax expenditure supporting social policy are well represented, with Canada’s child benefit, Ireland’s medical insurance relief, and the United Kingdom’s employment allowance all appearing among the respective countries’ ten largest tax expenditures.2
Of note in Figure 1, the sharp increase in reported Dutch tax expenditure revenue forgone between 2003 and 2004 reflects an expansion and harmonisation of the tax expenditure annex, including the first full reporting and re-estimation of indirect tax expenditures (Tweede Kamer, 2004[6]).
2.1.2. The impact of tax expenditures on budget deficits
Direct expenditure is easier to estimate, monitor, and control relative to tax expenditure. While both direct and tax expenditure can be difficult to predict ex ante, once in place, direct expenditure can be clearly measured and compared with initial estimates. By contrast, tax expenditures can only be estimated relative to a counterfactual, as the fiscal impact represents revenue that the government could have received in the absence of the measure.
From a budget perspective, this means that the fiscal impact of tax expenditures must be inferred relative to a benchmark, rather than directly observed. In contrast, direct expenditure is quantified by programming expenditure at a detailed level in departmental baselines and reports are prepared to a similar level of detail on the costs incurred. While in theory it is possible to impose limits on tax expenditures, this can be challenging from a policy and administrative perspective and can mean that potential recipients of a tax expenditure are unsure whether they will actually be able to benefit from a programme (Michael, 2018[8]).
Fiscal control of tax expenditures is made more difficult by the large number of tax expenditures that exist in a country. As shown by Figure 2, most of the case study countries (as well as Korea) have over 200 tax expenditures. This leads to complex tax expenditure systems, meaning it can be difficult to detect changes in the cost of a tax expenditure in terms of the revenue foregone, and the government’s capacity to evaluate the performance of each tax expenditure is spread thinly over the large number of tax expenditures.
Furthermore, as shown in OECD country reviews, definitions of tax expenditure benchmarks vary, meaning a country can define the various components of a benchmark relative to the specifics of the policy objectives relating to the tax expenditure and the design of that country’s tax system. For example, a lower rate of VAT on goods such as food are treated as tax expenditures in some countries, while in others they are considered part of the standard tax system. These elements make it difficult for parliaments, scrutinising bodies, such as independent fiscal institutions and supreme audit institutions, and the public to fully understand the impact of tax expenditures on the budget and the government’s financial statements.
2024 or latest available year
Note: The counts in this figure are derived from counting all active line items in the Global Tax Expenditures Database (GTED) which are still active. This can present some issues. For example, in some cases, the same measure can be counted multiple times when there is a line item for each tax base, while in other cases data cannot be disaggregated for multiple tax bases, meaning the measure appears as a single item. This can mean that country’s own counts differ from those presented in GTED. For example, the total count in the 2026 Canadian Report on Federal Tax Expenditures records 229 tax measures (see Canada case study).
The significant budgetary cost of tax expenditures, combined with their complexity and lack of transparency, mean that tax expenditures can pose risks for the fiscal sustainability of public budgets. To be clear, fiscal sustainability is not determined by tax expenditure, but limited mechanisms to effectively control tax expenditures can result in the fiscal cost of the tax expenditures increasing at higher rates than expected, eroding the tax base and constraining fiscal flexibility. This can contribute to and undermine long-term fiscal sustainability.
From 2007-2023, within the four case study countries in this paper, fiscal balances have frequently been in deficit (Figure 3). Tax expenditures are a significant contribution to these deficits, and reducing tax expenditures, where possible and desirable, could improve the fiscal position of these governments. In a context where OECD Member countries are facing multiple spending challenges, including structural pressures on pension and healthcare systems, alongside slowing economic growth and long-term productivity rates (OECD, 2026[9]), analysis of this form of fiscal consolidation is increasingly relevant to restoring public finances.
2007-2023
Source: (OECD, 2025[10]).
2.2.1. Definition and measurement of tax expenditure figures in budgets
Calculating the cost of tax expenditures
Tax expenditures do not have a cost in the way direct expenditures do. Rather, they represent revenue that the government could have received if it had not introduced the tax expenditure. As this is not an observed fiscal outlay, alternative methods to estimate the cost of tax expenditures are needed for forecasting and budget purposes.
The most common way that countries calculate the cost of tax expenditures is through the revenue foregone approach, which quantifies the loss from removing the revenue. In practice, this approach presents several issues. It is a static calculation, in that it does not capture changes in behaviour that the introduction of such an expenditure may induce. For example, if a government introduced a tax deduction for those investing money into certain types of pension funds, the revenue foregone approach would not account for the fact that such pension funds would likely see an increase in usage by the people who are eligible to receive the deduction. Furthermore, the approach estimates the impact of each tax expenditure measure in isolation, whereas the inter-dependencies between tax expenditures may mean that the revenue foregone from one tax expenditure impacts on the revenue foregone from another tax expenditure (OECD, DIAN and Minhacienda, 2021[11]).
Defining the benchmark
A tax expenditure benchmark defines what a country’s standard tax system looks like without giving effect to policies through tax expenditures. Anything outside of this benchmark counts as a tax expenditure. Countries can have different criteria to determine what should form the benchmark. For example, in the United Kingdom, many schools benefit from a wide range of tax reliefs, including an exemption from value added tax (VAT) as well as exceptions from capital gains tax (CGT) and stamp duty land tax. While in many countries such policies would classify as tax expenditure, the United Kingdom government classifies these measures as structural tax reliefs, i.e. part of the benchmark (Aliu and Redonda, 2024[12]). Other areas where variation is common across countries include the treatment of interest income, the simplification of tax regimes for smaller businesses, standard deductions provided to salaried taxpayers, treatment of agricultural income, treatment of gifts and donations, treatment of employee fringe benefits, and depreciation of physical assets. This variation means that it is challenging to compare benchmarks on a cross-country basis to inform budget forecasts.
In 2021, the OECD’s Centre for Tax Policy and Administration developed key criteria to inform the design of a benchmark (Box 1). The criteria show that it is important to consider country-specific variables.
Relevant criteria to consider when developing an effective benchmark.
Well defined and transparent: The benchmark should be well-defined and transparent so that policymakers and the public understand the underlying assumptions that have been made.
No discrimination: The benchmark should represent the standard taxation treatment that applies to similar taxpayers or types of activity (horizontal equity). Discriminatory elements in the tax code should be qualified as tax expenditures. This does not apply to the progressive structure of the personal income tax, which typically is included as part of the benchmark (to support vertical equity).
Avoid negative tax expenditures: Tax provisions that increase the tax burden do not constitute a tax reduction (i.e. an “expenditure”) but rather result in a tax increase. Tax expenditure assessments should try to avoid negative tax expenditures as much as possible, by making these provisions part of the benchmark. Even so, there might be cases where negative tax expenditures are informative, for instance, if the government disallows certain deductions that would be common in a benchmark tax system.
Consistent: The benchmark should be consistent across taxes and make explicit reference to how it treats measures that relieve double taxation (integration). For example, if there is integration of corporate and personal taxation, the benchmark system should be consistently applied to both the assessment of tax expenditures in the CIT and in the PIT.
Actionable: The benchmark should be defined in a way that the resulting list of tax expenditures informs policymakers about reform options. At least, the tax expenditure list provides a starting point for an evaluation of tax concessions
Facilitate international comparability: The benchmark may be chosen such that it follows the approaches in other countries, although these may vary in various dimensions and details. By using a similar benchmark used elsewhere, the tax expenditures might be compared with those other countries, although caution remains important in doing so.
2.2.2. Treatment of tax expenditures vs direct expenditures in the budget
The Global Tax Expenditures Database states that as of 2025, 116 jurisdictions publish generally good data on tax expenditures. Within the OECD Membership, Denmark, Greece, Luxembourg, Norway, Portugal and Türkiye have started reporting on tax expenditures in the last decade. According to the database, a revenue forgone estimate is available for at least one year for all OECD Member countries since 2018. The number of provisions reported has also increased, with the average number of tax expenditure provisions reported within the OECD rising from 98 to 162.
One potential reason for this proliferation is because tax expenditures can be viewed by line ministries as a low-cost way to achieve their policy objectives. In contrast to direct expenditure, the fiscal cost of such measures is reflected in reduced revenue rather than departmental spending, and is not subject to the same budget ceilings or scrutiny processes. As a result, tax expenditures can provide a policy instrument for line ministries to pursue policy goals, while the administrative burden rests with tax authorities and taxpayers.
Even so, reports and figures on tax expenditures do not necessarily contribute to the budget process. This can mean that the information is not available to inform decision making during the budget process, or that the information is reported separately and is not identified as a measured component within the aggregate budget figures. In both instances, this reduces the information available on the cost of the measures within the government’s budget. This reduces budget transparency, as it means that the cost of tax expenditures as revenue foregone is not explicitly stated, nor is the role that tax expenditures perform in the budget to meet policy objectives.
Forecasting the revenue impact of policies that deploy tax expenditure are also challenging. An increase in the cost of a tax expenditure, which is an increase in the revenue foregone, can indicate that a tax expenditure is working well, but can also indicate that it is being used in ways that were not intended or constitute a full windfall. Changes to macroeconomic variables such as economic growth and inflation, as well as changes to underlying tax rates, can all impact the cost of a tax expenditure. The underlying reasons for a change in cost can therefore be difficult to ascertain. In 2020, an analysis by the National Audit Office in the United Kingdom found that actual costs exceeded forecast costs for several tax expenditures, most notably for the research and development (R&D) expenditure credit, where between 2013 and 2018 costs exceeded forecasts by 55% (National Audit Office, 2020[13]). HMRC has stated that discrepancies between official statistics and published forecasts can be due to reductions in the costs of some tax expenditures where taxpayers switch to new ones, economic impacts of a new tax expenditure, and the impact of public finances of related policy changes made at the same time.
Treatment of refundable tax credits
Refundable tax credits are a specific type of tax expenditure that provide a refund when a taxpayer owes no tax, or when the credit exceeds the amount of taxes owed. This means that eligible individuals can receive money from the government. While the tax expenditure appears as part of the government’s tax code, in reality the tax expenditure functions as a cash transfer. Refundable tax credits can be delivered through both the corporate income tax system, e.g. R&D credits, or through the personal income tax system, e.g. child tax credits.
As refundable tax credits depend on uptake, i.e. are demand driven, a budget ceiling or fixed appropriation rarely applies, as this would mean that some recipients would not receive a payment when the ceiling is reached even though they are eligible to do so. The absence of a ceiling or cap on the total cost to the government can have larger than expected consequences for the budget if demand is greater than forecast. While cash payments through tax expenditures reduce tax liabilities, the payments also meet social policy objectives. For example, the Canada Child Benefit is a refundable tax expenditure which helps support families with the cost of raising children, and the United Kingdom’s Universal Credit supports low-income people, those who are out of work or those who cannot work.
2.2.3. Fiscal rules and objectives related to tax expenditures
OECD Member countries recognise that setting fiscal objectives and rules are key to achieving fiscal goals. In terms of fiscal objectives and rules, there are generally four types (OECD, 2025[14]):
Revenue objectives/rules set limits on the tax-to-GDP ratio and constrain how governments can use the revenues they raise.
Expenditure objectives/rules limit the amount of government spending, or the rate of growth in government spending. They can be expressed in nominal terms, real terms or as a percentage of GDP.
Budget balance objectives/rules set requirements related to the budget balance, be that requirement to attain a surplus, to run a balanced budget, or to not exceed a defined deficit limit. They can be expressed in nominal terms or as a percentage of GDP.
Debt objectives/rules limit the amount of government debt that can be accumulated. They can be expressed in nominal terms, as a debt-to-GDP ratio, or as a target reduction of the debt-to-GDP-ratio.
Most countries commit to a range of fiscal objectives and rules. Of the 34 countries surveyed in the Senior Budget Officials Survey on Budget Frameworks, 25 have at least three objectives or rules. Limits on the budget balance are most common, with 34 OECD Member countries having a fiscal objective or fiscal rule that limits the budget balance (OECD, 2025[14]).
Tax expenditures are captured in some, but not all, fiscal rules. As they result in a loss of revenue, they are implicitly captured in revenue objectives or rules, as well as rules that look at the ratio between revenue and expenditure (i.e. debt and budget balance rules/objectives). Despite this, they are not captured by expenditure rules, unless they are explicitly designated as equivalent to direct expenditure in budget processes.
2.2.4. Reporting on tax expenditures in budgeting
Reporting on tax expenditures has seen several developments in recent years. In 2021, the German Institute of Development and Sustainability, alongside the Council on Economic Policies launched the Global Tax Expenditures Database, which collects the information that national governments publish on tax expenditures. The OECD’s databases include the OECD R&D tax incentives database, which monitors availability, key design features, costs, and impacts of tax expenditures used by governments to promote R&D and innovation, and the investment tax incentives database that compiles information on the design and targeting of investment tax incentives available across mainly developing and emerging economies (OECD, 2024[15]) and (Celani, Dressler and Wermelinger, 2022[16]).
All OECD Member countries release a tax expenditure report in some form. Despite this, the quality of reporting can vary. Key relevant information is often not included in reports, such as targeted beneficiaries, legal references, policy objectives, and number of beneficiaries. Furthermore, data granularity levels vary considerably.
Information on beneficiaries can be particularly valuable, as much research has identified the risk that tax expenditures delivered through the personal income tax system can disproportionality benefit high-income recipients. For example, research by Avram (2017[17]) in six European countries finds that tax allowances are generally regressive, as the tax deduction value increases with tax rate. On the other hand, tax credits are often proportional or mildly progressive. The paper notes that tax expenditures are poorly suited to targeting low-income households, as access depends on tax liability, and higher incomes have a larger taxable base, particularly in progressive systems where liability is higher for higher income levels. In 2024, the OECD highlighted that where tax reliefs are provided for private services (i.e. private pension savings or private education), low-income groups that cannot access these services cannot benefit (OECD, 2024[18]). Furthermore, when not well-designed, tax incentives risk supporting investment that would have occurred anyway (OECD, 2026[19]).
2.2.5. Evaluation of tax expenditures
For public resources to be used efficiently, expenditure must be carefully managed. As the cost of tax expenditures can fluctuate, it is important to systematically evaluate their impacts and fiscal costs to ensure sound fiscal management. Evaluation, in this context, means a process that aims to inform policymakers on whether a tax expenditure should be introduced or maintained based on the available quantitative and qualitative data on their effects. It should focus on the stated policy objective of a tax expenditure, and whether it has effectively contributed to the achievement of that policy objective (IMF, 2022[20]). In other words, it should focus on whether a tax expenditure has a causal impact, rather than simply considering the take-up rate (OECD, 2026[19]).
It is useful for evaluations to be based on a framework requiring all tax expenditures to be evaluated on a regular basis. This can help ensure consistency and comparability of evaluations over time. In cases where too many tax expenditures exist to evaluate them all regularly, an alternative is to put in place a prioritisation system to ensure that the larger and more important tax expenditures are evaluated on a more regular basis. It also helps governments build a longer-term picture of tax expenditure costs and impacts, enabling governments to plan around their costs.
Sunset clauses are provisions within tax expenditure legislation that set an automatic expiration date for a tax expenditure. When they are in place, they force policymakers to periodically evaluate whether the provision still warrants its fiscal cost, and make them explicitly justify reasons for keeping them if they so choose. Nevertheless, sunset clauses in and of themselves are not enough to ensure only beneficial tax expenditure are kept in place. Without a comprehensive and effective evaluation framework in place, sunset clause-motivated reviews may be cursory and ineffective.
One form of evaluation is a spending review. These are tools that analyse a government’s existing expenditure, aiming to identify savings and improve the effectiveness and efficiency of existing programmes and policies. These can be useful to counteract the tendency to focus on new and additional spending and thus supporting fiscal sustainability (OECD, 2025[14]). Spending reviews have become increasingly common in OECD Member countries, as evident from Figure 4. In practice, spending reviews are unlikely to be sufficient on their own, and can benefit from separate, systematic ex-post evaluations of the costs and impacts of tax expenditures.
2011-2023
Percentage of OECD Member countries
Note: The number of OECD Member countries has increased over time. The data is for survey respondent OECD Member countries at that point in time. The 2020 dataset refers to all current 38 OECD Member countries. In 2023, 34 OECD Member countries out of 35 survey respondents implemented spending reviews.
Source: OECD 2023 Survey on Spending Reviews, Question 1.
2.2.6. Oversight, scrutiny and audit of tax expenditures
Parliament’s role in scrutinising the budget is crucial for the legitimacy of government spending plans. To be effective, parliaments should have oversight of all important decisions related to public finances, including the development of a fiscal policy framework, economic and fiscal forecasts, and oversight of budget execution. To do this, they should receive budget information in a timely manner, and have adequate analytical support. Independent fiscal institutions also have an important role in enhancing parliamentary oversight, helping cost fiscal measures, highlighting how policy measures impact public finances, and even making recommendations around policy plans (OECD, 2025[14]).
Tax expenditures can be as financially significant, but risk escaping the routine scrutiny applied to budget expenditure. Unless tax expenditures are subject to systematic oversight by parliament and independent fiscal institutions, tax expenditures may proliferate unchecked, or persist long after their original rationale fades. Independent oversight helps evaluate whether each tax expenditure achieves its policy objective in a cost effective manner, and does not create undue distortions or unfair advantages. It also ensures transparency, enabling legislators to see the full scale and impact of tax expenditures, and compare them with direct spending alternatives.
Audit is also an essential institution of public finance. While supreme audit institutions are generally associated with scrutiny of financial transactions, they can also conduct performance audits, where they consider the value for money of specific programmes or types of spending. These can be particularly useful in the context of tax expenditures, as they can help verify whether revenue foregone by the tax expenditure was justified by actual policy outcomes. Furthermore, supreme audit institutions can conduct audits of governance processes or management frameworks. Again, these can be useful for tax expenditures, helping ensure that existing review and evaluation processes accurately capture all relevant measures.
← 1. These figures are based on the Global Tax Expenditures Database (GTED) for all five countries. This database does not cover all tax expenditures, excluding those for which no data is available, those under the value of USD 500 000, and those not published for confidentiality reasons.
← 2. The largest tax expenditures for each country, alongside the monetary amounts, can be found in the country case studies in the annexes.