All forms of expenditure should be on an equal footing in the annual budget process. This includes spending authorised through appropriations or standing legislation, as well as tax expenditure. It also includes the issuance of loans and guarantees, which have become key tools for OECD countries to support policy objectives and mitigate the impact of macroeconomic shocks. These instruments can be harder to manage relative to cash-based appropriations and create contingent liabilities. It requires effective ex ante analysis and approval, transparent reporting and careful management of contingent liabilities and other fiscal risks.
Quality Budget Institutions
Developments in OECD Countries
Abstract
All forms of expenditure include appropriations authorised in the annual budget, expenditure authorised in standing legislation (“entitlements” or “compulsory” expenditure) and expenditure authorised in tax legislation (“tax expenditure”). In addition, governments carry out significant activity through loans, guarantees and other contingent liabilities that may have budgetary implications in the future. These may not impact immediately on government revenue or expenditure in cash terms. Nonetheless, such activities often carry a default risk or low rate of interest, which should be made explicit.
Governments are making increasing use of these various forms of expenditure. Social assistance has been expanding in most OECD countries, with many countries using a mix of direct grants and tax expenditures to deliver support to households. Loans and guarantees have been used for different purposes – including financing higher education – but the importance of such instruments has come to the fore as a tool for crisis response, first during the Global Financial Crisis and then the COVID-19 pandemic (Moretti, 2021[1]).
A key challenge for OECD countries is to ensure that budget frameworks and practices are effective across the different forms of expenditure used to fund public policies, and that ministers are able to choose the best instrument for achieving policy goals. Governments are responding by strengthening budget controls over tax expenditures and over loans and guarantees. They are also reinforcing the management of fiscal risks with a view to the long-term sustainability of public finances.
Governments finance expenditure policies though different means. Historically, public spending was financed through taxation and borrowing and authorised through the annual budget. Today, governments actively use the tax system and their balance sheets to achieve social and economic goals.
7.2.1. Spending authorised through appropriations
Government can progress policies by spending directly, using the funds authorised by parliament in an appropriation act or budget act. This grants the government authority to incur expenditures for a specific purpose over the next financial year. Once the financial year ends, or the appropriation lapses, the government must return to parliament to request a new authority to spend. Appropriations have historically been the basis for spending control and accountability. However, as noted, appropriations are increasingly used alongside other instruments which can receive different treatment in the budget process.
In Canada the Federal budget estimates form the basis for appropriation by parliament. The budget estimates include “voted” expenditure which will be appropriated, and “statutory” expenditure which is authorised through other legislation.
Historically, most expenditure in the Main Estimates was statutory, but the split has become more balanced in recent years. In 2014-15 statutory expenditure was around two-thirds of total spending. By 2023-24 the share had fallen to just over half. Statutory expenditures included transfer payments such as elderly benefits and the Canada Health Transfer.
Some statutory expenditure is excluded from the Main Estimates. For example, Employment Insurance is provided through a contributory social security fund. Some tax expenditures are legislated through the Income Tax Act, such as the Canada Child Benefit, so are not included in the Estimates.
The Public Accounts of Canada offer a consolidated picture of government spending, including Employment Insurance and tax expenditures that are classified in the accounts as an expense – such as Canada Child Benefit. Other tax expenditures are reported through the Federal Report on Tax Expenditure.
Source: Department of Finance, Canada.
7.2.2. Spending authorised through standing legislation
Standing legislation can also create expenditure obligations, including for social security benefits. Such legislation can make up a significant share of government expenditure. On average, OECD cash-based social expenditure rose from 9.3% of GDP in 1980 to 11.4% of GDP in 2019 (OECD, 2023[2]). Pensions and old age benefits make up the largest share and are forecast to grow as populations age.
Expenditure commitments created by standing legislation are not always included in the appropriation or budget act approved by parliament. Canada, for example, excludes some transfer payments (e.g. elderly benefits) and special funds (e.g. employment insurance benefits). Even when entitlements are included in the appropriated expenditure limits, the spending commitments can only be changed by amending the standing legislation that determines the eligibility and value of each benefit.
7.2.3. Spending authorised through tax laws
Rather than spending directly, governments can provide benefits and incentivise behaviour by reducing taxes for specific groups or economic activities. These “tax expenditures” are authorised through tax laws. Examples include reducing value added tax rates on some items; exempting pension contributions from income tax; and crediting businesses for research and development.
The cost of tax expenditure to the government is difficult to estimate and compare across countries. However, tax expenditure was calculated at 2.9% of GDP in France in 2023 and 15% of GDP in the Netherlands in 2022 (van Opstal, 2024[3]; Ecalle, 2024[4]). These costs are often driven by a small number of policies (Box 7.2). However, the proliferation of tax expenditures can add complexity to the tax system.
Tax expenditures are also less constrained by traditional budget controls. New tax expenditures will be announced in the budget with an estimate of how they will impact government revenue, but the total cost of tax expenditures is not usually limited by fiscal rules or expenditure ceilings. Different governance arrangements for tax and spending also creates challenges for co-ordinating tax expenditures and spending programmes that have similar goals (National Audit Office, 2020[5]).
Australia: In 2023/24, the government reported over 300 tax expenditures The estimated costs exceeded AUD 200 billion, roughly equal to one third of total government spending. The largest tax expenditures related to superannuation contributions and earnings (AUD 49 billion); exemptions of residential housing from capital gains tax (AUD 48 billion); and zero and reduced rating of sales tax (AUD 30 billion).
The Netherlands: A total of 119 tax expenditures added up to the equivalent of 41% of government expenditure in 2022. The ten largest items accounted for nearly three quarters of this total; and nearly half of the total was driven by the Earned Income Tax Credit, the General Tax Credit and the treatment of pensions savings
The United Kingdom: In 2024, the government identified 344 tax expenditures, of which 296 had been costed. The three most significant items were: the exemption of pensions contributions from income tax and national insurance; reduced and zero rating of value added tax for food and new dwellings; and the exemption of sales of private residences from capital gains tax.
7.2.4. Government loans and guarantees
Governments issue loans and guarantees to deliver a range of policies, including to support businesses, state enterprises, as well as individuals for instance for home ownership and education. They have also become a principal tool for supporting the people and the economy during economic crises. During the COVID-19 pandemic, for example, the stock of loans and guarantees rose in many OECD countries, in some cases by 4% of GDP or more (Figure 7.1).
The rise of loans and guarantees as policy instruments is not always backed by strong evidence on their effectiveness, particularly outside of an economic crisis. Yet loans and guarantees continue to be attractive to governments, as they can be deployed quickly and support existing ways of assisting businesses and households, such as through access to borrowing facilities (Moretti, 2023[7]).
The cost of loans and guarantees on the public finances is not always immediate. When a government issues a loan, it increases public debt and creates an asset on the government balance sheet, both of which will decline as the loan is repaid. If a loan is not repaid it becomes an expense for the government and is charged to an appropriated budget line, but this may only happen many years after the loan was issued.
Guarantees create a contingent liability for the government. By issuing a loan guarantee, the government promises to pay all or part of a loan if a borrower defaults. Guarantees are often excluded from the budget and the government balance sheet, unless there is a high likelihood that the guarantee will be called. If a guarantee is called, and the government has to meet the terms of that guarantee, the expenditure must be appropriated.
A further challenge with these tools is that, once issued, most loans and guarantees cannot easily be changed. Governments will carry the risk that economic conditions will worsen, which may lead to higher default rates on government loans and more calls on outstanding guarantees. This places particular importance on procedures that help governments decide which loans and guarantees to issue, and on what terms. It also means governments should monitor the stock of loans and guarantees along with other fiscal risks.
Note: Panel A: Data for 31 OECD countries based on OECD statistics on loans for general government (710. Financial balance sheets – consolidated -SNA 2008, Transaction: loans). OECD countries not represented in this graph are those for which loans or GDP data was not available for 2019 or 2021. Panel B: The figure is showing the difference between 2019 government guarantees as a percentage of GDP and 2021 government guarantees as a percentage of GDP. Direct comparisons can only be made between EU countries. For other OECD countries, the data is based on national definitions and methodologies, which may vary significantly from one country to another. The government guarantees data for EU countries refers to general government, data for Australia, Canada, Mexico, Switzerland and the United States refer to federal government and data for New Zealand refer to central government. OECD countries not represented in this figure are those for which loans or GDP data was not available for 2019 or 2021.
Source: Panel A: OECD staff estimates based on OECD National Accounts Statistics (database); Panel B: OECD staff estimates based on government guarantees in national currency: Eurostat for EU Member States, national sources for other OECD countries; for GDP in national currency: OECD National Accounts Statistics (database) on 2 May 2023.
7.2.5. Challenges for budget control
Instruments such as loans, guarantees and tax expenditure provide advantages for meeting the government’s public policy objectives. However, as just outlined, they can reduce the transparency of government spending and the effectiveness of fiscal objectives and rules (Moretti, 2023[7]).
When a grant to households is proposed, and then approved by parliament, the costs to the budget are relatively clear. The cash flow is estimated and included in the appropriations for the next financial year. This is subject to the constraints set by the government’s fiscal objectives and expenditure ceilings. After the close of the financial year, spending is reported in the audited financial statements. Parliament both approves and then reviews expenditure on the policy.
In contrast, tax expenditures, loans and guarantees are harder to measure and compare. Tax expenditures are an estimate of reduced cash inflows. The true costs of a loan or guarantee may be delayed until there is a default or a call is made. The differences in budget treatment, timing and costing methodology can complicate the comparison of policies using different forms of spending. Tax expenditures, loans and guarantees are also less likely to be visible in the budget aggregates which means they are also less constrained by the government’s fiscal objectives and fiscal rules.
Choosing the right expenditure instrument can improve policy. It can help deliver better results or lower the cost of administration, but bringing all forms of spending onto more equal footing in the budget process is not simple. The next two sections look at how countries have strengthened over the management of tax expenditures and loans and guarantees. The approaches vary, but often rely on three common elements:
Better identification, costing and disclosure of activities under different forms of expenditure.
Greater alignment with budget controls and new rules for policy design, monitoring and evaluation.
Improved co-ordination of different forms of expenditure.
Governments are also recognising the challenges of controlling different forms of expenditure in their oversight of fiscal risks, which is discussed in the final section.
Tax expenditures are widely recognised as being easier to introduce, harder to control and less likely to be subjected to scrutiny than direct spending. Most OECD countries have improved controls over tax expenditure by supplying more and better quality information on tax expenditures for ministers and for parliament. While the measurement of tax expenditures precludes direct comparisons with other forms of expenditure, countries made significant progress in disclosure and alignment with the budget process, in strengthening management and evaluation, and improving policy co-ordination and scrutiny.
7.3.1. Estimating the cost of tax expenditure
Estimating the magnitude of tax expenditure is a specialist task and presents some specific challenges. Tax expenditures are usually identified as a deviation from the “benchmark tax system” which can be subject to different interpretations. Tax expenditures are estimates of revenues foregone that can be calculated using different methodologies. These are well-recognised challenges that are regularly disclosed in tax expenditure reports (OECD, 2010[8]).
Any tax system will have numerous reliefs that reduce the amount of tax an individual or business must pay. These are usually classified based on which tax is being reduced – for example an income tax or a tax on goods and services. Tax reliefs can be provided in different ways (OECD, 2010[8]):
allowances: amounts deducted from the benchmark to arrive at the tax base
exemptions: amounts excluded from the tax base
rate relief: a reduced rate of tax applied to a class of taxpayer or taxable transactions
tax deferral: a delay in paying tax, and
credits: amounts deducted from tax liability.
Some tax reliefs are integral to the scope and structure of taxation. They form part of the “benchmark tax system” which describes the normal base, rate and timing for calculating a tax liability. Other reliefs are programmatic because they meet a specific social or economic objective, like incentivising businesses to spend more on research and development.
Definitions of tax expenditure vary, but most relate only to programmatic reliefs. This can require judgements about which reliefs are structural, and which are programmatic, and countries take different approaches to making this distinction. Publishing information on the benchmark tax system and lists of both structural reliefs and tax expenditures can support better understanding and oversight.
Tax expenditures are not directly observed, but can be estimated using different methodologies (Box 7.3). The most common approach is to estimate the revenue foregone as a result of the tax relief, assuming that taxpayer behaviours remain unchanged. However, the choice of method can vary for each tax expenditure.
These estimates provide the basis of reports that supplement the budget reports to parliament. They are, however, not directly comparable to expenditures reported in appropriations. Finance ministries and tax authorities typically note in their tax expenditure reports that:
Estimates of individual tax expenditures do not reflect changes in taxpayer behaviour or the economy that would result if a tax relief was removed.
Total tax expenditures may be calculated by adding up estimates of specific policies, but this does not take into account how a change in one tax might affect another.
Countries use different definitions and methodologies for estimating tax expenditures, which influences the reliability of international comparisons.
In some cases, the government may not produce a cost estimate for some tax expenditures due to insufficient data. In other cases, tax expenditures will be calculated, but not disclosed because the costs are insignificant or because of rules associated with disclosure in official statistics.
Tax expenditures are an estimate of revenues that the government has foregone as a result of a relief. These are estimated for individual tax expenditures, assuming all other taxes remain unchanged, rather than for the tax system as a whole.
The three common approaches for calculating tax expenditures are:
Initial revenue foregone: Estimates the reduction in revenue due to the tax relief, assuming taxpayer behaviour does not change.
Final revenue gain: Estimates the additional revenue the government would collect if the tax relief was removed, factoring in changes in taxpayer behaviour and the effects on other taxes.
Outlay equivalent: Estimates how much direct spending would be needed to achieve the same after-tax benefit for the taxpayer, factoring in how a cash subsidy or transfer would be taxed.
Tax expenditures are usually recorded as a revenue foregone. That means they are reflected within the estimate for the appropriate tax and are reported separately in the budget documents. A notable exception is “non-wastable” tax credits which are paid to an individual, business or household even if the payment exceeds the value of their liability to pay a tax. Child tax credits are designed in this way in some OECD countries and may be recorded as a grant in the expenditure part of the budget.
Source: OECD (2010[8]).
Despite the limitations, the estimates of tax expenditures serve as an important benchmark for policymaking. They offer an indication of the relative costs of different policies and help to identify possible budget savings to finance new initiatives.
7.3.2. Integrating tax expenditure with the budget reporting
Responses to the 2022 OECD Financial Management and Reporting show that 21 out of 34 OECD countries disclose information on tax expenditures either through the budget or in a supporting document. This creates a focus for analysis in government, supports parliamentary oversight and invites scrutiny from supreme audit institutions and independent fiscal councils.
To support the oversight of policy trade-offs, the information on tax expenditures should be presented so that it can be compared to other related spending policies in the budget. Good practices include:
Identifying related policies. Canada’s annual Report on Federal Tax Expenditures includes in its summary of each policy information on the objective, beneficiaries and “other relevant government programmes”.
Aligning tax expenditure forecasts with expenditure baselines and policy costing. Australia’s Tax Expenditures and Insights Statement includes the forecast costs of each tax expenditure for the full multi-year period covered by the Forward Estimates.
Distinguishing between policy changes and forecast changes. The Netherlands published an annex of tax expenditures with the annual Budget Memorandum. This includes projected costs that distinguish between baseline forecasts and policy changes.
Reporting both costs and benefits of tax expenditures. Volume II or France’s Ways and Means report (“Voies et Moyens”) describes and evaluates tax expenditures by budget mission and programme. It includes performance indicators and a programme for evaluating selected tax expenditures.
7.3.3. Strengthening the management of tax expenditure policy
A number of countries have tightened the rules and processes that govern tax expenditures. These typically focus on trying to limit the adoption of new tax expenditures while also evaluating existing policies more systematically.
Only in rare cases have OECD countries established a firm limit or direction on tax expenditures. France, for example, legislated in 2009 that a new tax expenditure be compensated by the ending of an old one. This was updated in 2014 to subject tax expenditures to an annual cap (Moretti and Kraan, 2018[9]). Germany’s Cabinet has previously issued non-binding guidelines that subsidies should be paid as a grant rather than as a tax expenditure (OECD, 2010[8]).
Most countries improve controls by supplying more, and better quality, information on tax expenditures for ministers and for parliament. From this perspective an effective system of tax expenditure management involves (National Audit Office, 2020[5]):
Thorough policy design, underpinned by good evidence, clear objectives and a rigorous appraisal of options for delivering those objectives. In the Netherlands, the Ministry of Finance submits an appraisal to parliament when it proposes a new tax expenditure (or changes to an existing one).
Closely monitoring the costs, benefits and risks of tax expenditures. This includes understanding whether the actual costs and benefits are deviating from the initial expectations. In the United Kingdom, HM Treasury uses an internal monitoring template to keep track of information on tax expenditures.
Evaluating the costs and benefits of tax expenditures systematically. Germany has established a process of formal reviews of tax expenditures, using a standard evaluation framework.
7.3.4. Co-ordinating tax expenditure
The co-ordination of budget arranges for tax expenditures involves the CBA, the tax policy unit within the ministry of finance and the tax administration. Within the finance ministry, policy design is usually the responsibility of a tax policy unit that is separate from the CBA that manages public spending. The effective co-ordination of these functions recognises the expertise of each of the contributing parties, but even with this, there can be differences in accountabilities, IT systems, and reporting requirements.
Some countries have strengthened co-ordination of tax expenditures by designating clear focal points for each tax expenditure (both for policy and delivery) and by creating small central teams that can consolidate information and standardise practices for analysis, reporting and evaluation. Examples include Australia, Canada, and the Netherlands.
The governance of tax expenditures creates challenges for the scrutiny of public spending, as the traditional separation of “spender” (line ministries) and “guardian” (finance ministry) requires an additional step in its co-ordination with the tax administration. The challenges of co-ordination within government increase the risks that tax expenditures are less subject to scrutiny and so result in weaker fiscal control.
The Supreme Audit Institution can play an important role by highlighting areas for improvement. Performance audits have been used in both Canada and the United Kingdom for this purpose, with audits in the latter contributing to improvements in reporting on tax expenditures since the first audit was published in 2014.
Loans and guarantees have become an important policy instrument for addressing market failures and in the response to economic shocks. However, the true costs of a loan or guarantee can be delayed and difficult to calculate. They are often issued outside cash limits and without first securing parliamentary approval. This complicates budget controls, but governments have responded by improving costing and reporting and by tightening the procedures that authorise loans, guarantees and other contingent liabilities (Moretti, 2023[7]).
7.4.1. Estimating the costs of loans and guarantees
A number of OECD countries report the full stock of outstanding loans and the maximum exposure to guarantees. While useful to highlight the scale of policies delivered through these instruments, this should be supplemented with information that is a better measure of the real cost of the policy to government.
As a basic principle, budget decisions should be guided by the full lifetime costs of a policy that will be delivered by issuing loans or guarantees. This should focus on estimating the subsidy involved – a practice known as budgeting “for subsidy cost” or “for grant-equivalence”. These subsidies should be estimated upfront at the time of granting the loan and guarantees, and fully funded when granted within established expenditure ceilings. Key assumptions underpinning the estimates should also be disclosed and explained (Table 7.1).
When choosing a measurement method, governments can use IPSAS as reference. For loans and guarantees classified as “financial instruments”, the government should calculate the net present value of all expected cash flows for the policy. The net benefit or loss represents the cost to the government. For guarantees that are not financial instruments, the costs are related to the risk that the guarantee will be called. Where a programme will issue multiple loans or guarantees, the costs should be estimated for programme as a whole for each year of the relevant budget period (annual appropriation or multi-year spending plans). This approach allows fair comparison with direct spending in the budget.
The final estimates will carry some uncertainty. Costing depends on how the government chooses to evaluate the risks of defaults and calls and the discount rate that is used. Projections of the share of students who will repay their student loans in full, for example, will depend on the prospects for future earnings and changes to the interest rate that will be charged on outstanding debt.
The sensitivity of estimates to different assumptions represents a fiscal risk and highlights the importance of revising estimates routinely with other expenditure baselines. In the United States, government agencies are required to create auditable documentation to explain assumptions used. This facilitates review by the Office of Management and Budget and Congress, as well as the agency’s internal audit team. As another approach, the Office of Budget Responsibility in the United Kingdom has used scenarios to illustrate up-side, central and down-sides estimates for the risks of default rates during times of crisis (Moretti, Braendle and Leroy, 2021[10]).
7.4.2. Maintaining budget controls
One way that governments can impose limits on loans and guarantees is to include the costs within the limits set through budget ceilings and annual appropriations. Another approach is to strengthen the rules, policies and procedures for issuing new loans and guarantees.
As a starting point, countries should clearly specify a taxonomy of loans and guarantees that can be issued by the government. They should also set out the budgetary implications of each main category of loans and guarantees and where this will be recorded. Sweden and the United States have passed legislation that does this (Moretti, 2023[7]).
The United States Federal Credit Reform Act (1990) was introduced as part of an effort to reduce the federal budget deficit. The reform maintained the cash-based appropriation system, but shifted the approval of loans and guarantees onto an accrual basis (Box 7.4). The reform has had a number of benefits, including:
Greater control, by requiring the government to budget up front for the full value of cost of loans and guarantees
Improved allocation, by allowing more direct comparison between loans, guarantees and other spending programmes.
Better monitoring, by generating incentives for agencies to collect additional data on the cost and risks of loans and guarantees.
Governments can also review policies for approving new loans and guarantees, and other contingent liabilities as well. In the Netherlands, for example, the government has adopted a “no unless” policy towards contingent liabilities (Moretti, 2021[1]). No new loans, guarantees or other contingent liabilities will be created unless there is an overriding reason to do so. This applies not just to new policies but also the relaxation of restrictions under existing schemes. The policy stance is supported by a framework which includes:
A ceiling for contingent liabilities, in particular guarantees. This ceiling may be a limit to the new guarantees that can be issued annually or a set ceiling which cannot be exceeded in any year.
An appraisal process involving:
A standardised assessment of a new contingent liability, which is shared with parliament, that includes a question about the appropriateness of the instrument.
An external opinion on the governance of risks and the proposed premium that will be charged to the beneficiary of a guarantee (requested in case of large and complex risks).
A requirement for the government to assess policies involving contingent liabilities no longer than five years after the start of their implementation.
The federal budget in the United States includes provision of loans and loan guarantees. In 2019, the OMB estimated that direct loans by the federal government exceeded USD 1 trillion and loan guarantees exceeded USD 3 trillion.
To calculate the estimated cost, federal agencies were required to estimate the cost of the loans and loan guarantees on a net present value basis, which allows the cost to be compared with other forms of federal spending. The net present value identifies the extent to which the federal government is subsidising the cost of the loans and loan guarantees. The subsidy cost is included in the President’s Budget and federal agencies require an appropriation for that cost before they can issue a loan obligations or loan guarantee. The administrative expenses for the instruments also require an appropriation and are shown separately.
Over a sustained period, OMB has improved the guidance to agencies on modelling and estimating the subsidy cost. The guidance is in Circular A.11 and refers to the Federal Credit Reform Act. When economic conditions require, OMB has also issued short-term guidance to provide flexibility under the Act when preparing the estimates, as was the case in 2021 to calculate the subsidy from the Pay-check Protection Program.
Source: Anderson and Burke (2021[11]).
7.4.3. Tailoring reporting for decision-making
Ex ante budget decisions guided by more complete cost estimates can be reinforced by improvements in the reporting of loans and guarantees as part of the budget process (Moretti, Braendle and Leroy, 2021[10]).
The adoption of accrual accounting standards for financial reporting in most OECD countries means that more information is published on the costs of loans and guarantees . As an indication of this shift, the 2022 OECD Survey on Financial Management and Reporting shows that 31 out of 34 OECD countries provide information on contingent liabilities in the year-end financial report. This reporting supports budget oversight and provides a safeguard for the government against making poor decisions (Moretti and Youngberry, 2018[12]).
However, government and parliament can still make more effective use of information from accrual accounting reforms. That requires regular analysis and disclosure of balance sheet information, for example to assess fiscal risks or to compare the expected and actual costs of policies financed through loans and guarantees. The Balance Sheet Review Report in the United Kingdom, for example, published analysis by HM Treasury of the Whole of Government Accounts, which influenced decisions to introduce new controls and more centralised monitoring of guarantees and other contingent liabilities.
Other good practices for tailoring reporting the reporting of loans and guarantees to support decision-making and oversight include:
Publishing monthly or quarterly reports on the flows of new loans and guarantees and associated costs.
Revising the estimated costs of loans and guarantees at least semi-annually with updates to the economic assumptions used for budgeting.
Providing more contextual information in the budget documents, including whether changes in the value of financial assets and liabilities is due to government decisions or changes in the economy.
7.4.4. Co-ordinating loans and guarantees
CBAs should take an active role in co-ordinating the budget management of loans and guarantees across government:
Providing policy advice, including by bringing budgetary issues with loans and guarantees to the attention of policymakers.
Setting and monitoring good budgeting practices, including standards for costing, appraisal, accounting and data collection.
Supporting line ministries in implementing good budget practices, either by reviewing the analysis of line ministries or by centralising appraisal and monitoring within the finance ministry.
Where countries decide to create actual or notional accounts in relation to contingent liabilities, the finance ministry plays a role recording and monitoring the cash flows in and out of these accounts. This may be a function given to the budget office or to the treasury function within the ministry of finance (Moretti, 2023[7]).
Establishing strong central capacity to produce or verify costings helps manage the risks associated with loans and guarantees. This requires a range of different skills and expertise including actuaries, credit risk experts and legal expertise.
Central teams may be located in the ministry of finance or shared with other ministries or agencies. For example, the Australian Department of Finance has a team that supports line ministries as a complement to the tools and guidance available on the department’s website. In contrast, the United Kingdom created a multidisciplinary advisory unit in United Kingdom Government Investments (a state-owned enterprise under the HM Treasury).
Fiscal risks are “potential deviations from the government fiscal forecasts” (Moretti, 2021[1]). They reflect uncertainties which may impact on revenues, expenditures, the fiscal balance or the government’s assets and liabilities. Government loans and guarantees are a source of fiscal risk, due to the uncertainty over how much of a loan will be repaid and whether a guarantee will be called. However, countries must contend with risks from many other sources that range from litigation costs to the potential of future global macroeconomic shocks.
Governments should monitor and manage their portfolio of loans and guarantees, as well as fiscal risks that do not meet the recognition criteria of a contingent liability, with a view to the long-term sustainability of public finances. This requires a strong risk management framework that helps the government to identify relevant fiscal risks and take action before they materialise – either by putting in place measures that will contain the risk, or by building sufficient buffers for the government to intervene appropriately.
Following early examples, including by Australia and New Zealand, many OECD countries established a systematic approach to fiscal risk management after 2008. By 2023, the OECD SBO Survey on Budget Frameworks found that 13 out of 36 OECD countries reported on fiscal risks as part of the budget process.
7.5.1. Strategies and institutional arrangements for managing risks
The purpose of a fiscal risk framework is to promote the resilience of budgetary plans and to mitigate the potential impact of fiscal risks, and thereby promoting stable and sustainable public finances.
Governments employ a combination of three broad strategies for managing fiscal risks:
intervene to mitigate specific risks (e.g. by imposing new spending controls)
provision for the costs of fiscal risks within budget ceilings (e.g. through a budget reserve), or
accommodate risks by adopting more prudent fiscal policy (e.g. by building fiscal buffers).
Which strategy is used will vary depending on the nature of the risks that have been identified and prioritised. Where possible, governments should take steps to prevent risks from arising or minimise their impact if they do. However, even with an effective mitigation strategy, risks will not be fully eliminated and governments must accept a degree of risk associated with most policies and investments. A contingency fund or other form of “margin” can set aside resources to cover risks within the approved budget ceilings, should they crystalise. This is unlikely to be sufficient to respond to larger shocks which raises questions about the appropriate fiscal policy course to adopt in the budget.
A key challenge for fiscal risk management in the government is to centralise enough information to guide overall fiscal policy, while acknowledging that the capacity to identify risks, and the authority to manage them, will be spread across the public sector. This means it is important for the finance ministry to establish clear roles, responsibilities and procedures to embed risk management across ministries, agencies and other public entities. This may require collaboration with the centre of government which will consider policy and operational risks for the government.
In most countries, the finance ministry will maintain a register of significant fiscal risks that will be analysed and monitored centrally by the government. This information is used to advise ministers, factor risks into budget planning and report to parliament. The finance ministry will also provide guidance and tools that standardise the approach to defining, categorising and managing fiscal risks. However, responsibilities for identifying and managing fiscal risks are almost always delegated to the organisations with the best knowledge of different risks. Figure 7.2 illustrates these arrangements.
The finance ministry will usually lead on identifying and managing macroeconomic risks, while line ministries and agencies will be responsible for risks associated with their activities. In rare cases, the ministry of finance will share this role with an IFI that has been given responsibilities for preparing official economic forecasts. In the Netherlands, for instance, this is focused mostly on the macroeconomic forecasts and related risks.
In most other countries, the IFI supports the oversight of fiscal risks by parliament, as part of the routine scrutiny of the government’s forecasts and fiscal plans. This independent perspective contributes to parliamentary and public debate, and can raise the profile of risks associated with specific policies as well as longer-term risks that governments may have fewer incentives to address in the short run.
The opportunities for parliament to review fiscal forecasts and associated risks have generally been growing. Alongside the usual approval of the budget, parliaments debate updates of the fiscal forecasts in a pre-budget or mid-term report and in long-term fiscal sustainability reports. Improvements in the coverage of year-end financial statements supports the ex post review of contingent liabilities and other risks. In some countries, including Australia and the United Kingdom, sectoral committees can also challenge individual ministries and agencies on how they are managing risks.
Officials have noted that strong parliamentary scrutiny can strengthen the culture of risk management in government. SAIs can support improvements in the governance of fiscal risks through a combination of financial and performance audits.
Source: Moretti (2021[1]).
7.5.2. Identifying and classifying fiscal risks
OECD countries use fiscal risk frameworks to identify and classify risks in order to make decisions on how to manage them. These classifications help to identify risks that are most likely to undermine fiscal controls, and to communicate these risks to parliament and the public (Moretti, 2021[1]).
The most common approach is to categorise risks based on where they arise. This includes:
Macroeconomic risks that arise from changes in the international or national economy, such as changes in the economic cycle.
Policy risks that are linked to the design, implementation or cost control of a specific government policy, for example, there could be uncertainty over the take-up of a new energy subsidy or a risk that costs will change if energy prices rise.
Uncertain budgetary claims (indemnities and contingent liabilities) including the risk that government guarantees will be called.
Balance sheet risks that may change the value of government assets or liabilities, such as the future costs of public sector pensions.
Some countries also categorise risks based on their likely impact by estimating the probability that a risk will crystalise and the impact this would have on the public finances in the future. This encourages governments to look beyond immediate threats to the budget and focus attention on risks that may arise in the future. It also helps decide which risks will be monitored centrally.
Risks may also be classified to clarify responsibility. Australia presents fiscal risks by department and agency to establish who is accountable for managing the various risks. Some countries, including the United Kingdom, have considered the degree of control that the government has to manage different risks – for example by differentiating risks over which the government has no control or some varying degree of control.
A growing trend is to identify both explicit and implicit risks, where the government may be socially or politically obliged to make if a particular event occurs. Finland’s Overview of Central Government Risks and Liabilities has analysed possible implicit liabilities to local government, state-owned enterprises, natural disasters and the banking sector These can be difficult to identify and governments are naturally weary of moral hazard and creating false expectations of when it will intervene in a crisis. However, major financial crises and public health emergencies have shown the importance of recognising and planning for implicit risks.
7.5.3. Measuring and analysing fiscal risks
Fiscal risks should be measured and analysed, ideally in quantitative terms to support a comparison to the budget forecasts and estimates. Data from the 2023 OECD SBO Survey on Budget Frameworks shows that most OECD countries identify and monitor similar types of risks, but also that these risks are more likely to be identified than measured. This partly reflects the nature of the risks themselves, but there remains scope in many countries to make fiscal risk management more comprehensive and systematic.
The most common risks identified by OECD countries are uncertain budgetary claims from guarantees (29 out of 36 countries, 81%); debt risks (26 out of 36 countries, 72%); and balance sheet risks from public sector pensions (25 out of 36 countries, 69%). These are also the risks that are most likely to be measured, reflecting the availability of data and well-established methodologies that finance ministries can use to quantify impact.
Over half (58%) of OECD countries also identify risks from the financial sector. This reflects a growing focus on larger and less certain shocks which are more likely to have a material impact on the government’s fiscal strategy. However, these can also be harder to measure. While 19 OECD countries identify geopolitical risks for instance (53%), just four have quantified the likely impact on the public finances (11%). Where risks cannot be measured reliably, governments like Australia and New Zealand prepare narrative information to guide decisionmakers and communicate risks to the public.
Notably, 13 OECD countries (36%) identify risks from state-owned enterprises and only 11 countries (31%_ identify risks from subnational governments, public private partnerships or balance sheet risks (e.g. changes to the valuation of government loans).
Share of countries for different types of risk, 2023
Note: Data for Lithuania and Mexico are not available. Other fiscal risks include risks as a result of demographic changes, growth in export demand and oil and gas prices.
Source: OECD (2023), Senior Budget Officials Survey on Budget Frameworks, Question 33.
The growing interest in large, uncertain risks after the Global Financial Crisis was initially through the use of “stress tests”. These recognised that shocks impact on the economy and public finances through different channels, such as, a recession that lowers profits, employment and tax revenue, and can result in higher calls on government guarantees.
Stress tests have been widely used by ministries of finance, independent fiscal institutions and central banks. For instance, the New Zealand Treasury deployed stress tests on a routine basis in the Investment Statement published every four years. Other countries including Finland, the Netherlands, and the United Kingdom have also employed stress tests to support planning for future macroeconomic shocks (Table 7.2).
Stress tests rely on being able to design relevant scenarios. New Zealand drew its scenarios from existing risk registers. The 2018 scenarios covered: a severe earthquake; an agricultural disease outbreak; and an international economic downturn. However, COVID-19 took many countries by surprise and raised questions about whether stress tests were sufficient. More generally, it seems that countries have elaborated upon stress tests by adding scenarios to the fiscal forecasts.
7.5.4. Reporting fiscal risks
Information on fiscal risks should be published routinely alongside the budget and updates to the government’s economic and fiscal forecasts. This gives parliament and the public a better understanding of the government’s fiscal plans and the uncertainties that surround its official forecasts.
Fiscal risks are usually summarised in the budget documents themselves. Australia and New Zealand include analyses of macroeconomic risks alongside the summary of macroeconomic and fiscal forecasts. Other fiscal risks, such as those presented by outstanding guarantees, are disclosed in other chapters or annexes. The assessment of fiscal risks may be published in a separate report (e.g. Chile) or integrated into budget papers (e.g. Australia and Ireland). Other OECD countries publish stand-alone reports. In Finland this is published by the ministry of finance. In the United Kingdom, reporting is shared between the Office for Budget Responsibility and HM Treasury.
While governments should aim to disclose their full assessment of fiscal risks, some exemptions are needed. For example, litigation against the government that is currently before the courts, or information that is likely to prejudice substantial economic or defence interests of the country. Regardless of the exemptions, a basic principle is for countries to adopt a consistent and transparent basis for not disclosing fiscal risks, and for that basis to be reviewed on a regularised basis.
References
[11] Anderson, B. and K. Burke (2021), “Budgeting for loans and guarantees: The United States Federal Credit Reform Act”, OECD Journal on Budgeting, Vol. 2021/3, https://www.oecd.org/content/dam/oecd/en/publications/reports/2021/12/oecd-journal-on-budgeting-volume-2021-issue-3_62a3ab64/7ef61932-en.pdf (accessed on 27 January 2025).
[4] Ecalle, F. (2024), Tax expenditures country report: France, German Institute of Development and Sustainability (IDOS) and Council on Economic Policies (CEP), https://doi.org/10.23661/cr4.2024.
[6] HM Revenue & Customs (2024), Non-structural tax relief statistics (December 2024), https://www.gov.uk/government/statistics/main-tax-expenditures-and-structural-reliefs/non-structural-tax-relief-statistics-december-2024#large_reliefs.
[7] Moretti, D. (2023), “Rise of non-standard policy Instruments: Can budgeting catch up?”, OECD Journal on Budgeting, Vol. 2023/3, https://www.oecd.org/en/publications/oecd-journal-on-budgeting-volume-2023-issue-3_ec1d2b4e-en.html.
[1] Moretti, D. (2021), “Managing fiscal risks: Lessons from case studies of selected OECD countries”, OECD Journal on Budgeting, Vol. 2021/1, pp. 81-128, https://www.oecd.org/content/dam/oecd/en/publications/reports/2021/04/oecd-journal-on-budgeting-volume-2021-issue-1_3cbbf46d/a4b2ab8d-en.pdf.
[10] Moretti, D., T. Braendle and A. Leroy (2021), “Balanced sheet based policies in COVID-19: How to improve transparency and risk analysis”, OECD Journal on Budgeting, Vol. 2021/2, https://www.oecd.org/content/dam/oecd/en/publications/reports/2021/11/oecd-journal-on-budgeting-volume-2021-issue-2_3e4b0986/132dd5df-en.pdf.
[9] Moretti, D. and D. Kraan (2018), “Budgeting in France”, OECD Journal on Budgeting, Vol. 2018/2, https://doi.org/10.1787/budget-v18-2-en.
[12] Moretti, D. and T. Youngberry (2018), “Getting added value out of accruals reforms”, OECD Journal on Budgeting, Vol. 2018/1, https://doi.org/10.1787/budget-v18-1-en.
[5] National Audit Office (2020), The management of tax expenditures, https://www.nao.org.uk/reports/the-management-of-tax-expenditures/ (accessed on 13 January 2025).
[2] OECD (2023), OECD Social Expenditure Database (SOCX), https://data-explorer.oecd.org/s/33z (accessed on 16 November 2023).
[8] OECD (2010), Tax Expenditures in OECD Countries, OECD Publishing, Paris, https://doi.org/10.1787/9789264076907-en.
[3] van Opstal, R. (2024), Tax expenditures country report: The Netherlands, https://doi.org/10.23661/cr1.2024.