VoxEU Column Taxation

Tax expenditures: The hidden side of government spending

Tax expenditures – such as tax incentives for firms, tax deductions for households, and lower tax rates for specific goods and services – are both widespread and costly, but are generally not subject to the same level of scrutiny in the budget process as direct spending. This column introduces the Global Tax Expenditures Database, which aims to improve reporting and enhance scrutiny. As governments worldwide face growing funding needs to respond to the pandemic, they cannot afford to lose revenues to ill-designed tax breaks. Comprehensively assessing tax expenditures is crucial, and estimating and reporting their fiscal cost is a vital first step.

Tax expenditures are generally not subject to the same level of scrutiny in the budget process as direct spending, and are hardly ever assessed in terms of their costs and benefits (CBO 2012, CRS 2019). Given their fiscal relevance, it is time to change this.

Tax expenditure reporting enhances transparency in fiscal management (Heady and Mansour 2019) and is needed to assess whether those measures are value for money, to inform the political and public debate, and thus ultimately to drive evidence-based tax expenditure reform. 

Significant fiscal cost

Tax expenditures come in many shapes and forms such as tax incentives for firms, tax deductions for households, and lower tax rates for specific goods and services. They are not only widespread, but also costly.

According to the newly released Global Tax Expenditures Database (GTED),1 revenue forgone through tax expenditures has been relatively stable over time, with global averages varying between 3.3% and 4.9% of GDP during the last 30 years. 

Figure 1 Global unweighted average revenue forgone and number of reports


Source: Global Tax Expenditures Database (
Note: The number of reports for 2018, 2019 and 2020 are likely to be incomplete because some countries report tax expenditure data with a certain time lag.

Beyond these averages, there is a large variation across economies. Many countries report significantly higher estimates of revenue forgone (Figure 2). 

Figure 2 Countries with revenue forgone-to-GDP ratios higher than 6%


Source: Global Tax Expenditures Database (
Note: Based on latest available year.

The Russian Federation has the largest TE-to-GDP ratio (almost 14.5%). In the US, income-related tax expenditures reduce federal government revenue by more than 1.4 trillion dollars a year. That is almost 7% of GDP and roughly one third of federal government spending. Similarly, the revenue forgone through tax expenditures accounts for more than 8% of GDP in Australia, more than 6% in Canada, and more than 7% in Japan and the UK. The average fiscal cost of tax expenditures across the EU is quite close to the worldwide average, lying around 4% of GDP. In some cases, though, it can be significantly higher, exceeding 12% of GDP in the cases of Finland and the Netherlands. 

Tax expenditures in developing and emerging economies are also significant. For instance, they amount to more than 4% of GDP in Brazil and South Africa, come close to 8% in Colombia and Mauritania, and exceed 10% in Jordan. 

Striking lack of transparency

Although significant in themselves, the figures reported above are likely to underestimate the full dimension of tax expenditures  for a few reasons. 

First, despite their magnitude and the fact that they have similar effects on public budgets as direct spending, the use of tax expenditures is characterized by a striking lack of transparency and accountability (Burman and Phaup 2011). In fact, as we found in building the GTED, only 97 out of 218 jurisdictions have reported official revenue forgone estimates at least once since 1990. 

Second, the quality, regularity, and scope of those reports vary significantly. Some countries, such as Australia and Morocco, publish comprehensive documents, providing not only revenue forgone estimates at the level of individual tax expenditures but also information on tax base, type of tax expenditure, policy objectives, and beneficiaries. 

In countries such as Germany, Canada, France, Ukraine and Italy, the information for some provisions even includes the number of taxpayers benefitting from a specific tax expenditure. In many other cases, however, governments report only a limited number of aggregate estimates. Portugal and Costa Rica, for instance, only provide overall estimates aggregated by tax base. Crucially, as shown in Figure 3, more than half of the roughly 21,000 tax expenditure provisions gathered by the GTED do not have any information regarding the policy objective they are supposed to serve. 

Figure 3 Number of provisions and revenue forgone by policy objective


Source: Global Tax Expenditures Database (
Note: Number of provisions are counted over the total time period. For revenue forgone estimates, we first calculate country averages across years, and then the global average across countries.

Moreover, not every government issues tax expenditure reports on a regular basis. Some countries have produced only one or very few reports during the last decades. For example, Switzerland has reported only one comprehensive tax expenditure report (in 2011), Senegal has published only two reports since 2014, and Turkey three reports since 2007. 

Finally, apart from the overall quality and regularity of tax expenditure reports, it is difficult to judge how close the data provided are to being representative, since governments often provide information on a subset of existing provisions only. In the US, for instance, the  official report includes revenue forgone estimates at the provision level, but only for income-related tax expenditure provisions.  

In general terms, richer countries tend to report more information on tax expenditures. However, there is heterogeneity. Morocco and India are examples of developing and emerging countries that publish relatively comprehensive tax expenditure reports on a regular basis. They provide not only revenue forgone estimates at the provision level, but also certain metadata on the tax base, policy objective, and target groups of each tax expenditure provision. 

Across the 46 G20 and OECD economies, two (China and Saudi Arabia) do not publish any official tax expenditure information, and 11 do not report provision-level data, providing only aggregate estimates. Likewise, among the 27 EU member states, three (Croatia, Cyprus and Malta) do not report on tax expenditures at all, and ten others only provide limited information. Both groups of countries breach the Council Directive 2011/85/EU on requirements for budgetary frameworks, which explicitly states that “Member States shall publish detailed information on the impact of tax expenditures on revenues” (European Union 2011: p. L 306/47).

The situation is even bleaker when it comes to low- and lower-middle-income countries, where in many cases tax expenditure reporting is still in its infancy. This is due to a variety of reasons, including data constraints, insufficient human and financial resources and weaker institutional frameworks (Kassim and Mansour 2018). Of the 79 low and lower-middle income economies, 45 do not report on tax expenditures and 8 countries report aggregate estimates only.

Tax expenditure reform

Existing evidence suggests that tax expenditures often end-up missing their stated policy goals and having negative side effects, such as exacerbating inequality (Redonda and de la Feria 2020) and contributing to climate change. The regressive impact of the preferential rates on long-term capital gains and qualified dividends (Toder and Baneman 2012) and of tax benefits for pension savings (Redonda and Axelson 2021), as well as the environmental impact of fossil fuel subsides – two thirds of which are granted as tax benefits (OECD 2021) – are cases in point.

If well designed, though, tax expenditures can attain stated goals. For instance, some tax benefits have proven to be effective in supporting the unemployed to get back into the labour force and support the worse-off. The Earned Income Tax Credit – one of the largest tax/transfer policies in the US, implemented as a refundable tax credit that benefits millions of low- to middle-income families – has increased mothers’ work hours without triggering adverse effects on children (Bastian and Lochner 2021).

With heightened urgency in view of the challenges from COVID-19, the reform of tax expenditure is vital, particularly in those countries that are being hit harder by the crisis (Redonda and Gupta 2021). The magnitude of the long-term social and economic costs of the pandemic are still uncertain. Yet, it is clear that the 3.5% output contraction observed in 2020 is unprecedented, and only comparable to the one triggered by WWII (Levy Yeyati and Filippini 2021, IMF 2021). 

As governments worldwide face growing funding needs to respond to the pandemic, they cannot afford to lose revenues to ill-designed tax breaks. They cannot afford to continue offering tax deductions that disproportionally benefit higher-income households. And they should not maintain tax incentives that harm the environment and thus the sustainability of their societies. To build back better, reforming tax expenditures is critical. Comprehensively assessing tax expenditures is crucial. Estimating and reporting their fiscal cost is a vital first step.

Overall, improving tax expenditure systems can play a crucial role to increase domestic revenue mobilisation, finance the Sustainable Development Goals and align tax systems with governments’ policy objectives (Redonda and Neubig 2018, von Haldenwang et al. 2021).


Bastian, J and L Lochner (2020), "The Earned Income Tax Credit: Increasing mothers’ work hours without adverse effects on children",, 23 January.

Burman, L and M Phaup (2011), “Tax Expenditures: The Big Government Behind the Curtain”,, 17 November.

CBO – Congressional Budget Office (2012), “Tax Expenditures Have a Major Impact on the Federal Budget”.

CRS – US Congressional Research Service (2019), “Spending and Tax Expenditures: Distinctions and Major Programs”.

European Council (2011), “Council Directive 2011/85/EU of 8 November 2011 on Requirements for Budgetary Frameworks of the Member States”.

Heady, C and M Mansour (2019), “Tax Expenditures and Their Use in Fiscal Management: A Guide to Developing Countries”, IMF How to Note 19/01.

IMF (2021), “After-Effects of the Covid-19 Pandemic: Prospects for Medium-Term Economic Damage”, in World Economic Outlook, October 2021.

Kassim, L and M Mansour (2018), “Tax Expenditures Reporting in Developing Countries: An Evaluation”, Revenue d’Économie du Développement 2018/2.

Levy Yeyati, E and F Filippini (2021), “Pandemic divergence: The social and economic costs of Covid-19”,, 12 May.

OECD (2021), “OECD Companion to the Inventory of Support Measures for Fossil Fuels 2021”.

De la Feria, R and A Redonda (2020), “Tackling Inequality Through Tax Expenditure Reform”, Think 20 (T20).

Redonda, A and C Axelson (2021), “Assessing Pension-Related Tax Expenditures in South Africa. Evidence from the 2016 Retirement Reform”, WIDER Working Paper 54/2021, UNU-WIDER.

Redonda, A and S Gupta (2021), “COVID-19 and Seizing the Opportunity for Reforming Tax Expenditures in Africa”, Council on Economic Policies.

Redonda, A and T Neubig (2018), “Domestic Revenue Mobilization and Tax Expenditures in Developing Countries”, Council on Economic Policies.

Redonda, A, C von Haldenwang and F Aliu (2021), “The Global Tax Expenditures Database (GTED)”, Companion Paper.

Toder, E and D Baneman (2012), “Distributional Effects of Individual Income Tax Expenditures: An Update”, Research Report, Urban-Brookings Tax Policy Center.

von Haldenwang, C, A Redonda and F Aliu. (2021), Shedding Light on Worldwide Tax Expenditures, GTED Flagship Report 2021. 


[1] The GTED is a joint initiative led by the Council on Economic Policies (CEP) and the German Development Institute (DIE). It gathers the official and publicly available tax expenditure data worldwide in an online database, to improve reporting, enhance scrutiny, and, ultimately, contribute to the design of effective and fair tax expenditure systems across the world. For more details, see Redonda et al. (2021).

4,094 Reads