A service of the

Download article as PDF

This article is part of The Future of EU Public Finances

EU financing can in principle draw from two sources – from “own resources” as well as from “other revenues”.1 No binding definition exists as to what kind of resources can qualify as own resources. With no substantive limitation, own resources can draw from various sources. In the past, member state contributions were the predominant sources, while EU taxes or borrowing are increasingly taking centre stage. This article addresses both – taxes and borrowing – and emphasises the legal leeway and limitations on using these financial resources for the EU budget.

The claims put forward in this contribution relate to the suitability of taxes and borrowing to finance the EU budget. First, not only does the EU have limited taxation power and there is no taxation power falling into its competence for public finance purposes; most tax proposals currently envisaged as own resources create “unreal” tax revenues on the basis of statistical values which cover the fact that they are nothing other than ordinary member states’ budget contributions running under the fake title of a tax. Second, repeating off-budget EU borrowing akin to NextGenerationEU (NGEU) is generally possible but faces the constraint that the space for additional EU borrowing is limited until NGEU repayments have brought “other revenue” back to magnitudes that are only marginal in relation to the amount of “own resources”. Third, EU borrowing on-budget for the EU budget would be unprecedented but possible, though with severe limitations, in particular associated with the requirement that all debt service having to do with EU borrowing must be backed, by legal requirement, by unconditional non-borrowed own resources.

Real versus unreal EU taxes

With respect to taxes, a distinction must be made between taxes identified as “own resources” in the Own Resource Decision (ORD) and taxes that are actually implemented (at the EU level or at the member state level). This distinction is important because the “own resources” (the current as well as probably much more future ORDs) draw from “imaginary”, statistical-based taxes that oblige member states to pay the EU a virtual tax from their national budgets without this tax actually being implemented. In order to avoid misunderstandings, one should therefore distinguish between real und unreal taxes.

By real taxes, I refer to taxation power that is backed by an actual EU competence to tax or by a specific member state tax which is then passed on to the EU budget. There is indeed leeway for the EU to implement environmental taxes2 as well as energy-related taxes.3 But these tax regimes are not taxes that can be raised for public finance purposes – they are tied to environmental or energy objectives. In other words, these taxes cannot be raised in order to finance the EU budget – this function can only be a side effect of the primary policy-specific purpose of these taxes. The same applies to the EU’s powers related to the internal market: the EU is allowed to harmonise taxes if this is necessary for the establishment and functioning of the internal market or required in order to eliminate distortions of competition in the internal market. Again, this cannot serve the EU to pursue public finance purposes. It is only permitted if harmonisation primarily pursues this objective.

“Unreal taxes” can be referred to as those identified as “own resources” – statistical-based revenues that determine the amount that member states must transfer to the EU budget. These sources are not necessarily levied in practice. Take the “plastic tax”, which is currently an “own resource” to the 2021-2027 EU budget – it is a statistical-based tax which many member states do not implement. The tax revenue of the plastic tax is hypothetically computed and member states pay this national contribution to the EU from their domestic budgets. The same applies to the various other tax revenues that have been in the policy debate, such as a new corporate tax based on operations and levied on companies.

The tax policy debate on “own resource” is thus a “ghost debate” in a certain way – it introduces imaginary taxes for which the EU has no compentece and which the EU cannot oblige member states to implement. The invention of statistical-based taxes for the purpose of generating “own resources” is even misleading to the extent that they engender tax increases other than the one for the “own resources”. This is so because a statistical-based own resource implies an increased transfer in the national contributions to the EU budget, for which each member state has to find financial cover. Practically, this implies that member states must consider domestic tax increases to ensure the transfer to Brussels (unless they are able to cut other expenditures).

Overall, the tax revenue debate for “own resources” should acknowledge that we are not talking about genuine EU taxes, nor do we necessarily talk about taxes that are actually implemented. The Union is not able to raise taxes for the purpose of budget financing (hence to finance the EU budget). The few tax powers the EU has are confined to their purpose to deal with sectoral policy objectives (i.e. climate, energy). Drawing from these sources for the purpose of “own resources” must remain a side benefit of sectoral taxation.

Repeating NGEU

From taxes, we turn to debt as a funding source for the EU budget. Two possible avenues for debt financing of EU public goods can be distinguished. First, debt financing can be used for the purpose of repeating a temporary, “one-off” and “off-budget” fund like NGEU that was set up to borrow (and spend) for a specific purpose. Second, we look into the EU engaging in borrowing in order to fund the regular EU budget, hence creating a permanent, “on-budget” debt-financing capacity. While this avenue has been used on various occasions in the past (basically for back-to-back lending operations), allowing debt as an own resource would be a major innovation under EU budget practice. This has occassionally been employed on a small scale by exploiting the budgetary headroom or margin under the EU budget, although only featuring a back-to-back funding mechanism (the European Financial Stability Mechanism is the most important example).

NGEU was built on unprecedented legal architecture that engaged the issuance of bonds with a quasi-mutualiasing effect, which had previously been ruled out given its distributive nature. There is no general barrier to adopting an NGEU-type approach for the purposes of financing specific future expenditures of the EU. This would require an amended ORD, which would authorise borrowing up to a maximum amount and for a specific purpose, and adjust the own resources ceiling to ensure that borrowing can be repaid.

However, repeating NGEU meets at least two limitations. The first barrier results from the Treaty’s expectation that own resources are the primary source for financing the EU. As mentioned above, there are in general two sources – “own resources” and “other revenues”.4 NGEU was introduced as “other revenue” off-budget and as externally assigned revenue into the EU finances. The primacy of “own resources” as the main sources of revenues would be challenged if a large and increasing portion of EU expenditure were to be financed off-budget via “other revenues”, including borrowing, rather than “own resources”. Put differently, “other revenues” must be small in relation to the own resources. A budgetary framework in which off-budget financing in the form of other revenue exceeds the financing from own resources would not comply with this requirement of the Treaty – no matter whether the economic purpose supports off-budget expenditure. Given the sizeable magnitude of off-budget NGEU resources, the expectation of the legal requirements is that other revenues will decline to a fraction of own resources until NGEU is repaid entirely in 2058. Against this background, while repeating NGEU is generally possible, doing so in the near future would significantly reduce the permissible amount of off-budget borrowing (given the still existing amounts of NGEU funds).

The legal requirement of “other revenues” to be only a fraction of the budget has been articulated by the German Constitutional Court. Some are tempted to argue that it is only the European Court of Justice (ECJ) that gives authoritative interpretations of EU law (which is formally correct), and thus one should ignore the interpretation of a domestic court. This attitude – which seems popular among those fed up with a German court that constantly opposed the various anti-crisis measures adopted in the Economic and Monetary Union in the past decade – would be disregarding the political repercussions. A German government that is bound to both the rulings of the ECJ as well as the domestic constitutional court would find itself in an extremely precarious situation, and the spillover effects for Europe are certain to be negative. As long as the ECJ itself has not ruled on certain Treaty provisions (such as the relationship between “other revenues” and “own resources”), there is (political) wisdom in paying attention to the concern expressed by national constitutional courts. Consequently, if the NGEU model were to be replicated to finance public goods in the coming years – as proposed, for example, by the European Central Bank in the form of an EU Climate Fund – the quantitative limit becomes binding, leaving only limited space for debt financing programmes for the budgetary period.

The second point of contention is the “exceptional” character of NGEU. The “exceptionality” and the “temporary” character of NGEU were explicity stipulated in the current ORD. In the case of NGEU, the exceptional character was built on solidarity to respond to the uneven effects of the COVID-19 shock on member states (which legally translated into the use of the infamous solidarity clause of Article 122, which gives member states much leeway and circumvents the European Parliament). One question that lawyers have been discussing is whether repeating NGEU would again be linked to a solidarity situation such as the pandemic. However, there are convincing reasons to distinguish between the revenue and the spending side.

The revenue side is secured through the ORD – creating “off-budget” external revenues (within the quantitative limitations mentioned above), which require unanimity in the Council and even member states’ ratification (i.e. through many national parliaments). For the expenditure side, it is not strictly necessary to limit a possible NGEU-successor to a solidarity-scenario akin to NGEU. This is in line with the previous borrowing practice of the EU: in borrowing for back-to-back lending for member states, the EU used a plethora of different justifications in addition to solidarity and emergency scenarios. Clearly, the EU is not entirely free to choose how it intends to use the revenues that it borrows. It must strictly apply with the legal core “principle of conferral”, which allows the EU to act only where it has a legal basis in the EU Treaties. There is a number of possible policy fields where the expenditure could be used to attain policy objectives – for example, in the area of cohesion policy,5 for environmental purposes,6 for trans-national infrastructure,7 or for trans-European research.8 Programmes pursuing objectives of cohesion akin to macroeconomic programmes addressing cross-border cooperation may be considered more generally as the climate emergency or environmental spending programmes.

Debt financing as an own resource?

Thinking one step further means considering an unprecedented move: allowing debt financing to be integrated as revenue into the general EU budget rather than borrowing funds for specific purposes as off-budget “other revenue”. While borrowing under the EU budget is not a new practice, scholarship and jurisprudence are divided on whether the EU may finance its general budget with debt. There are good arguments to consider EU borrowing for the general budget to be compatible with the legal requirements, but there are also legal risks associated with it (just the same way as many European institutional innovations such as building an European Stability Mechanism or setting up NGEU came with residual legal risk). These risks can be mitigated by a restrictive practice of allowing borrowing.

Specifically, the Treaties neither deny nor explicitly empower the EU to finance its budget with debt. While the ORD and the EU Financial Regulation reflect the preferences of the EU legislators at the time of their adoption, it is undisputed that the Treaty does not contain an absolute prohibition against raising debt. The EU would need to add a new category of own resources in the ORD that allows borrowing. Also, there are in principle no quantitative limits on borrowing, but two major limitations impede the use of debt proceeds as own resources.

First, the EU must have adequate means to meet its debt service in any year, which must be secured by a sufficient amount of (non-borrowed) own resources. This flows from the Treaty-based balanced budget requirement. To that end, given that borrowed money does not become own resources indefinitely, there must be a safeguard to ensure the repayment of the debt. Thus, there is a need for a counterbalancing asset in order to ensure such a “irrevocable, definitive and enforceable guarantee of payment” (Council Legal Services, 2020) provided by the member states. What matters is budget neutrality – the resulting debt must be matched by a claim allowing the Union to cover the debt service. This must be ensured through definite, non-borrowed own resources – the EU must, for example, increase the amount of the GNI-ceiling in order to guarantee a balanced budget every year.

Second, the ORD, which requires ratification by all EU countries, must specify the permissible amount of borrowing. When the proceeds of debt financing become a new category of own resource, there is no other way than regulating the amount that will be issued in the ORD. The upfront specification is necessary for two reasons: in order to determine the precise amount of guarantee that is necessary to back “borrowed own resources”, and in order to satisfy domestic (e.g. German) requirements emphasising that any financial transfer from a domestic to the EU budget must be ex ante foreseeable and quantifiable.

Borrowing on the regular budget rather than off-budget can build on several further advantages. The European Parliament is directly involved as co-legislator und must approve the EU budget – “on-budget” constructions thus enjoy greater legitimacy than “off-budget” solutions. On-budget solutions are fully transparent and subject to oversight by the European Court of Auditors. Finally, not only EU level legitimacy would be ensured through the European Parliament, but also national parliaments would remain in full control of the EU’s revenue from borrowing operations via the ORD. With the Commission tied to the ex ante defined borrowing in the ORD, member states have full foresight of the risk that they subscribe to with the budget.

Economists are fond of revolving debt, and the question here is whether outstanding EU debt may be refinanced by issuing new EU debt. Under NGEU, the EU is not allowed to roll over debt, with the legal authorisation only empowering the raising of debt for the specific purpose described in the ORD. Whether this would be possible with respect to borrowing proceeds that are categorised as own resources is less clear and poses difficult legal questions. However, with the maximum possible borrowing specified ex ante in the ORD, revolving debt appears possible. What matters from an EU primary law perspective is that member states create sufficient non-borrowed own resources to repay the liabilities.

Finally, what can debt-financed “own resources” be spent on? Different considerations apply for on-budget EU debt than in the case of funding off-budget debt. Unlike off-budget debt, there is no strict requirement for earmarking expenditures. Rather, the budgetary universality and non-assignment rule applies, which means that revenues shall be used without distinction to finance all expenditure entered in the Union’s annual budget. With this core budgetary principle, EU-borrowed funds can generally be spent on any budgetary item, provided that the expenditure is in line with an existing EU competence (as it is required for all EU expenditure irrespective of the funding type). However, one could generally consider an earmarking of on-budget debt-financed expenditure. This could be a sensible option in view of the German Constitutional Court’s reservation to acknowledge that EU debt can finance the general EU budget. In order to accommodate the restrictive perspective, the EU budgetary lawmakers would need to lift the universality principle in order to allow for an earmarking of EU debt.


The policy debate on own resources does not lack creativity in identifying possible financial sources. However, the identification of tax instruments seems particularly misleading, because of its insufficient distinction between “unreal” and “real” taxes, with the former determining hypothetical statistical-based taxes for which the EU has no authority to collect, nor can the EU require member states to implement these taxes. These taxes imply a tax collecting and public finance power that does not exist. In fact, the EU has very limited taxing power, and no authority to tax for public finance purposes. Rather than creating new “unreal” taxes, the debate should focus on which genuine taxing powers the EU should gain for public finance purposes. That goes beyond singular sectoral taxing competences such as in energy and climate, and makes Treaty changes indispensable.

In turn, the debate can benefit from more creativity with respect to debt-financing the EU – this contribution highlighted leeway and limitation of replicating NGEU and debt-financing the regular EU budget. Repeating NGEU for other purposes requires an amendment to the ORD to borrow other revenue (external assigned revenue) and create an off-budget item. Unlike for the pioneering NGEU, a replication would face significant size restrictions. “Other revenues” must be marginal compared to “own resources” in order to comply with the EU legal framework, which makes a repetition of this instrument in the near future unlikely because NGEU debt must converge towards marginality in relation to own resources. Any future NGEU-like fund must likewise demonstrate it is a one-off and temporary measure.

Debt-financing the regular budget is not per se prohibited. Clearly, a new ORD would have to be adopted, with the legitimacy enhancing requirements accorded through unanimity and national ratification, and with the involvement of the European Parliament, unlike under NGEU-like off-budget solutions. However, the economic potential of borrowing for the EU budget would be severely impaired by the limitation that all debt service arising from the borrowing must be backed by non-borrowed own resources (e.g. through an increased GNI-ceiling like under NGEU) as well as by the predefined maximum amount of borrowing.

Spending is subject to less constraints than funding. Certainly, repeating NGEU would need to comply with the exceptional and temporary character of off-budget constructs and using the borrowing exclusively for predetermined purposes is indispensable. There is more flexibility under on-budget debt. All EU expenditure must comply with EU primary law, which suffices as a limitation to expenditure. Alternatively, if politically desired and in order to address remaining legal concerns, the earmarking of borrowed debt to certain on-budget EU expenditure is feasible.

* Some elements of this article build on Grund and Steinbach (2023).

  • 1 Article 311 of the Treaty on the Functioning of the European Union (TFEU).
  • 2 Article 192 para. 2 subpara. 1 a) TFEU.
  • 3 Article 194 para. 3 TFEU.
  • 4 Article 311 TFEU.
  • 5 Article 175 TFEU.
  • 6 Article 192 TFEU.
  • 7 Article 171 TFEU.
  • 8 Articles 179 and 173(3) TFEU.


Council Legal Services (2020, 24 June), Opinion of the Legal Service on Proposals on Next Generation EU, 9062/20, https://data.consilium.europa.eu/doc/document/ST-9062-2020-INIT/en/pdf (30 November 2023).

Grund, S. and A. Steinbach (2023), European Union debt financing: leeway and barriers from a legal perspective, Bruegel Working Paper, 15/2023.

Treaty on the Functioning of the European Union (2012), Official Journal of the European Union, C 326, 47-390.

© The Author(s) 2023

Open Access: This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/).

Open Access funding provided by ZBW – Leibniz Information Centre for Economics.

DOI: 10.2478/ie-2023-0064