Background

The multiannual financial framework (MFF) for the period 2028-2034, worth almost €1.8 trillion in commitments (2025 prices),¹ envisages a complete overhaul of the current MFF architecture. It aims to make 'the EU budget and its programmes simpler, more flexible and strategically attuned to the priorities of today'. The proposed new MFF structure is meant to reduce fragmentation, complexity, overlaps and the lack of flexibility in the EU's current funding instruments.

Source: M. Pari and S. Pradier, EU budget 2028-2034: Overview of the Commission's proposal, EPRS, European Parliament, 2025; Graphic by Lucille Killmayer.

Announced in the 2025 Commission work programme (CWP), the Commission released its proposals for the post-2027 MFF in two batches, on 16 July and 3 September 2025. The accompanying communication, 'The road to the next multiannual financial framework', announced that the MFF would be both simpler and 'more impactful'. In the new MFF architecture, the Commission proposes grouping programmes under four headings (see Figure 1).

The European Parliament, in its resolution of 7 May 2025 on a revamped long-term budget, backed the idea of simplification as such. However, it called for more transparency, accountability and scrutiny, control of spending, and for reducing the risks of double funding, misuse and fraud. Regarding better law-making, Parliament regretted that the Commission had in the past 'not systematically conducted thorough impact assessments … for all legislation involving spending through the budget', insisting on a change of this practice.

Coverage of the MFF-related impact assessments

The Commission's preparatory work for the 2028-2034 MFF includes seven impact assessments. These support, either directly or indirectly, 18 MFF programme proposals and the proposed horizontal framework for budget expenditure tracking and performance (see Table 1).² Three cases can be distinguished:

Further programme proposals under the 2028-2034 MFF were presented entirely without an IA. Instead, they were accompanied by a staff working document taking the form of an ex-ante evaluation. This concerns the proposal, under the ordinary legislative procedure, on the Pericles V programme (i.e. the EU's programme for the protection of the euro against counterfeiting);⁴ the three MFF proposals pertaining to the Euratom Treaty;⁵ and the proposed Council decision on the association of the overseas countries and territories with the EU. The choice of conducting an ex‑ante evaluation (rather than an IA) is in line with the requirements set out by Article 34 of the EU's Financial Regulation and Tool #9 of the BR Toolbox, according to which spending programmes with a 'relatively small budget' that provide continuity in terms of scope and structure do not require an impact assessment but rather an ex-ante evaluation. Meant to support the preparation of the programmes in question, and drawing on ex-post evaluations of previous programmes, ex-ante evaluations analyse the issues to be addressed, the EU added value, objectives, and the expected effects of different options and monitoring and evaluation arrangements.

Methodological approach – EU better regulation framework

Based on the Interinstitutional Agreement on better law-making (IIA-BLM), the Commission's Better Regulation Guidelines (BRG) and Toolbox lay down the rules for carrying out IAs, which are considered an essential tool for high-quality EU law-making. IAs are designed to facilitate transparent political decision-making by explaining the reasoning behind a legislative proposal and providing a solid evidence base for the proposed measures.

The above-mentioned seven IAs relating to the proposals for the 2028-2034 MFF programmes do not follow the standard EU IA approach. They all acknowledge from the start a deviation from the scope and standard IA methods as defined by the Commission's BRG; this affects their quality in terms of better regulation and transparency to varying extents.

The IAs justify the deviation from the better regulation framework by invoking Tool #9 of the BRG relating to the preparation of spending programmes, which recognises that 'the special case of preparing a new multiannual financial framework is a unique process'. This admission constitutes progress in terms of transparency compared with the IAs accompanying the 2018 proposals for the MFF 2021-2027,⁶ which also differed significantly in form and substance from the standard BRG requirements, but did not provide any explanation in this respect. The 2025 MFF IAs also show progress in terms of substance, as the 2018 assessments focused on presenting the expected positive effects of the MFF proposals rather than assessing and comparing relevant impacts of alternative policy options.

However, Tool #9 does not detail what the 'specific approach as regards scope and depth of analysis' means in practice for MFF IAs. At the same time, it does not say either that the better regulation principles would not apply. This lack of specific guidance suggests that the Commission has discretion over the implementation of the better regulation principles when preparing an IA for an MFF proposal, which makes it difficult to formally assess those IAs' quality. The application of better regulation standards in the 2028-2034 MFF IAs varies considerably – however, they have some important common features.

First, a particularity of the 2028-2034 MFF IAs is their common focus on 'how to streamline the architecture of the EU budget'. Instead of developing objectives based on a specific problem definition, as would be the case in a standard IA, their approach is driven by the horizontal objectives of Commission President von der Leyen's political guidelines and the communication on the next MFF (reduce complexity and administrative burden and increase the MFF's flexibility). Moreover, as most of the IAs point out, 'policy aspects are considered in the analysis of the context, the problem definition and the objectives', but are not, or only sporadically, discussed in the sections on the options and their impacts, which usually are key parts of any IA and address the specific subject matter of the relevant legislative proposal in detail.

Second, the IAs state that the proposed fundamental changes to the MFF architecture and the uncertainty about the future budget allocations would not allow for the development of scenarios. This is surprising, as IAs relating to spending programmes usually make assumptions in order to build alternative scenarios (typically assuming no change, higher and lower budget allocations) and to enable a comparative assessment of the scenarios' potential impacts.⁷ In addition, the 2028‑2034 MFF IAs argue that the impossibility of developing scenarios also prevents quantified analysis, resulting in the IAs being mainly qualitative. However, Tool #9 does not exclude quantifications, and some of the MFF IAs do indeed complement their qualitative assessment with (limited) quantifications, whereas others provide no quantification at all (see section on data quality and methods below).

Altogether, the absence of clear rules for MFF IAs, the varied application of the better regulation standards (and tools) in the seven IAs for the 2028-2034 MFF, and the horizontal approach to essential parts of these IAs make it challenging to compare their quality. In the following, this briefing pinpoints common (or similar) features identified across the seven IAs, providing an overview of their strengths and weaknesses.

Problem definition

An evidence-based and coherent problem definition is of particular importance for an IA's quality, as it lays the groundwork for a pertinent intervention logic and the ensuing impact analysis: if the problem is poorly defined, the resulting definition of objectives and options – and their assessment – risk being flawed, as well.

In many of the sampled IAs, the problem definition seems to lack coherence, clarity and, for some of them, also sufficient and transparent substantiation. The IAs provide general lines of problematic elements in the current situation (baseline), covering a vast range of policy areas. However, they neither address policy details nor the precise nature of the problems in relation to the main affected stakeholder groups, as requested by the BRG. Moreover, the links between the problems, the underlying drivers and the consequences of the problems are not systematically spelled out, with problems and drivers occasionally overlapping, and little information provided on their respective consequences. In addition, most of the IAs do not indicate whether all problems and drivers are equally relevant or weigh differently (and if so, how).

Notwithstanding these weaknesses, the description of the nature of the problems seems to be overall satisfactory in most MFF IAs. However, the frequently observed lack of sufficient evidence, combined with the limited quantification, affects the analysis of the problems' magnitude, which is addressed either vaguely or not at all. This exacerbates the above-mentioned weaknesses, because the overall relevance of the problems remains vague and with it, the intervention logic of the IAs (e.g. the IAs on Global Europe and on the Single Market and Customs Programme). It seems that, in most cases, the problem analysis was performed with a specific policy option in mind - a process known as 'backward engineering'. This process is explicitly rejected by the BRG (Tool #13), with the Commission cautioning that 'this not only undermines the quality of the analysis, but it also hurts the credibility of the whole impact assessment process'. Finally, the question of whether and how the problems are expected to persist (or aggravate) without the proposed interventions could have been explored more thoroughly in most of the sampled IAs.

The problem definition in most MFF IAs is mainly based on stakeholder feedback and ex-post or interim evaluations of the programmes they cover, in line with the BRG's 'evaluate first' principle. Some IAs draw on a wide range of sources, pertinent reports and studies, while others would have benefited from a larger amount of external input and a more consistent use of data already featured in the evaluations, in particular quantified information.

Subsidiarity / proportionality

Most of the MFF IAs maintain the legal bases of the numerous existing programmes they propose to merge. The wide range of programmes and policy areas covered by each IA is reflected in a high number and variation of legal bases in most IAs. As noted, contrary to the BRG, most of the MFF IAs do not assess and/or substantiate sufficiently whether and how the identified problems would persist in the absence of (new) EU action, although a thorough analysis of a dynamic baseline scenario would have been important to explain the need for and added value of EU action. Several IAs refer to evaluations of previous programmes, some of which are not (yet) publicly available, meaning that the assertions cannot be checked.

Most IAs explain the need for EU action in the policy areas they encompass by citing the priorities of the political guidelines, i.e. to be more flexible, effective and efficient amid rapid changes and newly emerging internal and external challenges, in particular regarding cross-border activities (e.g. in the customs union, in cohesion policy and in civil protection). They state that better cooperation and coordination would entail the added value of limiting inefficiencies and administrative burdens, for instance by reducing suboptimal investments (IA on the Competitiveness Fund) or by limiting regional disparities within and between Member States (IA on the NRPPs).

Overall, it appears that subsidiarity analysis could have been more thoroughly substantiated in most of the IAs. None of them includes a subsidiarity grid, which could have supported and guided a more exhaustive analysis, as recommended by the BRG for 'politically sensitive and important proposals' (only the IA on a budget expenditure-tracking and performance framework falls under exclusive EU competence, and therefore no subsidiarity assessment was required in that case). In addition, most of the IAs do not assess proportionality, neither in the context of subsidiarity nor in the assessment of the options. The IAs that do discuss it provide limited qualitative reflections in the comparison of the options. Only the IA on AgoraEU and Erasmus+ features a rather comprehensive proportionality analysis.

In the context of their subsidiarity check, national parliaments issued three reasoned opinions on three of the total 19 legislative proposals, in accordance with Article 6 of Protocol No 2 on the application of the principles of subsidiarity and proportionality (see Table 2).

In addition, national parliaments raised concerns in the context of the political dialogue. Contributions were received on each of the 19 proposals – 73 in total. They came from the Spanish Cortes Generales (19), the German Bundesrat (12), the Czech Senate (10), the Portuguese Parliament (9), the Romanian Senate (7), the Italian Chamber of Representatives (5), the Irish Houses of Oireachtas (5), the Italian Senate (3), the Polish Senate (2), and the Dutch Senate (1). The highest number of contributions – seven each – was submitted on the proposals regarding the national and regional partnership plans and the Common Fisheries Policy .

Objectives of the initiative

Unlike standard IAs applying the better regulation criteria (where the objectives of an initiative are to be developed in the context of an intervention logic, from an evidence-based problem definition, to corresponding objectives, to specific policy options), the MFF IAs' overarching objectives are based on the common political goals of all MFF proposals: to reduce complexity and administrative burden of EU funding programmes, and to increase the MFF's flexibility. This affects the definition of the objectives in the seven MFF IAs.

The definition of the general objectives' quality varies between these IAs. In some cases, they are Treaty-based and directly linked to the problem and/or problem drivers, as requested by the BRG, while in others, they are more loosely related and diverse. In all cases, there is an overlap between general and specific objectives, often owing to a lack of specificity of the latter. This is contrary to the BRG, according to which an IA's general, specific and operational objectives should be clearly distinct and increasingly specific. In the same vein, none of the IAs presents tangible operational objectives as to what should be achieved by whom, how and when. Some of the IAs provide other measurable deliverables (e.g. the IA on the UCPM) or examples on how the achievement of objectives could be measured (the IA on the AgoraEU and Erasmus+ programmes). All in all, most of the IAs' objectives appear to only partially fulfil the better regulation 'SMART' criteria (according to which objectives should be specific, measurable, achievable, relevant and time-bound).⁸ This lack of clarity and specificity in the objectives' definition risks hampering the future monitoring and measuring of their achievements (see also the section on monitoring below).

As regards the United Nations (UN) Sustainable Development Goals (SDGs), most MFF IAs state briefly, either in the text or in Annex 3, that their preferred option would help achieve several SDGs in the policy areas concerned. Some briefly describe how they would support these SDGs (e.g. the IA on Agora EU and Erasmus+), while others do not corroborate this assertion.

Range of options considered

According to the BR Toolbox (Tool #16: How to identify policy options), an IA first needs to construct a baseline, i.e. a no-intervention scenario against which different policy options (and their impacts) are to be compared. It should then develop 'a wide variety of policy options', and at least two realistic alternative policy options in addition to the baseline to address the problems, their drivers and the identified objectives. The options are to be described 'in reasonable detail', to allow for an in-depth analysis of the expected impacts. This description should include the different policy measures they comprise and an indication on how, and by whom, the options would be implemented, monitored and enforced. In addition, Tool #16 explicitly states that policy options that 'are built around each other' should be avoided.

All seven IAs in question present a baseline scenario, albeit to a varying degree of analytical depth. In most of the seven MFF IAs, the range of options is limited to two or three policy options, which corresponds to the BRG by a narrow margin. Solely the IA on the NRPPs displays more complexity and thus considers more options. Some IAs present options that differ only in the extent to which they integrate current funding programmes and instruments (e.g. Single Market and Customs Programme) and are thus, strictly speaking, not self-standing alternative options.

Shortcomings in the examined IAs' policy option sections vary from one IA to another, with many featuring the following limitations:

Moreover, in some cases, it appears questionable whether the described policy options present fully realistic and/or self-standing alternatives to the baseline, and/or whether the IA identifies the full range of options to address the problem drivers.

On discarded policy options, the BRG require that policy options initially considered but discarded at an early stage of the IA process need to be explained. While most MFF IAs are overall transparent about discarded options, the description of the discarded options and the justification for discarding them could be more detailed in some cases.

Assessment of impacts

Once the retained policy options are identified, an IA should, according to the BRG, 'carry out a robust assessment of their economic, social and environmental impacts and of who will be affected and to what degree'. Furthermore, the IA should inform policymakers to what extent the identified policy options meet the set objectives, 'with what benefits, at what cost, with what implications for different stakeholders, and at what risk of unintended consequences'. The selection of a preferred option should be based on a thorough assessment of the retained options against the criteria effectiveness, efficiency and coherence.

Although the assessment of impacts is generally a core section of every IA, the depth and comprehensiveness of that section appears uneven in the seven IAs supporting the MFF proposals. While it seems well-substantiated in some IAs (e.g. Competitiveness Fund), it is superficial in others. For example, in the case of the UCPM and the Single Market and Customs Programme, the IA discusses economic, social and environmental impacts cursorily at best. These limitations appear to be linked to the above-noted approach of the assessments rather being horizontal than policy-specific, and to the lack of specificity in the definition of the policy options.

In all sampled IAs, regardless of their analytical depth, the assessment of impacts remains predominantly qualitative, with quantifications largely lacking. Even cost-benefit analyses are mostly qualitative, and - in addition - the effects of cost and benefits are not systematically assessed relative to the different stakeholder groups (as requested by the BRG). Remarkably, although the IAs in question relate to EU spending programmes, they lack an analysis of budgetary impacts, since, as noted above, the Commission chose not to include funding scenarios. This political choice appears at odds with the associated legislative proposals, which naturally operate with budgetary ceilings, allocating an indicative financial envelope to each of the proposed programmes for the 2028-2034 funding period.

Consideration of further impacts (beyond economic, social and environmental) remains scarce in the sampled IAs, although some IAs touch on fundamental/human rights impacts ⁹ (NRPPs, Global Europe, AgoraEU and Erasmus+); impacts on gender equality (horizontal performance framework, AgoraEU and Erasmus+); digitalisation (e.g. Competitiveness Fund, AgoraEU and Erasmus+); and democracy and equality (AgoraEU and Erasmus+).

Some of the EPRS briefings analysing the MFF IAs (see list in endnote 1) noted a lack of assessment of impacts that would have been relevant for a specific initiative. For instance, assessing the territorial impacts would have been appropriate in the IA on the NRPPs, given their obvious regional dimension. Similarly, the IA on the Single Market and Customs Programme could have addressed impacts on third countries, fundamental rights and consumer protection.

All sampled IAs contain an analysis of how the policy options compare in terms of effectiveness, efficiency and coherence, albeit with some variations. While the efficiency criterion remains unaddressed in the IA on the NRPPs, further criteria are added in other IAs. In particular, the IA on Global Europe also compares the options against 'economic, social and environmental sustainability'; the IA on the UCPM adds 'simplification, synergies, and the "ability to act"'; and the IA on AgoraEU and Erasmus+ also performs a brief proportionality assessment.

Monitoring and evaluation

The above-noted weaknesses in defining distinct, clear and specific objectives also affect the IAs' provisions for monitoring and evaluation. In line with the BRG, IAs should analyse monitoring arrangements, including systematic data collection, and identify pertinent indicators to measure the success of an initiative under the IA's preferred option (usually in relation to its specific or operational objectives). Given that the MFF IAs do not present operational objectives, and that their specific objectives often overlap with general objectives, a precise definition of targeted and measurable indicators appears challenging to achieve and is, indeed, missing in nearly all IAs.

Most of the MFF IAs do not present any monitoring indicators. Instead, they all point to the IA on the performance framework, which provides for an implementation report of each programme but does not discuss specific indicators for the other MFF files. Therefore, it remains unclear how the achievement of the MFF initiatives' objectives would be measured, which data would be collected, and who would be involved in the process. Not even the IA on the performance framework presents specific monitoring indicators in relation to its own objectives; instead, it proposes tracking two cross-cutting aspects over time (in the framework of a study looking at the relevance/fitness of the intervention field and list of indicators, and at the administrative costs of implementing performance provisions), without exploring further details. The IA on the AgoraEU and Erasmus+ programmes does not provide monitoring indicators either, but presents examples on how success of the initiative could be measured (in Annex 10). The IA on the UCPM appears to be the only MFF IA with a more detailed analysis of monitoring arrangements. It includes a list of suggested headline indicators that specify clear targets in line with the performance framework for the post-2027 budget (policy area (level 1): crisis management), and also describes data collection methods.

Finally, most of the MFF IAs announce that an ex-post evaluation will be carried out in accordance with the BRG,¹² without however indicating a timeline or other specifics in this respect.

Stakeholder consultation

As required by the BRG, all MFF IAs present a synopsis report of the consultation activities undertaken for the respective initiatives in a dedicated annex. The synopsis report informs about the methodology (i.e. consultation strategy), the stakeholder groups and the consultation results.

For each of the seven initiatives, the Commission carried out a questionnaire-based open public consultation running from 12 February 2025 until 7 May 2025, thus respecting the mandatory 12‑week period. The number of valid responses received ranged from 555 replies for the questionnaire on the performance framework, to 5 845 replies for the education, training, culture and civil society programmes. Beyond the replies to the questionnaires, the public consultations also yielded a significant number of position papers, going into the hundreds.

For each file, the synopsis report presents a breakdown of stakeholders by category. The extent and clarity to which stakeholders' views are described and systematically referenced throughout the IAs vary. The same holds for the substantiation of stakeholder support for the preferred option. Overall, consultation activities appear to have been heavily centred on the open public consultations. A call for evidence was launched for none of the initiatives.

While significant targeted consultation activities have reportedly taken place for some initiatives (e.g. Competitiveness Fund), none are mentioned for the others. Over half of the synopsis reports include a reference to the European Citizens' Panel on the new EU long-term budget, held in Brussels in May 2025. This format gave 150 randomly selected citizens from across all 27 EU Member States the opportunity to share their views.

Finally, some of the IAs report on the use of artificial intelligence (AI) for analysing the responses and position papers to the open public consultations. Relying on a hybrid approach, the Commission combined Large Language Model (LLM)-driven topic modelling and expert human validation.

Supporting data and analytical methods used

Most of the 2028-2034 MFF IAs are based on stakeholder consultations and ex-post or interim evaluations of the programmes they cover, in line with the BRG 'evaluate first' principle. Apart from this commonality, the IAs differ considerably in terms of quality and transparency of the evidence base. While some are largely based on a wealth of transparently referenced and linked data sources, reports and studies (the IAs on AgoraEU/Erasmus+, civil protection and the Competitiveness Fund), others show more limited substantiation and/or poor referencing of sources (the IAs on the NRPPs, Global Europe, the performance framework and the Single Market and Customs Programme – the latter with consistent lack of precise referencing and systematic absence of links). Transparency is also affected where the frequently cited evaluations had not yet been finalised at the time of drafting the IAs; and were therefore not publicly available for scrutiny. Moreover, even in IAs with more solid substantiation overall, a lack thereof was noted in important parts (e.g. in the comparison of the options and the selection of the preferred option in the IA on the UCPM).

All IAs acknowledge certain limitations and point to a lack of data (which is not always specified). Furthermore, the large absence of quantification entails limitations in key sections of the IAs, notably the problem definition and the assessment of potential impacts. Nonetheless, while some IAs do not feature any quantification (including in the 'one in, one out' approach to administrative burden,¹³ as is the case e.g. in the IA on the Single Market and Customs Programme), others do provide relevant – albeit limited – quantified estimates; for instance, estimates of economic impacts in the IA on the Competitiveness Fund and estimated administrative costs in the IAs on the NRPPs and the performance framework – whereby in the latter, 2021-2027 data was extrapolated to the entire EU budget, given that 2028-2034 allocations were unknown at the time of drafting the IAs. Another example suggesting that most IAs could have done more in terms of quantified information is the IA on the AgoraEU and Erasmus+ programmes: without carrying out own quantifications, it uses at least some quantified estimates from relevant evaluations (e.g. for impacts on SMEs in the audiovisual sector) to substantiate the analysis.

More could also have been done in terms of transparency regarding information on methods and assumptions underpinning the analysis. Some MFF IAs offer comprehensive and specific information, including on sensitivity analyses carried out to improve the IA's robustness (e.g. the IAs on the performance framework and on the AgoraEU and Erasmus+ programmes), whereas others do not provide any, or little, information in this respect.

Finally, three IAs mention the use of modelling – the IAs on the Competitiveness Fund and on the NRPPs refer to the RHOMOLO model, which is featured in the Commission's Modelling Inventory and Knowledge Management system (Midas), as is the SOCRATES model, referenced in the IA on the Agora EU and Erasmus+ programmes.

Follow-up to Commission Regulatory Scrutiny Board opinion

The Regulatory Scrutiny Board (RSB) issued profoundly critical opinions on the draft versions of each MFF‑related IA. 'Exceptionally', the Board refrained from giving a qualification to these opinions, instead of the usual 'positive', 'positive with reservations' or 'negative', as set out in BRG Tool #3. This is a novelty.

While acknowledging the applicability of Tool #9 for the preparation of new MFF proposals, the Board found that the draft IAs lacked 'several key elements', and noted 'significant shortcomings' across most sections of the IAs. In conformity with Tools #3 and #8, all IAs in question provide – albeit to a varying degree of depth – an explanation of how the RSB's recommendations were addressed. It appears that, despite some efforts on the Commission's side, the Board's comments were merely partially taken into account, with considerable shortcomings remaining unresolved in the revised (i.e. final) IAs.

Finally, Tool #3 requires the Board's opinion to be published once the related proposal(s) have been adopted by the College of Commissioners. This is indeed the case for six of the IAs in question. Only the RSB opinion on the IA on the NRPPs is not retrievable in the register of Commission documents; it is solely linked in the file-specific record on the 'Have your say' portal.

Coherence between the Commission's proposals and IAs

The legislative proposals appear to be largely coherent with the preferred option identified in each of the IAs in question – with one difference: as stated above, none of the IAs analyse budgetary impacts, arguing that it was impossible to develop funding scenarios. By contrast, the associated MFF proposals allocate expenditure ceilings to the different programmes.

This briefing analyses whether the principal criteria laid down in the Commission's own Better Regulation Guidelines appear to be met by the IAs accompanying the MFF proposals for the 2028-2034 period. It does not attempt to deal with the substance of the respective proposals.