Introduction
Despite the relatively modest scale of the EU budget compared to national European budgets, there is a growing recognition of the important role EU funding plays in shaping and steering policy decisions (De Witte Reference De Witte2023; Schramm and Terranova Reference Schramm and Terranova2024; Van Wolleghem Reference Van Wolleghem2019). This influence is also visible in contentious fields such as migration and asylum, where EU investments, such as through the EU Emergency Trust Fund for Africa (EUTF) launched in 2015 or the Asylum, Migration and Integration Fund (AMIF), may mediate polarised national policies within Europe (ECRE and UNHCR 2022; Zardo Reference Zardo2022) but also affect those beyond the EU borders (Sayed Reference Sayed2025; Tittel-Mosser Reference Tittel-Mosser2018). Reflecting the expansion of its policy agenda, the European Union has scaled up its financial commitments to migration and asylum. The Multiannual Financial Framework (MFF) for 2021–2027, the European Union’s tool for long-term budgetary planning, introduced a chapter – the Heading 4 – dedicated to migration and border management.
As the EU’s reliance on funding instruments has deepened, academic interest has followed suit. Scholars have increasingly focused their attention on the impact of EU funding on migration and asylum (Davitti and La Chimia Reference Davitti and La Chimia2017; den Hertog Reference den Hertog2016; Tsourdi and Zardo Reference Tsourdi and Zardo2025; Zardo Reference Zardo2022). Nevertheless, for all this growing momentum, the empirical evidence and sources used in these studies are neither comparable nor comprehensive across time and scope. In light of this, researchers (UNHCR and ECRE 2018), advocacy groups and civil organisations continue to call for efforts to ‘follow the money’ to understand better how financial resources are allocated, how funding programmes are conceived and negotiated, both within EU institutions and in cooperation with non-EU countries and ultimately, how these funds are spent.
The MigFund datasetFootnote 1 addresses this data shortage by providing an unprecedented empirical resource, tracing EU financial allocations for migration and asylum from 2000 to 2023. Encompassing around 1,700 funding measures across various funding programmes, MigFund captures both the internal and external dimensions of EU migration governance. Its detailed categorisation system allows researchers to systematically assess not only the financial magnitude of these measures but also their policy orientations, whether restrictive or open toward migration. Crucially, MigFund operates on the understanding that policy instruments, including funding allocations, are neither neutral tools nor technical decisions or reflections of the policy level alone (Lascoumes and Le Gales Reference Lascoumes and Le Gales2007). Rather, they reflect strategic and normative choices, mirroring how actors and bureaucrats in charge of designing and managing these financial tools conceptualise and address migration and asylum.
The article proceeds as follows. The first section highlights the importance of tracing and studying EU funds related to migration and asylum. The second section describes the dataset’s structure and the coding strategy to code funding programmes along a restrictive-open trajectory. The third section describes the evolution and nature of EU external and internal funds and reflects on their implications for our understanding of migration governance. In the final section, the article discusses the relevance of the MigFund dataset and possible applications.
The importance of studying European Union funding for migration and asylum
Unlike binding directives or laws, funding operates as a non-regulatory tool, a form of ‘governance by money’ that steers behaviour through resource allocation. This instrumentality means that funding brings with it values, political choices, and institutional dynamics throughout the entire policy implementation process (Capano and Lippi Reference Capano and Lippi2017). The social construction of tools (Lascoumes and Le Gales Reference Lascoumes and Le Gales2007) occurs at all stages of the design process, and a focus on instruments provides insight into the broader policy process, where tools may even diverge from decisions made at the policy level. In migration policy, for instance, policymakers might strategically allocate funding to limit immigration into Europe. Yet, the ultimate design of funding programmes can either strengthen or weaken this overarching goal. This is because the process of designing funding programmes is separate from – though interacting with – policy-making, granting these instruments a degree of autonomy. It is a highly bureaucratic exercise ‘where institutional battles converge and path dependency is high’ (Di Ciommo and Jones Reference Di Ciommo and Jones2019, 13), leading to programmes that do not necessarily align with the dominant policy narratives or the emphasis on specific policy aspects. Even when policy objectives and funding measures converge, we may expect variations in the pace at which change unfolds, according to an incrementalist logic often observed in budgetary decisions (Citi Reference Citi2015).
Over the past two decades, the European Union and countries in the so-called Global North, such as the United States and Australia, have significantly increased the amount allocated to govern migration, within and outside their borders (Tsourdi and Zardo Reference Tsourdi and Zardo2025). Between 2015 and 2017 alone, the EU’s budget directly allocated for migration management increased by EUR 10.1 billion, totalling EUR 17.6 billion. The EU budget for 2021–2027, established for the first time, includes a specific heading for migration and raises the planned expenditures to more than EUR 30 billion. Besides the budgetary increase, the EU also varied its portfolio of funding through instruments such as the EU Trust Fund for Africa or the EU-Turkey Facility and modified some of the existing funding programmes for migration, asylum, and border management.
Following the broad priorities defined by the EU Commission, Parliament, and Council within multi-annual frameworks (the EU’s seven-year budget plan), specific Directorates General (DG), such as DG Migration and Home Affairs (HOME), DG for International Partnerships (INTPA), or the former DG for Neighbourhood and Enlargement Negotiations (NEAR), operationalise these goals into detailed programmes. Programming involves extensive consultations through multiple drafting rounds, engaging stakeholders, EU member states, NGOs, international bodies, and, for external programmes, EU Delegations and non-EU governments (European Court of Auditors 2023). Variations between policies and funds do not necessarily prevent the overall achievement of the objectives. However, they greatly influence the practical outcomes when projects are planned and executed. The breadth or narrowness in conceptualising a particular funding programme or project can constrain or empower the implementing actors, such as small NGOs or local organisations in target countries. Moreover, misalignments can feed back into the policy-making process. The design and understanding of tools influence policy formulation, creating a cyclical relationship where practical experiences and challenges during implementation inform subsequent policy decisions (Bali, Howlett, Lewis et al. Reference Bali, Howlett, Lewis and Ramesh2021). Policy tools create ‘instrument constituencies’ that, like epistemic communities or advocacy coalitions, become part of the overall process of governance change (Simons and Voß Reference Simons and Voß2018, 14). Therefore, understanding this expansion helps assess the overall change of EU migration policies (Bonjour, Servent and Thielemann Reference Bonjour, Servent and Thielemann2018), their coherence, and impact.
The MigFund dataset
The MigFund dataset represents the most comprehensive effort to trace the European Union’s committed funding in the area of migration and asylum from 2000 to 2023. Specifically, it systematically records both the amounts and the content of EU migration-related programmes spanning from the year 2000, when the EU significantly expanded its competencies in this realm following the entry into force of the Amsterdam Treaty in 1999, up to the current funding period. From a budgetary perspective, this means looking at four Multiannual Financial Frameworks: i) 2000–2006; ii) 2007–2013; iii) 2014–2020; and iv) 2021–2027.
The dataset consists of two main components, reflecting the internal and external dimensions of EU migration policy. By ‘internal dimension’, we refer to policies falling within the EU’s Area of Freedom, Security and Justice and addressing asylum, integration, security, and border management inside the European Union, while ‘external’ usually refers to those policies pertaining to the EU’s external relations and development cooperation (Czaika, Bohnet and Zardo Reference Czaika, Bohnet and Zardo2024). The internal component includes funding programmes managed by the Directorate General for Migration and Home Affairs (see Table 1), while the external one tracks migration-related interventions in development programmes (also referred to as migration management aid, Norman and Micinski Reference Norman and Micinski2023). Here, ‘interventions’ captures the ‘action’ supported through EU funding, such as capacity building, research, and direct assistance to migrants and refugees. It provides detailed information about the financial allocation, the areas of intervention (eg return and readmission, migration and development, legal mobility, border control), the targeted migrant population (eg all migrants, irregular migrants, asylum seekers, vulnerable people), the geographic regions and countries and the degrees of restrictiveness. As described below, restrictiveness is conceptualised as the extent to which a planned intervention is more focused on curtailing migration or providing mobility options (legal channels) towards Europe.
EU-funded programmes included in the MigFund database. Internal and external dimensions
Table 1. Long description
A table listing EU-funded programs categorized into internal and external dimensions of migration policy. The table has two main sections: Internal dimension and External dimension. The Internal dimension section includes 19 programs, each with a specific focus area and time frame. Examples include European Refugee Fund 2000–2004, ARGO Programme, European Fund for the Integration of third-country nationals 2007–2013, and Asylum, Migration and Integration Fund 2021–2027. The External dimension section includes 17 programs, such as AENEAS programme, CARDS Programme, MEDA Programme, and Facility for Refugees in Turkey. Each program is listed with its respective time frame and focus area. The table provides a comprehensive overview of the financial allocation and areas of intervention for these programs.
The MigFund dataset records committed amounts rather than actual spending. Focusing on commitments provides a clear view of how actors in charge of designing funding programmes have translated policy objectives into funding priorities, identifying which concepts, target groups, or types of activities were deemed more important. Actual spending tells us what happened on the ground, but commitments speak to intent: they show us what policymakers hoped to achieve and where they believed the greatest impact could be made. Therefore, analysing commitments helps identify potential shifts in policy focus, budgetary constraints, and the intended scale of support over time.
Data collection
The data collection aimed to answer five main questions: How much funding was committed? What area of migration policy did it target? Who were the intended beneficiaries? Where were the interventions directed? And by whom were they implemented? Additionally, it examined whether each measure leaned more towards restricting migration or providing mobility options towards Europe.
Each entry in the dataset represents a funding measure programmed between 2000 and 2023 within EU migration-related programmes, both internal and external, as summarised in Table 1. A measure refers to a distinct component of an EU-funded programme, typically focused on a broad area of intervention, such as border control, return or legal migration, or a specific set of policy objectives. Measures usually define the scope of the intervention, outline its objectives, allocate specific funding, and establish the mode of implementation. In some instances, this information is available in what we refer to as a sub-measure. Sub-measures break down broader objectives into more detailed, actionable components. For example, a measure from the 2005 AENEAS Programme is described in the dataset as: ‘Support the development of national policies according to a unified approach in the fields of migration and asylum and visa’, targeting ‘Refugees, asylum seekers, and other vulnerable people’ in ‘Eastern Europe, Southern Caucasus, and Central Asia’, with an indicative amount of ‘1 million Euros’ for ‘Direct assistance to migrants and refugees’.
The codes and values were constructed to match the type of information needed for each question; for example, numeric values were used to capture funding amounts, predefined categories helped classify policy areas, target groups, and regions, and a scale from -1 to +1 was used to reflect the level of restrictiveness in each measure (see Table 2). Each entry also specifies the MFF it falls under, the name of the programme, the total amount allocated to the entire programme, the overall amount assigned to the corresponding EU budget Heading, and, in cases of joint management, the managing partner responsible for implementation.
MigFund coding scheme
Table 2. Long description
A table detailing the coding scheme for migration funding questions and values. The table has six rows and three columns. The columns are labeled Questions, Codes, and Values. Row 1: Questions, How much?; Codes, Committed amount; Values, Million euros. Row 2: Questions, What?; Codes, Sub-policy area; Values, Border & land management; legal migration, integration, and mobility; irregular migration & forced return; asylum & migration management; protection; migration-development; crisis prevention. Row 3: Questions, Who?; Codes, Target group/migrant category; Values, Regular migrants; irregular migrants; all migrants; refugees, asylum seekers, and other vulnerable people; migrants, asylum seekers, and refugees; diaspora. Row 4: Questions, Where?; Codes, Geographic scope; Values, Regions/countries. Row 5: Questions, How?; Codes, Type of intervention; Values, Capacity building/institutional development; local development/investment; development: diaspora & remittances; research and data collection; direct assistance to migrants and refugees. Row 6: Questions, How?; Codes, Implementation mode; Values, Direct, indirect, jointly with international organisations, shared. Row 7: Questions, Does the measure place more emphasis on curbing migration (restrictive, -1) or on providing protection and mobility options towards Europe (open, +1)?; Codes, Restrictiveness; Values, -1; -0.5; 0; +0.5; +1.
We first relied on the European Commission’s implementing decisions on financing Annual Action Programmes/Plans (AAPs) as a source of information. These documents translate the EU’s strategic priorities into concrete actions and projects. These legally binding documents bring the EU’s strategic priorities to life by translating them into specific actions and projects to be carried out each year. The process begins with proposal development, often in collaboration with EU Delegations and partner countries and involves rounds of consultation with key stakeholders, including EU member states and the European Parliament. Once a consensus is reached, the Commission formally adopts the implementing decision, which sets out the budget allocations, the content and scope of each action, and the practical arrangements for implementation, such as timelines, monitoring systems, and the roles of partner organisations.
When these documents were unavailable or inaccessible, we turned to alternative sources to gather information at the (sub)measure level. These included Calls for Proposals, Multiannual Indicative Programmes, annual reports and staff working documents from the European Commission, and relevant EU regulations on standard rules and procedures. From these sources, we extracted key details such as the objectives of each measure, the expected results and indicators, the types of activities planned, the management and implementation approach, and the geographic areas targeted by the intervention.
Measuring the restrictiveness of EU funding programmes
A substantial body of research has shown that migration policies are not simply situated on a linear scale from liberal to restrictive, but are often internally inconsistent, combining openness in some areas (eg high-skilled migration) with restrictiveness in others (eg asylum or family reunification) (de Haas, Natter and Vezzoli Reference de Haas, Natter and Vezzoli2016; Ruhs Reference Ruhs2013). This literature further demonstrates that what is labelled as ‘liberal’ policy may still involve significant mechanisms of control, conditionality, and selectivity (Bonjour Reference Bonjour2011; de Haas, Natter and Vezzoli Reference de Haas, Natter and Vezzoli2016). The underlying assumption of the MigFund dataset is that, much like legislative instruments, a funding programme can meaningfully be categorised as either restrictive or open. A programme’s orientation is never incidental; it is embedded in decisions about which objectives to prioritise, which target groups to support, and where resources are allocated. These choices shape the direction of migration governance through what is financed and what remains unfunded, as well as what is enabled, supported, or constrained. In this way, the act of financing becomes a form of governance in itself, carrying a distinct logic of inclusion and exclusion (Tsourdi and Zardo Reference Tsourdi and Zardo2025). As such, funding programmes lend themselves to analysis through conceptual lenses similar to those applied in the study of migration policy development (de Haas, Natter and Vezzoli Reference de Haas, Natter and Vezzoli2015).
Based on this assumption, each measure in the dataset has been coded along a restrictive-open trajectory to describe whether it emphasised curbing migration (restrictive) or providing mobility options towards Europe (open). I used a content-based coding procedure, assigning values ranging from −1 to 1 to each programmed measure across five intensity levels. These levels encompass measures entirely focused on curbing migration (−1), primarily focused on curbing migration (−0,5), equally focused on both curbing migration and providing mobility options (0), mainly focused on providing mobility options (0,5), and measures entirely focused on providing mobility options (+1). The coding was based on a manual content analysis, which involved a close reading of each measure’s general objectives, specific objectives, suggested activities, target groups, and geographic scope. For example, a measure with the general aim of addressing migration and development but specifically promoting the circular migration of skilled workers to increase voluntary return was coded as 0, reflecting an equal balance between restrictive and open goals. Likewise, measures focused on protection or legal migration were not automatically assigned an open value (+1) if their specific objectives or activities were also geared toward reducing irregular migration, promoting return, or enhancing border control.
Additionally, since MigFund´s goal is to assess the relative impact of each measure (whether restrictive or open) on the overall programme, we introduced a Restrictiveness Index. This index not only captures the absolute level of restrictiveness of a measure, but also considers its proportional weight within the entire programme. The weight of each measure is calculated as the share of its committed amount within the total committed amount of the programme to which it belongs. This weight is then multiplied by the measure’s restrictiveness score (ranging from −1 to +1). Summing these weighted scores across all measures within a programme yields the programme’s overall Restrictiveness Index. The index helps the user understand how those responsible for programming translate broad policy goals into specific objectives and reveals how much funding they are willing to allocate to that specific objective and area of intervention (sub-policy area) as a proportion of the overall programme.
In coding each measure, we stayed as close as possible to the language used in the official documents, prioritising what was explicitly stated over any assumptions about implicit political intention. When the wording was too vague or broad to allow a clear categorisation, we turned to additional sections of the AAPs, such as the ‘rationale’ or ‘intervention logic’, to better understand the scope of the funding measure. In cases where a measure combined seemingly contradictory objectives (eg supporting both border surveillance and search and rescue), we assigned a neutral score of 0. However, when the source provided sufficient detail, such as the frequency of specific terms or a disaggregated budget distribution, we applied a more fine-grained analysis to reflect the dominant emphasis. Coding was first carried out in full by one researcher following a shared codebook with the decision rules outlined above and was then entirely reviewed and revised by the project’s principal investigator. Discrepancies identified during this review were discussed by both researchers and resolved jointly, ensuring that the final coding reflected a shared understanding of the criteria across the datasetFootnote 2.
Methodological constraints
While we consider MigFund to be the most comprehensive effort to date to compile data on EU-funded migration-relevant measures, the dataset is not exhaustive, and several scope decisions shape what it can and cannot show. A first limitation concerns instruments in the internal dimension, such as the European Social Fund (ESF), the European Regional Development Fund (ERDF) and the Fund for European Aid to the Most Deprived (FEAD), which are not captured in MigFund because they lacked a dedicated migration-relevant monitoring framework between 2000 and 2020. A comparable challenge applies to the external dimension: commitments under the European Civil Protection and Humanitarian Aid Operations (ECHO) are particularly difficult to trace, since ECHO’s Annual and Multiannual Work Programmes do not report funding for forcibly displaced persons’ aid-related measures as a distinct category, organising budgets instead by management mode, implementation arrangement, and geographic scope (within or outside the Union).
We also made deliberate decisions to exclude certain instruments. The European Development Fund (EDF) is excluded for 2000–2020 because, for most of this period, it was funded outside the EU budget through voluntary Member State contributions. While the European Commission managed it, the design of interventions fell under Member States’ responsibility and reflected their specific interests, making it conceptually distinct from the EU-budget instruments that are MigFund’s focus. Other instruments with migration-relevant components, namely the European Instrument for Democracy and Human Rights (EIDHR), the Instrument Contributing to Security and Peace (IcSP), and the European Defence Fund, were excluded, both because recent estimates suggest they allocate comparatively minimal resources to migration-relevant measures (Davis Reference Davis2021; Woollard, Liebl, Davis et al. Reference Woollard, Liebl, Davis and Casajuana2022) and because of resource constraints that led us to prioritise the most substantively relevant external-dimension instruments.
Finally, the dataset excludes financial allocations for the staffing, administrative, and operational costs of EU decentralised agencies involved in migration management, such as European Asylum Support Office/European Union Agency for Asylum (EASO/EUAA), Frontex, European Union Agency for the Operational Management of Large-Scale IT Systems in the Area of Freedom, Security and Justice (eu-LISA), and Europol. This decision reflects a concern that assessing restrictiveness on the basis of founding regulations or annual mission plans alone risks introducing substantial bias. MigFund does, however, include a separate analysis of the EU’s annual budgets, offering a broader perspective on the development and financial weight of these agencies over time.
Insights and applications of the MigFund dataset
At the start of the 21st century, the European Union’s investment in migration policy was modest, in line with the gradual and uneven supranationalisation of migration and asylum policies (Guiraudon Reference Guiraudon2000). The MigFund dataset shows well how, over the ensuing decades, this landscape has transformed. Year by year, the EU expanded its financial commitment to managing migration, both within its borders and through partnerships beyond. By 2027, what began as a collection of relatively small funds has evolved into a substantial budget for migration governance across the Union and beyond.
The aggregated data on the internal funding managed by the DG for Migration and Home Affairs (DG Home) shows that it grew across successive MFFs. During the 2000–2006 period (MFF1), funding for these issues (€2 billion) was only a fraction of what it would become. Each subsequent budget cycle brought a major increase in resources. By the 2014–2020 MFF (MFF3), internal funding for migration-related programmes had surged to roughly €20 billion, and in the 2021–2027 cycle (MFF4), it reached new heights at over €30 billion. Committed EU external funding for migration-related measures in development programmes, which MigFund now traces longitudinally for the first time until 2023, followed a similar upward trajectory (Figure 1). In the early 2000s, EU external aid earmarked for migration in EU development programmes (excluding the EDF) was relatively small, amounting to only a few hundred million euros. By the 2014–2020 period, external funding linked to migration management had expanded to tens of billions of euros (see Figure 1). Initial allocations for 2021–2027 remain substantial.
Committed amounts for migration-related activities in the multiannual financial frameworks. MigFund database. For the MFF4 (2021–2027), data for the external dimensions are shown only up to 2023, reflecting the cut-off date of MigFund’s data collection. Later commitments for the external dimension under the MFF4 are therefore not yet captured in this article. The programming of the internal dimension was, however, entirely available at the cut-off date and included in the dataset, excluding amendments.
The EU thus increased spending on migration abroad in parallel with its internal funding, engaging more with third countries to manage migration flows. This twofold expansion challenges potential expectations of an internal/external trade-off (Lavenex Reference Lavenex2006). The EU invested in domestic capacity (asylum systems, border control, integration programmes) while simultaneously funding external initiatives (development projects tackling the so-called root causes of migration, partnerships with third countries on border control, protection of migrants along the migration route). From this perspective, the dataset is a resource for scholars investigating broader EU budgetary politics. It provides an empirical basis for tracking shifts in EU commitments across budgetary cycles and capturing how policy priorities evolve over time as a result of complex negotiation processes within EU institutions. EU budgetary politics are characterised by intense coalition-building and ideological negotiations in the Council and Parliament, factors that have been identified as key drivers of budgetary stability and change (Citi Reference Citi2015; de Wilde Reference de Wilde2012). Additionally, the database includes data on mid-term budgetary amendments and enables understanding of how the EU adjusts its financial commitments in response to unforeseen crises or political shifts, a dynamic that is increasingly relevant in contemporary EU governance, also beyond the migration realm (De Witte Reference De Witte2023; Schramm and Terranova Reference Schramm and Terranova2024).
An important contribution of the MigFund database lies in its potential to support the analysis of implementation gaps in EU migration policy from a public policy and administration perspective. By offering disaggregated data (see Figure 2) on EU financial commitments broken down by specific programmes and sub-policy areas (eg integration, return, border control), it enables tracking of ‘intent versus outcome’ across different instruments and over time, highlighting where funds remain unused or under-utilised and broadening prior work on funds for migration and integration (Van Wolleghem Reference Van Wolleghem2022). Figure 2 illustrates this potential with one example: the disaggregation of AMIF 2014–2020 commitments by sub-policy area. It is presented here as an illustrative starting point for the kind of disaggregated analysis the dataset enables, rather than as a direct measure of implementation gaps, which would require linking these commitment figures to data on actual disbursement or absorption. The dataset enables similar comparisons across time, within and across instruments.
Committed amount for the AMIF programme 2014–2020 broken down by sub-policy area. MigFund database.
Moreover, as described above, each entry in the dataset was coded based on general and specific objectives to identify the language and intentions behind individual funding measures. This structure helps uncover patterns in how EU actors have conceptualised migration over time and designed responses to it, across different programmes and within both the internal and external policy dimensions. In doing so, it speaks not only to migration scholars, but also to those interested more broadly in how public problems are constructed, categorised, and acted upon within the field of public policy. A preliminary analysis, for instance, shows how internal programmes such as AMIF and the Border Management and Visa Instrument (BMVI), along with their predecessors, tend to frame migration crises as strains on domestic governance capacity and a test of European solidarity. In this context, the goal of funding is to strengthen Member States’ asylum systems, reception infrastructure, facilitate relocation, and expand integration programmes. Migration is thus conceptualised as a challenge to be managed collectively within the EU, as highlighted in Figure 3, where measures targeting crisis management (see, for instance, the MFF 2014–2020) are on average coded as open (mostly focused on mobility and integration). By contrast, in the external dimension of EU migration funding, crisis is conceived not primarily as a strain for the countries concerned but as a trigger of migration towards Europe. This framing of crisis in external programmes leads to commitments for capacity building for security forces or actions to prevent secondary movements of forcibly displaced people. These interventions are therefore coded as restrictive. Or, unlike other dimensions of migration policies, the goal of protecting migrants along the route appears to be relatively insulated from more restrictive framings (Zardo Reference Zardo2025). It is less often tied to a containment logic and more often framed in neutral or positive terms, leading to proposed activities in the funding programmes such as the provision of assistance to vulnerable groups or services aimed at upholding human rights.
Level of restrictiveness of the funding measure. MigFund database (internal dimension). A lower value (−1) means a very restrictive framing of the funding measure.
A preliminary combined geographic and thematic analysis reveals that between 2007 and 2013, the EU directed the majority of migration-relevant funding in its development cooperation programmes for the Eastern Neighbourhood towards border management, committing approximately €532 million to this area. Cooperation programmes in Africa during the same period were to a large extent oriented towards controlling irregular migration and supporting forced-return operations. The 2015 migration crisis represented a turning point, with the establishment of ad-hoc Trust Funds that reshaped the structure and focus of cooperation programmes. In the Middle East, the EU Regional Trust Fund for Syria channelled funding almost exclusively into protection-related activities for displaced populations in the Region while in Africa, the EU Emergency Trust Fund for Stability and Addressing Root Causes of Irregular Migration (EUTF) shifted the emphasis toward addressing the so-called ‘root causes’ of migration and funding allocations moved toward local development initiatives alongside increased support for the protection of displaced populations.
Figure 4 extends this geographic analysis by visualising the weighted restrictiveness score of EU external migration funding by targeted country across the full 2000–2023 period. Unlike conventional analysis of EU migration governance and externalisation, this distribution shows that the EU has committed more funding for measures coded as open in countries in Sub-Saharan Africa and South and Southeast Asia compared to partner countries in the South Caucasus, Central Asia, and parts of Latin America, which show more restrictive scores. This geographic distribution is partly explained by the development and protection-oriented framing of the Trust Funds established after 2015: the EU Emergency Trust Fund for Africa, the Facility for Refugees in Turkey and the EU Regional Trust Fund for Syria, pulling the weighted scores of their primary recipient countries in an open direction. These patterns are nonetheless aggregated across two decades and multiple programming cycles, and the figure is best read as a starting point for further regional comparative research, drawing on the multiannual and annual programme documents linked to the recorded MigFund measures.
Weighted restrictiveness of migration-relevant measures in MigFund (external dimension).
The data provided by the MigFund dataset can thus support both macro- and meso-level analyses that have so far informed discussions on migration policies, as it offers a detailed and longitudinal empirical basis to trace financial priorities.
Conclusion
When we first embarked on the MigFund project, a central question was whether EU funding programmes could be analytically treated as policy instruments and, if so, how they might be meaningfully characterised as either restrictive or open, in line with ongoing research on the evolution of migration policies (de Haas, Natter and Vezzoli Reference de Haas, Natter and Vezzoli2016; Thielemann and Hobolth Reference Thielemann and Hobolth2016). Developing the dataset confirmed the analytical value of this approach. Funding measures are not neutral vehicles of implementation but embedded choices, reflected in how target groups are defined, how regions are prioritised, and how issues like ‘crisis’ or ‘protection’ are framed. These features express institutionalised understandings of migration, as well as the bureaucratic and political negotiations that shape them. Financial instruments, therefore, function as autonomous tools of governance (Tsourdi and Zardo Reference Tsourdi and Zardo2025) rather than as direct expressions of formal policies.
Categorising a funding measure as restrictive or open is a conceptual intervention into the broader debate on how the governance of international migration is developing over time. This binary, while imperfect, helps to highlight the tensions between control and mobility, as well as between securitised and rights-based approaches. Applied to financial tools, this lens shows how instruments can advance restrictive objectives such as return, deterrence, or containment through ostensibly technical programming or, conversely, support protection, legal pathways or integration even when framed within broader narratives of control. Different conceptualisations often coexist within single instruments or programmes and entail internal contradictions during the implementation process or the measurement of cooperation outcomes. Financial tools thus become a site where broader policy logics are negotiated, constrained, or reinterpreted.
Like any dataset, MigFund has limitations that delineate its appropriate uses. First, by design, it records committed rather than disbursed amounts. It describes the intentions and priorities of EU´s policymakers rather than what was ultimately spent on the ground, and it cannot on its own be used to assess implementation performance or absorption rates. It provides, however, the basis for such comparisons, especially since data on expenditures for later years are becoming increasingly available. Second, the dataset captures the content and stated objectives of funding measures as reflected in official programming documents; it is not designed to reconstruct the internal negotiation processes between EU institutions, member states, and third countries that shaped these documents, although the variation it documents across time and instruments serves as an empirical starting point for research that combines MigFund with process-tracing or archival methods. Third, the depth and consistency of the underlying source documentation vary across instruments and periods: Annual Action Programmes/Plans, the primary source, were not uniformly available for all instruments and years. Where they were missing (approximately 2% of the data), secondary sources such as Calls for Proposals and Multiannual Indicative Programmes were used instead. Fourth, as discussed above, several internal and external instruments are either partially or fully excluded from the dataset due to data availability constraints or substantive scope decisions, meaning MigFund should be read as the most comprehensive available account of EU migration-relevant commitments rather than an exhaustive one. These limitations notwithstanding, the dataset offers, to our knowledge, the most comprehensive longitudinal resource currently available for studying EU migration funding.
Looking ahead, the MigFund dataset offers a valuable basis for political scientists to engage empirically with the political economy of EU migration governance. It enables the study of agenda-setting, institutional inertia, and shifting coalitions across time, space, and policy domains. It also offers new ground for analysing how supranational actors translate contested goals into operational instruments and how these instruments interact with crises, policy feedback loops, and implementation practices. The approach described in this contribution can also be extended, mutatis mutandis, to other policy areas. Moreover, it opens underexplored lines of inquiry, such as how conceptual frames differ across the EU’s internal and external budgetary headings or how budgetary choices produce structural advantages for certain actors and priorities over others. Ultimately, it prompts a shift in how we perceive the role of money in governance. It brings financial programming to the forefront of migration research, as a process that shapes, enables, and delimits the trajectories of EU migration and asylum governance.
Data availability statement
The dataset described in this article and supporting the results of this study is publicly available at: MigFund Dataset | Zenodo.
Acknowledgements
I am deeply indebted to my research assistant, Dr. Raphael Tservenis, for establishing the MigFund dataset. His tireless efforts in collecting and organising the data were integral to the project’s success, and the work would not have been possible without his commitment. Any errors or shortcomings in the dataset are solely my responsibility.
Financial support
This work was supported by the Austrian Science Fund through the grant agreement ESPRIT 144-G.
Competing interests
No potential conflict of interest was reported by the author(s).
Use of AI
No generative AI was used to write this article. Grammarly Pro was used for editing.
Open access
