The EU’s cohesion policy is one of the biggest items in the EU budget and a central sticking point in the upcoming negotiations over the next seven-year funding cycle. Based on data from more than 2.4 million households, we study the distributional impact of the program and show that it often fails to reduce regional and social inequalities. The funds frequently target places that are not particularly needy and mainly benefit people at the upper end of the income distribution. Cohesion policy thus requires root-and-branch reform. This includes a sharper focus on truly disadvantaged areas, a more precise definition of local economic challenges, improved funding access for small municipalities and companies, and redirection of resources now allocated to wealthy member states towards an EU-level investment instrument.
The European Union’s (EU) cohesion policy is one of the biggest and politically most important items in the European budget. Over the years, it has served a swath of different goals—from supporting economic catch-up in poor regions, to strengthening EU competitiveness to fighting climate change and, most recently, as a war chest to respond to various short-term crises. At its treaty-enshrined core, however, the policy is supposed to strengthen economic, social, and territorial cohesion by reducing economic disparities between regions at different levels of development (Art. 174 TFEU).
Via a new study, we show that cohesion policy as it stands fails to support this fundamental goal. We collect income data from more than 2.4 million households, use newly available data on where and when cohesion money has been dispensed over the last 30 years, and study who exactly benefits from the program. Our analysis shows that cohesion policy has a positive impact on regional output and growth but suffers from two fundamental problems.
Put simply, it often targets the wrong places and reaches the wrong people. First, cohesion policy aims to reduce economic disparities between large European regions. However, our data shows that the specific regional inequalities addressed by the policy contribute little overall inequality in Europe and have decreased over time. A lot of money, therefore, simply goes to the wrong places. Second, it reaches the wrong people. While cohesion policy spending raises average incomes in the targeted regions, these gains go almost exclusively to relatively wealthy households. Instead of aiding the needy, the policy often ends up enriching affluent households in areas that are not particularly poor.
Against this background, cohesion policy requires root-and-branch reform. The upcoming EU budget negotiations offer a rare opportunity to achieve this. An ideal reform should effectuate a clearer differentiation between spending that genuinely addresses the pressing political need to reduce regional and social inequalities and investments aimed at other and broader EU policy objectives. This includes four elements: a sharper focus on truly disadvantaged areas; a more precise definition of local economic challenges; improved funding access for small municipalities and companies; and, finally, a redirection of resources now allocated to wealthy member states towards EU-level instrument that is more geared towards supporting common priorities, such as investments in shared infrastructure and industrial policy.
1) Why getting Cohesion Policy right is important
Getting cohesion policy right is important. First, the EU allocates substantial funds to its regional policy. For the 2021-2027 funding period, structural and investment funds in the European budget amount to €392 billion, averaging €56 billion annually. Consequently, in this funding period, close to a third of every euro disbursed under the regular Multiannual Financial Framework (MFF) falls under the cohesion umbrella.
This is not a recent development. A true EU-level regional policy was introduced in the late 1980s and has since expanded in economic scope. In the 1990s, the largest recipient regions saw funds accounting for 2-3 percent of their local GDP. By the 2010s, many regions were receiving EU funds exceeding 5 percent. By way of comparison, this exceeds the allocations that states like Berlin and Saxony-Anhalt receive from the German interstate fiscal equalization scheme.
How Cohesion Policy currently works
In the 2021-27 funding period, cohesion policy comprises five funds. The European Regional Development Fund (ERDF) and the European Social Fund Plus (ESF+) continue to be by far the largest instruments (see Figure 1). The ERDF aims at reducing economic disparities between regions and, at the same time, funnels spending in five thematic concentrations such as innovation and competitiveness, climate neutrality and social inclusion. The ESF+ is also a regional fund but aims more specifically at reducing unemployment, promoting social inclusion and supporting worker training. Allocations under both funds follow broadly the same rules. Most resources go to less developed regions with GDP per capita below 75% of the EU average. However, transition regions (GDP p.c. between 75-100% of the EU average) and more developed regions (GDP p.c. above the EU average) also receive funding (see Figure 2). On top of regional GDP, other factors such as regional unemployment, youth unemployment and greenhouse gas emissions have a minor impact.
Figure 1 – Distribution of Cohesion policy resources across funds and member states. Source: European Commission (2024)
The other three funds are substantially smaller and more sharply targeted. The Cohesion Fund (CF) aims at supporting investments in transport and environment infrastructure in member states with a Gross National Income below 90% of the EU average and has no subnational allocation key. The Just Transition Fund is the latest addition to the cohesion toolbox and is supposed to support territories most affected by the transition towards climate neutrality. Again, funds are allocated to all member states based on factors such as industrial emissions, employment in industry and overall level of development. In a second step, member states identify smaller NUTS-3 level regions most in need within their jurisdiction. Finally, a subset of funds from the ERDF is reserved for projects that cut across regions and member states (Interreg). Cohesion policy, thus, embraces a variety of funds and aims. However, in practice these funds are negotiated jointly and cuts in one financial instrument are often compensated for by increases in others.
Figure 2 – Distribution of spending under the ERDF and ESF+ across different categories of NUTS2 regions.
All cohesion policy spending is regulated by the Common Provision Regulation (CPR) and follows similar principles. First, it falls under shared management between the Commission and member states. This means that the EU level provides rules for the allocation of funds and broad policy goals, but member states (and their regions) are in charge of designing the specific programs following consultation with a range of stakeholders. It also means that the Commission does not allocate funding to specific regions but only monitors that the national program follows EU allocation rules on funding across the three categories of regions. Moreover, all cohesion policy is co-financed. While the EU per se provides some funding, member states must chip in as well with different co-financing rates set for different programs and categories of regions.
Importantly, cohesion funds are not limited to impoverished regions. The primary allocation rule has until today consistently directed the majority of funding to areas where GDP per capita falls below 75 percent of the EU average. However, since 2007, the EU has distributed funds across all European regions, including the wealthiest. In the current budget cycle, approximately €84 billion—around 27 percent of all funds disbursed under the European Regional Development Fund (ERDF) and European Social Fund (ESF)—has been allocated to regions where GDP per capita exceeds the 75 percent threshold. Remarkably, €30.5 billion of this is directed towards areas with a GDP per capita above the EU average.
Second, cohesion policy is one of the few ways in which the EU can underwrite its policy priorities with real money. Above all, cohesion policy aims to reduce economic disparities within the EU. This involves aiding poorer regions to catch up and helping them adapt to the rigors of the single market. In addition, the treaties compel the EU to use this policy to foster social cohesion by improving the situation of the most vulnerable and by helping those at risk of poverty. At its core, the program is designed to spread the gains of European economic integration and reduce economic inequalities.
However, given the absence of other investment instruments, the EU pours various political priorities into its regional policy. In 2007, regional innovation and competitiveness were emphasized as key policy goals under the Lisbon Strategy. Since 2014, new spending priorities such as the digital transition and climate investments have been folded in. Today, regional programs must entail minimum benchmarks for investments in a range of different thematically linked fields, from climate via digital to youth unemployment. This is particularly true in transition and more developed regions, where EU provisions mandate that 70% to 85% of funds target broader investment objectives such as innovation and the green transition. What's more, the rigidity built into the overall EU budget has frequently turned cohesion policy into a budgetary lifeline to tackle short-term spending needs related to issues like migration, the energy crisis or the EU’s response to the US Inflation Reduction Act.
Cohesion policy has developed a complex governance machinery to achieve these goals. Regional funds are managed collaboratively, with member states and regional authorities drawing up multiannual programs incorporating input from a broad spectrum of regional and local stakeholders and civil society. The Commission then reviews these plans to ensure they meet regulatory requirements before national and regional administrations begin to put them to work. Supporters of this partnership approach hail it as a core way to sideline central governments and directly link European regions and communities with the EU. It also meets the prerequisite of ensuring that investments meet the real needs of what are often vastly different local contexts. Critics, on the other hand, see it as a bureaucratic juggernaut contributing to the notoriously slow administration and absorption of cohesion funds.1
2) Wrong places, wrong people – current issues of cohesion Policy
The wide range of policy goals that have been tacked onto cohesion policy make the benchmarks for assessing its success somewhat fuzzy. In a new study, we focus on evaluating to what degree the ERDF and the ESF+ as the largest cohesion programs achieve the policy's most basic function: reducing social and economic disparities between European regions and people. Our findings show that, from a cohesion perspective, the policy has two fundamental problems.
Cohesion policy addresses the wrong kind of inequality
First, cohesion policy often ends up supporting the wrong kind of places. The program allocates funds from the ERDF and the ESF to so-called NUTS2 regions. These administrative units correspond to, for example, a Regierungsbezirk in Germany, a région in France, and a comunidad autónoma in Spain (see also Figure 1). These are relatively large units. According to the EU definition, each country's average NUTS2 region is supposed to have 0.8-3m inhabitants.
Methods Box 1 – New regional income data
To understand who benefits from the EU’s place-based policies, we compiled the first comprehensive panel data set on the distribution of income across and within European subnational regions between 1989 and2017. We collected household-level income data from a large set of national household surveys and more than 2.4m individual respondents. This includes national surveys compiled by the Luxembourg Income Study (LIS), the EU’s Statistics of Income and Living Conditions (SILC) and several national sources. Altogether, we based our data on 260 household surveys. To compare incomes across households of different sizes, we divide household incomes by the square root of household members. To compare incomes across countries and over time, we adjust them to 2011 international dollars at purchasing power parity (PPP). For a detailed description of the data sources and the steps we took to collect, harmonize and process the data see Lang, Redeker & Bischof (2024).
Importantly, these regions also house very different economic realities. Our newly compiled data set is based on individual income data from more than 2.4m households covering over 30 years across the EU. It allows us, for the first time, to move beyond average statistics like GDP per capital and zoom into the details of how incomes are distributed within the geographical areas which cohesion policy targets (see Methods Box 1). Figure 2 plots the disposable income of different percentiles of the within-region income distribution across NUTS2 regions. The regions are ordered by mean disposable household income. The richest regions include Luxembourg, the greater Paris area (“Ile-de-France”) and regions in Southern Germany. Among the poorest (with data) are regions in Poland, Hungary, Southern Italy, and the Baltics.
Figure 3 - Annual equivalized disposable household income of various percentiles of the intra-regional
income distribution, latest available year.
It shows that income inequality within the targeted regions is vast. Even within Europe’s wealthiest regions, many people have lower disposable incomes than the median in relatively poor areas. And even in less developed regions that receive most EU funding, such as Molise in Italy or XXX in XX, households in the upper third of the income distribution have higher incomes than those in the poorer third of affluent places. In transition or more developed regions that still get money from the cohesion pot, households in the upper deciles even figure among the EU's top income earners.
Another way to show this is that inequality between regions targeted by cohesion policy is not what drives overall inequality in Europe. Our data allow us to decompose the share of total inequality in Europe that is driven by disparities between NUTS2 regions versus inequality within them. Figure 3 illustrates the evolution of these two components from the late 1990s to the late 2010s. The data reveal that inequality within large regions has always been the dominant factor while its significance has risen over time. Currently, only X% of total income inequality in Europe arises from differences between places that are the core targets of the redistributive efforts of cohesion policy.
Figure 4 - Between-region and the within-region component of European inequality. The height of the bars indicates the level of inequality as measured by the mean log deviation. To ensure comparability over time, the sample of regions is fixed.
The fact that economic inequalities within the regions which cohesions funds target are so large is a problem. There is no regulation or general rule stipulating how EU funding is distributed within the receiving regions and money tends to flow to the most prosperous local hubs. Some of this is driven by agglomeration effects. However, previous research also shows richer municipalities are blessed with better administrative capacities to secure European funding. As a result, even when EU funding targets economically deprived NUTS2 regions (which in many cases it does not), it often ends up in cities and hubs that, by European standards, are not particularly poor.
Cohesion policy reaches the wrong people
Second, cohesion policy often reaches the wrong people. With our data, we can study who specifically benefits from cohesion spending under the ERDF and ESF. Two elements make this analysis especially interesting. First, our statistical models allow us to study causal effects for different income deciles. We can, thus, analyze how EU funds impact income growth for different parts of the local income distribution (see Methods Box 2). Second, we use newly available data on the precise timing of EU fund expenditures. This enables us to estimate how much GDP growth is generated by each euro of EU spending.
Figure 5 – Estimated impact of funding on annual growth of total income by regional decile-group along with 95% confidence-intervals.
The good news is that EU regional funds spur growth. EU spending has a positive impact on regional output and average household incomes and comes with a reasonable fiscal multiplier. On average, every euro the EU invests under the ERDF and the ESF lifts regional GDP by about 0.93 to 1.47 EUR (see Lang, Redeker, Bischof (2024), p. 21). While the requirement for national co-financing might lead to a slight overestimation of the growth impact in our models, the overall effect remains robust. EU funds successfully drive up economic activity and seem reasonably efficient.
Methods Box 2 – Identifying the effect of EU spending on local outcomes
We study the causal effect of EU funds on different economic and political outcomes in the receiving regions. The statistical problem is that these funds are not allocated randomly. The amount of money that a region receives is driven by a range of known and unknown factors. Simply studying the correlation between the EU spending and local outcomes would not allow us to disentangle whether the results are driven by EU spending or simply linked to other local factors. We solve this issue by taking advantage of the fact that the allocation has always followed the rule that regions with a GDP per capita below 75 of the EU average qualified for substantially larger amounts of funding than others. This provides a natural experiment that we can leverage for a so-called regression discontinuity design (RD). The intuition behind this technique is that regions just above the 75 percent threshold are very similar to regions just below them but still receive substantially lower funding for purely administrative reasons. This provides variation in EU spending which is as good as random. Specifically, we estimate a range of sharp and fuzzy RD models for various outcomes and conduct several robustness tests and alternative identification strategies. The specific estimation techniques and all further tests can be found in Lang, Redeker & Bischof (2024).
The bad news is that most income gains end up in the pockets of relatively wealthy households. Our analysis reveals that the benefits of EU funds are heavily concentrated at the upper end of the income distribution (see Figure 3). The richest 30 percent in the recipient regions experience a large and statistically significant boost in income growth. In contrast, the gains for middle-income groups are much smaller, and for poorer households there is no significant effect at all. Importantly, there is also no evidence that income gains trickle down to poorer households over time. This distributional pattern holds steady whether we look at annual changes or averages over the regular seven-year (MFF) funding period.
We do not find any evidence that this distributional pattern is due to sectoral biases. EU funds stimulate investment and employment in a wide range of industries, not just the high-paying ones. Moreover, EU money does not predominantly go to capital owners or local landlords who might benefit from higher house prices as a result of more people moving to the places that benefit from generous EU funding. Instead, the adverse distribution pattern stems mainly from high-skilled workers reaping significantly larger income gains compared to their less educated counterparts. It seems that the EU cohesion funding machinery currently privileges skills over needs.2
We can only speculate about the mechanisms driving this trend. One likely reason is that securing EU funds is complex. Firms must be aware of these subsidies, navigate lengthy bidding and application processes, and ultimately succeed. Previous research suggests that larger, more productive firms are, therefore, naturally better positioned to secure EU funding, particularly in less developed regions. These firms often employ more highly skilled and well-paid workers, which could contribute to the observed pattern.
Figure 6 – Estimated impact of funding on EU support by regional occupational group along with 95% confidence-intervals.
Whatever the reason, the results matter economically and politically. Economically, they mean that EU funds significantly increase local income inequality and often fail to reach low and medium-income households. The fact that the policy mainly lifts wages at the upper end of the income distribution is particularly problematic in richer places from a social cohesion perspective. Especially in transition and more developed regions, the main beneficiaries of EU money tend to be in the group of top European income earners.
The policy’s uneven economic benefits also limit its political appeal. EU regional funds are often heralded as a defense against the rise of right-wing populism and anti-EU sentiments. Indeed, numerous studies indicate that EU funds lower average levels of Euroscepticism in recipient regions. However, our analysis shows that once more these averages mask important differences. In line with our economic finding, it is mainly highly educated people there that know about EU funds and say that they benefit from them. Similarly, ongoing work with Johannes Lattmann suggests that the statistically significant positive impact that cohesion policy funds have on support for the EU is confined to highly skilled workers, managers and business owners (Figure 4). The policy, thus, seems to fail to convince broader segments of the population that the EU’s place-based program is lifting their tide.
3) How to make cohesion better – principles for fundamental reforms
EU cohesion policy needs fundamental reform. As the next budget negotiations for the MFF 2027-34 approach, early discussions are beginning to take shape. So far, the political buzz has largely centered on whether payments should follow the model of the Recovery and Resilience Facility (RRF), making funds conditional on member states undertaking domestic reforms and achieving specific investment milestones (the academic and expert discussions are more sophisticated, e.g. here and here). While this is an important consideration, our findings indicate that the EU has bigger fish to fry when it comes to its place-based policies. What is needed is a principled discussion on where and why the EU spends cohesion funds.
So far, every attempt to reform cohesion policy has been unhappy in its own way. The policy’s size and reach have created a large group of national and regional actors that stand to lose from major adjustments. More significantly, the redistributive, formula-based allocation of funds enables each member state to calculate precisely how much they will pay into and receive from the common pot. To national negotiators, this predictability often makes cohesion policy funds more appealing than other EU instruments.
This time could be different. The EU faces enormous budgetary pressures, with rising investment needs for the green transition, competitiveness and (economic) security. Additionally, the debt financing costs for NextGenerationEU will cut deep into the EU’s purse from 2028. The EU will, thus, require a bigger budget and, at the same time, have to think twice about how and where it spends its resources. On top, the prospect of EU enlargement means that significantly poorer member states could join the Union in the next budget cycle, drastically shifting the allocation of cohesion funds. Simply sticking with the status quo is becoming increasingly untenable.
The EU should seize this opportunity to clarify the role of cohesion policy within its overall investment framework. First and foremost, it needs a clearer division of labor between spending that genuinely enhances social and territorial cohesion and investments aimed at broader EU policy objectives. An ideal reform would pursue two goals. First, cohesion spending under the ERDF and the ESF+ should be refocused on poor areas in poor member states and reformed to reach broader parts of the income distribution. Second, resources currently allocated to relatively wealthy regions should be redirected into an EU-level investment instrument , where they can more effectively target EU-wide priorities.
Making cohesion policy work for cohesion
First, spending under the ERDF and ESF should be refocused on poor places and poor people. For one thing, this requires a more fine-grained definition of local economic grievances. Our findings indicate that NUTS2 regions are too large and economically diverse to effectively allocate funds to those areas truly in need of economic support. Resources should be distributed to smaller geographic areas to ensure that funds avoid ending up in relatively affluent hubs within poorer regions alone. Moreover, research shows that the economic benefits of place-based development programs are substantially greater in distressed areas than in more developed ones (see, for example, here, here and here). Therefore, targeting smaller, more economically deprived regions would likely enhance both the equity and efficacy of EU investments.
Moreover, cohesion policy needs to adjust its instruments to reach poorer households. The European Commission should think hard on how to make its place-based instruments more inclusive. Existing research already points to the fact that a lot of funding goes to big and capital-intensive firms. Our findings show that this also translates into disproportionate benefits for high-income households. The question is how to change that. One way is simplifying the administration of funds at the firm level to make it easier for smaller companies and their workers to access EU support. Additionally, placing a stronger emphasis on building local infrastructure and supporting public services like childcare, education, and training, could result in broader growth effects. Linking firm support to social conditions, such as wage minimums or job creation, is another option. Solutions are likely to differ from place to place. The key point is therefore that reducing the high-income bias in the current program needs to be a priority in the next cycle.
Improving investments on other policy goals
Second, the EU should stop spending cohesion money in places where funding clearly does not serve territorial or social cohesion. Our findings show that a significant portion of ERDF and ESF investment is currently funneled back to affluent regions in rich member states, benefiting households among the EU's top income earners. If the goal is to reduce regional and social inequalities, this approach is clearly misguided. If the aim, on the other hand, is to invest in EU-wide priorities, resources should be spent where they generate the biggest bang for the buck. Channeling them through regional quotas and shared management systems that leave the interpretation of these priorities and investment decisions largely to member states and regional bodies is unlikely to achieve that.
To better address the broader investment goals currently baked into cohesion policy, ERDF and ESF resources allocated to wealthy member states should, therefore, be redirected. These funds should flow into an EU-level vehicle focused on common priorities like infrastructure, climate initiatives, and competitiveness. This reallocation means that some rich countries will lose out from the revamped cohesion pot. However, projects in wealthier member states are likely to benefit more from central EU funding given the concentration of economic activity extant in these areas. Consequently, it is unlikely that net contributors would lose out. Importantly, such transfers from cohesion funds alone will not suffice to meet all EU-level investment needs; they should, thus, be viewed as just one financing component of a robust fiscal instrument for tackling joint priorities.
In our view, the outlined reform agenda is the most logical way forward. It may not be the politically most viable one. Within the Commission there are ideas on modelling the next version of cohesion policy more closely on proven RRF lines. However, following this approach comes with its own set of governance issues and would not necessarily solve the distributional issues described above. One important question going forward, then is if and how these ideal reforms could be integrated into a broader RRF-style governance reboot of cohesion policy (see Lindner & Redeker, forthcoming).
Conclusion
The EU faces big economic and social challenges. It needs a bigger – and a better – budget; this includes a cohesion policy that genuinely fosters cohesion and investment strategies that deliver the greatest value for money. The most pressing risk, therefore, is that the EU will squander the upcoming opportunity for reforms and fall back into its usual cohesion rut. The next Commission has only a couple of months before it must present its proposal for the next seven-year budget. It should use this time to push for a meaningful makeover.
1 At the end of 2020, for example, only slightly more than half of the total available ERDF, CF, ESF and YEI resources from the 2014-2020 budget period had been paid out. In 2024, only about 5 percent of all funds from the current budget cycle has been paid, though some of this is also caused by administrative backlog from the implementation of the Recovery and Resilience Facility. As a rule, member states can access funds for up to three years into the new budget cycle.
2 As explained, the ERDF and ESF focus on different goals. It would therefore be interesting to see whether the two programs have different distributional effects. However, given that spending for both of them follows the same allocation rules, we cannot study them separately.