Spain has long faced persistent challenges in its fight against poverty and social exclusion, especially among children, migrants and vulnerable communities. Covid-19 has further amplified these inequalities. In response, the Spanish government launched a nationwide minimum living income scheme (IMV by its Spanish acronym) in May 2020 to combat poverty and social exclusion. But poverty is a multidimensional problem that requires complementary policies, and it is still unclear which types of policies are most effective in reducing it.

To assess the real impact of these policies, the Ministry of Inclusion, Social Security and Migration (MISSM) set up the Inclusion Policy Lab in 2021 to produce rigorous and actionable evidence on 32 social inclusion programmes using randomised evaluations. These policies focused on issues related to work and entrepreneurship, social support and social benefit take-up, education, digitalisation, and housing and energy poverty. They targeted IMV beneficiaries and other populations at risk of poverty or social exclusion. The policies were implemented throughout Spain by regional governments, city councils and non-governmental organisations (NGOs). The programmes were funded by the NextGenerationEU Funds as part of the Recovery, Transformation and Resilience Plan, while the evaluations were conducted by the MISSM in collaboration with a team of researchers coordinated by CEMFI and J-PAL Europe.

The evaluations have provided rich insights on the effectiveness of a broad range of anti-poverty strategies and contributed to the institutionalisation of evidence-based policymaking in the country. Most policies showed several significant effects on important aspects of the social inclusion and well-being of the participants.
 

The share of the population at risk of poverty or social exclusion, commonly referred to as the AROPE rate, includes those who are in at least one of the following three situations: (a) at risk of poverty, in other words, with an equivalised disposable income below the at-risk-of-poverty threshold; (b) severe material and social deprivation, namely, not being able to afford at least 7 out of 13 items that are considered by most people to be desirable or even necessary to lead an adequate quality of life; or (c) living in a household with very low work intensity, in other words, where adults worked for 20% or less of their total combined work-time potential during the previous twelve months.

Spain has some of the highest rates of poverty and social exclusion in Europe. In 2024, its AROPE rate was 24.5%, the fourth largest in the European Union (EU), while the EU average rate was 21%, according to Eurostat. Furthermore, its AROPE rate reached a staggering 34.7% among people under the age of 16.



Although Spain spends significant amounts on pensions and unemployment benefits, social protection programmes remain underfunded, accounting for 18.5% of GDP in 2023, compared to an EU average of 39.3%, according to Eurostat. Regional disparities in policy implementation exacerbate this problem. In 2020, the covid-19 pandemic led to a 3% drop in employment and highlighted the fragility of the foundations of the Spanish welfare system. The new minimum living income scheme aimed to fill these gaps through targeted cash transfers. 

In 2023, the annualised expenditure of the IMV amounted to €2.5 billion, and it reached 17% of households at risk of poverty (with a threshold of 40% of the median). Nonetheless, doubts remain about its scope and impact. Most notably, individuals face administrative barriers when applying for the IMV, resulting in an estimated non-take-up rate of 56% among potentially eligible households (AIReF, 2024).

The Inclusion Policy Lab is a collaborative initiative to generate rigorous evidence on social inclusion policies through randomised evaluations. Launched in 2021 by the MISSM’s General Secretariat for Inclusion (GSI), it is part of Spain’s Recovery, Transformation and Resilience Plan and is funded by the European Union (EU) through NextGenerationEU. The Lab operates through partnerships between the GSI and implementing organisations, including the governments of 13 autonomous communities and of the autonomous city of Ceuta, four local governments and 13 non-governmental organisations (for more details, see MISSM, 2024).

A unique feature of the Lab is the collaboration between the GSI and two scientific institutions, namely CEMFI and J-PAL Europe. CEMFI is the Bank of Spain’s foundation and a leader in research and postgraduate training in economics in Europe. The Abdul Latif Jameel Poverty Action Lab (J-PAL) is a global research organisation whose goal is to reduce poverty by ensuring that policies are based on scientific evidence. J-PAL co-founders, Abhijit Banerjee and Esther Duflo, were awarded the 2019 Nobel Prize in Economics for their pioneering approach to alleviating global poverty. J-PAL work is anchored by a global network of more than 900 professors who have completed more than 2,300 randomised evaluations in 121 countries, and at least 850 million people have benefitted from programmes based on evaluations conducted by J-PAL researchers.

The Lab’s scientific team was led by Samuel Bentolila (CEMFI) and Mònica Martínez-Bravo (originally at CEMFI) together with Ana García-Hernández (J-PAL Europe). More than 18 researchers from eight Spanish and foreign institutions have provided scientific support to ensure methodological rigour. The team has cooperated with the GSI throughout the project lifecycle, from design to implementation, evaluation and dissemination.

A total of 32 evaluations were carried out with an investment of €212 million, benefitting more than 175,000 people directly and indirectly. The Lab itself serves as a model for the development of evidence-based policymaking. Specific products for evidence use include a detailed evaluation report for each project, an interactive database on the Lab website showing the impact indicators for all projects and a summary report highlighting ten key policy lessons (MISSM, 2024).
 

Although the body of rigorous research on poverty reduction has grown in recent decades, important evidence gaps remain, particularly in Europe and in the area of social inclusion. When policymakers must act without solid local evidence, decisions are often made based on intuition. But intuition is often wrong and based on ideological positions, and general principles, even when correct, are too vague to inform policy. Well-designed randomised evaluations can test specific hypotheses about policies and estimate their causal impact. They also allow us to rigorously test individual elements of a policy and provide insight into the mechanisms that lead to change.

Attempting to estimate a programme’s impact simply by comparing the results of a group of participants with those of a group of non-participating individuals runs the risk of introducing bias into the impact estimation. Although the observable characteristics of the people in the sample can be taken into account when estimating the effect of a treatment, it could still be incorrectly estimated if unobserved characteristics that affect the outcome of interest are not balanced between the two groups. For example, people enrolled in an educational programme may be more motivated to learn than those who are not. Randomly assigning people in a sample to a treatment group (those to whom the policy in question is applied) and a comparison group (those to whom it is not) is a powerful tool for estimating the causal impact of policies. Random assignment helps to ensure that there is no systematic bias in the composition of the two groups in terms of observed or unobserved characteristics. The differences between the two groups observed after the treatment can therefore be attributed to the causal impact of the programme or policy. 

Even when utilising rigorous methods of evaluation, a policy lab such as the one launched in Spain often faces a number of challenges. First, it is necessary to adapt to political cycles and dynamics, as changing political environments often require swift action, while rigorous evaluation takes more time. Second, it is necessary to collaborate with high-level researchers to ensure the quality of the analysis. Despite the methodological robustness of randomised evaluations, they present a number of challenges in practice that require sound scientific advice through the evaluation cycle. A third requirement is to align the timelines, priorities and work dynamics of academics and policymakers, which can easily become misaligned. Fourth, the model’s sustainability requires securing a structural budget and internal administrative capacity. Finally, it is essential to ensure researchers’ access to individual administrative data, as these can be very useful in addressing limitations in field data collection and allow for the assessment of long-term impact without the need for costly follow-up surveys.

The success in managing these challenges at the Inclusion Policy Lab was due to a number of enabling factors. These included strong political will to generate rigorous evidence and the availability of sufficient resources, the intense participation of researchers affiliated with CEMFI, J-PAL and other institutions, and investment in capacity of public sector human resources. The last point played a key role in improving the trust of policymakers in the methodology and their capacity to incorporate the results into their policy agenda. Finally, a key condition is to partner with implementing organisations that are familiar with the relevant context and target populations, especially when working with vulnerable communities.

The Lab’s evaluations cover key social dimensions: work and entrepreneurship (14 programmes), social support and social benefit take-up (13), education (9), digitisation (7) and housing (1). Key target populations among people at risk of poverty or social exclusion include children and young people, single mothers, the unemployed, migrants, the Roma population and people with disabilities. Since the results of all projects (see MISSM, 2024) cannot adequately be presented, this section highlights a couple of them.

Roma communities across Europe face long-standing disadvantages in education, labour market participation and social inclusion. Roma youth in Spain consistently show significantly lower educational attainment and higher dropout rates, with poverty rates exceeding 80% and unemployment rates over 50%. Nevertheless, many economic studies highlight that education is a powerful tool for reducing inequality and improving economic prospects.

A randomised evaluation tested a programme implemented by the Fundación Secretariado Gitano across multiple regions in Spain designed to bridge the educational gap for Roma youth and, more broadly, to promote social inclusion among the Roma community. The programme combined remedial education with personalised social support for families and a digital skills component. The educational intervention combined remedial group classes conducted outside the school with personalised educational counselling and community-level interventions.

A total of 399 families were randomly assigned to either a treatment or control group. Families in the treatment group received a comprehensive package of educational and social services, whereas those in the control group did not receive any intervention during the study period. The programme was carried out over the course of one school year.

The results reveal that the programme led to substantial gains in academic performance, average school grades improved by 0.11 standard deviations, standardised test scores improved by 0.88 standard deviations – a magnitude considered large in the education literature. The results broken down by subject show significant effects in language-related subjects, with an increase of 0.15 standard deviations in school grades in both Spanish and English. In standardised tests, treated students scored 0.58 and 0.89 standard deviations higher in reading and mathematics, respectively.



The intervention significantly raised students’ educational aspirations, although it had no detectable effect on overall motivation and had a small negative impact perceived teacher expectations, self-efficacy, and peer social relations. This may reflect increased self-awareness and salience of perceptions about teachers, peers and themselves, or more truthful and self-critical assessments due to a heightened perception of scrutiny among treated students. Notably, the programme substantially boosted parental engagement with their children’s education. 

People with intellectual and developmental disabilities (PIDs) have much lower rates of labour force participation than the general population, as well as lower social inclusion and well-being. To test the effectiveness of an innovative support methodology aimed  at helping PIDs find work, the GSI partnered with an NGO (Confederación Plena inclusión España) and a research institute (IDOCAL).

The evaluation compared two approaches. The first policy was customised employment, a method developed by the US Department of Education in 2001. In this method, a social worker supports the PID in four sequential phases. (1) Discovery: exploring the PID’s strengths, preferences and needs by involving their family. (2) Planning: identifying potential employers. (3) Negotiation: arranging interviews between the PID and potential employers as well as developing a job creation proposal. (4) Support: maintaining ongoing contact with the PID and the employer if a job is created. The second policy was the standard supported employment, a less intensive, comprehensive approach that aims to equip PIDs with certain skills to meet the demands of the labour market and only seeks existing job offers.

The intervention was implemented by 44 entities associated with Confederación Plena inclusión España located in 12 autonomous communities and the autonomous city of Ceuta. It included 512 PIDs, aged between 20 and 52, with a degree of disability between 33% and 78% who were either unemployed or working less than 20 hours per week. They were randomly assigned to one of the two policies. The intervention took place between October 2022 and July 2023.

The results show that at the end of the intervention period, participants in the customised employment group were 7 percentage points (pp) more likely to be employed and worked 1.4 hours more. However, these effects dissipated four months later, highlighting the need for ongoing institutional support to maintain gains. They also experienced a 0.11 increase in the number of contracts and a 10 pp increase in their participation in training over a 12-month period, and these effects are even larger four months later (see Bentolila & Domenella, 2025).



Overall social inclusion and well-being indices compiled from several survey questions show large increases and the likelihood of participating in volunteer activities during the past 12 months increases by 20 pp, with all these impacts remaining over time.

In 2021, the Ministry of Inclusion, Social Security and Migration launched 32 pilot projects to promote social inclusion as a complement to the minimum living income scheme. A total of €212 million was allocated to the Inclusion Policy Lab projects,benefiting more than 175,000 people. The policies addressed issues related to work and entrepreneurship, social support and IMV take-up, education, digitalisation, and housing and energy poverty. They were implemented throughout Spain, in collaboration with autonomous community governments, city councils and non-governmental organisations.

The projects were developed as randomised evaluations. They were implemented and evaluated in cooperation between the MISSM and a team of academic researchers coordinated by CEMFI and J-PAL Europe. The results of many of the projects form the basis of a series of scientific articles that are currently underway.

This is a pioneering initiative due to its ambition and scale, the collaboration between three levels of government and NGOs, and its adherence to rigorous evaluation methods through a partnership with academic researchers. The Lab has provided rich insights on the effectiveness of a broad range of inclusion policies, contributed to the institutionalisation of evidence-based policymaking in the country and, last but not least, found some significant positive effects on important aspects of the social inclusion and well-being of the participants in most programmes. Six of the projects are being scaled up. In view of these results, it is hoped that this example will lead to a more widespread use of random evaluation of public policies and the use of evidence in Spain.

This article provides an overview of the characteristics, functioning and results of the Inclusion Policy Lab, an innovative initiative to evaluate a large number of projects that implement social inclusion policies in Spain through randomised evaluations. It is part of Spain’s Recovery, Transformation and Resilience Plan, funded by the European Union through NextGenerationEU. The coordination of the research team was funded by CEMFI, J-PAL Europe and the Social Observatory of the ”la Caixa” Foundation.

BENTOLILA, S.; Y. DOMENELLA (2025): ‘Disability and Labor Inclusion: Experimental Evidence from Spain’, CEMFI working paper 2528.

CALLE, M.; A. GARCIA-HERNANDEZ; P. GIANNATTASIO; I. TORRES-ROJAS (2025): ‘Educational Support Program for Roma Families’, working paper.

INDEPENDENT AUTHORITY FOR FISCAL RESPONSIBILITY (2024): ‘Third Opinion. Minimum Income Scheme’, Opinion 2/24, Madrid.

MINISTRY OF INCLUSION, SOCIAL SECURITY AND MIGRATIONS (2024): ‘Levers for Inclusion. Recommendations from the Inclusion Policy Lab based on Scientific Evidence’, Madrid.