Swedish development aid should be evidence-based. Initiatives that work and deliver results are to be promoted, whilst others to be phased out. This is a key element of the Government’s reform of development aid. But how does one base an initiative on evidence – the best currently available knowledge of what works? Report from a conference held in spring 2026.
Foreword by EBA
This report is about “evidence”. It is written against the backdrop of the government’s steering towards more evidence-based Swedish official development assistance. The report draws from, and provides a summarising discussion of, the conversations held during the conference “Evidence Based Aid in a New Era” in Stockholm on 19 February 2026.
The conference was jointly organised by the Ministry for Foreign Affairs and EBA. Participants included international experts and senior staff and administrators from Swedish government agencies. The aim was to strengthen networking, dialogue and learning within and between Swedish actors engaged in Sweden’s official development assistance.
The question of evidence touches on processes of learning and knowledge utilisation that are genuinely complex. Swedish actors, including the Ministry for Foreign Affairs and EBA, need time and opportunity to reflect on how this ecosystem of knowledge production, learning and policy formation functions. How can the system be developed so as to best inform the design and implementation of future aid? The conference was one way for the organisers to contribute to this learning. It is our ambition that this summary should also serve as a basis for reflection and learning. Large parts of the conference were filmed and are available on EBA’s website.
EBA’s remit is to contribute to the development of Sweden’s international aid policy and implementation through evaluation, analysis and communication. It is our intention to continue the dialogue between EBA, the participating organisations and other stakeholders in various ways.
This summary has been written by Mats Hårsmar and Jan Pettersson at EBA’s secretariat. This version was translated from the Swedish original using AI. The authors are solely responsible for the content, including any misunderstandings and misinterpretations.
Stockholm, June 2026
Jan Pettersson, Managing Director
Introduction
On 19 February 2026, the Ministry for Foreign Affairs and EBA jointly hosted a full-day conference on evidence and effectiveness in aid. The background was the government’s stated ambition that Swedish aid should be grounded in evidence – research, evaluation and proven experience – together with EBA’s remit to contribute knowledge in this area.
The conference, “Evidence-based aid in a new era”, brought together approximately 120 specially invited individuals from government agencies and state-owned companies whose activities are financed through Sweden’s official development assistance. The conference was not focused on individual interventions or sectors but remained at a strategic and conceptual level.
The overarching aim was to strengthen ongoing dialogue and networking between Swedish public-sector actors on effective development cooperation. A number of inspirational and knowledge-sharing lectures and panel discussions were conducted by and with some ten international experts (the programme is included as an appendix).
This working paper summarises key messages and lessons from the conference and places them in a broader context. In addition to serving as documentation, the report is intended to provide a basis for continued dialogue between EBA, the participating organisations and other stakeholders.
What is evidence?
A report addressing a concept must begin by clarifying what the concept denotes. Evidence derives from the Latin evidentia, meaning “obviousness” or “clearness”, something that is clear or apparent. In everyday language as well as in academia, the concept of evidence is used to mean that there is support for something, that a phenomenon or a relationship has been clearly demonstrated.
How, then, can we think about what “evidence” is and what kind of evidence is needed? Interpretations of evidence differ, and reference is sometimes made to “weak” or “strong” evidence. The dividing lines involve both epistemological perspectives and methodological approaches. One strand places great emphasis on the use of specific methods for producing knowledge, aiming to establish causal relationships as precisely as possible – to demonstrate aid’s specific contribution to an outcome. Other analysts argue that most situations which aid seeks to influence must be analysed using a broader range of methods, and that these methods too can demonstrate causal relationships. There are thus epistemological differences between different actors – that is, varying views on how human beings can obtain knowledge about the world around them. A shared view, however, is that the concept of evidence refers to the best knowledge currently available.
During the conference, EBA’s Chair Torbjörn Becker emphasised that EBA uses the concept of evidence to mean our best current understanding – from evaluation, research and proven experience – about what we have good reason to believe constitutes well-functioning aid that leads to intended results. It is therefore not solely about effects or cost-effectiveness, even though such knowledge, through impact evaluations for example, is in many cases of great importance.[1]
EBA’s starting point is not to exclude or take sides between scientific disciplines and their often differing views on method and scientific rigour.[2] The starting point is to avoid methodological reductionism (the notion that quality in research or evaluation coincides with a specific method). For EBA, questions come first and the choice of scientific method follows. Sometimes the questions lead to quantitatively oriented or experimental methods being used in our studies, and sometimes to more qualitatively oriented methods such as comparative analyses, ethnography or case studies. Regardless of approach, however, the requirement is that the method is applied with rigour in the pursuit of new knowledge.
Evidence for whom and for what?
Having defined what evidence is, the natural questions become: who should be responsible for producing what type of evidence, how can evidence be used, by whom, and in what contexts?
How strongly should evidence drive decisions? The Minister for International Development Cooperation and Foreign Trade Benjamin Dousa opened the seminar by emphasising that the question of evidence is one of the most important elements of the government’s reform of aid. He highlighted the importance of distinguishing between, on the one hand, questions that need to be resolved politically – such as the choice of partner countries – and on the other hand, aid’s concrete interventions. The latter should be guided by the methods that work best. This becomes even more important as aid funding decreases. Things that work less well should be abandoned, whilst those that deliver better results should be scaled up.
The seminar as a whole clearly demonstrated that aid is expected to contribute to development at various levels. Contributing to economic growth at the national level presupposes long-term influence on fundamental institutions and governance. At the same time, aid is also expected to deliver more immediate results in concrete, local programmes and projects. Knowledge requirements differ for different societal levels or sectors.
To ground something in evidence is an active act – deliberately finding the currently best available basis for action. The opposite of grounding in evidence is to act arbitrarily, to act on the basis of loose assumptions, ideological convictions or untested practice.[3] At the same time, there must be room for experimentation and innovation, and above all, politics in a democratic society must be allowed to be values-based rather than a technocratic expert rule. This is not the same as lacking knowledge, but it is possible to find arguments for certain decision-making to be evidence-informed rather than evidence-based or evidence-governed. Being able to account for what decisions are based on is a matter of transparency, not least towards citizens.[4]
Evidence can be produced both for learning and for accountability. The production of evidence itself tends to occur ex post. An event has been observed, interpreted and explained, and this new knowledge can help to establish who can be held accountable for what happened. More pertinent to this report is how this new knowledge is used to get things right in the future – as a source of learning and improvement.
A recurring conclusion at the conference was that a lack of use of evidence is often a greater problem than a lack of access to evidence. This ‘political economy of evidence use’ involves a range of perspectives, from incentives to accessibility, knowledge and mandate.[5]
Different levels within ‘the system’ need knowledge of different kinds. Evaluation and research of individual interventions lead over time to accumulated (synthesised) knowledge about types of interventions or sectors. This accumulated knowledge must then be translated to a level that becomes relevant for political priorities and the formulation of policies (or strategies, which are one of the principal steering instruments in Swedish aid). Policies must then be translated again so that the individual agency (for example Sida or the Folke Bernadotte Academy) can support activities that function effectively in relevant contexts. There are thus two pathways of ‘policy translation’ in need of intermediaries, or ‘knowledge brokers’ if you will. But there is also a need for knowledge produced directly for each level in the system, in the form of evaluation of policies and governance.
Processes of learning and knowledge use are genuinely complex. People need to be given time and opportunity to reflect on and discuss how this ecosystem of knowledge production, learning and policy formation should develop in order to best ensure that available evidence is used in shaping future aid. This conference was an attempt to contribute to such reflection.
The following sections summarise the main content of the conference according to a conceptual logic that moves from broad development perspectives, to country examples, followed by a discussion of impact evaluation, the use of evidence for the strategic direction of aid, and questions about how evidence work should be organised.
The Role of Institutions in Development
This section summarises the contribution from Professor Simon Johnson. What do we believe we know, at a general level, about effective development strategies?
Socio-economic development has several distinct dimensions, which ultimately concerns how and in what way a good or well-functioning society can emerge. What characterises ‘a good society’ is in many respects value-driven and ideologically grounded. One cannot therefore hope for consensus on such questions within the aid community.
On the basis of what can be described as a strong mainstream view – that economic growth is a necessary precondition for good societal development – it is possible to reason about how aid can contribute to economic growth. This does not imply that economic growth is a sufficient precondition for socio-economic development.
Simon Johnson, Professor at MIT and recipient of the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel in 2024, spoke under the title “Why Are Not All Countries Rich?” Professor Johnson noted at the outset that aid can be very successful in some cases but sometimes fails to have its intended effects. According to him, a country’s long-term economic development depends to a large extent on whether the country has ‘inclusive institutions’.
Since the 1940s, the differences in life expectancy between rich and poor countries have narrowed considerably. This is due to the countries themselves having invested in health, aided by international assistance. With regard to economic development, we do not see the same convergence. GDP per capita between rich and poor countries has not drawn closer in the same way. Why is this?
Johnson’s explanation concerns the interaction between economic development and a country’s type of institutions – the formal and informal norms, rules and laws that govern how people in a society behave towards one another in economic and political life. The two extremes are termed inclusive institutions and extractive institutions respectively. Inclusive institutions enable a greater number of the population to participate in, and benefit from, economic (and political) life. Extractive institutions instead primarily benefit an elite and restrict the economic and political rights of the majority. Johnson points to a very strong positive correlation between inclusive institutions and economic development.
Why, then, becomes the natural follow-up question, do not all countries choose to develop their institutions in a more inclusive direction? Johnson argues that institutions are heavily path dependent. They are slow-moving and resistant to change, and countries therefore tend to retain the same type of institutions over long periods of time. Institutions also influence how the fruits of economic development are distributed. Technological advances create both winners and losers. Johnson gave as an example from the United States how the mechanisation of the cotton industry led to the spread of slavery, both geographically and in numbers, between 1790 and 1860. At that time, according to Johnson, the American South had some of history’s most extractive institutions directed against African Americans (who in most southern states constituted around 50 per cent of the population). Technological development there led to a reinforcement of local extractive institutions even as the United States overall was characterised by relatively inclusive institutions. This is an example of how a country can grow economically on average whilst certain groups are actively excluded.
What, then, is required for prosperity to be shared by the majority? Technological development can help, if it creates more jobs requiring skilled labour. In countries with inclusive institutions, people can protest when created wealth is distributed unequally. This usually leads to conflicts that must be resolved, typically through a more equitable distribution. In autocratic societies with extractive institutions, such voices can instead largely be suppressed. Today’s technological development, not least in relation to AI, is instead leading to greater automation with lower demand for skilled labour. This is pushing wages down and causing income to be distributed more unequally than before.
In countries with extractive institutions, one must, according to Johnson, have realistic expectations about the extent to which aid can lead to inclusive economic development. Aid in these countries has limited capacity to contribute to major changes, even if the modest influence may be beneficial in itself.
What does Johnson say about countries like China, which have grown economically despite authoritarian rule? Johnson argues that the combination of politically extractive and economically more inclusive institutions has historically not proved to be a sustainable combination over time. He is therefore sceptical about the prospect of China managing a transition to more politically inclusive institutions.
The relationships Johnson points to are, in other words, long-term. The slow-moving institutions he has studied cannot explain short-term fluctuations in economic growth. Most poor countries exhibit periods of ‘stop-go’ growth – rapid growth followed by downturns and often negative growth. Complementary explanations are therefore needed concerning what drives economic growth in the shorter term.
Further reading
Acemoglu, D., S. Johnson, J.A.Robinson (2025) ”Varför är inte hela världen rik?”, Volante Förlag.
Besley, T., R. Burgess, D. Donaldson (2006): “Confronting Global Poverty: The Role of Institutions, Expanding Opportunities and Market Liberalization”, Expert Group for Development Issues (EGDI), Swedish MFA / Almqvist&Wiksell International, Stockholm
Sen, K (2013) The Political Dynamics of Economic Growth, World Development, Volume 47, July 2013, Pages 71-86
Can Aid Strengthen Institutions?
This section summarises the contributions from Professor Njuguna Ndung’u and Professor Tymofiy Mylovanov. What can aid do to strengthen institutions? Experiences from Kenya and Ukraine.
Is the conclusion then that aid should be conditioned on institutional quality? Or should aid work to improve institutions?
Njuguna Ndung’u, Professor at the University of Nairobi and former Central Bank Governor and Finance Minister of Kenya, argued for the latter. He spoke about ‘what aid can and cannot do’. Ndung’u himself had his doctoral training financed through Swedish aid. Drawing on that experience and a number of examples, he argued that individual development can in turn lead to capacity building of and within institutions.
Constraints to development are country and context specific. A common limitation for most poorer countries, however, is weak formal institutions, in both politics and public administration. The laws and regulations that are to be upheld by supervisory authorities also need to be such that they create incentives for development.
Ndung’u argued that aid can play a positive, decisive role in two central areas: building capacity for oversight and contributing to effective regulatory frameworks. He gave three examples of such capacity-strengthening interventions.
The African Economic Research Consortium (AERC) was founded in 1988 with support from a group of development actors: bilateral donors (Swedish Sida, Canadian CIDA, American USAID, British DfID), philanthropic organisations (Ford, Rockefeller, MacArthur) and multilateral institutions (World Bank, African Development Bank). Initially the activities were coordinated by the Canadian International Development Research Centre, IDRC. Through economics research and doctoral training, African public administration has over the years been strengthened at all levels. At one point, a majority of African central bank governors and finance ministers had received their training through the AERC. This has, according to Ndung’u, contributed to better economic policy in several countries.
A second example was the African network for developing countries’ financial systems, the Financial Sector Deepening Trust (FSD Network). The network (with Sida as one of its founders in 2005) works to make countries’ financial institutions and markets more accessible and efficient, and to harmonise regulations between countries.
The FSD made a strong contribution to the regulatory foundation for the third example, the electronic payment system M-Pesa (initially supported by British DfID). This system was the first of its kind in the world. A decade or so later, it was followed by equivalents including Sweden’s Swish.
M-Pesa created fast and verifiable transactions for ordinary people and small businesses, in both the formal and informal sectors. This has helped to create links between sectors in the segmented market structure that characterises Kenya and many poor countries. But the contribution is greater than that. It has come to change people’s way of thinking about money and their own finances. Through the ability to save and access credit, individuals have gained greater financial security, and smaller businesses have found a pathway into the financial system.
According to Ndung’u, this demonstrates how aid-supported capacity building can lead to better financial institutions. A better financial system in turn contributes to economic development and, ultimately, to poverty reduction.
A further argument for institution building – and that institutions need to be built over the long term – came from Ukraine, with Professor Tymofiy Mylovanov, President of the Kyiv School of Economics. His account was very concrete. Despite the harsh wartime winter of 2026, with temperatures some 20 degrees below zero, constant Russian missile and drone attacks and severe energy shortages, his university is the only academic institution in Kyiv to be maintaining in-person teaching. Students who cannot remain in their homes are accommodated in hotels. Classrooms cannot be adequately heated, which has given rise to new methods involving learning during physical activity. Continuing teaching even under these conditions is absolutely crucial, Mylovanov argued, for maintaining the social contract in Ukraine. Part of this contract involves training a new generation of leaders and thereby ensuring the continuity of various societal functions over time.
The necessary adaptations to the university’s teaching are made possible with the support of a large international network that has been built up over decades, and this network’s assistance is essential in planning for the next and subsequent cold wartime winters.
Aid is often provided solely to address the acute crisis situation. One of the most important lessons from Ukraine in this regard, Mylovanov argued, is the critical need for long-term support of resilience and adaptability. You cannot build the boat while sailing it.
Further reading
Ndung’u (2021), “A Digital Financial Services Revolution in Kenya: The M-Pesa Case Study, African Economic Research Consortium” (ISBN: 978 9966 61 112 3)
Ndung’u (2019), “Digital Technology and State Capacity in Kenya”, CGD Policy Paper 154, August 2019; Center for Global Development, Washington D.C.
Evidence from Impact Evaluation
This section summarises the discussion between Martina Björkman-Nyqvist and Jos Vaessen. How can impact evaluation contribute to more effective development aid?
Methodological questions sometimes appear to be the preserve of the initiated – but they are of great importance. How do we find out what is likely to work, when, for whom, and under what circumstances? How these questions are answered is largely determined by the methods used to generate knowledge.
At the outset, it is worth noting that certain terms are both used and understood in different ways.[6] Impact evaluation is used in this section to refer to evaluations of an intervention’s causal effect. What is being affected can be different types of results: outcomes (short- and medium-term effects), impact (long-term, transformational effects), but also an intervention’s outputs.
There are different methods for impact evaluation. Some (such as EBA) use terms such as quantitative, qualitative and theory-based impact evaluation. Others use the term impact evaluation solely for quantitative studies. In this discussion, ‘Impact Evaluation in a Changing Aid Landscape’, the latter definition was used. Qualitative methods were instead described as ‘rigorous methods for causal claims’.
Martina Björkman-Nyqvist, Professor at the Stockholm School of Economics, emphasised the importance of robust impact evaluation from an accountability perspective – that is, answering the question of whether interventions achieve what they set out to achieve. But she also highlighted how knowledge about interventions’ effects should be used for planning new interventions. She highlighted experimental approaches as strong and rigorous, primarily because they study the counterfactual outcome: what has been achieved in comparison with what would have happened had the intervention not taken place? With such approaches, it is possible to estimate how much difference an intervention actually makes, particularly in situations where randomised controlled trials (RCTs) can be conducted.
The logic behind experimental methods is, in brief, that populations can be randomly divided into different groups. Since the groups are randomly divided, they are on average similar. When one group is reached by an intervention (‘is treated’) and another is not (a control group, which in some situations may receive a placebo), measured differences between the groups can be attributed to the treatment (see Olofsgård, 2014, for a non-technical overview).
Björkman-Nyqvist also pointed to the usefulness of ‘natural experiments’ or ‘quasi-experimental’ methods when active random assignment cannot be used. This opens up more areas of application. In brief, natural experiments involve something that ‘happens to occur’ in some situations but not in others that are otherwise similar. For example, locust swarms may land randomly in specific places, or human errors may occur randomly – creating natural comparison groups. In quasi-experiments, comparison groups are instead created deliberately, for example by identifying groups that were similar before a reform was implemented.
Björkman-Nyqvist emphasised that these methods cannot be applied to all aid questions, but that there is great value in organisations such as Sida supporting more such studies. There is also much to learn from the many studies that have already been conducted.
Professor Jos Vaessen, Chief Evaluation Officer in the World Bank’s Independent Evaluation Group (IEG), highlighted the importance of a varied approach. Different methods have their advantages and disadvantages and are suited to different situations. The central question becomes which method should be used for which type of situation or question. He stressed the importance of rigour both in the choice of method and in the design of impact evaluations (ensuring that it is actually possible to analyse causal relationships) and in their implementation (data collection, analysis, etc.). Rigorously answering a question often requires using several methods in the same evaluation.[7]
How can and should organisations such as Sida make use of these methods? First and foremost, argued Björkman-Nyqvist, aid organisations themselves should not to any significant extent conduct their own randomised controlled trials. These are costly and complicated and should be carried out (by researchers) where there are knowledge gaps or where there is reason to revise what is believed to work. The most important thing is to use knowledge from already completed studies to guide the choice of interventions.
Björkman-Nyqvist gave two examples of learning from RCTs. The view of microfinance has over time shifted from glorified to sobered. This is the result of impact evaluations finding a limited and context-dependent effect of microfinance on poverty. Such interventions are today less common and better designed. Another example is pupils’ learning, where textbooks alone have not been shown to have a positive effect on learning, whilst deworming interventions (as an unexpected effect of a health intervention) have in some situations contributed to improved pupil outcomes.
Vaessen gave an example from Malawi during Covid-19, where cash support was provided to families. Rumours arose that the support was leading to increased domestic violence against women. Theoretically, cash support can lead to both increased and decreased gender-based violence. How could one know whether the rumour was true? In this case there was no baseline study available, but by (rigorously) conducting a large cross-sectional study, a current picture was created. Thereafter it was possible, through (rigorously conducted) in-depth interviews with women, to understand the causes of gender-based violence in the home, and the evaluation found that the support programme actually led to a reduction in domestic violence. By using two different complementary methods, it was thus possible in this case to substantiate a causal claim (or rather to dispel a rumour) about the intervention’s effects.
Further reading
Befani, B. (2016): “Pathways to change: Evaluating development interventions with Qualitative Comparative Analysis (QCA)”, Expertgruppen för biståndsanalys, EBA 2016:05
Dahlgren, S., L. Johansson de Château, J. Molander, S. Olander, J. Pettersson (2024), “Engelsk-svensk ordlista över nyckeltermer inom biståndsutvärdering” (“English-Swedish glossary of key terms in aid evaluation”), Underlagsrapport oktober 2024, Expertgruppen för biståndsanalys.
Olofsgård, A. (2014): “Randomized controlled trials: strengths, weaknesses and policy relevance”, Expertgruppen för biståndsanalys, EBA 2014:01.
Vaessens, J., S. Lemire and B. Befani (2020) “Evaluation of International Development Interventions: An Overview of Approaches and Methods”, Independent Evaluation Group. Washington, DC: World Bank.
What to Support with Reduced Funding?
This section summarises the contribution from Professor Rachel Glennerster. What strategy can donors adopt to conduct effective development cooperation under shrinking resources and increasing needs?
Impact evaluations can thus be of great use in development cooperation. Professor Glennerster based her contribution largely on evidence from impact evaluations of the kind described in the previous section.
At the outset, Glennerster noted that the title ‘aid in a new era’ alludes in part to the new geopolitical landscape, including declining aid volumes, but also to the fact that a new context has been created because development and aid have been a ‘spectacular success’, through reduced extreme poverty, improved health outcomes and the large reduction in the number of low-income countries. Even though Professor Johnson emphasised that poor countries have not converged economically with the rich, many of them have nonetheless managed to advance from low- to middle-income countries. Those development successes should also influence how one works with aid, argued Glennerster. Most extremely poor people now live in middle-income countries, which themselves have both the responsibility and the economic means to reduce poverty. Evidence-based cooperation with those countries concerns, according to her, prioritising trade and innovation through knowledge transfer for more effective use of their own resources, rather than continuing grant aid.
Glennerster gave examples from India and Indonesia of how donor-funded impact evaluations have resulted in increased cost-effectiveness in national poverty programmes of up to 25 per cent.
As the poorest countries are in many respects in a worse situation than before, grant aid should be reserved for low-income countries. Professor Glennerster also argued in these cases for a radical simplification of interventions. Donor organisations need not only to specialise within fewer areas of expertise, but the interventions themselves should be made simpler (they were described as sometimes resembling Christmas trees) and larger in scale. This leads to greater cost-effectiveness.
She emphasised that this type of evidence is not confined to sectors that allows for simple divisions into groups of ‘treated’ and ‘controls’ (such as health and education) but also, for example, in democracy support. She also pointed out that it is a misconception to think that impact evaluations should be conducted on every intervention – this is not the right way to think about the production of evidence. Knowledge is generated through solid research and high-quality impact evaluation of certain interventions. Based on the learning from these, effective aid can then be designed. In this way, good production of evidence can be paired with good use of the same.
In her presentation, Glennerster pointed to three questions where different types of evidence and knowledge are needed: What development problem needs to be solved? What do effective measures look like? What concrete interventions are needed in a given context? We return to this framework later.
Further reading
J-PAL: Evaluations | The Abdul Latif Jameel Poverty Action Lab,
IPA: Innovations for Poverty Action
World Bank, Development Impact Evaluation Data Catalog
Evidence in Crisis and War
Aid to Ukraine was, understandably, an important part of the discussions during the conference. In situations of crisis and war, the question of evidence is brought into sharp focus. The situation changes constantly, and new decisions must be made continuously. This requires that expertise has been built up in advance, and that it is both present and operating locally.
Torbjörn Becker emphasised the importance of long-term commitment. It would not have been possible to support higher education in wartime Ukraine had the collaboration with educational actors in Sweden not been established long in advance. There is a need for interplay between more acute support and longer-term development aid. It is not always possible to foresee when and how what is being done will prove significant, but in general, argued Becker, there are good reasons to support capacity building, education and research collaboration.
The capacity question also has a spatial dimension. The capacity that is built must be local. Those who live closest to the problems also generally have the closest access to solutions, highlighted Megan Kennedy-Chouane, head of the secretariat for the OECD-DAC evaluation network, EvalNet. Decisions must be taken by people with an in-depth knowledge of their country. Those directly affected by aid interventions financed by Sweden also need to be involved in the work of grounding those interventions in evidence. Becker too emphasised the importance of Ukraine’s priorities being decisive in determining what support is given and how.
From previous situations of aid in conflict settings, OECD evaluations and studies show several lessons that can be drawn for Ukraine as well, even though each situation has its unique characteristics, Kennedy-Chouane noted. Becker highlighted several features that distinguish Ukraine: it is a European middle-income country under attack with a direct impact on donors’ own security, where military aid is many times greater than humanitarian and development aid. Kennedy-Chouane argued, however, that the assumption of Ukraine’s unique situation has led to a failure to learn from a number of general insights. Some of these are as follows:
In conflict situations, donors tend to try to emphasise their visibility. Whilst demonstrated solidarity is important, and can itself have positive effects, and whilst cost-effective solutions may sometimes legitimately take second place, there is a risk that support thereby becomes less needs based.
When an armed conflict breaks out, donors rarely take a realistic view that support will be needed for a very long period of time. Refugees, for example, will in all likelihood be displaced for decades. Yet instead of proceeding step by step on the basis of a long-term plan, support is rushed.
There are enormous problems with incentives when there are large amounts of money to be spent in a short time. The capacity to absorb resources is often limited. At the same time, a great deal is invested in short-term activities with quantitative targets for disbursement rather than long-term effects, and the main purpose of the support risks being lost.
The effects of aid differ for different groups and different parts of society, and external involvement will always have unintended consequences. Support can also cause harm. It must be implemented with great conflict sensitivity. Sometimes it may be best to refrain from providing support.
Examples of unintended consequences include certain types of support – such as large quantities of clothing – leading to border crossings becoming overloaded, meaning that more critical materiel, weapons, ammunition and essential supplies, took longer to come through, as Tymofiy Mylovanov recounted. Indirectly, this led to people dying, at least during the early years of the war.
At the same time, he argued that it is important to get people in other countries to feel that they are contributing. Relationships and bonds must be built for support to continue. Ukrainians care about those who provide support; foreign taxpayers must feel that their support matters, Mylovanov underlined.
He also emphasised that it is more important to get things done than to get them done in exactly the right way. There is a ‘cost of emergency’ that cannot be avoided. Mylovanov referred to NATO representatives who, drawing on experience from Afghanistan and Iraq, argue that in the acute phase of a conflict situation, a certain degree of embezzlement must simply be accepted. Corruption exists not only, or even primarily, on the recipient side but arises among donors and in supply chains. Substandard quality is delivered, deliveries are substituted or disappear, but the important thing is to see the bigger picture and be highly pragmatic. Until the war is over, when tolerance can be reduced and leakages plugged, one must ensure that things happen quickly and that the overarching objectives are met.
Torbjörn Becker confirmed, in response to a direct question, that as far as he is aware there is not much research into how this type of leakage can be plugged. What can be done is improved coordination between donors, maintaining good control systems and whatever else belongs to good donorship. There are no simple solutions to these problems.
The challenge of being pragmatic also includes being able to adapt to changing circumstances. For example, rather than sending older types of weapons, the international community should focus on building the capacity to produce new types of weapons suited to modern warfare, both in Ukraine and abroad.
Inefficiencies exist in the conflict itself as well as in the donor system. Mylovanov gave examples of how the donor system sometimes operates inefficiently, such as being forced to improvise around the procurement of body armour because procurement procedures took too long, or the difficulty of protecting the education system when it took years for UN bodies and civil society organisations to secure funding.
But how can pragmatic and flexible action be combined with the need for long-term commitment? An effective aid instrument in such situations is budget support, according to Kennedy-Chouane. Budget support has proved particularly effective when governments use it both to act flexibly and simultaneously to ensure that longer-term reform programmes continue to be driven forward.
Mylovanov highlighted two other central requirements for managing the combination of flexibility and long-term commitment:
The first is establishing arenas and channels of communication between Ukraine and taxpayers in Sweden. Relationships and connections are needed in order to sustain solidarity and the will to provide support. The importance of enabling taxpayers in donor countries to feel and understand that their support makes a difference is absolutely essential for the long-term support that is needed.
The second is recognising what happens during a typical fourth year, a fourth winter, of a war. Increasing numbers begin to grow accustomed to people dying, and to the war simply continuing. The situation becomes normalised. At that point, systems and organisations tend to become bureaucratised whilst at the same time there remain great needs to find new, more effective solutions. There is therefore a great need for donor organisations to continue to experiment on a smaller scale, in order to then scale up what works.
Further reading
Becker, T., A. Olofsgård, M. Perrotta Berlin, J. Roine (2025), ”Svenskt Ukrainastöd i en internationell kontext: offentligfinansiella effekter och framtidsscenarier” (“Swedish Ukraine Support in an International Context: fiscal effects and future scenarios”), Rapport till Finanspolitiska rådet, 2025:01.
OECD (2024), “Report on the implementation, dissemination and continued relevance of the DAC Recommendation on the HDP Nexus”, DCD/DAC/INCAF(2023)1/FINAL, OECD, Paris.
SIGAR (2021), “The Risk of Doing the wrong things perfectly: Monitoring and Evaluation of Reconstruction Contracting in Afghanistan”, Special Inspector General for Afghanistan Reconstruction, July 2021, Arlington Virginia, USA.
Zürcher, C. (2022), “Evidence on aid (in)effectiveness in highly fragile states: A synthesis of three systematic reviews of aid to Afghanistan, Mali, and South Sudan, 2008-21”, WIDER Working Paper Series wp-2022-160, World Institute for Development Economic Research (UNU-WIDER).
Evidence in Practice
How evidence is actually used at different levels of the aid system – from overarching priorities to practical implementation – is largely a matter of how the work is organised. How can actors be encouraged to make greater use of the knowledge that already exists?
During the panel discussion entitled ‘The Use of Evidence for Steering and Managing for Results’, Jillian Popkins, Chief Commissioner of the British Independent Commission for Aid Impact (ICAI), Professor Jörg Faust, Director of the German Institute for Development Evaluation (DEval), and Professor Rachel Glennerster discussed opportunities and barriers for policymakers and practitioners to increasingly absorb and use existing evidence.[8]
Both Popkins and Faust stressed the importance of impartial evaluation for producing knowledge. Impartiality needs to be actively safeguarded by both evaluators and commissioners. Three areas were identified as important for how an organisation can produce and communicate evidence.
The first concerns the mandate of the evaluating organisation. In both the British and German cases, the evaluation bodies report to both government and parliament, meaning that governments can be held to account. DEval also has status as a federal research institute, enabling it to work on methodology development. It also has a remit to work on institution building in the field of evaluation in partner countries. According to Faust, it is crucial to strengthen partner countries’ own capacity to evaluate, in order to create ownership over the production and use of evidence in the countries where aid should make a difference.
Another area concerns the funding of the evaluating organisation. If it is financed by government, clear firewalls for independence need to be built, argued Faust.
A third area is the reporting and communication of evaluation findings. All agreed that there is an enormous amount of evidence available on aid in the form of RCT studies, meta-evaluations and knowledge syntheses. At the same time, Glennerster argued, it is important to identify the (significant) evidence gaps that exist in important areas. The question is therefore not only about using existing knowledge, even though this is a central problem in both the design and implementation of aid.
The discussion about how politicians and practitioners can better use evidence centred primarily on incentives and time constraints. Different recipients have different needs and sometimes conflicting demands. Popkins argued that those who produce evidence must adapt their communication to different audiences. Different users have different objectives and therefore different needs. It is about finding the relevance for the intended user and communicating accordingly. To succeed, one must understand what government, parliament and administration need the evidence for. This applies not only to evaluation but also to research. As EBA’s Chair Torbjörn Becker noted in his opening remarks, the bridge between research and practice is often one-way. Research needs to learn more from the conditions of policy and practice in order to produce relevant knowledge.
Faust argued that the logic needs to be reversed. Evaluators tend to focus on disseminating knowledge. It is more important to create incentives for users to demand evidence. Glennerster shared her experiences as Chief Economist of the British aid agency DfID. Incentives for actively acquiring knowledge were created through requirements on projects and programmes to refer to existing evidence in the area. Projects deemed insufficiently grounded in evidence were returned for revision. Staff and managers who wanted to get projects approved were thus incentivised to refer actively to available evidence. Faust argued that clear follow-up on whether recommendations are acted upon or commitments fulfilled is an important tool for assessing after the event whether and how evaluations are used.
Different phases of development cooperation require different types of evidence. Use can be structured into different steps, argued Glennerster:
The first step is to clarify which development problem needs to be solved. This is an analytical question where the primary need is for descriptive data and information about relevant factual conditions.
Once the problem is well defined, knowledge is needed about which instruments and interventions work to address and resolve it. This requires evidence that clarifies causal relationships. Such evidence can be produced through experimental or theory-based methods. But often relevant existing knowledge is already available.
The third step concerns concrete implementation. In this application, Glennerster noted, there is always a gap between available knowledge and user needs. Interventions can rarely be copied directly; there are never ready-made solutions to be taken off an evidence shelf. Evidence must be translated in relation to the current need. Studies should be interpreted in terms of what they say about human behaviour. In this way they can contribute guiding principles – evidence – for how interventions can be designed, including to steer incentives. Studies within, say, the education sector can thereby become relevant and usable within, say, the health sector. This requires deep knowledge of the countries and situations in which the aid is to be implemented.
Jörg Faust noted that two distinct perspectives on what aid is and how it should be conducted had crystallised during the conference. One perspective focuses on poverty reduction where evidence at the intervention/micro level is expected to drive which interventions are carried out. The other perspective holds that it is inclusive institutions at the national level that determine whether development – and thus poverty reduction – takes place. These institutions are slow-moving, which means that aid should be long-term and engage with political processes and structural questions. The effects of such work are likely to be transformative, but also difficult to measure. How to prioritise between these two – both well-founded – accounts of effective aid in a situation of declining resources is, and perhaps should be, a choice for political decision-makers.
Concluding Remarks
The sessions offered many perspectives and an engaged audience participated with nuancing, challenging questions and additional insights. The animated discussions during the conference lead to several conclusions and observations.
Grounding aid in evidence is difficult. Genuinely so. The question cannot be reduced to a value-free technical exercise. It concerns what should work, for whom and under what conditions. Aid can – and should – be understood from the two different perspectives raised during the conference. The ambition to influence development at the national level can coexist with the ambition to deliver concrete and cost-effective results at the local level for intended target groups. Prioritising between these is not straightforward and partly concerns the question of when aid is supposed to “work”. Systems that depend on human action need to take account of different contexts, incentives, trade-offs, goal hierarchies, and in particular whose perspective is placed at the centre.
The concept of evidence is given different meanings in different contexts, which can create ambiguity. For EBA, the decisive question is not which specific method has been used to produce knowledge, but with what rigour the respective method has been applied. Different types of evidence are needed depending on which question is to be answered. It is helpful to think in terms of which development problem needs to be solved, which types of interventions have been shown to deliver results, and how the local context functions where the intervention is to be implemented. Each of these questions requires different types of knowledge to be answered; the intended user becomes important here – is it politics, policy and strategy formation, strategy implementation, the concrete design of interventions, or the implementation of interventions that is to be grounded in evidence?
Evidence must be interpreted and understood. This requires facilitative and ‘translating’ measures – between evaluation, research and practice – so that relevant knowledge can be applied by the intended user in the relevant context.
Grounding aid in evidence is an active, ongoing process in which the currently best available evidence underpins decisions and actions. But we never know everything, and the evidence base is shifting and incomplete. There is a need to identify what we do not know today, to continually question what we know, and simultaneously to make use of what we currently believe we know. A clear conclusion from the conference is that insufficient use of existing evidence is often a greater problem than the availability of evidence itself.
Grounding decisions and actions in evidence presupposes an organisation that deliberately steers towards this knowledge base. Incentives and decision rules need to promote evidence-grounding. Relevant expertise must be in place and, above all, sufficient time must be able to be allocated to taking relevant decisions and shaping relevant actions based on the specific questions that need to be answered.
A final observation from the conference is the great commitment and strong will among Swedish actors to work for effective evidence-based aid, both in the sense of preserving and developing what works today and in finding new ways to design and implement future aid. EBA’s ambition is to contribute to this work.
References
Acemoglu, D. S. Johnson, J.A. Robinson (2025), ”Varför är inte hela världen rik? Om makt, framsteg och teknik – genom historien och i framtiden”, Volante Förlag.
Becker, T., A. Olofsgård, M. Perrotta Berlin, J. Roine (2025), ”Svenskt Ukrainastöd i en internationell kontext: offentligfinansiella effekter och framtidsscenarier”, Rapport till Finanspolitiska rådet, 2025:01.
Befani, B. (2016), “Pathways to change: Evaluating development interventions with Qualitative Comparative Analysis (QCA)”, Expertgruppen för biståndsanalys, EBA 2016:05
Besley, T., R. Burgess, D. Donaldson (2006), “Confronting Global Poverty: The Role of Institutions, Expanding Opportunities and Market Liberalization”, Expert Group for Development Issues (EGDI), Swedish MFA / Almqvist & Wiksell International, Stockholm.
Dahlgren, S., L. Johansson de Château, J. Molander, S. Olander, J. Pettersson (2024), “Engelsk-svensk ordlista över nyckeltermer inom biståndsutvärdering”, Underlagsrapport oktober 2024, Expertgruppen för biståndsanalys.
EBA (2026), ”Policy samt riktlinjer för säkerställande av kvalitet i studier”, tillgänglig på www.eba.se.
Goldman, I., M. Pabari, (2020), ”Using Evidence in Policy and Practice”, Routledge, Oxford, UK.
Hedlin och Lokatt (2024), ”Det svenska biståndets transparens”, EBA-rapport 2024:04
Kristensson Uggla, B. (2019), ”En strävan efter sanning – vetenskapens teori och praktik”, Studentlitteratur, Lund.
Ndung’u (2021) “A Digital Financial Services Revolution in Kenya: The M-Pesa Case Study”, African Economic Research Consortium (ISBN: 978 9966 61 112 3)
Ndung’u (2019) “Digital Technology and State Capacity in Kenya”, CGD Policy Paper 154, August 2019
OECD (2024): “Report on the implementation, dissemination and continued relevance of the DAC Recommendation on the Humanitarian-Development-Peace Nexus”, DCD/DAC/INCAF (2023)1/FINAL, OECD, Paris.
Olofsgård, A. (2014): “Randomized controlled trials: strengths, weaknesses and policy relevance”, Expertgruppen för biståndsanalys, EBA 2014:01.
Sen, K (2013) The Political Dynamics of Economic Growth, World Development, Volume 47, July 2013, Pages 71-86
SIGAR (2021) “The Risk of Doing the wrong things perfectly: Monitoring and Evaluation of Reconstruction Contracting in Afghanistan”, Special Inspector General for Afghanistan Reconstruction, July 2021, Arlington Virginia, USA.
Vaessens, J., S. Lemire and B. Befani (2020) “Evaluation of International Development Interventions: An Overview of Approaches and Methods”, Independent Evaluation Group. Washington, DC: World Bank.
Zürcher, C. (2022) “Evidence on aid (in)effectiveness in highly fragile states: A synthesis of three systematic reviews of aid to Afghanistan, Mali and South Sudan, 2008-21” https://ideas.repec.org/s/unu/wpaper.html World Institute for Development Economic Research (UNU-WIDER).
Appendix: Conference Programme
Evidence Based Aid in a New Era
19 February 2026
08:00-18:30
Lunden, Ministry for Foreign Affairs of Sweden
Drottninggatan 4, Stockholm
The Ministry for Foreign Affairs and the Expert Group for Aid Studies (EBA) are jointly organising a conference on how evidence-based methods can enhance the effectiveness of Swedish aid. The conference will be led by the Minister for International Development Cooperation, Benjamin Dousa.
Swedish aid will be designed and implemented in accordance with the government’s aid policy. Efforts will be supported by evidence in the form of research, evaluation and proven experience. But how can we develop and use evidence on aid effectiveness? How can evidence and learning be integrated into governance and implementation? What drives poverty reduction and prosperity at a country level?
These questions, and others, will be addressed at the conference, which will bring together decision-makers and change agents from government agencies and state-owned companies involved in Swedish aid (Team Sweden). Delegates will gain knowledge and insights from leading international experts in evaluation, evidence-based policy and aid effectiveness.
The conference aims to establish a foundation for ongoing dialogue and networking within Team Sweden, as well as between Team Sweden and academia and leading experts in evaluation, evidence-based policy, and aid effectiveness. In addition, this will create conditions for continued collaboration, particularly between EBA and relevant organizations regarding the achievement of effective aid-funded work and cooperation.
Agenda
Opens 08:00, Closes 08:50 Registration with breakfast, coffee/tea
Brief introduction on practicalities and an overview of the day’s schedule: Moderator Malin Oud (EBA Vice-chair), Jan Pettersson (EBA Managing Director)
Welcome and opening remarks: Mr Benjamin Dousa, Minister for Trade and Development
What is evidence? Torbjörn Becker, EBA Chair, Director SITE
Aid in a New Era: How to be Strategic. Professor Rachel Glennerster, President Center for Global Development, former Executive Director, J-Pal. Under decreasing global volumes of foreign aid, what strategic policy options for donors can ensure transparency and cost-effective impact?
What aid can and cannot do: Njuguna Ndungú (ex-Minister of Finance, ex-Governor of the Central Bank of Kenya). Insights from Kenya on how aid has complemented financial sector development, private enterprise, trade and broader development processes
10:35 – 11:10 Coffee/tea
Aid in Conflict. Lessons from and for aid to Ukraine. Professor Tymofiy Mylovanov (President Kyiv School of Economics), Torbjörn Becker (chair EBA, director SITE), Megan Kennedy-Chouane (head OECD/DAC-EVALNET). What can aid to Ukraine learn from past experiences, from the Marshall plan to Afghanistan? What does “effectiveness” mean in crisis, and how should evidence be used in conflict settings?
12:10 – 13:30 Lunch
Special Keynote: Why are Not All Countries Rich? (online) Professor Simon H. Johnson (MIT, Nobel Laureate in Economics 2024). On the importance of institutions and economic transformations. Exploring the role of evidence-based policy in shaping such institutions and if (and if so, how) development assistance can assist in these processes.
14:20 – 14:45 Coffee/tea
Impact Evaluations in a Changing Aid Landscape. Professor Martina Björkman Nyqvist (Stockholm School of Economics), Professor Jos Vaessen (Chief Evaluation Officer, Independent Evaluation Group, World Bank). A critical look at how impact evaluations have been conducted and how their role may evolve as aid priorities shift.
The use of Evidence for Steering and Managing for Results. Jillian Popkins (Chief Commissioner, Independent Commission for Aid Impact), Professor Jörg Faust (Director, German Institute for Development Evaluation (Deval), Professor Rachel Glennerster (President Center for Global Development, former Executive Director, J-Pal). To what extent is evidence used to guide aid governance and implementation? What are the barriers and enablers for integrating research and evaluation into decision-making? How can effective systems for results-based management be built?
Concluding reflections: Director-General for International Development Cooperation Ms Camilla Nevstad Bruzelius
16:30 – 18:30 Reception, Blå Salongen
Reflections of the day: Mr Benjamin Dousa, Minister for International Development Cooperation and Foreign Trade
- That evidence cannot be reduced solely to a narrow understanding of cost-effectiveness can be illustrated by the need for redundancy in certain systems, such as military defence or climate adaptation. Redundancy may be seen as an antithesis to cost-effectiveness but can be essential for achieving objectives. ↑
- See for example EBA’s quality policy (EBA 2026). ↑
- Kristensson Uggla, B. (2019), ”En strävan efter sanning – vetenskapens teori och praktik” (”A Pursuit of Truth – Theory and Practice of Science”), Studentlitteratur. ↑
- A more developed argument is provided in Hedlin, P., and C. Lokatt (2024), ”Det svenska biståndets transparens” (”The transparency of Swedish aid”), EBA report 2024:04. ↑
- See e.g. Goldman, I., M. Pabari, (2020), “Using Evidence in Policy and Practice”, Routledge, Oxford, UK, for an overview of factors that support and hinder the use of evidence. ↑
- See also EBA and Sida’s English-Swedish glossary of evaluation terms in aid (Dahlgren et al., 2024). ↑
- Examples of such qualitative methods are ‘qualitative comparative analysis’ (QCA, see Befani, 2016) and ‘process tracing’. ↑
- ICAI and DEval are two of EBA’s four sister organisations within the OECD. The others are Norway’s NOREC-Eval and France’s CEAPD. ↑
Please refer to the present report as: Hårsmar, M., J. Pettersson (2026), Evidence for Aid: Report from a Conference, Working Paper June 2026, The Expert Group for Aid Studies (EBA), Sweden.
Working Paper, June 2026 to The Expert Group for Aid Studies (EBA)
The EBA Working Paper Series constitutes shorter overviews, surveys, mappings and analyses that have been undertaken to bring about discussion and advance knowledge of a particular topic. Working Papers are not subject to any formal approval process by the Expert Group. Just as in the EBA reports, authors are solely responsible for the content, conclusions and recommendations.