This chapter examines how development partners can strengthen their ability and those of partner governments to understand and increase the impact of poverty and inequality reduction interventions. The people most affected by deprivation and disparities are often the least visible in the systems designed to track progress. The chapter starts by considering strengthening poverty and inequality data and research to close these gaps at both global level and in partner countries. It then explores how to embed poverty and inequalities into results frameworks and monitoring systems, as well as how to apply people-centred evaluation approaches that drive learning and accountability. Finally, it considers the institutional structures and cultures that can deliver for those furthest behind.
OECD Guidance on Reducing Poverty and Inequalities through Development Co‑operation
Abstract
Translating commitments to poverty and inequality reduction into impact requires data, research, and monitoring and evaluation systems to track whether interventions are working, for whom and under what conditions. These are not technical add-ons to development co-operation; they are the infrastructure through which commitments, strategies, objectives and action are translated into results and accountability. By centring people facing poverty and inequalities in these systems, development partners can better understand who benefits from interventions; who is left behind and why; and how development co-operation programming interacts with the wider local context and structural drivers of deprivation and exclusion.
The people most affected by deprivation and disparities are often the least visible in the systems designed to track and assess progress. The evidence base for poverty and inequality reduction remains fragile in many low and middle-income countries (L&MICs) where statistical systems are often under-resourced and fragmented. Census undercounts are widespread and systematic, disproportionately affecting marginalised and hard-to-reach populations. Meanwhile, results and monitoring frameworks can oversimplify complex poverty and inequality dynamics, reflecting what is easy to measure rather than what matters most, and masking the distributional differences that lie at the heart of inequality. The ability of evaluation systems to assess transformative and long-term impacts on poverty and inequalities remains limited, particularly where changes involve deep structural barriers, power relations and intersecting forms of disadvantage.
The people most affected by deprivation and disparities are often the least visible in the systems designed to track and assess progress.
Development partners are well-placed to close these gaps by investing in the systems that make poverty and inequality reduction visible, measurable and actionable – and to ensure that the people who are furthest behind are reflected in the evidence that shapes decisions. Strong data and evidence systems can expose the realities of marginalised groups who might be invisible in official statistics, while qualitative and participatory approaches reveal the lived experiences and hidden barriers that shape unequal outcomes. Robust monitoring and evaluation can identify which interventions reduce poverty and narrow inequalities, and which unintentionally reinforce them. And continuous learning and use of evidence allow development partners to adapt to changing contexts, respond to the needs of people experiencing poverty and inequalities, and direct limited resources to where they will have the most impact. Realising this potential requires making deliberate choices: what data to produce and use, and how to fill gaps for the most marginalised; how results frameworks and monitoring systems are designed to track interventions’ impact on multiple and intersecting dimensions of poverty and inequalities across individuals and groups; and how evaluations are conducted and used to generate people-centred learning that improves practice and accountability.
This chapter provides development partners with guidance on how to strengthen data, monitoring, evaluation and learning for enhanced accountability and impact on poverty and inequalities. Section 5.1 examines how development partners can invest in inclusive data, evidence and research systems, strengthening statistical capacity globally and in partner countries, addressing the data exclusion of the most marginalised, and shifting research ownership toward the L&MICs. Sections 5.2 and 5.3 consider development partners' own systems and approaches, with Section 5.2 addressing integrating a poverty and inequality lens into results frameworks and monitoring, and Section 5.3 examining how to apply that lens to evaluations and use findings to drive learning. Finally, Section 5.4 considers the institutional conditions that make a stronger focus on poverty and inequality reduction possible, including how staffing, performance systems, skills development and flexible resourcing models can embed these priorities across development partners’ organisations.
Reliable data, analysis and research are essential and mutually reinforcing foundations for reducing poverty and inequalities. Without accurate measurement of living standards and disparities across individuals and groups, governments cannot effectively design, target or evaluate inclusive policies. But data alone are not enough – they only drive change when rigorously analysed and translated into policy-relevant evidence that is actively used in country-led decision making. This in turn requires political commitment, research capacity, local expertise and systems that actively connect poverty and inequality evidence to decision making.
Significant challenges persist across all three dimensions in many L&MICs. Statistical systems often remain weak, especially in low-income countries (LICs), with fragmented survey programmes, under-resourced national statistical offices and line ministries, and limited co-ordination among institutions that produce, share and use official data. There may be limited political will to generate and use evidence on excluded and marginalised groups, particularly where inequalities are deeply entrenched or politically sensitive. Although there has been a significant expansion in the availability of poverty- and inequality-related statistics, much of this evidence remains underutilised (Landau, Jütting and MacFeely, 2025[1]). Analytical capacity gaps further mean that even where data exist, they can fail to inform policy decisions. Finally, research ecosystems across many L&MICs face significant capacity constraints, and research on poverty and inequality remains deeply unequal.
Invisibility in data means invisibility to policy, reinforcing the very cycles of exclusion that development efforts aim to break. Available estimates suggest that over 800 million people lack official proof of identity – limiting their ability to access services, work or participate in civic life – with over half of these children who were never registered at birth (UNICEF, 2024[2]). Entire groups – the homeless; people in institutions like care homes, prisons or hospitals; mobile populations; and those in informal settlements – often remain invisible, with the most recent estimates suggesting 300-350 million of the world’s poorest people are missing from population counts and surveys (Carr-Hill, 2013[3]). The share of the global population covered by up-to-date censuses has declined, while the census results for countries representing around a quarter of the global population remain unpublished (Espey, Tatem and Thomson, 2025[4]).
Beyond data production, challenges remain in ensuring data quality, accessibility and interoperability. Data are often insufficiently detailed: persistent gaps in disaggregation by income, gender and location limit the ability to identify who is being left behind. Most partner countries disaggregate at least some results indicators by gender, geographic area and age, but only around 14-19% do so consistently across all indicators, and disaggregation remains much less common for ethnicity, Indigenous status and migration status.
International agreements have increasingly recognised these challenges. The Cape Town Global Action Plan, the Bern Data Compact and the Medellín Framework all reinforce the need to strengthen statistical capacity and invest in disaggregated and inclusive data systems (UNSTATS, 2025[5]). These align with UN commitments under the 2030 Agenda and the UN’s human right-based approaches to data, which emphasises that data must reveal inequalities beyond national averages, ensure the meaningful participation of marginalised populations and be used transparently to strengthen accountability (OHCHR, 2016[6]). Box 5.1 details two initiatives to ensure more inclusive data gathering.
Development partners play an important role in financing data, evidence and research systems, yet investment remains insufficient and is increasingly at risk. DAC members disbursed USD 550 million for statistics in 2023 – just 0.25% of total ODA – while many LICs depend heavily on this external financing to sustain core statistical functions (PARIS21, 2025[7]). ODA cuts in 2025 have strained national statistical systems and reduced support for key data collection programmes such as the Demographic and Health surveys, which generate critical data including on child and maternal health, family planning, nutrition and women’s empowerment (UNSTATS, 2025[8]).
This section examines how development partners can strengthen data, evidence and research systems to support more effective and inclusive poverty and inequality reduction. It starts by examining how they can help strengthen the architecture of national, subnational and global data systems and considers how evidence is used in decision making. It then outlines how development partners can provide support to governments and other development actors to address the invisibility and exclusion of hard-to-reach and undercounted populations. Finally, it lays out how they can invest in poverty and inequality research in ways that shift knowledge production toward the Global South to support context-specific, evidence-based action.
The Inclusive Data Charter
Convened by the Global Partnership for Sustainable Development Data, the Inclusive Data Charter is a global initiative that mobilises political commitment to ensure no one is left behind in data. It supports governments and civil society organisations (CSOs) to build inclusive data systems that accurately represent marginalised and vulnerable populations, and to develop time-bound national action plans that translate five core principles into concrete reforms (IDC, 2026[9]): These principles are: i) all populations must be included in data; ii) data should be disaggregated wherever possible to accurately reflect different groups; iii) data should be drawn from all available sources; iv) those responsible for collecting and producing data must be accountable; and v) human and technical capacity to collect, analyse and use disaggregated data should be strengthened through adequate and sustainable financing.
The Copenhagen Framework on Citizen Data
The endorsement of the Copenhagen Framework on Citizen Data by the UN Statistical Commission in 2025 encourages national statistical systems to make greater use of citizen-generated data produced by communities and civil society actors (CSAs), helping to fill critical data gaps on marginalised groups (Collaborative on Citizen Data, 2024[10]), such as tracking timely data on topics not officially collected like gender-based violence. As CSAs are embedded in local communities and have significant contact with marginalised groups, participatory citizen data can provide nuanced and intersectional insights.
5.1.1. Strengthening the production and use of poverty and inequality data in global, national and local systems
Development partners play a critical role in strengthening the data infrastructure that underpins poverty and inequality measurement. These span from global databases to national statistical systems and subnational data. Without sustained investment in this architecture, the evidence base for effective policy design, targeting and accountability will remain fragile. Investing in both quantitative and qualitative data is important to support poverty and inequality reduction efforts. Quantitative data establish the scale and distribution of poverty and inequalities, and enable comparisons across groups and over time, while qualitative data reveal the lived experiences, social norms, power dynamics and structural mechanisms that make structural barriers clear, explaining why some groups remain persistently left behind.
Investing in global data systems
Development partners can strengthen global data systems on poverty and inequalities by supporting the databases that make cross-country comparisons and long-term monitoring of progress possible. Table 5.1 highlights the key global databases on poverty, inequalities, fiscal redistribution and inclusion. These initiatives host and maintain centralised databases, develop standardised methodologies and harmonise country-level datasets. Supporting these systems enables globally comparable estimates on poverty and inequalities, improves data quality and credibility through rigorous validation, and provides a shared evidence base for resource allocation decisions – helping development partners prioritise where needs are greatest (see Section 3.1). Global standardised metrics can also serve as an accountability tool, with evidence suggesting they incentivise policymakers to strengthen development efforts in order to keep pace with peer countries (PARIS21, 2025[11]). However, cross-country comparisons are inherently limited and should be made with caution: significant methodological variations – including differences in survey design, sample scope and the use of income versus consumption metrics – and differences in data timeliness can distort comparisons both across countries and over time (Mahler, Foster and Tetteh-Baah, 2024[12]).
Strengthening national and subnational data systems and use of evidence
Development partner support to strengthening national statistical systems is critical for building integrated data systems that underpin poverty and inequality analysis. Such support also helps state-led monitoring and reporting for greater accountability, for example through Voluntary National Reviews. Development partners can strengthen this foundation by providing long-term, predictable and co-ordinated funding, aligned with national statistical strategies and action plans. The World Bank’s Statistical Performance Indicators can help development identify weaknesses in L&MICs statistical systems (World Bank, n.d.[21]).
Systems strengthening also includes support to systems recording births, deaths and other key life events, enabling the identification of people experiencing poverty and exclusion (see Section 5.1.2). Household surveys, censuses, civil registration and vital statistics systems and administrative data are the primary sources of poverty and inequality measurement. Supporting national statistics offices to use these to produce high-quality consumption, income and multidimensional deprivation data is essential. This includes standardised surveys such as the Demographic and Health Surveys and the Living Standards Measurement Study, which provide internationally comparable data across countries and over time on key dimensions of deprivation.
Development partners can also support countries to develop national multidimensional poverty indices (MPIs) and other country-owned measures to capture multiple dimensions of deprivation, providing a richer and more policy-relevant picture of poverty. A major limitation in L&MICs is the limited availability and quality of data on income, assets and tax records, particularly among high-income earners, which enables income and wealth inequality analysis (see Section 3.2). Given the particular reliance of gender and equity data on external funding, targeted investments are needed to institutionalise these within national systems rather than relying on ad hoc project funding (PARIS21/UN Women, 2025[22]). Development partners can support the modernisation of data collection through digital tools and high-frequency surveys to improve coverage, timeliness and reliability, while promoting interoperability across data systems to enable more coherent action across government institutions.
Supporting national registration and identification and inclusive data in Malawi
The National Registration and Identification Systems Project, implemented by the United Nations Development Programme (UNDP) in partnership with the Government of Malawi, the UK Foreign, Commonwealth and Development Office (FCDO), the EU, Norway, Irish Aid and UNICEF, established a permanent national registration and identification system in Malawi. A mass registration campaign in 2017 registered over 9.1 million citizens (UNDP, 2018[23]). Since then, the project has focused on registering remote and underserved communities, resulting in over 13.5 million registered citizens (UNDP, 2026[24]). In Malawi, the national ID has become a critical tool for enabling access to basic rights and services, including social protection programmes, as well as broader participation in economic and civic life. The system has been integrated with a digital civil registry and linked to voter and tax registration systems, helping to improve public service delivery. By providing accurate and up-to-date demographic information, it enables government institutions to plan more effectively, allocate resources more efficiently and better target services to those in need. Malawi is also developing its Inclusive Data Charter action plan with development partner support, focusing on improving data on marginalised groups such as people experiencing homelessness and children (GPSDD, 2026[25]). However, progress has been slowed by limited resources and scepticism about the value of inclusive data.
The West African Economic and Monetary Union’s (UEOMA) support to regional harmonised household surveys
UEMOA, with World Bank support, has conducted two rounds of harmonised household surveys across all eight member states during 2016-23 and has started a third round (UEMOA, n.d.[26]). Covering monetary and non-monetary poverty, multidimensional poverty, labour markets, human capital and household vulnerability, the surveys have produced comparable poverty estimates and welfare indicators across the region. The World Bank has provided technical assistance as well as financing to strengthen the capacity of national statistical offices to conduct surveys to international standards using digital data collection technology (World Bank, 2016[27]). Microdata have been made publicly available online, broadening access for researchers and policymakers. Importantly, the UEMOA Council of Ministers has institutionalised the surveys as a permanent mechanism on a three-year cycle, representing a significant step toward sustainable, regionally owned data systems.
Institutionalising citizen-generated data in Kenya
Kenya has a vibrant civil society ecosystem that generates valuable data for policymaking. Development partners have played a key role in strengthening collaboration between CSAs and the Kenya National Bureau of Statistics to address data gaps. PARIS21 (2022[28]) supported Kenya in integrating citizen-generated data into their National Strategy for the Development of Statistics, developing a quality assurance framework and supporting the validation of citizen data published on the official statistical website, while also fostering dialogue between official and non-state data producers. Other development partners worked with CSAs to develop guidelines for inclusive data production and strengthen their capacities through training and workshops. These efforts led to the integration of citizen-generated data into official statistics, improving evidence on underserved groups. For example, the production of citizen-generated education data provided insights into children's home environments and learning conditions not captured by conventional education statistics (PARIS21, 2023[29]) .
In addition to supporting data generation, development partners can help ensure that evidence is effectively used to design, target and evaluate policies that improve outcomes for the poorest and most marginalised populations. Data produced does not automatically leads to data used. Often, they can end up in “data graveyards” where data are not made widely available, buried in non-machine-readable reports and statistical websites, difficult for policymakers to interpret or not timely enough to inform policy formulation, planning and budgeting processes. Supporting the translation of poverty and inequality data into actionable insights, strengthening governments' analytical capacities, improving statistical literacy and aligning evidence with policy cycles are therefore essential to bridge the gap between data collection and policy impact. National MPIs provide a powerful example of data used as a policy tool: when embedded into policy frameworks, they can help governments identify those furthest behind, tailor interventions to specific regions and population groups, strengthen inter-ministerial co-ordination, and support high-level commitments (OPHI, 2024[30]). Other support can be directed to ensure the accessibility, timeliness and transparency of statistical outputs as well as fostering trust in and awareness of official statistics through effective communication and targeted data dissemination strategies to diverse actors – areas historically underfunded in L&MICs. This support for increased data usage is considered a low-cost, high-impact entry point, as it focuses on maximising the use of existing data rather than investing in new data collection (PARIS21, 2025[11]).
Data produced does not automatically lead to data used. Beyond support to data generation, development partners can help ensure that evidence is effectively used to design and evaluate policies.
Between data generation and policy action lies a critical but often overlooked "missing middle" comprising the systems, capacities, incentives, political will and institutional processes needed to both systematically collect data and use it as evidence for action. In many contexts, limited analytical capacity, weak links between data producers and policymakers, lack of prioritisation or political will on the topic, and challenges in communicating findings can constrain the use of data regardless of its quality. Gender data have been shown to be more directly impactful in countries where there is already awareness and political will to tackle gender inequality (PARIS21, 2025[11]). Poverty and inequality data are often overshadowed in policy and decision making by established macroeconomic indicators, such as GDP growth. The collection and use of poverty and inequality data may face additional political sensitivities, including concerns about exposing disparities or challenging existing power structures, or acceptance of disparities due to social norms (PARIS21, 2025[11]). Development partners can support advocacy, awareness raising and political engagement to strengthen commitment to both the production and use of data on poverty and inequalities. They can also provide technical assistance to develop tools and institutional mechanisms, such as dashboards, data portals, evidence repositories, monitoring frameworks and embedded labs, that support the long-term integration of data and analysis into policymaking and decision making on poverty and inequality outcomes.
The World Bank’s Umbrella Facility for Poverty and Equity (UFPE) supporting the use of inclusive data and in policymaking in Central Asia
Through technical assistance and analytical support, the UFPE helps strengthen L&MIC government systems and capacity to generate, analyse and use evidence on poverty and equity (World Bank, 2025[31]). It funded the Effective Governance for Economic Development programme which helps governments in Central Asia integrate evidence into planning, budgeting and policy reforms and programmes. For example, in the Kyrgyz Republic, more than 2 800 local government officials were trained to use a digital data planning tool, improving the efficiency and equity of subnational resource allocation. It also helped governments strengthen policy design and equitable targeting by making better use of existing data through tools such as improved inflation forecasting models, fiscal incidence analysis, welfare measurement approaches that better capture middle-class dynamics and social protection diagnostics. Another initiative, the high-frequency, participatory “Listening to Citizens” surveys provided governments across the region with timely information on socio-economic trends through the experiences of marginalised people and those experiencing poverty. In Georgia, these findings were disseminated through dashboards, open-access platforms, policy dialogues and public engagement activities, bringing both quantitative evidence and the lived realities of those furthest behind to the attention of policymakers and wider audiences, strengthening public and political support.
Innovation for Poverty Action’s (IPA) Embedded Evidence Labs institutionalising evidence use in government
Embedded Evidence Labs provide a model for institutionalising the use of data and evidence within government decision making (IPA, 2026[32]). Co-designed with government partners, these labs consist of dedicated teams embedded in government agencies combining technical assistance, learning-by-doing and capacity strengthening to help institutionalise evidence use within government, with leadership and ownership gradually transferred to government institutions over time. At the policy level, they help governments test, evaluate and scale effective interventions. At the systems level, they strengthen the capacities, processes and institutional arrangements needed to embed the sustained use of evidence in decision making over the long term. Labs are tailored to country priorities and institutional needs. For example, Evidence Labs have been established within education ministries in Peru, Rwanda and the Philippines, while Nigeria hosts a National Poverty Intelligence Lab and Colombia has labs focused on family welfare and inclusive finance. The integration of Evidence Labs into government plans, legal frameworks and budgets has helped sustain the use of data and evidence in decision making beyond political cycles. Results include the scaling up of evidence-based programmes; for example, in Peru, an evidence-based information campaign reduced secondary school dropout rates by nearly 19% in urban areas.
Development partners can also support investments in subnational data systems to reveal the geographic concentrations of poverty and spatial inequalities often concealed by national averages. Subnational governments – municipal, district, provincial or regional – are often responsible for delivering and adapting policies, yet they frequently lack the granular data needed to identify local needs, target interventions and monitor progress. Development partners can support local statistical capacity and data systems that integrate administrative, survey and community-based data sources to better align with local planning and service delivery needs (PARIS21, 2016[33]). Supporting interoperability across datasets – through integrated data platforms and common standards that allow data from central and local governments to be linked – enables co-ordination across authorities and service delivery sectors, reducing duplication and supporting more coherent policymaking and effective implementation.
Data availability is often constrained by limited geographic disaggregation, masking significant local disparities, especially in small, remote or mountainous communities (PARIS21, 2025[11]). Small-area estimation techniques are particularly valuable, enabling the production of reliable poverty estimates at granular levels and revealing hidden subnational disparities that would otherwise go unaddressed (World Bank, 2021[34]). This is illustrated by the UN’s support to Indonesian local authorities at the provincial, district and municipal levels to establish local mechanisms generating more reliable and timely data (UN Indonesia, 2024[35]). Development partners can also help develop and scale innovative approaches to generating geographically disaggregated data – such as satellite data, local participatory indices and mobile phone data – and can play a direct role in producing such data, particularly in data-scarce or low-capacity contexts (see Section 5.1.2).
Supporting disaggregated and intersectional data
Aggregate data can obscure how characteristics such as gender, ethnicity, disability, location and socio-economic status intersect to create distinct and compounding forms of exclusion. Disaggregating by a single characteristic is often insufficient to capture this complexity: a woman living in a rural area with a disability faces a combination of disadvantages that neither gender data nor disability data alone would reveal. Disaggregated data therefore make different and intersecting forms of marginalisation visible and can help uncover their underlying causes and the potential barriers to transformative impact (UNFPA, 2022[36]). Yet, significant data gaps remain for many marginalised populations, including Indigenous Peoples and caste-affected communities. These gaps reflect not only technical and capacity constraints but also the political sensitivities associated with documenting and exposing inequalities. It is important to identify and prioritise gaps in data, capacity and processes, and to understand the underlying causes of these gaps, such as capacity constraints, political and security barriers, and social norms (UNSDG, 2022[37]). This includes asking critical questions about why data are missing, who is not being counted, and how exclusion, stigma or fear may prevent certain groups from being represented (see Section 5.1.2).
Disaggregated data can make different and intersecting forms of marginalisation visible and can help uncover their underlying causes and the potential barriers to transformative impact.
Across global, national and subnational levels, development partners can support disaggregated and intersectional data that reveal overlapping deprivations and inequalities. They can invest in improving survey designs, administrative data systems and statistical methods so that disaggregated datasets can be analysed across intersecting characteristics. This includes enabling self-identification and integrating qualitative and citizen-generated data alongside quantitative sources (see Section 5.1.2) and supporting innovation by funding the development and testing of new methodologies and tools to advance intersectional and disaggregated data. Development partners can further help ensure that disaggregated and intersectional data feed into policy processes, including support to embed an intersectional lens throughout the data value chain. Requiring disaggregated and intersectional data to be embedded into development partners’ financing frameworks and national action plans can help ensure sustained commitment beyond ad hoc investment.
Colombia’s institutionalisation of intersectional data within national statistical systems
Colombia’s National Administrative Department of Statistics has embedded intersectionality into national data systems, supported by UN Women’s development partner-funded Women Count Programme. It takes an “intersectional and differential” approach that considers how overlapping identities such as gender, ethnicity, disability and socio-economic status interact to produce distinct forms of exclusion. It goes beyond simple data disaggregation, creating a dedicated team, methodological guide and national training programme to embed intersectionality across the data cycle, grounded in engagement with marginalised groups. This approach has enabled intersectional data to directly inform targeted programmes for rural women, Indigenous populations and LGBTIQ+ groups. A key example is the Wayúu Multidimensional Register, a localised data system developed with Indigenous participation to capture multiple culturally relevant dimensions of well-being (Open Data Watch, 2025[38]). UN Women also co-produced a flagship report with Colombia’s Department of Statistics on gender gaps, which guided national policy including the National Gender Policy (2022) and the National Development Plan 2023–2026 (UN Women, 2023[39]).
UN Economic and Social Commission for Asia and the Pacific’s (UNESCAP) intersectional and disaggregated data platform
UNESCAP's Leave No One Behind (LNOB) Platform helps policymakers identify the population groups most disadvantaged in access to basic services and opportunities, including health, education, water, sanitation, energy and financial inclusion (UNESCAP, n.d.[40]). It applies classification and regression tree analysis to disaggregate data into progressively smaller groups based on intersecting characteristics – such as income, location, gender and education – revealing how multiple disadvantages combine to drive unequal outcomes. A complementary Dissimilarity Index quantifies how unevenly opportunities are distributed and what redistribution would be needed to achieve equality. UNESCAP also supports capacity strengthening through training and technical workshops, enabling government, statistical and civil society actors to conduct intersectional analysis and integrate the findings into policy design and monitoring. It also produces knowledge products on development in Asia and the Pacific using intersectional data and analysis.
The International Women's Development Agency's intersectional approach to understanding poverty
The International Women's Development Agency's Individual Deprivation Measure, supported by development partners and operationalised through the Equality Insights approach, captures poverty at the level of each person rather than the household, collecting data across 15 dimensions including health, education, work, voice and time use (Crawford et al., 2021[41]). By surveying all adults in a household, it identifies the "invisible poor" – those who may live in relatively better-off households but still experience significant deprivation. This individual-level design enables detailed disaggregation by sex, age, disability and other characteristics, as well as intersectional analysis of overlapping disadvantages. The measure is scalar rather than binary, showing degrees of deprivation rather than a simple poor/not poor classification, enabling more precise identification of those furthest behind. For example, the Equality Insights Rapid Report in Tonga was conducted to inform COVID-19 recovery efforts and the response to the January 2022 volcanic eruption (Equality Insights, 2023[42]). It adapted to time sensitivity and connectivity constraints in the Pacific by using shorter, phone-based surveys, with support from UN agencies to capture individual-level data to inform urgent policy.
Embedding ethical and responsible data practices
Development partners can help governments embed ethical and responsible practices into data systems from the outset, recognising that poor data governance can itself deepen poverty and inequalities. This means strengthening safeguards for privacy, consent and data protection. It is particularly important when collecting data on sensitive characteristics such as ethnicity, disability, sexual orientation or migration status, where misuse can expose already marginalised groups to discrimination, surveillance or harm. Development partners can build capacity to identify and mitigate biases in data collection and analysis and promote transparency and accountability in how data is used. Embedding "do no harm" and human rights-based approaches across all supported data programmes helps ensure that the expansion of data systems does not inadvertently reinforce the exclusion of the groups they aim to reach. As governments increasingly adopt artificial intelligence (AI) and data-driven systems for service delivery and targeting, ensuring these tools are accompanied by robust inclusion safeguards is essential, a point taken up further in the next section.
5.1.2. Addressing data exclusion and invisibility of marginalised communities
Deliberate efforts to address data exclusion are an important complement to the investments in data infrastructure discussed in the previous section. Even well-resourced national statistical systems can systematically exclude the poorest and most marginalised (Kidd, Athias and Mohamud, 2021[43]). Within countries, data identification gaps mainly affect marginalised groups along age, income, gender, education level, employment and geographical lines (Metz, Casher and Clar, 2024[44]). The groups most systematically missed include stateless persons, displaced people, Indigenous Peoples, persons with disabilities, informal and migrant workers, homeless and highly mobile or nomadic populations, individuals in institutional settings such as prisons or care homes, and others outside standard household frameworks (Carr-Hill, 2013[3]). These gaps arise from a combination of factors: exclusion by design, when data systems simply do not account for certain groups; exclusion for structural reasons, including discrimination, lack of legal status or the practical barriers that poverty itself creates; and exclusion for methodological reasons, when sampling frames, questionnaire design or analytical categories fail to capture the diversity of lived experiences. Marginalisation occurs across the full data value chain: from who is identified and counted, to whose experiences are captured when collecting data and whose realities inform analysis and decision making. Development partners need to understand these structural drivers and tailor their support accordingly, with different forms of exclusion requiring different responses.
Supporting active search and adapted data collection for hard-to-reach populations
Development partners can support governments to reduce the barriers that exclude some of the most marginalised from official registration systems. Section 5.1.1 identified civil registration and vital statistics systems as core data infrastructure, but their value depends on whether the most marginalised can actually access them. National ID and digital registration systems can significantly advance inclusion by underpinning access to services, economic participation and civic life, but also carry risks if poorly governed. In some contexts, they can be used to deny marginalised groups access to rights and services. Realising their potential therefore requires registration processes to be accessible to all population groups, as well as robust legal frameworks and strong accountability mechanisms to prevent misuse. Development partners can support the removal of direct and indirect costs of registration and assist governments in waiving birth certificate requirements for displaced and stateless persons. Where barriers persist, active search programmes – in which governments proactively go out to identify and register those who have not come forward themselves, through approaches such as door-to-door visits and mobile outreach units – are equally important.
Development partners can also support adapted data collection approaches designed to reach groups that fall outside official systems (see Box 5.5). This includes funding survey methodologies tailored to groups such as stateless persons, nomadic communities, displaced populations and informal workers. Development partners can also assist governments in addressing structural blind spots by supporting time-use surveys that capture unpaid care work and informal labour – dimensions of poverty and inequality that standard household surveys systematically miss (Charmes, 2019[45]). Longitudinal data help reduce the invisibility of marginalised and left-behind populations by capturing how exclusion and disadvantage evolve over time, including the impacts of life transitions, shocks and opportunities for resilience.
Finally, marginalised groups can face additional barriers that reduce their willingness to participate in surveys or registries, rooted in fear of stigma, retaliation or political surveillance. Depending on context, characteristics such as ethnicity, religious affiliation or sexual orientation can expose participants to harm if data are not handled carefully or if findings are not adequately anonymised (Busher and Godfrey-Faussett, 2022[46]). Such risks undermine both the safety of marginalised populations and their propensity to engage in data collection exercises. Development partners can help address under-reporting of sensitive topics – while maintaining safety – by supporting engagement with local communities, leaders and civil society organisations and investing in privacy-protecting data standards that build trust and encourage participation (Sannon and Forte, 2022[47]).
Reaching marginalised groups through an active search approach in Brazil
Busca Ativa is an active search initiative that proactively registers low-income and marginalised households into Brazil's Unified Social Registry, enabling access to social protection programmes including the Bolsa Família cash transfer. Using door-to-door visits, mobile outreach units and community-based identification, it reaches Indigenous Peoples, Afro-descendant communities, remote populations, homeless populations and other hard-to-reach groups that standard registration processes miss (Lindert et al., 2020[48]). With World Bank technical and financial support, Busca Ativa registered over 1.7 million additional families with extremely low incomes in six years, with the number of marginalised households registered growing from around 200 000 in 2011 to 1.9 million by 2015 (World Bank, 2019[49]). Between 2023 and 2024 alone, a further 3.2 million households were registered (Government of Brazil, 2024[50])
Norway’s role in strengthening statelessness and displacement statistics globally and nationally
Statistics Norway and Norad support the strengthening of data on historically excluded people: refugees, internally displaced persons, forcibly displaced and stateless populations in co-operation with partners including the United Nations High Commission on Refugees (UNHCR) and the African Union. At the country level, Statistics Norway supports partner countries to produce and use statistics on these groups through technical assistance and capacity development in survey methods and implementation, and census inclusion (Statistics Norway, 2026[51]). The resulting data have informed national policy decisions, including citizenship reforms to include previously stateless groups, which led to transformative results by granting individuals legal identity and protection, and by enabling their participation in society, their access to essential services and their right to work. Globally, Statistics Norway helped initiate the Expert Group on Refugees, Internally Displaced Persons and Statelessness Statistics and contributed to developing international statistical standards. Since these efforts to strengthen evidence for the inclusion of stateless began in 2014, an estimated 500 000 stateless individuals have acquired citizenship (Statistics Norway, 2024[52]).
Leveraging innovative data sources to reach the invisible
Development partners can support innovative data sources to identify and reach populations that fall outside official systems (OECD, 2020[53]). Satellite imagery – drawing on indicators such as night-time light intensity, infrastructure density and land-use patterns as proxies for living standards – can estimate poverty incidence and map deprivation at granular levels that household surveys do not reach. For example, in Colombia, the use of satellite imagery generated 70 times more data points on poverty than census-based approaches (PARIS21, 2023[54]). Mobile phone data offer a complementary source, capturing patterns of movement, connectivity and economic activity that can help identify vulnerable populations and track welfare changes in near real time. This is particularly valuable in fragile and conflict-affected situations where traditional household surveys can be unviable and the need for timely, localised data on vulnerable populations is greatest, for example to monitor real-time migration flows to better allocate resources (ESA, n.d.[55]). Development partners can support public-private partnerships to access and use private sector mobile and satellite data responsibly for public purposes and help establish global standards and governance frameworks for their use in L&MICs to ensure that the benefits of data innovation reach the countries and communities that need them most.
AI and machine learning tools offer significant additional opportunities to enhance the analytical power of these data sources. Development partners are already supporting governments to test inclusive AI use cases, from identifying social protection beneficiaries to translating community priorities into development planning. However, poorly governed AI-driven systems carry significant risks of exacerbating exclusion and inequities. Systems trained on incomplete or biased data can embed and scale exclusion – leading to "algorithmic exclusion" from services and assistance (UNDP, 2025[56]). Biometric failures, language barriers, limited digital connectivity and system designs that reflect dominant groups can further reinforce existing inequalities. Development partners can therefore help ensure that AI investments are accompanied by inclusion safeguards – bias testing, disaggregated performance metrics, human oversight and appeal mechanisms – and support technology transfer and capacity strengthening to prevent AI benefits from becoming concentrating in more technically advanced economies.
The Asian Development Bank’s (ADB) support for satellite imagery in mapping the spatial distribution of poverty in Thailand
Through its Data for Development initiative, the ADB supported Thailand to improve poverty data by introducing satellite imagery and machine learning to map the spatial distribution of poverty at fine geographic scales (ADB, 2021[57]). High-resolution satellite images are analysed to detect features associated with economic activity such as housing conditions, infrastructure, land use and night-time light intensity, and then combined with survey data to estimate poverty levels in areas where traditional data are limited. Because satellite data are collected frequently, this enables more timely monitoring than infrequent census rounds. ADB also strengthened the National Statistical Office's capacity in small-area estimation and geospatial analytics, improving Thailand's ability to identify vulnerable populations, design targeted programmes and allocate resources more efficiently.
Supporting the ethical and inclusive integration of AI in data systems and public service delivery
Governments in L&MICs are increasingly using AI to improve data and public services while embedding inclusion through bias testing, disaggregated metrics and human oversight in decision making (e.g. welfare eligibility). Development partners are supporting these efforts by helping governments pilot inclusive AI, strengthen capacity and co-develop locally relevant, people-centred systems. For example, UNDP is supporting AI-assisted participatory planning tools in Indonesia to capture community priorities for village development planning; AI chatbots in Nepal to expand access to public services, particularly in remote areas; and AI-driven beneficiary identification Malaysia’s social protection systems to strengthen data use in response to shocks (UNDP, 2025[56]). However, poorly governed AI risks deepening existing data gaps and biases, leading to barriers to accessing services. To help address this, UNESCO’s Global AI Ethics and Governance Observatory, funded by the EU, supports rights-based, human-centred AI governance through capacity strengthening and the development of practical tools for government use, such as an Ethical Impact Assessment to evaluate how AI systems may reproduce or amplify bias and exclusion (UNESCO, 2023[58]).
Supporting participatory, community-led, qualitative and other alternative data sources
Development partners can help move beyond just collecting data “about” communities to collecting data “with and by” them, by funding approaches that give people experiencing poverty a say in defining what gets measured and how. This means supporting participatory data governance across the full data cycle – from defining what is measured to how findings are used – not just at the point of collection (GPSDD, 2022[59]). As discussed in Section 4.3, community-led mapping and participatory wealth ranking complement official statistics by drawing on local knowledge to identify and reach those that formal systems miss. Qualitative approaches, such as surveys, participatory poverty assessments, focus groups and community dialogue, generate rich evidence on lived experiences of exclusion, social norms and power dynamics, revealing underlying structural barriers that may not be apparent from quantitative data alone. This strengthens policymaking by helping to explain why some groups are persistently left behind, contextualising disparities, building empathy among decision makers, and identifying root causes (PARIS21, 2025[11]). Citizen-generated data further expand the scope of evidence beyond what official systems capture, connecting actors such as marginalised people, local communities, civil society and researchers – although care is needed to ensure quality, representativeness and uptake by national statistics offices. Participatory data approaches can also involve broad dissemination, directly engaging people with the data and sharing findings in ways that encourage dialogue, reflection and collective interpretation – helping to reshape social norms and strengthen community support for poverty and inequality reduction.
Development partners can also support the use of data sources that provide evidence on marginalisation and exclusion that official statistics may not capture. Human rights data are a particularly valuable source of development intelligence: country-specific recommendations from UN human rights mechanisms provide authoritative evidence on priority inequalities and the groups facing the most severe discrimination. They can be used to inform country strategies, identify gaps in official data and design more inclusive, rights-based policies (OHCHR, 2016[6]). Complementing this, perception-based surveys – such as the World Values Survey and real-time monitoring surveys of people experiencing poverty and inequalities – capture social norms, attitudes and lived experiences of discrimination and exclusion that official statistics rarely reach.
UNDP’s Integrated Local Multidimensional Poverty Index+ supporting inclusive local policies
This index extends the traditional MPI by incorporating locally defined indicators, qualitative insights and participatory inputs, producing localised, context-specific evidence that supports targeted local interventions (UNDP, 2024[60]) . Through participation, it enables communities and local actors to shape and validate poverty measures so that results reflect lived experiences, local priorities and cultural perspectives.
In Burkina Faso, the analysis revealed strong links between conflict, displacement and deprivation in basic services. Participation processes were conducted across diverse regions, including rural, urban and conflict-affected areas with varying rainfall levels and involving different demographic and livelihood groups such as internally displaced persons and community leaders. Local MPI data were piloted to strengthen poverty targeting design and programming, allocate resources, identify local initiatives to support, and improve the equity of fiscal systems.
Indigenous-led framing of poverty through community-generated data in Peru
Supported by the EU, the Indigenous Navigator is a framework and set of tools used for and by Indigenous Peoples to monitor their rights through community-generated data (Indigenous Navigator, n.d.[61]). It aims to strengthen communities’ ability to claim their rights and to informs political decision-making. The Navigator’s data on the Wampis in Peru illustrate how poverty can be both mismeasured and poorly addressed without cultural adaptation and participation in policy design (Indigenous Navigator, 2020[62]).
The Peruvian state classifies all Amazonian Indigenous peoples as extremely poor solely based on ethnic identity, regardless of their actual conditions, thereby institutionalising stigma that equates Indigeneity with deprivation. Drawing on survey data, interviews and participatory meetings, Wampis communities emphasised that well-being is rooted in self-sufficiency, subsistence livelihoods, access to territory and cultural practices, rather than income alone – although the communities do have monetary needs (such as school materials, tools, fuel and other essentials). Despite this, national social protection programmes rely primarily on quantitative, monetary-based approaches that may not fully capture these Indigenous perspectives of well-being. This example highlights how Indigenous community‑owned data can generate evidence for culturally relevant multidimensional deprivations that reflects local values of well-being, which can be used to advocate for the recognition and protection of rights.
Using community radio and data to tackle gender-based violence (GBV) in Zanzibar
Community radio stations in Zanzibar are using monthly GBV data, disaggregated by district, age, disability status and type of violence, to engage communities, raise awareness and challenge harmful social norms (UN Women, 2025[63]). By sharing clear, verified statistics through news programmes, talk shows and interactive call‑ins, they create safe spaces for open discussion that reduce stigma and break long‑standing silence around violence. This community-driven approach helps people better understand GBV, recognise abuse and feel empowered to speak out. As attitudes shift, more people are reporting cases and supporting survivors. In 2024, 1 809 GBV cases were recorded, a 28% increase since 2020, while listener feedback shows growing knowledge of reporting channels and services. Many now feel confident helping others in their communities. The data also inform targeted services locally and guide national policy, resource allocation and broader GBV prevention efforts.
Source: Authors’ compilation based on UNU (2019[64]; 2019[65]), UNICEF (2020[66]) and Thinyane (2018[67]).
5.1.3. Investing in poverty and inequality research in L&MICs
Poverty and inequality reduction requires evidence-based research to identify what works, for whom and under what conditions and inform relevant stakeholders in policy planning, implementation and evaluation. Development partners play an important role in funding poverty and inequality research globally, from large-scale longitudinal studies to impact evaluations, research capacity strengthening and policy-oriented analysis in L&MICs. This research serves a dual purpose: informing partner country governments' own policies and programmes, and strengthening the evidence base that development partners themselves draw on to design, target and evaluate their interventions.
Yet despite this investment, global research on poverty remains deeply unequal: most studies are conducted in HICs, published in English and shaped by donor priorities rather than local needs (Kanmodi, Amzat and Nnyanzi, 2023[68]; GDN, 2017[69]). This limits both the relevance and the ownership of the evidence base in the countries that need it most. Policymakers are more likely to use evidence they perceive as contextually relevant, with research showing greater receptiveness to findings from comparable L&MICs than from HICs, as differences in political and institutional contexts can reduce perceived applicability and limit cross-country learning (Richards, 2025[70]). Where research is locally led, it tends to produce more relevant, owned and sustainable policy solutions. This leaves development partners with both an opportunity and a responsibility to support a better balance of knowledge production between the Global North and South through equitable partnerships (Danida Fellowship Centre, 2025[71]).
Strengthening the quality, inclusivity and local ownership of poverty and inequality research
Development partners can strengthen both the quality and inclusivity of the evidence base on poverty and inequalities. They can fund longitudinal research to track how poverty evolves over time and across generations, mixed-methods approaches to capture both its scale and lived experience, and impact evaluations to generate rigorous evidence on what policies and programmes actually work (see Section 5.3 on evaluations). Such research tends to add greatest value when co-designed with local researchers, addressing locally defined questions, building local capacity alongside generating findings and ensuring results are accessible in relevant languages and formats. Development partners can therefore embed expectations for local co-ownership, South-based researchers’ involvement and capacity strengthening in their research funding – supporting a shift toward models where L&MICs are not only subjects of research but active producers of it. Research that includes historically excluded voices and lived experiences tends to be not only more equitable but more accurate, generating findings that better reflect the complexity of poverty and inequalities on the ground.
Research that includes historically excluded voices and lived experiences tends to be not only more equitable but more accurate, generating findings that better reflect the complexity of poverty and inequalities on the ground.
Strengthening local research capacity
Development partners can invest in research capacity strengthening to support locally led poverty and inequality research in L&MICs. This requires a systemic approach that responds to different capacity needs at different levels – from strengthening research skills, career incentives and pathways, to improving institutional contexts, research management and ICT infrastructure, as well as the broader enabling environment of research policy and funding (GDN, 2017[69]; FCDO, 2019[72]). Development partners can support change at both the individual level, by investing in researchers through training, mentorship and career development, and at the institutional level, by entering into long-term partnerships with local think-tanks, research centres and universities and engaging with governments on national research strategies and funding frameworks that sustain locally led research. Strengthening research capacity tends to be most effective through equitable North-South partnerships in which Southern researchers lead and set agendas, and the involvement of diverse national stakeholders (Chadwick El-Ali et al., 2022[73]). Development partners can also support open science initiatives that expand access to research and address language barriers, empowering researchers in L&MICs to engage with the international scientific community and promote the dissemination of findings beyond academic networks – to policymakers, development partners, CSOs, the media and the private sector – ensuring that evidence on poverty and inequalities reaches those best placed to act on it.
Institutionalising evidence use and embedding researchers into decision making structures
Even high-quality, locally led research will not reduce poverty and inequality unless it influences the people and institutions that make decisions. Development partners can play an active role in building these connections, starting within their own organisations. Strengthening links between research and operational teams ensures that evidence generated through research actively shapes their operations. Development partners can also support structured connections between researchers, knowledge translators and policymakers in L&MICS and globally – building communities of practice and policy dialogues that foster the trust and mutual understanding evidence uptake requires (Vicente-Crespo et al., 2021[74]). Evidence dissemination benefits from being tailored to the local context; for example, in Brazil, a policymaker highlighted their preference for informal WhatsApp networks, where researchers and policymakers share and discuss new evidence in real time, finding them more timely and practicable compared to formal briefings (Richards, 2025[70]). Embedded programmes, which place researchers directly in government institutions, offer a valuable mechanism for bridging the gap between evidence and policy (Woodall, Potts and Brown, 2024[75]). Flexible funding that allows research programmes to respond to policy windows, rather than fixed deliverables, can increase the likelihood that findings will inform policy decisions. Development partners can also support CSOs to analyse poverty and inequality evidence and advocate on behalf of marginalised groups, ensuring research influences not just governments but the broader range of actors working on inclusive development.
FCDO’s support to better understanding of the causes and dynamics of poverty and inequalities
FCDO’s research and evidence approach draws on national and international expertise to inform policy and decision making, including generating insights on the underlying causes of poverty and the dynamics faced by those living in poverty and facing marginalisation. The Development Engagement through Evidence and Policy programme, a global consortium launched by FCDO in 2018, aims to improve data and analysis on the drivers of extreme poverty and identify context-specific policy responses, particularly in Sub-Saharan Africa, fragile and conflict-affected settings, and among marginalised groups (FCDO, 2025[76]). Through the programme’s Challenge Fund, FCDO supports locally led research. For example, in Madagascar, research explores how repeated flooding and internal displacement contribute to urban poverty and how artisanal gold mining shapes rural livelihoods and poverty reduction (DEEP, 2026[77]).
FCDO also funds global longitudinal programmes that generate in-depth participatory qualitative and quantitative insights by engaging adolescents and their wider communities. The Gender and Adolescence: Global Evidence (GAGE) initiative, supported alongside Irish Aid, UNICEF and Canada, tracks the lived experiences of marginalised adolescents, including refugees, young parents and people with disabilities, through repeated interviews and surveys, highlighting how gender norms and intersecting vulnerabilities shape life courses (FCDO, 2025[78]). The Young Lives at Work project has followed 12 000 children in 4 countries since 2002, capturing their perceptions, attitudes, aspirations and experiences, and examining how social and environmental factors influence educational access and outcomes, skills development and transitions into labour markets, as well as overall life trajectories (FCDO, 2025[79]). This long-term research has fed into evidence used by policymakers. In Ethiopia, GAGE evidence helped inform a national public hearing and a draft law on female genital mutilation, while Young Lives research has influenced national policy debates and laws on banning child marriage in Peru and India (FCDO, 2025[79]).
Advancing country-led research ecosystems through support to the African Centre of Excellence for Inequality Research (ACEIR)
ACEIR, hosted by the University of Cape Town and supported by the EU-AFD Facility on Inequalities, places African scholars at the forefront of inequality research (AFD/World Bank, 2025[80]). It strengthens research and data capacity through training, scholarships, institutional partnerships and data dissemination for postgraduate students, early-career researchers, policymakers and civil society for a stronger African-led inequality research ecosystem. ACEIR focuses on poverty, growth and the drivers of different types of inequalities, analysing the conditions of those facing poverty, the middle classes and elites, as well as the dynamics between them (ACEIR, n.d.[81]).
It has developed a Handbook on Inequality Measurement for Country Studies in partnership with the EU-AFD Research Facility, which provides a practical framework for country-level, policy-relevant inequality analysis and has been applied in a series of Inequality Diagnostics across Africa (AFD, 2026[82]). ACEIR's African Inequality Review is a multi-country research and policy programme that works with partners including the London School of Economics, FCDO, UNU-WIDER and the World Bank to generate comparable, policy-relevant evidence on inequality across Africa through cross-country analysis and in-depth country studies, with a focus on multidimensional inequality and perceptions of inequality (UNU-WIDER, 2026[83]).
Results and monitoring frameworks help development partners structure what they aim to achieve for people experiencing poverty and inequalities, track progress and correct course when needed. Results frameworks articulate the link between individual interventions and overarching objectives, while monitoring frameworks allow continued tracking of implementation and delivery. Without them, efforts can fail to reach those furthest behind or even inadvertently widen the gaps they aim to close. These frameworks therefore play a critical role in translating strategic commitments on inclusive and equitable development into a clear pathway for action and generating evidence to improve interventions. Monitoring – the continuous process of collecting and analysing data to compare how well an intervention is being implemented against expected results – supports transparency and accountability and can feed evidence into decision making during implementation (OECD, 2022[84]). Evaluation, which often draws on monitoring evidence to assess whether interventions delivered their intended outcomes or impact after implementation to inform future programming, is covered in Section 5.3.
Integrating a meaningful people-centred poverty and inequality lens into results and monitoring frameworks requires deliberate attention. Fundamentally, results frameworks serve multiple and sometimes competing functions spanning accountability, learning, communication and resource allocation. These tensions can play out in poverty and inequality reduction efforts. For instance, frameworks optimised for accountability to development partners can crowd out learning or meaningful accountability to affected populations. Results chains can oversimplify the multi-causal nature of deprivation and disparities; indicators may reflect what is easy to measure rather than what matters most and may mask distributional differences; and monitoring systems in high-poverty contexts are often limited to recording outputs, precisely where impacts on the lives of people experiencing poverty and inequalities are least visible. Power dynamics, political sensitivities and data gaps – particularly for the most marginalised – further complicate the picture.
Results frameworks serve multiple and sometimes competing functions: frameworks optimised for accountability to development partners can crowd out learning or meaningful accountability to affected populations.
Even the most well-designed monitoring and evaluation architecture will only drive change if institutional conditions, culture and incentives exist to embed it into practice (OECD, 2023[85]; OECD, 2024[86]). When these incentives encourage rigorous and transparent reporting, results and monitoring frameworks can generate evidence that accurately reflects the realities of poverty and inequalities, rather than rewarding risk avoidance or a focus on demonstrating success. Ensuring that monitoring systems translate into adaptive learning requires strong feedback loops that link evidence directly to decision making, to enable interventions to respond to people’s needs in real time. Adaptive management approaches, combined with participatory and people-centred monitoring, can strengthen responsiveness and accountability by grounding actions in the lived realities of people experiencing poverty and inequality.
This section provides guidance on how development partners can address these challenges and better integrate poverty and inequality considerations into results and monitoring frameworks. It starts by outlining the considerations involved in building a poverty- and inequality-sensitive results framework and then examines how to select and use indicators that track who benefits and who is left behind. It considers what is involved in creating fit-for-purpose monitoring systems, and how monitoring can drive adaptive learning, including through participatory and people-centred approaches. The effectiveness of these frameworks ultimately depends on the institutional conditions explored in Section 5.4, including leadership commitment, dedicated expertise and incentive structures that reward equity-based outcomes.
5.2.1. Operationalising the strategic intent of poverty and inequality reduction interventions through a results framework
A results framework spells out how and why specific interventions lead to desired impacts. It provides a structured way to link poverty and inequality objectives with corresponding inputs, activities, outputs, outcomes and impacts. It typically includes tools like logical frameworks (logframes), which map the results chain from inputs to impacts, and theories of change, which articulate the causal logic and assumptions behind the framework, explaining why and under what conditions poverty and inequality outcomes are expected to materialise. When well designed, these tools can shape the design of individual interventions so that they connect to poverty and inequality reduction goals and support adaptive learning and accountability. Table 5.3 presents a simplified theory of change for poverty and inequality reduction; in practice, as elaborated below, such frameworks should reflect the full complexities involved.
Note: This example would apply to a country context characterised by high needs and low capacity.
Source: (Sumner, unpublished[87]).
Theories of change have different strengths and limitations for tackling poverty and inequalities at different development co-operation levels. At the institutional or portfolio level, theories of change can support strategic alignment around an organisation’s poverty and inequality goals and promote coherence across sectors and countries. However, they may remain too abstract to effectively guide the design and implementation of specific interventions. At the country level, they enable context-specific approaches aligned with partner country priorities and strategies for inclusive development but may create fragmentation across the portfolio if not grounded in a higher-level framework. At the intervention level, theories of change help ground design and monitoring in concrete terms. However, they can risk narrowing the focus to immediate causes, while insufficiently addressing the structural drivers of poverty and inequalities that may sit outside the scope of the intervention. Linking theories of change across institutional, country and intervention levels can help offset some of these limitations and strengthen overall coherence in tackling poverty and inequalities.
Designing results frameworks for poverty and inequality reduction requires careful attention to several common pitfalls. Linear results chains can oversimplify the dynamic, multi-causal nature of poverty and inequalities, obscuring unintended distributional effects that are central to how inequalities are produced and reproduced (Stein and Valters, 2012[88]). Frameworks that treat power relations, elite capture and discriminatory institutions as background noise risk missing uneven impacts across groups and resistance to change (Eyben et al., 2015[89]). Finally, overly rigid logframes can make it difficult to adapt interventions when the underlying context changes or when marginalised populations prove harder to reach than expected. Results frameworks can be seen as a part of ongoing review and learning processes that allow for course correction, rather than rigid blueprints – particularly in complex programme or policy settings.
It is equally important to ensure that poverty- and inequality-related results frameworks are actually used in practice (see also Section 5.2.4 on monitoring and adaptive learning). Even well-designed results frameworks that avoid the pitfalls identified above risk sitting idle if there are insufficient incentives for their use, including the organisational culture and staff capacities needed to act on the results (OECD, 2024[86]).
5.2.2. Selecting indicators to track progress along the results chain
Understanding who is benefiting from an intervention, how and to what extent, and who is being negatively affected depends on having the right indicators. These translate poverty and inequality reduction objectives into specific, measurable expressions of change that can be tracked over time. Different levels of the results chain have different indicators (see Table 5.4).
As with objective setting (see Section 2.4.2), indicator selection involves balancing potentially competing considerations. Indicators need to reflect the dimensions of poverty and inequalities identified as most binding through context analysis (see Section 2.4.1). There is also the need to maintain coherence with partner-country results frameworks and internationally agreed frameworks such as the Sustainable Development Goals (SDGs) to avoid fragmentation and parallel reporting burdens, while being mindful that partner countries’ frameworks may themselves have quality limitations. Finally, indicators need to align with development partners' own results frameworks and institutional priorities. Data availability adds a further practical constraint: indicators need to be shaped by what poverty and inequality data already exist, how frequently they are updated and whether additional collection is feasible, with development partners able to support this where gaps exist (see Section 5.1) – balanced against the risk of measuring only what is easy to measure (e.g. beneficiaries reached) at the expense of what matters most (e.g. genuine reductions in poverty and inequalities). In practice, full alignment with all these considerations is rarely achieved; the objective is to identify workable points of convergence.
The principle of country ownership means partner countries’ results frameworks serving as the primary reference point as far as possible. Results frameworks reflect a choice of what counts as progress and so aligning with partner countries' frameworks is not merely a technical exercise, but a normative one about ensuring their voices shape the measures of success (OECD, 2019[90]). Where possible, co-defining indicators with local actors – especially people experiencing poverty and inequalities – strengthens the legitimacy and relevance of the results architecture, grounding it firmly in lived realities and priorities (OECD, 2024[86]) (see Section 4.3).
International frameworks can provide a useful common language and a global basis for indicator selection. The SDGs elaborate a set of goals, targets and indicators reflecting shared commitments to reduce multidimensional poverty (SDG 1) and inequalities (SDGs 5 and 10) to which both provider and partner countries have subscribed. Germany, for instance, adopted 83 key indicators drawn from the 2030 Agenda, including on poverty and inequality reduction (Federal Government of Germany, 2025[91]). The UN's emerging Framework for measuring equitable, inclusive and sustainable well-being and progress, proposes an "Equity and Inclusion" pillar covering wealth and income inequality, poverty, work inclusion, regional disparities and overlapping deprivations (see Table 1.2). As this framework matures and gains broader international traction, it may help development partners align how they track distributional outcomes with an emerging global consensus on measuring equity and inclusion, while avoiding a proliferation of reporting frameworks.
It is important to distinguish between types of indicators. Indicators may be poverty-focused or inequality-focused; both may track progress toward the same overarching goal, but differ in how they measure it. Poverty-focused indicators capture absolute disadvantage, whereas inequality-focused indicators capture relative disadvantage between groups (see Table 5.5. ). Indicators tracking whether individuals or groups are being left behind can be targeted (focusing on a specific group, e.g. the proportion of girls without access to education), disaggregated (broken down to enable comparisons between at least two groups, e.g. the proportion of girls without education by income level, disability, migratory status and ethnic origin) or group-difference based (assessing outcomes between a disadvantaged group and the rest of the population, e.g. the percentage-point difference in education between girls and all other children) (UNICEF, 2021[92]; OECD, 2024[86]).
Disaggregated data collection and analytical systems enable development partners to monitor how well their interventions are reaching different population groups and tackling different dimensions of poverty and inequalities. Data can be disaggregated across various dimensions (e.g. income, ethnicity or gender), but disaggregation across all of them remains rare, given the limitations of results frameworks and lack of data in partner countries, and needs to be handled with care to minimise harm (see Section 5.1). A pragmatic approach is to prioritise the dimensions most relevant to the intervention’s objectives, context or development partner’s comparative advantage and policy tradition. For example, UNFPA’s integrated results and resources framework presents outcome indicators disaggregated by sex and by one or more “left furthest behind” characteristics (UNFPA, 2021[94]), while Finland’s results framework includes multiple indicators on gender equality and disability inclusion, and requires sex- and disability-disaggregation across many of its other indicators (MFA Finland, 2023[95]). More broadly, incorporating disaggregation mechanisms early in intervention design can strengthen the ability to track results effectively and allows for an intersectional perspective, capturing how overlapping characteristics shape differentiated experiences of poverty and inequality.
To determine an intervention’s expected level of impact and to guide monitoring processes, the Spanish Agency for International Development Cooperation (AECID) considers the following guiding questions (AECID, 2024[96]):
Are there specific indicators to measure the reduction of inequalities? Will it be possible, by the end of the intervention, to measure the “gap” between the most disadvantaged target population group and the reference group?
Will an impact assessment be carried out to analyse developments and factors that may have influenced the outcome?
Are there indicators (or will they be developed) to monitor population groups that may be more disadvantaged (in terms of opportunities, access to services and employment) or at risk of exclusion, disaggregated by relevant characteristics (e.g. age, sex, gender, ethnicity, disability, religion)? Are there geographic or spatial indicators (e.g. rural vs urban)? Are there economic indicators (e.g. related to levels of income, taxation)?
Are there plans to create or strengthen spaces for dialogue between stakeholders on the intervention and the underlying drivers of inequality in the relevant sector, country or region?
Quantitative and qualitative indicators play complementary roles in illustrating poverty and inequality dynamics. Quantitative indicators, such as household surveys and administrative data, capture the scale and distribution of outcomes, while qualitative indicators, including experience-based measures and focus groups, capture how people categorise or report their own experiences, perceptions and behaviours. Combining both in a holistic approach helps ensure that monitoring is comprehensive. This is reflected in Switzerland’s guidance on monitoring the various facets of LNOB, as it emphasises the importance of using both quantitative and qualitative data to reflect the heterogeneity of people experiencing poverty and inequalities (SDC, 2018[97]).
The International Finance Corporation’s Inclusiveness Index
This index estimates how the benefits of its projects are distributed, particularly for poor and vulnerable populations (Ray, Martin and Fernandez, 2016[98]). It combines quantitative and qualitative data to assess the distributional impacts of its interventions, in the absence of dedicated ex-ante indicators to measure poverty and inequality effects. The Index draws on a project's ex-post economic rate of return alongside qualitative descriptions of benefits to those experiencing poverty across four dimensions: employment and entrepreneurial opportunities, access to goods and services, access to finance, and improved productive capacity.
Food and Agriculture Organization’s (FAO) Voices of the Hungry
This approach uses the Food Insecurity Experience Scale to measure food insecurity based on people’s own reported experiences (FAO, 2026[99]). As an experience-based qualitative measure, it captures both material deprivation and the psychological dimension of food insecurity, such as uncertainty about future access to food. This provides a richer, more human-centred perspective but also introduces an element of subjectivity relative to objective measures. Perceptions can vary across contexts and individuals, particularly at moderate levels of food insecurity and in relation to cultural contexts and societal norms.
Ensuring indicator systems are adaptable and flexible enables them to respond to changes in context and poverty and inequality trends. Multidimensional poverty and inequality dynamics are inherently influenced by the evolving local context, as well as by political, economic and environmental shocks, among other factors. For this reason, development partners need to allow for periodic reviews and adjustments of their indicator systems based on evolving circumstances, without losing sight of their longer-term impacts and goals (OECD, 2024[86]). This flexibility is also what makes indicator systems useful inputs for adaptive management and learning via monitoring and evaluation. A practical illustration of this is the State Accountability and Voice Initiative, a six-year governance programme operating across ten Nigerian States. Rather than treating its indicators as fixed, the initiative identified “bedrock indicators” at the outcome level that remained stable, but allowed output-level indicators to change as the programme progressed (Barr, 2015[100]).
Effective indicator systems balance long-term consistency with the flexibility to respond to changing contexts, shocks and poverty and inequality dynamics.
5.2.3. Tracking poverty and inequality results using a fit-for-purpose monitoring system
A poverty- and inequality-sensitive monitoring system tracks whether interventions are achieving their objectives, how outcomes vary across individuals and groups, whether unintended effects are occurring, and whether interventions remain relevant. All development co-operation interventions, not just those explicitly targeting poverty and inequalities, can have intended and unintended effects on people experiencing different forms of deprivation and exclusion, ranging from direct harm to transformative positive change. At the same time, poverty and inequality dynamics can shift as context evolves. At minimum, it is essential that development partners:
1. Assess the extent to which the intervention is achieving its poverty and inequality objectives, analysing how outcomes vary across individuals or groups and intersecting forms of deprivation and exclusion (e.g. based on socio-economic status, gender, disability, ethnicity or geographic location).
2. Identify and address unintended negative effects on individuals or groups targeted by the intervention, as well as on those who are not.
3. Monitor changes in the operational environment that may affect the relevance and reach of the intervention.
To address inequality through its interventions, the European Commission encourages a monitoring system that serves as a learning tool to assess whether and how an intervention facilitates change in the drivers of inequality, and why (EC, 2021[101]). The system was designed through a participatory process involving stakeholders in its design, application and use. Indicators are tailored to reflect the contextual drivers of inequality and designed to track changes in these drivers over time. It uses both quantitative and qualitative indicators and focuses on four key dimensions:
Relevance: verifying that the forms of inequality of greatest concern are identified and addressed by the intervention, and that the most significant drivers are influenced;
Added value: ensuring that the intervention is complementary to others and draws greater attention to the issue of inequality;
Risks: assessing the tensions created by the intervention and any potential negative consequences for the intended beneficiaries;
Opportunity: identifying new channels for dialogue created by the intervention and assessing whether they could provide new channels for advancing action on inequality.
The poverty and inequality continuum (see Section 2.4.3 and Infographic 2.2) provides a blueprint for poverty- and inequality-sensitive monitoring. The continuum classifies interventions according to the type of change they produce – harmful, unresponsive, sensitive, responsive or transformative – and can help development partners structure their monitoring of poverty and inequality results.
At the negative end of the continuum (i.e. harmful), monitoring exercises can look for early warning signs of whether an intervention risks deteriorating living conditions (e.g. by reducing access to essential services), deepening exclusion or increasing vulnerability to shocks for specific groups. These signals can emerge even at the output level of the results chain – for instance an output indicator showing skewed reach across groups.
In the middle of the range (sensitive or responsive), monitoring can focus on the reach of interventions, examining whether they reach disadvantaged groups as intended. This can, for instance, entail tracking disaggregated output indicators across population segments.
At the transformative end, monitoring can focus on whether an intervention is contributing to longer-term, structural change for marginalised group, such as by reducing exclusion, fostering greater access to resources and opportunities and strengthening resilience. While transformative change itself is slow-moving and unlikely to be fully visible within a single programme or project cycle, output and outcome indicators showing improvements on indicators for those further behind or on measures of systemic inequalities can provide meaningful signals of progress toward transformative impact.
Poverty and inequality are shaped by long-term structural drivers that extend beyond the scope of a single monitoring system, but monitoring can still be designed with these in mind. The structural disadvantages underpinning poverty and inequality unfold over time horizons that typically exceed the project cycle. Monitoring systems can, however, track intermediate signals or proxy indicators that point to whether conditions are beginning to shift for marginalised groups, including early warning signs of emerging risks. For example, increased uptake of services by disadvantaged groups may indicate that access constraints are starting to ease, while the introduction or enforcement of restrictive laws or policies may signal increasing inequalities. Such tracking requires, where possible, going beyond output-level indicators to medium-term outcome indicators, which can provide a more meaningful picture. One practical approach is to start from impact-level indicators at the design stage and work backwards through the results chain to ensure that what is monitored during implementation remains oriented towards structural poverty and inequality reduction goals (Guerrero-Ruiz, Schnatz and Verger, 2021[102]).
Monitoring systems can track intermediate signals that point to whether long-term structural conditions are beginning to shift for marginalised groups, including early warning signs of emerging risks.
5.2.4. Ensuring that monitoring translates to adaptive learning, including through participatory and people-centred approaches
Using learning loops to link monitoring to decisions ensures it strengthens accountability to people experiencing poverty and inequalities while enabling adaptive programming. One objective of monitoring is to collect information on project or programme performance to allow development actors to report progress to relevant stakeholders, supporting accountability and transparency (OECD, 2024[86]). However, focusing exclusively on reporting risks overlooking monitoring’s potential to improve ongoing programmes in ways that better respond to the needs and experiences of people facing poverty and inequalities. Strengthening the learning loop, where timely and relevant data feeds directly into decision making, enables programmes to adapt and correct course while they are still being implemented (IPA, 2025[103]). This mirrors the importance of feeding evaluation learning into future strategies and programming (see Section 5.3.3) and highlights that improving outcomes depends not only on filling data gaps but also on using data effectively (see Section 5.1.1).
An adaptive learning approach can strengthen the link between the monitoring of poverty and inequalities and decision making. The institutional enablers of adaptive learning include structured flexibility in results frameworks, tolerance for plan revisions, alignment of monitoring timelines with decision making cycles, and designing monitoring outputs for those who can act on them, rather than solely for upward reporting. For poverty and inequality reduction programming, this requires clear institutional pathways to respond when monitoring flags up barriers, emerging risks or unintended negative effects on people facing poverty or marginalisation to ensure “do no harm” principles are followed. For example, the World Bank's Iterative Beneficiary Monitoring approach, applied in Mali and Zambia to social protection programmes targeting the poorest women, used rapid, repeated participant feedback loops, explicitly designed to trigger project adjustments rather than serve upward reporting (World Bank, 2017[104]).
However, poverty and inequality reduction programming presents specific challenges for adaptive learning. Because poverty is multidimensional, monitoring data may show progress in one dimension while masking deterioration in another (e.g. income gains alongside worsening intra-household dynamics). In such cases, adaptive decisions become complex and less straightforward. Interventions aimed at addressing deeper structural barriers such as power imbalances or institutional reform often change things slowly and unevenly, making it difficult to distinguish between a programme that is off track and one operating along a longer causal pathway. Moreover, since the most excluded populations are the hardest for data collection to reach, feedback loops designed to trigger course corrections may systematically overlook the very groups whose exclusion most urgently needs to be addressed.
Monitoring systems that engage people experiencing poverty and inequalities strengthen insights into how implementation is working in practice. Participatory approaches shift monitoring from treating communities as subjects of measurement to engaging them as contributors to data collection and interpretation, recognising their lived experience as valid evidence (see Section 4.3). Frequent, real-time and ongoing engagement captures nuanced changes in people’s lives, as well as contextual dynamics and structural inequalities that conventional and aggregate data often miss (SDC, 2018[97]), making them well suited to understanding the non-linear and complex nature of poverty and inequalities.
These approaches are also critical for strengthening feedback loops for those least likely to be heard or empowered through other processes. People facing chronic and entrenched deprivation, intersecting disadvantages, and deep marginalisation often experience barriers related to exclusion, stigma and power relations that shape how interventions work in practice but remain invisible in aggregate data (see Sections 4.3 and 5.1.2). This creates a blind spot, as these overlapping constraints influence how effectively interventions reach and serve them. Participatory monitoring helps uncover data gaps and can reach populations often excluded from conventional monitoring systems, providing more granular insights to improve implementation.
Development partners can use participatory or people-centred mechanisms to monitor interventions targeting poverty and inequalities for more responsive and accountable programming (see Table 5.6). For instance, a pilot of community scorecards in rural Bangladesh found that bringing service users and providers together in structured dialogue improved transparency and accountability in health service delivery, contributing to improved utilisation among participating communities (Hanifi et al., 2020[105]). Ritshidze, a community-led monitoring system for people living with HIV and other marginalised groups in South Africa used community-generated data to drive improvements in waiting times, access to medicines and facility staffing (Govender et al., 2024[106]). JICA’s (2025[107]) Smallholder Horticulture Empowerment Promotion project in Malawi used a high-frequency monitoring system through short, phone-based surveys collecting real-time data on income, food security and coping strategies from people experiencing poverty. The monitoring system captured household resilience to emerging shocks through survey questions on the impacts of natural disasters and rising inflation, while high-frequency data revealed the temporal dynamics of poverty, including seasonal fluctuations linked to maize harvest cycles not captured in traditional data. These mechanisms serve as a useful complement to typical logframe-style indicator frameworks, but if they are to usefully inform decisions, they need to be well-integrated into institutional monitoring and learning systems.
Engaging people with poverty and inequalities through monitoring requires an intentional focus on safety and trust. As mentioned in Section 5.1.2, managing risks to marginalised individuals related to stigma, retaliation and political surveillance (e.g. through anonymisation) is essential to uphold “do no harm” principles. Grievance mechanisms also ensure that people can safely report concerns, complaints or harmful impacts of interventions (HRW, 2015[108]). Trust is vital for sustaining meaningful engagement with people experiencing poverty and inequalities, and for generating reliable and honest information about their circumstances. It can be shaped by how monitoring is conducted: enumerators who share the social characteristics of respondents tend to generate more accurate and complete responses, particularly from marginalised groups (Rodriguez-Segura and Schueler, 2023[109]).
Evaluation is essential to ensuring that development co-operation delivers meaningful, equitable and sustainable results for people facing poverty and marginalisation. It helps assess whether interventions – understood in this context as a project, programme, policy, strategy, thematic area, technical assistance, policy advice, institution, financing mechanism, instrument or other activity – are relevant, coherent, effective, efficient, impactful and sustainable (OECD, 2023[110]). Evaluation findings can enable adjustments to ongoing interventions, strengthen future interventions and contribute to the wider evidence base on what works for poverty and inequality reduction. This is particularly important given that development partners often face challenges in translating these commitments into practice (OECD, 2024[111]); and they increasingly operate in a context of tightening budgets and competing priorities. However, the influence of evaluations ultimately depends on whether organisations have the incentives, systems and culture needed to use evaluation findings and evidence to inform decisions and learning (OECD, 2023[85]; OECD, 2024[86]).
Poverty and inequality reduction are overarching development co‑operation goals, and all interventions affect them, whether intentionally or not. Therefore, interventions across sectors and portfolios can be evaluated in terms of their approach to, and contribution to, poverty and inequality reduction. For interventions where poverty and inequality reduction is a primary objective, the full application of the considerations outlined in this section can be considered to centre evaluations on people facing the highest levels of poverty or marginalisation. For other interventions, development partners can at minimum ensure evaluations assess whether a “do no harm” approach was followed and whether any unintended effects on people facing deprivation and marginalisation occurred. A more thorough application may be justified for large-scale or high-risk interventions, or where there are clear opportunities to strengthen learning and improve outcomes for people experiencing poverty and inequalities.
The focus on poverty and inequalities in development co-operation evaluations has been inconsistent over time, reflecting changing priorities, funding availability and policy attention. An internal review of DEReC, the database of DAC members' central evaluation reports, shows that while evaluations historically centred on poverty reduction, this emphasis has declined since the Millennium Development Goals era – with poverty increasingly treated as a secondary outcome rather than a central focus (OECD, unpublished[112]). At the same time, development partners have shifted attention towards horizontal inequalities in strategic evaluations, while evaluations explicitly centred on poverty have become less common (OECD, 2024[111]).
This section provides guidance on evaluating development co-operation’s approach to, and impact on, people experiencing poverty and inequalities. It starts by examining how to apply a poverty and inequality lens to the OECD-DAC evaluation criteria (the “what”). It then outlines approaches to assessing the focus on poverty and inequalities at the institutional and project levels, as well as inclusive and people-centred evaluation methodologies (the “how”). Finally, this section focuses on how learning from evaluations can be institutionalised to inform future poverty and inequality reduction strategies and interventions (the “what next”). While important, development partners’ contributions to strengthening country or locally led evaluations and supporting national and local evaluation systems is beyond the scope of this section.
5.3.1. Applying a poverty and inequality lens to the OECD DAC evaluation criteria
The OECD DAC evaluation criteria provide a core normative framework for assessing development interventions, with each criterion serving as the basis for evaluative judgement. The six criteria – relevance, coherence, effectiveness, efficiency, impact and sustainability – are intended to be applied flexibly, depending on the purpose, context and scope of the evaluation, not as a rigid checklist. They can be used across the intervention cycle, from design to ex-post evaluation, helping to inform theories of change, assumptions and data collection strategies (OECD, 2021[113]).
The OECD guidelines Applying a Human Rights and Gender Equality Lens to the OECD Evaluation Criteria provide practical guidance to support evaluations to assess whether and how interventions contribute to realising human rights and gender equality – both for interventions with explicit human rights and gender equality objectives and those without. Applied to each criterion, this lens shifts the focus to the rights, needs and priorities of rights-holders; alignment with human rights and gender equality; differential outcomes across groups; equitable resource allocation; transformative change; and sustainable enabling conditions for inclusion (OECD, 2023[114]). Applying Evaluation Criteria Thoughtfully (OECD, 2021[113]) also highlights how to integrate inclusion when applying the criteria.
Building on these guidelines, this section provides entry points for integrating a poverty and inequality lens more broadly into each of the six criteria. It outlines key considerations and guiding questions that development partners, evaluators and evaluation managers can draw on and that can be used in the context of interventions which primarily focus on poverty and inequality reduction, as well as in other contexts where these objectives are not explicitly prioritised or targeted.
Note: This infographic presents illustrative examples of poverty and inequality (P&I) considerations relevant to the OECD DAC evaluation criteria and is not a prescriptive checklist.
Source: Authors’ compilation influenced by OECD (2023[114]; 2019[115]).
Relevance: Is the intervention doing the right things for people experiencing poverty and inequalities?
Assessing relevance through a poverty and inequality lens focuses on the extent to which an intervention's objectives and design respond to the needs and priorities of people facing deprivation and marginalisation. This includes which aspects of poverty and inequalities – rights, resources, opportunities and resilience across economic, social, political and environmental dimensions (see Chapter 1) – are most relevant to address in a given context. While relevance also considers the poverty and inequality reduction priorities of governments, civil society and other partner institutions, equitable partnership principles require that these do not overshadow those of people facing deprivation and marginalisation, recognising how power dynamics shape who participates in design and decision making. Evaluations can examine whether needs and priorities were identified at the outset through context analysis, whether the intervention remained responsive over time, and whether it enabled the meaningful participation and ownership of relevant actors, particularly those most marginalised. A key challenge is that the perspectives of the poorest and most marginalised are often insufficiently captured in programme design or documentation, particularly in interventions not explicitly targeting poverty and inequalities. Intersectionality is important here, as overlapping forms of deprivation shape needs, priorities and structural barriers in ways that differ across individuals and groups.
Understanding the intervention’s wider context is central to assessing relevance. This includes examining whether the intervention was informed by a robust contextual analysis of poverty and marginalisation, including their underlying drivers, root causes, systemic and immediate barriers and local dynamics. (see Sections 2.4.1 and 4.3.2). Evaluations can also assess relevance to locally grounded concepts of poverty and inequalities such as Ubuntu in South Africa, which emphasises interdependence and shared well-being (Government of South Africa, 2022[116]). Relevance also depends on adaptability: how well risks were anticipated and managed, and whether interventions responded to shifts in the wider context, such as shocks that alter patterns of marginalisation. This is particularly important as people living in poverty often have limited capacity to cope with shocks.
Coherence: How well does the intervention fit with broader poverty and inequality reduction efforts?
Internal coherence examines how an intervention aligns with the development partner’s poverty and inequality related policies, objectives, commitments and other interventions. It focuses on the extent to which interventions are harmonised, complementary and non-duplicative across an institution, and whether they collectively reinforce or undermine efforts and contributions to overarching policy goals on poverty and inequalities. It also examines whether the intervention aligns with relevant institutional commitments and obligations related to poverty and inequalities, including human rights frameworks, international conventions and agreements, and principles such as LNOB, do no harm, gender equality and non-discrimination. Internal coherence between poverty reduction and inequality reduction is essential, as efforts to reduce poverty may fail to address structural root causes such as marginalisation based on ethnicity, or horizontal equality initiatives may fail to account for overlapping forms of deprivation and exacerbate vertical inequalities (see Chapter 1 and Section 2.4.2). The depth of the analysis of coherence may vary depending on the intervention type and evaluation context. A risk-based approach can help to prioritise attention in sectors or situations where risks of incoherence or harm to people experiencing poverty and inequalities are greatest – such as extractives, infrastructure, land and labour policies, cross-sectoral trade-offs between economic growth and equity, or contexts where interventions may reinforce exclusion, power imbalances or unequal access to resources.
External coherence focuses on how an intervention aligns with and complements other actors’ policies, initiatives and systems that address poverty and inequalities – whether it adds value, avoids duplication and makes use of existing co-ordination mechanisms. This can include alignment with global commitments such as the SDGs, in particular SDGs 1 and 10; with partner countries’ own poverty and inequality reduction strategies and policies; and with efforts driven by people experiencing poverty and inequalities themselves, and grassroots and civil society organisations directly representing them. In this sense, external coherence can help examine whether principles of equitable partnership are upheld – supporting rather than duplicating or undermining the efforts of local actors, communities and partner countries (see Chapter 4). Beyond this, coherence also considers the extent to which an intervention is sensitive to local norms, cultural values and social practices, ensuring that poverty and inequality reduction efforts are contextually appropriate.
Effectiveness: Is the intervention achieving its poverty and inequality reduction objectives with equitably distributed results?
Assessing effectiveness through a poverty and inequality lens focuses on whether the results of an intervention are equitably distributed. Analysing differential impacts is essential, as positive aggregate results may mask an unequal distribution of benefits, with gains disproportionately captured by more powerful or better-connected groups or unevenly distributed across regions and underserved areas. Particular attention can be paid to the extent to which the intervention was grounded in ex-ante distributional analysis (see Section 2.4.4) and whether it has met the needs of those furthest behind or experiencing chronic poverty or marginalisation. This requires disaggregating results across groups as much as possible; however limited availability of disaggregated data means that a pragmatic approach is required (see Section 5.2) and that evaluators may need to adapt data collection methods.
Another critical element of effectiveness is the identification of unintended effects, both positive and negative. On the positive side, interventions can create unexpected spillovers, such as strengthening the capacity and systems of implementing partners (GIZ, 2024[117]). On the negative side, interventions may reinforce structural inequalities or increase marginalisation even when objectives are achieved – a frequent unintended consequence, particularly in fragile contexts (GIZ, 2024[117]). Evaluators can examine whether safeguards were in place to prevent and mitigate potential negative effects, and whether accountability mechanisms were adequately integrated into design and monitoring (see Section 5.2). This is particularly relevant for interventions not explicitly targeting poverty and inequalities, such as infrastructure or private sector programmes, which may inadvertently reinforce existing patterns of exclusion.
Evaluations can also assess how the intervention’s process influenced outcomes, including whether people experiencing poverty were meaningfully involved and whether the design and implementation of the intervention were tailored and responsive to the immediate and structural barriers to participation that they face, and how their participation contributed to effectiveness (see Section 4.3).
Efficiency: How well are resources being used to reach people experiencing poverty and inequalities?
Evaluating efficiency through a poverty and inequality lens requires looking beyond cost efficiency to examine whether resources are used in ways that genuinely improve outcomes for the poorest and most marginalised people. An intervention may appear efficient if it reaches many people cheaply, but this can mask a fundamental limitation: approaches that are efficient on average may fail to reach those facing the greatest barriers, such as people in remote areas, informal workers or socially excluded groups (see Table 4.5). Reaching these populations often requires more time, targeted outreach and additional investment, but it can still represent good value for money if it leads to more transformative outcomes. Evaluating efficiency therefore involves asking whether trade-offs were made between scale and inclusion, whether programme design and implementation were adapted to overcome structural barriers that keep people in deprivation and marginalisation, and how resources were allocated across groups. A key challenge is that data on resource allocation and costs are rarely disaggregated by group, making it difficult to assess efficiency at the group level.
Evaluations can also approach operational efficiency from a poverty and inequality perspective, which involves assessing whether implementation and delivery processes were adapted to the accessibility needs of people facing deprivation and marginalisation or whether feedback and participation mechanisms were used to improve delivery for the most marginalised.
An intervention may appear efficient if it reaches many people cheaply, but average efficiency can mask a failure to reach those facing the greatest barriers, such as people in remote areas, informal workers or socially excluded groups.
Impact: What significant or transformative difference does the intervention make for people experiencing poverty and inequalities?
Understanding impact through a poverty and inequality lens goes beyond the immediate results captured under the effectiveness criteria and focuses on the overall significance of an intervention for the people it targets. This encompasses both observable and emerging effects, from intermediate changes in people's lives, to longer-term and potentially transformative shifts in systems, norms and inequalities. Significance can be assessed across key dimensions of poverty and inequalities, including changes in rights, agency and power; access to resources and outcomes; opportunities; and exposure to risks and resilience (see Table 1.3). The poverty and inequality continuum can be used to evaluate impact by classifying approaches along a spectrum from harmful to transformative (see Section 2.4.3).
Assessing transformational and systemic impact involves evaluating the extent to which an intervention contributes to enduring changes. These can be contributions to wider economic, social, political and cultural systems, and to the structural barriers, norms, power dynamics and historical legacies that drive the root causes of poverty and inequalities – especially of chronic poverty and marginalisation (AECID, 2024[96]; Ray, Martin and Fernandez, 2016[98]). Short-term improvements often do not persist when systemic factors are left unaddressed. This makes it important to examine change at multiple levels – individual (e.g. improved agency or resilience), community (e.g. increased inclusion), societal (e.g. shifting norms) and institutional (e.g. policy changes or improved service delivery systems) – while recognising that transformative change is often non-linear and unevenly distributed across groups. Examining distributional impacts and unexpected outcomes of transformative change is therefore critical (Government of South Africa, 2022[116]). Evaluators can also assess whether the intervention was grounded in a theory-based or systems approach that clarifies causal pathways, reflects multidimensional drivers and accounts for political economy factors (see Table 5.14).
Assessing transformative impact presents inherent methodological challenges. First, the structural changes that drive poverty and inequality reduction often take time to materialise, creating a mismatch with typical intervention cycles that tend to focus on shorter-term results. For example, the benefits of transformative primary education may only become evident later in the life course. Second, the multidimensional and interconnected nature of poverty and inequalities means that outcomes are shaped by multiple interacting factors. Third, external shocks – such as economic cycles, policy changes or climate events – further complicate causal inference. These factors make attribution difficult and require impact assessments to rely on contribution analysis and plausible pathways of change rather than direct attribution, while recognising uncertainty and external influences (OECD, unpublished[112]). Adopting a systems approach can further support understanding of how outcomes are interconnected and shaped by context (see Table 5.14).
The structural changes that drive poverty and inequality reduction often take time to materialise, creating a mismatch with typical intervention cycles that tend to focus on shorter-term results.
Sustainability: Will the benefits for people experiencing poverty and inequalities last?
Evaluating sustainability from a poverty and inequality perspective involves assessing whether an intervention’s impacts are likely to endure over time. Building on the assessment of transformational and systemic impact, this includes examining whether the intervention has contributed to an enabling environment that will sustain benefits at individual (e.g. reduced stigma, increased resilience and awareness of rights), institutional (e.g. improved accountability, access to services, inclusive mechanisms) and systemic levels (e.g. strengthened laws, policies and systems supporting inclusion). It also requires examining whether an intervention contributed to addressing underlying systemic drivers, rather than only immediate outcomes – “treating the illness, not just symptoms” (OECD, 2021[113]).
From an equitable partnership perspective, sustainability also depends on whether local actors – from community to partner country government level – have the ownership, capacity and leadership to sustain outcomes beyond the intervention. Evaluators can assess whether clear and adaptable exit strategies were in place, with appropriate financial and capacity-strengthening measures for sustained impact for people experiencing poverty and inequalities.
Sustainability also involves assessing the long-term trade-offs and risks that may undermine sustained progress. This includes the balance between immediate impacts and longer-term effects or costs, as well as across economic, social, political and environmental dimensions. For example, support to critical health care may save lives in the short term but risks creating dependency on external funding, while growth strategies centred on labour-intensive sectors may boost GDP in the short term but risk entrenching low-wage employment. Longitudinal approaches can be particularly valuable for evaluating sustainability, as they allow evaluators to assess whether outcomes persist, strengthen or reverse over time, including in response to life events, shocks and changing contextual conditions. Risks to sustained outcomes include broader contextual and systemic factors – such as shocks, persistent exclusionary norms, changing political agendas and human rights backlashes – which lie beyond the direct control of the intervention.
Table 5.13 provides examples of development partners applying a poverty and inequality lens to the findings of specific evaluations; and Box 5.12 illustrates how this can be done within institutional evaluation guidelines and criteria to foster alignment with overarching strategic objectives on poverty and inequalities.
The International Fund for Agricultural Development (IFAD) defines the purpose of its evaluation function as improving the effectiveness of its efforts to enhance the well-being of people facing poverty in rural areas (IFAD, 2022[120]). To align with this core mandate, IFAD’s evaluation criteria guide all IFAD evaluations at the project and country levels to systematically assess whether interventions reach, benefit and transform the livelihoods of deprived and marginalised groups.
Cross referencing these with the OECD DAC criteria, this would mean examining:
Relevance: whether interventions are tailored to very deprived or marginalised groups;
Coherence: whether partnerships include organisations representing marginalised groups and expand economic opportunities for rural populations;
Effectiveness: whether results are achieved equitably across groups, accounting for gender, youth and socio-economic differences;
Efficiency: how resources are converted into outcomes across target groups;
Impact: whether very deprived or marginalised groups have substantially benefited, including improvements in income, assets, food security and nutrition;
Sustainability: whether benefits endure, with adequate engagement, ownership and resilience of rural populations experiencing poverty.
IFAD also includes a separate criterion on gender equality and women’s empowerment, assessing how interventions address root causes, intersectionality and changes in social norms and power relations. Evaluations are also expected be responsive to social justice by assessing whether programmes identify and address inequalities and examine inclusion, non-discrimination and differential outcomes across groups.
5.3.2. Adopting people-centred and equity-focused evaluation approaches
While the previous section focused on which questions evaluations seek to answer, this section focuses on how those questions can be answered. Integrating poverty and inequality considerations into evaluations requires deliberate choices, from how they are scoped and designed to how findings are generated, used and shared. This section outlines how development partners can examine poverty and inequalities through evaluations at both the institutional and project levels, applying appropriate methodologies to assess how interventions contribute towards these objectives.
Assessing the focus on poverty and inequalities at the institutional and project level
Evaluations relating to poverty and inequalities can take a wide range of forms, depending on their purpose, context and evaluating objective. They might range from ex-ante assessments of expected impacts to mid-term reviews and ex-post evaluations of final outcomes, and from formative and process evaluations to strategic, impact and economic evaluations – conducted across institutional, thematic, regional, country, programme or project levels. This section illustrates the process at institutional and programme level.
Different types of institutional evaluation offer different entry points for examining poverty and inequality integration. Strategic evaluations are particularly valuable for assessing whether poverty and inequality goals are systematically embedded into overarching policies and strategies – including through examining institutional capacities such as leadership, resource allocation, organisational culture and coherence – to determine whether the institution as a whole is oriented and equipped to deliver on these commitments (OECD, 2023[121]; UNEG, 2018[122]). Sector, programme and country programme evaluations offer a more contextualised lens, examining whether poverty and inequality considerations are tailored to specific fields (e.g. health or social protection) or national contexts, and whether these contribute to overall portfolio coherence. Institutional policy instruments (e.g. context analyses, frameworks or markers) can be assessed through both strategic and process evaluations: the former examining their influence on policy orientation, the latter how they function in practice.
Strategic poverty- and inequality-oriented evaluations of bilateral development partners have consistently highlighted a gap between strong policy commitments and the slower progress in effectively mainstreaming these priorities in practice (Meeks, unpublished[123]). Key lessons also emphasise the importance of adopting intersectional approaches, the risk that framing poverty and inequality reduction as a cross-cutting issue rather than an overarching goal may lead to it not being prioritised, the challenge of achieving and measuring transformative change, and the need for stronger monitoring and evaluation systems.
Project level evaluations can help examine whether a specific development activity or effort was designed and implemented to achieve its poverty- and inequality-related objectives. They assess whether the project reduced poverty and inequalities or exacerbated them, and can highlight how local factors such as power relations, culture and geography affected who benefited and who was left out. Evidence from these evaluations helps decision makers refine interventions, allocate resources more equitably and strengthen future monitoring and evaluation systems to track equity-focused outcomes. Process evaluations are particularly useful at this level, examining how a project or programme was implemented in practice – including management practices, tools, instruments and delivery mechanisms – and how these shaped poverty and inequality outcomes.
Formative evaluation of UNDP’s integration of the principle of leave no one behind (LNOB)
UNDP conducted an evaluation (2022[124]) of how it integrated the LNOB and “reach the furthest behind first” principles into its programmes and operations between 2018 and mid‑2022. As a formative evaluation, its primary purpose was to support organisational learning and improvement rather than to judge final impact, focusing on how well policies, systems and practices are being developed and implemented. There was also a strategic evaluation dimension, as it assessed UNDP’s overall strategic approach and positioning (including niche, value added and LNOB principles), coherence, efficiency and effectiveness in addressing inequality, exclusion and intersectionality across its global portfolio, as well as relevance to the priorities of the furthest behind. The evaluation included an institutional assessment of policies and systems, a programmatic review, in‑depth case studies using process tracing to examine political and policy influence, benchmarking against comparator organisations, and a survey of civil society actors to capture external perspectives. Following the evaluation, UNDP strengthened the operationalisation of LNOB by emphasising equality, non-discrimination and intersectionality; improving internal capacities, co-ordination and partnerships; and refining funding and measurement tools. However, gaps remain, including underdeveloped impact evaluation systems, funding constraints and uneven implementation across programmes.
Evaluating the institutionalisation of Sida’s Multidimensional Poverty Analysis (MDPA)
The evaluation of Sida’s MDPA (see Box 2.7) adopted a theory-based approach focused on understanding how the framework shapes Swedish development co-operation rather than measuring direct poverty outcomes (Ternström, Tvedten and Uggla, 2025[125]; Sida, 2024[126]). It combined a theory of change and contribution analysis to examine both formal influence and more indirect effects, such as learning and internalisation of a multidimensional view of poverty. The purpose of the evaluation was to assess whether, and to what extent, Swedish bilateral aid is influenced by multidimensional poverty analyses and how the MDPA contributes to guiding aid towards poverty reduction. It examined the institutionalisation of the approach within Sida, the outcomes it generated and its value as a steering tool, considering both practical impacts on programming and broader effects on staff understanding and organisational coherence. The methodology and process relied on mixed methods, including document review, interviews, a survey and six country case studies. Data were triangulated and analysed against a theory of change to assess MDPA’s influence on decision making and institutional practices over the evaluation period.
Korea International Cooperation Agency’s (KOICA) evaluation of the civil empowerment and self-reliance project for vulnerable people in Rwanda
KOICA’s evaluation of this self-reliance project illustrates how quasi-experimental and participatory approaches can be effectively combined (KOICA, 2025[127]). The evaluation relied on quasi-experimental methods comparing vulnerable target and non-target households to estimate impacts on outcomes (savings behaviour and asset accumulation among vulnerable households). Participatory methods were central to the evaluation design, with extensive stakeholder engagement through interviews, focus group discussions and surveys involving local government officials, project staff, community leaders and vulnerable populations participating in the project. The evaluation confirmed that activities and facilities were successfully transferred in co-ordination with the local government and that self-help groups and co-operatives continued to operate and expand, demonstrating sustained local ownership. This mixed-method approach enabled the evaluation to capture both causal evidence on economic outcomes and community perspectives.
Applying a poverty and inequality lens to evaluation methodologies
Conducting evaluations that integrate poverty and inequality considerations require deliberate methodological choices. In line with the OECD Evaluation Quality Standards, the credibility and usefulness of an evaluation depends on selecting fit-for-purpose approaches and methods that are appropriately aligned with the evaluation questions, the characteristics of the intervention, the intended use of the findings and the context in which it is implemented, through an evaluation process that is ethical, participatory, transparent and independent (OECD, 2010[128]). The evaluation methodology refers to how an evaluation is designed and conducted, including the formulation of evaluation questions, criteria and indicators; the selection of analytical tools, data collection methods and data sources; and the principles guiding the process. When evaluation questions relate to poverty and inequality reduction, the methodology can reflect these dimensions throughout by using disaggregated data and tools that capture differential impacts across population groups. Evaluations focused on specific forms of poverty and inequalities may require particular methodological adaptations. For example, disability-focused evaluations may need accessible formats, assistive technologies and targeted outreach to groups typically excluded from surveys, alongside examination of structural barriers such as inaccessible infrastructure or discrimination (UNFPA, 2020[129]). There are a range of guidelines on integrating types of poverty and inequalities into evaluations, including human rights and gender (OECD, 2023[114]), disability (UNFPA, 2020[129]), transformative equity (Government of South Africa, 2022[116]), feminist evaluations (Government of Canada, 2025[130]) and LNOB (UNFPA, 2022[36]).
As a first step, development partners can ensure that they ground their evaluations in principles and ethical standards that account for poverty and inequality considerations. This includes principles governing the evaluation process itself, such as ethical conduct, transparency, equity, inclusiveness, respect for human rights, LNOB and sensitivity to people experiencing poverty and inequalities. There may also be a specific set of principles for the intervention evaluated (see principles for participation in Table 4.6). Ethical standards are especially important when engaging with vulnerable or marginalised groups or when in sensitive contexts due to heightened risks. Evaluators can carefully consider power imbalances, protect confidentiality and data security; obtain free and informed consent; and use culturally- and context-sensitive methods to ensure they “do no harm”.
Development partners can consider a range of evaluation approaches to ensure high quality, credible evaluations of poverty and inequalities, including participatory approaches. Evaluation approaches are broader ways of thinking and overarching frameworks that guide how an evaluation is conceptualised, designed, conducted and used, with the choice of approach depending on the evaluation questions, the intervention, its intended use and the context. Different approaches can be combined to provide a more holistic evidence base, drawing on a range of methods and tools to better capture the complexity of poverty and inequalities. Participatory evaluations deserve particular emphasis: meaningfully engaging people experiencing poverty and inequalities across the full evaluation cycle – from identifying relevant stakeholders and rights-holders at the outset, through collecting and interpreting findings, to formulating conclusions and recommendations – keeps evaluations centred on people and grounded in lived realities (see Section 4.3). Choosing who is included and excluded carries important methodological and ethical implications and demands a thorough understanding of local context, power dynamics and potential risks of exclusion or harm.
Note: Section 4.3 provides more details on the meaningful participation of people experiencing poverty and inequalities
Source: Authors’ compilation based on EC (2025[131]), OECD (2023[132]) and IEG (2020[133])
The markets and seeds access project in Zambia and Zimbabwe is a long-term intervention that supports vulnerable and marginalised smallholder farmers by strengthening local and indigenous seed systems for drought‑resilient crops while linking farmers to markets to improve income, food security and resilience to climate change (Muthengo Development Solutions, 2024[134]). The mid-term evaluation used participatory methods – including through focus group discussions and case studies – allowing farmers to articulate both benefits and constraints of the project in their own words.
Farmers described tangible gains, for example noting that “I have never held K12,000 in my hands before… after selling my produce I was able to buy goods and land”, while women and people living with HIV found that the new crops were less physically demanding and detrimental to their health. Community voices also revealed ongoing vulnerability – with participants calling for solutions like irrigation, fertiliser, loans and diversification. It highlighted unexpected systemic barriers beyond the project’s scope, such as severe drought, repeated elephant raids and government unresponsiveness to smallholders. One key informant also highlighted the challenge of shifting farmers and private sector actors from high-value crops like maize to more resilient but lower-yielding small grains and legumes. Participation was however uneven, with barriers limiting inclusion of some groups. Youth were particularly difficult to mobilise, in some areas either men or women were under-represented, limiting the possibility of disaggregating perspectives across different groups. Findings informing the second phase of the project included more deliberate youth participation, improved responsiveness to droughts, drought mitigation practices and more of a focus on nutrition.
Poverty- and inequality-focused evaluations can use both quantitative and qualitative analytical or evidence collection tools (see Table 5.15). Quantitative analytical tools allow for measurable, comparable analysis of scale, impact and efficiency, enabling causal analysis of whether interventions reduced poverty inequalities (IEG, 2020[133]). However, evaluators face significant challenges with poverty and inequality data, as it may be limited or unavailable, outdated, expensive to gather and reflect biases especially when marginalised groups are concerned (see Section 5.1). Qualitative tools capture the lived experiences of those affected and provide rich contextual understanding of the often invisible social, cultural and structural mechanisms that explain how and why changes in poverty and inequality occur. They however usually have a smaller sample size and can be more costly (EC, 2025[131]). Mixed-method approaches leverage the strengths of both quantitative and qualitative evidence, enabling more nuanced and comprehensive understanding of how interventions affect different population groups through triangulation of different methods (Ray, Martin and Fernandez, 2016[98]).
Source: Adapted from (Ray, Martin and Fernandez, 2016[98]; OECD, 2023[132]; IEG, 2020[133]; EC, 2025[131])
Meaningfully integrating poverty and inequality considerations into evaluations requires sufficient resources – including institutional guidance, capacity support and learning opportunities – and can be costly (OECD, unpublished[112]; SDC, 2022[135]; UNDP, 2022[124]). Development partners can operationalise this through organisational commitments, dedicated budget lines for poverty and inequality integration in monitoring and evaluation, and resourced knowledge-sharing hubs available to both headquarters and country/regional offices. Evaluation teams would benefit from expertise strengthening and inputs from specialised experts on tools and approaches that are key for poverty- and inequality-sensitive evaluations. Where possible, engaging local experts from country or region programmes – including local feminist researchers and evaluators with contextual knowledge and language skills – can support culturally responsive approaches and strengthen participation (Government of Canada, 2025[130]).
5.3.3. Institutionalising learning from evaluations to inform future poverty and inequality reduction strategies and interventions
Transforming evaluation findings into learning and actionable improvements can inform the design and implementation of more effective interventions. By analysing what worked and why, for whom, and under what conditions, development partners can refine strategies and targeting, ensuring future interventions build on experience, avoid past mistakes and better meet the needs of marginalised populations. As discussed above, evaluating efficiency from an equity perspective requires going beyond average cost effectiveness to assess whether approaches reach those experiencing the deepest poverty and marginalisation – and feeding these insights back into future programme design is itself a critical dimension of learning.
By analysing what worked and why, for whom, and under what conditions, development partners can refine strategies and targeting, ensuring future interventions build on experience and better meet the needs of marginalised populations.
The influence of evaluations depends not only on the evidence produced and the evaluation policies and frameworks in place, but also on the broader institutional environment. This includes whether the incentives, processes, dedicated resources and organisational cultures support the systematic use of evaluation findings in decision making (OECD, 2023[85]; OECD, 2024[86]). Relevant mechanisms include mandatory management responses to evaluations, evidence review processes before programme approvals or budget allocations, and requirements to demonstrate how evaluation findings have been incorporated into strategies and investments. Integrating evaluation recommendations into annual planning, results frameworks and performance management systems can help ensure that evidence is not treated as a standalone product but becomes part of routine institutional practice.
Where evaluations demonstrate success, the findings can provide the evidence base to scale up effective solutions and guide their adaptation across contexts. Development partners can test innovative approaches at smaller scale, using ex-post evaluations to inform their expansion and identify elements that can be transferred to other settings (OECD, 2024[136]). Evaluating sustainability provides insights into how interventions can deliver lasting results, supporting decisions on scalability and replicability. Development innovation funds apply a venture-style approach to testing, evaluating and scaling cost-effective innovations with high social impact.
AFD’s Fund for Innovation in Development is an independent structure that supports innovative development interventions aimed at reducing poverty and inequalities, anchored in principles of global solidarity (FID, 2026[137]). The fund defines innovation broadly as any technological, social, organisational or financial solution that improves outcomes for vulnerable populations. Its approach combines experimentation with rigorous, counterfactual-based impact evaluations to generate evidence on “what works”, prioritising innovations based in – and led by – people from L&MICs with high impact potential, cost-effectiveness and feasibility to scale. Through staged financing, evidence-based decision making and capacity strengthening in partnership with L&MIC governments, the fund supports the scaling of cost-effective solutions to reduce poverty and inequalities.
One example is a school-based well-being programme implemented by the NGO Labhya in partnership with the Government of Tripura, India, which was evaluated using a randomised controlled trial (RCT) with support from the Abdul Latif Jameel Poverty Action Lab (FID, 2025[138]). The study assessed the impact of the Saharsh Curriculum – integrating daily 30-minute sessions focused on mindfulness, socio-emotional learning and reflective activities into primary school schedules – across 100 public schools with over 29 000 students. The evaluation found that investing in students’ social and emotional development improved learning outcomes in a cost-effective way and the programme is now planned to expand regionally to millions.
Finally, development partners can strengthen organisational learning on poverty and inequality reduction by investing in the generation and dissemination of accessible knowledge products. Short, well-structured outputs with clearly presented key lessons are particularly valuable, as they are more likely to be used by policymakers to inform and influence policy, while also being accessible to other stakeholders, development partners and local actors (for example, by identifying “best buys” or grouping findings by themes). More comprehensive products such as meta-evaluations, evaluation syntheses, rapid evidence assessments and evidence reviews can draw on the breadth of existing evaluations thereby strengthening collective learning and supporting more effective and equitable policy and practice, as illustrated in Table 5.16. Dissemination can take diverse forms, including videos, summaries, briefs, infographics and workshops, which also create opportunities for peer learning and for incorporating local insights into poverty and inequality dynamics. Finally, to strengthen accountability, development partners can share results and management responses with people experiencing poverty and inequalities involved in the project in accessible ways, tailored to overcome any language, education or other barriers (SDC, 2022[139]).
Source: Based on DEval (2025[140]), Sida (2021[141]), UNDP (2020[142]), IFAD (2016[143]) and GEEAP (2023[144]).
Ensuring development partners’ institutional frameworks and culture are aligned with poverty and inequality reduction goals, as well as able to deliver on these objectives requires a twin-track approach. As with shaping overarching goals (see Section 2.2.1) this requires both dedicated action and efforts mainstreamed throughout internal structures (including human resources management), led by management and supported by poverty and inequality experts and advisers, and implemented by all staff (OECD, 2001[145]; OECD, 2022[146]). The considerations in this section apply to development agencies, development ministries and other ministries involved in co-operation and development finance institutions, depending on each development partners’ set up. These arrangements can be calibrated to each development partner's institutional context and resources – where staffing and budgets are under pressure, flexible and shared resourcing models can help preserve necessary expertise.
Leadership commitment is a crucial starting point to establish a shared understanding of poverty and inequality reduction and foster an institutional culture that promotes these objectives (UNRISD, 2024[147]; OECD, 2001[145]). Leadership and senior management can make their commitment explicit through both words and actions, for example by ensuring that performance frameworks reward reaching the most marginalised, not just aggregate delivery targets, and by setting specific targets such as the proportion of senior staff completing poverty- and inequality-related training.
Senior management can make their commitment explicit through both words and actions, for example by ensuring that performance frameworks reward reaching the most marginalised, not just aggregate delivery targets.
Internal structures and ways of working can integrate poverty and inequality reduction across multiple levels – from overarching development goals strategies and policy frameworks to the design and appraisal of interventions monitoring, evaluation and learning. Poverty and inequality targets and standards can also be embedded into contractual arrangements and partnership agreements with external agencies, NGOs and consultants. However, internal working methods – which determine how staff interact across organisational structures – are also key to promoting the integration of these objectives. For example, facilitating staff collaboration and cross-sharing of expertise between different multidisciplinary teams helps to break down siloes and promote intersectional expertise across diverse dimensions of poverty and inequalities (e.g. between staff working on climate and social development, or on gender inequality or other types of marginalisation). Effective action on poverty and inequalities also depends on combining explicit knowledge generated through monitoring, evaluation, research and data systems, as outlined in Sections 5.1-5.3, with the tacit knowledge of staff, encompassing the skills, intuition, networks and familiarity with relevant tools and resources developed through experience and engagement on these issues (OECD, 2024[86]).
Staffing arrangements are critical to ensure adequate expertise and responsibility for poverty and inequality reduction. Appointing dedicated poverty or inequality experts helps mainstream these priorities across organisations and avoid the risk of diluting responsibility among staff with competing objectives (OECD, 2024[111]). It is important for specialists to hold relatively senior or strategic roles to be able to leverage their influence and increase the issues’ uptake and to be present at both headquarters and country offices, engaging local expertise where possible. Appointing staff “conductors” can help to connect field-based knowledge with policy analysis at headquarters.
Other complementary arrangements that development partners can consider include: i) establishing dedicated poverty or inequality units with strong mandates and explicit support from senior leadership; ii) appointing champions or focal points to raise awareness, provide guidance and foster collaboration; iii) creating or engaging with existing communities of practice or networks that promote knowledge exchange and peer learning across teams, institutions and development partners; and iv) setting up dedicated helpdesks to provide on-demand technical advice and support to staff. For example, the EC Directorate-General for International Partnerships has appointed a senior-level Advisor to the Director-General to help maintain inequality reduction high on the political agenda, as well as a network of inequality focal points across EU Delegations to support in-country mainstreaming and a dedicated HelpDesk to provide technical support and guidance on addressing inequalities. AECID has set up an internal “Inequality Network” of 40 focal points across the organisation to support the implementation of its guidelines on mainstreaming inequality reduction. Within Sida, the Chief Economist’s team is the focal point for issues related to multidimensional poverty and multidimensional poverty analysis. FCDO offers dedicated helpdesk support on selected issues of poverty and inequalities, such as disability inclusion.
Human resources policies and practices that promote staff diversity contribute to more inclusive policies and programming (OECD, 2018[148]). Greater staff diversity helps not only to ensure that different perspectives and experiences are reflected in poverty- and inequality-reducing interventions, but also to enhance development partners’ credibility and trust with partner countries, by signalling strong internal commitment to inclusion (OECD, 2022[146]). Development partners can take proactive steps to address barriers to recruiting, promoting and retaining staff from diverse backgrounds and under-represented groups – for example through measures that promote better work-life balance and equal pay, as well as strong safeguarding and protective measures to prevent discrimination, sexual harassment and abuse, in line with the 2019 DAC Recommendation on Ending Sexual Exploitation Abuse and Harassment in Development Co-operation and Humanitarian Assistance (OECD, 2019[149]).
Greater staff diversity helps ensure that different perspectives are reflected in poverty- and inequality-reducing interventions, and enhances development partners’ credibility and trust with partners.
Performance management and incentive systems reinforce both institutional and individual accountability to poverty and inequality reduction. This is particularly important where institutional incentive structures otherwise reward throughput (e.g. disbursement rates or aggregate beneficiary numbers) over equity-based outcomes such as reaching the hardest-to-reach or addressing structural barriers to inclusion. Including poverty- and inequality-related assessment criteria within performance reviews (e.g. including specific objectives contributing to the organisation’s overarching poverty and inequality reduction goals) can increase staff motivation and commitment to contribute to these objectives, while signalling the organisation’s priorities, values and performance expectations to both staff and external stakeholders (OECD, 2014[150]). Incentives may include both tangible, monetary benefits (e.g. pay awards or promotions) and intangible, non-monetary ones (e.g. professional recognition, career advancement opportunities, specialised training, international assignments or greater flexibility in job placements).
Investing in skills development is essential to ensuring that staff are sufficiently equipped to work towards poverty and inequality reduction goals. Knowledge products can contribute to this, ranging from guidelines to implementing institutional poverty and inequality strategies (see Table 2.2), to regular reports that strengthen staff’s knowledge on poverty and inequality dynamics. For instance, Norad publishes an annual report on poverty (Norad, 2025[151]). Dedicated training and learning opportunities allow staff to develop the necessary knowledge and skills to be able to integrate poverty and inequality dimensions into their work. Where possible, training can include implementing partners and be tailored to different roles and expertise levels. Making training materials publicly available also promotes the pooling of resources among development partners. Topics such as vertical (economic) inequality and intersectionality would benefit from particular attention, as both are relatively newer issues in development co-operation and have often been overlooked. For example, the Social Protection, Labour and Employment Expert Facility (SOCIEUX+) – an EU co-funded facility bringing together partners such as Belgium, France and Spain – developed an online course on “Acting upon inequality” designed to sensitise and empower both expert partners and the broader public to adopt a do no harm approach and actively contribute to reducing social and economic inequalities (SOCIEUX+, n.d.[152]) as well as a course on designing equitable social protection in health (SOCIEUX+, n.d.[153]).
The Directorate-General for International Partnerships (DG INTPA) has introduced the following organisational changes to support the mainstreaming of inequality reduction: i) a high-level senior position (Advisor to the Director-General) to keep the objective high on the political agenda; ii) a network of inequality focal points across all EU Delegations to support in-country mainstreaming; iii) a dedicated help desk providing technical support and guidance to Delegations on addressing inequalities; and iv) a public Capacity4Dev page on inequality (EC, 2026[154]), as well as training materials and learning opportunities shared on the INTPA Academy.
As part of the European Commission’s Diversity and Inclusion Strategy, DG INTPA has implemented its Diversity and Inclusion in the Workplace Action Plan 2023–2024 (EC, 2023[155]). The wider Strategy aims to promote a diverse and inclusive workplace that reflects shared European values, taking into consideration factors such as nationality, cognitive diversity and socio-economic background. To advance these priorities, DG INTPA has established a Diversity, Equality, Equity and Inclusion Pilot Team, responsible for delivering impactful change for colleagues at both headquarters and delegations, updating DG INTPA’s Action Plan with clearly defined responsibilities and timelines, and establishing key indicators and monitoring protocols. The Pilot Team’s three working groups focus on anti-discrimination, anti-harassment and bias; microaggressions; and work-life balance, considering six key areas: gender, sexual orientation, age, racial and ethnic origin, religion and belief, and disability.
References
[81] ACEIR (n.d.), About us, African Centre of Excellence for Inequality Research, https://aceir.uct.ac.za/about-us (accessed on 9 April 2026).
[57] ADB (2021), Mapping the Spatial Distribution of Poverty Using Satellite Imagery in Thailand, Asian Development Bank, https://doi.org/10.22617/TCS210112-2.
[93] AECID (2025), Tools for the Inequality-Reduction Approach, Spanish Agency for International Development Cooperation, https://www.aecid.es/documents/d/guest/2-herramientas-aecid-final-12-12-24_ingles_adelante.
[96] AECID (2024), Guide for Mainstreaming the Inequality-Reduction Approach in AECID, Spanish Agency for International Development Cooperation, https://dominicana.aecid.es/documents/d/guest/1-guia-aecid-final-12-12-24_ingles-adelante.
[82] AFD (2026), Handbook on Inequality Measurement for Country Studies, Agence Française de Développement, https://www.afd.fr/en/resources/handbook-inequality-measurement-country-studies-updated-version.
[80] AFD/World Bank (2025), Inequalities in Sub-Saharan Africa: Multidimensional Perspectives and Future Challenges, Agence française de développement, https://www.afd.fr/en/ressources/inequalities-sub-saharan-africa.
[100] Barr, J. (2015), Monitoring and Evaluating Flexible and Adaptive Programming, ITAD, https://www.itad.com/article/monitoring-and-evaluating-flexible-and-adaptive-programming/.
[119] Brasington, A. and C. Connal (2025), Evaluation of the APHRC-Funded Project “Challenging the Politics of Social Exclusion”, Sida, https://cdn.sida.se/app/uploads/2025/10/16092301/62807_DE2025_10_Ev.-of-the-APHRC-funded-project-CPSE_webb.pdf.
[46] Busher, H. and T. Godfrey-Faussett (2022), “Participatory research and the ethics of anonymisation”, Education Sciences, Vol. 12/4, https://doi.org/10.3390/educsci12040260.
[3] Carr-Hill, R. (2013), “Missing millions and measuring development progress”, World Development, Vol. 46, pp. 30-44, https://doi.org/10.1016/J.WORLDDEV.2012.12.017.
[73] Chadwick El-Ali, A. et al. (2022), Research Capacity Strengthening: Lessons from UK-Funded Initiatives in Low- and Middle-Income Countries, UK Collaborative on Development Research, https://ukcdr.org.uk/resource/research-capacity-strengthening-lessons-from-uk-funded-initiatives-in-low-and-middle-income-countries/.
[45] Charmes, J. (2019), The Unpaid Care Work and the Labour Market, International Labour Organization, https://www.ilo.org/sites/default/files/wcmsp5/groups/public/@dgreports/@gender/documents/publication/wcms_732791.pdf.
[10] Collaborative on Citizen Data (2024), The Copenhagen Framework on Citizen Data, UN Statistics Division, https://unstats.un.org/UNSDWebsite/citizenData/events/The-Copenhagen-Framework-on-Citizen-Data-new-design.pdf.
[41] Crawford, J. et al. (2021), “Equality Insights and individual-level, gender-sensitive measurement of multidimensional poverty and inequality”, Guidance Note, International Women’s Development Agency, https://equalityinsights.org/resources/guidance-note-equality-insights/.
[71] Danida Fellowship Centre (2025), Development Research: Annual Report 2024, Danida Fellowship Centre, https://dfcentre.com/wp-content/uploads/2026/05/DFC-Development-Research-Annual-Report-2024.pdf.
[77] DEEP (2026), DEEP Challenge Fund: Madagascar, Data & Evidence to End Extreme Poverty, https://povertyevidence.org/challenge-fund/madagascar-challenge-fund/ (accessed on 4 June 2026).
[140] DEval (2025), The Effects of Capacity Strengthening Interventions on Food Security and Nutrition: Evidence Synthesis of Development Cooperation in Sub-Saharan Africa, German Institute for Development Evaluation, https://www.deval.org/fileadmin/Redaktion/PDF/05-Publikationen/Berichte/2025_Food_Security_Study/2025_DEval_Report_Food_Security_web_-_Bf.pdf.
[154] EC (2026), Fighting Inequalities through development cooperation and stronger partnerships, European Commission: Directorate-General for International Partnership, https://capacity4dev.europa.eu/groups/fighting-inequalities-through-development-cooperation-and-stronger-partnerships_enguid.
[131] EC (2025), Evaluation Handbook - 2025 Edition, Publications Office of the European Union, https://data.europa.eu/doi/10.2841/0700577.
[118] EC (2025), Ex-post evaluation of Sindh Union Council and Community Economic Strengthening Support, European Commission, https://international-partnerships.ec.europa.eu/publications-library/ex-post-evaluation-sindh-union-council-and-community-economic-strengthening-support-success_en.
[155] EC (2023), Fact Sheet on the Diversity and Inclusion in the Workplace Action Plan 2023-2024, European Commission: Directorate-General for International Partnerships, https://commission.europa.eu/system/files/2023-09/fact-sheet-diversity-inclusion-in-workplace-action-plan-2023-2024_en_0.pdf.
[101] EC (2021), Addressing Income Inequalities Through Development Cooperation. Volume 3, Guidelines for Mainstreaming the Reduction of Inequality in Interventions, Publications Office of the European Union, https://data.europa.eu/doi/10.2841/269301.
[42] Equality Insights (2023), Equality Insights Rapid Report: Tonga Survey 2022, Equality Insights, https://equalityinsights.org/resources/eir-report-tonga-survey-2022/.
[55] ESA (n.d.), Fragility Conflict Security, European Space Agency, https://gda.esa.int/thematic-area/fragility-conflict-security/ (accessed on 4 June 2026).
[4] Espey, J., A. Tatem and D. Thomson (2025), “Disappearing people: A global demographic data crisis threatens public policy”, Science, Vol. 388/6753, pp. 1277-1280, https://doi.org/10.1126/science.adx8683.
[89] Eyben, R. et al. (2015), The Politics of Evidence and Results in International Development: Playing the Game to Change the Rules?, Routledge.
[99] FAO (2026), About the Food Insecurity Experience Scale (FIES), Food and Agriculture Organization of the United Nations, https://www.fao.org/measuring-hunger/access-to-food/about-the-food-insecurity-experience-scale-(fies)/en (accessed on 4 June 2026).
[78] FCDO (2025), Annual Review – GAGE 2025, Foreign, Commonwealth & Development Office, United Kingdom, https://iati.fcdo.gov.uk/iati_documents/S20352917.odt.
[79] FCDO (2025), Annual Review Template: Young Lives at Work, Improved Understanding of the Effect of Educational Opportunities on Labour Market Outcomes, Foreign, Commonwealth & Development Office, United Kingdom, https://iati.fcdo.gov.uk/iati_documents/S30110815.odt.
[76] FCDO (2025), Data and Evidence to Tackle Extreme Poverty, Foreign, Commonwealth & Development Office, United Kingdom, https://devtracker.fcdo.gov.uk/programme/GB-1-204836/summary.
[72] FCDO (2019), Research Capacity Strengthening in LMICs: A Rapid Evidence Assessment, Foreign, Commonwealth & Development Office, United Kingdom, https://assets.publishing.service.gov.uk/media/5ef4af0f86650c1294cf233b/SRIA_-_REA_final__Dec_2019_Heart_.pdf.
[91] Federal Government of Germany (2025), Transformation gemeinsam gerecht gestalten: Deutsche Nachhaltigkeitsstrategie, Weiterentwicklung 2025 (Shaping Transformation Together, Fairly: German Sustainability Strategy, 2025 Update), Federal Government of Germany, https://www.bundesregierung.de/resource/blob/975228/2335292/3962877378d74837d4f4c611749b6172/2025-05-13-dns-2025-data.pdf?download=1.
[137] FID (2026), Impact evaluation to support innovation, Fund for Innovation in Development, https://fundinnovation.dev/en/approach (accessed on 3 June 2026).
[138] FID (2025), A classroom program in India aimed at improving students’ well-being for better academic outcomes, Fund for Innovation in Development, https://fundinnovation.dev/en/projects/classroom-program-india-improving-students-well-being-better-academic-outcomes (accessed on 3 June 2026).
[69] GDN (2017), Doing Research Assessments: Understanding Research Systems in Developing Countries, Global Development Network, https://www.gdn.int/sites/default/files/Doing%20Research%20Assessment.pdf.
[144] GEEAP (2023), Cost-Effective Approaches to Improve Global Learning, Global Education Evidence Advisory Panel, World Bank Group, https://documents1.worldbank.org/curated/en/099420106132331608/pdf/IDU0977f73d7022b1047770980c0c5a14598eef8.pdf.
[117] GIZ (2024), Unintended Effects of Project Interventions Cross-Sectional Analysis, Deutsche Gesellschaft für Internationale Zusammenarbeit, https://www.giz.de/en/about-us/results-evaluation/overarching-evaluations.
[106] Govender, K. et al. (2024), “Community-led monitoring and the role of Ritshidze in improving the quality of primary healthcare in South Africa”, South African Medical Journal, p. e2043, https://doi.org/10.7196/samj.2024.v114i6.2043.
[50] Government of Brazil (2024), Busca ativa garante a entrada de mais de 3,21 milhões de domicílios no Bolsa Família, https://www.conam.com.br/busca-ativa-garante-a-entrada-de-mais-de-321-milhoes-de-domicilios-no-bolsa-familia/ (accessed on 17 March 2026).
[130] Government of Canada (2025), Planning and managing feminist evaluations, https://www.international.gc.ca/world-monde/funding-financement/planning-planification.aspx?lang=eng.
[116] Government of South Africa (2022), Integrating a Transformative Equity Criterion into Evaluations for Promoting Transformative Systemic Change, Department of Planning, Monitoring and Evaluation, South Africa, https://www.dpme.gov.za/keyfocusareas/evaluationsSite/Evaluation%20Reports/Transformative%20Equity%20Guideline_v5.pdf.
[25] GPSDD (2026), Inclusive Data Charter monitoring form - Malawi, Global Partnership for Sustainable Development Data, https://www.data4sdgs.org/sites/default/files/2026-04/Malawi_IDC%20annual%20monitoring%20form_%202026.pdf.
[59] GPSDD (2022), Participatory data governance: How small changes can lead to greater inclusion, Global Partnership for Sustainable Development Data, https://www.data4sdgs.org/blog/participatory-data-governance-how-small-changes-can-lead-greater-inclusion.
[102] Guerrero-Ruiz, A., J. Schnatz and C. Verger (2021), “A baseline survey of the guiding principles on managing for sustainable development results”, OECD Development Co-operation Working Papers, No. 93, OECD Publishing, Paris, https://doi.org/10.1787/f8974d89-en.
[105] Hanifi, S. et al. (2020), “Do community scorecards improve utilisation of health services in community clinics: Experience from a rural area of Bangladesh”, International Journal for Equity in Health, Vol. 19/1, https://doi.org/10.1186/s12939-020-01266-5.
[108] HRW (2015), At Your Own Risk: Reprisals against Critics of World Bank Group Projects, Human Rights Watch, https://www.hrw.org/report/2015/06/22/your-own-risk/reprisals-against-critics-world-bank-group-projects.
[9] IDC (2026), Six Years of the Inclusive Data Charter, Inclusive Data Charter, https://www.data4sdgs.org/sites/default/files/2025-04/6%20year%20report%20-%20IDC.pdf.
[133] IEG (2020), Evaluation of International Development Interventions: An Overview of Approaches and Methods, Indpendent Evaluation Group, World Bank, https://ieg.worldbankgroup.org/methods-resource/evaluation-international-development.
[120] IFAD (2022), 2022 IFAD Evaluation Manual, International Fund for Agricultural Development, https://ioe.ifad.org/en/w/evaluation-manual-third-edition.
[143] IFAD (2016), IFAD’s Engagement with Indigenous Peoples: Evaluation Synthesis, International Fund for Agricultural Development, https://ioe.ifad.org/en/w/ifad-s-engagement-with-indigenous-peoples-evaluation-synthesis.
[62] Indigenous Navigator (2020), Poverty Measurement, Social Programmes and Amazonian Indigenous Peoples, Indigenous Navigator, https://indigenousnavigator.org/publication/poverty-measurement-social-programmes-and-amazonian-indigenous-peoples-a-peruvian-case.
[61] Indigenous Navigator (n.d.), What is the Indigenous Navigator?, https://indigenousnavigator.org/what-is-the-indigenous-navigator (accessed on 23 March 2026).
[32] IPA (2026), IPA Embedded Evidence Lab Program, Innovations for Poverty Action, https://poverty-action.org/embedded-evidence-labs (accessed on 16 July 2026).
[103] IPA (2025), Why Learning-focused Monitoring, Evaluation, and Learning is the Best Investment You Can Make Right Now, Innovations for Poverty Action, https://poverty-action.org/why-learning-focused-monitoring-evaluation-and-learning-best-investment-you-can-make-right-now.
[107] JICA/World Bank (2025), Monitoring the Smallholder Horticulture Empowerment Promotion Project in Malawi using the Survey of Well-Being via Instant and Frequent Tracking, Japan International Cooperation Agency, https://www.jica.go.jp/activities/evaluation/__icsFiles/afieldfile/2025/03/26/MASHEPSWIFTrepot.pdf.
[68] Kanmodi, K., J. Amzat and L. Nnyanzi (2023), “The state of global research on poverty after SDG declaration”, Public Health Challenges, Vol. 2/3, https://doi.org/10.1002/PUH2.110.
[43] Kidd, S., D. Athias and I. Mohamud (2021), “Social registries: A short history of abject failure”, Working Paper, Development Pathways, https://www.developmentpathways.co.uk/publications/social-registries-a-short-history-of-abject-failure/.
[127] KOICA (2025), End-of-project Evaluation Report on the Self-Reliance Project for Vulnerable People Based on Civil Empowerment in Rwanda (2018–2020), Korea International Cooperation Agency, https://www.koica.go.kr/sites/evaluation_en/article/view/1477.
[1] Landau, A., J. Jütting and S. MacFeely (2025), “How the SDGs increased data production and what this means for policy making”, PARIS21 Discussion Paper 18, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/sites/default/files/media/document/2025-01/sdg-and-data-production-working-paper.pdf.
[48] Lindert, K. et al. (2020), Sourcebook on the Foundations of Social Protection Delivery Systems, World Bank Group, Washington, DC, https://doi.org/10.1596/978-1-4648-1577-5.
[12] Mahler, D., E. Foster and S. Tetteh-Baah (2024), How improved household surveys influence national and international poverty rates, Washington, DC, World Bank, https://doi.org/10.1596/42225.
[123] Meeks, P. (unpublished), Desk review of DAC members’ approaches to reducing poverty and inequalities through development co-operation.
[44] Metz, A., C. Casher and J. Clar (2024), ID4D Global Dataset 2021: Volume 2-Digital Identification Progress & Gaps, World Bank, Washington, DC, http://documents.worldbank.org/curated/en/099020824141510923.
[95] MFA Finland (2023), Theories of Change and Aggregate Indicators for Finland’s Development Policy, Ministry for Foreign Affairs of Finland, https://um.fi/publications/-/asset_publisher/TVOLgBmLyZvu/content/theories-of-change-and-aggregate-indicators-for-finland-s-development-policy-2020.
[134] Muthengo Development Solutions (2024), Markets and Seeds Access Project (MASAP) Mid Term Review, Swiss Development Co-operation, https://www.aramis.admin.ch/Texte/?ProjectID=56898.
[151] Norad (2025), The World’s Poor in Global Headwinds: Can Africa’s Young Population Become a Driving Force for New Progress?, Norwegian Agency for Development Cooperation, https://www.norad.no/en/news/news/2025/new-norad-report-839-million-people-live-in-extreme-poverty/.
[136] OECD (2024), DAC Guidance on Scaling Development Outcomes, Best Practices in Development Co-operation, OECD Publishing, https://doi.org/10.1787/621810cc-en.
[111] OECD (2024), Development Co-operation Report 2024: Tackling Poverty and Inequalities through the Green Transition, OECD Publishing, Paris, https://doi.org/10.1787/357b63f7-en.
[86] OECD (2024), Effective Results Frameworks for Sustainable Development: Achieving Impact by Design, OECD Publishing, Paris, https://doi.org/10.1787/aaa4bb52-en.
[114] OECD (2023), Applying a Human Rights and Gender Equality Lens to the OECD Evaluation Criteria, Best Practices in Development Co-operation, OECD Publishing, https://doi.org/10.1787/9aaf2f98-en.
[110] OECD (2023), “Evaluating development co-operation”, Development Co-operation Tips, OECD, https://www.oecd.org/en/publications/development-co-operation-tips-tools-insights-practices_be69e0cf-en/evaluating-development-co-operation_5602e1f7-en.html.
[85] OECD (2023), Evaluation Systems in Development Co-operation 2023, OECD Publishing, Paris, https://doi.org/10.1787/a255365e-en.
[121] OECD (2023), Evaluation Systems in Development Co-operation 2023, OECD Publishing, https://doi.org/10.1787/a255365e-en.
[132] OECD (2023), Glossary of Key Terms in Evaluation and Results-Based Management for Sustainable Development (Second Edition), OECD Publishing, https://doi.org/10.1787/632da462-en-fr-es.
[146] OECD (2022), DAC Guidance for Development Partners: Gender Equality and the Empowerment of Women and Girls, OECD Publishing, Paris, https://doi.org/10.1787/0bddfa8f-en.
[84] OECD (2022), Glossary of Key Terms in Evaluation and Results-Based Management: 2nd Edition - Prepublication version, OECD, https://one.oecd.org/document/DCD/DAC/EV(2022)2/en/pdf.
[113] OECD (2021), Applying Evaluation Criteria Thoughtfully, OECD Publishing, Paris, https://doi.org/10.1787/543e84ed-en.
[53] OECD (2020), “Key trends in development co-operation for national data and statistical systems”, OECD Development Policy Papers, No. 31, OECD Publishing, https://doi.org/10.1787/1ce044d2-en.
[115] OECD (2019), Better Criteria for Better Evaluation: Revised Evaluation Criteria Definitions and Principles for Use, OECD Publishing, Paris, https://doi.org/10.1787/15a9c26b-en.
[149] OECD (2019), DAC Recommendation on Ending Sexual Exploitation, Abuse, and Harassment in Development Co-operation and Humanitarian Assistance: Key Pillars of Prevention and Response, OECD, Paris.
[90] OECD (2019), Managing for Sustainable Development Results: Guiding Principles, Best Practices in Development Co-operation, OECD Publishing, https://doi.org/10.1787/44a288bc-en.
[148] OECD (2018), OECD Toolkit for Mainstreaming and Implementing Gender Equality: Implementing the 2015 OECD Recommendation on Gender Equality in Public Life, OECD Publishing, Paris, https://doi.org/10.1787/afd32b5a-en.
[150] OECD (2014), Mainstreaming Cross-cutting Issues: Seven Lessons from DAC Peer Reviews, OECD Development Co-operation Peer REviews, OECD Publishing, Paris, https://doi.org/10.1787/9789264205147-en.
[128] OECD (2010), Quality Standards for Development Evaluation, DAC Guidelines and Reference Series, OECD Publishing, Paris, https://doi.org/10.1787/9789264083905-en.
[145] OECD (2001), Poverty Reduction, The DAC Guidelines, OECD Publishing, Paris, https://doi.org/10.1787/9789264194779-en.
[112] OECD (unpublished), Meeting Summary: Second thematic discussion of the OECD-DAC Community of Practice on Poverty and Inequalities.
[6] OHCHR (2016), A Human Rights-Based Approach to Data - Leaving No One Behind in the 2030 Development Agenda, Office of the High Commissioner for Human Rights, https://www.ohchr.org/sites/default/files/Documents/Issues/HRIndicators/GuidanceNoteonApproachtoData.pdf.
[38] Open Data Watch (2025), Counting Everyone: Applying an Intersectional Lens to Data Systems in Colombia, Open Data Watch, https://opendatawatch.com/wp-content/uploads/2025/publications/Counting-Everyone-Colombia-Case-Study.pdf (accessed on 2 June 2026).
[30] OPHI (2024), How to Use National MPIs as a Policy Tool: From Metrics to Policy, Oxford Poverty and Human Development Initiative, https://ophi.org.uk/sites/default/files/2025-01/Handbook_2024.pdf.
[15] OPHI (n.d.), Oxford Poverty and Human Development Initiative (OPHI), https://ophi.org.uk/ (accessed on 20 July 2026).
[11] PARIS21 (2025), Data-Driven Policymaking? The Politics, People and Practices Behind Gender Data Use, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/sites/default/files/media/document/2025-10/gender-data-use.pdf.
[7] PARIS21 (2025), The PARIS21 Partner Report on Support to Statistics 2025: Maintaining Momentum to Protect Hard-Won Gains, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/knowledge-base/press-2025-paris21-partner-report-support-statistics-2025-maintaining-momentum.
[29] PARIS21 (2023), A Question of Quantity and Quality: How a participatory approach turned citizen-generated data into official statistics in Kenya, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/news-center/news/question-quantity-and-quality-how-participatory-approach-turned-citizen-generated (accessed on 26 June 2026).
[54] PARIS21 (2023), Using satellite images for a more detailed picture of Colombia, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/our-impact/using-satellite-images-more-detailed-picture-colombia (accessed on 21 July 2026).
[28] PARIS21 (2022), Citizen-Generated Data for SDG Reporting in Kenya, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/knowledge-base/citizen-generated-data-sdg-reporting-kenya.
[33] PARIS21 (2016), Strategic Development of Subnational Statistics, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/knowledge-base/strategic-development-subnational-statistics.
[22] PARIS21/UN Women (2025), Legal Frameworks for Gender Data across Countries Status and Opportunities, Partnership in Statistics for Development in the 21st Century, https://www.paris21.org/knowledge-base/legal-frameworks-gender-data-across-countries-status-and-opportunities.
[98] Ray, R., F. Martin and A. Fernandez (2016), Poverty, Inequality, and Evaluation: Changing Perspectives, World Bank Group, Washington, DC, https://doi.org/10.1596/978-1-4648-0703-9.
[70] Richards, C. (2025), Strengthening the Connection Between Global Evidence Synthesis and Local Users, On Think Tanks, https://onthinktanks.org/wp-content/uploads/2025/03/strengthening_the_connection_between_global_evidence_synthesis_and_local_users.pdf.
[109] Rodriguez-Segura, D. and B. Schueler (2023), “Assessors influence results: Evidence on enumerator effects and educational impact evaluations”, Journal of Development Economics, Vol. 163, p. 103057, https://doi.org/10.1016/J.JDEVECO.2023.103057.
[47] Sannon, S. and A. Forte (2022), “Privacy research with marginalized groups: What we know, what’s needed, and what’s next”, Proceedings of the ACM on Human-Computer Interaction, Vol. 6, pp. 1-33, https://doi.org/10.1145/3555556.
[139] SDC (2022), Participatory Assessment: A Methodology to Leave No One Behind, Swiss Agency for Development and Cooperation, https://www.sdc-pge.ch/dam/en/sd-web/UVD6aWWyYOxr/PGE-Slides-PA-LNOB-Launch-2022-Full-Final_EN.pdf.
[135] SDC (2022), Participatory Assessments Focusing on Intended Target Groups: A Methodology to Leave No One Behind, https://www.sdc-pge.ch/en/participatory-assessment.
[97] SDC (2018), SDC Guidance: Leave no One Behind, Swiss Agency for Development and Cooperation, https://www.sdc-pge.ch/en/leave-no-one-behind-in-practice.
[126] Sida (2024), Multidimensional Poverty Analysis (MDPA): How to Analyse Poverty, Swedish International Development Cooperation Agency, https://cdn.sida.se/app/uploads/2024/06/13135253/62693_MDPA-How-to-analyse-poverty_WEB.pdf.
[141] Sida (2021), What Works? To Reduce Poverty Through Social Protection, Swedish International Development Cooperation Agency, https://cdn.sida.se/app/uploads/2022/02/14123650/10205454_Sida_What_Works_Social_protection_Aug_2021_webb.pdf.
[152] SOCIEUX+ (n.d.), E-learning courses: Acting Upon Inequality, https://socieux.eu/elearningcourses/acting-upon-inequality/.
[153] SOCIEUX+ (n.d.), How to design equitable social protection in health?, https://socieux.eu/elearningcourses/how-to-design-equitable-social-health-protection/ (accessed on 8 June 2026).
[51] Statistics Norway (2026), Annual Report 2025: International Development Cooperation in Statistics Norway, Statistics Norway, https://www.ssb.no/en/omssb/ssbs-virksomhet/planer-og-meldinger/annual-report-2025.international-development-cooperation-in-statistics-norway.
[52] Statistics Norway (2024), International Statistics on Statelessness, https://www.ssb.no/en/befolkning/innvandrere/artikler/international-statistics-on-statelessness (accessed on 4 June 2026).
[88] Stein, D. and C. Valters (2012), “Understanding theory of change in international development”, JSRP Paper, No. 1, LSE, London, https://meguidelines.factory.iomdev.org/sites/g/files/tmzbdl2306/files/2023-03/JSRP1-SteinValtersPN.pdf.
[87] Sumner, A. (Sumner, unpublished), Why Inequality and Poverty Persist.
[125] Ternström, I., I. Tvedten and F. Uggla (2025), Evaluation of Sida’s Application of Multidimensional Poverty Analyses, The Expert Group for Aid Studies, https://eba.se/en/reports/mdpa/22679/.
[67] Thinyane, M. (2018), Engaging Citizens for Sustainable Development: A Data Perspective, United Nations University, https://sdghelpdesk.unescap.org/sites/default/files/2019-05/UNU-CS-DataSD-Report-Engaging-Citizens-20181004-compressed.pdf (accessed on 4 June 2026).
[26] UEMOA (n.d.), Programme d’Harmonisation et de Modernisation des Enquêtes sur les Conditions de Vie des ménages dans les Etats membres de l’UEMOA, https://phmecv.uemoa.int/ (accessed on 4 June 2026).
[17] UN (n.d.), SDG Global Database, https://unstats.un.org/sdgs/dataportal (accessed on 20 July 2026).
[35] UN Indonesia (2024), New UN programme to support utilization and availability of data for development at the local level, United Nations, https://indonesia.un.org/en/285978-new-un-programme-support-utilization-and-availability-data-development-local-level (accessed on 5 June 2026).
[63] UN Women (2025), How community radio stations in Zanzibar are using data to confront gender-based violence, https://data.unwomen.org/features/how-community-radio-stations-zanzibar-are-using-data-confront-gender-based-violence (accessed on 4 June 2026).
[39] UN Women (2023), Women Count: 2022 Annual Report: Consolidating Gains, Expanding Reach, UN Women, https://data.unwomen.org/sites/default/files/documents/Publications/2023/Women-Count_Annual-Report-2022_web.pdf.
[24] UNDP (2026), Malawi Surpasses 13.5 million National ID Registrations, United Nations Development Programme, https://www.undp.org/malawi/news/malawi-surpasses-13-million-national-id-registrations (accessed on 21 July 2026).
[56] UNDP (2025), The Next Great Divergence: Why AI may widen inequality between countries, United Nations Development Programme, https://www.undp.org/sites/g/files/zskgke326/files/2025-12/undp-rbap-the-next-great-divergence-background-papers.pdf.
[60] UNDP (2024), LMPI PLUS: Integrated Local Multidimensional Poverty Index for more inclusive local policies and programmes Synthesis Report, United Nations Development Programme, https://www.undp.org/africa/publications/lmpi-plus-integrated-local-multidimensional-poverty-index-more-inclusive-local-policies-and-programmes.
[124] UNDP (2022), Formative Evaluation Of the Integration by UNDP of the Principles of “Leaving No One Behind”, United Nations Development Programme, https://www.undp.org/evaluation/publications/evaluation-integration-lnob.
[142] UNDP (2020), Lessons from Evaluations: UNDP Support to Empowering Marginalized Groups, United Nations Development Programme, https://www.undp.org/evaluation/publications/empowering-marginalized-groups.
[23] UNDP (2018), Malawi tips African countries on national and civil registration, United Nations Development Programme, https://www.undp.org/malawi/news/malawi-tips-african-countries-national-and-civil-registration (accessed on 2 June 2026).
[14] UNDP (n.d.), Human Development Data, United Nations Development Programme, https://hdr.undp.org/data-center (accessed on 20 July 2026).
[122] UNEG (2018), Guidance on Evaluating Institutional Gender Mainstreaming, United Nations Evaluation Group, https://www.unevaluation.org/uneg_publications/guidance-evaluating-institutional-gender-mainstreaming.
[40] UNESCAP (n.d.), Leaving No One Behind, https://lnob.unescap.org/ (accessed on 2 June 2026).
[58] UNESCO (2023), Ethical Impact Assessment: A Tool of the Recommendation on the Ethics of Artificial Intelligence, United Nations Educational, Scientific and Cultural Organization, https://doi.org/10.54678/YTSA7796.
[36] UNFPA (2022), Guidance on Integrating the Principles of Leaving No One Behind and Reaching the Furthest Behind in UNFPA Evaluations, United Nations Population Fund, New York, https://www.unfpa.org/admin-resource/guidance-integrating-principles-leaving-no-one-behind-and-reaching-furthest-behind.
[94] UNFPA (2021), UNFPA Strategic Plan, 2022-2025, United Nations Population Fund, New York, https://www.unfpa.org/sites/default/files/board-documents/DP.FPA_.2021.8_-_UNFPA_strategic_plan_2022-2025_-_Annex_1_-_Integrated_results_and_resources_framework_-_FINAL_-_2Aug21.pdf.
[129] UNFPA (2020), Guidance on Disability Inclusion in UNFPA Evaluations, United Nations Population Fund, New York, https://www.unfpa.org/sites/default/files/admin-resource/2020_Guidance_on_Disability-Inclusive_Evaluation_FINAL.pdf.
[2] UNICEF (2024), Birth registration steadily increases worldwide, but 150 million children still ‘invisible’ - UNICEF, United Nations Children’s Fund, https://www.unicef.org/press-releases/birth-registration-steadily-increases-worldwide-150-million-children-still-invisible (accessed on 2 June 2026).
[92] UNICEF (2021), Water, Sanitation and Hygiene: A Guidance Note for Leaving No One Behind, United Nations Children’s Fund, https://www.unicef.org/media/102136/file/LNOB-in-WASH-Guidance-Note.pdf.
[66] UNICEF (2020), Producing Disability-Inclusive Data: Why it Matters and What it Takes, United Nations Children’s Fund, https://data.unicef.org/resources/producing-disability-inclusive-data-why-it-matters-and-what-it-takes/ (accessed on 26 March 2026).
[147] UNRISD (2024), Sectoral Guidance Notes for Reducing Inequalities: A Checklist for Reducing Inequalities in Development Policy and Practice, United Nations Research Institute for Social Development, Geneva, https://www.unrisd.org/en/library/publications/spotlight-2024-sectoral-guidance-notes-for-reducing-inequalities-reducing-inequalities-in-developmen.
[37] UNSDG (2022), Operationalizing Leaving No One Behind: Good Practice Note for Country Teams, United Nations Sustainable Development Group, https://unsdg.un.org/sites/default/files/2022-04/Operationalizing%20LNOB%20-%20final%20with%20Annexes%20090422.pdf.
[8] UNSTATS (2025), A Rapid Assessment of Impact and Resilience on National Statistical Systems Amid Funding Gaps and Disruptions, United Nations Statistics Division, https://unstats.un.org/iswghs/documents/funding_impact_NSO_report.pdf.
[5] UNSTATS (2025), Medellín Framework for Action on Data for Sustainable Development, United Nations Statistics Division, https://unstats.un.org/sdgs/hlg/Medellin%20Framework%20for%20action%20on%20data%20for%20sustainable%20development.pdf.
[65] UNU (2019), A Typological Framework for Data Marginalisation, United Nations University, Macau, https://i.unu.edu/media/cs.unu.edu/page/4453/UNU-MACAU_Data_Marginalization_Flyer.pdf.
[64] UNU (2019), Data Marginalization and Exclusion - Operationalizing the ’Leave No One Behind’ Principle in Social Indicators Mmonitoring, United Nations University, Macau, http://staging.unu.edu/devmedia/cs.staging.unu.edu/page/4453/UNU-MACAU_Data_Marginalization_Brochure-min.pdf.
[83] UNU-WIDER (2026), The African Inequality Review – closing the knowledge gap on inequality in Africa at a critical moment, https://www.wider.unu.edu/publication/african-inequality-review-closing-knowledge-gap-inequality-africa-critical-moment (accessed on 4 June 2026).
[19] UNU-WIDER (n.d.), World Income Inequality Database (WIID), https://www.wider.unu.edu/project/wiid-world-income-inequality-database (accessed on 20 July 2026).
[74] Vicente-Crespo, M. et al. (2021), “Institutionalizing research capacity strengthening in LMICs: A systematic review and meta-synthesis”, AAS Open Research, Vol. 3, p. 43, https://doi.org/10.12688/AASOPENRES.13116.3.
[75] Woodall, J., A. Potts and S. Brown (2024), “Embedded researchers in public health: A critical assessment”, Perspectives in Public Health, https://doi.org/10.1177/17579139231223711.
[18] World Bank (2026), World Development Indicators, World Bank Group, https://databank.worldbank.org/source/world-development-indicators (accessed on 13 July 2026).
[31] World Bank (2025), Umbrella Facility for Poverty and Equity FY25 Annual Report, World Bank Group, Washington, DC, https://documents1.worldbank.org/curated/en/099809412152579464/pdf/IDU-21c78957-b91a-449b-b43e-92b4cffa583e.pdf.
[34] World Bank (2021), World Development Report 2021: Data for Better Lives, World Bank Group, Washington, DC, https://www.worldbank.org/en/publication/wdr2021.
[49] World Bank (2019), Implementation Completion and Results Report for the Second Bolsa Familia Project, World Bank Group, Washington, DC, https://documents1.worldbank.org/curated/en/578141568314136393/pdf/Brazil-Second-Bolsa-Familia-Project.pdf.
[104] World Bank (2017), Project monitoring in fragile places does not have to be expensive, World Bank Blogs, https://blogs.worldbank.org/en/nasikiliza/project-monitoring-in-fragile-places-does-not-have-to-be-expensive (accessed on 5 June 2026).
[27] World Bank (2016), Western Africa - Regional Program to Harmonize and Modernize Living Conditions Surveys Project, World Bank Group, Washington, DC, https://documents.worldbank.org/pt/publication/documents-reports/documentdetail/854311467998818616.
[13] World Bank (n.d.), Geospatial Poverty Portal, https://pipmaps.worldbank.org/ (accessed on 2026 July 20).
[16] World Bank (n.d.), Poverty and Inequality Platform, 2026, https://pip.worldbank.org/ (accessed on 13 July 2026).
[21] World Bank (n.d.), Statistical Performance Indicators (SPI), World Bank Group, https://www.worldbank.org/en/programs/statistical-performance-indicators/about-spi (accessed on 4 June 2026).
[20] World Inequality Lab (n.d.), World Inequality Database, https://wid.world/ (accessed on 20 July 2026).