SOCIAL PROTECTION DISCUSSION PAPER No. 2510 | MARCH 2025 State of Social Protection Report 2025 The 2-Billion-Person Challenge Background Paper #3 Wake-Up Call for Social Assistance? An Unfinished Mission to Reach the Poor and Beyond Yuko Okamura Hrishikesh TMM Iyengar Colin Andrews © 2025 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: +1 (202) 473 1000; Internet: www.worldbank.org. This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. RIGHTS AND PERMISSIONS The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: +1 (202) 522 2625; e-mail: pubrights@worldbank.org. Abstract This paper examines the current state of play and trends in social assistance in terms of spending, coverage, incidence, benefit size, and poverty reduction impacts using administrative data and household survey data from about 70 countries worldwide. This paper is the fourth edition of the State of Social Safety Nets Report, following the publications in 2018, 2015, and 2014. It also serves as a background paper for the first edition of the State of Social Protection Report 2025: The 2-Billion-Person Challenge. As a major driver and component of social protection, social assistance has made considerable progress in expanding its coverage over the past decade. However, the work is still unfinished, as is evident in the persistent gaps in coverage and financing, particularly in low-income countries and those affected by fragility, conflict, and violence, where the need for support is greater. As a result, the adequacy of benefits remains low, undermining the impacts of social assistance in reducing poverty. The paper highlights potential opportunities and priorities for further investment to address these challenges and progressively expand the effectiveness of social assistance which can better support not only the poor but also broader populations in the face of shocks. For example, these include leveraging new technologies, strengthening referrals to other programs, and reforming fiscal policies. These investments are relevant and critical to make adaptive and integrated social protection systems, covering wider social protection pillars – including social insurance, and labor market and employment programs – and to ensure adequate social protection support both during normal and crisis periods. JEL Codes: D63, H53, I38, J18, O15 Keywords: Social protection, adaptive social protection, social assistance, social safety nets, coverage, incidence, adequacy, spending, financing, poverty reduction, cash transfer, subsidy reform. i Acknowledgments This paper was prepared by the Social Protection Global Department at the World Bank as a part of their flagship report, State of Social Protection Report 2025: The 2-Billion-Person Challenge (World Bank 2025). This background paper for the flagship report was authored by Yuko Okamura, Hrishikesh TMM Iyengar, and Colin Andrews, with contributions and excellent research assistance from Ana Sofia Martinez Cordova, Maria Belen Fontenez, Muhsine Senart, Johanna Estefania Andrango Brito, Ola Hesham Moustafa Hossni, and Usama Zafar. The paper was prepared under the guidance of Emil Daniel Tesliuc, Claudia Patricia Rodríguez-Alas, and Jamele Rigolini, who led the State of Social Protection Report. The authors are also grateful for the input of the authors of the other background papers on social protection, social insurance, the labor market, and gender. This paper was made possible thanks to data from the Atlas of Social Protection: Indicators of Resilience and Equity (ASPIRE), which was the primary data source for the analyses. As acknowledged in the State of Social Protection Report, the authors would like to echo and recognize the concerted efforts of nearly 200 colleagues—the ASPIRE central team, the ASPIRE regional focal points, and the country teams—to collect, update, and validate the data, which was a huge undertaking. In addition, the following colleagues kindly validated additional country data on the evolution of cash transfer programs: Juan Martin Moreno, Raquel Tsukada, Tiago Falcao Silva, Josefina Posadas, Marcela Ines Salvador, Luz Stella Rodriguez, Christabel E. Dadzie, Cornelia M. Tesliuc, Ekki Syamsulhakim, Rada Naji, Jeries Shahin, Zaineb Majoka, Gul Najam Jamy, Amjad Zafar Khan, Ruth Rodriguez, Yoonyoung Cho, Boban Varghese Paul, Victoria Monchuk, Samik Adhikari, Andrea Martin, Aziz Ben Ghachem, and Olfa Hamza Ep Hila. The team would also like to thank Ugo Gentilini, Emil Daniel Tesliuc, and Emma Wadie Hobson for allowing us to use data from their forthcoming publications (Hobson et al. forthcoming; Tesliuc et al. forthcoming). The team also would like to express our gratitude to the World Bank G2Px initiative for their valuable support in the data collection and analysis of the G2Px-ASPIRE indicators on the payment methods and gender. The team is also deeply grateful for the guidance provided by the peer reviewers: Aline Coudouel, David Coady (IMF), Carolina Diaz-Bonilla, Phillippe Leite, Anita Schwarz, Joana Silva, and Ruslan Yemtsov at the concept stage and Brooks Evans (IMF), Ugo Gentilini, Ruth Hill, Harry Edmund Moroz, and Joana Silva at the decision stage. In addition, the team is also grateful for additional comments shared by Achim Daniel Schmillen, Asha M. Williams, Christina Louise ii Acknowle dg me nts iii Lowe, Daniel Garrote Sanchez, Maddalena Honorati, Rebekka E. Grun, Stephanie Brunelin, Thomas Walker, and P. Facundo Cuevas. The team would like to thank many other colleagues who provided valuable inputs and feedback throughout the process, including Afrah Alawi Al-Ahmadi, Mohamed Bubaker Alsafi Almenfi, Sarang Chaudhary, Melis U. Guven, Georgina Marin, and Victoria Strokova. The report was produced under the guidance of Iffath Sharif (Global Director, Social Protection), Michal Rutkowski (former Global Director, Social Protection), and Loli Arribas-Baños (Practice Manager, Social Protection). The team would also like to thank the regional practice managers for their continuous support: Paolo Belli, Yasser El-Gammal, Paolo Gottret, William Wiseman, Anush Bezhanyan, Cristobal Ridao-Cano, Stefano Paternostro, Cem Mete, Robert Chase, Suleiman Namara, Christian Bodewig, and Camilla Holmemo. Lastly, the team benefited from valuable support from Fiona Mackintosh (Acquisitions Editor), Matthew Naumann (Editor), Helena Makarenko (Program Assistant), Agnes Nderakindo Mganga (Program Assistant), Amy Lynn Grossman (Production Editor), and Marcela Sanchez- Bender (Senior External Affairs Officer). Fact Sheet The analysis based on sample data highlights the following key facts. Coverage • Social assistance expanded over the recent decade by 9 percentage points, reaching 42 percent of the total population globally. It benefits 58 percent of the poorest quintile (i.e., the bottom 20 percent of the population). • Social assistance coverage represents over four-fifths of all social protection coverage (of 51 percent of total population. Social assistance coverage is more than double of social insurance coverage (of 19 percent of total population). • Coverage is significantly lower – slightly over one-fifth – in low-income countries (LICs) and fragile, conflict, and violence (FCV) settings. • On average, 58 percent of social assistance beneficiaries are female. Cash transfers have a higher female share (63-75 percent). • Rural households have a 10-percentage point higher likelihood of receiving assistance than urban households. In terms of beneficiary distribution, social assistance beneficiaries are almost evenly split between urban and rural areas due to higher urban population. Spending • Global social assistance expenditure has plateaued at around 1.5 percent of GDP since 2010. • Social assistance spending represents 28 percent of total social protection spending, at around 5.3 percent of GDP. Social assistance spending is less than half of social insurance spending, at around 3.7 percent of GDP • Social assistance spending is positively correlated with a country’s income level. High-income countries allocate more to social assistance (about 2 percent of their GDP or US$495 in per capita terms) compared to LICs (0.8 percent of their GDP or US$16 in per capita terms). • Social assistance spending increased by 50 percent during the COVID-19 pandemic years, demonstrating its countercyclical nature. By 2022, spending levels have largely returned to pre- pandemic levels. iv Fact S h e e t v • Around 70 percent of total social assistance expenditures are transferred through cash in the forms of social pensions and (un)conditional cash transfers. • Domestic resources are the main source of financing for social assistance globally (67 percent of total social assistance spending), particularly in high-income countries that entirely finance their social assistance through domestic sources. • External grants play a primary role in countries that are both LICs and affected by FCV, accounting for 77 percent of their social assistance expenditures. Adequacy • For social assistance beneficiaries, program transfers account for 11 percent of their total income or consumption which is less than two-fifths of total social protection benefits (27 percent of total income or consumption). For poor beneficiaries (i.e., those belonging to the poorest quintile), the transfers account for 18 percent of their total income or consumption. By contrast, social insurance adequacy is approximately 36 percent of total income or consumption. • In dollar terms, the average daily benefit amount is about US$1.1 (2017 PPP), translating to approximately US$33 per month. Benefits for the poorest quintile are slightly higher at US$1.4 per day or US$42 per day. • Benefit adequacy significantly varies across countries, ranging from 1 percent to 94 percent of total consumption or income of all beneficiaries. • Benefit adequacy is influenced by social assistance instruments. The most generous is the social pension with 17 percent, which is still only almost half the contributory pensions. The adequacy of unconditional cash transfers follows next (10 percent). Poverty Reduction • Poverty Headcount  : Social assistance transfers help 37 percent of extremely poor individuals (below US$2.15/day) and 11 percent of relatively poor individuals (bottom quintile) to move above their respective poverty lines. This impact is very limited in LICs (0-2 percent) owing to the scale and adequacy of programs. • Poverty Gap : Social assistance reduces the poverty gap by 45 percent for extreme poverty and 20 percent for relative poverty. The impact is larger (15-39 percent) in higher-income countries and regions with bigger coverage and adequacy. • Rural vs urban: The poverty reduction impact is greater in rural areas than urban areas, by a 6 percentage point for poverty headcount and a 9 percentage point for poverty gap compared to urban, reflecting higher coverage and adequacy in rural settings. Contents Abstracti Acknowledgments ii Fact Sheet iv 1 Introduction 1 2 Data Sources 3 3 Coverage 5 3.1 Substantial Progress Has Been Made in Extending Coverage, but There Is Still a Long Way to Go  5 3.2 Social Assistance Prioritizes the Poor but Also Benefits the Non-poor  11 3.3 Social Assistance is Key to Responding to Various Shocks When Anchored by Adaptive Delivery Systems 13 4 Spending  21 4.1 Relative Spending on Social Assistance Has Remained Virtually Unchanged  22 4.2 Spending Is Countercyclical 23 4.3 Social Assistance Spending is Dominated by Cash-based Interventions 27 4.4 There Are Stark Contrasts in Financing Outlook Affected by Country Income and (Un)stability Levels 30 5 Adequacy  33 6 Impact on Poverty  43 7 Conclusions – Summary of Main Findings and Ways Forward 49 References52 Annex 1 The Evolution of Social Assistance Coverage, Spending, and Adequacy 56 Annex 2 Economic Inclusion Implemented with Social Assistance  61 Annex 3 Flagship and Temporary Cash Transfer Programs Presented in Figure 3.7 63 Annex 4 Cash Transfers as an Enabler for Subsidy Reforms 65 vi Conte nts vii Boxes 1 Call for Better and More Data 4 2 How Social Assistance Promotes Gender Equity 19 3 Who Benefits from Social Pensions? 29 4 Who Are the Implementers of Social Assistance Programs? 32 5 How to Set the Benefit Amounts of Social Assistance Programs? 33 6 What Approaches are Used to Select Target Populations? 47 Figures 3.1 Coverage of Social Assistance as a Share of Total Population, 2006–2014 to 2015–2022 6 3.2 Coverage of Social Assistance as a Share of Total Population and the Poorest Quintile and Poverty Incidence, 2015–2022 7 3.3 Social Assistance Beneficiaries by Urban and Rural Areas and by Level of Urbanization, 2015–2022 9 3.4 Coverage of Social Assistance and Social Insurance by Welfare Quintile, 2010–2022 10 3.5 Distribution of Social Assistance by Welfare Quintile, 2015–2022 12 3.6 Distribution of Social Assistance Beneficiaries by Welfare Quintile and Instrument Type, 2015–2022 13 3.7 Coverage of Flagship and Temporary Cash Transfer Programs During Crises, 2001–2024 14 3.8 Composition of Social Protection Measures in Response to the Inflation Crisis by Instrument, 2022–2023 16 3.9 Coverage of Social Assistance as a Share of the Poorest Quintile Before and During/After COVID-19, 2015–2022 18 4.1 Social Assistance Spending as a Share of GDP and in Absolute Per Capita Amount in 2017 US dollar PPP, 2022 23 4.2 Social Assistance Spending Before, During, and After COVID-19, 2017–2022 24 4.3 Social Assistance Spending on Existing and Temporary Programs by Year, 2017–2022 25 4.4 Country-level Social Assistance Spending as a Share of GDP by Year, 2017–2022 26 4.5 Composition of Social Assistance Spending by Instrument, 2022 28 4.6 Share of Social Assistance Spending Allocated to Each Instrument, All Countries versus Only Those Countries that Use the Instrument 29 4.7 Composition of Social Assistance Spending by Financing Source, 2017–2022 31 5.1 Social Assistance Adequacy for Total Beneficiaries and for Beneficiaries in the Poorest Quintile, 2015–2022 34 5.2 Social Assistance Benefit Amounts in US$ 2017 PPP for Total Beneficiaries and Beneficiaries in the Poorest Quintile, 2015–2022 36 5.3 Social Assistance Adequacy for Beneficiaries in the Poorest Quintile in 2015–2019 (before COVID-19) and 2020–2022 (during/after the Pandemic) 37 5.4 Social Assistance Adequacy by Instrument, 2015–2022 38 5.5 Adequacy of Selected Social Assistance Programs for Total Beneficiaries and for Beneficiaries in the Poorest Quintile, 2015–2022 40 6.1 Reduction in Poverty Due to Social Assistance Transfers by Extreme and Relative Poverty Lines, 2015–2022 44 6.2 Reduction in Poverty and Inequality from Social Assistance Transfers for Relatively Poor Beneficiaries, 2015–2022 45 viii State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond 6.3 Reduction in the Poverty Headcount and the Poverty Gap due to Social Assistance Transfers for Relative Poverty by Urban and Rural Areas, 2015–2022 46 A1.1 Changes in the Coverage of Flagship Cash Transfer Programs, 2019–2022 56 A1.2 Coverage of Social Assistance among the Total Population and the Poorest Quintile, by FCV and Non-FCV contexts and in LICs and LMICs, 2015–2022 57 A1.3 Spending on Social Safety Nets and Share of Spending Outside of the Poorest Quintile 57 A1.4 Cash and Subsidy Measures Introduced in Response to Inflation Crisis (2022–2023) 58 A1.5 Composition of Social Protection Response to Inflation Crisis by Instrument, 2022-2023 59 A1.6 Pre-COVID-19 Social Assistance Spending and Country Income Level, 2019 59 A1.7 During COVID-19 Social Assistance Spending and Country Income Level, 2020–2021 60 Tables 2.1 Number of Countries Used for the Analysis, by Region and Income Group 4 4.1 Number of Social Assistance Programs included in the ASPIRE administrative Database 21 A2.1 Coverage of Social Assistance Beneficiaries 62 1 Introduction As a key pillar of social protection, social assistance plays an indispensable role in helping people to cope with poverty and vulnerability. As a non-contributory pillar of social protection systems, social assistance interventions or social safety nets focus on poor and vulnerable populations who are often unable to afford to join a contributory scheme (that is, social insurance). Social assistance has the widest coverage among all types of social protection pillars, especially in low- and lower-middle-income countries (LICs and LMICs) and among low-income households. Social assistance plays a critical role not only in addressing poverty but also in helping households respond to various risks, crises, and shocks. Social assistance has a solid track record of contributing to poverty reduction and human capital accumulation through progressive, cost- effective, and gender-sensitive policy instruments (Andrews, Hsiao, and Ralston 2018; Bastagli et al. 2016; Crosta et al. 2024; Light, Nwaobia, and Nwobia 2024; Thota et al. 2024; World Bank 2018). In recent years, social assistance has become even more essential as a series of natural disasters and crises have occurred around the world with higher frequency and intensity. For example, during the 2020 pandemic, countries scaled up their social assistance interventions to an unprecedented level, including the largest temporary expansion of cash transfers in modern history (Gentilini et al. 2022a). The world is only halfway to achieving the Sustainable Development Goals (SDGs) and is facing a wake-up call with regard to social assistance. Specifically, SDG-1.3 aims for the implementation of nationally appropriate social protection systems for all, with the goal of achieving substantial coverage of the poor and the vulnerable by 2030. With only five years to that deadline, notable progress in social assistance has been achieved over a few decades. This is illustrated by a larger number of national flagship programs, the development of delivery systems, and the integration of social protection in many national development strategies. However, the analysis from this paper shows that about three in five people worldwide still lack access to social assistance. There are clear gaps not only in coverage but also in financing and the adequacy of benefits. Globally, investment in social assistance has plateaued at around 1.5 percent of gross domestic product (GDP), and the level of social assistance benefits has remained virtually unchanged at around 11 percent of their welfare level (pre-transfer) over the past decade. 1 2 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Continuing with business as usual will be insufficient to address existing gaps and limitations, especially in low-income and fragile settings where challenges and needs are greatest. Based on current trends, it would take several decades to achieve universal coverage, and it is estimated that an additional US$311 billion would be required to lift those in the bottom 20 percent of the population out of poverty even after increasing the efficiency of current social assistance spending (World Bank 2025). While the challenges are real and difficult to overcome, there are also opportunities and innovations that have the potential to accelerate progress, such as leveraging new technologies, strengthening referrals to other programs and services, reforming fiscal and public financing, and piloting innovative financial mechanisms. Efforts to strengthen social assistance in adaptive social protection systems show promise and a pathway forward. Recent crises have underscored the critical need for countries to have strong and adaptive social protection system1 in place that can ensure continued investments, stronger institutional collaboration, and effective delivery, even during crisis periods. In some countries, together with the country’s foundational systems (for example, ID systems, interoperability, and payment systems), large-scale national cash transfer programs have served as the backbone of their adaptive social protection systems, backed up by a vision to design and deliver various interventions beyond social assistance (for example, those related to health and nutrition, economic inclusions, and social services) that can support the resilience and productivity of different populations. This paper is structured as follows. Section 2 describes the data used in the analyses, namely samples from household surveys and administrative data from the World Bank’s global database on social protection. Section 3 tracks progress in coverage over time, space (regional distribution and urban versus rural areas), and population groups (by income and gender). Section 4 examines social assistance spending—its level, its changes over time, its distribution by regions and by country-income levels, its instruments, and its sources (domestic versus external). Section 5 explores the adequacy of social assistance benefits, with some country and program examples. Section 6 estimates the impacts on social assistance in reducing poverty and inequality. The paper concludes with a set of closing remarks and ways forward. 1. See Bowen et al. (2020) for building blocks of Adaptive Social Protection system. 2 Data Sources The main source of data used for this paper is the Atlas of Social Protection: Indicators of Resilience and Equity (ASPIRE), the World Bank’s global social protection and labor database. The ASPIRE database consists of administrative program-level data and/or national household surveys from 153 countries, which can be used to generate key indicators. The authors used two subsets of the ASPIRE database: (a) household survey data from 68 countries,2 for which comparable data are available for 2006–2014 and for 2015–2022 and (b) administrative data from 76 countries for which the latest information was collected for 2017–2022. Administrative data are used for analyzing expenditures (Section 4), and other sections (Sections 3, 5, 6 for coverage, adequacy, and poverty reduction, respectively) are based on household survey data. Table 2.1. presents the distribution of the two sets of data by region and income level. (Note that certain regions or income groups [for example, Middle East and North Africa3] have a small sample of countries and hence may not be representative of such subcategories. These data challenges call for better and more data (see box 1).) Although these two datasets were primarily used for most of the analyses presented in this paper, the authors clearly flag any analyses that used different data from these base samples. 2. Out of 72 countries used in the State of the Social Protection Report (World Bank 2025), this paper uses the data from 68 countries because the four remaining countries did not have information on social assistance. 3. For example, the household survey data is available for only three countries in the Middle East and North Africa region, namely Egypt, the West Bank and Gaza, and Tunisia and therefore may not be representative of the entire region. 3 4 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Number of Countries Used for the Analysis, by Region and Income Group TABLE 2.1   The Paper uses administrative data from 76 countries and household survey data from 68 countries. Region and country income level Countries with Countries with Overlap Administrative data Household Survey data Region  East Asia and Pacific 8 5 3  Europe and Central Asia 23 15 9  Latin America and the Caribbean 12 16 9   Middle East and North Africa 8 3 2   South Asia 6 6 4   Sub-Saharan Africa 19 23 8 Income  Low Income Countries 11 10 5  Lower Middle Income Countries 21 22 10   Upper Middle Income Countries 31 30 17  High Income Countries 13 6 3 Total 76 68 35 Source: Original table for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey and administrative data. BOX 1  Call for Better and More Data The analysis presented in the paper highlights the need for more and better data for use in programming and decision-making. Good policies need to be based on accurate and reliable data, but many information gaps remain, especially in LICs and in FCV settings. There is a need to increase the availability, open-accessibility, and relevance of data from both administrative sources and household surveys while also improving their quality. The minimum amount of information needed for countries to be able to monitor the size, composition, and evolution of their social protection systems is the number of beneficiaries of and expenditures on all of their large and medium-sized social protection and labor programs, including social assistance interventions. However, there are several significant barriers to collecting these kinds of administrative and household survey data in many developing countries, including institutional fragmentation, the lack of a clear designation of institutional responsibility for monitoring, and the existence of many small, ad hoc, and short-term social assistance programs. Therefore, governments will need to make greater systematic efforts and investments to gather more and better data going forward. 3 Coverage Globally, social assistance remains the bedrock of social protection coverage. Social assistance reaches over four-fifths of all social protection beneficiaries.4 In LICs, the coverage of social assistance is essentially equivalent to the coverage of all social protection,5 but this coverage remains low compared to the vastness of the need. Using household survey data from 68 countries as its main source, this section discusses the state of social assistance coverage, both historically and currently, its distribution by welfare level, and its recent scaling up in response to the shocks. Substantial Progress Has Been Made in Extending Coverage, but 3.1  There Is Still a Long Way to Go Globally, social assistance has expanded across different country contexts—all income levels and regions—over the recent decade. Comparing the two periods 2006–2014 (that is, circa 2010) and 2015–2022 (that is, circa 2020), the proportion of the population that receives social assistance increased by 9 percentage points, from 33 percent to 42 percent, based on a sample of 68 countries (Figure 3.1). While various factors (such as political economy and leadership, institutional capacity, financing, and the social contract) influence the scope of social assistance in any given nation, countries at all income levels have managed to substantially increase the coverage of their social assistance programs (10 to 15 percentage points), with the exception of UMICs where the increase was limited to only 3 percentage points. LICs increased their coverage by 10 percentage points (that is, 80 percent increase between circa 2010 and circa 2020) and LMICs by 15 percentage points (that is, 60 percent increase). 4. Global social protection coverage is 51 percent, with social assistance contributing the most by far— 42  percent compared to 19 percent for social insurance and 5 percent for the labor market (see Figure 2.2 in World Bank 2025). This social assistance coverage contribution rate reduces to two-thirds when including social insurance contributors. 5. The difference between the coverage of social protection and of social assistance increases with a country’s income level. The difference is small in LICs (25 percent and 22 percent) and in LMICs (42 percent and 40 percent) and grows to 16 percentage points in UMICs and HICs (see Figure 2.2 in World Bank 2025). 5 6 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE 3.1  Coverage of Social Assistance as a Share of Total Population, 2006–2014 to 2015–2022 (n = 68 countries) Global coverage expansion across a variety of contexts and income levels. Social assistance coverage (% of total population) 70 70 66 60 57 55 53 50 49 45 42 40 40 42 40 36 33 32 33 33 30 29 23 25 22 19 20 12 10 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 68) (n = 5) (n = 15) (n = 16) (n = 3) (n = 6) (n = 23) (n = 10) (n = 22) (n = 30) (n = 6) Circa 2010 Circa 2020 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Notes: Based on a sample of 68 countries (n = no. of countries). Circa 2010 takes the closest year to 2010 between 2006 and 2014, and Circa 2020 takes the latest coverage data available between 2015 and 2022. MNA has three countries (Egypt, West Bank and Gaza, and Tunisia) and may not be representative of the entire region. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. The creation of national flagship cash transfer programs in many countries has been a hallmark of recent decades, which has contributed to the expansion of social assistance coverage. The emergence and evolution of cash transfers6 was often prompted by major crisis events such as the Asian financial crisis in 1997, the global financial crisis of 2007–2009, the COVID-19 pandemic from 2020 to 2021, Russia’s invasion of Ukraine in 2022, and the inflation crisis of 2022–2023. Also, the evolution of cash transfers often occurred in a broader context of social protection sector and subsidy reforms, which meant not only the implementation of more progressive cash transfers per se but also the establishment and strengthening of country’s delivery systems, which can better identify, enroll, manage, and pay to beneficiaries of various social programs beyond social protection. 6. Only a few countries globally (Bangladesh, Brazil, and Mexico) had conditional cash transfers in 1997 (Fiszbein and Schady 2009). C ove rage 7 Although progress has been made, significant coverage gaps persist, especially in LICs.7 The  current reach of social assistance is limited to 42 percent of the total global population and 58 percent of those in the poorest quintile based on a sample of 68 countries (Figure 3.2). However,  social assistance covers a smaller fraction of the population where the needs are greatest, as can be seen by the inverse relationship between poverty incidence and social assistance coverage. These two measures—social assistance coverage and poverty rates—diverge widely, especially in LICs, where 43 percent of the population is poor, while social assistance only covers 22 percent of the population. The fragility, conflict and violence (FCV) status is another factor that affects the size of social assistance in LICs and LMICs where its coverage is 9 to 11 percentage points lower than in their peer non-FCV countries (see Annex Figure A1.2).8 FIGURE 3.2  Coverage of Social Assistance as a Share of Total Population and the Poorest Quintile and Poverty Incidence, 2015–2022 (n = 68 countries) Social assistance reaches the fewest people where the need is greatest. Social assistance coverage (% of the total population and bottom 20% of population [i.e., Q1]) 100 85 83 79 80 70 66 66 60 58 57 51 53 53 49 45 42 44 40 40 40 33 36 33 27 22 20 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 68) (n = 5) (n = 15) (n = 16) (n = 3) (n = 6) (n = 23) (n = 10) (n = 22) (n = 30) (n = 6) SA ALL SA Q1 % of pop < NPL % of pop < $2.15 % of pop < $3.65 % of pop < $6.85 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Poverty incidence from World Bank’s Poverty and Inequality Platform. Notes: Based on a sample of 68 countries (n = no. of countries), except for the national poverty line, which is based on 67 countries, as there are no data for Brazil. Middle East and North Africa has three countries (Egypt, West Bank and Gaza, and Tunisia) and may not be representative of the entire region. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. SA = Social Assistance, ALL = total population, Q1 = Poorest or First Quintile, “% of pop <” = Share of population below. NPL = National Poverty Line. $2.15, $3.65 and $6.85 are poverty lines for extreme poor, LMICs and UMICs, respectively. 7. Around 9 percent and 50 percent of the global population live in extreme poverty and poverty, respectively. This means that 712 million people live in extreme poverty (World Bank 2024). Also, while extreme poverty is highly concentrated in LICs, poverty—including relative poverty and inequality—persists in HICs, underscoring the relevance of social assistance in all country contexts, regardless of their economic level. For example, the poverty rate in the countries of the Organisation for Economic Co-operation and Development (OECD) ranged from 6 percent to 21 percent in 2022 (OECD 2024). 8. Based on data from 32 LICs and LMICs consisting of 9 FCV countries and 23 non-FCV countries. 8 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Social assistance is the main pillar of social protection in many countries, but in some high- income countries, social insurance is the main driver. In Sub-Saharan Africa and South Asia, the reach of social assistance is limited to only around 30 percent of the total population (and less than half of the poorest quintile). Such low levels of social assistance coverage are a concern because these regions also have the lowest rates of social protection coverage, at 38 and 43 percent, respectively (see World Bank 2025). On the contrary, a similar level of low social assistance coverage in Europe and Central Asia countries (33 percent) is less concerning due to their extensive social protection reach (84 percent), supported by more mature social insurance systems. While Latin American and the Caribbean countries have a similarly high level of social protection coverage (84 percent) as Europe and Central Asia countries, their systems are geared more toward providing social assistance like the two remaining regions, which is also the case in East Asia and the Pacific and in the Middle East and North Africa. The recent rate of expansion and the current level of coverage imply that it would take several decades to achieve universal coverage if nothing were to change (World Bank 2025). Social assistance still has a long way to go to fully cover the bottom 20 percent of the population, as more than two in five of these people currently have no access to it. The task is even harder in lower- income countries where the current coverage rate is the lowest and the need (and relevance) for social assistance is the highest, far beyond the bottom 20 percent of the population. Social assistance continues to prioritize rural areas, with economic growth and urbanization directly influencing the distribution of beneficiaries between rural and urban areas. Based on the disaggregated data for coverage by rural and urban areas from 60 countries,9 the likelihood of rural households receiving social assistance is 10 percentage points higher than that of urban households (Figure 3.3a). This rural-urban gap varies between 4 and 14 percentage points depending on the country’s level of economy and the region where it is located. However, while looking at coverage in terms of absolute numbers (rather than shares), there are more social assistance beneficiaries in urban areas than in rural areas, especially in UMICs and HICs, because of the higher populations in urban areas,10 resulting from population migrating to cities in search of jobs and a better standard of living (Figure 3.3b).11 9. Out of 68 countries with coverage data, 8 countries were excluded either because there were no data on the rural population or because the survey sample was small (fewer than 60 households) and non- representative of the rural or urban population receiving social assistance. 10. For example, Colombia has a population of almost 52 million individuals and is highly urbanized, with only 18 percent of its population living in rural areas. Social assistance programs cover around 52 percent and 25 percent of the rural and urban populations, respectively. Since more than 80 percent of the population lives in urban areas, more people receive social assistance in urban areas (11 million) than in rural areas (5 million). However, people in rural areas have greater chance of receiving the transfer, as one in every two individuals in rural areas is likely to be receiving a transfer compared to only one in four individuals in urban areas. 11. As of 2021, more than half of the global population live in urban areas, and this trend of urbanization is expected to increase to 68 percent by 2050 (UN Habitat 2022). C ove rage 9 Social Assistance Beneficiaries by Urban and Rural Areas and by Level of Urbanization, 2015–2022 FIGURE 3.3   (n = 60 countries) Social assistance continues to focus on rural poor, while urbanization leads to more beneficiaries in cities. a. Coverage of social assistance as a share of total population, by urban and rural areas (with share of rural population) Social assistance coverage (% of the population) 80 75 69 71 66 64 69 65 60 55 57 59 54 51 45 47 41 46 42 46 41 41 41 36 37 39 40 35 28 30 30 37 22 34 20 24 26 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 60) (n = 5) (n = 12) (n = 15) (n = 3) (n = 6) (n = 19) (n = 8) (n = 21) (n = 25) (n = 6) Rural Urban % of pop living in rural areas b. Distribution of social assistance beneficiaries and population, by urban and rural areas Share of social assistance coverage and share of population, by urban and rural areas 100 23 31 36 33 31 80 41 38 49 48 46 55 53 55 59 61 63 61 66 70 76 72 74 60 40 77 69 64 67 69 59 62 51 52 54 45 47 45 41 20 39 37 39 34 30 24 28 26 0 Ben Pop Ben Pop Ben Pop Ben Pop Ben Pop Ben Pop Ben Pop Pop Pop Ben Pop Ben Pop Ben Pop Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 60) (n = 5) (n = 12) (n = 15) (n = 3) (n = 6) (n = 19) (n = 8) (n = 21) (n = 25) (n = 6) Rural Urban Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Population data from World Development Indicators. Note: Based on a sample of 60 countries (n = no. of countries). Middle East and North Africa has three countries (Egypt, West Bank and Gaza, and Tunisia) and may not be representative of the entire region. Pop = Share of Population, by rural and urban areas; Ben = Share of Beneficiaries, by rural and urban areas. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. 10 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Social assistance coverage inversely relates to social insurance, with people in higher-welfare categories and higher-income countries benefiting more from social insurance than social assistance. Globally, social assistance coverage is more than double of social insurance coverage (42 percent versus 19 percent). In short, as countries get richer, the coverage of social protection also evolves. To look at the relationship of these two pillars of social protection, Figure 3.4 uses data from 61 countries for which distributional information on the coverage of social assistance and social insurance were available. The coverage of social assistance decreases as households become richer, but social insurance coverage—those who either contribute to or receive benefits (for example, pensioners) from the system—has a positive correlation with households’ welfare levels. Given that the aggregate of the two rates hovers around 30–40 percent in LICs and 60–70 percent in LMICs, a significant proportion of their population—across all welfare quintiles—is clearly being excluded from any social protection mechanism. In addition, in higher-income countries, there is a notable overlap between social assistance and social insurance, indicating that many households can avail a mix of different services and benefits when they need.12 The point on the welfare distribution at which social insurance reaches more people than social assistance does not occur until the top quintile (the richest 20 percent of the population) in LICs (see Figure 3.4a), whereas this point gradually and systematically moves down into the lower quintiles as a country’s income level rises (see Figures 3.4b, c, and d). For HICs, this is as low as the second poorest quintile (see Figure 3.4d). FIGURE 3.4  Coverage of Social Assistance and Social Insurance by Welfare Quintile, 2010–2022 (n = 61 countries) Higher welfare means less access to social assistance and more access to social insurance. a. LICs (n = 11 countries) b. LMICs (n = 27 countries) SA-SI coverage (% of population in each quintile) SA-SI coverage (% of population in each quintile) 100 100 80 80 60 60 45.5 45.3 40 40 27.4 20.7 20 20 25.3 2.8 19.3 17.9 0 0 Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Social assistance (SA) Social insurance (SI) (continued) 12. Governments can strengthen their social protection systems which can provide layered support by offering different programs to support individuals and households throughout the life cycle and across the income spectrum. This also includes creating pathways for people to access and adjust benefits as their needs change (World Bank 2022b). C ove rage 11 FIGURE 3.4  Coverage of Social Assistance and Social Insurance by Welfare Quintile, 2010–2022 (n = 61 countries) (continued) c. UMICs (n = 20 countries) d. HICs (n = 3 countries) SA-SI coverage (% of population in each quintile) SA-SI coverage (% of population in each quintile) 100 100 89.1 93.7 80 69.4 69.4 80 60 60 40 40 50.8 25.7 33.6 20 30.4 20 0 0 Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Social assistance (SA) Social insurance (SI) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Notes: Based on a sample of 61 countries (n = no. of countries). Social insurance coverage is based on both beneficiaries (recipients) and contributors, capturing all those affiliated with the contributory scheme. The most recent survey with data on social assistance (SA) and social insurance (SI) coverage from 2010–2022 is used. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries. Q1 = Poorest or First Quintile, Q2 = Second Quintile, Q3 = Third Quintile, Q4 = Fourth Quintile, Q5 = Fifth Quintile. Social Assistance Prioritizes the Poor but Also Benefits the Non-poor 3.2  Overall, social assistance continues to be progressive, favoring poor people, but it also benefits the non-poor. Not all social assistance programs are designed to or aimed at reducing poverty.13 Based on the country’s context, priority, and policy needs, the objectives of these programs can vary. For example, they might aim to support specific population categories (such as the elderly and people with disabilities), improve food and nutrition security, health and education outcomes, or provide emergency or temporary income support. In practice, nearly half of all social assistance programs use categorical criteria to select their beneficiaries, whereas household welfare assessments are used much less often (see Box 6. in Section 6). As a result, the beneficiaries of social assistance tend to be concentrated in the poorest two quintiles, accounting for more than half of all beneficiaries, while spreading across all welfare levels (Figure 3.5). This suggests that many countries have the scope and potential to better target their social assistance programs to reach the poor more effectively, even within existing budget constraints, thereby maximizing the impact on poverty reduction (see Grosh et al. 2022 for more information). 13. In addition, perfect targeting with no inclusion and exclusion errors is not feasible. 12 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE 3.5  Distribution of Social Assistance by Welfare Quintile, 2015–2022 (n = 68 countries) Social assistance primarily focuses on the poor, but also benefits the non-poor including the rich. Distribution of social assistanace beneficiaries 100 11 14 9 9 10 11 14 15 12 9 13 12 16 14 14 80 15 15 16 17 17 17 18 17 19 19 20 18 19 60 20 19 20 20 21 24 24 25 24 25 25 24 40 23 22 20 23 20 37 34 34 30 26 30 30 27 27 29 25 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 68) (n = 5) (n = 15) (n = 16) (n = 3) (n = 6) (n = 23) (n = 10) (n = 22) (n = 30) (n = 6) Q1 Q2 Q3 Q4 Q5 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 68 countries (n = no. of countries). MNA has three countries (Egypt, West Bank and Gaza, and Tunisia) and may not be representative of the entire region. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Q1 = Poorest or First Quintile, Q2 = Second Quintile, Q3 = Third Quintile, Q4 = Fourth Quintile, Q5 = Fifth Quintile. The progressiveness of social assistance depends on a country’s policy objectives and program’s targeting performance, with distributional impacts varying substantially by country-income level, region, and choice of instrument. For example, the share of social assistance beneficiaries from the poorest quintile among total beneficiaries is the highest in Europe and Central Asia (37  percent), followed by Latin America and the Caribbean (30  percent), and South Asia (30 percent). However, it is only 26–27 percent in Sub-Saharan Africa, and East Asia and the Pacific. Of the various social assistance instruments that exist, conditional cash transfers (CCTs) and social pensions are the most pro-poor instruments, with fee waivers being one of the least pro-poor (Figure 3.6). CCTs—many of which are poverty-targeted—continue to outperform other instruments by a wide margin, with 45 percent of beneficiaries being in the poorest quintile, and nearly three-quarters of beneficiaries being in the poorest 40 percent of the population. C ove rage 13 FIGURE 3.6  Distribution of Social Assistance Beneficiaries by Welfare Quintile and Instrument Type, 2015–2022 (n = 68 countries) Pro-poorness significantly varies by instrument, with CCT being a steadfast champion. Distribution of social assistance beneficiaries by instrument 100 4 9 9 9 9 8 13 11 9 14 12 13 14 15 80 16 15 15 18 17 18 20 21 19 60 18 27 25 24 25 25 26 22 24 40 20 45 35 38 35 33 30 31 30 0 CCT UCT Social PWP In-kind School Fee Social (n = 22) (n = 55) pension (n = 10) (n = 39) meals waiver assistance (n = 37) (n = 28) (n = 29) (n = 68) Q1 Q2 Q3 Q4 Q5 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 68 countries (n = no. of countries) out of which 10–54 countries implement each instrument. CCT = Conditional Cash Transfer, UCT = Unconditional Cash Transfer, PWP = Public Work. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Q1 = Poorest or First Quintile, Q2 = Second Quintile, Q3 = Third Quintile, Q4 = Fourth Quintile, Q5 = Fifth Quintile. Social Assistance is Key to Responding to Various Shocks When 3.3  Anchored by Adaptive Delivery Systems Other than helping reduce chronic poverty during normal times, social assistance has increasingly been playing an important role in protecting people from shocks through temporary expansion. During the COVID-19 pandemic, social assistance—particularly unconditional cash transfer (UCT) intervention—grew at an unprecedented scale, though the magnitude of the temporary assistance provided varied widely across countries (Gentilini et al. 2022a). Annex Figure A1.1 show the extent to which cash transfer programs were scaled up during the COVID-19 pandemic using administrative data. Marin and Palacios (2022) also estimate that 1.7 billion people in low- and middle-income countries (or roughly one in five people in the world) received emergency assistance payments, over half of them for the first time. In several countries, national cash transfer programs have been serving as the backbone of adaptive social protection responses. Figure 3.7 depicts the coverage of the flagship cash transfer programs in 12 countries plus the coverage of emergency/temporary transfers provided during shocks. It shows that these countries have managed both gradual expansions in coverage and swift expansions to adapt to crises. While a few countries (such as Ghana 14 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond and South Africa) gradually increased their coverage, the majority temporarily scaled-up cash transfer interventions during COVID-19. Some countries expanded their already existing programs (for example, Pakistan), while other countries not only expanded their existing programs but also introduced new programs to reach new beneficiaries (for example, Brazil, Argentina, and Colombia). Reflecting the nature of the shocks, the scale-up of cash transfers from temporary programming was followed by a scale-back in the succeeding years in all these countries, except for South Africa, which implemented the Social Relief of Distress Grant (R350) and has one of the few known COVID-19 programs that are still active today (at the time of writing this report). FIGURE 3.7  Coverage of Flagship and Temporary Cash Transfer Programs During Crises, 2001–2024 (n = 12 countries) Investment in delivery systems during normal times was crucial for adaptability during shocks. Coverage (percentage of total population) of flagship or emergency cash transfer programs 70 60 50 40 30 20 10 0 00 20 1 02 03 04 05 06 07 08 09 10 11 12 13 14 24 15 16 17 18 19 20 20 1 22 23 0 2 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Brazil (PBF/AE/AB/nPBF) Argentina (AUH/IFE/CT Retires and Pensioners) Colombia (MFA/SI) Pakistan (BISP/EEC) Philippines (4Ps/SAP) Indonesia (PKH/BLT DD/BLT minyak goreng) Jordan (Takaful/NAF old CT/ECT) Tunisia (PNAFN/AMEN PCT/FA/TCT) South Africa (CSG) Sierra Leone (Ep Fet Po/ECT/COVID Ep Fet Po) Ghana (LEAP/ECT) Chile (Single Family Allowance/IFE) Source: Original figure for this publication using data from the World Bank’s country teams, Hobson et al. (forthcoming), the World Bank’s price shock tracker (Gentilini et al. 2023), the World Bank’s COVID-19 tracker (Gentilini et al. 2022a), the ASPIRE administrative database, and various other sources including government agencies, multilateral organizations (such as the IMF, the OECD, and the World Bank), and humanitarian agencies (United Nations [UN] global and regional agencies including the UN Economic Commission for Latin America and the Caribbean [UNLAC]). Note: Based on a review of programs in 12 countries (n = no. of countries). 4Ps = Pantawid Pamilyang Pilipino Program; AB = Auxílio Brasil; AE = Auxílio Emergencial; AMEN PCT = AMEN Permanent Cash Transfer; AUH = Asignación Universal por Hijo para Protección Social; BISP = Benazir Income Support Programme; BLT DD = Bantuan Langsung Tunai Dana Desa; CSG = Child Support Grant; CT = Cash Transfer; ECT = Emergency Cash Transfer; EEC = Ehsaas Emergency Cash Transfer; FA = Family Allowance; IFE = Ingreso Familiar de Emergencia; LEAP = Livelihood Empowerment Against Poverty; MFA = Más Familias en Acción; NAF = National Aid Fund; nPBF = New Bolsa Familia Program; PBF = Bolsa Familia Program; PKH = Program Keluarga Harapan; PNAFN = Program me National d’Aide aux Familles Nécessiteuses; SAP = Social Amelioration Program; SI = Ingreso Solidario; TCT = Temporary Cash Transfer. See Annex 3 for details. C ove rage 15 To design and provide social assistance, cash transfers in particular, countries have invested in delivery systems that can also be used by other programs. Some of these delivery systems, such as beneficiary registries, were developed during the initial scale-up phases of cash transfer pilots, beginning with the collection of the necessary data. Over time, these countries expanded their registries beyond cash transfer programs itself (that is, larger than the program caseload) to be used for future program scale-ups or used by other programs by integrating them with other databases to expand its coverage and improve assessment and monitoring. For example, when the Colombian government created a new cash transfer program, Ingreso Solidario or Solidarity Income in response to COVID-19, it pre-set the number of eligible recipients at 3  million households who were already registered in the registry but were not benefiting from any of the three major already existing programs (Blofield, Lustig, and Trasberg 2021). Argentina and Brazil relied on self-targeting and demand-driven mechanisms to reach a large share of households not covered by existing social protection schemes, even if they were not included in the existing administrative registries. Tunisia, which has a long history of providing cash transfers since the 1980s, began reforming its social protection following the Arab Spring in 2011, which included developing a social registry and revamping its targeting approach. When the pandemic hit, it drew on the registry, which only covered the beneficiaries of existing programs. Therefore, in parallel, it set up an application platform to reach those in need who were not in the registry, particularly the self-employed. Jordan, another country from the MNA region, had established a national unified registry with broad coverage and has been modernizing and expanding Takaful, its main cash transfer program since 2018. These systems were the backbone of the emergency support that it provided during COVID-19 and subsequent crises including inflation, and it has now standardized the eligibility criteria for all cash transfer programs. Pakistan, which has a dynamic registry as a result of decades of investment, was able to deliver a transfer to 44 percent of the population in only 10 days during a catastrophic flood in 2022, which was much faster than its COVID-19 transfers, which took about two months to organize and deliver. Even with little to no infrastructure, some countries were able to set up or scale up emergency support successfully. Despite the lack of a social registry, some fragile African countries— namely the Democratic Republic of Congo, Togo—relatively quickly created welfare-targeted cash transfer programs from scratch, leveraging novel data sources, particularly call detail records and mobile-phone-based outreach strategy (Okamura, Ohlenburg, and Tesliuc 2024). Building upon its previous experience with epidemics, such as Ebola in 2015, Sierra Leone opted for more traditional approaches, using the lists of employees to identify beneficiaries for emergency transfers, while also leveraging its existing cash transfer programs (Ep Fet Po). Efficient payment systems are also key to facilitating the delivery of transfers to recipients. Many countries have been making gradual improvements in their financial and payment infrastructure which can ensure greater transparency, fiduciary control, financial inclusion, and efficiency of cash transfers. And when the pandemic hit, it accelerated the adoption of mobile- based registration and enrollment, and mobile wallets, facilitated by relaxed KYC (Know Your Customer) requirements (Marin and Palacios 2022). For example, during COVID-19 the Philippines expedited payments for 11 million beneficiaries of Social Amelioration Program 2 by bulk registering them into ‘restricted’ transaction accounts (Cho et al. 2021).14 14. The Social Amelioration Program, a new temporary unconditional cash transfer initiative, was executed in two rounds (SAP 1 and 2). It aimed to reach 11 million beneficiaries, including 1.4 million from the regular cash transfer program (4Ps). 16 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Social assistance also played a central role during the recent global inflation crisis in 2022 and 2023, for which many governments initially adopted subsidies to cushion its impacts. To tackle rapidly surging prices in 2022–2023, most countries responded by implementing price subsidies on food, fuel, and fertilizer as well as utilities (for example, electricity, water and heating services).15 While subsidies were the most common measures adopted initially, their share decreased from 79 percent to 36 percent of the total social protection response after five months (see Figure 3.8, from April 2022 to September 2022). This decrease was accompanied by the increased adoption of more targeted, more effective, and progressive measures—social assistance and tax measures together almost doubled from 17 percent to 36 percent over the same period.16 FIGURE 3.8  Composition of Social Protection Measures in Response to the Inflation Crisis by Instrument, 2022–2023 (n = 84–178 countries) Gradual shift responses from subsidies to social assistance. Composition of social protection measures 100 3 1 12 17 22 21 19 3 2 75 6 6 21 10 5 6 4 5 6 4 50 24 29 31 79 61 25 36 34 33 0 April 2022 July 2022 Sept 2022 Dec 2022 June 2023 (n = 84) (n = 121) (n = 159) (n = 170) (n = 178) Subsidies Social assistance Social insurance Labor market programs Trade related measures Tax measures Source: Adapted based on data from versions 1–5 of the World Bank’s Price Shock trackers (Gentilini et al. 2022b, c, d, e, 2023) Note: Based on a sample of 84 to 178 countries (n = number of countries). Annex Figure 1.5 shows the breakdown by region and income levels. 15. After a declining trend in the last couple of years, energy subsidies re-emerged rapidly at much higher levels than the pre-COVID-19 period (Gencer and Akcura 2022). 16. As the inflation prolonged over several months, the responses gradually shifted to social assistance measures (see Figure 3.8). For example, Egypt introduced a total of 17 social protection measures to the price shock, including five social assistance measures and six subsidy measures (four food subsidy, one fuel subsidy, and one agriculture subsidy). All the subsidy measures, except one, started at the onset of the inflation crisis. On contrary, all the social assistance responses came with a time lag of several months (that is, mostly in the third and fourth quarter of 2022), except for a horizontal expansion of their national flagship cash transfer program (Takaful and Karama) in March 2022. Senegal, which implemented eight measures (five were subsidies, two were social assistance, and one was through tax measures), followed a similar sequence, starting with subsidies and followed by social assistance (Gentilini et al. 2023). C ove rage 17 Furthermore, social assistance has been playing a vital role in supporting subsidy reforms more broadly. General subsidies are expensive – an estimated 8 percent of global GDP is being spent on fossil fuels, agriculture, and fisheries (Damania et al. 2023). They also tend to be regressive as richer households benefit more than poorer ones, besides other negative effects on the environment and market. As a part of the government’s efforts to reduce or eliminate such harmful subsidies, cash transfers have played an instrumental role in terms of mitigating the impacts of price increases, thereby achieving more progressive and sustainable fiscal expenditures (Mukherjee et al. 2023) (see Annex 4 for more information on how social assistance supported subsidy reforms). Household survey data are used to quantify the recent changes in social assistance coverage related to the COVID-19 pandemic and the inflation crisis. While the expansion of social assistance during the pandemic has been well documented, most of these analyses have been based on administrative data, which are more widely and frequently available than household survey data. To assess and quantify changes in social assistance coverage over time using welfare metrics, the analysis used data from 28 countries (out of our 68 sample countries) for which panel data was available for two periods: the pre-COVID-19 period (2015–2019) and for the years during or after the pandemic (2020–2022) (see Figure 3.9). This restricted sample disproportionally consists of Latin American and the Caribbean countries because household surveys are fielded more frequently than average in these countries. Social assistance was scaled up during or after the pandemic in most countries in different ways and to different extents. All but two of the countries in our sample (Mexico and the Kyrgyz Republic) expanded their social assistance coverage of the bottom 20 percent of the population in a wide range from minimal to almost 70 percentage points. Viet Nam had the largest increase, more than tripling their social assistance coverage from 26 percent before COVID-19 to 93 percent during or after the pandemic. Other countries increased their coverage of their poorest quintile of the households by 20–25 percentage points, including Brazil, Burkina Faso, and Indonesia. Since living standard measurement household surveys are typically designed to capture permanent programs rather than temporary emergency responses, these changes may be an indication that countries expanded their mainstream social assistance programs between 2020 and 2022, rather than attributing them to temporary scale-up. 18 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE 3.9  Coverage of Social Assistance as a Share of the Poorest Quintile Before and During/After COVID-19, 2015–2022 (n = 28 countries) Most countries expanded their social assistance during or after the COVID-19 at different magnitudes. Burkina Faso 55 76 LICs Niger 30 46 Bolivia 90 100 Vietnam 26 93 Bangladesh 57 71 LMICs Senegal 40 43 22 Kyrgyz Republic 21 Tunisia 19 19 Bhutan 3 5 Georgia 97 98 Thailand 94 97 Peru 90 97 Mongolia 96 97 Costa Rica 87 95 South Africa 94 95 Paraguay 91 95 Brazil 67 92 UMICs El Salvador 90 91 Argentina 68 82 52 Indonesia 75 Armenia 55 63 Dominican Republic 62 63 Mexico 68 62 Kazakhstan 49 59 Ukraine 55 55 Colombia 43 54 Panama 91 95 HICs 75 Romania 76 0 20 40 60 80 100 Social assistance coverage (% of population in the poorest quintile [i.e., Q1]) Pre-COVID-19 (2015–2022) During/Post COVID-19 (2020–2022) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 28 countries (n = no. of countries), which is the subset of 68 countries for which comparative data exist for both before and during/after COVID-19. Pre-COVID-19 coverage is based on the data from the most recent household survey for the period 2015–2019, while coverage during/after COVID-19 is based on the household surveys with highest coverage published during the period 2020–2022. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries. C ove rage 19 BOX 2  How Social Assistance Promotes Gender Equity Social assistance programs deliberately foster gender equality in their design. Recognizing the gender gap in accessing social services, economic opportunities, and intra-household decision-making power, most social assistance programs are intentionally designed to address the gaps and enhance women’s and girls’ capability to meet their basic needs, and access essential services and financial resources. For instance, some use gender-related criteria such as pregnant women and mothers or female-headed households to target their recipients. Some prioritize adult female members of beneficiary households as the recipients of cash transfer payments. The evidence shows that carefully designed social assistance programs can enhance women’s empowerment and gender equality outcomes, including access to finance, giving them more agency at the household and community levels and increasing their labor force participation (Peterman et al. 2024; Gavrilovic et al., 2022). Cross-country analysis using household and administrative information from the ASPIRE database confirms that social assistance programs are gender-responsive. On average, the proportion of female recipients of social assistance programs is 58 percent according to household survey data (see below table) and 68 percent17 according to administrative data. In the household survey data, the proportion of women recipients in LICs and LMICs tends to be around 50 percent, whereas the share jumps to around 60 percent for UMICs and HICs. In regional terms, South Asia, Sub-Saharan Africa, and the Middle East and North Africa hover just above 50 percent, while the shares in Latin America and the Caribbean, Europe and Central Asia, and East Asia and the Pacific range from 57 to 65 percent. Proportion of Female Recipients of Social Assistance Programs (N = 214 programs from n = 43 countries) Almost three-fifths of social assistance recipients are female. Income level % Region % LICs 49.9 EAP 61.7 LICS 52.5 ECA 57.1 UMICs 59.5 LAC 59.8 HICs 60.7 MNA 64.5 Total 57.6 SAR 52.2 SSA 51.2 Total 57.6 Source: Original table for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on sample of 214 programs (N = no. of programs) in 43 countries (n = no. of countries), which is a subset of 68 countries, for which information by gender was available. On average, 63 percent of the social assistance programs have proportions of female recipients ranging from 40 to 60 percent. About one-third of programs have a female share over 60 percent, with 12 percent even exceeding 80 percent. Only 7 percent of the programs have a female recipient share below 40 percent of total beneficiaries. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. (continued) 17. Based on a sample of 39 active social assistance programs (UCT, CCT, Public Works, and social pension) from administrative data in 39 countries containing coverage/expenditure data from 2015 onwards. 20 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond BOX 2  How Social Assistance Promotes Gender Equity (continued) The type of program also influences the gender balance. CCT programs average 74.8 percent of beneficiaries who are female, followed by UCTs (62.9 percent). In nearly half of the CCTs, more than 80 percent of beneficiaries are female. The female share is around 55 percent for public works and fee waivers, depending on the type. Typical labor-intensive public works tend to have a larger share of male recipients due to the need to lift heavy weights. The gender distribution for social pensions, and food and in-kind transfers is 53.3 percent and 52.6 percent, respectively, depending on each program’s objective and target groups. The female share is below 50 percent for school meals (48.7 percent) as the gender distribution depends on school enrollment. Share of Female Recipients, by Instrument Type (N = 214 programs from n = 68 countries) CCTs have the highest share of female recipient. Share of female recipient 100 5 5 7 12 90 18 16 11 26 25 80 48 19 50 70 27 60 50 100 13 68 75 40 68 61 63 30 48 50 20 39 10 7 8 4 9 6 1 0 3 4 5 CCT UCT Social Public Food and School Fee waivers Other Social (N = 23) (N = 44) pensions works in-kind meals and targeted social assistance (N = 38) (N = 2) (N = 28) (N = 16) subsidies assistance (N = 214) (N = 19) (N = 44) <20 20–40% 40–60% 60–80% 80–100% Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on sample of 214 programs (N = no. of programs) across 43 countries, which is a subset of 68 countries. CCT = Conditional Cash Transfer, UCT = Unconditional Cash Transfer. 4 Spending Globally, countries spend an average of almost 5.3 percent of their GDP on social protection, of which 1.5 percent is spent on social assistance, 3.7 percent on social insurance, and only 0.2 percent on labor market programs (Figure 4.1a). Social assistance spending has remained essentially unchanged at this level since 2012, even though it increased temporarily in response to different crises. This means that there is a huge financing gap compared to need, particularly in LICs and MICs. Looking at financing together with coverage, social assistance reaches many more people (for much less expenditure) than social insurance. Social assistance coverage is double of that of social insurance (42 percent versus 19 percent of total population) yet contributory social insurance expenditure accounts for almost two and half times the social assistance spending. To a large extent, this reflects the evolution of social protection expenditures as countries grow richer (Banerjee et al. 2024). This section presents an analysis of social assistance expenditure from 2017 to 2022. The sample for this analysis consists of the 76 countries for which comprehensive expenditure information was available on their social assistance programs in the ASPIRE administrative database. The sample contains 1,244 social assistance programs, which translates to an average of 16 programs per country. As can be seen in Table 4.1, the most popular instrument is the UCT, which is implemented in nearly 90 percent of the countries for various objectives and target groups, followed by social pension, and food and in-kind transfers are each used in nearly 65 percent of the countries. Public works and fee waivers are each used by around 60 percent of the sample countries. Number of Social Assistance Programs included in the ASPIRE administrative TABLE 4.1   Database (N = 1,244 programs from n = 76 countries) A diverse array of social assistance instruments totaling over 1,200 programs from 76 countries. Social assistance category Number of programs Number of countries Unconditional cash transfers 432 (35%) 69 (91%) Conditional cash transfers 47 (4%) 29 (38%) Social pensions 162 (13%) 49 (64%) (continued) 21 22 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Number of Social Assistance Programs included in the ASPIRE administrative TABLE 4.1   Database (N = 1,244 programs from n = 76 countries) (continued) Social assistance category Number of programs Number of countries Public works 55 (4%) 43 (57%) Food and in-kind 116 (9%) 49 (64%) School feeding 39 (3%) 34 (45%) Fee waivers 108 (9%) 45 (59%) Others social assistance 90 (7%) 37 (49%) Social care 117 (9%) 37 (49%) Non-Contributory Health Insurance 78 (6%) 37 (49%) Total 1,244 76 Source: Original table for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Note: Based on a sample of 76 countries (n = no. of countries). UCT includes Poverty targeted programs, Family, child or orphan allowance; Funeral grants, burial allowance; Emergency support (including support for refugees and migrants); Public charity (including Zakat). Social pensions are for the elderly, survivors, persons with disabilities, veterans. Public works are also often called cash-for-work, food-for-work (including food-for- training), and food-for-assets. Food and in-kind transfers also include food stamps, rations, and vouchers; nutrition programs; school supplies; in-kind or non-food emergency support; and other in-kind transfers. Fee waivers are for health and education, food subsidies, housing, utilities, electricity, agriculture inputs, and transportation benefits. Others consists of scholarships, education, benefit transfers for caregivers, and tax exemptions. Relative Spending on Social Assistance Has Remained 4.1  Virtually Unchanged Global social assistance spending has plateaued at around 1.5 percent of GDP since 2010.18 Using the 2022 expenditure data from 72 countries, Figure 4.1a shows that the average spending on social assistance has remained at around 1.5 percent of GDP, which amounted to 28 percent of total social protection expenditures in 2022. The level of social assistance spending varies significantly across countries, by both region and income level. Countries in Latin America and the Caribbean, and Europe and Central Asia allocate similar levels of GDP (1.9–2.0 percent), although social assistance spending as a share of social protection budget is much lower in ECA (20 percent) than in LAC (31 percent). This reflects the greater role played by social insurance in ECA. Sub-Saharan Africa, South Asia, and East Asia and the Pacific spend less than the global average at 0.8 percent, 1.2 percent, and 1.3 percent of GDP, respectively. By income level, HICs and UMICs spend 1.7–2.0 percent of their GDP on social assistance, while LICs and LMICs spend only around 0.8–1.0 percent. The average per capita global spending on social assistance is US$244 per year, with high heterogeneity across income level. Figure 4.1.b presents per capita spending defined as annual total spending divided by the total population of each country (not the total number of beneficiaries) expressed in US$ 2017 purchasing power parity (PPP). HICs invest over 30 times more than LICs (US$495 versus US$16) per capita, while LMICs spend US$66 and UMICs spend US$331. 18. Based on the World Bank’s second edition of the State of Safety Net report in 2015 (Honorati, Gentilini, and Yemtsov 2015), with the caveat that the sample countries in the current report and the 2015 report are different. S p en di ng 23 Social Assistance Spending as a Share of GDP and in Absolute Per Capita Amount in 2017 US dollar PPP, FIGURE 4.1   2022 (n = 72 countries) Social assistance investment has been stagnant at 1.5 percent of the GDP over the past decade. a. Social assistance spending as a share of GDP Average spending (% of GDP) 9 8.4 8.3 8 7 6.6 6.1 6.4 6 5.3 5 4.2 4 3.7 3 2.5 1.9 2.0 2.0 2 1.5 1.6 1.7 1.6 1.7 1.3 1.2 0.8 0.8 1.0 1 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 72) (n = 8) (n = 23) (n = 10) (n = 8) (n = 5) (n = 18) (n = 9) (n = 21) (n = 29) (n = 13) b. Social assistance per capita spending in US$ 2017 PPP Per capita spending ($ 2017 PPP) 3,000 2,608 2,500 2,000 1,938 1,500 1,133 1,109 1,003 1,000 869 710 495 500 356 392 244 185 239 216 281 331 112 84 146 16 30 66 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 72) (n = 8) (n = 23) (n = 10) (n = 8) (n = 5) (n = 18) (n = 9) (n = 21) (n = 29) (n = 13) Social assistance Social protection Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Notes: Based on a sample of 72 countries (n = no. of countries), which is a subset of 76 countries. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Spending Is Countercyclical 4.2  In 2020 and 2021, countries scaled up their social assistance spending countercyclically to support their populations during the global pandemic and related economic shocks. The global average of social assistance spending increased by 50 percent, from 1.4 percent before COVID- 19 (2019 or before) to 2.2 percent during the pandemic either in 2020 or 2021 (see Figure 4.2). In 2022, it fell back to 1.5 percent, which was only just slightly higher than the pre-pandemic level. This inverted U-shaped trend is consistent across regions and country income groups. 24 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Among the sample of 76 countries, those in the East Asia and the Pacific, and Latin America and the Caribbean regions more than doubled their social assistance budgets at the peak of the pandemic compared to before COVID-19, whereas the increases were more moderate (ranging from 33 to 55 percent) in other regions. Wealthier countries (HICs and UMICs) had higher levels of social assistance spending before the pandemic, and their marginal increase during the COVID-19 period was also larger than developing countries. HICs and UMICs increased their social assistance budgets by more than 1 percentage point of their GDP, while the rise in LICs and LMICs was limited to 0.4 to 0.6 percentage points of their GDP. Even though the change in absolute terms was small, LICs increased their budget allocation to social assistance by 75 percent, marking the highest jump among all country income groups. While there are many different complex determinants of a country’s social assistance budget envelope (including political will and socioeconomic factors), there was a stronger correlation between the social assistance budget and the country’s income level during the pandemic years compared to before the pandemic during normal times (see Annex Figure A1.6 and A1.7). Social Assistance Spending Before, During, and After COVID-19, 2017–2022 (n = 72–76 countries) FIGURE 4.2   Social assistance scaled up and reversed back, serving as stabilizers during the pandemic. Average social assistance spending (% of GDP) 4 3.8 2.9 3 2.7 2.4 2.2 2.2 2.2 2.0 2.0 2 1.8 1.9 1.7 1.6 1.7 1.7 1.7 1.5 1.6 1.6 1.4 1.3 1.4 1.3 1.1 1.2 1.2 0.8 0.8 0.8 0.9 1.0 1 0.8 0.8 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC Pre COVID-19 (2017-2019) During COVID-19 (2020-2021) Post COVID-19 (2022) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Notes: Based on a sample of 76, 76, and 72 countries for the periods 2017–2019 (Pre-COVID), 2020–2021 (COVID-19), and 2022 (Post COVID), respectively (n = no. of countries). Overall trend remained the same when the results were replicated using the same sample of 72 countries for all three periods. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Countries temporarily scaled up their social assistance provision both by adapting existing programs and by establishing new or exceptional programs equally. As the pandemic began in 2020, existing programs in many countries not only continued to support their existing beneficiaries but were also able provide them additional benefits and support new beneficiaries whether by vertical expansion, horizontal expansion, administrative adaptation, or some combination of these. In parallel, many countries delivered temporary support by creating new programs. Out of the increase in social assistance expenditure in 2020 (0.6 percentage points), more than half resulted from the expansion or adaptation of existing programs, with the remaining half being spent on new temporary measures. These temporary or exceptional measures had almost vanished by 2022 (0.04 percent of GDP) as the world emerged from the pandemic (see figure 4.3). S p en di ng 25 Social Assistance Spending on Existing and Temporary Programs by Year, FIGURE 4.3   2017–2022 (n = 72–76 countries) Countries leverage existing platforms for temporary expansion of social assistance. Average social assistance spending (% of GDP) 3 1.98 2 1.75 0.37 0.28 1.50 1.34 1.31 1.34 0.04 1 1.61 1.46 1.46 1.34 1.31 1.34 0 2017 (n = 74) 2018 (n = 76) 2019 (n = 76) 2020 (n = 76) 2021 (n = 72) 2022 (n = 72) Existing program Exceptional meaures during COVID-19 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data and the World Bank’s COVID-19 trackers (Gentilini et al. 2022a) Notes: Based on sample of 74, 76, 76, 76, 72, and 72 countries for 2017, 2018, 2019, 2020, 2021, and 2022, respectively (n = no. of countries). Furthermore, the level and scale of the budgets that countries allocated to social assistance during the COVID-19 pandemic varied widely. Figure 4.4 presents countries’ annual social assistance spending between 2017 and 2022. The elevated level of social assistance expenditure during the pandemic is clear and consistent in most countries.19 Out of a total of 76 countries, 16 countries20 more than doubled their social assistance spending during COVID-19 compared to 2019, although most of these countries had been spending less than the global average on social assistance in 2019 (1.34 percent as seen in Figure 4.3). For example, Chile and Mongolia more than tripled their social spending during the pandemic. Chile spent the highest share of GDP on social assistance (11 percent) in 2021 to finance two new emergency unconditional cash transfers. Mongolia scaled up its social assistance spending from 1.8 percent in 2019 to 6.8 percent in 2021, mostly attributed to the scale-up in the flagship Child Money program and a new one-off transfer program to citizens during lockdown (called One-time MNT300K).21 In addition, Liberia, which spent a negligible amount on social assistance before COVID-19 (0.04 percent of its GDP) allocated 0.7 percent during the pandemic. The majority of the financing for Liberia’s measures was provided by external partners and was used to implement a new in-kind transfer to the vulnerable population, in combination to more than tripling their national cash transfer program. 19. This trend is evident in 69 out of 76 countries. The remaining countries either reduced their social assistance spending or maintained it at the same level. For example, Hungary, an HIC, maintained its spending level during the pandemic compared to the level in 2019. This was also the case in the Central African Republic and Zimbabwe, which are an LIC and an LMIC, respectively. 20. Azerbaijan, Brazil, Chile, the Dominican Republic, Ethiopia, Indonesia, Liberia, Malaysia, Mongolia, Morocco, Nigeria, Pakistan, Panama, Peru, the Philippines, and Tonga spent above the global average in 2019. 21. The expenditure on the flagship Child Money program increased from 0.6 percent in 2019 to 3 percent of the GDP during the pandemic, plus expenditure on the new one-off program accounted for 2.3 percent of the GDP. 26 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE 4.4  Country-level Social Assistance Spending as a Share of GDP by Year, 2017–2022 (n = 72–76 countries) Spending levels during COVID-19 increased but there are significant variations across countries. a. All countries (n = 72–76 countries) SA expedniture as % of GDP 12 10 8 6 4 1.98 1.75 1.31 1.34 1.50 2 1.34 0 2017 2018 2019 2020 2021 2022 (n = 74) (n = 76) (n = 76) (n = 76) (n = 72) (n = 72) b. Countries spending less than 3 percent of GDP (n = 59–63 countries) SA expedniture as % of GDP 3 1 1.98 1.75 1.50 1.34 1.31 1.34 2 0 2017 2018 2019 2020 2021 2022 (n = 61) (n = 63) (n = 63) (n = 63) (n = 59) (n = 59) LICs LMICs UMICs HICs Total Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data and the World Bank’s COVID-19 trackers (Gentilini et al. 2022a). Notes: Based on samples of 74, 76, 76, 76, 72, and 72 countries for 2017, 2018, 2019, 2020, 2021, and 2022, respectively (n = no. of countries). LICs = Low Income Countries, LMICs = Lower Middle Income Countries, UMICs = Upper Middle Income Countries, HICs = High Income Countries. S p en di ng 27 Following the scaling-up of spending during the crisis, most countries seem to have returned to their pre-COVID-19 investment levels. Around three-quarters of the countries that scaled up their social assistance spending during the pandemic (2020 and 2021) have scaled back their spending levels closer to their pre-COVID-19 levels (including Pakistan, Indonesia, Tajikistan, and Tunisia). However, higher spending continued after the pandemic in around one-quarter of the countries (including Iraq, Jordan, Mauritania, Mexico, Moldova, Montenegro, Rwanda, Tonga, and Uzbekistan22), albeit with a smaller increase between 2021 and 2022. 4.3  Social Assistance Spending is Dominated by Cash-based Interventions Cash transfers receive the most resources, followed by social pensions, with some emerging insights. Figure 4.5 shows how countries allocate their social assistance budgets to different instruments. Globally, around 70 percent of total social assistance spending is disbursed through the three top cash-based instruments (notably UCT, CCT, and social pensions), with nearly half of the budget is disbursed through cash transfers while recently, there has been a shift away from CCTs to UCTs. The proportion of UCTs has jumped in recent years, while that of CCTs has decreased. This reflects the need for more temporary income support for broader segments of the population in recent years, particularly because of COVID-19 and inflation. For instance, in Europe and Central Asia, cash transfers are now almost all unconditional. This is a noteworthy change since the last edition of the World Bank’s State of Safety Nets report in 2018 (World Bank 2018), which showed that ECA spent 36 percent of their social assistance expenditure on UCTs and 10 percent on CCTs, with the caveat that the 2018 report and this paper used different countries in their analyses. Social pensions continue to be vital not only for the low-income elderly population but also for families with special needs including those with disabilities (see Box 3). This is reflected in their one-quarter share of global social assistance spending. Social pensions are the second largest social assistance budget item in countries of all income levels and regions except in LICs, in the Middle East and North Africa, and in Sub-Saharan Africa where spending on food aid and in-kind transfers exceed spending on cash transfers. Furthermore, public works continue to play an important role in supporting households, particularly in South Asia (19 percent) and also in Sub-Saharan Africa and Europe and Central Asia (7–9 percent). Historically, public works have been allocated much lower spending from the social assistance budget (0–9 percent) and have been more prevalent in LICs and in the SSA region (18 percent and 10 percent, respectively). Last but not least, social care services23 have been on the rise over the past years in middle- and high-income countries. While its share is still small (3 percent globally), this is a growing trend particularly in the Middle East and North Africa, Europe and Central Asia, and Latin America and the Caribbean. 22. Given that the analysis is based on the data till 2022, the trend of higher spending post-pandemic does not necessarily hold, including for these example countries, as of writing. 23. Social care services can consist of a variety of social and care services (non-cash interventions) to support individuals through the life cycle including vulnerable families and families with special needs (for example, illness, disabilities). 28 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE 4.5  Composition of Social Assistance Spending by Instrument, 2022 (n = 72 countries) More than 70% of social assistance spending are disbursed through cash-based programs. Total (n = 72) 43 6 25 3 2 5 4 2 3 6 EAP (n = 8) 41 10 22 3 1 4 4 10 13 ECA (n = 23) 56 1 24 6 11 5 2 3 2 LAC (n = 10) 27 21 16 3 4 3 10 3 3 9 MNA (n = 8) 48 6 10 10 17 30 6 7 SAR (n = 5) 34 7 44 4 2 3 3 20 1 SSA (n = 18) 31 1 6 10 8 19 4 11 2 8 LIC (n = 9) 25 3 3 18 9 26 0 7 0 8 LMIC (n = 21) 43 5 12 2 4 11 5 8 3 8 UMIC (n = 29) 36 9 28 3 2 3 6 2 3 8 HIC (n = 13) 63 0 15 9 11 7 0 40 0 10 20 30 40 50 60 70 80 90 100 Compostion of expenditure on social assistance, by instrument UCT CCT Social pensions Public works School feeding Food and in-kind Fee waivers Other SA Social Care NC health Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Notes: Based on a sample of 72 countries (n = no. of countries), which is a subset of 76 countries with spending data. CCT = Conditional Cash Transfer, UCT = Unconditional Cash Transfer, SA = Social Assistance, NC Health = Non-Contributory Health Insurance. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. S p en di ng 29 Share of Social Assistance Spending Allocated to Each Instrument, All Countries FIGURE 4.6   versus Only Those Countries that Use the Instrument (n = 72 countries) Countries implementing CCT, in-kind transfers, and health fee waivers spend 13-15 percent of their social assistance spending on each instrument. Share of expenditure 50 46 43 40 30 28 25 20 15 14 13 10 9 10 7 8 6 5 5 6 3 4 3 2 2 0 UCT CCT Social Public In-kind School Fee Other Social NC pension works meals waivers SA care health All countries Only those countries that use the instrument Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Notes: Based on a sample of 72 countries (n = no. of countries), which is a subset of 76 countries with spending data. CCT = Conditional Cash Transfer, UCT = Unconditional Cash Transfer, SA = Social Assistance, NC Health = Non-Contributory Health Insurance. Who Benefits from Social Pensions? BOX 3   Types of Social Pension (N = 162 programs from n = 49 countries) Social pensions support a range of policy Social pensions have various objectives or target groups. objectives and target populations. The figure on the right shows their main target 9 groups based on 162 social pension programs recorded in the ASPIRE administrative 32 database. There are three main target groups, which together account for at least 85 percent 24 of recipients. These groups are (a) people with disabilities (32 percent), (b)  the elderly (27 percent), and (c) war veterans and survivors (24 percent). While some programs specifically 6 target caregivers, they are mainly those who 2 are taking care of elderly family member(s) or 27 people (including children) with disabilities. Disabled Elderly Elderly and disabled Caregiver Civilian war invalids/victims; (elderly, disabled, etc.) Martry, war, survivors, veterans Miscellaneous Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Notes: Based on a sample of 162 programs in 49 countries, which is a subset of 76 countries. 30 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond There Are Stark Contrasts in Financing Outlook Affected by 4.4  Country Income and (Un)stability Levels This subsection examines a relatively unexplored topic on how social assistance programs are financed. The analysis is based on a sample 627 programs24 in 65 countries for which data are available on their sources of financing, combining two sources. One is the ASPIRE database and the other is information shared from four UN agencies as part of the Tesliuc et al. (forthcoming) to obtain a better picture of humanitarian assistance programs in FCV countries. Out of these 65 countries, 59 countries are part of 76 countries of our ASPIRE sample, while six countries were additionally introduced.25 In the interest of simplicity, the financing sources are classified in three broad categories: (a) grants (labeled as ‘external’); (b) others (labeled as ‘domestic’), which consist of general revenue, earmarked funding, and concessional loans;26 and (c) a combination of the two. The analysis presented in this subsection is adapted from an upcoming publication (Tesliuc et al., forthcoming) which provides additional information. Most countries use domestic resources to fund at least two-thirds—potentially closer to four- fifths—of their total social assistance budgets, indicating concrete national commitment (Figure 4.7). In other words, grant financing from international donor agencies provides only 21  percent of the social assistance budget, while the remaining 12 percent consists of a combination of domestic and external sources. Income level is a strong determinant of how countries fund their social assistance programs. HICs fully finance their social assistance programs on their own without any external financing defined as grant appropriation. The share of domestic resource used is also notably high for MICs— at least 85–86 percent in non-FCV countries and 59–66 percent in FCV settings.27 However, this percentage decreases sharply in LICs (33 percent in non-FCV countries and only 7 percent in FCV countries). LICs rely heavily on grants, for example, from donor organizations or NGOs, as their main source of financing for their social assistance programs. As an LIC becomes more economically developed, its reliance on these funding sources falls or disappears, except for the 24. Including 57 programs which were funded by four humanitarian agencies (the Food and Agriculture Organization [FAO], United Nations High Commissioner for Refugees [UHCR], United Nations Children’s Fund [UNICEF], and the World Food Programme [WFP]) in 16 FCV countries. Exclusively for this analysis, the data on these 57 programs were appended to ASPIRE’s administrative database to provide a more holistic picture on financing. None of these 57 programs exist in the ASPIRE database. Of these 16 FCV countries, 10 countries are part of our ASPIRE administrative sample of 76 countries, and 6 countries were newly introduced. 25. Mali, Mozambique, Myanmar, Somalia, Syria, and the Republic of Yemen. 26. Tesliuc et al. (forthcoming) presents more detailed results and discussions, including the breakdown of domestic financing sources by general revenue, earmarked funding, and concessional loans. ‘Earmarking’ here means setting aside a specified amount to finance social assistance programs. Concessional loans that offer favorable terms (by rate and duration) are provided by various entities, including government agencies, international financial institutions, development banks, private sector investors, nongovernmental organizations (NGOs), and donors. We classified concessional loans as domestic resources, given that governments often integrate the loan and the repayment requirement into their public financing, national budget processes, and fiscal policy. 27. Under the World Bank’s 2023 classification, 17 countries experienced conflict. However, South Sudan was excluded due to the unavailability of GDP data. S p en di ng 31 significantly larger provision of grants to UMICs compared to LMICs in FCV settings. By region, Sub-Saharan African countries rely heavily on grants for at least half of their social assistance spending, while countries in the Middle East and North Africa and in South Asia finance at least 32 percent and 23 percent, respectively, of their social assistance spending from grants. LICs and MICs affected by fragility, conflict, and violence receive a high proportion of grants from the international community to complement their limited domestic capacity to fund social assistance. Estimates show that 59 percent of the extreme poor around the world will be concentrated in FCV-affected countries by 2030 (World Bank 2024). Currently, FCV countries are receiving a much larger influx of grants (money without repayment obligations) to fund their social assistance—and humanitarian assistance—programs than non-FCV countries in the same country income group. For example, in LICs, full grants account for 77 percent of total social assistance spending in FCV countries compared with 10 percent in non-FCV countries. FIGURE 4.7  Composition of Social Assistance Spending by Financing Source, 2017–2022 (N = 627 programs from n = 65 countries) Income level and FCV are two key determinants of social assistance financing patterns. Total (N = 627) 67 21 12 EAP (N = 30) 100 0.1 ECA (N = 59) 93 7 0.2 LAC (N = 267) 89 0.1 11 MNA (N = 60) 58 32 10 SAR (N = 39) 55 23 21 SSA (N = 172) 33 51 16 LIC-FCV (N = 99) 7 77 15 LIC-non-FCV (N = 11) 31 10 59 LMIC-FCV (N = 45) 66 25 9 LMIC-non-FCV (N = 113) 86 2 12 UMIC-FCV (N = 31) 59 29 12 UMIC-non-FCV (N = 239) 85 0.1 15 HIC-non-FCV (N = 89) 100 0 10 20 30 40 50 60 70 80 90 100 Composition of social assitance expenditure, by financing type Domestic External Combination of domestic and external Source: Adapted from Tesliuc et al. (forthcoming) using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data and additional data from UN agencies. Note: Based on a sample of 627 programs in 65 countries. Of this sample, 59 countries were part of our ASPIRE sample, while an additional six countries were added exclusively for this analysis. See footnote 24 for more info. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. FCV = Fragile, Conflict and Violence. 32 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Who Are the Implementers of Social Assistance Programs? BOX 4   Government agencies are responsible for implementing most social assistance programs globally, while UN agencies also play a key role as implementors, especially in FCV-affected countries. The table below shows the distribution of 731 programs based on their implementing agency, as the sustainability of social assistance programs can differ, depending on the implementing arrangement and the level of government leadership and/or of the integration of these programs into the national systems. As is the case with funding agencies, in non-FCV setting, governments implement all the programs, either independently (as in HICs) or in collaboration with other agencies outside the government (in the case of MICs and LICs). In FCV settings, UN agencies not only provide funding but also support as implementers with a strong presence in the field. They directly deliver close to half of the 731 programs, mostly in low-income FCV countries, but they also implement around 25–30 percent of programs in middle-income FCV countries as well. In FCV countries, governments implement more programs than they fund, which means that programs funded by external resources are implemented by either external partners or the government. Notably, in FCV countries, governments or authorities operate at least nearly half to three-quarters of social assistance programs. Composition of Social Assistance Program by Implementing Agencies (N = 731 programs from n = 65 countries) Governments play a central role in implementing social assistance with support from partners. LICs LMICs UMICs HICs Total FCV Non-FCV FCV Non-FCV FCV Non-FCV Non-FCV (N = 731) Implementation agency (N = 115) (N = 22) (N = 53) (N = 134) (N = 35) (N = 274) (N = 98) Government 46% 91% 75% 93% 54% 97% 100% 85% Others (for example, UN 43% 23% 1% 29% 0% 10% agencies, NGOs, private sector, foundations) Combination 10% 9% 2% 5% 17% 3% 5% Grand Total 100% 100% 100% 100% 100% 100% 100% 100% Source: Original table for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Note: Based on a sample of 731 programs (n = number of programs) in 65 countries. Of this sample, 59 countries were a part of our ASPIRE sample, while an additional six countries were added exclusively for this analysis. See footnote 24 for more information. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries. UN = United Nations, NGO = Non-Governmental Organization. 5 Adequacy As discussed earlier, social assistance covers twice as many people as social insurance with half the resources. This means that the per-beneficiary amount provided by non-contributory social assistance programs are much smaller than that of contributory social insurance benefits. Globally, the adequacy of social protection programs averages to 27 percent of the total welfare of its beneficiaries, with social assistance programs only providing just over one-fourth of the average adequacy of social insurance programs (11 percent vs. 36 percent)(World Bank 2025). This section assesses the adequacy of social assistance programs using household survey data from 54 countries (out of a sample of 68 countries). The remaining 14 countries either do not have any data on the monetary value of the benefits provided by their social assistance programs or do not have a sufficient number of beneficiary households. BOX 5  How to Set the Benefit Amounts of Social Assistance Programs? In principle, social assistance programs set their benefit amounts high enough to incentivize the target population to participate but not too high to cause any unintended consequences, including disincentives to work. There is no scientific cut-off point or shared consensus to define ‘appropriate’ or ‘adequate’ levels of social assistance benefits, and social assistance transfers are not always sufficient to lift all poor beneficiaries out of poverty. In reality, the benefit level is also influenced by program objectives and availability of budget. As a guide for ‘large’ transfers, some studies suggest that around 20–30 percent of the beneficiaries’ income or consumption and a much higher proxy are used for ‘big push’ transfers in the context of economic inclusion interventions. So, the bottom line is that the cut-off choice is somewhat arbitrary. Furthermore, it is rare for social assistance programs to systematically revise benefit amounts over time, unlike social insurance which are more likely to have an automatic mechanism to adjust the benefit amount based on the price index, wages, or other macroeconomic indicators (Gentilini et al. 2024). The lack of such a mechanism weakens the effectiveness of social assistance benefits due to the erosion of their purchasing power over time. In the context of high inflation and currency depreciation, implementers of social assistance programs, whether governments or development partners, are increasingly exploring better mechanisms to adjust the level of the benefits for cash-based programs. 33 34 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond As in the case of coverage and financing, little progress has been made in increasing the adequacy of social assistance benefits. Benefit levels remain low in LICs, particularly in Sub- Saharan Africa and South Asia. Figure 5.1 shows the adequacy of social assistance transfers as captured in household surveys from 54 countries, with values being calculated as a share of beneficiaries’ daily post-transfer welfare.28 Social assistance adequacy is calculated for two reference groups: one for the total beneficiaries and another for beneficiaries belonging to the poorest 20 percent of the population (Q1). This means that the benefit of a given monetary amount can have a higher ‘adequacy’ for Q1 beneficiaries than for total beneficiaries, given that ‘adequacy’ is defined as a share of their total income or consumption. On average, globally, the adequacy of social assistance has not changed much since 2010.29 For the Q1 beneficiaries, social assistance benefits account for 18 percent (or one-fifth) of their total welfare measured by either income or consumption (Figure 5.1). However, this average masks Social Assistance Adequacy for Total Beneficiaries and for Beneficiaries in the Poorest Quintile, FIGURE 5.1   2015–2022 (n = 54 countries) Social assistance adequacy accounts for almost one-fifth (and one-tenth) of total welfare for the beneficiaries in the poorest quintile (and for total beneficiaries). Social assistance adequacy (daily benefit amount as a % of beneficiary’s daily consumption or income) 35 30 30 26 25 23 20 18 16 17 17 15 15 14 11 12 12 12 11 11 10 10 8 8 8 9 9 6 5 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 54) (n = 4) (n = 13) (n = 16) (n = 2) (n = 5) (n = 14) (n = 6) (n = 17) (n = 26) (n = 5) For total beneficiaries For beneficiaries in Q1 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 54 countries (n = number of countries). LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. SA = Social Assistance. Q1 = Poorest or First Quintile. 28. Adequacy is defined as the percentage of transfer amounts received by beneficiaries of a group relative to the total welfare of the beneficiaries of that group, with the quintile distribution being based on the post-transfer welfare aggregate. 29. Based on the World Bank’s State of Safety Net report in 2015 (Honorati et al. 2015), with a caveat that the sample countries in the current report and the 2015 report are different. Ade quacy 35 the wide variation that exists across countries, ranging from a minimum of 1.1 percent (in Côte d’Ivoire) to a maximum of 93.6 percent (in Georgia). The adequacy of benefits for all recipients (the total population of beneficiaries) is 11 percent (or around one-tenth) of their total welfare, and this too varies, though not as widely, from a minimum of 0.7 percent (in Colombia) to a maximum of 55.2 percent (in The Gambia). LICs and LMICs have particularly low adequacy, about half of the global average,30 with a large disparity to the rest, but the benefit level of social assistance increases with the income level of the country. In dollar terms, on average, the daily amount that beneficiary households receive from social assistance is about US$1.1 per day in 2017 PPP (Figure 5.2a). This daily amount translates into about US$33 per month. There are substantial variations across countries, ranging from US$0.10 (in Bangladesh, Cameroon, Colombia, Eswatini, Guinea, and Malawi) to US$4.80 (in Mauritius). Furthermore, the social assistance benefit amount received by beneficiaries in the poorest quintile is slightly higher at US$1.4 per day (Figure 5.2b) than the US$1.1 per day for all beneficiaries. This may indicate that certain social assistance programs might favor poorer households (for example, these households might be entitled to the benefits of more social assistance programs or the programs might offer higher benefits to poorer households based on their characteristics, such as household size). There is a sign, albeit small, that benefits have been increasing in adequacy over recent years, but it is still premature to determine whether this increase will be permanent. Figure 5.3 compares the adequacy of social assistance benefits before COVID-19 (2017–2019) and during or after COVID-19 (2020–2022) based on a restricted sample of 23 countries (with a high concentration of Latin American countries) for which household surveys exist for both periods. In the interest of assessing the extent of adjusting the benefits of social assistance, responding to a series of crises since 2020, the maximum adequacy figure from 2020-2022 is compared with the adequacy before 2020 (using the latest from 2015 to 2019). All but five of these countries increased the adequacy of their social assistance benefits by a varying degree, ranging from 2–30 percentage points. Five countries in particular—Colombia, Thailand, Costa Rica, Brazil, and Burkina Faso—more than doubled the adequacy of their social assistance, and six other countries (Argentina, Armenia, Bolivia, Panama, Viet Nam, and the Dominican Republic) increased adequacy by at least 50 percent. While these increases could be due to the temporary expansion of social assistance during the latter period of the global pandemic, they might also indicate that the level of social assistance benefits is on an upward trend, given that household surveys tend to capture more permanent and regular programs. Finally, during the same period, adequacy declined in five countries in the sample—Bangladesh, Mexico, Paraguay, Senegal, and Ukraine—ranging from small to significant decreases. Again, these results are based on a restricted sample to assess the maximum increase in social assistance benefits during 2020-2022, therefore additional and updated data is required to examine further the latest trends of social assistance adequacy.31 30. Among LICs, the adequacy was particularly high for one country, namely Gambia. Excluding Gambia, the adequacy for total social beneficiaries was 7 percent (instead of 15 percent) in LICs. 31. For example, some of the countries included in the analysis here have further adjusted the benefit level of their social assistance programs. For instance, Armenia has changed its adequacy from 34 percent in 2019 to 62 percent in 2020, but it reverted close to its pre-COVID-19 level (32  percent) in 2021. While Senegal was included as a country where the adequacy declined during 2020-2022, the country 36 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Social Assistance Benefit Amounts in US$ 2017 PPP for Total Beneficiaries and Beneficiaries in the FIGURE 5.2   Poorest Quintile, 2015–2022 (n = 54 countries) Average monthly transfer to beneficiaries is around US$33 (US$ 2017 PPP). a. For total beneficiaries, 2015–2022 (n = 54 countries) Daily social assistance benefit amount ($ 2017 ppp) for total beneficiaries 5 4.8 4.8 4.8 4.0 4.0 4 3.0 3 2.6 2.0 2.1 2 1.7 2.1 1.3 1.7 1.5 1.4 1 1.1 1.0 1.2 0.8 1.3 1.0 0.8 0.7 0.6 0.1 0.6 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 54) (n = 4) (n = 13) (n = 16) (n = 2) (n = 5) (n = 14) (n = 6) (n = 17) (n = 26) (n = 5) b. For beneficiaries in the poorest quintile, 2015–2022 (n = 54 countries) Daily social assistance benefit amount ($ 2017 ppp) for beneficiaries in the poorest quintile (Q1) 7 5.8 5.8 5.8 5.8 5.8 6 5 4.5 4 3.5 3 2.8 2.6 2.6 2.8 2.7 2 2.2 1.6 1.6 1.4 1.3 1.3 0.9 1.2 1 1.1 0.7 0.7 0.4 0.6 0.0 0.2 0.0 0.1 0.0 0.1 0.0 0.0 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 54) (n = 4) (n = 13) (n = 16) (n = 2) (n = 5) (n = 14) (n = 6) (n = 17) (n = 26) (n = 5) Maximum Average Minimum Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 54 countries (n = number of countries). LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. increased the benefit amount of their main safety net program by 40 percent (The quarterly transfer amount was raised from 25,000 CFA to 35,000 CFA) in 2023. Ade quacy 37 Social Assistance Adequacy for Beneficiaries in the Poorest Quintile in FIGURE 5.3   2015–2019 (before COVID-19) and 2020–2022 (during/after the Pandemic) (n = 23 countries) Recent years indicate early and potential signs of an increase in social assistance adequacy. LICs Burkina Faso 01 Kyrgyz Republic 24 26 Bolivia 11 19 Vietnam 11 17 LMICs Tunisia 8 14 Bangladesh 4 3 Senegal 4 2 Georgia 83 94 Armenia 34 62 Costa Rica 24 55 Brazil 26 52 Argentina 20 35 Kazakhstan 26 30 Mexico 30 28 UMICs Mongolia 19 23 Colombia 1 19 Peru 15 19 Dominican Republic 8 14 Paraguay 29 14 Thailand 3 7 Ukraine 7 6 Panama 25 36 HICs Romania 18 21 0 20 40 60 80 100 Social assistance adequacy (daily benefit amount as a % of beneficiary’s daily consumption or income) Pre-COVID-19 (2015–2019) During/Post COVID-19 (2020–2022) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 23 countries (n = no. of countries). The latest data were used for the pre-COVID-19 period (2015–2019) while the maximum value was used for the COVID-19 period (2020–2022). Bangladesh in SAR, Senegal in SSA, Paraguay and Mexico in LAC, and Ukraine in ECA had lower adequacy during or after COVID-19 (2020–2022) than pre-COVID-19 (2015–2019). LICs = Low Income Countries, LMICs = Lower Middle Income Countries, UMICs = Upper Middle Income Countries, HICs = High Income Countries. 38 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Different types of social assistance instruments have widely different adequacy levels, with social pensions offering the highest benefits. On average, social pension benefits are equivalent to 17 percent of the total welfare of all beneficiaries and over 30 percent of the welfare of beneficiaries from the poorest quintile. Social pensions cater to different groups, primarily (poor) elderly people who are not part of contributory pension system and people with disabilities (Box  3). However, this benefit amount is still less than half of the pension amount provided through the contributory social insurance schemes (Hartley and Abels 2025). UCTs and other social assistance transfers32 come in second and third, respectively, with their benefit levels being close to the average adequacy of social assistance. These are followed by CCTs and fee waivers, whose benefit levels are close to half of the average figure. Social Assistance Adequacy by Instrument, 2015–2022 (n = 54 countries) FIGURE 5.4   The benefit level and therefore its impact varies across program type. Adequacy (daily benefit amount as a % of beneficiaries’ daily consumption or income) 40 30 30 20 18 17 18 12 11 10 10 9 10 7 8 6 5 5 3 4 2 2 0 UCT CCT Social Public In-kind School Fee waiver Other social Social (n = 34) (n = 16) pension works (n = 16) meals (n = 12) assistance assistance (n = 26) (n = 2) (n = 5) (n = 35) (n = 54) For total beneficiaries For beneficiaries in Q1 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 54 countries (n = no. of countries) that may or may not implement all the instruments. Public works are based on a very small sample of two, so might not be representative. Q1 = Poorest or First Quintile. UCT = Unconditional Cash Transfer, CCT = Conditional Cash Transfer. 32. These mostly include scholarship programs. Ade quacy 39 Even for the same instrument, the adequacy can vary widely across countries and programs. Figure 5.5 presents adequacy data for 83 flagship programs in 37 countries. These consist of 26 social pensions, 33 UCTs, 20 CCTs, and 4 public works. To take social pensions as example, some programs offer generous transfers accounting for 60 to 80 percent of total consumption or income of Q1 beneficiaries (such as Georgia’s Retirement Pension, Lesotho’s old-age pensions, and Brazil’s BPC33), whereas benefit level is limited to below 10 percent of welfare for social pensions in the Philippines, Sri Lanka, Indonesia, and Bangladesh, for example. As for UCTs, those in the upper range provide around 35–45 percent (Serbia’s social allowance, Uzbekistan’s childcare allowance, and Georgia’s Assistance for IDPs) while others provide less than 10 percent.34 As highlighted earlier, CCTs and public works have a much lower adequacy level than social pensions and UCTs, and they vary across programs much less. Among our sample, two CCT programs in LAC, Ecuador’s Bono de Desarrollo Humano and Brazil’s Bolsa Familia, have the highest adequacy at around 23–24 percent, followed by Mexico’s Jovenes Escribiendo el Futuro, Philippines’ Pantawid Pamilyang Pilipino Program (4Ps), and Costa Rica’s Avancemos, which range from 14 to 17 percent). Even well-established and well-known programs have benefit levels as low as 1 percent of total welfare, including Colombia’s Más Familias en Acción and Jóvenes en Acción and Bangladesh’s student stipend for primary education. Similarly, the adequacy of the selected public work programs ranges between a lower limit of 2 percent of the poorest quintile (Malawi’s Public Works Programme) and an upper limit of 8 percent (Ethiopia’s Productive Safety Net Program). 33. BPC: Beneficio de prestacao Continuada. 34. Such as Nigeria’s National Cash Transfer Programme, Chile’s Aporte Familiar Permanente, Russia’s poverty-targeted cash transfer and child allowances, Gabon’s Family Allowance, Chile’s Asignación Familiar, Jamaica’s PATH, and Russia’s unconditional allowances and other social payments for people entitled to receive social support. 40 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE 5.5  Adequacy of Selected Social Assistance Programs for Total Beneficiaries and for Beneficiaries in the Poorest Quintile, 2015–2022 (N = 83 programs from n = 37 countries) a. Social pension Retirement pension (Georgia) 40 81 Old-age pension (Lesotho) 30 75 Beneficio de prestacao Continuada (BPC) (Brazil) 39 61 Old age pension (Basic Retirement Pension) (Mauritius) 27 46 Social pension (Poland) 28 45 Disability pension (Georgia) 23 45 Pensiones del Régimen no Contributivo (Costa Rica) 22 41 Renta Dignidad (Bolivia) 9 38 Aporte Previsional Solidario (APS) de Vejez (Chile) 26 36 Widows’ and children pension (Mauritius) 23 35 Elderly pay (Turkey) 23 34 Social pensions including disability, survivorship (Russia) 21 33 Retirement benefits (Maldives) 20 33 120 a los 65 (Panama) 17 32 Programa Pensión 65 (Peru) 13 28 Adulto mayor (Paraguay) 17 27 Pensión para el Bienestar de Adultos Mayores (Mexico) 11 27 Social assistance pension (Romania) 9 25 Social welfare: Disability pension (Mongolia) 19 24 Social welfare pension (Mongolia) 18 22 Special benefits for disable persons (Romania) 17 20 Senior citizen pension (Philippines) 4 9 Elderly payment (Sri Lanka) 5 8 Cash for elderly (Indonesia) 5 7 Cash for persons with disabilities (Indonesia) 4 6 Old age allowance (Bangladesh) 3 4 0 10 20 30 40 50 60 70 80 90 100 Adequacy of transfer (% of total income/consumption of beneficiaries) For total beneficiaries For beneficiaries in Q1 (continued) Ade quacy 41 Adequacy of Selected Social Assistance Programs for Total Beneficiaries and for Beneficiaries in the FIGURE 5.5   Poorest Quintile, 2015–2022 (N = 83 programs from n = 37 countries) (continued) b. Unconditional cash transfers 37 Social allowance (Serbia) 46 Childcare allowance (Uzbekistan) 17 39 14 Assistance for IDPs (Georgia) 36 18 Alimony fund (Poland) 29 19 Targeted social assistance (Kazakhstan) 29 Programa subsidio social basico (PSSB) (Mozambique) 10 28 Family allowance (Uzbekistan) 13 27 Monthly social benefits (Kyrgyz Republic) 18 25 Low-income families single monthly cash allowance 16 (Kyrgyz Republic) 25 State benefits to families with children (Kazakhstan) 16 24 Guarantee minimum income (Romania) 17 24 16 Ajutor Social (Moldova) 23 10 Direct support PSNP (w/o labour activites) (Ethiopia) 19 Universal child allowance (Romania) 6 17 Child welfare grant (Lesotho) 6 16 Social welfare for single parent with 3/more children 13 (Mongolia) 15 Child money (Mongolia) 7 15 Child allowance during parental leave (Poland) 10 14 Allowance for care of 0-3 y.o. (Moldova) 12 13 11 Bono base familiar (Chile) 13 8 Cash transfers from government (Malawi) 13 Children allowances (Serbia) 8 12 8 Family benefits (Poland) 12 Samurdhi (Sri Lanka) 7 12 Unconditional allowances and other social payments for 5 people entitled to receive social support (Russia) 7 5 Benazir income support program (BISP) (Pakistan) 6 Public assistance [Programme of Advancement Through 3 Health and Education, PATH] (Jamaica) 5 2 Asignación Familiar (Chile) 5 Family allowance (Gabon) 2 3 2 Child allowances (Russia) 3 Poverty targeted cash transfers (Russia) 2 2 1 Aporte Familiar Permanente (Chile) 2 National cash transfer programme (Beta Don Come) 1 (Nigeria) 2 0 10 20 30 40 50 60 70 80 90 100 Adequacy of transfer (% total income/consumption of beneficiaries) For total beneficiaries For beneficiaries in Q1 (continued) 42 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE 5.5  Adequacy of Selected Social Assistance Programs for Total Beneficiaries and for Beneficiaries in the Poorest Quintile, 2015–2022 (N = 83 programs from n = 37 countries) (continued) c. Conditional cash transfers Bono de Desarrollo Humano (Ecuador) 14 24 Bolsa Familia (Brazil) 12 23 Jóvenes Escribiendo el Futuro (Beca para el Bienestar 7 Benito Juárez de Educación Superior) (Mexico) 18 Pantawid Pamilyang Pilipino Program (4Ps) (Philippines) 9 14 Avancemos (Costa Rica) 8 14 Beca Universal de Educación Media Superior Benito 5 Juárez (Mexico) 11 Asignaciones Familiares (Uruguay) 6 11 Red de Oportunidades (Panama) 8 11 Beca de Educación Básica para el Bienesta Benito 5 Juárez (Mexico) 11 Juntos (Peru) 6 10 Bono Niño (Peru) 4 10 Tekopora (Paraguay) 6 9 Maternity Allowance Programme for the Poor 4 Lactating Mothers (Bangladesh) 7 Rice subsidy (Philippines) 4 6 Subsidio familiar (SUF) (Chile) 3 6 Bono Vida Mejor (Honduras) 2 5 Bono Juancito Pinto (Bolivia) 12 Student stipend for primary education level (Bangladesh) 1 1.4 Jóvenes en Acción (Colombia) 11.4 Más Familias en Acción (Colombia) 1 1.1 0 10 20 30 40 50 60 70 80 90 100 Adequacy of transfer (% of total income/consumption of beneficiaries) d. Public work programs Productive safety nets program (PSNP) (Ethiopia) 8 14 Cash for work (Fato Fato) (Lesotho) 5 12 Food/cash-for-work programme (Malawi) 4 7 Public works programme (MASAF) (Malawi) 2 5 0 10 20 30 40 50 60 70 80 90 100 Adequacy of transfer (% of total income/consumption of beneficiaries) For total beneficiaries For beneficiaries in Q1 Source: original figure for this publication using atlas of Social Protection indicators of Resilience and equity (aSPiRe) household survey data. Note: Based on a sample of 83 programs (N = no. of programs) in 37 countries (n= no. of countries). Q1 = Poorest or first Quintile. 6 Impact on Poverty By providing direct transfers to targeted poor households, social assistance helps to reduce poverty. However, the actual impact depends on the program’s coverage and benefit level, as well as on how well the program is targeted and implemented. In other words, how many people receive how much and how beneficiaries are selected. Given that poverty reduction is an important goal of social assistance, the objective of this section is to quantify the impact of social assistance in terms of reducing poverty and inequality using household surveys. To achieve this objective, the per capita income or consumption of beneficiary households is measured before and after they receive the transfer to assess whether they have risen above the poverty line (to become non-poor) or to what extent they moved closer to the poverty line. Social assistance transfers make a significant contribution to raising people up out of poverty. Figure 6.1 uses household survey data from 54 countries with monetary information to estimate the impact of social assistance on reducing both absolute poverty35 and relative poverty (people in the bottom 20 percent of the welfare distribution, Q1).36 It estimates that social assistance transfers raised 37 percent of extremely poor people who were previously living below the absolute poverty line of US$2.15, cross this line to be out of extreme poverty. In terms of relative poverty, social assistance helped almost one in nine relatively poor individuals who were previously in the bottom quintile (Q1) to move up to the next quintile (Q2). Some countries have experienced sharp reductions in poverty as a result of the provision of social assistance transfers. For example, Georgia reduced its relative poverty headcount and poverty gap by 52 percent and 75 percent, respectively. Poland also reduced its relative poverty headcount and poverty gap by 38 percent and 67 percent, respectively. However, as discussed earlier, different social assistance programs have different objectives, and not all of them are designed or implemented to pull all beneficiaries out of poverty. 35. The global absolute poverty line specifies people living below US$2.15 per person per day (in US$ at 2017 PPP) as the extreme poor. 36. People in the bottom 20 percent of the welfare distribution (that is Q1) are considered to be the relative poor. 43 44 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Reduction in Poverty Due to Social Assistance Transfers by Extreme and Relative FIGURE 6.1   Poverty Lines, 2015–2022 (n = 41–54 countries) Social assistance directly contributes to poverty reduction, but not enough to lift all poor out of poverty. Reduction in poverty (as a % of poor beneficiaries previously living below the absolute and relative poverty lines) 50 45 40 37 30 20 20 11 10 0 Poverty headcount Poverty gap Poverty headcount Poverty gap reduction reduction reduction reduction Absolute poverty (n = 41) Relative poverty [poorest quintile/ Q1] (n = 54) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 41 and 54 countries with data on adequacy and absolute and relative poverty reduction, respectively (n = no. of countries). Social assistance transfers are reducing the extent to which households are poor, but they may not always be sufficient to lift all beneficiaries out of poverty. In other words, social assistance transfers are helping beneficiaries residing below the poverty line to move up in the welfare distribution, closer to the poverty line (reducing poverty gap), even if the amount of the cash transfers is not large enough for all poor beneficiaries to cross the poverty line (not reducing poverty headcount).37 The analysis shows that social assistance transfers are estimated to reduce the poverty gap by 45 percent when using the absolute poverty line, and 20 percent when using the relative poverty threshold (see Figure 6.1). The change in the Gini inequality index was 4 percent when using the relative poverty line (see Figure 6.2). These poverty impacts are likely to be underestimates, as household surveys do not capture the whole universe of social assistance programs or all beneficiaries in any country. Therefore, it can be inferred that the real impacts are likely to be even larger. Current coverage and adequacy levels limit the poverty reduction impacts of social assistance programs. As a country’s income level rises, social assistance is able to lift more people out of poverty. While there are many reasons why the poverty impact varies across countries, the main reasons relate to the adequacy, coverage, and targeting of the country’s social assistance programs. Also, it is important to remember that, by design, not all programs are just for the poor and benefit level is not set to lift their beneficiaries out of poverty (also see Box 6). 37. Depending on how far each beneficiary household is from the poverty line, transfer may be enough or not enough to help each household reach non-poor status. Impact o n Pove rty 45 Reduction in Poverty and Inequality from Social Assistance Transfers for Relatively Poor Beneficiaries, FIGURE 6.2   2015–2022 (n = 54 countries) Social assistance has larger impacts in reducing poverty in HICs, and very little in LICs. Reduction in poverty and inequality (as a % of poor beneficiaries previously living below the relative poverty lines) 40 39 32 31 30 26 21 22 20 20 18 17 14 15 11 12 11 9 8 10 7 8 6 6 6 5 4 3 2 3 2 1 1 2 0 1 0 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 54) (n = 4) (n = 13) (n = 16) (n = 2) (n = 5) (n = 14) (n = 6) (n = 17) (n = 26) (n = 5) Poverty headcount reduction Poverty gap reduction Inequality reduction (Gini) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 54 countries (n = no. of countries). Due to its small sample (two countries), MNA is included only for illustrative purposes. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Comparing the countries in different income groups, the average reduction  in the relative poverty headcount is only 1 percent in LICs. But this increases to 6–15 percent in MICs, and further increases to 22 percent in HICs. In the case of the poverty gap, the average reduction in LICs is 2 percent, while in LMICs and UMICs, it is 11 and 26 percent, respectively, and in HICs, it is 39 percent. The Gini inequality index is less affected by social assistance transfers than poverty, but they do reduce inequalities to some extent, ranging from 0.2 percent in LICs to 8 percent in HICs. In terms of regions, social assistance reduces poverty more in ECA, EAP, and LAC than in SAR and SSA, reflecting the lower coverage, financing, and adequacy of social assistance in those latter two regions. The impact of social assistance in reducing poverty is larger in rural areas than in urban areas. Aligned with the earlier findings (Figure 3.3), which showed that social assistance coverage is higher in rural areas, the reduction in poverty headcount was higher in rural areas than urban areas by 6 percentage points (Figure 6.3a). The reduction in the poverty gap was also higher in rural areas by 9 percentage points (Figure 6.3b). To ensure comparability with previous figures 6.1 and 6.2, the analysis is based on the bottom 20 percent of the population. However, it could also have used different poverty lines for urban and rural areas related to differences in the cost of living in the two geographic areas. Social assistance is generally but not always pro-poor, and poverty reduction is not always the primary program objectives. In fact, some programs are customized to support specific disadvantaged and marginalized groups who may not necessarily be the poorest in terms of 46 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond income or consumption. In this case, social assistance programs commonly use categorical characteristics to select beneficiaries as their main targeting method. Box 6 discusses the range of targeting approaches used for selecting beneficiaries of social assistance programs, though it is important to remember that there is no perfect targeting method that can identify and enroll the target population with 100 percent accuracy. Reduction in the Poverty Headcount and the Poverty Gap due to Social Assistance Transfers for Relative FIGURE 6.3   Poverty by Urban and Rural Areas, 2015–2022 (n = 41 countries) Social assistance has larger impacts in reducing poverty in rural areas, compared to urban areas. a. Reduction in the poverty head count Reduction of poverty from social assistance transfers (percent) 30 26 25 22 20 20 20 20 20 17 15 14 15 13 11 11 10 10 11 10 5 5 5 5 2 2 1 0 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 41) (n = 4) (n = 12) (n = 14) (n = 2) (n = 5) (n = 4) (n = 2) (n = 12) (n = 22) (n = 5) b. Reduction in the poverty gap Reduction of poverty from social assistance transfers (percent) 50 45 40 37 36 37 32 33 29 30 30 26 23 20 19 19 20 18 17 10 9 10 10 3 4 2 1 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 41) (n = 4) (n = 12) (n = 14) (n = 2) (n = 5) (n = 4) (n = 2) (n = 12) (n = 22) (n = 5) Rural Urban Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on a sample of 41 countries (n = no. of countries) for which data were available that disaggregated urban versus rural areas. LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Impact o n Pove rty 47 What Approaches are Used to Select Target Populations? BOX 6   Taking advantage of ASPIRE’s ongoing efforts in collecting information on key design features, this box presents preliminary findings on the targeting approaches used by social assistance programs. Of the sample of 1,173 active social assistance programs in 68 countries for which targeting information was collected, nearly two-thirds (743 programs) use a single targeting method, while the rest use a mix of methods. Of the latter group of 430 programs, 71 percent (306 programs) use a combination of two targeting methods, usually categorical and means tests (31 percent), categorical and proxy means tests (14 percent), or categorical and geographical (19 percent). The rest use a combination of three or more methods, with the most common being categorical, geographic, and either means tests or proxy means tests. Categorical characteristics—particularly those related to demography—appear to be the most popular targeting criteria in almost all country groups. Of the programs that use a single targeting method, categorical targeting is by far the most popular (70 percent or 522 out of 743 programs in 63 countries). This is followed by means tests (12 percent or 96 programs in 29 countries), which tend to be used most in UMICs and HICs in ECA, EAP, and LAC. Geographic, proxy means testing (PMT) and universal targeting are rarely used alone. Targeting Methods Used for Social Assistance Programs (N = 1,173 programs from n = 68 countries) Nearly two-thirds of social assistance programs use a single method in selecting beneficiaries. Method Number of programs Proportion of total active Number of countries SA programs Single targeting method 743 63% Categories 549 47% 64 Demographic/Categorical 522 45% 63 Geographical 27 2% 18 Welfare targeting 131 11% 49 Means/income test 92 8% 29 Proxy means test 27 2% 15 Community-based 9 1% 5 Hybrid means test 3 0% 2 Universal 30 3% 14 Self-targeting 20 2% 8 Other (incl. Pension-tested) 13 1% 8 Mixed targeting method 430 37% Two targeting methods 306 26% Three or more targeting methods 124 11% Total active SA programs 1,173 100% 68 Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Note: Based on a sample of 1,173 programs in 68 countries (N = no. of programs and n = no. of countries). SA = Social Assistance. The results are in line with similar earlier work (see: Grosh et al. 2022) 48 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond What Approaches are Used to Select Target Populations? (continued) BOX 6   Single Targeting Methods Used by Social Assistance Programs (N = 743 programs from n = 66 countries) Categorical criteria are widely used in selecting beneficiaries. Composition of social assistance programs by targeting methods 100 0 3 3 3 4 3 3 2 100 13 1 11 11 5 1 10 3 1 5 1 3 1 11 4 11 4 10 3 5 90 4 5 6 5 4 21 3 11 4 7 4 5 4 7 4 11 80 2 6 8 7 3 12 5 11 4 3 17 10 5 70 40 10 60 3 50 10 40 80 76 73 73 73 70 67 67 65 30 56 20 40 10 0 Total EAP ECA LAC MNA SAR SSA LIC LMIC UMIC HIC (n = 66; (n = 6; (n = 19; (n = 14; (n = 14; (n = 6; (n = 18; (n = 9; (n = 15; (n = 31; (n = 10; N = 743) N = 46) N = 290) N = 208) N = 20) N = 39) N = 140) N = 74) N = 136) N = 385) N = 148) Categorical Means/income test Universal Geographical Proxy-means test Self-targeting Other Community-based Hybrid means test Pension-tested Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Note: Based on a sample of 743 programs in 66 countries (N = no. of programs and n = no. of countries). LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Economic inclusion measures designed to boost earning opportunities for poor beneficiaries have the potential to significantly enhance poverty reduction. However, the current coverage of such programs is very small. Despite growing interest to scale-up, economic inclusions interventions reach only 70 million people either directly or indirectly according to the latest estimate (Arévalo-Sánchez et al. 2024), and only a small fraction of social assistance programs include such interventions (0.03 percent according to the ASPIRE data). This undermines their potential effectiveness in reducing poverty but also highlights the potential for using them more widely (and at a larger scale). Particularly in LICs, some governments are implementing bundled inclusion packages of business support, mentoring, and coaching along with social assistance programs (see Annex Table A2.1 for more details). 7 Conclusions – Summary of Main Findings and Ways Forward As a key component of social protection systems, social assistance has been making progress which is notable but still insufficient. Over the past decades, the emergence of large-scale national programs and delivery systems has been a particular feature of the evolution of social assistance in various contexts. Also, the occurrence of shocks and crises, including the COVID-19 pandemic, has led to the widespread adoption of emergency responses to meet acute and temporary needs beyond supporting the chronically poor. Despite the progress and needs, social assistance faces notable gaps in all key metrics on coverage, adequacy, and financing. The paper concludes by recapitulating the main findings yielded from the World Bank’s ASPIRE database, while highlighting selected recommendations that can address existing challenges to improve the effectiveness of social assistance and social protection in going forward. First, clear coverage gaps remain despite progress: the world is only just over halfway to achieving its minimum and most fundamental mission of reaching the poorest. Currently, more than two in five people in the poorest 20 percent of the population worldwide lack access to social assistance, indicating that there is still a long way to go in achieving the Sustainable Development Goal (SDG-1.3) of implementing appropriate social protection for all. Without accelerated efforts, it will take an additional two decades for social protection to reach the poorest 20 percent of the population (World Bank 2025). Larger coverage gaps remain, therefore longer time will be required in places where the need is greatest (such as low-income and FCV countries). Countries can reach more poor people by enhancing delivery systems and institutional capacity, even with existing budgets. Efficiency gains can be made through strengthening the ability to identify and deliver benefits to target populations. There is abundant scope for such efficiency gains by improving targeting by prioritizing the poorest (where applicable) and by minimizing targeting errors. For example, only 54 percent of current social assistance beneficiaries belong to the bottom 40 percent of the population. Many social assistance programs use categorical criteria – instead of welfare metrics – to select their beneficiaries. Besides better targeting, efficiency gains can be achieved by modernizing and digitizing delivery systems – digital payments; digital, dynamic, interoperable social registry; digital ID; monitoring, recertification and exit mechanisms. This will enhance government’s institutional capacity to directly and swiftly deliver social assistance programs to the beneficiaries. This will also result in cost savings through lower admin costs and leakages (including corruption). 49 50 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Second, besides coverage, the benefit level of social assistance is another crucial parameter, which remains low. Globally, the adequacy of social assistance benefits has not essentially changed since 2010. For poor beneficiaries belonging to the bottom 20 percent of the population, social assistance benefits constitute 18 percent of their total welfare (which is measured by either income or consumption). When using all beneficiaries as a reference group, the benefit level decreases to 11 percent of their total welfare. In dollar terms, on average, the amount that beneficiary households receive from social assistance is about US$1.1 per day (2017 PPP), with heterogeneity across countries and social assistance instruments. The lower benefit size and adequacy, particularly result in smaller impacts on poverty reduction (headcount). For example, in low and lower-middle income countries, social assistance lifted only 1 to 6 percent of the bottom quintile of the population to the next quintile. Ensuring adequate support requires not only a better system to regularly revisit benefit amounts but also a better mechanism that can refer people to other services. Unlike social insurance, which typically embeds an indexation system to regularly adjust its benefit amount, it is not a norm for social assistance to systematically or periodically revise its transfer amounts to ensure adequacy over time. To protect purchasing power hence program impacts, it is important for social assistance programs to incorporate such adjustment mechanisms, particularly in highly inflationary contexts. Countries also need to invest in better case management and referral system which can connect households to different programs and services they need, based on their profiles and challenges. In this process, social assistance programs – particularly cash transfers facilitated by the social registry – can serve as a gateway to provide accompanying measures and link to other programs or services, including social or special services and economic inclusion programs. In addition, scaling up active labor market programs and voluntary saving schemes could pave the way for certain segments of vulnerable households to transition to better and more self-reliant economic status, off the payroll of social assistance programs. Third, financing is a critical binding constraint, with relative spending on social assistance remaining largely unchanged for nearly a decade and a half. Spending on social assistance has stayed at around 1.5 percent of GDP since 2010, with temporary boosts in countercyclical spending in response to economic shocks or the global pandemic. The income level of a country is a strong determinant of its social assistance budget. High income countries allocate more than 30  times on social assistance than low income countries in per capita terms. LICs spend an average of US$16 per capita annually on social assistance, while MICs spend US$196 (US$66 by LMICs and US$331 by UMICs), and HICs spend US$495. Social assistance is largely financed through domestic resources, except for low-income and FCV countries, where the financing landscape heavily relies on external grants. Countries use domestic sources to finance at least two-thirds of their total social assistance budgets, highlighting concrete national commitments for social assistance. Grants from donors constitute 21 percent, while the remaining 12 percent comes from the combination of two financing sources (such as grants and non-grants, or external and domestic sources). However, the capacity for mobilizing sizable domestic revenues in low-income and FCV settings is limited. For example, among LICs, full grants account for 77 percent of total social assistance spending in FCV countries compared to only 10 percent in non-FCV countries. C o nclus ions – Summ ary of Main Findi ngs an d Ways Forwa r d 51 For more and better spending, countries can leverage both existing and new avenues of funding to increase their investment in social assistance. As discussed above, countries can increase the efficiency of social assistance by improving their delivery systems, which can lead to reaching more of the target population and lower operational costs. In terms of creating fiscal space, abolishing regressive (and often environmentally damaging) subsidies and reforming tax policies (e.g., VAT “sin tax” on tobacco, and alcohol and other beverages) are prime candidates for freeing up resources to be reallocated to progressive social assistance interventions. Some countries even have established sovereign wealth funds (through the revenues from natural resources, for example), which are leveraged to finance social causes or tap into innovative financing sources, particularly when objectives or outcomes are related to climate change or the environment (for example, Payment for Ecosystem Services). Additionally, unconventional and newer external funding avenues have emerged, such as Diaspora Funds, Climate Funds, and debt-for-social protection swaps. However, further exploration is needed and the magnitude of funding from these sources remains unknown. As a way forward, the world needs a renewed recognition of social assistance which has untapped potential not only to support the poor but also to contribute to universal social protection. A wealth of evidence demonstrates that social assistance is a successful, relevant, and enduring part of national social protection strategies around the world. As a result, countries are increasingly looking to social assistance to meet a wide variety of objectives. There is a growing practice on how countries can leverage social assistance as part of their adaptive social protection systems in good and bad times, not only by delivering benefits more effectively but also as a conduit for protecting and promoting people’s livelihoods and for increasing equity, resilience, and opportunity in the country as a whole. A clear set of priorities involves continued investment to make social protection systems more progressive, adaptive and inclusive. This can be achieved by continuing to invest in robust social protection systems encompassing institutions, financing, data, and programs. Being progressive is a foundational priority that involves prioritizing the populations in greater need, including and particularly the poorest. Being scalable means having the flexibility to scale up support quickly and efficiently to protect broader populations from temporary shocks and scale it down once the crises are over. As crises become more frequent, there is also a need to develop adaptive social assistance mechanisms to protect vulnerable groups and to increase their resilience by integrating social protection, disaster risk reduction, and climate change adaptation measures. In moving forward, these investments are critical to complement wider social protection pillars – including social insurance, and labor market and employment programs – and to ensure adequate social protection support in good times and bad. References Ajayi, Kehinde, Robin Audy, Thomas Bossuroy, Ayodele Fashogbon, and Oyebola Okunogbe. Forthcoming. “National Social Safety Net Project.” Working Paper. World Bank, Washington, DC. Andrews, C., A. Hsiao, and L. Ralston. 2018. “Social Safety Nets Promote Poverty Reduction, Increase Resilience, and Expand Opportunities.” In Realizing the Full Potential of Social Safety Nets in Africa, edited by K. Beegle, A. Coudouel, and E. Monsalve, 87–139.  Africa Development Forum; Washington, DC: World Bank. http://hdl.handle.net/10986/29789. Arévalo-Sánchez, Inés, Janet Heisey, Sarang Chaudhary, Timothy Clay, Victoria Strokova, Puja Vasudeva Dutta, and Colin Andrews. 2024.  “The State of Economic Inclusion Report 2024: Pathways to Scale.” World Bank, Washington, DC. http://hdl.handle.net/10986/42408. Arze del Granado, Francisco Javier, David Coady, and Robert Gillingham. 2012. “The Unequal Benefits of Fuel Subsidies: A Review of Evidence for Developing Countries.” World Development 40 (11): 2234–2248. Banerjee, Abhijit, Rema Hanna, Benjamin A. Olken, and Diana Sverdlin Lisker. 2024. “Social Protection in the Developing World.” Journal of Economic Literature 62 (4): 1349–1421. Bastagli, F., J. Hagen-Zaker, L. Harman, B. V. Sturge, and T. Schmidt. 2016. Cash Transfers: What Does the Evidence Say? A Rigorous Review of Programme Impact and of the Role of Design and Implementation Features. London: Overseas Development Institute [cited 2018 Nov 06].  The Evolution of Unconditional Cash Transfers in South Africa. Blofield, M., N. Lustig, and M. Trasberg. 2021. „Social Protection During the Pandemic: Argentina, Brazil, Colombia, and Mexico.” Centre for Global Development. https://www.cgdev.org/sites/default/files/Note​ -Blofield_Lustig_Trasberg-Revised_0.pdf. Bossuroy, Thomas, Markus Goldstein, Bassirou Karimou, Dean Karlan, Harounan Kazianga, William Pariente, Patrick Premand, Catherine C. Thomas, Christopher Udry, Julia Vaillant, and Kelsey A. Wright. 2022. “Tackling Psychosocial and Capital Constraints to Alleviate Poverty.” Nature 605 (7909): 291–297. Bossuroy, Thomas, Dean Karlan, William Parienté, Patrick Premand, Christopher Udry, Julia Vaillant, and Kelsey Wright. 2024. “Impact des mesures d’inclusion productive du programme Yom Kom Kom au Sénégal.” Banque Mondiale, Washington, DC. Bowen, Thomas, Carlo del Ninno, Colin Andrews, Sarah Coll-Black, Ugo Gentilini, Kelly Johnson, Yasuhiro Kawasoe, Adea Kryeziu, Barry Maher, and Asha Williams. 2020. “Adaptive Social Protection: Building Resilience to Shocks.” International Development in Focus series. World Bank, Washington, DC. doi:10.1596/978-1-4648-1575-1. Brollo, F., David Coady, Samir Jahan, and Riki Matsumoto. 2024. “Challenges Facing SSNs in Emerging and Developing Economies: An Illustration Using the World Bank’s ASPIRE Database.” IMF Working Paper No. 24/96, International Monetary Fund, Washington, DC. https://www.imf.org/en/Publications​ /WP/Issues/2024/05/03/Challenges-Facing-SSNs-in-Emerging-and-Developing-Economies-548582. 52 Re f e r e nce s 53 Crosta, Tommaso, Dean Karlan, Finley Ong, Julius Rüschenpöhler, and Christopher R. Udry. 2024. “Unconditional Cash Transfers: A Bayesian Meta-Analysis of Randomized Evaluations in Low and Middle Income Countries.” NBER Working paper 32779. Damania, Richard, Esteban Balseca, Charlotte de Fontaubert, Joshua Gill, Kichan Kim, Jun Rentschler, Jason  Russ, and Esha Zaveri. 2023.  “Detox Development: Repurposing Environmentally Harmful Subsidies.” World Bank, Washington, DC. http://hdl.handle.net/10986/39423 Fiszbein, A., and N. Schady. 2009. Conditional Cash Transfers: Reducing Present and Future Poverty (English). A World Bank Policy Research Report, World Bank, Washington, DC. http://documents1.worldbank.org​ /curated/en/914561468314712643/pdf/476030PUB0Cond101Official0Use0Only1.pdf. Gavrilovic, M., M. Rubio, F. Bastagli, R. Hinton, S. Staab, R. G. Goulder, C. Bilo, et al. 2022. “Gender- Responsive Social Protection Post-COVID-19.” Science March 11, 2022 (6585): 1111–1113. doi:10.1126​ /­science.abm5922. https://www.cgdev.org/publication/digitaltechnology-social-assistance​-transfers-covid-19​ -relief-lessons​-selected-cases. Gencer, D., and E. Akcura. 2022. “Amid Energy Price Shocks, Five Lessons to Remember on Energy Subsidies.” World Bank Blogs. May 13. https://blogs.worldbank.org/en/energy/amid-energy​-price-shocks​ -five-lessons-remember-energy-subsidies. Gentilini, Ugo, Mohamed Almenfi, Hrishikesh TMM Iyengar, Yuko Okamura, John Austin Downes, Pamela Dale, et al. 2022a. “Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures.” World Bank, Washington, DC. https://openknowledge.worldbank.org/entities/publication​ /3bc00930-8388-5d60-86a9-a579de8a5b28. Gentilini, Ugo, Mohamed Almenfi, Hrishikesh TMM Iyengar, Yuko Okamura, Emilio Raul Urteaga, Giorgia Valleriani, Jimmy Muhindo, and Sheraz Aziz. 2022b. “Tracking Global Social Protection Responses to Price Shock: Living Paper, v.1.” Social Protection and Jobs Discussion Paper Series, World Bank, Washington, DC. http://hdl.handle.net/10986/37441. Gentilini, Ugo, Mohamed Almenfi, Hrishikesh TMM Iyengar, Yuko Okamura, Emilio Raul Urteaga, Giorgia Valleriani, Jimmy Muhindo, and Sheraz Aziz. 2022c. “Tracking Global Social Protection Responses to Price Shock: Living Paper, v.2.” Social Protection and Jobs Discussion Paper Series, World Bank, Washington, DC. http://hdl.handle.net/10986/37441 Gentilini, Ugo, Mohamed Almenfi, Hrishikesh TMM Iyengar, Yuko Okamura, Emilio Raul Urteaga, Giorgia Valleriani, and Sheraz Aziz. 2022d. “Tracking Global Social Protection Responses to Price Shock: Living Paper, v.3.” Social Protection and Jobs Discussion Paper Series, World Bank, Washington, DC. http://hdl​ .handle.net/10986/37441. Gentilini, Ugo, Mohamed Almenfi, Hrishikesh TMM Iyengar, Yuko Okamura, Emilio Raul Urteaga, Giorgia Valleriani, and Sheraz Aziz. 2022e. “Tracking Global Social Protection Responses to Price Shock: Living Paper, v.4.” Social Protection and Jobs Discussion Paper Series, World Bank, Washington, DC. http://hdl​ .handle.net/10986/37441. Gentilini, Ugo, Mohamed Almenfi, Hrishikesh TMM Iyengar, Yuko Okamura, Emilio Raul Urteaga, Giorgia Valleriani, Sheraz Aziz, Mohammad Farid Al Azim Bin Noruzi, and Margret Chu. 2023. “Global Social Protection Responses to Inflation: Living Paper, v.5.” Social Protection and Jobs Discussion Paper Series, World Bank, Washington, DC. http://hdl.handle.net/10986/37441. Gentilini, U. 2023. “Where has all the social protection gone… and where will it come from?” Brookings, November 14, 2023. https://www.brookings.edu/articles/where-has-all-the​-social-protection-gone​-where​ -will-it-come-from/. Gentilini, U., H. TMM Iyengar, G. Valleriani, S. Aziz, H. R. Arimbi, M. A. T. Nogueira, and C. Cheng. 2024. “Keep The Pace. How Inflation Erodes Cash Transfers And What to Do About It.” World Bank. https:// documents.worldbank.org/en/publication/documents-reports/documentdetail/099112124114560336​/­p17 65851a664ed0c6196651ca8e74d0aea0. 54 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Grosh, Margaret, Phillippe Leite, Matthew Wai-Poi, and Emil Tesliuc. 2022. Revisiting Targeting in Social Assistance: A New Look at Old Dilemmas. Human Development Perspectives series, World Bank, Washington, DC. http://hdl.handle.net/10986/37228. Hartley, G. R., and M. Abels. 2025. “Riding the Demographic Wave: Pensions and Retirement Income in an Aging World.” World Bank. Hobson, E. W., H. TMM Iyengar, G. Valleriani, and M. Almenfi. Forthcoming. Delivery systems readiness and the scale-up of social assistance. In How Scale-Up Happens: Explaining the Role of Financing, Political Economy and Delivery in Expanding Social Assistance, edited by U. Gentilini, 234–289. Honorati, Maddalena,  Ugo Gentilini, and Ruslan  G. Yemtsov. 2015. The State of Social Safety Nets 2015 (English).  Washington, DC: World Bank Group.  http://documents.worldbank.org/curated​ /­en/415491467994645020/The-state-of-social-safety-nets-2015 ILO (International Labour Organization). 2018. Care Work and Care Jobs for the Future of Decent Work. ILO. https://www.ilo.org/publications/major-publications/care-work-and-care-jobs-future-decent-work. IMF (International Monetary Fund). 2024. “Macroeconomic Developments and Prospects for Low-Income Countries - 2024,” IMF Policy Paper No. 2024/011, International Monetary Fund, Washington, D.C. https://www.imf.org/en/Publications/Policy-Papers/Issues/2024/04/02/Macroeconomic​ -Developments-and-Prospects-For-Low-Income-Countries-2024-547064. Lawson, Cina, Morlé Koudeka, Ana Lucía Cárdenas Martínez, Luis Iñaki Alberro Encinas, and Tina George Karippacheril. “Novissi Togo: Harnessing Artificial Intelligence to Deliver Shock-Responsive Social Protection.” Social Protection and Jobs Discussion Paper No. 2306, World Bank, Washington, DC. http:// hdl.handle.net/10986/40405. Light, C., G. E. Nwaobia, and L. I. Nwobia. 2024. “Effects of Conditional and Unconditional Cash Transfers on Poverty Reduction, Education, and Health Outcomes in Sub-Saharan Africa—A PRISMA Approach.” Journal of Poverty. https://www.tandfonline.com/doi/pdf/10.1080/10875549.2024.2426807. Marin, G., and R. Palacios, 2022. “The Role of Digital in the COVID-19 Social Assistance Response.”  World  Bank,  Washington,  D.C.  https://documents1.worldbank.org/curated/en/09983000​ 9302217091/pdf/P1731660f8c52f062092ac00d53c648bac7.pdf. Mukherjee, A., L. Bermeo, Y. Okamura, J. Vulembera, and P. Bance, 2023. “Digital-first Approach to Emergency Cash Transfers: STEP-KIN in the Democratic Republic of Congo (English).” Social Protection and Jobs Discussion Paper No. 2302. World Bank, Washington, DC. https://documents1.worldbank.org​ /­curated/en/099935104272316767/pdf/IDU05debcd9500bf004a580a48b0c6c201068bdc.pdf. Mukherjee, Anit, Okamura Yuko, Ugo Gentilini, Mohamed Almenfi and Adea Kryeziu. 2022. “How Can Cash Transfers Support Energy Subsidy Reforms? Emerging Lessons from Four Case Studies.” Draft. Social Protection and Jobs Global Unit, World Bank, Washington D.C. OECD. 2024. “Poverty Rate Definition.” OECD iLibrary. https://doi.org/10.1787/7f420b4b-en. Okamura, Yuko, Tim Ohlenburg, and Emil Tesliuc. 2024. “Scaling up Social Assistance Where Data is Scarce: Opportunities and Limits of Novel Data and AI.” Social Protection and Jobs Discussion Paper No. 2402. World Bank, Washington, DC. http://hdl.handle.net/10986/41548.  Peterman, Amber, Jingying Wang, Kevin Kamto Sonke, and Janina Steinert. 2024. “Social Safety Nets, Women’s Economic Achievements and Agency: A Systematic Review and Meta-Analysis.” CGD Working Paper 684, Center for Global Development, Washington, DC.  https://www.cgdev.org/publication​ /­social-safety-nets-womenseconomic-achievements-and-agency-systematic-review-and-meta. Pople, A., R. V. Hill, S. Dercon, and B. Brunckhorst, 2021. “Anticipatory Cash Transfers in Climate Disaster Response.” Working Paper No. 6, Centre for Disaster Protection, London. Rodríguez Alas, C. P., Lopez, V., and I. Mujica. 2025. Unlocking the Potential of Household Surveys to Measure Women’s Access to Social Protection. World Bank. Sepulveda, M. 2019. “Biometric Technology and Beneficiary Rights in Social Protection Programmes.” International Social Security Review 72: 3–28. 10.1111/issr.12219. https://www.researchgate.net​ /­p ublication/337947534_Biometric_technology_and_beneficiary_rights_in_social_protection​ _programmes. Re f e r e nce s 55 Sulemana, M., B. F. Malongza, and M. Abdulai, 2018. “Assessment of the Livelihood Empowerment against Poverty Programme in Karaga district, Ghana.” Development in Practice 29 (4): 437–447. https://doi.org/10 .1080/09614524.2018.1551859. Suri, T., and W. Jack. 2016. “The Long-run Poverty and Gender Impacts of Mobile Money.” Science 354 (6317): 1288–1292. Tesliuc, E., C. Rodrigues-Alas, U. Zafar, and A. S. M. Cordova. Forthcoming. “The State of Financing Social Assistance in Low- and Middle-Income Countries.” In How Scale-Up Happens: Explaining the Role of Financing, Political Economy and Delivery in Expanding Social Assistance, edited by U. Gentilini, 66–95. Tesliuc, E., and A. S. Martinez. 2025. Mind the Gap: Coverage, Adequacy and Financing Gaps in Social Protection for the Extreme Poor and the Poorest Quintile. World Bank. Tesliuc, E., and M. B. Fontenez. 2025. Adaptive Social Protection Agenda: Lessons from Responses to the COVID-19 Shock. World Bank.  Thota, A., R. de Groot R., T. Palermo, K. Roelen, V. Petit, N. Tirivayi, V. Groppo, B. Hickler, J. Yablonski, and S. Bakrania. 2024. “What Works to Improve Outcomes for Children? Evidence on Cash Transfers with Social and Behaviour Change (SBC) Programming in Low- and Middle-Income Countries.” UNICEF Innocenti – Global Office of Research and Foresight. Florence, Italy. https://www.unicef.org/innocenti/ media/3271/file/UNICEF-Innocenti-What-Works-Improve-Outcomes-Children-2024.pdf. Ubah, U. T., A. Mukherjee, B. Webster, W. Cunningham, and G. Marin. 2023. “Delivering Digital G2P Payments to Urban Informal Populations. Lessons and Future Policy Implications from COVID-19 Responses.” World Bank, Washington, DC. https://documents1.worldbank.org/curated/ en/099092023102013224/pdf/P1691710b71ebd01608e180e11a3beed63d.pdf. UN Habitat. 2022. World Cities Report 2022: Envisaging the Future of Cities. United Nations Human Settlements Programme, Nairobi, Kenya. https://unhabitat.org/world-cities-report-2022​-envisaging-the-future​-of​-cities. World Bank. 2018.  The State of Social Safety Nets 2018.  Washington, DC: World Bank. http://hdl.handle. net/10986/29115. World Bank. 2022b. Charting a Course towards Universal Social Protection: Resilience, Equity, and Opportunity for All. World Bank, Washington, DC. http://hdl.handle.net/10986/38031. World Bank. 2024. “Poverty and Inequality Platform (PIP).” pip.worldbank.org. World Bank. 2025. State of Social Protection Report 2025: The 2-Billion-Person Challenge. Washington, DC: World Bank. Annex 1 The Evolution of Social Assistance Coverage, Spending, and Adequacy Following the massive scaling up of social assistance during the COVID-19 pandemic and the crises that followed, early evidence shows that countries are now scaling back their cash transfers closer to pre-crisis levels. Figure A1.1 plots the coverage rate of the largest cash transfer—whether regular or emergency—in each country over the past four years according to administrative data. Following the unprecedented COVID-19 social protection scale-up, coverage seems to be back FIGURE A1.1  Changes in the Coverage of Flagship Cash Transfer Programs, 2019–2022 (n = 71 countries) Cash transfers coverage (% pop) 100 80 60 40 20 0 Pre-COVID-19 COVID-19 Post-COVID-19 (2019 circa) (2020–2021) (2022) Sources: Gentilini (2023) using Gentilini et al. (2022a, 2023), ASPIRE data, and World Bank staff estimates.  Note: Based on sample of 71 countries (n = no. of countries). Coverage was computed based on the largest cash-based program in terms of coverage in the country. Cash-based programs include unconditional cash transfers, conditional cash transfers, social pensions, and public works. Population data are from 2021. Based on a sample of 71 cash transfer programs with available data for ‘before COVID’ (circa 2019), ‘COVID’ (2020–2021), and ‘post-COVID’ (2022). 56 The Evolutio n of S ocial Assistance Cove r age , Sp ending, a nd Ade quacy 57 to normal or pre-COVID-19 levels. The fact that countries have been knee-deep in short-term inflation responses has contributed to a slight bump in coverage compared to pre-pandemic levels (currently global coverage is at 12 percent compared to 10 percent before COVID-19). FIGURE A1.2  Coverage of Social Assistance among the Total Population and the Poorest Quintile, by FCV and Non-FCV contexts and in LICs and LMICs, 2015–2022 (n = 9–23 countries) Social assistance coverage (as % of total population and poorest quintile) in FCV and non-FCV countries 50 47 40 38 38 30 27 20 10 0 FCV (n = 9) Non-FCV (n = 23) Total population Poorest quintile Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Note: Based on sample of 32 countries, consisting of 23 non-FCV and 9 FCV countries (n = no. of countries), in LIC and LMICs. LIC = Low Income Countries, LMIC = Lower Middle Income Countries. FCV = Fragile, Conflict and Violence. Spending on Social Safety Nets and Share of Spending Outside of Poorest FIGURE A1.3   Quintile Share of spending outside of poorest quintile 100 BFA ZMB UGA TLK TCD 80 NER SEN MDV CMR MWI NPL CIV KEN ETH HND BWA SLE LBR 60 COD TZA ZWE UZB BGD KGZ 40 GHA MMR DJI 20 0 0 1 2 3 Spending in social safety nets (percent of GDP) (continued) 58 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond Spending on Social Safety Nets and Share of Spending Outside of the Poorest FIGURE A1.3   Quintile (continued) 6 5 4 3 2 Average spending in the poorest quintile 1 0 hio e My erooe a n nz ar ZimZam ia ba bia Ke we Ho Nig a nd er Gh ras Dj ana an i da Rw DRC jik da Uz C tan Ba eki ad lad n Sie Sen esh L l rk ibe i Ky na F ria yz N so Tim ep al -L c Le este o ld ia Maives Ug out rra ega Bu L law Et eon or ubli ny ng sta m ir th an ep Ma p Ta nm a Ta an b h Ca d'Ivo b u is so ib R i te Co rg SSN spending in poorest quintile SSN spending outside poorest quintile Sources: IMF (2024) using ASPIRE data, World Economic Outlook, and IMF staff calculations. FIGURE A1.4  Cash and Subsidy Measures Introduced in Response to Inflation Crisis (2022–2023) Number of programs announced 100 83 80 60 40 28 20 1 0 Au 1 Se 1 Oc 1 No 1 De 1 Ja 1 Fe 2 Ma 2 Ap 2 Ma 2 Ju 2 22 Au 2 Se 2 Oc 2 No 2 De 2 Ja 2 Fe 3 Ma 3 Ap 3 3 l-2 2 2 t-2 v-2 c-2 2 2 r-2 r-2 2 2 r-2 r-2 y-2 l-2 2 2 t-2 v-2 c-2 g- p- n- b- n- b- n- g- p- Ju Ju Cash Subsidies Source: Original figure for this publication using data from versions 1–5 of the World Bank’s Price Shock trackers (Gentilini et al. 2022b, c, d, e, 2023). The Evolutio n of S ocial Assistance Cove r age , Sp ending, a nd Ade quacy 59 FIGURE A1.5  Composition of Social Protection Response to Inflation Crisis by Instrument, 2022-2023 (n = 178 countries) Composition of social protection measures 100 4 5 16 16 14 19 19 20 21 21 20 25 6 5 13 41 13 1 75 6 7 3 36 23 11 32 33 32 31 36 36 38 50 36 8 17 30 1 6 6 7 3 7 42 2 1 7 6 9 7 3 3 2 3 9 7 2 25 4 47 6 31 31 32 3 35 31 31 34 28 24 3 23 13 0 Total ECA LCA EAP MNA North SAR SSA LIC LMIC UMIC HIC (n = 178) (n = 47) (n = 38) (n = 17) (n = 21) America (n = 5) (n = 47) (n = 24) (n = 44) (n = 46) (n = 64) (n = 3) Social assistance Social insurance Labor market programs Subsidies Trade related measures Tax measures Source: Gentilini et al. (2023). Notes: Based on data from 178 countries (n = no. of countries). LIC = Low Income Countries, LMIC = Lower Middle Income Countries, UMIC = Upper Middle Income Countries, HIC = High Income Countries, EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Pre-COVID-19 Social Assistance Spending and Country Income Level, 2019 (n = 76 countries) FIGURE A1.6   Social assistance spending (as % of GDP) 8 y = 0.3268x – 1.7079 R² = 0.0696 6 4 2 0 6 7 8 9 10 11 12 Country income level (natural log of GDP per capita, constant US dollar PPP 2017) EAP ECA LAC MNA SAR SSA Linear (All countries) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Note: Based on sample of 76 countries (n = no. of countries) with social assistance expenditure data for 2019. EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. 60 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond FIGURE A1.7  During COVID-19 Social Assistance Spending and Country Income Level, 2020–2021 (n = 76 countries) Social assistance spending (as % of GDP) 8 y = 0.5538x – 3.0351 R² = 0.0793 6 4 2 0 6 7 8 9 10 11 12 Country income level (natural log of GDP per capita, constant US dollar PPP 2017) EAP ECA LAC MNA SAR SSA Linear (All countries) Source: Original figure for this publication using Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) administrative data. Note: Based on sample of 76 countries (n = no. of countries) with social assistance expenditure data for 2020-21. EAP = East Asia and Pacific, ECA = Europe and Central Asia, LAC = Latin America and the Caribbean, MNA = Middle East and North Africa. SAR = South Asia Region, SSA = Sub-Saharan Africa. Annex 2 Economic Inclusion Implemented with Social Assistance Economic inclusion programs aim to unlock the productive potential of the poorest and most vulnerable individuals, gradually integrating them into broader economic development processes. By taking a multidimensional approach, these programs address the multiple constraints faced by the poor and vulnerable. Participants receive business capital, consumption support, life and business skills training, and rigorous coaching to help them to start household enterprises and bring their products and services to the market. In some contexts, participants are provided with job search assistance, technical and vocational training, and referrals to integrate them into the labor market. Despite the global economic slowdown that began in 2020, a renewed surge in the number of economic inclusion programs is under way globally. The World Bank’s State of Economic Inclusion Report 2024 presents data on 405 economic inclusion programs across 88 countries. The programs support over 15 million households and benefits over 70 million individuals, most of whom live in LICs. The report highlights the 53 government-led programs in 45 countries that have built their economic inclusion efforts onto their existing social protection programs, including social assistance and labor market programs. These programs currently serve around 4 million households directly and 18.6 million individuals either directly or indirectly. Forty-one of these programs focus exclusively on supporting self-employment opportunities, diversifying income sources, and boosting productivity. Key programs at scale include the Productive Social Safety Net Project in Tanzania, which reaches over 330,000 households, and the Social Support for Resilient Livelihoods Project in Malawi, which reaches 350,000 households. While coverage so far is modest, there is potential to scale up these economic inclusion programs based on impact evidence. Currently, only 0.03 percent of social assistance programs globally include economic inclusion interventions according to ASPIRE data (see Annex Table 4.1), which suggests that there are untapped opportunities for extending their coverage. In Niger, adding economic inclusion measures to a national safety net program increased business revenues by 102 percent due to the creation of new off-farm business activities (Bossuroy et al. 2022). In Nigeria, a livelihood package provided as part of the country’s National Social Safety Nets Program increased participants’ total household earnings by 24 percent and raised profits from household enterprises by 45 percent (Ajayi et al. forthcoming). Senegal’s Yook Koom Koom program, implemented in urban areas, yielded a 22 percent increase in business revenues for 61 62 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond participants (Bossuroy et al. 2024). These programs are also proving to be cost-effective and can be high-return investments. For example, in Niger, the benefit-cost ratio of the economic inclusion measures was 127 percent 18 months after the intervention. TABLE A2.1  Coverage of Social Assistance Beneficiaries Country income group Share of PEI with SA component in total SA coverage (%) Low income 2.56 Lower middle income 0.17 Upper middle income 0.005 High income 0.00 Total 0.03 Source: Original table for this publication using data from World Bank’s Partnership for Economic Inclusion (PEI) and Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) household survey data. Annex 3 Flagship and Temporary Cash Transfer Programs Presented in Figure 3.7 Argentina - Asignación Universal por Hijo para Protección Social (AUH) for 2009–2019 and 2021; Ingreso Familiar de Emergencia (IFE), a COVID-19 response, for 2020; and Cash to Retirees and Pensioners, an inflation crisis response, for 2022. Brazil - the Bolsa Familia Program (PBF) for 2004–2019; the Auxílio Emergencial (AE), a COVID-19 response, for 2020–2021; and Auxílio Brasil (AB) for 2022. Chile - Single Family Allowance, which was created in 1981, for 2002–2019 and 2022; Ingreso Familiar de Emergencia or Emergency Family Income (IFE), a COVID-19 response, for 2020–2021. Colombia - Más Familias en Acción (MFA) for 2001–2022; and Ingreso Solidario or Solarity Income (IS), a COVID-19 response, for 2020-2022. During COVID-19, the IS program coverage was not supposed to have any overlap with the MFA program beneficiaries, but in reality, there might have been some overlap. Ghana - Livelihood Empowerment Against Poverty (LEAP) for 2008-2023; and the COVID-19 Emergency Cash Transfer, a COVID-19 response, for 2021–2022. Indonesia - Program Keluarga Harapan or Family Hope Program (PKH) for 2009–2021; Bantuan Langsung Tunai (BLT) Dana Desa,  a COVID-19 response, for 2020; and Bantuan Langsung Tunai (BLT) Minyak Goren, an inflation response, for 2022. Jordan - Takaful for 2019–2024, the National Aid Fund (NAF), an old cash transfer program, for 2017–2024; Emergency Cash (a COVID-19 response) for 2020–2021. Pakistan - the Benazir Income Support Program (BISP) for 2009–2019 and 2022–2023; the Ehsaas Emergency Cash program (EEC) for 2020-21, a COVID-19 response program that was previously known as BISP unconditional cash transfer. Philippines - Pantawid Pamilyang Pilipino Program (4Ps) for 2007–2019 and 2021; the Social Amelioration Program (SAP), a COVID-19 response, for 2020; Targeted Cash Transfer (TCT), an inflation response) for 2022. Sierra Leone - Ep Fet Po for 2014–2019; the COVID-19 Ep Fet Po, a COVID-19 response for 2020–2022; the Emergency Cash Transfer (ECT), a COVID-19 response, for 2020–2022. 63 64 State of Social Safety Nets 2025: Unfinished Mission to Reach the Poor and Beyond South Africa - Child Support Grant (CSG), created in 1998, for 2007–2024. South Africa also has a famous COVID-19 response called the Social Relief of Distress Grant (R350), which is still ongoing. The coverage of that program is not included here because the overlap between CSG and R2350 program wasn’t clear. Tunisia - Program me National d’Aide aux Familles Nécessiteuses or National Program of Assistance to Needy Families (PNAFN), created in 1985, for 2010-2024. The PNAFN was recently renamed as the AMEN Permanent Cash Transfer (PCT). The AMEN Temporary Cash Transfer (AMEN TCT), a COVID-19 response, for 2020-21, and the AMEN Family Allowance (AMEN FA) for 2022-24. Annex 4 Cash Transfers as an Enabler for Subsidy Reforms Recent crisis responses have highlighted the need to reform generalized subsidies, which are regressive, inefficient, and expensive. Subsidy measures are very expensive for governments and often do not achieve their objectives. Also, they can induce economic, environmental, and social distortions that yield unintended consequences. A recent study by Damania et al. (2023) showed that explicit and implicit subsidies on fossil fuels, agriculture, and fisheries exceed US$7 trillion, which is equivalent to around 8 percent of global GDP. Generalized subsidies tend to be regressive as they benefit richer families more than poorer families. A study by Arze del Granado, Coady, and Gillingham (2012) on the welfare impact of energy subsidy reform in 20 low- and middle-income countries across the globe found that the richest 20 percent of households capture six times more of the value of subsidies than the poorest 20 percent of households. Social protection can play a vital role in substituting for subsidies. Targeted social assistance (and cash transfers in particular) can help governments to manage the potential distributional and welfare effect of subsidy reforms and help build public trust in—and support for—speedy implementation of subsidy reform. Investing in social protection delivery systems can enable countries to expand their social protection coverage far more rapidly, effectively, and efficiently, including identifying and targeting the populations who are most in need. In recent decades, several countries have successfully reformed their energy subsidies (the Dominican Republic and Ukraine), food subsidies (Egypt and Indonesia), and agriculture subsidies (Guinea, Mali, and Niger). The Dominican Republic, for instance, undertook a series of energy subsidy reforms over nearly two decades and moved from a generalized price subsidy for liquefied petroleum gas and electricity to a system of targeted transfers covering nearly 40 percent of all households in the country. Also in Indonesia, the government switched its rice subsidy program modality in 2017 from the in-kind delivery of subsidized rice to eligible low-income families to electronic food vouchers paid directly into the beneficiaries’ bank accounts.38 It also removed subsidized prices, requiring beneficiaries to use their vouchers to purchase food items at the market price. 38. In 2022, delivery Indonesia’s BPNT/Sembako food assistance was switched from fully e-voucher distribution mechanism to a current mixed modality with partial cash distribution at post offices or bank account via electronic payment card. 65 ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, we leverage electronic publishing options and print-on-demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consumption, chemical use, greenhouse gas emissions, and waste. We follow the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC)–certified paper, with nearly all containing 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine–free (PCF), or enhanced elemental c ­ hlorine–free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/corporateresponsibility. Social Protection & Jobs Discussion Paper Series Titles 2025 No. Title 2510 State of Social Protection Report 2025: The 2-Billion-Person Challenge. Background Paper #3: Wake-Up Call for Social Assistance? An Unfinished Mission to Reach the Poor and Beyond 2509 State of Social Protection Report 2025: The 2-Billion-Person Challenge. Background Paper #2: Adaptive Social Protection Agenda: Lessons from Responses to COVID-19 Shock 2508 State of Social Protection Report 2025: The 2-Billion-Person Challenge. Background Paper #1: Mind the Gap: Coverage, Adequacy and Financing Gaps in Social Protection for the Extreme Poor and the Poorest Quintile 2507 Service Integration and Case Management for People on the Move: A Review of Selected International Practices 2506 Impact of Climate Change and the Green Transition on Human Capital: A Review of the Evidence from Europe and Central Asia 2505 A slippery slope: the opportunities and risks of digital approaches and technology in Social Protection Systems 2504 De Jure and De Facto Coverage of Parental Benefits in Nepal 2503 Awareness, Access, and Perceptions around Parental benefits among Urban Argentinians 2502 Regulating Markets So More People Find Better Jobs 2501 São Tomé and Príncipe Unpacking Migration Dynamics: Critical Issues and Policy Recommendations To view Social Protection & Jobs Discussion Papers published prior to 2021, please visit www.worldbank.org/sp. ABSTRACT This paper examines the current state of play and trends in social assistance in terms of spending, coverage, incidence, benefit size, and poverty reduction impacts using administrative data and household survey data from about 70 countries worldwide. This paper is the fourth edition of the State of Social Safety Nets Report, following the publications in 2018, 2015, and 2014. It also serves as a background paper for the first edition of the State of Social Protection Report 2025: The 2-Billion-Person Challenge. As a major driver and component of social protection, social assistance has made considerable progress in expanding its coverage over the past decade. However, the work is still unfinished, as is evident in the persistent gaps in coverage and financing, particularly in low-income countries and those affected by fragility, conflict, and violence, where the need for support is greater. As a result, the adequacy of benefits remains low, undermining the impacts of social assistance in reducing poverty. The paper highlights potential opportunities and priorities for further investment to address these challenges and progressively expand the effectiveness of social assistance which can better support not only the poor but also broader populations in the face of shocks. For example, these include leveraging new technologies, strengthening referrals to other programs, and reforming fiscal policies. These investments are relevant and critical to make adaptive and integrated social protection systems, complementing wider social protection pillars – including social insurance, and labor market and employment programs – and to ensure adequate social protection support in good times and bad. JEL CODES D63, H53, I38, J18, O15 KEYWORDS Social protection, adaptive social protection, social assistance, social safety nets, coverage, incidence, adequacy, spending, financing, poverty reduction, cash transfer, subsidy reform. ABOUT THIS SERIES Social Protection & Jobs Discussion Papers are published to communicate the results of The World Bank’s work to the development community with the least possible delay. This paper therefore has not been prepared in accordance with the procedures appropriate for formally edited texts. For more information, please visit us online at www.worldbank.org/socialprotection