In early April, the World Bank published its State of Social Protection report (2025), earlier known as the ‘State of Social Safety Nets’ series. These reports use global-level data about social protection programs to reflect on trends in social security and to ascertain future requirements of social security.
What is intriguing about this report was the advocacy of a ‘context-specific’ approach to social protection policy. Policymakers are prompted to think about unique contextual factors that may or may not contribute to success in achieving favourable outcomes. This represents a departure from how multilateral organisations usually prescribe policy for the industrialising world and so, merits further enquiry.
Key findings from the report
The report relies on the Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) database as its main source of data. The database consists of indicators measuring the scope and performance of social protection programs, derived from program-level administrative data and national household survey data from 146 countries.[1]
The report comments on global gaps in the coverage, adequacy, and financing of social protection in those countries. Its major findings are as follows:
Coverage[2]
- Approximately 50% of the world’s population has some form of social protection, but coverage varies widely by region and income level. Social protection benefits reach only about a quarter of people in LICs[3].
- Social assistance[4] schemes are the most prevalent social protection instruments, and the increase in social protection coverage over the past few years is almost entirely attributable to the expansion of such schemes in LICs and LMICs.
- The most common social assistance programs in terms of coverage are cash transfers and school feeding programs
- 4 women receive a social assistance transfer for every male beneficiary
- Social insurance[5] schemes are far less prevalent in LICs and LMICs than in UMICs and HICs. The presence of social insurance schemes is directly related to the proportion of labour force in formal employment, and hence, it is in Europe and Central Asia that most such programs are concentrated.
- Coverage of labour market assistance[6] programs is considerably low across countries at all income levels. Less than 5% of the global population is covered by such programs.
Adequacy[7]
- Social protection benefits account for 16 percent of household welfare in LICs and 37 percent in HICs[8].
- The poverty alleviating impact of social assistance schemes is much stronger in HICs and almost non-existent in LICs. Social assistance transfers bring 82 percent of poor households up to the poverty line in HICs. In comparison, these schemes only bring 1 percent and 11 percent of poor households out of poverty in LICs and LMICs respectively.
Financing[9]
- On average, countries spend about 5.3 percent of their GDP on social protection programs.
- HICs spend about 85 times more on social protection per capita than LICs.
- In most countries, social insurance schemes account for the largest share of social protection expenditure. So, differences in social protection spending across countries are determined by the factors which influence social insurance payments (such as size of the formal sector, old-age population, and design of the contributory system).
- In LICs which are also characterised by conflict and violence (primarily in Sub Saharan Africa and the Middle East and North Africa), social protection is funded mostly through international grants.
The report concludes with the need to expand both coverage and adequacy of social protection, especially in LICs and LMICs through the following pathways:
- Improve targeting protocols to ensure that programs are focused on the poorest quintiles of the population.
- Undertake large-scale subsidy reform to create the fiscal space needed to expand social protection coverage and adequacy. That is, re-think subsidies (especially in LICs and LMICs) for fossil fuels, agriculture, and fisheries which may be regressive.
- Strengthen social protection delivery systems by leveraging real-time data, referral protocols, and governance structures using technology.
- Leverage social protection (labour market programs, economic inclusion initiatives) to improve employment opportunities and entrepreneurial activities.
What does a context-based approach really mean?
The report encourages policymakers to consider their respective countries’ specific needs, implementation capacity, and fiscal space while deliberating on design choices and financing options in social protection policy. This endorsement of context-based policy design is much welcome in view of typical templatised approaches to social policy development which uniformly apply policies to widely varying contexts.
The report acknowledges how standardised policy applications can fail due to contextual factors. For instance, it quotes how:
- Labour market programs often underperform because of a lack of context-specific design to meet the needs of vulnerable workers in the unorganised sector
- Contributory social insurance programs have failed to adapt to the realities of life for informal workers with irregular incomes and no formal contracts
- Subsidy reform may fail because of the macroeconomic context in which it is implemented
The learning here is that policies grounded in a) specific needs of the beneficiary group and b) political-economic context that is a co-determinant of those needs, will be more likely to have favourable impacts. Intuitively, it makes sense that countries with starkly different socio-political characteristics, fiscal spaces, and cultural values might need to take different approaches to reach the same goal of universal social protection.
Having made the case for context, however, the report devolves to only a singular dimension of context, namely per capita income level in the country. With regard to this particular aspect of context, the recommendations are as follows:
- LICs are advised to scale up foundational non-contributory programs like cash transfers or food aid for the poor, to improve their shock responsiveness. The recommendation is in view of the highest coverage gaps occurring in LICs.
- Since LMICs have moderate coverage, but usually provide inadequate benefits, they are recommended to expand social insurance and increase benefit levels for social assistance. It is also suggested that they begin developing labour market systems to encourage employment (such as self-employment support, public works programs, wage subsidies, etc.).
- UMICs have low-to-moderate coverage gaps but face challenges in adapting to emerging risks such as rapidly ageing populations which are leading to labour shortages and large impending pension payouts. Accordingly, they should improve social insurance systems by developing elder care services and actuarially fair contributory pension programs.
While distinct recommendations on the basis of per capita income is one way to account for context, we should remember that a country’s per capita income is itself a function of multiple variables, such as demographic characteristics, political institutions, the nature of economic system and configuration of economic policy, industrial composition, cultural values, etc. Perhaps these underlying, more primitive variables, should enter the space of contextual considerations, in place of per capita income, as these variables make for significant differences between countries. Srinivas (2020) [10] describes these differences as the variation in the norms and rules that govern the institutions which bind actors, activities, and objects together in economic activity, and she calls this variation ‘institutional variety’. Development policy should be designed to confront and internalise such variety, instead of endorsing ‘one-size-fits-all’ approaches (Srinivas, 2020) – even for countries belonging to the same income grouping.
The primary arena in which social protection policies play out is that of social systems, which are complex, multivariate, constantly adapting and necessarily uncertain (Ang, 2025)[11]. Emerging alternative models of development are appreciative of the complexity of social systems and treat this institutional variety as endogenous components of the development process. This implies that countries’ social protection policies should account for the evolutionary history of institutions, or as Ang (2025) simply puts it, countries should ‘use what they have, not what they want’.
The institutions that govern and administer social protection policies in LICs, and the norms they embody, are likely to function fundamentally differently than their HIC counterparts. The report seems not to account for this. If it did, it may appreciate that in non-HIC countries, policy instruments to ensure that citizens are able to withstand and respond to shocks may look quite different from those seen to commonly work in HIC contexts. For instance, the report prescribes that LICs strengthen their social assistance offerings to mitigate exposure to risks but does not reconcile the fact that the poverty reduction effects of social assistance programs are much stronger in HICs but nearly non-existent in LICs. Why then saddle LICs with instruments that either a) don’t work in their contexts or b) are challenging to fiscally accommodate? An appreciation of institutional variety and the different pathways of institutional evolution may lead to the conceptualisation of social protection policies that build on the institutions that work well in LICs. For instance, the household finance literature demonstrates how, in many developing countries, households rely on localised and ‘informal’, community-based solutions to bolster their shock-responsiveness (Morduch, 1999; Mobarak and Rosenzweig, 2013). [12], [13],Could formal social protection policy systems learn from these social realities and be designed to mimic them in some manner? This is a tentative example, but it points to the necessity of a proper investigation into the institutional variety that supports the design and delivery of social protection policies across countries. Such an investigation could be very revealing and encourage the country to think beyond ‘traditional’ (that is, HIC) policies.
In view of the above discussion, we may formulate the following prompts to help a policymaker reflect upon the context-specificity of social protection policies in their country:
- What institutions and norms are responsible for the design and delivery of social protection policies in my country? What is the historical trajectory of such institutions? What do these institutions do well, and what do they struggle with? What factors enable/hinder good performance? What can I learn from studying other country cases with other kinds of institutional variety?
- How is the administration of social protection managed ‘vertically’, i.e., how are responsibilities devolved from the Centre downwards? Is there variation in patterns of decentralisation across countries and what can that teach me about the preferred mode of decentralization in my country ?
- What are the most pressing risks at the local level in my country? What are the local, ‘informal’ solutions that emerge in response to these risks? What are the enabling factors for their success, and can social protection policy formulated at the Centre learn anything from such solutions?
- How do the people of a given region in my country relate to social protection policy? What role does it serve in their economic and cultural lives?
The World Bank acknowledging the need for context-specific social protection policy is a promising development towards enabling social security for all who need it. However, expanding the scope of what this ‘context’ refers to can put us on the track towards policy solutions which are truly tailored to local requirements and are more likely to be successful.
Footnotes:
[1] ASPIRE: The Atlas of Social Protection Indicators of Resilience and Equity
[2] The report measures coverage as all people living in a household in which at least one member receives a social protection benefit. It may be noted that typically, coverage is measured as the number of people who directly receive a benefit (or contribute to social insurance). So, the report’s measurements of coverage are likely to be overestimations.
[3] The report categorises nations as belonging to one of the following income groupings: low-income countries (LICs), lower-middle income countries (LMICs), upper-middle (UMICs), and high (HICs). While the report does not provide explicit definitions the categories are presumed to align with the World Bank’s income groupings as specified here.
[4] Social assistance schemes, also known as social safety nets, are non-contributory (that is, tax- or donor-financed) and offer cash or in-kind benefits and services to poor or vulnerable recipients.
[5] Social insurance includes contributory pensions and benefits (usually related to formal employment).
[6] Labor market assistance programs include economic inclusion programs (which provide cash transfers, skills training, business capital, coaching, and market access), active labour market programs (which actively keep workers employed/bring workers into the labour force), and unemployment insurance.
[7] Adequacy is measured as the level of transfers received by a population group divided by the total income or consumption level (presumably based on the availability of data, though the report does not explicitly say this) of the beneficiaries within the group. It may also be noted that the report does not specify any threshold for adequacy, but describes the adequacy of social protection (at a global average of 27% of household income) as being ‘low’.
[8] Since “welfare” refers to income for some countries and consumption for others the authors appear to have arrived at the number for each group by averaging across countries within that group.
[9] Refers to countries’ investments/public expenditures on social protection.
[10] Institutional variety and the future of economics | Review of Evolutionary Political Economy
[11] muse.jhu.edu/pub/1/article/954433/pdf
[12] Between the State and the Market: Can Informal Insurance Patch the Safety Net? | The World Bank Research Observer | Oxford Academic
[13] Informal Risk Sharing, Index Insurance, and Risk Taking in Developing Countries – American Economic Association;
One Response
Thought provoking piece, Aishwarya. Enjoyed reading it. I think your criticism of accounting for institutional variety and articulating how this shapes the social protection systems is valid. I think the per capita income is a good lens to classify countries, because per capita income is often correlated strongly with other institutional and cultural factors that you talk about. Recently, I learnt that there is a strong positive correlation between countries’ per capita income and its cultural values. Countries with high per capita incomes align with high levels of secular-rational and self-expression values, whereas countries with low per capita income align more towards traditional and survival values. Similarly, countries with high per capita income by large might have more effective institutions and governance. So I think per capita income is a good proxy for several things that are non-monetary in nature. But yes, agree that an articulation of what such a lens misses is equally crucial.