Attachments
Abhijit V. Banerjee
Effective implementation is the missing link in the drive toward inclusive, reliable social protection. In recent decades, social protection systems have expanded rapidly across the developing world. Today, more than 2.5 billion people benefit from programmes like cash transfers, food subsidies, and social pensions. Yet despite this progress, the goal of social protection for all remains out of reach. Too many eligible people still slip through the cracks—excluded by complex procedures, poor targeting, or weak delivery systems.
Drawing on evidence from large-scale randomized evaluations, including the work of the Abdul Latif Jameel Poverty Action Lab (JPAL), I argue that simple improvements in how programmes are delivered can dramatically increase their reach and impact. But implementation doesn’t happen in a vacuumThe brief also highlights challenges in identifying who needs support, trade-offs in programme design (e.g., conditional vs. unconditional transfers), and the growing need to expand beyond poverty relief into social insurance. By focusing on practical, scalable solutions—especially those grounded in rigorous evidence—policymakers can move from social protection in principle to social protection in practice.
Reaching the right people
One of the most pressing challenges in Global South countries is accurately identifying who needs help. Unlike high-income settings, where governments can use tax and employment data to determine eligibility, most people in low- and middle-income countries work in the informal sector. Relying on self-reported income can lead to similarly flawed data. This makes income difficult to observe and undermines the reliability of standard targeting tools.
In response, governments use alternative approaches such as proxy means tests (PMTs), which estimate poverty based on visible household characteristics like assets or housing conditions. While widely used, PMTs are often outdated or imprecise (for instance, someone with a large home may have a low income).
Geographic targeting is another option, where support is directed to poorer regions, though this method risks missing poor individuals who live in better-off areas. Other strategies include self-selection mechanisms and community-based targeting, which leverage local knowledge but can be vulnerable to bias or manipulation. All of these methods come with trade-offs—between inclusion errors (helping those who don’t need support) and exclusion errors (missing those who do). Evidence from Indonesia shows that no single method is best across all settings. Combining approaches, and tailoring them to local conditions, often yields better results.
Implementation and governance challenges
Even the most accurate targeting, however, will fail if programmes are poorly delivered. Across many countries, beneficiaries face delays, bureaucratic hurdles, and outright leakages. In some cases, eligible households are discouraged from applying after repeated failed attempts. These problems are often symptoms of deeper governance challenges: limited administrative capacity, outdated infrastructure, and weak accountability systems. For social protection to work, delivery systems must be transparent, responsive, and resilient to interference or misuse.
Solutions that work
Encouragingly, recent evidence from randomized controlled trials (RCTs) shows that specific implementation reforms can make a big difference—and many are both affordable and scalable. First, technology can streamline delivery. Digital tools like biometric ID systems and mobile payments enhance transparency and reduce leakage. In India, smart card-based payments for pensions and public employment schemes cut payment delays and leakage nearly in half. In Indonesia, shifting from physical rice distribution to electronic food vouchers improved both targeting accuracy and benefit delivery.
Second, providing clear and timely information can dramatically boost programme take-up. A simple intervention in Indonesia—mailing ID cards to eligible households—increased access to food subsidies by 21 percentage points. Such low-cost informational ‘nudges’ empower households to claim their rights and reduce exclusion.
Third, simplifying application procedures is key to increasing participation, particularly among the poorest. Complex requirements—multiple documents, forms, or travel to government offices—can act as serious deterrents. Evidence from Chile, India, and Indonesia shows that while offering application support helps, reducing the complexity of the process itself (for example, through automatic enrolment or pre-filled forms) is more effective.
Importantly, these delivery innovations are not confined to small pilot programmes. In India, the biometric payment reform reached millions of people. In Indonesia, the government scaled up e-voucher systems and informational interventions nationwide. With political support and rigorous testing, these reforms can be institutionalized and replicated, even in low-capacity settings.
Designing smarter programmes
While improving delivery is essential, programme design also matters. Governments must decide whether to condition transfers on certain behaviours—such as school attendance or clinic visits—or to distribute them unconditionally. Conditional Cash Transfers (CCTs) can support long-term human capital development, as shown in Nicaragua, but they risk excluding the most vulnerable if conditions are difficult to meet. Unconditional Cash Transfers (UCTs) are simpler and more inclusive, with demonstrated benefits for consumption and household well-being. Some programmes, like labeled cash transfers, have shown that simply indicating the intended use of funds can shape behaviour, offering a low-cost alternative to formal conditionality. The optimal design depends on local goals and administrative capacity; in contexts with weaker systems, simpler programmes may be more effective and equitable.
Social insurance
Finally, the social protection agenda must expand beyond poverty relief to include social insurance. This includes support for people facing job loss, health shocks, disability, and other life risks. Yet in most Global South countries, the majority of workers—especially those in the informal economy—have no access to formal insurance systems. As informal work grows, traditional contributory models become harder to sustain. Innovations like subsidized insurance schemes, mobile-based enrolment, and flexible contribution models offer new ways to reach excluded populations. Integrating social insurance with existing cash transfer programmes can build a more complete safety net and increase resilience to future shocks, including those driven by climate change and pandemics.
Conclusion
Expanded social protection is one of the great policy achievements of the past two decades. But to realize its full potential, governments must focus not just on who is covered, but on how programmes are delivered and designed. By investing in better implementation—rooted in evidence, technology, and simplicity—countries can build systems that are not only broader, but fairer, faster, and more inclusive. Social protection for all is possible—but only if systems work for those who need them most.