A solid foundation for adaptive social protection in Liberia

Social registries, defined as information systems that support processes of enrollment, registration, and eligibility assessment for social-protection programs, are being developed and expanded across Sub-Saharan Africa.

In an environment where information is scarce and informal, collecting and centralizing data on households or individuals offers tremendous programmatic potential. Social registries streamline referral processes and program implementation, placing countries on a more effective path toward poverty reduction.

For example, during the COVID-19 pandemic, countries that already had social registries responded more quickly and reached more people than those without them.

In Liberia—where household-level surveys are rare—establishing a social registry through door-to-door data collection provides a unique opportunity to inform government agencies, donors, and development partners interested in social-protection program management, design, planning, and service delivery across multiple sectors.

However, using a door-to-door data-collection approach is no easy task, as it often requires covering large geographic areas with challenging terrain—whether densely populated, like Lagos, or sparsely populated, like the Sahel.

One example is development of the Liberia Household Social Registry (LHSR) under the Liberia Social Safety Nets Project (LSSNP) from 2016 to 2023. The Ministry of Gender, Children and Social Protection led the project with World Bank support.

More than 250,000 households—about 20 percent of the population—are now enrolled in the LHSR. Although this meets the LSSNP’s target, major hurdles had to be overcome.

Initial delays in Liberia

To correct early mistakes, the World Bank and the LSSNP Project Implementation Unit decided to revisit households that were inadvertently skipped. Further analysis, however, revealed misclassification errors linked to building codes, complicating efforts to revisit those structures.

Early delays stemmed from political transitions and difficulty finding a suitable data-collection firm. The onset of COVID-19 further disrupted the project just as data collection was about to start. Fieldwork began three years after project approval and concluded in April 2021 with data on 200,000 households.

Those data, however, showed significant quality issues, raising concerns about their reliability and usefulness. To address the problem, the team developed an ad-hoc deep-learning model to identify misclassified households. The model—“Automatic Correction of Inaccurate Household-Structure Data for Liberia”—detects errors by clustering nearby households that, based on GPS coordinates and photographs, are most likely in the same building.

Processes

Although the cleaning process took several months, it succeeded in identifying and reassessing the most problematic structures, resulting in a far cleaner and more usable social registry.

The following lessons from this and other project experiences may help anyone seeking to establish a national social registry:

1. Put quality and standards first: look for red flags and assume nothing. Most challenges encountered were related to data quality and productivity—common issues in data collection. Yet the volume of data and its intended use require detailed protocols and weekly (or more frequent) reviews of each data package to spot defects and improve quality.
2. Treat the project as a platform for practical, data-driven innovation. The LSSNP became a vehicle and opportunity to foster new developments in Liberia’s social-protection sphere—such as the machine-learning application mentioned above, field-work planning with public source building footprints (OpenBuildings), and digitalizing LHSR ID cards to identify resource-constrained families and structures.
3. Seek early-adoption opportunities to ensure sustainability. In January 2020, Liberia’s Cabinet approved the LHSR as the primary tool for managing social-protection programs nationwide. The LSSNP also developed a Management Information System (MIS) that supports user enrollment and electronic payments, among other functions.

The LHSR and MIS are already used by three government-led social-protection programs, and more are expected as systems continue to be strengthened under the REALIZE project.

Information source: blogs.worldbank.org

\