Reading Sisbén IV: what the score doesn’t tell you

Colombia’s social targeting system is one of the most sophisticated in Latin America. It is also, for a specific population, systematically blind.

What is Sisbén IV?

Sisbén IV (Sistema de Identificación de Potenciales Beneficiarios de Programas Sociales) is Colombia’s main instrument for targeting social programs. It assigns every registered household a composite score based on multiple variables — housing conditions, asset ownership, educational attainment, employment status, and more. Households below the C09 threshold are eligible for a range of social programs, from subsidized health insurance to cash transfers.

The system is technically impressive. It processes millions of records, weights dozens of variables, and produces a single number that is meant to summarize a household’s socioeconomic position. It represents a significant advance over earlier, cruder targeting methods.

What does the score miss?

The score is built to identify who is poor. It is not built to identify who has the markers of non-poverty but is experiencing hidden precarity. These are structurally different problems, and the same instrument cannot solve both.

Consider a household in Teusaquillo — stratum 4, with a mortgaged apartment, a child in private school, and one formally employed adult. Sisbén IV sees: housing (asset), education (asset), formal employment (asset). Score: above C09. Classification: non-poor. Reality: 65% of income goes to mortgage payments. Private school fees are financed by a credit card. The second adult works informally in underemployment. The household is one medical emergency away from collapse.

This household will never appear in Sisbén IV as a potential beneficiary. Not because it is not poor — but because its poverty does not look like poverty.

The ILSIS compound problem

The invisibility is compounded by ILSIS — the Índice de Localización de Servicios de Integración Social, published by Bogotá’s Secretaría Distrital de Integración Social. ILSIS aggregates social need indicators at the UPZ (Unidad de Planeamiento Zonal) level to guide territorial allocation of social services.

The problem: UPZ-level aggregation washes out within-zone variation. A UPZ like Teusaquillo may contain some blocks with concentrated hidden poverty, but when averaged across the entire zone, the signal disappears. ILSIS sees a non-poor zone and allocates resources accordingly. The blocks with hidden poverty receive nothing.

Our research maps this precisely: using Sisbén IV microdata at block level and cross-referencing with ILSIS classifications, we can identify where the two instruments — each individually defensible — combine to produce systematic blind spots. The result is 193,406 households invisible at UPZ scale in Bogotá alone.

These are not households the system has failed to reach. They are households the system was not designed to see.