Data Wealth Trapped in Public Sector Silos

Data Wealth Trapped in Public Sector Silos

Apr 25 , 2026. By Birhanu Beshah (PhD) ( Birhanu Beshah (PhD), ( birhanu.beshah@aait.edu.et) is an associate professor in the School of Mechanical & Industrial Engineering of Addis Abeba University (AAU). )

Ethiopia’s data architecture is currently shaped by institutional mandates rather than a unified national strategy. Unlike land or minerals, data is a non-rivalrous good. A use by one agency does not reduce its availability to another. However, its reusability should be intentionally designed. Without common standards and secure protocols for exchange, data remains an underused byproduct of administration.

Data is everywhere in the public sector, yet nowhere near as useful as it should be. From biometric registrations to tax filings, from school records to utility payments, public institutions are gathering information. Officials and policymakers increasingly repeat a familiar slogan that “data is the new oil.” But this assertion obscures a harder, unsettled question.

When, exactly, does data become a resource?

The answer lies less in how much data is collected than in how it is used. Data becomes a genuine economic and governance resource only when it moves beyond its first purpose, when a dataset can serve several functions across systems, institutions and decisions. Today, much of Ethiopia's data ecosystem remains stuck in a linear model. It is collected, used once for a specific administrative task, and then left in a silo.

For instance, the pattern is clear in biometric data collection. Several institutions, including the national ID initiative, Fayda, the immigration office, civil registration systems, the Ministry of Labour & Skills, and city administrations such as Addis Abeba, have independently collected overlapping datasets. Each system is built around its own mandate, with little interoperability. The result is redundancy, inefficiency and lost opportunity. Instead of becoming a reusable national asset, data is divided into institutional islands.

The problem reaches beyond biometrics when we examine the education systems, which hold records on students. Health institutions maintain patient information, while tax authorities collect financial and compliance data, and utilities monitor consumption patterns. Each dataset has value on its own. But its greater value lies in connection, not in being locked up in a silo. Linking education and labour data could guide workforce planning, and combining utility and tax data could sharpen economic forecasting. Integrating health and demographic data could strengthen public health interventions.

Yet such use across sectors remains the exception, not the rule, due in large part to a failure of design as much as technology.

Ethiopia’s data architecture is still shaped largely by institutional mandates rather than national strategy. Each agency builds around its own needs, often starting from scratch, repeating data collection efforts, and paying too little attention to interoperability standards. In doing so, the country bears not only direct financial costs, but also opportunity costs, the lost value of data that could have been used again.

In theory, data has the qualities of a non-rivalrous good. Use by one institution does not reduce its availability to another. That is what separates it from traditional resources such as land or minerals. But unlike natural resources, data does not become useful through extraction alone. Its reusability has to be designed. Without rules for sharing, standards for compatibility, and systems for governance, data remains underused.

Moving from data as a byproduct to data as a resource requires a shift in policy and mindset. Interoperability has to become a national priority, with common data standards, unique identifiers and secure protocols for data exchange. If used well, the Fayda ID system could serve as a layer that links different datasets without repeated collection.

Data governance should also balance access with privacy and security. The rise of data protection and privacy regulations offers a solid base for responsible use. These rules should not be viewed as obstacles but treated as enablers, helping build trust among citizens and institutions and creating the conditions for safe data sharing. Trust is the currency of any functioning data ecosystem.

Nor does Ethiopia need to wait for artificial intelligence (AI) before treating data as a resource. Even analytical tools can create value. Basic matching, trend analysis and cross-referencing can produce insights that improve service delivery and policy choices. Identifying overlaps between social protection beneficiaries and tax records could improve targeting efficiency. Far more than futuristic ideas, these are practical steps within immediate reach.

However, the path ahead cannot be without its limitations. Institutional inertia, disputes over data ownership, and limited technical capacity can all slow progress. There is also a risk of over-centralisation, in which data integration concentrates power without enough safeguards. A federated data architecture may offer a better balance, allowing institutions to retain control over their own data while following shared standards.

Other options deserve attention, too. Public-private partnerships could help build data infrastructure and strengthen analytics capacity. Incentives for data sharing, whether through performance metrics or budget allocations, could push institutions beyond siloed operations. Building skills, especially in data engineering and governance, remains equally important.

For Ethiopia to use data as a national resource, it has to move from accumulation to utilisation, from fragmentation to integration, and from isolated systems to a connected data ecosystem. As digital transformation shapes economic competitiveness and the quality of governance, the value of data lies not in its collection but in its connections.

PUBLISHED ON Apr 25,2026 [ VOL 27 , NO 1356]