Towards Data-Driven Governance in Yemen: A Statistical Framework for Addressing Data Gaps and Evaluating Public Service Effectiveness in Post-Conflict Phases

This working paper proposes a statistical framework for data-driven governance in post-conflict Yemen. The study addresses the critical challenge of incomplete and missing public service data, which hinders evidence-based policymaking during institutional rebuilding phases. Using advanced imputation techniques for handling missing values in matrices — including Multiple Imputation by Chained Equations (MICE) and K-Nearest Neighbors (KNN) methods — the paper develops a practical methodology that Yemeni ministries and international donors can adopt to evaluate public service effectiveness despite data gaps. A simulated case application on education enrollment rates across Yemeni governorates demonstrates the framework's reliability in producing statistically sound estimates from incomplete datasets. The paper concludes with seven actionable policy recommendations focused on building national statistical capacity, establishing open data infrastructure, and training local data analysts. This research contributes to the emerging literature on post-conflict state-building by bridging data science methodologies with public governance needs in fragile contexts.