This study investigates the causal impact of Nigeria’s Conditional Cash Transfer (CCT) programme on household welfare in the Federal Capital Territory (FCT), Abuja, focusing on Abaji, Kwali, and Kuje Area Councils. Using primary household survey data collected between 2021 and 2025, complemented by baseline information from 2020. The study used Double/Debiased Machine Learning (DML) as the primary statistical and causal estimation method to analyse the data. DML was adopted because the Conditional Cash Transfer (CCT) beneficiaries were not randomly selected, creating selection bias and high-dimensional confounding. The method combined machine learning algorithms (Random Forest, Gradient Boosting, and regularized regression/LASSO) with orthogonalization and cross-fitting to estimate the Average Treatment Effect (ATE) and Heterogeneous Treatment Effects (HTE). The study also conducted robustness checks using Inverse Probability Weighting (IPW) and Doubly Robust (DR) estimators for validation. The analysis reveals that CCT participation leads to a statistically significant increase in household welfare by approximately 0.29 standard deviations. The strongest impacts are observed in education and healthcare utilisation, while food security improves moderately and asset accumulation remains limited. Heterogeneous treatment effects indicate that female-headed households, larger households, and those receiving regular payments benefit disproportionately. The study contributes to the literature by integrating causal machine learning techniques with primary data in a subnational African context and provides policy-relevant insights on programme design, targeting, and implementation efficiency. The findings suggest that while CCTs are effective in protecting human capital, their long-term poverty reduction potential depends critically on complementary livelihood interventions and institutional effectiveness.
Ohiro Leo Atakpu (2026) studied this question.