Statistical Analysis of Heterogeneous Treatment Effects in Nigeria’s National Social Investment Programme Using Quantile Regression

Social intervention programme evaluations in Nigeria typically report average treatment effects that mask substantial heterogeneity across the welfare distribution, and quantile regression methods reveal distributional treatment effect heterogeneity that mean-based approaches cannot detect. This study analyses heterogeneous treatment effects of Nigeria's National Social Investment Programme (N-SIP), specifically the Conditional Cash Transfer (CCT) component, using quantile regression applied to a household survey of 1,600 beneficiary and matched non-beneficiary households in the North Central and North East zones. Consumption expenditure, child school enrolment, and healthcare utilisation were the primary outcomes. Propensity score matching created comparable treatment and control groups. Quantile regression estimated CCT effects at the 10th, 25th, 50th, 75th, and 90th consumption percentiles. Average treatment effects from OLS showed N2,800 per month consumption increase (p < 0.001). Quantile regression revealed pronounced distributional heterogeneity: consumption effects were largest at the 10th percentile (N4,200, p < 0.001) and smallest at the 90th percentile (N1,100, p = 0.08), confirming progressive programme targeting effectiveness. Child school enrolment treatment effects were uniformly significant across quantiles. Healthcare utilisation effects were concentrated in the lower three quartiles. The study demonstrates that N-SIP CCT effects are strongest for the poorest households and recommends quantile regression as the standard evaluation method for Nigerian social protection programmes to document distributional impact. Keywords: quantile regression, heterogeneous treatment effects, social investment programme, Nigeria, conditional cash transfer