Spatial Heterogeneity in Returns to Agricultural Extension and Its Implications for Extension Targeting Policy in Oyo State, South West Nigeria

Agricultural extension investment decisions are typically guided by average returns to programme participation, masking substantial spatial heterogeneity in returns that has important implications for efficient resource targeting. In Oyo State, where the agricultural extension system serves a diverse landscape of peri-urban, transition zone, and deep rural farming communities, understanding where extension investment generates the highest marginal returns is critical for cost-effective service delivery. This study investigated spatial heterogeneity in returns to agricultural extension and its implications for extension targeting policy in Oyo State using a spatial panel dataset of 460 farming households geo-referenced at the plot level across 48 communities. Geographically weighted regression and spatial Bayesian heterogeneous treatment effects models were applied to estimate community-level heterogeneous average treatment effects of extension contact on crop productivity. Results revealed substantial spatial variation in extension returns, ranging from near-zero effects in well-integrated peri-urban communities with high baseline technology access to returns of over 35 percent in remote communities with constrained information environments. Market access, agro-ecological conditions, and baseline technology adoption rates were the primary spatial determinants of heterogeneous returns. The study makes an original methodological contribution to extension targeting literature by demonstrating the welfare and efficiency gains achievable through spatial return-informed extension resource allocation in Oyo State, with specific recommendations for geospatially targeted extension investment strategies. Keywords: spatial heterogeneity, extension returns, targeting policy, Oyo State, geographically weighted regression.