This paper aims to investigate whether the promotion of high-value agriculture (HVA) in China contributes to rural income inequality and to identify the mechanisms through which such impacts may arise. The study focuses on the national advantageous agricultural product zones (NAAPZ) policy as a key policy intervention to promote HVA.
The study uses the staggered implementation of the NAAPZ policy as a quasi-natural experiment. It draws on four waves of micro-level panel data from the China Family Panel Studies covering 2016 to 2022, comprising over 29,000 observations. A double machine learning framework is employed to control for high-dimensional household and regional characteristics flexibly, enhancing the robustness of causal inference.
The findings show that NAAPZ policy significantly increases rural household income inequality, with the household-level Kakwani index rising by an average of 0.033. Mechanism analyses confirm an elite capture effect, as households with higher education and stronger digital capacity disproportionately benefit from entrepreneurial and non-farm opportunities. Spatial disparities also emerge: households in township centers or county seats gain more than those in peripheral villages. Further, sectoral heterogeneity is evident – crop-oriented programs are strongly associated with rising inequality, whereas livestock initiatives show no significant effect. Finally, governance differences matter: city-level initiatives exert more substantial inequality impacts than county-level ones.
These findings highlight a policy paradox: although NAAPZ is designed to foster inclusive rural growth, it may unintentionally amplify disparities. Policy measures should therefore include targeted training for disadvantaged households, microcredit access for resource-poor farmers, infrastructure investment in peripheral regions and safeguards against elite capture to ensure more equitable outcomes.
This paper provides the micro-level causal evidence on the distributional impacts of HVA in China, combining a quasi-experimental design with advanced machine learning methods. It reveals individual-level and spatial mechanisms – specifically elite capture and spatial differences – through which agricultural modernization policies may influence income inequality. The insights offer valuable guidance for other developing economies seeking to balance economic upgrading with social equity.
