Rice-area specialization and young women’s NEET rates in African countries: dynamic persistence, institutional heterogeneity, and import dependence

Abstract This paper examines the conditional dynamic associations between rice-area specialization and the rate of young women who are not in employment, education, or training (NEET) in African countries. The analysis uses an annual unbalanced panel of 41 countries over 2009–2023, comprising 583 country-year observations in the main sample, and two-step System Generalized Method of Moments (System GMM) estimates based on collapsed internal instruments and finite-sample-corrected standard errors. The female youth NEET rate is highly persistent. In the preferred institutional specification, the alternative specification, and the male-NEET sensitivity model, rice-area intensity is negatively associated with the female youth NEET rate, while rice import dependence also has a negative coefficient. These estimates lose precision or change sign when deeper lags and stricter instrument reduction are imposed. Institutional interactions are heterogeneous, and the exploratory long-run transformations reported in Supplementary File 1 have wide confidence intervals. The reported models do not reject the null hypotheses of the Arellano–Bond AR(2) and robust Hansen tests, and the number of instruments remains below the number of countries. However, identification relies on internal instruments, the panel includes only 41 country groups, the persistence coefficient is close to or above unity in several specifications, and the Sargan test rejects in the two most demanding stress tests. The results are therefore interpreted as model-dependent conditional associations, not as evidence that greater rice specialization or import dependence causally reduces the female youth NEET rate. Their policy relevance lies in evaluating rice-sector transformation with directly gender-disaggregated indicators of employment, education, training, and access to opportunities in downstream value-chain segments, rather than inferring inclusion from production growth alone.