How important are neighbourhood endowments of physical and human capital in explaining diverging
fortunes over time for otherwise identical households in a developing rural economy? To answer this question
we develop an estimable micro model of consumption growth allowing for constraints on factor mobility
and externalities, whereby geographic capital can influence the productivity of a household’s own capital.
Our statistical test has considerable power in detecting geographic effects given that we control for latent
heterogeneity in measured consumption growth rates at the micro level. We find robust evidence of geographic
poverty traps in farm-household panel data from post-reform rural China. Our results strengthen the equity
and efficiency case for public investment in lagging poor areas in this setting.