We examine the spatial determinants of the prevalence of poverty for small spatially defined populations in rural Malawi. Poverty prevalence was estimated using a small-area poverty estimation technique. A theoretical approach based on the risk chain conceptualization of household economic vulnerability guided our selection of a set of potential risk and coping strategies — the determinants of our model — that could be represented spatially. These were used in two analyses to develop global and local models, respectively. In our global model—a spatial error model — only eight of the more than two dozen determinants selected for analysis proved significant. In contrast, all of the determinants considered were significant in at least some of the local models of poverty prevalence. The local models were developed using geographically weighted regression. Moreover, these models provided strong evidence of the spatial nonstationarity of the relationship between poverty and its determinants. That is, in determining the level of poverty in rural communities, where one is located in Malawi matters. This result for poverty reduction efforts in rural Malawi implies that such efforts should be designed for and targeted at the district and subdistrict levels. A national, relatively inflexible approach to poverty reduction is unlikely to enjoy broad success.