Small area estimation of poverty indicators in Rwanda sectors: A case study in KARONGI district

Abstract

This research reflects on the small area estimation (SAE) with the principal objective of presenting the status of the poverty and extreme poverty at sector level. To accomplish this objective, we first present the theory of the small area estimation (SAE) technique. The SAE is concerned on the generation of believable estimations of characteristic of interest for small demesnes, starting on small or non samples coming from these demesnes; and the assessment of the estimate or error preduction. To improve direct estimations for a small demesne, SAE technique tries to ”borrow strength” (covariates) from other related data sets, either from similar areas, or relevant/auxiliary information obtained from a recent census or some other administrative records. The covariates used that are related to poverty indicators were the rates of household with no electricity, production, roof with local tiles, roof with other materials, and unimproved sanitation facilities. These five covariates were chosen as they were found to be of great impact on poverty status as published in the fourth integrated household living conditions (EICV4) report 2013-14 and in Poverty mapping report 2013-14. This study showed that, there is inequality in sectors of Karongi District. Bwishyura and Rubengera sectors were less suffering from poverty, with poverty incidence of 43.047% and 49.9902% respectively and with the extreme poverty incidence 27.7159% and 32.3044% respectively. Mutuntu sector was the most suffering from the poverty in Karongi district with the poverty incidence of 73.2498% and the extreme poverty of 49.2456% followed by Rwankuba sector with the poverty incidence of 70.2774% and the extreme poverty incidence of 46.6706%.