Disentangling Selection and Causality in Assessing the Effects of Health Inputs on Child Survival: Evidence from Demographic and Health Surveys in 14 African Countries

Gebrenegus Ghilagaber, Stockholm University

Public policy response to the problem of high childhood mortality in developing countries has primarily focused on encouraging prenatal care and institutional delivery. Since there are no randomized trials of standard prenatal care and hospital delivery, it is difficult to assess the impact of such health inputs on survival chances without accounting for selection processes in the utilization of health facilities. This paper examines the biasing effects of selection on the efficacy of health inputs. Analyses of Demographic and Health Survey (DHS) data from 14 African countries shows a clear selection process in the use health facilities. The extent and direction of the selection process is, however, different across countries. In some of them, failure to take due account for adverse selection process leads to underestimates the beneficial effect of health inputs. In others, such effect is overestimated due to a favorable selection process.