Governments in less developed countries face severe budgetary constraints. Given that public sector employment is a large part of modern sector employment, the government wage bill h as come under increased scrutiny. The central question is how do gove rnment wages compare with those in the private sector? In this paper, the authors develop and estimate a model to answer this question. An important aspect of this model is the endogenous treatment of sector choice. The estimation results (full information maximum likelihood) sound a strong warning against the use of ordinary least squares est imates that are based on sector-specific samples. Data are from the I vory Coast. Copyright 1988 by MIT Press.