The Long Term Effects of Rural Postings on Health Workers' Careers in Ethiopia: Results of a Natural Experiment

Abstract

International attention has recently been drawn to the problems of attracting, retaining, and motivating health workers in developing countries, particularly to more remote areas - the so-call human resources in health (HRH) crisis. Ethiopia, with less than one doctor per 100,000 citizens, and virtually no physicians outside urban areas, provides an acute manifestation of this crisis. Rural postings are viewed negatively for both "flow" and "stock" reasons: current living standards may be reduced if the quality of infrastructure, children's education, and general consumption possibilities is low in rural areas; but the medium- to long-term impact of a rural posting on a health worker's future career may represent an important additional cost. Current costs can be offset by higher wages, better housing conditions, etc. However, estimating the size of the longer-term impact - which is necessary if suitable incentives are to be designed - is fraught with selection-induced identification problems. In this paper, we use a unique feature of the Ethiopian health worker allocation mechanism to resolve this problem.

Until recently, new medical and nursing school graduates in Ethiopia were assigned to their first posting via a national lottery. We use this natural experiment to assess the long term effects of rural postings on the careers of doctors and nurses, using data from a new survey of about 1,000 health workers in Addis Ababa and two other regions of Ethiopia. We also investigate the impact of the end of the lottery system on the ability of the government to attract workers to remote areas, and on health worker absenteeism.