Panel Estimation of Non-Monetary Poverty in Ethiopia and its Driving Factors: A Multidimensional Approach

Abstract

By employing Alkire and Foster (2011) counting approach and Cerioli & Zani (1990)s’ unequal weighting method, this study estimates the non-monetary multidimensional poverty in Ethiopia for three waves with fourteen country-specific indicators and four dimensions in its first part; and tried to identify the presence of state dependence with unobserved heterogeneity using dynamic random effect probit model with unobserved heterogeneity estimation technique in its second part. The data for the three waves are from Demographic and Health Survey (DHS) of Ethiopia for the year 2005, 2011 and 2016. The panel estimation of non-monetary multidimensional poverty gives a higher level of multidimensional poverty compared to the reports from international organizations, mainly from OPHI and UNDP, in respective years. This is mainly due to the additional indicators and the weighting method used. Housing, cooking fuel, toilet and sanitation and electricity played a significant role compared with the other indicators for the high non-monetary multidimensional poverty in the country. It is also found that there is a substantial and positive impact of the previous poverty status of households for their future experience in poverty. Households living in rural areas are more prone to experience poverty and have low chance to exit from poverty compared to urban households. In addition to focus on infrastructure development in fighting poverty, the governments’ development policies and strategies should target households who are experiencing poverty trap mainly those who are living in rural areas.

Keywords: Panel Estimation, Non-monetary multidimensional poverty, Dynamic Random Effect Probit Model with Unobserved Heterogeneity, State dependence, Entry and exit probabilities

JEL Classification: I32, D10, D60, D69, B21