Welfare Impact of Urban Agriculture in Ethiopia: Evidence from a Multinomial Endogenous Switching Regression Model

Abstract

Using primary data collected from 384 urban households in North Shewa Zone, Oromia Region of Ethiopia, this study examined the impacts of urban agriculture (UA) on household welfare, as measured by consumption expenditure. A multinomial logistic regression model was applied to identify factors influencing participation in three UA types: horticultural crop production, livestock rearing, and both combined. To estimate the welfare effects while addressing selection bias and unobserved heterogeneity, a Multinomial Endogenous Switching Regression (MESR) model was used. The multinomial model results indicated that family size, gender of the household head, landholding size, access to water, improved inputs, extension services, and non-farm income significantly influenced participation. According to the MESR model, households engaged solely in horticulture or livestock earned average consumption gains of 2,557.24 ETB and 2,867.36 ETB, respectively, while those participating in both saw gains of 1,370.76 ETB. Counterfactual analysis revealed that non-participants could have gained up to 24,393.15 ETB or 27,400.32 ETB. These amounts are significant compared to the average expenditure of 6,733.33 ETB; highlighting UA’s potential to improve welfare. Enhancing access to land, credit, water, and support services is essential to promote wider participation.