Les stratégies de subsistance en milieu rural en Afrique du Sud: Implications pour la réduction de la pauvreté
Zerihun G. Alemu, Development Bank of Southern Africa
This paper attempted to tackle some related questions in the debate around rural development - identification of the poor in rural areas, the kind of livelihood strategies they pursue, and the constraints they face to take on high-return livelihood strategies. The study was based on a recent (2009) nationally-representative survey (GHS) obtained from Statistics South Africa. Guided by the review of the available literature, it grouped households into distinct livelihood strategies, it applied Stochastic Dominance Tests to rank livelihood strategies by some measure of welfare, and it used a multinomial logistic regression technique to identify barriers that poor households face to enter into high-return livelihood strategies. Households were classified into four livelihood strategies, namely, only farm, only non-farm, farm and non-farm, and non-labor.
Thereafter, the four livelihood strategies were sub-divided into seven specific livelihood strategies. Ordinal ranking of livelihood strategies using Stochastic Dominance Test gave that households that rely on non-farm sources of income (wage income) are relatively non-poor compared with other households.
The identified classifications, together with socioeconomic characteristics of households, were next used to estimate a multinomial logistic regression to identify factors that constrain poor households’ access to high-return livelihood strategies. Results indicate that poor households face a host of constraints. These include, inter alia, demographic (gender and age of household heads), human capital (labor endowment and education), and social infrastructure (access to basic services). The paper concludes by analyzing policy implications of the findings in the context of current policy discourse around issues of rural development in South Africa.
CEA 2011 - Programme (298 kB)
Brochure CEA 2011 (1,01 MB)
Rapport de recherche 2010 (3,15 MB)