Join us in person (refreshments provided; no booking required) or online (please register for webinar link: bit.ly/4dOcNP4)

Two main focuses of studies on multidimensional poverty pertain to measurement issues and using statistical techniques to identify factors that explain deprivation. This study extends the literature by proposing an approach to produce forward-looking projections of multi-dimensional poverty (MDP) indicators. The proposed approach uses multinomial logistic regression (MLR) techniques to establish statistical association between dimensions of poverty and demographic, economic and social factors, and uses the outcome of the estimation to add an MDP module to a linked macro-micro model. The final model is a policy tool that can be used to design anti-poverty policies and produce ex-ante assessment of their impact on MDP. To demonstrate the approach, we used a full General Household Survey of South Africa as the database to measure deprivation using education, healthcare, living conditions and assets as four dimensions of poverty, each measured through four indicators with low and high cut-off indicators. We then specified and estimated two multinomial logistic regression models for the two cut-offs, each with five deprivation outcomes as their categorically distributed dependent variable and a set of independent variables composed of demographic, economic and social indicators. The estimated MLR equations were used to build the MDP module of a South African linked macro-micro model, built by the Applied Development Research Solutions. In each period, the model’s projections of demographic, economic and social variables are used by the MDP module to generate projections of deprivation indicators at national level and by gender, race and region. The final MDP augmented model is used to establish the current trajectory for deprivation of various population groups in South Africa and to test the direct and indirect effects of five cumulative policy measures (i.e. fiscal policy, monetary policy, private investment, public employment and social grant scenarios) on dimensions of deprivation over the period from 2024 to 2030.