Multidimensional poverty in Botswana: leaving no one behind

The Leave No One Behind (LNOB) principle is at the core of the 2030 Agenda for sustainable development. It acknowledges that poverty is multidimensional and that it should be examined at the individual level. Notwithstanding this, poverty measurement remains heavily premised on monetary measurement and household-level multidimensional poverty measurement. However, monetary as well as household-level multidimensional poverty measures are likely to produce a biased assessment of individual poverty leading to an underestimation of the poverty levels of society. Using Botswana as a case study, the broad purpose of this thesis is to measure and analyse poverty in accordance with the LNOB principle. The objectives of my thesis are to (i) construct an individual-level multidimensional poverty measure for Botswana; (ii) examine multidimensional poverty profiles and inequality among the multidimensional poor; (iii) investigate poverty mismatch and overlaps between monetary and multidimensional poverty measures; (iv) examine targeting performance of social assistance programmes in reaching their intended beneficiaries; and (v) provide policy implications for using multidimensional poverty measure. The thesis utilised the 2015/16 Botswana multi-topic household survey (BMTHS) collected by Statistics Botswana. The results reveal that an estimated 46.2 per cent of individuals are considered multidimensionally poor based on individual-level analysis compared to 36.5 per cent when using household-level analysis. The results also reveal that monetary and multidimensional poverty measures are distinct constructs and identify different people as poor. Concerning targeting by implementation, results reveal high inclusion and exclusion errors. Regarding the targeting performance by design, the results reveal high under-coverage rates regardless of the poverty method used. However, results show higher leakage rates for the monetary measure than the multidimensional measure. These findings have policy implications. How poverty is measured reflects how it is understood, and this can significantly influence which policies should be implemented to address it.