Multidimensional Wellbeing in Mozambique Vincenzo Salvucci University of Copenhagen and MPD/DNEAP

Introduction Measuring multidimensional wellbeing involves decisions regarding: – Dimensions/Indicators – Deprivation cutoffs – Relative importance/weights – Methodology for aggregation/synthetic index

Data Census (1997 and 2007) IAF/IOF (1996/97, 2002/03, 2008/09) DHS (1997, 2003, 2009, 2011) Others (MICS, TIA, etc.) Different indicators, definitions, representativity

Some studies MPI 2013 and 2014 – Based on DHS 2009 and 2011, respectively – Alkire-Foster methodology – National and provincial level A few DNEAP studies – Based on Census 1997 and 2007, IAF/IOF 1996/97 and 2008/09, DHS 1997, 2003, 2009, 2011 – FOD, Alkire-Foster, Fuzzy set methodology – Provincial, District, and Posto Admnistrativo level

A FIRST ORDER DOMINANCE APPROACH TO POVERTY MAPPING FOD principles – The state (0,1,1,0,0) is unambiguously better than (0,0,1,0,0) because the former state is always at least equivalent and is better than the latter in one instance – However, the states (1,0,1,0,0) and (1,1,0,0,0) are indeterminate – The FOD criterion is strict – No judgment is made as to the relative importance of dimension three versus all other dimensions

Safe water: – For 1997, not deprived when the water source is piped water inside or outside the house or the water source is standpipes. – For 2007, the water source should be piped water inside or outside the house/yard, spring water, hand pumped well water, or mineral/bottled water Sanitation: – For 1997, not deprived when there is a flush toilet or a latrine. – For 2007, having access to flush toilet, toilet with septic tank, or an improved latrine Education: Not deprived for households where at least one household member has some education Electricity: Not deprived for households with electricity for lighting Radio: Not deprived for households with a functioning radio

(a) Headcount ratio 1996/97 (b) Headcount ratio 2008/09 (c) 1996/97 – 2008/09 change

(a) Spatial FOD index 1997 (b) Spatial FOD index 2007 (c) Temporal FOD index

(a) Headcount ratio 1996/97 and spatial FOD index 1997 (b) Headcount ratio 2008/09 and spatial FOD index 2007 (c) Headcount ratio 2008/09 and 1996/97(d) Spatial FOD index 2007 and 1997

Poor/non-poor: consumption poverty indicator from the small area poverty analysis substitutes the radio indicator among the five FOD variables

(a) Spatial FOD index 1997 (b) Spatial FOD index 2007(c) Temporal FOD index

Serviços básicos e bem-estar em Moçambique, Sometimes studying individual dimensions is more useful for policy makers In this study we present an atlas of deprivation for several indicators Together with a multidimensional poverty index (using Alkire-Foster methodology)

Dimensions: – House type, floor, wall, and roof conditions – Sanitation – Safe water – Electricity – Access to information (radio) Census 1997 e 2007 Posto Administrativo level

Tabela

Multidimensional Poverty Index (M0_40)

Multidimensional Poverty in Mozambique: An Estimation Based on Two Different Methodologies Using the DHS 1997, 2003, 2009, and 2011 Here we compare FOD and Alkire-Foster methodology results Based on DHS data Provincial level and comparison over time

Dimensions: – Water: Deprived if drinking water is from an unprotected well, river, dam, lake, ponds, stream, canal or irrigation – Sanitation: Deprived if no toilet facility or use the bush or field – Information: Deprived if no radio or television – Electricity: Deprived if no electricity – Education: Deprived if any of the HH members have not completed primary education – Housing: Deprived if floor is made from earth, wooden planks, adobe, clay, palm or bamboo

M0 (MPI) – ranking Average domination (spatial FOD) – ranking

Temporal FOD Change in MPI

High correlation over time

Other metodologies exist Fuzzy

Conclusions Choices are important to drive the results Important that they are made explicit and discussed, and confronted with the relevant literature Relative importance of indicators/dimensions changes over time and across space – From country to country – Nampula rural/Maputo urban – 1997/2007 – Income groups

Conclusions Measuring multidimensional wellbeing is not substitute for consumption poverty estimates It complements the info available (more stable over time) Both estimates in the 4 th National Poverty Assessment using the IOF 2014/15

Thanks

Appendix Mapa