Measuring multidimensional poverty in Latin America

1. Seminar on poverty measurement Geneva, 5-6 May 2015 Measuring multidimensional poverty in Latin America Xavier Mancero Statistics Division, ECLAC

2. Background • Income provides an incomplete assessment of living standards. • Possible bias characterizing poverty • Income poverty measure does not account for the impact of public policies in various areas of welfare • Multidimensional poverty is increasingly being monitored at the country level (Mexico, Colombia, Chile). • Social Panorama of Latin America 2014 presented a multidimensional poverty index, that: • Builds upon “Unmet Basic Needs” method • Includes deprivations in terms of employment, social protection and schooling gap, thus widening the set of dimensions commonly used; • Integrates monetary and non-monetary dimensions, so as to minimize errors of inclusion and exclusion in identifying the poor; • Includes new deprivation cut-offs that better reflect the current regional reality

3. Regional multidimensional poverty index

4. Regional multidimensional poverty index • Weights • Equal weights (7.4%), excluding social security (3.7%) and income (14.8%). • Deprivation of social protection  less associated with the traditional concept of poverty. • Income  income is itself a synthetic indicator of welfare. • Multidimensional threshold k = 25%. • Poor = deprivation in a complete dimension plus an indicator from other dimension; or deprivation in income and at least two additional deprivations. • No person who is deprived in only one dimension is identified as multidimensionally poor.

7. Source: ECLAC (2014), Social Panorama of Latin America 2014

8. Source: ECLAC (2014), Social Panorama of Latin America 2014

9. Source: ECLAC (2014), Social Panorama of Latin America 2014

10. Contribution of each dimension to overall poverty, around 2012 Source: Santos et al (2014), “A Multidimensional Poverty Index for Latin America”

11. Estimates for Different k Values es • robustness Source: Santos et al (2014), “A Multidimensional Poverty Index for Latin America”

12. Otro grafico Source: Santos et al (2014), “A Multidimensional Poverty Index for Latin America”

13. Exclusion discrepancy and multidimensional poverty rate “ Exclusion discrepancy”: percentage of total population that is multidimensionally poor but not income poor Source: Santos et al (2014), “A Multidimensional Poverty Index for Latin America”

14. Some results • Deprivations suffered by the poor vary from country to country in respect of intensity and the forms they take. • Income insufficiency is important, but it is not the only hardship that the poor suffer. • Income deprivation has the highest contribution, and it is higher in countries with low poverty rates. • Contribution of precarious housing, lack of energy and of durable goods is higher in high-poverty countries. • Multidimensional poverty yields similar headcount ratios to income poverty in most countries, but both methods do not necessarily identify the same population as poor. • Poverty reduction requires coordinated policies across multiple sectors.

15. Multidimensional poverty and data • Adopting a multidimensional index provides useful information for analysis of living conditions and to guide policy . • Current information is insufficient and lacks comparability. • Education: Indicators of access but not quality or competency in adults. • Housing: variables and categories not clearly linked to deprivations. • Health: not measured in most regular surveys of the region. • Current context (SDGs and data revolution) offers an opportunity to improve household surveys. • Moving towards the harmonization of certain basic dimensions. • More comprehensive (within the constraints of sample size and representativeness). • Taking advantage of existing survey programs, in the context of stronger and better coordinated National Statistical Systems