Multidimensional Poverty Index Milorad Kovacevic Human Development Report Office.

Multidimensional Poverty Index

Milorad KovacevicHuman Development Report Office


• The dimensions of poverty go far beyond inadequate income—to – poor health and nutrition, – low education and skills, – inadequate livelihoods, – bad housing conditions, – social exclusion and lack of participation

Workshop on HD Approach and Measurement for the GCC States, Doha, 9-11 May, 2011 2


• HDR 2010, in collaboration with Oxford University’s Poverty and Human Development Initiative, introduced a new Multidimensional Poverty Index (MPI)

– only for 104 developing countries (due to lack of comparable data)– The 104 countries include 92% of the population in 98 developing

countries – in 2011 – at most 120

• The MPI is an index of acute multidimensional poverty and is meant to complement monetary based measures


Dimensions of the Multidimensional Poverty Index

• MPI identifies overlapping deprivations at the household level • Composed of 10 indicators corresponding to the same 3 dimensions

as the HDI: Health, Education and Living Standards- Each dimension is equally weighted- Each indicator has equal weight within its dimension

4

Dimensions of the Multidimensional Poverty Index

• The MPI shows the average number of poor people and the average number of deprivations with which poor households contend

• A household is multidimensionally poor if it is deprived in at least 30% of the weighted indicators (2 to 6 indicators)

• The MPI reveals a different pattern of poverty than income poverty – it illuminates a different set of deprivations


Data Sources

Data: Household Surveys‐ Demographic & Health Surveys (DHS – 48 countries) ‐ Multiple Indicator Cluster Surveys (MICS – 35

countries)‐ World Health Survey (WHS – 19 countries)

‐ Additionally used 2 special surveys covering Mexico and urban Argentina

‐WHS 2003 for United Arab Emirates

‐ MPI is deeply affected by lack of comparable dataWorkshop on HD Approach and Measurement for the GCC States, Doha, 9-11 May, 2011 6

Methodology• MPI corresponds to the first measure of the Alkire & Foster (2007) family of multidimensional poverty measures, called M0

• It is constructed using the AF method:

H is the percentage of people who are poor - shows the incidence of multidimensional poverty: (H=q/n)

A is the average proportion of weighted deprivations people suffer at the same time - shows the intensity of multidimensional poverty:


• Two step procedure applied to identify who is multidimensionally poor, uses dual cutoff method:

1. Identify all individuals deprived in any dimension• Within dimension cutoff

2. Identify who is multidimensionally poor• Cross dimensional cutoff

– Deprived in at least 30% of the weighted indicators

Methodology


Measurement Indicators and Cutoffs

Health (each indicator weighted equally at 1/6 )

- Child Mortality: If any child has died in the family

- Malnutrition: If any interviewed adult in the family has low Body Mass Index; if any child is more than 2 standard deviations below the reference normal weight for age, (WHO standards) [WHS has male data but no child data; MICS has child data but no adult data]



Education (each indicator weighted equally at 1/6 )

- Years of Schooling: if no person in the household has completed 5 years of schooling

- Child Enrolment: if any school-aged child is out of school, where school-aged is an 8 year period from the national starting age

Workshop on HD Approach and Measurement for the GCC States, Doha, 9-

11 May, 201110


Standard of Living (each indicator weighted equally at 1/18)

- Electricity (no electricity is poor)- Drinking water (MDG definitions)- Sanitation (MDG definitions + not shared)- Flooring (dirt/sand/dung are poor)- Cooking Fuel (wood/charcoal/dung are poor)- Assets (poor if do not own a car/truck and do not

own more than one of these: radio, tv, telephone, bike, motorbike, or refrigerator)


11 May, 201111

Illustration

• Ali’s household is deprived in nutrition and child enrolment. Is Ali’s household multidimensionally poor?

10(1/6)+10(1/6) = 3.34 (> 3) Yes• Maira’s household is deprived in electricity, water, sanitation,

and has a dirt floor. Is Ali’s household multidimensionally poor?

10(1/18)+10(1/18) + 10(1/18)+10(1/18) = 2.20 (<3) No

• Tom’s household is deprived in years schooling, sanitation, assets, and cooking fuel. Is Tom’s household multidimensionally poor?

10(1/6)+ 10(1/18)+10(1/18) + 10(1/18)= 3.33 (>3) Yes


11 May, 201112

Missing Dimensions

Missing dimensions include:• Work• Empowerment• Safety from Violence (crime, conflict)• Political Freedom• Relationships (social capital, inclusion, dignity)• (Cultural/Spiritual/Subjective Well-being)

Data are not available to incorporate any of these into the MPI for 100+ countries


11 May, 201113

Patterns of Multidimensional Poverty

• 32% the population in 104 developing countries, about 1.75 billion people, are MPI poor– The 104 countries includes 92% of the population in 98

developing countries

• Regional rates vary from 3% in Europe and Central Asia to 65% in sub-Saharan Africa

• South Asia is home to the largest number of MPI poor, followed by sub-Saharan Africa


11 May, 201114


Eight Indian states are home to 421 million MPI poor people - more than the 410 million poor living in the 26 poorest African countries combined

Half the world’s MPI poor live in South Asia, but the intensity of MPI poor is highest in sub-Saharan Africa


11 May, 201115


Countries with higher multidimensional poverty headcounts tend to have more deprivations



0400800

MPI poor population (millions)

South Asia has highest incidence of multidimensional poverty in the world-ranging from 38.7% in Sri Lanka to 54.1% in Nepal

Sub-Saharan Africa has significant variation–ranging from 3% in South Africa to 93% in Niger

East Asia and the Pacific has relatively low rates of multidimensionalpoverty– but over half of Cambodians are MPI poor


11 May, 201117


0204060

MPI poor popu-lation (millions) Arab states MPI values are

generally below 7%, but as high as 52% in Yemen and 81% in Somalia

Latin America and Caribbean MPI values range from 2% (Uruguay) to 57% (Haiti)

Europe and Central Asia’s incidence of Multidimensional poverty is lowest of the developing country regions – close to zero in several countries, while Tajikistan is the highest with 17%


11 May, 201118


United Arab Emirates (WHS 2003)


11 May, 201119

MPI H A @ Risk (1)

At least one severe deprivationEducation Health Living

Standard0.002 0.6% 35.3% 2.0% 0.6% 5.4% 0.0%

(1) Suffering in 20% of weighted indicators


Within-countries– Nairobi is similar to

Dominican Republic, rural northeast is worse than Niger

Among ethnicities, religions and castes– MPI headcount in Kenya

ranged from 29% for the Embu to 96% for the Turkana and Masai

The MPI highlights significant variations:


11 May, 201120

Changes Over Time

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1M

PI -

2004

MPI -

2007

H -

2004

H -

2007

A -

2004

A -

2007

MPI -

2000

MPI -

2005

H -

2000

H-2

005

A -

2000

A -

2005

MPI -

2003

MPI -

2008

H -

2003

H -

2008

A -

2003

A -

2008

Bangladesh Ethiopia Ghana

MPI at two points in time in Bangladesh, Ethiopia and Ghana


Relevance to Country Level Work

• Can be adapted using indicators and weights that make sense for the region or the country

• Can be adopted for national poverty eradication programs

• It can be used to study changes over time


11 May, 201122

Policy Applications

• Allocate resources effectively• Target those with the greatest intensity of poverty

• Identify interconnections among deprivations• Helps in addressing MDGs strategically

• Design policy• Show which deprivations are most common in different

groups so that policies can be tailored to particular needs

• Show impacts• Reflects results of policy interventions quickly


11 May, 201123

Limitations of the Multidimensional Poverty Index

• Drawbacks mainly due to data constraints• Indicators include inputs, outputs and one stock indicator

because flow data unavailable in some instances• Health data relatively weak• Judgments necessary where data is missing

• Intrahousehold inequality is not captured• Does not measure inequality amongst the poor • Cross-country comparability limited


11 May, 201124

Criticism of the Multidimensional Poverty Index

Martin Ravallion’s recent criticism:• Arbitrariness of components and weights.• Income based measures aggregate consumption across a large number of goods.• Limited information about trade-offs.• Uncertainty about robustness of resulting rankings.

Responses• Giving income weight of one is no less arbitrary.• Prices reflect scarcities and current distribution.• Indices can and should enable reasoned public debate about the implicit weights

and trade-offs.• Contention of HDR approach is that since education and health are public

goods, current prices underestimate their social value.• MPI background research exhaustively evaluated robustness


11 May, 201125

Multidimensional Poverty Index Milorad Kovacevic Human Development Report Office.

Documents

dimensions of poverty

income poverty

hd approach

measurement indicators

gcc states

different pattern of

oxford universitys poverty

methodology mpi