Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Key and missing components of y g pmultidimensional poverty measurement
S bi AlkiSabina Alkire,
Bonn , 21 November 2011
I i l k h ‘Th MDG d B d’ DIE/PEGN BMZInternational workshop ‘The MDGs and Beyond’ DIE/PEGNet BMZ
Motivation for the MDGsMotivation for the MDGs“We will spare no effort to p
free our fellow men, women, and children,
from the abject and dehumanizing g
conditions of extreme poverty, to which more p y,than a billion of them are now subjected.”are now subjected.
Kofi Annan
Three Assumptionsp– Why: That post-MDGs will share the ethical motivation of
the MDGs, to reduce abject human suffering.
– Purpose: That post-2015 MDGs aim to create momentum –political & financial, and be a focal point for research, advocacy, and collective action to reduce suffering.
– How: The ‘tools’ of post-MDGs (data, indicators, measures, processes, reports) must catalyse and sustain results – cost
/effectively. Reports/measures are not ends in themselves.
M‘Growth’
Measurement Inputs
Measures DataInputs
ProcessesThis presentation:
Some on dataSome on measures
Missing data . Missing g
A key constraint in 2015?DataWork
Violence A key constraint in 2015?EmpowermentRelationships
– The MDGs focus on living standards, health, education, gender and the environmental conditions. We ignore well-known weaknesses in existing indicators.
– The number of proposed ‘ninth’ MDGs is not infinite.p p
– Proposals mainly fall into four categories, as above.
– Sources of proposals include poor people, experts, MDG Reports, PRSP documents, and research.
5– .
The Voices of the Poor found that poor people call these part of poverty.
Material Well-being (Work)
Fr d f Ch i & A ti n (E)Freedom of Choice & Action (E)
Security (Safety from Violence)
Social Well-being (Connectedness)
P h l i l W ll b iPsychological Well-beingBodily Wellbeingy g
Mental Wellbeing
The Sen-Stiglitz-Fitoussi Commission Q li f Lif ll fQuality of Life – all from same survey
S bj i ll b i• Subjective well-being• Health• Education• The Balance of Time• Political Voice & Governance• Social Connections• Social Connections• Environmental Conditions• Personal Security• Economic security including Worky g
BUT… Data on these dimensions are missinggfrom MDG surveys & other hh surveys
Most MDG data come from 4 survey instruments:
• Violence is nearly invisible. (Domestic V in some DHS)• Formal employment is covered in LSMS surveys but notFormal employment is covered in LSMS surveys but not
Informal employment or quality of work usually; DHS/MICS are weak.DHS/MICS are weak.
• Voice/Empowerment is systematically absent.Sh H ili i d I l i b• Shame Humiliation and Isolation are absent.
We might wish to look ahead as data will constrain 2015 baselineWe might wish to look ahead as data will constrain 2015 baseline
8
What’s needed: see interconnectionsWhat s needed: see interconnections
Dimensions Health Educa-tion Income Safety from
Violence Work Empow.
I di id l 1 NP P NP P P PIndividual 1 NP P NP P P P
Individual 2 NP NP P NP P NP
Individual 3 P P P NP NP NP
Individual 4 P P P P P P
5-8 min surveysurvey
modules
]
Motivation
• Intrinsic and instrumental value.• At least 5 countries identified ‘decent work’ as a ninth
MDG, or as a high priority target along the MDGs• Of 19 composite indices of poverty and well-being, 14 p p y g
included work-related indicators• Informal work is predominate among poor & among p g p g
women, so important to grasp. – 2 million people die each year from work-related accidents,
illnesses or wounds in formal work sites (WHO 2010)– 268 million non-fatal accidents cause three lost days of work
per injured worker;160 million cases of work-related illnesses.per injured worker;160 million cases of work related illnesses.All on formal worksites (informal higher).
11
Motivation• Safety from violence is clearly valued for its own sake.
• Recognition of its importance:− 19 countries stressed security from violence in their PRSP or plan− Two countries developed a 9th MDG around security− At least 17 more place violence as a pillar of their national strategy− At least 17 more place violence as a pillar of their national strategy
alongside meeting the MDGs.• An estimated 90 percent of all violence-related deaths occur in
d ddlow- and middle-income countries (Krug et al., 2002).• Fragile and failed states, and countries in conflict, are
disproportionately failing in progress on the MDGs. d sp opo t o ate y a g p og ess o the MDGs.• For every war death, more than 3 people die of crime &
homocide (WHO Burden of Disease 2004; 2008 update)
Shame, humiliation and isolation
• The stigma of poverty is a recurring theme among the poor• Can result in increasing isolation from services and support• Can undermine social relationships and provoke psycho-social
l di l lf i l l i h lmaladies: low self-esteem, poor interpersonal relations, school-related difficulties, delinquency, social phobia, etc.
• May fuel horizontal inequalities and spill over into conflict.• May discourage use of public services by poor
Missing Dimensions: Empowerment
Arab Spring – most potent indication
• Moving out of poverty 2009 found that 77.5% of those who exit
i h i ‘poverty cite their ‘own initiative’ as the most important reason forimportant reason for moving out of poverty.
•
New Goals
need new Data
ButBut…Focus is key
](data needed anyway)
Multidimensional MeasuresMeasures u t d e s o a easu esDashboards
Multdimensional
Reports & AnalysisWhat is useful in these multidimensional measures?multidimensional measures?
1. They show who is deprived in what at the same time. The MDGs do not.
Who is poor?Who is poor?
A l d ll f hA person is multidimensionally poor if they are deprived in 33% of the dimensions.
33%33%
Endah’s MPIThe MPI is built from each person’s own
profile of poverty, and keeps this information so we can zoom back to it later.
Seeing the Interconnections: Measures gMultidimensional MeasuresDashboards
Multdimensional
Reports & Analysis
What is useful in these dmultidimensional measures?
1 They show who is deprived in what at the same time1. They show who is deprived in what at the same time. The MDGs do not.
2 Th i i b h f ld l2. They give an overview, an above-the-fold, at-a-glance view of the trends – nationally and by regions.
The MPI Headcount Ratios and the $1.25/day Povertyy y103 of our 109 Countries in
2011 have $1.25 data; only 71 have ; y$1.25 poverty data within 3 years of MPI. $1.25 data ranges from 1992-
f2008; MPI from 2000-2010.
Niger
Ethiopia
Mali
tral African
Republic
Burundi
Liberia
Burkina Faso
Guinea
Rwanda
Mozam
bique
Sierra Leone
Comoros
DR Congo
Uganda
Malaw
iBenin
Timor‐Leste
Senegal
Madagascar
Tanzania
Nepal
Zambia
Chad
Cote d'Ivoire
Gam
bia
Bangladesh
Haiti
Togo
Nigeria
India
Cameroon
Yemen
Cambodia
Pakistan
Kenya
Lao
Swaziland
Republic of Congo
Gabon
Lesotho
o Tome and Principe
Honduras
Ghana
Djibouti
Nicaragua
Bhutan
Guatemala
Indonesia
Bolivia
Peru
Viet Nam
Tajikistan
Mongolia
Iraq
Philippines
South Africa
Paraguay
China
Morocco
Estonia
Turkey
Egypt
Syrian
Arab Republic
Colombia
Sri Lanka
Azerbaijan
Maldives
Kyrgyzstan
Dom
inican
Republic
Hungary
Croatia
Mexico
Argentina
Brazil
Jordan
Uzbekistan
Ecuador
Ukraine
Macedonia
Moldova
Uruguay
Thailand
Latvia
Montenegro
Albania
Russian Federation
Arm
enia
Serbia
nia and Herzegovina
Georgia
Kazakhstan
Belarus
Slovenia
CenSa
S
Bosn
Some MPI 2011 ‘Fast Facts’
• Extent: 32% of people in 109 countries are poor• R i n 50% f MPI p p pl li i S th• Region: 50% of MPI poor people live in South
Asia and 29% in SS Africa.• MICS: 69% of MPI poor live in Middle Income
Countries. • Rural: 83% of MPI poor live in Rural areas.• Range: MPI headcount in MICs ranges from 0-Range: MPI headcount in MICs ranges from 0
77%; MPI headcount in LICS: 5-92%.• Disparity: National aggregates hide disparities b• Disparity: National aggregates hide disparities by
region and ethnic group.
National MPINational MPI(109 Countries)
22
Sub-national MPI (66 countries)Sub national MPI (66 countries)(highest disaggregation available)
23
TunisiaMorocco
TunisiaMorocco
PakistJordan
Iraq
MauritaniaMauritania
United Arab Emirates
Egypt
Guinea
Chad
Gambia
Senegal
Mali Niger
Mauritania
DjiboutiGuinea
Chad
Gambia
Senegal
Mali Niger
Mauritania
DjiboutiBurkina Faso
Yemen
BeninCote d'Ivoire Togo
CameroonCentral African Republic
Liberia
Nigeria
Uganda
EthiopiaSierra Leone
Gambia
SomaliaBeninCote d'Ivoire Togo
CameroonCentral African Republic
Liberia
Nigeria
Uganda
EthiopiaSierra Leone
Gambia
Somalia
S T d i iGabon
Republic of Congo
Ghana
DR CongoTanzania
KenyaGabon
Republic of Congo
Ghana
DR CongoTanzania
KenyaBurundi
Sao Tome and Principe
Angola
Zambia
Zimbabwe
Malawi
Mozambique
Angola
Zambia
Zimbabwe
Malawi
Mozambique
Namibia
ZimbabweMadagascar
Namibia
ZimbabweMadagascar
24
SwazilandSouth Africa
Lesotho
SwazilandSouth Africa
Lesotho
B li
Dominican RepublicMexico
B li
Dominican RepublicMexico
Belize
Haiti
Guatemala
G
Trinidad and Tobago
Belize
Haiti
Guatemala
G
Trinidad and Tobago
Honduras
NicaraguaGuyana
SurinameColombia
NicaraguaGuyana
SurinameColombia
Peru
Ecuador
Peru
Ecuador
Brazil
Bolivia
Brazil
Bolivia
ParaguayParaguay
UruguayUruguay
25
ArgentinaArgentina
Nutrition (CH)Child Mortality (CH)
Bangladesh
Bhutan
Nepal
India
Pakistan
Bangladesh
Bhutan
Nepal
India
Pakistan
Bangladesh
Bhutan
Nepal
India
Pakistan
Bangladesh
Bhutan
Nepal
India
Pakistan
Nutrition (CH)
MyanmarMyanmar MyanmarMyanmar
Sri LankaSri Lanka Sri LankaSri Lanka
S f D i ki W (CH)S h l A d (CH)
Nepal
Pakistan
Nepal
Pakistan
Nepal
Pakistan
Nepal
Pakistan
Safe Drinking Water (CH)School Attendance (CH)
Myanmar
Bangladesh
Bhutan
India
Myanmar
Bangladesh
Bhutan
India
Myanmar
Bangladesh
Bhutan
India
Myanmar
Bangladesh
Bhutan
India
26Sri LankaSri Lanka Sri LankaSri Lanka
Seeing the Interconnections: Measures gMultidimensional MeasuresDashboards
Multdimensional
Reports & Analysis
What is useful in these dmultidimensional measures?
1 They show who is deprived in what at the same time1. They show who is deprived in what at the same time. The MDGs do not.
2 Th i i b h f ld l i f2. They give an overview, an above-the-fold, at-a-glance view of the trends – nationally and often by region.
3. Alkire Foster measures can be broken down by region and by indicator, to show how change happened.
Ghana, Nigeria, and Ethiopia2003-8 2003-8 2000-52003 8 2003 8 2000 5
70%
75%
MPI = A x H
Poorest Countries,Highest MPI Niger
65%
ty (A
)
Nigeria
Ethiopia
Mozambique
55%
60%
of P
over
t
Indonesia
Nigeria
ChinaAngola
50%
Bre
adth
IndonesiaDR Congo
UgandaBrazil
40%
45%
Ave
rage
Upper-Middle IncomeLower Middle Income
IndiaSouth Africa
BangladeshGabon
NamibiaGhana
35%
Lower-Middle IncomeHigh IncomeLow Income
Kyrgyzstan
Turkey
Ukraine
Tajikistan
30%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percentage of People Considered Poor (H)
Pathways to Poverty Reduction
2
Ghana Nigeria Ethiopia
1
2
Assets
Cooking
0
Cha
nge
nt
gFuel
Floor
-2
-1
bsol
ute
Cd
Hea
coun Water
Sanitation
-3
ualiz
ed A
bC
enso
red
Electricity
Nutrition
-4Ann
u C
Mortality
Child
-6
-5Child Enrolment
Schooling
Changes in MPI by sub-national region
0.5
1ty
(A
)
0-7 -6 -5 -4 -3 -2 -1 0 1%
Int
ensi
ty
-0.5
-7 -6 -5 -4 -3 -2 -1 0 1
atio
n in
%
-1
lute
Var
ia
-2
-1.5
nual
Abs
ol
-2.5
2
Ann
u
Annual Absolute Variation in % Headcount Ratio (H)Annual Absolute Variation in % Headcount Ratio (H)
Nigeria Ghana Ethiopia
Regions of Nigeria: How MPI changed
2
South East
South West South
North Central
North West
North East
1
2
Schooling
Attendance
0
hang
e nt
Attendance
Child Mortality
-2
-1
bsol
ute
Ch
d H
eaco
un Nutrition
Electricity
-3
2
ualiz
ed A
bC
enso
red
Improved Sanitation
Water
-4Ann
u C
Flooring
Cooking
-6
-5Cooking Flue
Assets
0,7
0,6
Upper confidence
0 4
0,5 interval
0 3
0,4Lower confidence interval
0,2
0,3
Point estimate
0,1Nigeria: Which of the 10 indicators h t ti ti ll i ifi t?
0
changes are statistically significant?
MPI Data UpdatesMPI Data Updates (best estimates)
• In 2012, between 30 and 50 new and updated datasets will be available and more than one third ofdatasets will be available and more than one third of the MPI countries’ data will date from 2010.
• By 2014, the MPI will probably have been calculated for over 100 new and updated datasets.for over 100 new and updated datasets.
33
Seeing the Interconnections: Measures gMultidimensional MeasuresDashboards
Multdimensional
Reports & Analysis What is useful in these multidimensional measures?
1. They show who is deprived in what at the same time. The MDGs can not (And it matters). The MDGs ca ot ( d t atte s).
2. They give an overview, an above-the-fold, at-a-glance view of the trends nationally and often by regionthe trends – nationally and often by region.
3. Alkire Foster measures can be broken down by region and by i di h h h h dindicator, to show how change happened.
4. The indicators, dimensions, cutoffs are flexible. You , ,choose them for your own purpose.
1. (data permitting)
Uses? Colombia launched an official national MPI to monitor progress towards their plan using the Alkire-Foster methodprogress towards their plan using the Alkire Foster method.
Poverty committee: monitoring poverty reduction▪ Leaders
– Counselor for the Presidency– National Planning Department
▪ Permanent members– Ministry of Health– Ministry of Labory– Ministry of Housing– Ministry of Agriculture– Ministry of Educations y o duca o– Ministry of Finance
No Deputies permittedNo Deputies permitted.
MANDATORY PRESENCEThe President of Colombia
35
Mexico’s AF measure shows extreme & moderate povertyextreme & moderate poverty
WithoutWith deprivations
Vulnerable people by
Ideal Situation
Without
DepE i
EWL
Vulnerable people by social deprivations
$1,921.7 U$1,202.8 R
privaMULTIDIMENSIONALLY POOR
Economic wellbeing line
EXTREME Multidimensional
Moderate MultidimensionalPoverty
Vulnerable people by
income
MWL
$1,202.8 R
$874.6 U$613.8 R
tion
MULTIDIMENSIONALLY POOR
Minimum MultidimensionalPoverty
03
income
5 24 16
sMinimum
wellbeing line
Social RightsDeprivations
035 24 16
Deprivations
Bhutan’s national MPI adjusts the i iglobal MPI using better national
indicators – with participatory inputindicators with participatory input.
A possibility?1. Create an internationally comparable measure with
some post 2015 MDGs as a ‘bell weather’ indicatorsome post-2015 MDGs, as a bell-weather indicator alongside the detailed single indicator dashboard.
2. Encourage countries/groups to make ‘National MPIs’ reflecting their own priorities, goals & values. (we already do this for income poverty)
I ’ h h k? N h i ll I i f iblIsn’t this too much work? Not technically. It is feasible.
Example: Colombia’s measure & report was made in 4 monthsExample: Colombia s measure & report was made in 4 months by 3 persons under the age of 33 who do not have PhDs.
Concluding Remarks:
1. MDG measures need to anticipate the data requirements especially for a ‘baseline’ in 2015requirements – especially for a baseline in 2015.
2 It is possible to include brief 5-min modules on:2. It is possible to include brief 5 min modules on: work, violence, empowerment, and relationships.
3. The value-added of a multidimensional measure is that it reveals coupled deprivations. MDGs don’t.
‘S i i ’ li i dd l i l MDG• ‘Synergistic’ policies address multiple MDGs; traps. • You can go inside a country to regions / groups
4. Consider an international MPI, and support for national MPIs if they will actually be used.
2 Dimensions and Indicators of MPI2. Dimensions and Indicators of MPI
3 Methodology: Aggregation3. Methodology: Aggregation
We construct the MPI using the Alkire & Foster M0:
H i h f l h I h
Formula: MPI = M0 = H × A
• H is the percentage of people who are poor. It shows the incidence of multidimensional poverty.
• A is the average proportion of weighted deprivations g p p g ppeople suffer at the same time. It shows the intensity of people’s poverty.p p p y
Policy implications...Country A: Country B:
Policy oriented to the poorest of the poorPoverty reduction policy
MultidimensionalHeadcount
(H)
Intensity of Deprivations
(A)
Multidimensional Poverty Index(MPI = H * A)
MultidimensionalHeadcount
(H)
Intensityof Deprivations
(A)
Multidimensional Poverty Index(MPI = H * A)
MultidimensionalHeadcount
(H)
Intensity of Deprivations
(A)
Multidimensional Poverty Index(MPI = H * A)
MultidimensionalHeadcount
(H)
Intensity of Deprivations
(A)
Multidimensional Poverty Index(MPI = H * A)
y p p(without inequaliy focus)
70.00
75.00
58.00
59.00
60.00
0.40
0.41
0.42
Before
(H) (A) (MPI = H A)
70.00
75.00
58.00
59.00
60.00
0.40
0.41
0.42
Before
(H) (A) (MPI = H A)
70.00
75.00
58.00
59.00
60.00
0.40
0.41
0.42
Before
(H) (A) (MPI = H A)
70.00
75.00
58.00
59.00
60.00
0.40
0.41
0.42
Before
(H) (A) (MPI H A)
65.00
55.00
56.00
57.00
0.36
0.37
0.38
0.39
65.00
55.00
56.00
57.00
0.36
0.37
0.38
0.39
65.00
55.00
56.00
57.00
0.36
0.37
0.38
0.39
65.00
55.00
56.00
57.00
0.36
0.37
0.38
0.39
55.00
60.00
52.00
53.00
54.00
0 32
0.33
0.34
0.35
55.00
60.00
52.00
53.00
54.00
0 32
0.33
0.34
0.35
55.00
60.00
52.00
53.00
54.00
0 32
0.33
0.34
0.35 After
55.00
60.00
52.00
53.00
54.00
0.32
0.33
0.34
0.35
After
50.00 50.00
51.00
0.30
0.31
0.32
50.00 50.00
51.00
0.30
0.31
0.32
50.00 50.00
51.00
0.30
0.31
0.32
50.00 50.00
51.00
0.30
0.31
0.32
Country B reduced the intensity of deprivation among the poor more. The final index reflects this.
(MPI satisfies Dimensional Monotonicity)