Multidimensional poverty: Global MPI and NITI Aayog's national MPI

Detail

1. What "multidimensional poverty" means

• Income poverty (the poverty-line method) asks one question: does a person's spending fall below a fixed rupee amount?
• Multidimensional poverty asks a wider question: what does the person actually lack in daily life?

• Examples: enough food, schooling, clean water, a toilet, electricity, a proper house.

• Overlapping deprivations means the same person lacks several of these things at the same time. A person without a toilet may be counted poor on one measure and not on another. The MPI adds up all the gaps faced by the same household.

• The Multidimensional Poverty Index (MPI) measures deprivations in health, education and living standards together.
• Two parts of the MPI:
• Incidence (H) is the share of people who are poor.
• Intensity (A) is how deprived the poor are, on average.

• MPI = H × A.

• Who publishes it:

• The Global MPI comes from UNDP (its Human Development Report Office, HDRO) and OPHI (Oxford Poverty and Human Development Initiative).
• India's national MPI is prepared by NITI Aayog.

2. Global MPI (UNDP HDRO and OPHI, 2010)

Origin

• It started in 2010.
• It replaced the Human Poverty Index (HPI), which was used from 1997 to 2009.
• The HPI used country-level averages.
• The MPI builds up from data on each household, so it can show who is poor and in what ways.

Method: the Alkire-Foster method

• The Alkire-Foster method is a counting method, named after Sabina Alkire and James Foster. It works in three steps: 1. Check each person against a list of indicators. The person is marked "deprived" or "not deprived" on each one. 2. Add up the weights of the indicators on which the person is deprived. This total is the person's deprivation score. 3. If the score crosses a cut-off, the person is counted as poor.

• It uses 3 dimensions. Each dimension gets an equal weight of 1/3. Together they hold 10 indicators.

Dimension (weight 1/3)	Indicators	Weight of each indicator
Health	Nutrition, child mortality	1/6 each
Education	Years of schooling, school attendance	1/6 each
Living standards	Cooking fuel, sanitation, drinking water, electricity, housing, assets	1/18 each

Cut-offs (the deprivation score needed for each label)

• Poor: deprived in at least 1/3 (33.3%) of the weighted indicators.
• Severely poor: deprived in at least 1/2 (50%).
• Vulnerable: deprived in 1/5 to 1/3 (20% to 33.3%). These people are not yet poor, but one shock (an illness, a crop failure) could make them poor.

Worked example: is one household poor?

• A child in the house has died, which counts as child-mortality deprivation. Weight = 1/6.
• The family cooks on wood and has no toilet. It lives in a kutcha house. Weight = 3 × 1/18 = 1/6.
• Deprivation score = 1/6 + 1/6 = 1/3. This reaches the 1/3 cut-off, so everyone in the household is counted as MPI-poor.
• Suppose the house were pucca instead. The score would be 1/6 + 2/18 = 5/18 ≈ 0.28. That lies between 1/5 and 1/3, so the household would be vulnerable, not poor.

The formula: MPI = H × A

• H (headcount ratio) = number of poor people ÷ total population.
• A (intensity) = the average deprivation score of the poor only.
• The MPI runs from 0 (nobody deprived in anything) to 1 (everybody deprived in everything).
• Why multiply?
• Suppose one poor person stops lacking one thing but is still poor. H does not change.
• A still falls, so the MPI falls.
• This means the index rewards help given to the poorest, not just moving people who were close to the line.

Worked example: the whole population

• Out of 100 people, 20 are poor. So H = 0.20.
• The average score of these 20 people is 0.50. So A = 0.50.
• MPI = 0.20 × 0.50 = 0.10.

Decomposition (breaking the MPI into parts)

• The MPI can be split by indicator, region (state or district) and social group (caste, religion, age).
• This shows which deprivation adds most to poverty, and where. Governments can then aim schemes at those gaps, for example cooking fuel in one state and nutrition in another.

India in the Global MPI

• The Global MPI 2023 found that 415 million Indians left multidimensional poverty between 2005-06 and 2019-21. The headcount fell from 55.1% to 16.4%.
• New editions come out around 17 October each year, which is the International Day for the Eradication of Poverty. Check the latest edition before quoting India's figure.

3. The World Bank's Multidimensional Poverty Measure (MPM) is a different index

• Do not confuse it with the Global MPI. The MPM puts money income inside the index itself.
• Three dimensions, 1/3 weight each [6]:
• Monetary: daily consumption or income below $3.00 per person a day (2021 PPP) [6]. PPP (purchasing power parity) adjusts money for price differences between countries.
• Education: 2 indicators of 1/6 each. One is a school-age child (up to grade 8) not enrolled. The other is no adult in the house with completed primary education [6].

• Basic infrastructure: 3 indicators of 1/9 each. They are drinking water, sanitation and electricity [6].

• Cut-off: a person is poor if deprived in at least one full dimension, or in indicators that add up to the same weight. That means 1/3 or more of the total weight [6].

• Latest global figures (around 2023): monetary poverty 11.2%, multidimensional poverty 16.7%. The data cover 100 economies with 67.5% of the world's people [6].

• So counting non-money needs shows poverty about 51% higher than money alone [6].

• The scaffold notes that the MPM's 2026 edition covers India (verify).

4. National MPI (NITI Aayog)

Institutional setting

• NITI Aayog is the nodal agency (the lead body responsible) for the MPI under the government's Global Indices for Reform and Growth (GIRG) initiative.

• Under GIRG, each global index India appears in is tracked by one ministry or body. That body uses the index to push reforms.

• It uses 12 indicators: the 10 global indicators plus 2 more:

• Maternal health, which sits in the health dimension.

• Bank account, which sits in the living-standards dimension.

• Data source: the National Family Health Survey (NFHS), rounds 4 (2015-16) and 5 (2019-21) [3][4].

• Identification rule: a person deprived in one-third or more of the 12 weighted indicators is counted as "MPI poor" [2][3].

The three releases

Release	Data	Key results
Baseline (Nov 2021)	NFHS-4 (2015-16)	25.01% poor. Bihar highest (51.91%), Kerala lowest (0.71%)
Progress Review (July 2023)	NFHS-4 → NFHS-5	24.85% (2015-16) → 14.96% (2019-21). About 13.5 crore people left poverty [4]. Rural 32.59% → 19.28%. Urban 8.65% → 5.27%
Discussion paper (Jan 2024)	2013-14 → 2022-23 (extrapolated)	29.17% → 11.28%. About 24.82 crore people left poverty [2][3]

Progress Review 2023: more detail

• Intensity (the average deprivation of the poor) fell from 47.14% to 44.39% between 2015-16 and 2019-21 [4].
• The MPI value almost halved, from 0.117 to 0.066 [4].
• Check with the formula:
• 2015-16: 0.2485 × 0.4714 ≈ 0.117.
• 2019-21: 0.1496 × 0.4439 ≈ 0.066.

• Both H and A fell. So fewer people were poor, and those still poor lacked fewer things.

• SDG link: SDG Target 1.2 asks countries to cut poverty "in all its dimensions" by at least half by 2030. NITI Aayog says India is on track to reach this target well before 2030 [4][5].

Discussion paper 2024: state-wise exits (2013-14 to 2022-23) [2]

• Uttar Pradesh: 5.94 crore, the largest.
• Bihar: 3.77 crore.
• Madhya Pradesh: 2.30 crore.
• Rajasthan: 1.87 crore.
• The states that were poorest at the start had the biggest exits. This means poverty is falling fastest where it was highest, so the gap between states is narrowing.

Faster decline recently

• The annual rate of decline is the average yearly percentage fall in the headcount ratio.
• It was 10.66% a year from 2015-16 to 2019-21.

• It was 7.69% a year from 2005-06 to 2015-16.

• The paper also covers the 2005-06 NFHS round, which gives a longer comparison [3].

5. Critiques of the national MPI

• Extrapolation
• No NFHS survey exists for 2013-14 or 2022-23.
• These figures are projected from trends between actual survey years, not measured directly [3].

• The "24.82 crore" figure is therefore an estimate, not a survey result.

• Access, not use or quality

• A latrine counts even if nobody uses it.
• An LPG connection counts even if the family cannot afford refills and still cooks on wood.

• A bank account counts even if it is dormant (not used for a long time).

• Weights and cut-off are judgement calls

• Equal 1/3 weights and the 1/3 cut-off are choices, not scientific facts.

• Different weights would put different people in or out of poverty.

• Missing dimensions

• Monetary poverty is not included. The World Bank MPM does include it [6].
• Learning outcomes are left out. A child counts as attending school even if the child cannot read.
• The quality of care is also missing, for example whether hospitals and teachers are any good.

Prelims Hooks

• MPI = H × A. H is the headcount ratio (incidence) and A is the average intensity among the poor. The value lies between 0 and 1.
• The Global MPI (2010) is published by UNDP HDRO and OPHI. It replaced the Human Poverty Index (1997–2009).
• Alkire-Foster method: 3 equally weighted dimensions and 10 indicators (global). Health 2, education 2, living standards 6.
• Cut-offs: poor ≥ 1/3; severely poor ≥ 1/2; vulnerable 1/5 to 1/3.
• NITI Aayog's national MPI has 12 indicators. The two extra ones are maternal health and bank account. Data come from NFHS, not NSSO consumption surveys.
• NITI Aayog is the nodal agency for the MPI under the GIRG initiative.
• Baseline 2021 (NFHS-4): Bihar highest (51.91%), Kerala lowest (0.71%).
• 2015-16 → 2019-21: headcount 24.85% → 14.96%. MPI value 0.117 → 0.066. Intensity 47.14% → 44.39% [4].
• 2013-14 → 2022-23: 29.17% → 11.28%; 24.82 crore exited. UP had the most exits (5.94 crore) [2].
• Trap: the World Bank MPM includes monetary poverty ($3.00/day, 2021 PPP) and has no health dimension [6]. The UNDP-OPHI Global MPI has no income indicator.

Mains Points

• Income and MPI answer different questions.
• Poverty-line methods measure means: can the person afford a basket of goods?
• The MPI measures outcomes: does the person actually have food, schooling, water and a toilet?

• Using both gives a fuller picture. The World Bank MPM shows that adding non-money needs raises measured poverty by about 51% [6].

• MPI as a policy tool (GS-II, governance).

• Because the MPI can be broken down by indicator and district, it tells states which gap to close.
• Examples are cooking fuel, sanitation and nutrition.

• This supports cooperative and competitive federalism, and tracking of SDG 1.2 [4][5].

• Credibility debate.

• The steep fall since 2015-16 matches the spread of household schemes (toilets, LPG, electricity, bank accounts).
• But the index counts access, not use or quality.
• The 2022-23 figure is extrapolated.

• So an answer should credit the progress and also call for newer surveys and outcome-based indicators, such as learning levels and actual LPG refills.

• Intensity matters as much as headcount.

• Intensity is still about 44% (2019-21) [4]. So those who remain poor lack many things at once.
• Policy should focus on the hardest-to-reach households, not only on those just below the cut-off.