Poverty Measurement — Explained
Detailed Explanation
What is being measured when we count poverty?
Poverty measurement converts a social condition into an explicit rule for identifying deprivation. The rule matters. A monetary measure asks whether income or consumption falls below a specified threshold. A multidimensional measure asks whether people experience a weighted combination of deprivations such as inadequate nutrition, schooling or living conditions. These approaches illuminate different aspects of hardship; their percentages are not interchangeable.
Begin any comparison with four questions: who is the unit, what information is observed, where is the threshold and when was the observation made? A household expenditure survey, a health survey and an extrapolated estimate can all inform policy, but they do not provide identical evidence. Treating their publication dates as their observation dates is a common source of misleading claims.
What does a headcount ratio show—and omit?
The headcount ratio is the share of people classified as poor under the chosen rule. In an illustrative population of 1,000, if 200 fall below the monetary threshold, the headcount ratio is 20 per cent. If their incomes fall further while the same 200 remain below it, the ratio stays unchanged even though deprivation deepens.
A poverty-gap measure addresses the shortfall below the line rather than merely counting people on either side. Distribution within the poor population also matters: equal average shortfalls can conceal different concentrations of severe deprivation. The policy implication is straightforward. A falling headcount is useful evidence, but it should be read alongside depth, vulnerability and differences across social groups and places.
How does India's national MPI work?
NITI Aayog's 2023 National Multidimensional Poverty Index uses three dimensions—health, education and standard of living—and twelve indicators. The dimensions receive equal weight; this does not mean that all twelve indicators receive equal weight. A person is identified as multidimensionally poor when the household's weighted deprivation score reaches at least one-third.
The index combines H, the incidence or headcount ratio, and A, the average intensity of deprivation among people identified as poor: MPI = H × A. If 30 per cent of people are poor and their average weighted deprivation is 40 per cent, the illustrative MPI is 0.30 × 0.40 = 0.12. It is not 70 per cent and not a claim that 12 per cent of the population is poor. The multiplication adjusts the headcount for intensity.
What do the 2023 report's changes actually compare?
The national report compares NFHS-4, conducted in 2015–16, with NFHS-5, conducted in 2019–21. It reports a decline in the multidimensional headcount ratio from 24.85 per cent to 14.96 per cent. Those are survey-period observations under the report's framework, not annual poverty estimates for 2023.
Subtracting the two percentages gives 9.89 percentage points. Calling this a 9.89 per cent relative reduction would be incorrect: the relative reduction divides the change by the original 24.85. An answer should name the measure before discussing progress. It should also examine which indicators and locations improved, since a national average can coexist with persistent deprivation in particular communities.
Why must the 2024 discussion paper be labelled differently?
NITI Aayog's January 2024 discussion paper provides estimates for 2013–14 and 2022–23 using interpolation or extrapolation. It gives estimated multidimensional headcount ratios of 29.17 per cent and 11.28 per cent respectively. These are not direct NFHS surveys carried out in those two years.
An estimate can be informative without being a direct observation. Its interpretation depends on the trend assumptions used to extend the underlying series. A student should therefore write “the discussion paper estimates” rather than “the 2022–23 NFHS found.” This distinction becomes particularly important when discussing disruptions or policy changes occurring outside the survey reference period. A projection cannot independently establish their realised effects.
How do international poverty lines differ from a market exchange rate?
The World Bank's June 2025 update set the international extreme-poverty line at USD 3.00 per person per day in 2021 international dollars, replacing USD 2.15 in 2017 purchasing-power-parity terms. Purchasing power parity accounts for differences in prices across economies; it is not the exchange rate used to buy dollars for travel.
Consequently, multiplying three dollars by today's rupee–dollar market rate does not produce the appropriate Indian poverty threshold for this exercise. Nor does the numerical increase from 2.15 to 3.00 prove that a household's real consumption changed by that proportion. Updating the price basis and benchmark changes the measuring framework. Compare estimates constructed consistently under the same framework before claiming a trend in living standards.
Can better infrastructure reduce multidimensional poverty without equal income growth?
Yes, the indicators can respond differently. Consider an illustrative household that gains reliable sanitation and electricity while its cash income changes little. Its deprivation score can fall even though a monetary measure shows limited improvement. Conversely, higher income may not immediately secure access to a functioning school or health service.
This is why the measures should be used together. Public provision, household resources and local access constraints interact, but they are not identical. A programme evaluation should link the intervention to the relevant outcome and compare suitable groups or periods. A simultaneous national improvement does not, on its own, isolate one scheme's contribution from other economic and social changes.
How should an answer turn measurement into policy?
Use the measure to diagnose the problem, then choose the response. Large nutrition deprivation requires a different operational inquiry from low school attendance or inadequate housing. A high headcount with moderate intensity may require a different mix of broad services and targeted support from a smaller population facing multiple severe deprivations.
Conclude with measurement discipline: disclose the source, survey period, threshold, unit and whether the figure is observed or estimated. Monetary poverty, multidimensional poverty and vulnerability to future poverty are related questions rather than competing slogans. A credible answer recognises improvement where the evidence supports it while retaining enough detail to identify who remains deprived and what public action would address the deprivation.