Multidimensional poverty is a condition in which people experience deprivation in several aspects of life, rather than solely a shortage of income or consumption. It treats poverty as involving limited access to essential opportunities, resources, and living conditions, including adequate nutrition, education, housing, sanitation, and electricity. The concept is broader than any particular index: different measurement frameworks select different dimensions, thresholds, and rules for deciding who is poor.
Conceptual foundations
Monetary poverty measures usually identify people whose income or consumption falls below a poverty line. These measures provide information about command over economic resources. Multidimensional approaches examine whether people actually achieve minimum conditions in several areas of life.
The distinction matters because resources do not translate uniformly into outcomes. People with similar incomes may face different public services, local prices, disabilities, household responsibilities, or environmental conditions. A household may have income above a monetary poverty line yet lack safe sanitation or accessible schooling. Conversely, publicly provided services may reduce some deprivations even where household income remains low.
An important intellectual foundation is the capability approach, associated particularly with Amartya Sen. It evaluates well-being in terms of the substantive opportunities people have to live lives they have reason to value. Multidimensional poverty measures generally capture selected observable achievements or deprivations, rather than the full range of a person’s capabilities.
The concept also draws on basic-needs approaches, social indicators, and research on exclusion and material deprivation. It overlaps with studies of living standards and quality of life, but focuses specifically on shortfalls rather than the entire distribution of well-being.
Multidimensionality is distinct from the distinction between absolute poverty and relative poverty. A multidimensional measure can use fixed minimum standards, standards defined relative to prevailing social conditions, or a combination of both.
Dimensions and indicators
A dimension is a broad aspect of life, while an indicator is a specific variable used to represent it. Education, for example, may be represented by school attendance and completed years of schooling.
Common dimensions include:
- Health and nutrition: indicators of nutritional deprivation, mortality, or access to essential care.
- Education: attendance, educational attainment, literacy, or learning.
- Housing and basic services: housing materials, crowding, drinking water, sanitation, and electricity.
- Economic security: employment, income, assets, or protection against economic shocks.
- Participation and personal security: social inclusion, exposure to violence, or the ability to participate in community life.
Not every measure includes all these dimensions. Selection depends on the purpose of the measure, the population concerned, public deliberation, and available data. Child-focused measures may distinguish age-specific needs, while national measures may include employment or services that are especially relevant in their setting.
An indicator also requires a deprivation threshold. Years of schooling, for instance, becomes a deprivation indicator only after a minimum level is specified. Such thresholds are normative choices informed by legal standards, policy commitments, evidence, and social priorities.
Measurement approaches
Dashboards and overlapping deprivation
A dashboard reports several indicators separately. It preserves detail and avoids imposing a single weighting system, but may not show whether disadvantages affect the same people.
This distinction is central to multidimensional poverty. Separate statistics on sanitation and school attendance do not reveal whether the households lacking sanitation are also those whose children are out of school. Measuring overlapping deprivation requires information linked at the relevant individual or household level.
Some approaches identify anyone deprived in at least one dimension as poor, known as a union rule. Others require deprivation in every selected dimension, known as an intersection rule. Intermediate rules identify poverty when deprivation reaches a specified number or weighted share of indicators.
The Alkire–Foster counting method
A widely used framework is the Alkire–Foster method. In its commonly used counting form, it applies two kinds of cutoff: thresholds determining deprivation in individual indicators, and an overall cutoff determining multidimensional poverty.
For person and indicator , let if the person is deprived and otherwise. Let be nonnegative indicator weights summing to one. The weighted deprivation score is
A person is identified as multidimensionally poor when , where is the overall poverty cutoff.
If of people are identified as poor, the headcount ratio is
The intensity of poverty, , is the average weighted deprivation score among poor people:
The adjusted headcount ratio is
When no one is identified as poor, is defined as zero. In survey applications, these quantities generally incorporate sampling weights.
For example, if 30 percent of a population is poor and poor people experience, on average, 40 percent of weighted deprivations, then . This is not a poverty rate of 12 percent: the headcount ratio remains 30 percent.
Unlike a headcount alone, can register a reduction in deprivation among people who remain poor. It can also be decomposed by population subgroup, while indicator contributions show the composition of poverty among those identified as poor.
International and national measures
The global Multidimensional Poverty Index, introduced in 2010 by the United Nations Development Programme and the Oxford Poverty and Human Development Initiative, is a prominent application of the Alkire–Foster method.
It uses three equally weighted dimensions—health, education, and living standards—represented by ten indicators:
- Health: nutrition and child mortality.
- Education: years of schooling and school attendance.
- Living standards: cooking fuel, sanitation, drinking water, electricity, housing, and assets.
Each health and education indicator has a weight of one-sixth, and each living-standards indicator has a weight of one-eighteenth. A person is identified as multidimensionally poor when the household’s weighted deprivation score is at least one-third. Because identification is household-based, members of a household share its poverty status under this measure.
National multidimensional poverty indices need not reproduce this design. They may use different indicators, weights, thresholds, and units of identification. Such adaptation improves relevance to national priorities but limits direct comparison between countries using different measures.
The World Bank also publishes a multidimensional poverty measure combining monetary poverty with selected education and basic-service indicators. It is distinct from the global Multidimensional Poverty Index.
Multidimensional poverty measures also differ from the Human Development Index, which summarizes average achievements in health, education, and income rather than identifying people experiencing overlapping deprivation.
Uses in research and public policy
Multidimensional measures can reveal disadvantages that national income averages, including gross domestic product, do not capture. Their uses include:
- Monitoring change: distinguishing changes in the number of poor people from changes in the intensity of their deprivation.
- Geographic and demographic analysis: identifying differences between regions, rural and urban populations, or population groups.
- Policy coordination: showing where education, housing, infrastructure, and social protection interventions address overlapping disadvantages.
- Budget analysis: informing discussion of whether public spending corresponds to observed deprivation.
- Programme assessment: tracking outcomes across several dimensions rather than income alone.
An index is descriptive, not by itself a demonstration of causation. A decline following a programme does not establish that the programme caused it; causal evaluation requires an appropriate research design.
Limitations and methodological debates
Normative choices and compensability
Dimensions, weights, and poverty cutoffs affect both who is classified as poor and the measured level of poverty. Equal weighting is transparent but is still a substantive choice. Results can be tested across alternative specifications to assess their robustness.
An additive deprivation score allows different combinations of disadvantages to yield the same total. This may obscure distinctions between severe deprivation in one area and several less severe shortfalls. Measures can instead impose additional requirements or report critical indicators separately.
Information loss
A single index simplifies communication but cannot represent every deprivation profile. Two populations can have the same index value while differing in which needs are unmet and how deprivations are distributed among poor people. The adjusted headcount ratio also excludes the deprivations of people below the overall poverty cutoff from its aggregate value.
For these reasons, an index is usually interpreted alongside its headcount, intensity, indicator breakdowns, and subgroup results.
Household assumptions and data constraints
Assigning one poverty status to all household members can conceal inequalities within households. Access to food, education, assets, or decision-making may differ by age, gender, or disability. Individual-level measures can reveal some of these differences, but require suitable data.
Household surveys may omit people without stable housing or underrepresent populations affected by conflict or displacement. Indicators may measure access rather than service quality: school attendance does not necessarily demonstrate learning, and an electricity connection does not establish reliable supply.
International comparisons are also affected by survey dates, questionnaire differences, missing observations, and changes in indicator definitions. Combining indicators from unrelated data sources without observing their joint distribution generally cannot establish who experiences overlapping deprivation.
Relationship to monetary poverty
Monetary and multidimensional measures answer related but different questions. Their poor populations need not coincide. Reporting them together can distinguish people facing inadequate economic resources, selected nonmonetary deprivations, or both. Neither approach fully replaces the other.