Frequently Asked Questions - Multidimensional Poverty Index (MPI)

The Multidimensional Poverty Index (MPI) identifies multiple deprivations at the household and individual level in health, education and standard of living. It uses micro data from household surveys, and—unlike the Inequality-adjusted Human Development Index—all the indicators needed to construct the measure must come from the same survey. Each person in a given household is classified as poor or non-poor depending on the number of deprivations his or her household experiences. These data are then aggregated into the national measure of poverty. The MPI reflects both the prevalence of multidimensional deprivation, and its intensity—how many deprivations people experience at the same time. It can be used to create a comprehensive picture of people living in poverty, and permits comparisons both across countries, regions and the world and within countries by ethnic group, urban or rural location, as well as other key household and community characteristics. The MPI offers a valuable complement to income-based poverty measures.
The 2015 Human Development Report (HDR) presents estimates for 101 developing countries with a combined population of 5.0 billion (75% of the world total). About 1.5 billion people in the countries covered—29% of their entire population—lived in multidimensional poverty between 2005 and 2014. We could not include other countries due to data constraints. Comparable data on each of the indicators were not available for other developing nations. There was also a decision not to use data from surveys conducted earlier than 2005.

The MPI identifies overlapping deprivations at the household level across the same three dimensions as the Human Development Index (health, education and living standards) and shows the proportion of poor people and the average number of deprivations each poor person experiences at the same time. For more details see Technical note 5, http://hdr.undp.org.

The MPI reflects the severe deprivations that people face at the same time. We have described the MPI as a measure of “acute” because it reflects overlapping deprivation in basic needs and also to avoid confusion with the World Bank’s measure of “extreme” poverty that captures those living on less than $1.90 (in 2011 $PPP) a day.

The critical review of the family of human development indices including the MPI was conducted during the two conferences on measuring human progress organized by the UNDP in February 2012 and March
2013. As an outcome of these critical reviews a number of adjustments of the MPI were made. They are justified on the grounds of being more in line with the MDGs. At the same time Oxford Poverty and Human Development Initiative (OPHI) has published their own estimates using their original 2010 specifications.

The difference between the two approaches stands in the definition of deprivations for several indicators. A household is now deprived in school attainment if no household member has completed 6 years of education (previously it was 5). Six years is the duration of primary education in most countries, so this change reinforces MDG 2 “Universal primary education.” School attendance – we allow a child of school-entry age one year late enrollment to avoid coding as deprivation a mismatch between the birthdate and the school start date. Further, in the health dimension, for nutrition – a household is deprived if there is a stunted child (instead of underweight child). Because, if a child is stunted, the damage is mostly irreversible. As Anthony Lake of UNICEF described it: “That child will never learn, nor earn, as much as he or she could have if properly nourished in early life.”
Similarly, the child mortality ‘experienced’ in the household is considered a deprivation in the health dimension if it has occurred within 5 years before the survey. Previously, there was no limit. This change captures recent improvements in child mortality. We also added ownership of arable lands and livestock into the living standard dimension to better capture rural poverty. Further details about the revised HDRO’s specifications can be found in the methodological note: http://hdr.undp.org/en/content/undp%E2%80%99s-multidimensional-poverty-i...

The MPI constitutes a set of poverty measures. These measures are explained as follows. Incidence of multidimensional poverty: the proportion of people who are poor according to the MPI (those who are deprived in at least one third of the weighted indicators). Average intensity of poverty: the average number of deprivations poor people experience at the same time. MPI value: The MPI value summarizes information on multiple deprivations into a single number. It is calculated by multiplying the incidence of poverty by the average intensity of poverty. These measures can be unpacked to show the composition of poverty both across countries, regions and the world and within countries by ethnic group, urban/rural location, as well as other key household and community characteristics.

One deprivation alone may not represent poverty. The MPI requires a household to be deprived in multiple indicators at the same time. A person is multidimensionally poor if she/he is deprived in one third or more of the weighted indicators. We also count persons who live near multidimensional poverty as those who are deprived in one fifth or more but less than one third of the weighted indicators. Those who are deprived in one half or more are considered living in extreme multidimensional poverty.

We could not include income due to data constraints. Income data come from different surveys, and these surveys often do not have information on health and nutrition. For most countries we are not able to identify whether the same people are income poor and also deprived in the MPI indicators.

We could not include empowerment due to data constraints. The Demographic and Health Surveys (DHS) collect data on women’s empowerment for some countries, but not every DHS survey includes empowerment, and the other surveys do not have these data. Data on men’s empowerment or political freedom are missing.

The MPI relies on two main databases that are publicly available and comparable for most developing countries: the ICF Macro Demographic and Health Survey (DHS) and the UNICEF’s Multiple Indicators Cluster Survey (MICS). For several countries, national household surveys with the same or similar content and questionnaires are used - Argentina, 2005 Encuesta Nacional de Nutrición y Salud (ENNys); Brazil, 2013 and 2012 Pesquisa Nacional por Amostra de Domicílios (PNAD); China, 2012 China Family Panel Studies; Ecuador, 2014 and 2006 Encuesta de Condiciones de Vida (ECV); Jamaica, 2010 Jamaica Survey of Living Conditions (JSLC); Libya, 2007 Pan Arab Population and Family Health Survey (PAPFAM); Mexico, 2012 and 2006 Encuesta Nacional de Salud y Nutricion (ENSANUT); Morocco, 2011 Pan Arab Population and Family Health Survey (PAPFAM); State of Palestine, 2006/2007 Pan Arab Population and Family Health Survey (PAPFAM); Syrian Arab Republic, 2009 Pan Arab Population and Family Health Survey (PAPFAM), and South Africa, 2012 and 2008 National Income Dynamics Study (NIDS). Tables 6 and 7 of the Human Development Report Statistical Annex indicate for each country if data come from the DHS, MICS or from a national survey.

The MPI relies on the most recent and reliable data available since 2005. The difference in dates limits direct cross-country comparisons, as circumstances may have improved, or deteriorated, in the intervening years. This is the reason why we do not rank countries based on the MPI value.

The MPI complements income poverty measures. It measures various deprivations directly. In practice, although there is a clear overall relationship between MPI and $1.25 per day poverty, the estimates do differ for many countries. This is a topic for further research, but some factors can include provision of public services, as well as different abilities to convert income into positive outcomes such as good nutrition.

The MPI, like the $1.25 per day line, is a globally comparable measure of poverty. It measures acute multidimensional poverty, and only includes indicators that are available for many countries. National poverty measures are typically monetary measures, and thus capture something different. The fact that there are differences does not mean that the national poverty number, or the MPI headcount is wrong— these simply measure different concepts of poverty. At the same time, just as national poverty measures, in contrast, are designed to reflect the domestic situation more accurately and often differ in very useful ways from the $1.25 measure, some countries may wish to build a national multidimensional poverty index that is tailored to their context, to complement this international MPI.

The MPI methodology shows aspects in which the poor are deprived and helps to reveal inter-connections among those deprivations. This enables policymakers to target resources and design policies more effectively. This is especially useful where the MPI reveals areas or groups characterized by severe deprivation.

The multidimensional poverty approach can be adapted using indicators and weights that make sense at the country level to create tailored national poverty measures. The MPI can be useful as a guide to help governments tailor a poverty measure that reflects local indicators and data. In 2009, Mexico became the first country to adopt a multidimensional poverty measure reflecting multiple deprivations at household level.

The MPI methodology can and should be modified to generate national Multidimensional Poverty Measures that reflect local cultural, economic, climatic and other factors. The global MPI was devised as an analytical tool to compare acute poverty across nations.

The MPI has some drawbacks, due mainly to data constraints. First, the indicators include both outputs (such as years of schooling) and inputs (such as cooking fuel) as well as one stock indicator (child mortality). Second, the health data are relatively weak and overlook some groups’ deprivations especially for nutrition, though the patterns that emerge are plausible and familiar. Third, in some cases careful judgments were needed to address missing data. But to be considered multidimensionally poor, households must be deprived in at least six standard of living indicators or in three standard of living indicators and one health or education indicator. This requirement makes the MPI less sensitive to minor inaccuracies. Fourth, intra-household inequalities may be severe, but these could not be reflected. Fifth, while the MPI goes well beyond a headcount to include the intensity of poverty experienced, it does not measure inequality among the poor, although decompositions by group can be used to reveal group-based inequalities. Finally, the estimates presented here are based on publicly available data and cover various years between 2005 and 2014, which limits direct cross-country comparability.

Yes. The global MPI estimates are constrained by need for comparability. National teams should use the indicators and weights that make sense. At the country level, however, the multidimensional poverty approach to assessing deprivations at the household level can be tailored using country-specific data and indicators to provide a richer picture of poverty at the country level.

We estimated the MPI for two or more data points for 61 countries for which suitable data were available. It seems that the MPI can be used to study the changes in poverty pattern over time providing that the data were available from the same survey conducted at different years. We advise the reader to carefully interpret the changes over time for a particular country because different indicators could be missing from the survey at different points in time (i.e. for Brazil, the 2006 data lacks information about cooking fuel while the 2012 and 2013 data lack information about nutrition and type of floor).

The effects of shocks are difficult to capture in any poverty measure. Because the standard survey data used to estimate the global measure are collected infrequently, the ability to detect changes is limited by the available data source. The MPI will reflect the impacts of shocks as far as these cause children to leave primary education or to become malnourished, for example. If more frequent data are available at the country or local level, this can be used to seek to capture the effects of larger scale economic and other shocks.

Since the 2014 HDR, all statistical programmes used to calculate the MPI are available in the HDRO’s website http://hdr.undp.org/en/content/mpi-statistical-programmes. Also, DHS and MICS data are publicly available online. Therefore, national governments, civil societies, and research communities can replicate the MPI results as well as adapt the programmes to their own country-specific poverty needs.