E-Paper
A richer measure of poverty: Why income alone is misleading
In the same way that GDP measures are flawed because they say little about our true state of wellbeing, income-based poverty measures miss a lot
Prefer HTon Google
Imagine a worker at a brick kiln in eastern Uttar Pradesh who is the sole wage-earner in a family of five. Although he earns enough to keep the household’s spending above the World Bank’s extreme poverty line of $3 per day, a look inside his house would show that his wife and children are underweight, the family cooks on a wood stove rather than on LPG, they share a latrine with neighbours and live under a thatched roof — all signs that we would associate with poverty. When we look closely, income-based poverty measures of the kind often reported miss a lot about the reality of people’s lives.
To address this concern, the Oxford Poverty and Human Development Initiative and the United Nations Development Programme developed the Multidimensional Poverty Index (MPI) that has been applied globally. Rather than focussing on income, the MPI looks at 10 indicators of health, education and living standards. The index works like a scorecard with 18 points. Each of the four health and education indicators is worth three points, and each of the six living-standard indicators is worth one. A household that scores six or more is counted as poor. In the case of the family in UP, the family loses three points because the mother is underweight, and one point each for cooking fuel, sanitation and housing. That makes six, exactly the poverty line, with any more deprivations exacerbating their poverty status. If the family is able to build a toilet with help from the government, its score would drop to five and it would be counted as having left poverty, even if the other deprivations are unchanged.
In the same way that GDP measures are flawed because they say little about our true state of wellbeing, which also depends on whether education and health are affordable, the air and water are clean, and if the streets are safe to walk on, income-based poverty measures miss a lot. The gap between income poverty and MPI can be high. About 5% of Indians live below $3 a day, but 16.4% are multidimensionally poor according to the 2026 global MPI. In Kenya, the opposite is true. 45.5% of Kenyans fall below the income line, yet only 25.4% are multidimensionally poor, because many households with little cash still have schooling, water and clinics within reach.
In truth, MPI is more amenable to government intervention. In communities where toilets, electricity connections, schools and nutrition programmes are provided, MPI can decrease even if there is no observable change in household income. Economic growth helps reduce multidimensional poverty, but the link is weaker than many assume. A cross-country study by María Emma Santos and colleagues found that GDP growth did reduce income poverty, its effect on MPI was much less impressive.
Within India, GDP growth at the state level between 2005-06 to 2015-16 was barely correlated with reductions in state-level MPI, according to Oxford researchers Sabina Alkire, Christian Oldiges and Usha Kanagaratnam. On NITI Aayog’s national measure which uses 12 indicators and differs from the global MPI measure, 11.7% of people in Gujarat, one of India’s richest large states, experienced multidimensional poverty in 2019-21, well above Andhra Pradesh at 6.1%, Haryana at 7.1% and Karnataka at 7.6%. In contrast, the MPI rate in Kerala was below 1% by 2015-16, while it was 2.2% in Tamil Nadu in 2019-21. Although these two states are not the richest in the country, they have long invested heavily in better schools, healthcare, public distribution, and local governance.
On the positive side, the greatest reductions in multidimensional poverty were achieved in the poorest states. Between 2015-16 and 2019-21, Bihar’s headcount fell by about 18 percentage points, from 52% to 34%, the largest decline of any state. Madhya Pradesh and Uttar Pradesh followed, with falls of about 16 and 15 points. The effect sizes were actually more impressive than these numbers indicate. In Bihar, the average deprivation among those who remained poor also fell, from 51% to 47%. Earlier work on 2005-06 to 2015-16 showed that by headcount the poorest states seemed to lag, but by MPI value they progressed fastest, because the intensity of deprivation among their poorest residents fell quickly.
However, there are exceptions. Meghalaya’s headcount fell only from 33% to 28%, and the intensity of poverty among its poor barely changed. Punjab, a prosperous state, moved only from 5.6% to 4.8%. Despite these improvements, Bihar still has a third of its people in multidimensional poverty, and the five most populous states account for just over half of India’s multidimensionally poor.
Now let’s look at the national level. According to the World Bank’s October 2026 estimates, 2.6% of Indians were below the poverty line of $3 a day and 16.1% below the $4.20-a-day line, using 2023 data. Multidimensional poverty has fallen from 55% in 2005-06 to 28% in 2015-16 and 16% in 2019-21. The rate of reduction has been roughly 2.5 to 2.8 percentage points a year in both periods, and there was no slowdown before the pandemic.
The experience of other countries shows that it is possible to reduce MPI quite dramatically. China halved its MPI value between 2010 and 2014, while Bangladesh’s MPI rate fell from 31% to 12% between 2014 and 2022. But progress has not been uniform. Mexico’s MPI rate has stagnated between 4% and 6% for more than a decade, and Nigeria went backwards between 2021 and 2024, adding nearly 18 million people to its multidimensional poor.
How do we achieve further decreases in MPI? In 2019-21, the largest deprivations among India’s poor were in cooking fuel, housing, nutrition and sanitation. Nutrition remains an important part of the unfinished agenda both because it counts for a lot and declines relatively slowly compared to some of the other measures. The NFHS-6 data released so far show stunting falling from 35.5% to 29.3%, but the share of underweight children has barely changed. Addressing nutrition has to be front and center of any policy. Providing better health and education have to receive a much higher priority that is currently the case.
Big reductions in MPI are within our reach, a more meaningful measure of progress which sadly gets much less attention in the media than how many billionaires India has created. A focus on reaching migrant workers, remote hamlets across the country where the delivery of government services is still weak, and where ill health is a major reason to fall into poverty would do a lot to reduce multidimensional poverty. A target of cutting the MPI rate in every state to below one percent should be part of any conversation on Viksit Bharat.
Ramanan Laxminarayan is the author of Bacteria to Buddha: How Cooperation Shaped Consciousness — and Makes Us Human. The views expressed are personal
(This is a Hindustan Times digital exclusive)
