Surjit Bhalla writes: World Bank's warped inequality numbers

Indian Express

16 hours ago

The facts are clear and unambiguous — consumption inequality in India, as measured by the Gini index, was the lowest (most equal) in the world in 2022-23. However, the debate on this is messy and bordering on sordid. What is interesting is that all the important participants in the discussion/debate (Government of India, World Bank, and Indian media) have got it wrong in one way or another. This is not idle armchair criticism, as I document below.
It all started with an essay published on the ORF website which says that, 'contrary to popular narratives, India is not merely the world's fourth-largest economy; it is also the world's fourth most equal society'. Unfortunately, this is false on both counts. India will not be the fourth-largest economy at least until March 2027; and India today is the most equal society — not fourth most equal — but only in terms of consumption. Its rank in terms of an income Gini is not known since India, to date, has not conducted an official income distribution survey.
The government has endorsed these wrong 'findings' through a PIB release; the BJP also wrongfully stated that India was the fourth most equal society — it only differed in stating that we ranked fourth in terms of social equality (whatever that index might be). The endorsers completely ignored the important, and correct, result that India was most equal only in terms of consumption inequality. Critics were quick to pounce, and correctly so, on the errors of interpretation. Which led to them compounding the error. And the debate goes on.
For example, Surbhi Kesar of SOAS has rightly claimed that the conclusion that India is amongst the most equal countries (as published in the GOI PIB report) was incorrect, since it compared apples with oranges or mixed consumption distributions with income distributions. 'A fair comparison would either be to compare India's income inequality with other countries' income Ginis, or compare India's consumption inequality with other countries' consumption Ginis.'
This makes perfect sense. Unfortunately, after outlining the right procedure, Kesar makes the same mistake as others. She chooses to use a third source of data on distribution — not consumption, not income, but synthetic estimates of income distribution as published by the private-sector World Inequality Database (WID). WID constructs estimates by its own assumptions — a new literature has developed in leading American economic journals about the unrealistic and questionable nature of WID's synthetic estimates for the US. Inequality has always been discussed in terms of either the distribution of consumption or of income. Everyone recognises — from survey organisations to policy makers to academicians — that the measurement of distributions is flawed, imperfect, imprecise. All recognise that household surveys are the best instrument for gauging inequality in a society.
Towards this end, the World Bank has laboured to construct the popular and respected Poverty Inequality Platform (PIP), which provides data for all the official consumption and income surveys in the world. PIP reports data on 167 countries, 2,258 distributions for years 1963 to 2024. These data are the only 'official' source of comparable income and consumption distributions. 'Official' because governments provide unit-level data to the World Bank for analysis by its army of researchers and analysts. There is only one exception to this rigid, and correct, World Bank requirement: China only provides summary ventile (5 percentile) distributions to the World Bank for analysis and subsequent publication. Perhaps BRICS can 'influence' China to comply with the open-data requirements of international organisations!
The PIP data does not contain any information on income distribution for India or South Africa — for good reason, because no official income survey exists for either of these countries.
According to PIP, the most unequal (consumption) country in the world, at least for the last 25 years, is South Africa with a Gini of 63.0. The most equal is India with a Gini of 25.5 in 2022.
Notwithstanding its own findings, and the lack of any official data, the World Bank felt compelled to gratuitously note in its April 2025 Poverty and Equity Brief for India: 'India's consumption-based Gini index improved from 28.8 in 2011-12 to 25.5 in 2022-23, though inequality may be underestimated due to data limitations. In contrast, the World Inequality Database shows income inequality rising from a Gini of 52 in 2004 to 62 in 2023.' The briefs report summary data for 130 developing countries, but only for India does the World Bank quote WID synthetic results on income distribution. To add insult to statistical injury, the World Bank believes that official surveys have data limitations (yes they do), but WID synthetic distributions have no data limitations! In India, we call this kind of inference the advent of Kalyug, the 'age of darkness'.
The World Bank has been a pioneer in the collection and verification of consumption and income distribution data since its creation. The PIP database is testimony to its intellectual honesty and expertise. However, the World Bank states that India has the lowest consumption inequality in the world; in the same breath (or paragraph) it states (via its endorsement of the WID synthetic results) that India has the highest income inequality. Such a paradoxical result has heretofore not been documented by any organisation in the world, let alone by a well-known and internationally renowned NGO. One of these estimates is clearly in huge error.
The memory banks of the World Bank will reveal that its own research (some 25 years ago) had documented that income Ginis were, on average, six Gini points higher than consumption Ginis. Would the World Bank care to explain how it finds credible an income WID Gini of 62 along with a consumption Gini of 26? What do these 'facts' say about the World Bank's credibility? About WID's credibility? The whole world is watching, and awaiting, a meaningful response from sister organisations on this important credibility question.
The writer is chairperson, Technical Expert Group for the first official Household Income Survey for India. Views are personal