All the Reasons Why the World Bank's Claims on Falling Inequality Are Wrong – Especially in India

Those rejoicing on discovering that India is one of the most ‘equal’ countries in the world – fourth from the top – should understand the difference between income and consumption surveys.

For India's figures, the World Bank has relied on India’s consumption survey of 2022-23 to arrive at India’s Gini score of 25.5, which is low. However, by consumption surveys, which have been done in India since 1977, the inequality in India has always been very low. More on this below.

However, first, let us understand the differences between income-based and consumption-based household surveys. Income-based surveys  are used mostly in high-income countries and in the Latin America and Caribbean, where governments typically conduct income-focused household surveys. Examples include OECD nations or data from the Luxembourg Income Study. Income Gini estimates, especially in wealthier economies, come largely from these income surveys. The reason for doing income (as opposed to consumption surveys) is that at least the rich countries all have highly formalised economies, and income estimates have credibility; in Latin America, too, the economies are relatively more formalised (though less so than in OECD countries). Thus, in Latin America, the Economic Commission for Latin America has published annual survey data on income inequality – Social Panorama – since the 1990s.

By contrast, consumption-based surveys  are common in low- and lower-middle-income countries, particularly across Sub-Saharan Africa, South Asia, and the Middle East and North Africa. Here, household surveys focus on expenditures – food, utilities, rent, own production – and the Gini is derived from consumption rather than income.

Why the difference in type of survey matters

Income-based Gini tends to be higher, since incomes are more unequally distributed than consumption (the rich save more). Consumption-based Gini is typically lower, understating inequality in regions where consumption data are used.

Therefore, direct comparisons across regions can be misleading – e.g., Sub-Saharan Africa’s inequality may seem lower than it really is when compared to Latin America. Because of the different data bases, analysts should note whether an inequality estimate is consumption-based or income-based before making cross-country or regional comparisons. If I were to advise the World Bank on this matter, I would say to them that they are reducing their own credibility by presenting league tables of countries that compare apples and oranges. That results in misleading conclusions to be drawn by readers of such data/analysis.

India, with per capita income or PCI at market exchange rates of $ 2500 per year, is among the vast majority of Low Middle Income Countries (LMICs). The High Income Countries (HIC) normally use an income survey (those with PCI of $ 14,005 per year, while the Upper Middle Income Countries (UMICs) use a mix of consumption or income surveys.

Also read: No, India Is Not the Fourth Most Equal Country. Here’s the Real Data

Three reasons income and consumption surveys are not comparable

There are essentially three reasons why income surveys show a higher income inequality than consumption inequality. The first reason is the effects of savings and wealth.

Thus, higher-income households save more, so consumption understates true income at the top. Consumption tends to smooth out fluctuations, making its distribution appear more equal. Academic literature has established this as a fact.

Second, there is top income underreporting. Thus, household surveys often miss the truly wealthy (underreporting or non-response), especially in income data, which biases Gini downward. Correcting with tax or national-account data typically increases measured inequality.

Thirdly, there are conceptual differences between the two types of surveys. Income includes all sources (capital gains and business earnings), while consumption is what is actually spent. Income-based Gini captures income from wealth and assets, making it inherently more unequal.

There are four types of inequality in any society. The World Bank estimate is only capturing one of them.

The first is social inequality (e.g. caste based in our society, between ethnic groups in other societies), which are endemic in our societies. These take decades to eliminate, even with well-meaning affirmative actions. This usually manifests itself in health, educational, and nutritional indicators between social groups – which remain abiding in India, despite all the affirmative action.

The second is income inequality, which is what is often the cause of the third type, wealth inequality. We will discuss both in turn, given that usually income is an important wealth indicator, though of course incomes are also generated when wealth gets transferred through inheritance.

It is only consumption inequality, that, as we noted above, is what is usually quite equally distributed.

In fact, from 1977 onwards, when Consumption Expenditure Surveys began in India, consumption inequality has demonstrated very low levels of inequality. So the .25 Gini for consumption that the World Bank finds in 2022-23 is nothing out of the ordinary. The consumption Gini in India has hovered around this level or slightly higher for decades. It might appear slightly lower, on account of government programmes such PM KISAN (a Union government cash transfer to owner cultivators in rural India) or state level cash transfers of all varieties, or the free wheat or rice that has been available under the Public Distribution System (it was always available at Rs 2 or 3 per kg for the two cereals, and for the last five years, it has been free).

However, the inferences being drawn about the World Bank analysis are flawed for at least four sets of other reasons as well: the first relates to rising wealth inequality; the second relates to rising income equality (as estimated from other sources); third, the growing evidence about rising concentration in India’s industry/service sectors; and fourth, the overwhelming evidence that is emerging about job growth as well wages. We shall discuss each in turn.

Wealth and Income Inequality

Here’s a refined summary of the 2024-25 findings by Piketty, Chancel, Somanchi and Bharti on wealth and income inequality in India (up to 2022–23), based on their working paper 'Income and Wealth Inequality in India, 1922–2023: The Rise of the Billionaire Raj'.

In the long-run trend (1922–2023) they find a post-Independence decline in inequality until the early 1980s, followed by a dramatic rise from the 2000s onward. Top-end inequality trends of income and wealth have moved in tandem since the 1960s.

The top 1% share of pre-tax income, at 2022-2023, at 22.6%, is among the highest globally, exceeding US, Brazil, and South Africa. The top 1% share of wealth is at 40.1%, up from around 12.5% in 1980. The top 0.1%, 0.01%, and 0.001% control around 30%, 22%, and 17% of overall wealth, respectively. The latter group (fewer than 10,000 people) hold nearly three times the wealth of India’s bottom half (around 460 million people).

The data sources include national accounts, tax records (since 1922), household surveys (NSSO or AIDIS), and Forbes and Hurun rich lists. The important point is that their estimates are likely conservative: wealth data excludes offshore assets like those held in Dubai, and poor-quality data suggests actual inequality is even higher.

The authors suggest implementing tax reforms, including a “super tax” on top billionaires (~2% tax on net wealth of the wealthiest 167 families), which could raise ~0.5% of national income to fund social investments. They also advocate for comprehensive wealth and inheritance taxes and larger public spending on health, education, and nutrition.

These estimates have prompted comparisons between India’s current “Billionaire Raj” and the former British Raj, with India now surpassing peers like Brazil and the US in income/wealth inequality. Rightly, Piketty has urged the Indian government to reinstate wealth and inheritance taxes – citing the top 1% owning more than in Western countries – though officials argue this could trigger capital flight. The capital flight argument is a red herring; if the data of the last 10 years is any guide, the absence of wealth and inheritance taxes has not prevented 23000 High Net Worth Individuals and Ultra High Net Worth Individuals leaving India for tax havens or other countries.

Growing concentration in India’s productive sectors

Both economic growth and the manufacturing narrative have been vitiated by the rising power of a limited number of large corporates in India – precisely at a time when formal job creation has stagnated. In 2023, Viral Acharya and Rahul Chauhan showed how market concentration has grown in India since 2015, and this would have further entrenched the capital intensity and import-intensity of formal manufacturing, as well as producer/consumer service sectors – with both impacting formal job creation adversely.

Acharya and Chauhan alone with S.J. Estrin Commander together establish some rather devastatingly damaging facts for the Indian growth ‘narrative’. They show that industrial concentration fell after the 1991 liberalisation but rose after 2015.

The Big Five, Reliance, Tata, Aditya Birla Group, Adani, and Bharti Telecom, have increased their shares horizontally across sectors, and then within those sectors deepened the concentration through mergers and acquiring other firms, even gaining at the expense of the Big Six to 10. They also show their contribution to core inflation was driven up by this greater market power, as the sales over variable costs for the Big Five fell from 1990 to 2011 but have risen since.

Moreover, concentration has gone hand in hand with the fall in the investment-to-operating income ratio. The entry of firms is shown to fall in that industry. This entry is more restricted where the Big Five have a larger share – meaning that competition fell.

The authors merge data on family owned companies to create a list of the top five Big family groups. In this list are four of the five above, excepting Bharti Telecom, but with O.P. Jindal group. These five account for 10% of sales, compared to 3% in developed countries. Adani had been present in only five sectors in 2000 but expanded to 38 by 2020, Jindal from nine to 37 sectors, Tatas from 43 to 69 sectors; and Birla from 44 to 65 sectors. They also show that rising concentration reduced firm entry into these sectors, and the biggest five raised their mark-ups from 1.45 in 2013 to 1.7 in 2020.

In one word, under the circumstances, should we be surprised if manufacturing contribution to the GDP has fallen, and manufacturing employment share in total jobs is not rising? Most of the manufacturing employment growth (in absolute terms) has been concentrated in the unorganised sector.

Also read: India's Poverty Decline Is Not What it Seems

More reasons

There is hard evidence that poverty fell very sharply between 2004 and 2014, while since 2015,  although it is likely that there is continuing fall in poverty it is nowhere close to the fall that occurred in the preceding 10 years. Here are the reasons why poverty falls would be slower since 2015. All references to data in this section of the piece are from my forthcoming book, India Out of Work, Bloomsbury, London, 2025.

The first is  the difference in GDP growth rates between the two periods. The growth rate was faster from 2004-14 due to higher savings and investment rates. Tthe savings rate had begun to rise since the demographic dividend set in in the early 1980s. Rising savings/GDP ratio (and corresponding growth in the investment to GDP rate) was the basis for the ensuing rise in the GDP growth rate. Thus by 2003-4 the savings rate had risen to 23% of GDP (and investment to 24% of GDP). However, appropriate macro-economic policies enabled this to be to translated into the investment to GDP between 32 and  38% over 2004-2024. This was the highest ever that India had (or has) achieved. 

Not only did the GDP growth rate fall to 5.8% pa over 2014-24, the investment rate has ranged between 26 and 32% pa over 2014 and 2024 (primarily as the savings rate has also fallen).

Second, the pattern of growth has varied. Tthe earlier period’s growth encompassed all sectors -  the unorganized and organized. Not surprisingly, aggregate demand was sustained, as all growth engines were firing (public and private investment, final consumption, exports, and government expenditure). Hence non-farm jobs grew at a rate of 7.5 million per annum, itself unprecedented. Except agriculture (where workers fell, a good thing), all sectors generated jobs, especially Construction (at least 26 million rise) and manufacturing (at least 8 mn rise) . 

By contrast, wanton policy-induced shocks to the economy  (demonetisation, poorly designed GST, national lockdowns) since 2016 caused unorganised and MSME sector jobs to collapse. 

Third, there is a real difference in job growth between the two periods. For the first time in India’s post-independence history, the absolute number of workers on farms actually fell after 2004, as non-farm job growth was high (7.5 million per annum). This had the effect of tightening the labour market in rural areas till 2014. The exact opposite happened after 2020, with 80 million workers added to agriculture, partly due to reverse migration. 

Fourth, the new non-farm jobs, and tightening rural labour market   raised real wages – which rose all the way till 2015. Finally, as real wages rose, private final consumption expenditure rose, especially of simple consumer goods. For the first time in India’s history, the absolute number of poor fell – that had never occurred until then from 1950. The incidence of poverty fell since 1973-4, but the total number of poor fell only between 2004 and 2012.

In short, not one of these life changing transformations in ordinary lives was sustained after 2015. Growth slowed, pattern of growth became K-shaped and more unequal; non-farm job growth fell; and hence real wages stagnated. Poverty may have fallen since 2015; by how much is still unknown.

Finally, we have shown earlier that the claim of falling inequality, just as the claim of falling poverty, between 2012 and 2022 is bogus, because the two household consumption surveys are not comparable, and hence their data, too, is not comparable.

All in all, the World Bank should stop publishing such estimates, in the interest of its own credibility.

Santosh Mehrotra was professor of Economics, Jawaharlal Nehru University, New Delhi. He is currently visiting professor at the University of Florence, Italy.

The Wire is now on WhatsApp. Follow our channel for sharp analysis and opinions on the latest developments.