THE World Bank has just come out with estimates of poverty in India which show that “extreme poverty” which afflicted 27.1 percent of the total population in 2011, now afflicts only 2.6 percent. Its estimate is based on the National Sample Survey data on consumption expenditure collected over the period 2022-23 and 2023-24.

The Modi government, using the same data, had claimed that extreme poverty had reduced to 5 percent of the population for these years. It is now crowing over the World Bank’s latest findings which come to an even lower poverty ratio using a different poverty line, namely 3 dollars per person per day; this is converted to rupees at an exchange rate that is about 28 percent of the market exchange rate (which it calls purchasing power parity adjustment). But the World Bank’s findings are not worth the paper on which they are written.

There are two quite distinct problems with the World Bank’s conclusion: first, there has been a substantial change in the method of data collection between 2011 and 2022-23 because of which the figures for the two years are not directly comparable. For instance, the NSS investigators earlier used to make one visit to a household during the month of investigation, now they make three visits; and the greater the number of visits the greater is the reported consumption expenditure by the household. Comparing 2011 and 2022-23 figures therefore has an in-built bias towards underestimating poverty in the later year compared to the earlier one.

The second problem with the World Bank estimates is conceptual: the very definition of poverty the Bank uses is a flawed one; let us see why. If a person consumed more of an identical set of commodities in a later year that he was consuming in the base year, then clearly he has become better off and is obviously less poor. But if a person consumes more of some and less of others then one cannot be sure if he is better off or worse off. The usual method of comparison here is to deflate the expenditure sum in the later year by a price index which is a weighted average of price-relatives, with the weights being the proportions of total expenditure on different commodities in the base year. In the case of poverty estimates this method is translated to updating the base-year poverty line for the later year by using such a base-weighted poverty index and then seeing how many people fall below this updated line.

One may not agree with this method, but it has some rationale. The problem arises however when the commodities consumed in the later year are not all the same as in the base year, and of some common commodities there is a reduction in consumption. The typical method for poverty estimation here too is to update the base year poverty line by a base-weighted price index, but this has no rationale whatsoever.

The rationale that is adduced for doing so is usually the following: if the old goods were being consumed in the later year then anyone above such an updated poverty-line would be considered non-poor; and if the person actually consumes a different set of goods then he must be preferring this latter set and hence be even better off. Using a poverty-line updated by a base-weighted price-index therefore is perfectly legitimate even in this case.

The problem here however is that observed choice does not necessarily indicate actual preference. Observed choice can also arise because of the non-availability of base year goods. An obvious example is public healthcare. Because of the pursuit of fiscal “austerity”, government expenditure on maintaining public hospitals and building new ones may go down over time and far more expensive private hospitals may come up in their place. Many more patients, even those who are indigent, would then be forced to go to these private hospitals. The actual price of healthcare in this case would have gone up considerably even though this increase would never be captured by any base-weighted price-index. The base-weighted price-index would capture at best the rise in public healthcare costs (if government hospitals alone existed in the base year), but not the rise in actual healthcare costs when such hospitals have become of secondary importance. The actual price-rise therefore would have been under-estimated and so would the updated poverty-line and hence the estimated poverty-ratio.

Exactly the same can be said of education and a range of other services, where the withdrawal of government services leads to an underestimation of the actual price-rise and hence of the poverty-line and poverty ratio. The importance of such withdrawal can be assessed from the fact that all over the world students and youth are currently protesting the effects of a decline in public services.

It is because of this in-built bias towards underestimating poverty because of the method of estimation, whenever there is a forced change in the available basket of goods, that Indian statisticians had persuaded the Planning Commission in 1973-74 to adopt an alternative method of estimating poverty. When the basket of consumed goods changes, not necessarily out of choice on the part of the majority, there are some goods which do not change in the basket and whose consumption is observed all over the world to go up as people become better off (at least over a whole range of incomes); the consumption of these goods in two periods which can be physically compared, can then be taken as a proxy for whether people are becoming better off or worse off.

Foodgrains are one such good and their total ingestion, both direct ingestion and indirect ingestion in the form of feedgrains entering into animal products, and processed grains, can be such a proxy. Since foodgrains are themselves quite diverse, one way of comparing their total ingestion on two different dates, is to take the calories being ingested, which can be physically measured without bringing in prices. This is not “calorie fundamentalism” and does not amount to claiming that calories alone matter; it is simply taking calories as a proxy for the standard of living without bringing in price-indices that are necessarily flawed.

Accordingly, for years the Planning Commission used to estimate poverty by counting how many persons in urban India do not have access to 2100 calories per person per day, and in rural India how many persons cannot access 2200 calories per person per day (the original benchmark for rural India was 2400 but this was reduced to 2200 for unknown reasons). On this criterion there had been a remarkable increase in the incidence of poverty in the neo-liberal period: the proportion of population defined to be “poor” on this criterion had increased from 57 percent to 65 percent in urban India between 1993-94 an 2011-12, and the proportion defined as “poor” in rural India had increased from 58 to 68 percent over the same period. The figures for 2017-18, the next quinquennial NSS survey, were so large, reportedly in excess of 80 percent for rural India, that the Modi government decided to withdraw the data from the public domain, to stop making calorie-intake estimates and to change the method of data collection under the NSS that was noted above and that makes the figures obtained from all subsequent surveys non-comparable with those of 2011-12.

To be sure, every method of estimating poverty can be legitimately criticized on one ground or another; but to claim that poverty has disappeared in India, which in effect is what the World Bank is doing, when hunger stalks the country, despite 5 kilograms of foodgrains per person per month supposedly being distributed free to about 800 million persons, is the height of absurdity. India currently occupies the 100^th rank among the 122 countries for whom data are available and the Hunger index can be computed. When the free foodgrains scheme, introduced to cope with the pandemic, will be terminated, which the government has already indicated will happen soon, matters will become even more serious. We shall then have a country where according to the World Bank poverty has disappeared, but where people go desperately hungry!

The government of India objects to the criteria on the basis of which the world hunger index is calculated: it argues that too much weight is given to child malnutrition but not enough to malnutrition in the population at large. But in India child malnutrition itself is a crucial indicator of deprivation, hence the government’s argument is not very consequential.

The World Bank, being a backer of neoliberalism, has a vested interest in prettifying its effects. It is obviously not in the data or the estimates based on them that this prettification occurs; it occurs in the very conception behind poverty. Likewise the Oxford Multidimensional Poverty Index has a conceptual underpinning that takes “modernity” as being indicative of lack poverty. One therefore has to join issue with these bodies at a conceptual level, which the debate in India on the claimed decline in poverty has not done until now.