Don’t let poverty of thought stop India from tracking deprivation
So long as we can define a cut-off for poverty, we can measure it, and its validity doesn’t require data trackers to be kept frozen
is former executive director at the International Monetary Fund and former part-time member of the Prime Minister’s Economic Advisory Council.
There is a rumour going around. That senior economists in India feel that the method of measuring consumption affects the construction of the poverty line. This is the Great Indian Poverty Debate III—that a comparison of the levels of living (and hence derivation of the headcount ratio of poverty, or simply poverty) between India’s two official Household Consumption Expenditure Surveys (HCESs) in 2011-12 and 2022-23 cannot be done. And so one cannot conclude that extreme poverty (based on the purchasing power parity or PPP-linked $1.9 poverty line) in India has been eliminated, as Bhalla-Bhasin have done (bit.ly/3Vz9Hc2). In the Telegraph, online on 2 March 2024, (bit.ly/3ISoHuj), former acting member and chairman of the National Statistical Commission P.C. Mohanan and retired Indian Economic Service officer K.L. Datta, author of Growth and Development Planning in India, said “Tendulkar’s parameters cannot be applied to the findings of the HCES 2022-23 to arrive at a poverty figure.” Why? “Because the Mixed Reference Period (MRP) methodology of the old HCES, whose findings provided the base for Tendulkar’s poverty cut-offs, was different from the Modified Mixed Reference Period (MMRP) methodology employed during HCES 2022-23”.
Its translation: We cannot arrive at any poverty estimate for 2022-23 comparable to the poverty estimate for 2011-12 because the consumption estimates of the two surveys are not comparable. How valid an objection is that? Never in world-poverty history has anybody asserted that a poverty line (poverty cut-off) is a function of how one measures per capita consumption. A reading of poverty literature in India and abroad for the last 80-plus years would suggest that the Mohanan-Datta view borders on defying logic.
The universally accepted accounting procedure for measuring poverty is to define (even imagine) a cut-off level of per capita expenditure (or income). Those above the cut-off are not-poor; those below are poor. The first poverty line was constructed by a working group of nine experts in the Planning Commission in July 1962. These experts noted that “on the basis of available data on distribution of population according to per capita expenditure, nearly half the Indian population in 1960-61 was below this national minimum level of ₹20 per capita per month,” and the commission called them poor.
In their classic study, Dandekar-Rath defined the poor in terms of an average consumption of 2,250 calories per capita per day, equivalent to ₹15 for rural areas and ₹22.50 for urban areas, both in 1960-61 prices. This became the poverty line for academic studies in India and around the world. At around the same time, in the early 1960s, US president Lyndon Johnson launched a war on poverty in the US. The identification of the poor, for policy purposes, was constructed to be equal to the money equivalent of families of three or more whose food expenditure was more than one-third of their total spending. Using this method, poverty in the US was estimated as 20% in the early 1960s, and then dropped rapidly to 12% by 1968. It has hovered in the low-teens since then.
In the early 1970s, the World Bank entered the poverty arena in a big way. In a paper published in 1979 titled ‘Growth and poverty in developing countries,’ authors Montek Singh Ahluwalia, Nick Carter and Hollis Chenery laid much of the groundwork for the poverty research that followed. Issues pertaining to definition, measurement and forecasts were discussed in detail. The authors explicitly rejected a calorie consumption approach and opted for a monetary poverty line; and lack of consumption-survey data for several developing countries made them reluctantly opt for an income rather than a consumption-derived poverty line. The poverty line chosen under the International Comparison Program (or ICP, later to be named PPP) was $200 per capita per year in 1970 prices. The authors tagged this poverty line to be the income of the 45th percentile of the Indian income distribution for 1975.
The circle was complete. Paraphrasing T.S. Eliot, the profession came back to where it started and knew how to measure poverty for the first time. Again, not an iota of discussion took place on whether the different methods or surveys were comparable. The profession, statisticians, economists and policymakers have not really cared about how the poverty cut-off line is derived, so long as it is known and defined in a recognizable manner. Calculation of the level of poverty, once a poverty line has been agreed upon and consumer expenditure (or income) survey data is available, is no more than a simple accounting exercise. Indeed, until the World Bank prevailed on India’s Planning Commission, the practice in the country was to arrive at a best estimate of survey consumption and then blow up all per-capita consumption by the ratio of mean consumption in national accounts and mean consumption in the survey.
Consider this. The World Bank estimates poverty for over 100 countries across 60 years for the same and different poverty lines. None of these calculations involves the method of measurement or number of questions asked, or the methods of collecting data (one visit or ‘n’ number of visits to a household).
In each country, the method of collection of data on per capita consumption has varied from one survey to another. More items are included and some are dropped. Think of consumer durables like typewriters, gramophones, tape decks, etc. The list is endless.
Note that this is a mix of several methods of measurement, all geared to arrive at a best estimate of consumption. Like the Invisible Hand, the Mohanan-Datta model of estimation is nowhere to be seen. How does one interpret their supposed findings and those who agree with their conclusion that the HCES surveys of 2011-12 and 2022-23 are not comparable? It is political, let’s face it.