No set of calculations yield an absolute poverty figure of 45% for India in 1996 as reported by the World Bank. A correct use of WB’s own data suggests that absolute poverty in India is close to 20%. This suggests that research on poverty is Big Business.

World Bank - We Have a Poverty Problem By Surjit S. Bhalla (Economic Times, Jan. 18, 2000) Poverty is Big Business - for (com)pliant academics, guilt-ridden do-gooder liberals, international organizations nee NGO's, politicians and bureaucrats who believe that they know better. Hence, the series of "recommended" government interventions to decrease poverty - schemes for fodder and saris, employment, fertilizer and food subsidies, education, health care etc. All with loco parentis, noblesse d'oblige and a noble goal. Unfortunately, it is a moot question whether the so-called intervention schemes actually help reduce poverty. Which makes one wonder why such simple poverty alleviation policies like economic growth, education and health voucher schemes, cash transfers and food stamps are not advocated by the poverty economists, or by the biggest NGO of them all - the World Bank.

The reason it is not is because it will throw the Poverty Businessman out of business. A related issue is the business of just counting the poor, something that the World Bank (WB), through its vast resources, has appropriated for itself. This business is so huge that analysis of national data e.g. National Sample Survey (NSS) data from India, first appear in World Bank documents. Most Indian researchers find it impossible to access NSS data, except, of course, the privileged rentiers.

Now that the WB has lost its role in providing developing countries with investment capital, it may feel the need to inflate the numbers on its "new" raison d'etre - removal of absolute poverty. The logic is straightforward - the more poor there are, the more resources (yours and my taxes) it can tap. And the more poor there are, at least in the capitalist nineties, the more the WB, and the we-know-better types, can authoritatively state that economic growth is not doing it, so what is needed is - you guessed it - more government intervention.

World Bank consumption, growth and poverty figures are reported in the table for the years 1987 and 1996, the latter year deliberately chosen well before the onset of the Asian crisis. The data are obtained from the recently published Global Economic Prospects (GEP) and the World Development Indicators - i.e. the latest WB data. The consumption growth figures for the non-China developing world are a healthy 8.7 % for the seven year period 1987-1996. For India, the increase in per capita consumption levels is an impressive 18 percent.

But how much poverty reduction took place in India with an 18 % increase in per capita consumption ? Precious little according to WB-GEP - from 44.9 % of the population in 1987 to 42.3 percent in 1996. These numbers are actually for South Asia; while China figures are reported separately, India's are not. Inference would suggest that poverty in India, according to WB, stayed relatively constant at 45 percent during the time-period 1987 to 1994.  The WB growth-poverty figures for India are shocking - and bizarre. An 18 percent growth in per-capita consumption and no decline in absolute poverty?

The political economy of these figures is best understood by noting that the Government of India (GOI), and several Indian researchers, have also been reporting flat-to-rising absolute poverty levels in the economic reform nineties. Hence, the politically correct? conclusion that liberalization has proceeded without a human face, has in fact increased absolute poverty etc. The fact that an East Asian financial crisis followed in 1997 and 1998 is additional gravy to those whose ideology dictates that more government and World Bank intervention is needed for human welfare improvements.

Going back to the figures. Note that once data on the distribution of consumption expenditure (i.e. which tenth or fifth of the distribution has what share of average consumption), and data on average consumption are available, then a simple non-rocket scientist calculation yields the absolute poverty figures. Several organizations (GOI Planning Commission, World Bank, WIDER) report that the share of consumption of the bottom 45 % of the Indian population has varied between 24 and 25 percent from 1973 to 1996. That is right - all the available data suggests that the share of consumption of the bottom 45 % has been a "constant". (NCAER income distribution data for India for 1974-75 and 1994-95 suggests that the share of income of the bottom 45 % (ranked by per-capita income) has also been relatively constant at 20-21 percent.)

Given knowledge about the relative constancy of the distribution (at least between the bottom and top half of the population), only the data on the level of consumption in 1987 and 1996 are needed to calculate trends in absolute poverty. Given that the Indian economy, according to all analysts, has grown by about 4 % per capita per year in the nineties, the estimate of 18 percent consumption growth from 1987 to 1996 is conservative. And these conservative figures suggest that absolute poverty in India decreased from about 30 percent in 1987 to 19 percent in 1996 ($ 1 PPP poverty line). Thus, far from staying constant, poverty has radically declined in India, according to GOI and World Bank data! Indeed, the WB published poverty levels are closer to a 50 percent higher poverty line of $ 1.5 per day. Can the WB figures be explained by a typo?

Is there any logical explanation for the World Bank figures on poverty in India, 1987 to 1996? One (admittedly far-fetched) explanation could be that the poor did not share in the observed growth, but the non-poor did. If this were true, then the share in consumption of the bottom 45 percent would have declined from 24 percent to 20 percent. Is this plausible? Again, WB's own data says not. For most countries (including Pakistan and Bangladesh) for which expenditure distributions are reported in WB's own World

Development Indicators, an increase in the share of the bottom 40 percent is indicated!

All the available WB data suggests that WB's inflated estimates of poverty cannot be replicated, at least for India. The moral of the story - an ideologically motivated hodge-podge of numbers leads to a morally incorrect mixup. Over-counting the poor may be good for the business of the rich. But is it good for those being counted ?