It is a mad, mad Indian world. The government has just reported that the percentage poor in India has fallen from 39 percent in 1987-88, and 36 percent in 1993-1994, to about 27 percent in 1999-2000. If MAD magazine were to review the situation, it would state that there are two camps – the what me worry types and the “But Me Worry” (BMW) types. Instead of creating some joy that there are less joyless in India, and therefore a reason for some celebration, the reaction among the BMWs in the Indian government (and their close relatives in Indian academia) is that the data are wrong. And why are the data wrong? Because in the National Sample Survey of 125,000 households, there were some (less than 10 percent) of households who were asked some expenditure questions in the wrong sequence. Sorry, I am not stating the position of the BMWs accurately.

In the latest survey, households were asked both about their food expenditures on a 7 day recall basis, as is the practice elsewhere in the world, and a 30-day recall basis, as has been the practice in India. The reason for the switchover (after this transition year) is because the NSS has (belatedly only 40 years later) found that the 30 day recall period for food expenditures leads to unreasonably low estimates of actual consumption of all households, poor and non-poor. NSS decided (correctly) to ask the 7-day recall questions after the 30-day recall questions for 1999-2000. However, due to predictable goof-ups, the order got reversed for about 10 percent of the respondents. Why does this matter? Because, the BMWs state, the data are wrong because the respondents will have a tendency to just multiply the weekly figure to arrive at the monthly figure - and this can result in over-estimation of actual food consumption.

The mad world is madder because what the BMW's don't realize is that the above error will result in consumption being less, not more, than the "true" figure. And therefore the poverty levels being reported are too high, not too low. This follows from the simple observation that most people will multiply the weekly figure by 4 to arrive at the monthly estimate, rather than by the correct multiple of 4.345. So monthly consumption levels will be understated by about 9 percent!

The BMW types - those who want to show that poverty does not decline with economic growth, or that global integration does not lead to higher economic growth, or that the oppressive hand of government is needed to guide all souls to nirvana, or that all actions are in the public interest as long as they are conducted by the public sector - are a bit behind reality. When all is done and counted, the NSS survey for 1999-2000, based on the more accurate 7 day recall period, is likely to show poverty levels around 15 percent. Yes, Sahil, poverty in India has declined, concomitant with economic growth. To be ahead of the curve, the BMW wallahs should start waiting for 2005. Poverty declines will now be more difficult as India approaches the more intractable, and government action needed part, of reduction in hard-core poverty.

But there is a possible escape route for the BMW's. Poverty may have gone down, but the poor are getting poorer in terms of health and education. Jean Dreze, "Poverty: beyond headcount ratios", The Hindu, Sept. 9, 2000 states that "the rate of decline of the infant mortality rate, for instance, was less than half as high in the 1990s as in the 1980s". But me worry about globalization, income growth, privatization, industrial delicensing and a decreased role for the government.

Indeed, infant mortality declines (and Kerala) are major weapons in the BMW's armor. Such declines are largely due to three factors - mother's health, mother's education and technology (medicinal advances). Income is a fourth factor. The table shows the infant mortality levels for 1981 and 1997 for various states of India, the actual and predicted (log) percentage decline, and the ratio of the two.

The results suggest that broad based medicinal advances are perhaps the most important factor behind the changes in infant mortality over the last two decades. Bihar can be taken as the base-case, pure technology, no policy (only God) scenario. Infant mortality there has declined from a very high level of 118 (per 1000 births) in 1987 to a still very high, but lower, 71 level in 1997. The best performers are Gujarat and UP (actual vs. predicted gains of over 50 percent), followed by Kerala, Bihar, Madhya Pradesh and Tamil Nadu (actual gains larger than predicted gains). The not so well performing states are the rest, with the worst performer being Karnataka - gain of less than half than that which should have occurred.

Note that what is being assessed here is performance between 1981 and 1997. In other words, the ranking is based on something new that happened, rather than historical. Kerala, with the lowest infant mortality rate in 1997 (at international developed country levels of 12) also had the lowest infant mortality rates in 1981 (at international developed country levels of 35). Though data are not available, it is likely that Kerala had similarly outstanding infant mortality levels in 1947. The relevant question is not why Kerala has a low infant mortality level in 1997, but why it had a low level in 1947. The answer is likely to be the emphasis on education via the influence of the Christian missionaries - a hundred plus years process. The answer is not as claimed by the World Bank in its "authoritative" World Development Report that politically correct, me worry state policies were responsible. "Kerala has life expectancies greater than those in Washington, D.C., despite vastly lower levels. The effectiveness of public action in Kerala has been attributed to its strong tradition of political and social mobilization." (p. 34, emphasis added).

The question remains - why do some officials, some academics, have the urge to worry? Could it possibly have something to do with their ideologies ? I have an ideology, and so do you. We, especially economists, cannot survive without it. The practice of economics attracts a bucket-full of policy wonks. By definition, policy wonks (and I consider myself a card carrying member of the tribe) like the feeling of power. Policy wonks are often opinionated (a necessity for being a successful wonk) and opinions are derived from ideology. Indeed, ideology is the guiding force - the sequence is from ideology, to research, to advocating the results of this research, to power, to celebrity fame. And broad political appeal is ensured if at the forefront of all this advocacy is concern for the poor.

Michel de Certeau, in his 1984 book The Practice of Everyday Life, captures this urge, and the phenomenon, very well. "The ministers of knowledge have always assumed that the whole universe was threatened by the very change that affected their ideologies and their positions. They transmute the misfortune of their theories into theories of misfortune". (p. 95-96, emphasis added).