Indian poverty: ideology and evidence


* I would like to thank Arindom Mookerjee for excellent research assistance. See for articles on poverty and related subjects.

back to issue

IDEOLOGY plays an important part in the formulation of policies to attack poverty, and in the study of evidence on poverty. On both counts, its role in the removal of poverty in India has been questionable.

Policies to remove poverty have been state intensive. Poverty in India is defined as per person consumption expenditure of less than Rs 51 per month (1973 prices) or close to Rs 400 per month at current (year 2000) prices. For a family of five this translates into Rs 24,000 per year, which is only half the level at which Indian citizens pay taxes. There are good arguments to suggest that both the first taxation level and the poverty line in India be raised by about 25%.

What the poverty line should be is dictated partly by the values pertaining to redistribution and partly by the ‘surplus’ income in society. This is fair. What is questionable are the policies derived from this goal, policies that are generally a license for state inefficiency and state corruption. For example, there is an elaborate machinery to procure foodgrains from farmers, store them, move them across states, and sell them in ration shops at ‘subsidized’ prices. This elaborate procedure ensures considerable leakage and corruption, and leads to the absurd reality that at present there are about 20 million tons of grains rotting in government godowns.

The ideology behind these policies is one of individuals articulating them knowing ‘better’, which gets translated into the state knowing better. The ideology is one of ‘loco parentis’, e.g. a parent is needed to guide the itinerant peasant. It is this same logic that leads some to proclaim that we should preserve the adivasis lifestyle, even if it means keeping them permanently in poverty. Have the adivasis ever been asked whether they want their children to go through the same lifestyle as themselves – or do they desire ever increasing opportunities and incomes, as the rest of us on this planet?

There is little logic to the plethora of policies – ration shops, food for work, fertilizer subsidies – that masquerade as policies meant for the poor. This is where ideology is present in large ‘in the name of the poor’ quantities. Given that the goal of poverty reduction and equal opportunity is universally accepted, and politically correct, it should follow that only those policies should be adopted which maximize redistribution to the poor at minimum cost.

One such efficient policy is income transfers. This policy is simple – after identifying the poor, they are given a cash transfer to bring them at par with the poverty line. The local panchayats can be involved in both the identification of the poor and in the estimation of their incomes. The poor would be more than anxious to assert their rights and demand their due. Leakages will be minimal, as will corruption.



Not that this policy will eliminate corruption; for example, the panchayat could bribe the poor to show a larger than actual poverty gap in order to receive a kickback. The Government of India has conducted various National Sample Surveys (NSS) since 1951. These surveys contain large amounts of information on poverty in a particular region, and its evolution over time. These village/district estimates can be used to cross-check estimates provided by panchayats.

The likelihood of this simple policy being adopted is low, all for the simple reason that vested interests (the bureaucracy, rich farmers, politicians) will lose considerably if the present system is junked. The gainers will be the poor, and they have little lobbying power.

Ideology also pervades the assessment of the number of poor in India, and what policies have, or have not, been successful in reducing poverty. Exaggeration has its due place in literature, but policies should not be based on hysterical examination of data.

The most egregious example of such behaviour is that of the novelist Arundhati Roy in her ‘concern for the poor’ pamphlet, The Greater Common Good. In it she claims that more than 50 million people have been displaced by the construction of large dams in India. Normally, outpourings of novelists do not deserve serious attention; but if they are of the right ideology, they can even influence the outpourings of so-called objective commissions of inquiry, e.g. the World Commission on Dams (WCD).

The WCD’s report is the latest example of ‘public policy via fiction’. This exhaustive, and otherwise thorough, report is flawed by its overtures towards those who want to exaggerate the number of people displaced. The WCD states that the ‘overall level of physical displacement could range from 40 to 80 million.’ Another statement, strewn throughout the report, and starting from the Chair’s preface, is that there are more than 45,000 large dams (defined as those above 15 m or with a reservoir volume of above 3 million cubic meters) in the world. So the average number of people displaced by large dams is 1333 (60 million divided by 45000).



For India and China, the Commission states the following: ‘Thus, in India and China together, large dams could have displaced between 26-58 million people between 1950 and 1990.’ China, according to the report, has close to 23000 dams and India 4200. So the average displaced in these two populous countries is 1544 per large dam; excluding India and China, the number displaced per dam is close to 1125. So the two populous countries – India and China – displace about 40% more people per dam than non-populous countries. The aggregate statistics have a ring of plausibility, and consistency to them.

But in an apparent seizure of ideology, the WCD genuflects to the ‘demands’ of Indian ‘analysts’. The WCD break-up between the two most populous countries, India and China, is revealing. India, with 4200 dams has displaced 16-38 million people; China, with six times as many dams – 23000 – has half as many displaced, 10-20 million! Nowhere in the report does the Commission attempt to explain this (ideological) discrepancy. Or the fact that if the Indian upper estimate is applied to China, then there would be about 226 million Chinese displaced by dams – or a third of the rural population of China.

If you believe that statistic, you will also believe that 56 million have been displaced by large dams in India – a statistic cited by Arundhati Roy in her recent polemic on globalization (Roy, 2000). The India Country Study from where Roy ostensibly gets her ideological statistic, does not contain any reference to 56 million.

As discussed in detail in Bhalla (1999a, 1999b) and Bhalla-Mookerjee (2000), the likely figure for people displaced per dam in India is about 1360, a number strikingly close to the world average figure given by WCD. Note that the WCD figure for displacement by all the dams in the entire world is close to Roy’s ideologically exaggerated figure for India alone.

The controversy over the displaced numbers raises the question as to why do people exaggerate with such intensity. The why is outside the domain of this paper – and more in the domain of psychiatrists wanting to discover how different people cope with the ‘guilt’ associated with not being poor. It has been mentioned here to alert readers, and researchers, to the possibility that the poverty numbers thrown up by different analysts may have an ideological bias. Perhaps ‘in the name of the poor’ studies should carry a statutory warning: ‘Reportage of "facts" may be injurious to the truth.’



An analogous, and perhaps ideological, problem exists with the ‘research’ studies on the extent of absolute poverty in India. The poverty line in India is defined according to the ‘base’ year, 1973, and defined as monthly per capita consumption below Rs 49 in rural areas and Rs 56.6 in urban areas. Table 1 reports the ‘official’ poverty lines for the different years; the 1973 levels have been updated using separate deflators for rural and urban areas; and the national poverty line is obtained by aggregating the data for the urban and rural population in the different years.

The table reports the percentage of the population that is poor in each of the years – this percentage is referred to as the ‘head count ratio’ or HCR. The table reports three different estimates of the poverty ratio – NSS7, NSS30 and national accounts modified (NAM). NSS7 and NSS30 refer to the National Sample Survey estimates, and NAM refers to an estimate obtained from national accounts data. (For a description of the NAM methodology, see Bhalla 2000f).



National accounts data yield estimates of mean consumption; the NSS data yield estimates of the distribution of consumption. Together, the two sets of data can yield ‘reliable’ estimates of trends in poverty. Indeed, until 1993, this was precisely the approach taken by the Government of India. For reasons that are not entirely clear, the GOI set up an expert group under the chairmanship of the late eminent scholar, D.T. Lakdavala. The officially stated reason for an expert study was as follows: ‘A number of methodological issues have been raised in respect of the estimates of poverty released by the Planning Commission. In view of the importance of poverty eradication as a social objective, wide-ranging references to the incidence of poverty in discussions relating to social problems as also their use in allocation of funds for poverty alleviation programmes, it was thought that all the issues relating to the estimation of poverty could be considered afresh by an expert group’ (Page 1, first paragraph, EG-GOI, 1993.)

In 1993, the group came out with its report entitled ‘Report of the Expert Group on Estimation of Proportion and Number of Poor’ (EG-GOI, 1993). One of the most significant policy conclusions contained in this report was the recommendation that national accounts data no longer be used for adjusting mean survey estimates, i.e. the National Sample Survey be used, in toto, to generate results on poverty trends.

It is likely that this single decision has been responsible for the considerable amount of noise, and ideological misinformation, that now exists about poverty in India. Table 1 reports several estimates of poverty in India. Some idea of the divergence between the official survey data (NSS) and national accounts data (NAS) can be gleaned from the fact that in 1973, the survey data reported aggregate expenditures which were three-fourths of the national average as given by the national accounts; in 1999, the survey is only able to capture 58% of national income. All incomes, including that of the poor, are understated in 1999, and understated by a large amount. Even if 1973 is taken as the ‘true’ figure, it still emerges that in 1999, with the same understatement as 1973, survey consumption expenditures are 25% lower than ‘actual’; and poverty about 20 percentage points higher.

The data reported in Table 1 illustrates the huge divergence in poverty estimates obtained by various researchers. The latest NSS data pertains to July-December 1999. The official estimate of poverty for this period is between 24 and 27% or around 240 to 270 million. Note that this very same data source indicated that only 18% of the population was poor four years ago – in other words, poverty has gone up substantially in the last five years! The NAM (modified national accounts) estimate for 1999 is that only 6% of the population was poor in 1999. The NCAER estimate for poverty in India in 1997 is 17%.



Note also that both NCAER and Deaton-Tarozzi obtain identical estimates for the poor in India in 1993 – 29%. Per capita consumption has grown by about 30% between 1993 and 1999. On a conservative estimate, this growth probably resulted in poverty being reduced by about two-thirds of the rise in income, i.e. 20%. (See Bhalla, 2000d and Bhalla, 2000f for details pertaining to the estimation of the trickle down elasticity, i.e. how much poverty goes down with increase in income.) A conservative forecast of the non-NSS 1993 estimates of poverty would suggest that poverty in India today is in the single digits, i.e. less than 100 million. This number is close to that yielded by the modified national accounts method – a method, incidentally, which was used till 1996 by the Government of India.



Different Estimates of Poverty in India



Poverty Line (Rs/month)

Urban Pop. (% of total)

Mean per capita expn. per month

Head Count Index






















































1993 (D-G)




1993 (D-T)































1998 (Jan-June)



















* NSS 1 refers to the 30-day recall method of estimating food consumption, and NSS 2 refers to the preferred, 7-day recall period for estimating monthly food consumption. Prior to 1994-95, all NSS data used the 30-day recall period.

* NAS refers to the author’s own estimates using per capita consumption from National Accounts data and distribution from NSS. National Accounts data have been adjusted for likely exclusion and under-estimation of income of rich households.

* Head Count Index is the per cent of population estimated to be absolutely poor according to the poverty line reported in column 2.

* D-T refers to estimates in Deaton-Tarozzi (1999); D-G refers to estimates by Dubey-Gangopadhyay (1998).

* For NCAER (MISH) data, see Natarajan et al, ‘Economic Reforms and Poverty Alleviation: A tale of two surveys’, 20 July 2000.




Bhalla, Surjit S. (2000a), Trends in World Income Distribution and Poverty: Some New, Some Different, Estimates. Paper Presented at Princeton University, 3 October 2000.

Bhalla, Surjit S. (2000b), ‘Bis-mil-muflis: In the name of the poor’, Business Standard, 16 September 2000.

Bhalla, Surjit S. (2000c), ‘Why We Don’t "Grow" Poor’, Business Standard, 22 July 2000.

Bhalla, Surjit S. (2000d), FAQ’s on Poverty in India. Paper Presented at a seminar in Delhi School of Economics, 20 July 2000.

Bhalla, Surjit S. (2000e), Trends in World Poverty: Ideology and Research, Paper presented at IMF, 28 June 2000.

Bhalla, Surjit S. (2000f), ‘Growth and Poverty in India – Myth and Reality’, in Govinda Rao (ed.), Poverty and Public Policy: Essays in Honour of Raja Chelliah, (forthcoming).

Bhalla, Surjit S. (2000g), ‘World Bank – We have a Poverty Problem’, Economic Times, 18 January 2000.

Bhalla, Surjit S. (1999a), ‘Don’t Confuse me With Facts’, Indian Express, 27 September 1999.

Bhalla, Surjit S. (1999b), ‘Going Wrong With Figures in a Big Dam Way’, Indian Express, 6 September 1999.

Bhalla, Surjit S. and Ravinder Kaur (1999), ‘Poverty in India – Towards New Policies’, in S. Gangopadhyay (ed.), Poverty in India, Rajiv Gandhi Foundation, December 1999.

Bhalla, Surjit S. and A. Mookerjee (2000), Big Dam Development: Facts, Figures and Pending Issues. Paper to be presented at forthcoming conference on Water Development and Poverty Alleviation, Ahmedabad, 29-31 January 2001.

Deaton, Angus and A. Tarozzi (1999). Prices and Poverty in India. Mimeo, Princeton University, December 1999.

Dubey, A. and S. Gangopadhyay (1998), Counting the Poor: Where are the Poor in India? Sarvekshana Analytical Report No.1. Department of Statistics, Government of India, February 1998.

Expert Group, Government of India (1993), Report of The Expert Group on Estimation of Proportion and Number of Poor. Perspective Planning Division, Planning Commission, July 1993.

Natarajan I, D.Lal and R. Mohan (2000), Economic Reforms and Poverty Alleviation: A Tale of Two Surveys. Paper presented at a seminar in Delhi School of Economics, 20 July 2000.

NCAER, Volume 1, Issue 2 (2000), ‘Micro Impacts of Macroeconomic and Adjustment Policies’ (MIMAP) – India, January 2000.

NCAER, Volume 1, Issue 1 (1999a), ‘Micro Impacts of Macroeconomic and Adjustment Policies’ (MIMAP) India, October 1999.

NCAER, (1999b), India Human Development Report of the Nineties, Oxford University Press, 1999.

Roy, A. (2000), ‘The Reincarnation of Rumpelstiltskin’, Outlook, 27 November 2000.

Roy, A. (1999), ‘The Greater Common Good’, first published in Outlook, and then released as a booklet by India Book Distributors, 1999.

World Commission on Dams (2000), Dams and Development – A New Framework for Decision-Making, EarthScan Publications, November 2000.