As India’s Planning Commission seeks to review its measurement of poverty, the issue has become a hot topic for public debate. This column argues that while poverty lines should be used as benchmarks for policy, they should not be used to decide who receives benefits and who doesn’t – nor should they distract us from the real issues of poverty.
Never has the debate over the number of poor people in India caught the popular imagination the way it has done in the past year. While the measurement of poverty, along with what causes it to decline, has long been a topic for Indian academics, this debate has never reached the middle classes – at least not in the misinformed and misguided way that it has done in the current context.
Credit for this should go to the Planning Commission – the agency assigned the task of estimating poverty in India. Last year the Planning Commission went to the Supreme Court in a bid to justify the use of poverty estimates for targeting those who should receive more support. Yet given the historical experiences of targeting in India, any such move to seek legitimacy for targeting was not only bound to create strong reactions but also to raise doubts over the motive for seeking such legitimacy. This was essentially the reason for the outburst from the middle class, which asked a legitimate and relevant question as to the foundations of the poverty estimates the Planning Commission was seeking to use. It is unfortunate that the collateral damage of all this was the credibility of the National Sample Survey Office (NSSO) consumption data and the poverty estimates, which enjoy widespread credibility among economists and academics not only in India but also internationally. .
The question of whether the poverty lines are adequate or not is best seen in the context of which poverty lines are used. At least in India, as elsewhere, the primary purpose of measuring poverty has been, and remains, for tracking of progress over time of the population at the bottom of the wealth distribution. However, if the purpose is limited to its use as a statistical benchmark then any reasonable poverty line is good enough so long as it is sufficiently low and captures the various dimensions of poverty. In fact, the idea that the poverty line used for such statistical analysis is always arbitrary but reasonable was explicit in the Tendulkar committee, set up by the Indian government to look at the methodology for estimating poverty. However, even though arbitrary, the fact that the poverty line is rooted in some notion of what is the minimum level of living allows it to be used as a yardstick for measuring how India’s economic growth is being distributed.
The other important and essential characteristic of a poverty line is that it should be adjusted correctly for variations in level of prices across regions (states and rural/urban in the Indian context, known as ‘spatial’ variation) and across time (known as ‘inter-temporal’ variation). The poverty line that existed before the Tendulkar committee was based on the recommendations of another government investigation under the Lakdawala committee (Expert Group 1993). However, there were several problems with the spatial as well as inter-temporal price indices that led to these poverty lines being questioned. The Tendulkar committee essentially tried to correct the weaknesses in these price indices by shifting to unit values from the NSSO consumption surveys even though it retained the all India urban poverty estimates based on the Lakdawala committee. The resultant poverty line for urban areas was the same as earlier but the correction in price indices meant that the rural poverty line was higher by 30% and the rural poverty estimates higher by 50% at the national level.
Even though the Tendulkar committee revised and corrected the poverty lines for spatial prices and effectively raised the rural poverty lines, it was criticised for being an underestimate. While there may not be much merit in the arguments of those who have been criticising the poverty line for being too low on a per person per day basis, their concerns remain valid. This is again entirely due to the double speak of the Planning Commission which has not shown clarity in making its stand clear on whether these poverty lines will be used for the targeting of beneficiary households or not. While it must be made clear that there is nothing wrong with the Tendulkar poverty lines as long as it is used only as a statistical benchmark to track progress over time, it must also be made clear that the Tendulkar committee did not recommend using the resultant poverty estimates for any targeting of beneficiaries nor was this ever recommended by any other committee set-up earlier to estimate poverty. To maintain the sanctity of poverty estimates, the least that was expected of Planning Commission was to maintain the distinction between use of poverty estimates as a tool of measuring progress over time and targeting of beneficiaries for social assistance programmes.
At the same time, even as a statistical tool, the absolute poverty line that is used by the Planning Commission has to satisfy the generally accepted norms of a minimum standard of living, as measured by such factors as spending and access to healthcare. The poverty lines recommended by the Tendulkar committee have been externally tested for most of these. The committee did accept the fact that the lines met most of the norms in the majority of states but not all. But even with all these checks, there may still be some validity of the claims that these poverty lines are an underestimate. Because of this, individual researchers have often used different yardsticks depending on the nature of enquiry to understand various dimensions of poverty and wellbeing. For example, even the World Bank while using the $1.25 poverty line as standard benchmark for poverty comparisons also gives estimates by different poverty lines such as $1.50 and $2.00 per capita per day.
Having said this, let’s also compare it with other poverty lines. The weighted average of Tendulkar committee rural and urban poverty lines for 2004-2005 turns out to be Rs 16.25. This is only marginally lower than the Rs. 20 used by the Arjun Sengupta committee that claimed that 77% of Indians live below this poverty line. The World Bank uses a poverty line of $1.25 per day in Purchasing Power Parity (PPP) terms. The current poverty line, as claimed by the Planning Commission in its affidavit, is Rs 26 for rural areas and Rs 32 for urban areas. The weighted average turns out to be Rs 28 in 2009-2010 prices. Using the current PPP exchange rate of Rs. 19 to a dollar, the Indian poverty line is higher than the World Bank poverty lines. What about other countries and their poverty lines? Most developed countries don’t use an absolute poverty line but use a relative poverty line pegged at 60% of median income/expenditure. The Tendulkar poverty lines, by contrast, are 92% of the median in rural areas and 69% of the median expenditure in urban areas in 2004-2005. In other words, the Indian poverty line is considerably higher than poverty lines used either in international comparisons or comparable poverty lines in other countries.
Overstepping the line
But does it then justify using these poverty lines to restrict benefits to the Below Poverty Line (BPL) households, particularly for basic rights such as food and health? The answer is an emphatic no. I have consistently argued for universal provisioning of these basic rights without recourse to any targeting (Himanshu 2010b, Himanshu 2011). The debate should not focus on what the poverty line is but on who is eligible for social benefits. On this issue there is wide ranging acceptance that this must be completely delinked from estimates of poverty based on expenditure norms. Poverty lines are benchmarks for policymakers and economists to understand how the country is progressing and cannot be used for inclusion and exclusion from government programmes.
But this controversy has robbed the country of the more important debate on what happened to poverty between 2004-2005 and 2009-2010, which was not only a period of acceleration of growth rates of the country as a whole but also saw exceptional changes in the performance of the state governments. While some of the hitherto poor states such as Bihar and Uttar Pradesh saw acceleration in growth rates, with Bihar achieving the distinction of second-fastest growing state, it was also the period of states reforming the delivery of public services notably the Public Distribution System (PDS) for delivery of subsidised food grain. Two states that have seen the highest reduction in poverty during 2004-2010, namely Orissa and Tamil Nadu, are also the states that have been recognised for reforms in the PDS.
This is also the period when issues such as impact of the National Rural Employment Guarantee Act (MNREGA) on poverty should have been evaluated based on concrete data. On the other hand, this period is also characterised by very high inflation, particularly in food, which coincided with the worst droughts in India’s independent history. Not to be ignored is the fallout of the international financial crisis and the subsequent recession. Moreover, by all estimates, this period has been accompanied by increasing inequality in many spheres of economic as well as social wellbeing. Clearly, the last decade should have been a time to focus on the the distributional consequences of the growth rather than the measurement of it.
The immediate task for policymakers should be to evaluate the robustness of these poverty estimates independently but also in comparison of the Lakdawala estimates. Not only in terms of its suitability as a yardstick of measuring poverty but also comparability across regions and across time.
Yet this task requires careful scrutiny of survey concepts as well as clarity on comparability of various poverty estimates. An obvious case in this regard is the addition of imputed mid-day-meal expenditures in the 2009-2010 survey round whereas previous surveys have not included these as part of private household expenditure. Preliminary estimates suggest that the exclusion of this item alone lowers the poverty decline by 1.5% points against the official estimates. While a case is made for inclusion of these free or subsidised expenditures since they do improve the welfare of recipients, proper comparison requires that earlier household surveys are also treated similarly.
More importantly, however, any evaluation of poverty estimates should also be used as an opportunity to decompose the poverty decline between 2004-2005 and 2009-2010 to evaluate the contribution of various features of economic growth outlined above. A more disaggregated and rigorous analysis will be useful in understanding the challenges to inclusive growth in India across social classes and across different states.
A call to researchers
The issue of poverty and inequality and its relation to economic growth is not only an academic concern but also a political concern. Unfortunately, politicisation of poverty estimates for targeting purposes or otherwise has neither helped the states objectives of effective public service delivery nor has it helped in building trust in these estimates. While the new committee set up by the Planning Commission to re-estimate poverty may be able to resolve the problem, it is left to independent researchers to get back to using poverty estimates for the purpose for which they are designed. That is to analyse the trends and underlying dynamics of social flux in our society in a period of significant change.
- Himanshu (2010a), “Towards New Poverty Lines for India”, Economic and Political Weekly, 45(1)
- Himanshu (2010b), “Food Entitlements should be Universal”, MINT 28th April
- Himanshu (2011), “Towards Universal Food Security” SEMINAR, January 2011
- Planning Commission (1993): “Report of the Expert Group on Estimation of Proportion and Number of Poor”, Perspective Planning Division, Planning Commission.
 For details see Planning Commission (1993).
 For example, external validation checks were carried out for the poverty line to satisfy minimum nutritional, educational and health expenditure norms along with the ability to capture dimensions of poverty such as occupational vulnerability, literacy and so on (Himanshu 2010a).
 The year was 2009-2010.