MacroScan (Printable Version) - The Tendulkar Report: A Small Step Forward

The report of the expert group on the estimation of poverty led by Professor Suresh Tendulkar has been submitted to the Planning Commission. Apart from issues of comparability of data across NSS rounds, the most important ToR for the committee was to ''review alternative conceptualizations of poverty, and the associated technical aspects of procedures of measurement and database for empirical estimation'' of poverty in India.

Poverty, of course, is a multi-dimensional concept. However, official statistics in India have always referred, arguably narrowly, to only income poverty (using the proxy measure of consumption expenditure from the NSSO surveys). In India, we have been using a calorie-based procedure to fix the poverty line as the minimum level of expenditure that would enable a person to purchase a specified food basket. A task force of the Planning Commission in 1979 defined the poverty line as that per capita expenditure at which the average per capita per day calorie intake was 2400 calories in rural areas and 2100 calories in urban areas. This task force used age-sex-activity specific calorie allowances recommended by a Nutrition Expert Group in 1968 to estimate the average daily per capita calorie requirements for rural and urban areas. Estimates of average expenditure incurred by that population in each State that consumed these quantities of calories as per the 1973-74 survey of NSSO were fixed as poverty lines.

Based on the observed consumer behaviour in 1973-74, it was estimated by the task force that an expenditure of Rs 49.09 per capita per month was associated with a calorie intake of 2400 per capita per day in rural areas and Rs 56.64 per capita per month with a calorie intake of 2100 per day in urban areas. These poverty lines were updated for future years by simply accounting for the changes in consumer price indices. As such, the all-India poverty lines updated for 2004-05 were Rs 356.30 in rural areas and Rs 538.60 in urban areas, all per capita per month. The shares of population below these poverty lines (the head count ratios; HCR) were estimated to be 28.7 per cent in rural areas and 25.9 per cent in urban areas.

These estimates of poverty threw up a number of controversies. First, the NSSO estimates of poverty were arrived at using poverty lines that were extremely low in levels. An amount of Rs 356.30 per month per person amounted to just Rs 11.90 per day in rural areas, which was at best a destitute income. The fact that about one-fourth of India’s population did not incur even this level of expenditure was in itself a revealing point.

Secondly, the levels of poverty and deprivation reported from independent surveys, including village surveys, were far higher than the NSSO estimates of poverty.

Thirdly, the NSSO estimates were at great variance with estimates of nutritional outcomes that other surveys like the National Family Health Survey (NFHS) provided. For instance, according to the NFHS-3 in 2005-06, the share of underweight children (under 3 years) in rural India was 44 per cent and the share of stunted children in rural India was 41 per cent. Among women in the age group of 15 to 49 years, 58 per cent were anaemic and 39 per cent had below normal body mass index (BMI).

Fourthly, there were major methodological issues involved in the use of consumer price indices, continuously re-weighted keeping the 1973-74 consumption basket unchanged, to update the poverty lines over time. The consumption basket of rural and urban persons had changed significantly after 1973-74. One striking absurdity that resulted was that in some States, urban poverty rates were estimated to be higher than the rural poverty rates.

The Tendulkar committee has reviewed the present methodology for measuring poverty and has suggested major changes for the future. These changes may be crudely summarised as follows:

Given the poor correlation between calorie consumption levels and nutritional outcomes, the calorie-norm for estimating the poverty line should be abandoned. Instead, the committee has suggested a new method.
This new method involves the consideration of the present all-India urban poverty line as the basis for every other poverty line. This consideration is justified on the basis of two independent validating reasons:

(a) The population that corresponded in 2004-05 to the poverty line expenditure in urban areas consumed 1776 calories per capita per day, which was close to the calorie norm of 1800 calories per capita per day suggested for India by the Food and Agriculture Organisation (FAO).

(b) The actual levels of per capita expenditure in urban areas in 2004-05 were also sufficient to meet a defined ''normative level of expenditure on education and health services''.
The all-India urban poverty line has to be consistently estimated based on the mixed reference period method (using a combined 365 days and 30 days recall) rather than the present uniform reference period method (using a uniform 30-day recall).
With the present all-India urban poverty line as the basis, the Committee has recommended the identification of its parity levels at the State-level for rural and urban areas separately. Thus, applying purchasing power parity (PPP), separate rural and urban poverty lines are to be estimated for each State at which the levels of consumption in the urban areas can be sustained.
It is thus postulated that the new poverty lines, fortuitously, meet not just food expenditure requirements, but also those of education and health that are important basic needs.

Using the above method, the new poverty lines for 2004-05 have been re-estimated by the committee as Rs 446.68 for rural areas and Rs 578.80 for urban areas (per capita per month). Using these poverty lines, the HCRs in 2004-05 were estimated as 41.8 per cent in rural areas and 25.7 per cent in urban areas. These new estimates are an upward revision in rural areas from 28.7 per cent as per the old method, and a slight downward revision in urban areas from 25.9 per cent as per the old method. The upward revision in the rural areas is due to the use of the PPP method, which has reduced the urban-rural price differentials implicit in the present method of estimation.

The upward revision of rural poverty by the committee is indeed a welcome step, as this would help States to expand their BPL coverage in the public distribution system (PDS) using grains from the central quota itself. It is also welcome that non-food expenditures like those on education and health have not just been included (in fact, there was an allowance in the earlier method too), but also that provisions have been made to update them across time. However, these steps solve only a part of the problem, as the system of targeting in welfare schemes like the PDS is likely to remain in place and large sections of poor people above the new poverty line would remain outside targeted welfare provisions.

Take an example: the new poverty line for rural areas has been revised from Rs 356.30 per capita per month to Rs 446.68 per capita per month. In daily terms, this means an increase from Rs 12 to Rs 15 per capita. This is just a meagre upward revision. For urban areas too, the increase is meagre; the revision of poverty line is from Rs 538.60 per capita per month (Rs 18 per day) to Rs 578.8 per capita per month (Rs 19 per day). In other words, the new poverty line continues to be extremely low in levels and keep a large section of the population outside the definition of the ''poor''.

Juxtapose this with the fact that 77 per cent of the population lived at less than Rs 16 per day with respect to expenditure in 2004-05. In 2004-05, the average MPCE of those households with expenditure less than double the poverty line (i.e., of the 77 per cent; the ''poor and vulnerable'', as classified by the NCEUS report) was only Rs 486, or Rs 16 per day. If the average expenditure stands at Rs 16 per day, there is likely to be a sizeable section of the population above the newly suggested poverty line of Rs 15 per day in rural areas and Rs 19 per day in urban areas. In a targeted welfare provision, these sections of the population would remain to be excluded.

Another central question is whether abandoning the calorie norm is a wise step or not. It is true that calorie intakes were poorly correlated with nutritional outcomes (as in the famous case of Kerala). However, abandoning the calorie norm altogether and taking solace from the fact that calorie intakes appear to be adequate at the new poverty lines is an overstretched and arbitrary proposition. It is unclear whether there is any basis, theoretical or empirical, for this relationship to hold at all the years to come.

The Tendulkar Committee report is the latest input to the ''Great Indian Poverty Debate''. The reason for the rising contestation around poverty data in the recent years is the use of HCRs to arbitrarily fix the number of households eligible for many important welfare benefits, such as the PDS. In this method, the questions of estimation of the number of poor and the identification of the poor remained separate processes, and were thus open to bizarre policy outcomes. It is for the absence of a reliable method of combining estimation and identification that political and social movements have been demanding universalisation of welfare provisions like the PDS. The Tendulkar Committee report itself is evidence to the fact that levels of poverty are extremely sensitive to even minor changes in the poverty line.

While the increase in the number of poor households, as suggested by the Tendulkar Committee, may indeed help expand the coverage of welfare schemes, it would still fall short of including all the needy sections from the ambit of such schemes. One would welcome the newly suggested methodology for arriving at a strictly technical measure of poverty. However, it is important to insist that the new estimates are not mechanically linked to the issue of eligibility to access major welfare schemes. In a country with such mass poverty as India, universalisation remains the most efficient tool for ensuring livelihood security.



Themes > Current Issues	23.12.2009


The Tendulkar Report: A Small Step Forward

R. Ramakumar

The report of the expert group on the estimation of poverty led by Professor Suresh Tendulkar has been submitted to the Planning Commission. Apart from issues of comparability of data across NSS rounds, the most important ToR for the committee was to ''review alternative conceptualizations of poverty, and the associated technical aspects of procedures of measurement and database for empirical estimation'' of poverty in India. Poverty, of course, is a multi-dimensional concept. However, official statistics in India have always referred, arguably narrowly, to only income poverty (using the proxy measure of consumption expenditure from the NSSO surveys). In India, we have been using a calorie-based procedure to fix the poverty line as the minimum level of expenditure that would enable a person to purchase a specified food basket. A task force of the Planning Commission in 1979 defined the poverty line as that per capita expenditure at which the average per capita per day calorie intake was 2400 calories in rural areas and 2100 calories in urban areas. This task force used age-sex-activity specific calorie allowances recommended by a Nutrition Expert Group in 1968 to estimate the average daily per capita calorie requirements for rural and urban areas. Estimates of average expenditure incurred by that population in each State that consumed these quantities of calories as per the 1973-74 survey of NSSO were fixed as poverty lines. Based on the observed consumer behaviour in 1973-74, it was estimated by the task force that an expenditure of Rs 49.09 per capita per month was associated with a calorie intake of 2400 per capita per day in rural areas and Rs 56.64 per capita per month with a calorie intake of 2100 per day in urban areas. These poverty lines were updated for future years by simply accounting for the changes in consumer price indices. As such, the all-India poverty lines updated for 2004-05 were Rs 356.30 in rural areas and Rs 538.60 in urban areas, all per capita per month. The shares of population below these poverty lines (the head count ratios; HCR) were estimated to be 28.7 per cent in rural areas and 25.9 per cent in urban areas. These estimates of poverty threw up a number of controversies. First, the NSSO estimates of poverty were arrived at using poverty lines that were extremely low in levels. An amount of Rs 356.30 per month per person amounted to just Rs 11.90 per day in rural areas, which was at best a destitute income. The fact that about one-fourth of India’s population did not incur even this level of expenditure was in itself a revealing point. Secondly, the levels of poverty and deprivation reported from independent surveys, including village surveys, were far higher than the NSSO estimates of poverty. Thirdly, the NSSO estimates were at great variance with estimates of nutritional outcomes that other surveys like the National Family Health Survey (NFHS) provided. For instance, according to the NFHS-3 in 2005-06, the share of underweight children (under 3 years) in rural India was 44 per cent and the share of stunted children in rural India was 41 per cent. Among women in the age group of 15 to 49 years, 58 per cent were anaemic and 39 per cent had below normal body mass index (BMI). Fourthly, there were major methodological issues involved in the use of consumer price indices, continuously re-weighted keeping the 1973-74 consumption basket unchanged, to update the poverty lines over time. The consumption basket of rural and urban persons had changed significantly after 1973-74. One striking absurdity that resulted was that in some States, urban poverty rates were estimated to be higher than the rural poverty rates. The Tendulkar committee has reviewed the present methodology for measuring poverty and has suggested major changes for the future. These changes may be crudely summarised as follows: Given the poor correlation between calorie consumption levels and nutritional outcomes, the calorie-norm for estimating the poverty line should be abandoned. Instead, the committee has suggested a new method. This new method involves the consideration of the present all-India urban poverty line as the basis for every other poverty line. This consideration is justified on the basis of two independent validating reasons: (a) The population that corresponded in 2004-05 to the poverty line expenditure in urban areas consumed 1776 calories per capita per day, which was close to the calorie norm of 1800 calories per capita per day suggested for India by the Food and Agriculture Organisation (FAO). (b) The actual levels of per capita expenditure in urban areas in 2004-05 were also sufficient to meet a defined ''normative level of expenditure on education and health services''. The all-India urban poverty line has to be consistently estimated based on the mixed reference period method (using a combined 365 days and 30 days recall) rather than the present uniform reference period method (using a uniform 30-day recall). With the present all-India urban poverty line as the basis, the Committee has recommended the identification of its parity levels at the State-level for rural and urban areas separately. Thus, applying purchasing power parity (PPP), separate rural and urban poverty lines are to be estimated for each State at which the levels of consumption in the urban areas can be sustained. It is thus postulated that the new poverty lines, fortuitously, meet not just food expenditure requirements, but also those of education and health that are important basic needs. Using the above method, the new poverty lines for 2004-05 have been re-estimated by the committee as Rs 446.68 for rural areas and Rs 578.80 for urban areas (per capita per month). Using these poverty lines, the HCRs in 2004-05 were estimated as 41.8 per cent in rural areas and 25.7 per cent in urban areas. These new estimates are an upward revision in rural areas from 28.7 per cent as per the old method, and a slight downward revision in urban areas from 25.9 per cent as per the old method. The upward revision in the rural areas is due to the use of the PPP method, which has reduced the urban-rural price differentials implicit in the present method of estimation. The upward revision of rural poverty by the committee is indeed a welcome step, as this would help States to expand their BPL coverage in the public distribution system (PDS) using grains from the central quota itself. It is also welcome that non-food expenditures like those on education and health have not just been included (in fact, there was an allowance in the earlier method too), but also that provisions have been made to update them across time. However, these steps solve only a part of the problem, as the system of targeting in welfare schemes like the PDS is likely to remain in place and large sections of poor people above the new poverty line would remain outside targeted welfare provisions. Take an example: the new poverty line for rural areas has been revised from Rs 356.30 per capita per month to Rs 446.68 per capita per month. In daily terms, this means an increase from Rs 12 to Rs 15 per capita. This is just a meagre upward revision. For urban areas too, the increase is meagre; the revision of poverty line is from Rs 538.60 per capita per month (Rs 18 per day) to Rs 578.8 per capita per month (Rs 19 per day). In other words, the new poverty line continues to be extremely low in levels and keep a large section of the population outside the definition of the ''poor''. Juxtapose this with the fact that 77 per cent of the population lived at less than Rs 16 per day with respect to expenditure in 2004-05. In 2004-05, the average MPCE of those households with expenditure less than double the poverty line (i.e., of the 77 per cent; the ''poor and vulnerable'', as classified by the NCEUS report) was only Rs 486, or Rs 16 per day. If the average expenditure stands at Rs 16 per day, there is likely to be a sizeable section of the population above the newly suggested poverty line of Rs 15 per day in rural areas and Rs 19 per day in urban areas. In a targeted welfare provision, these sections of the population would remain to be excluded. Another central question is whether abandoning the calorie norm is a wise step or not. It is true that calorie intakes were poorly correlated with nutritional outcomes (as in the famous case of Kerala). However, abandoning the calorie norm altogether and taking solace from the fact that calorie intakes appear to be adequate at the new poverty lines is an overstretched and arbitrary proposition. It is unclear whether there is any basis, theoretical or empirical, for this relationship to hold at all the years to come. The Tendulkar Committee report is the latest input to the ''Great Indian Poverty Debate''. The reason for the rising contestation around poverty data in the recent years is the use of HCRs to arbitrarily fix the number of households eligible for many important welfare benefits, such as the PDS. In this method, the questions of estimation of the number of poor and the identification of the poor remained separate processes, and were thus open to bizarre policy outcomes. It is for the absence of a reliable method of combining estimation and identification that political and social movements have been demanding universalisation of welfare provisions like the PDS. The Tendulkar Committee report itself is evidence to the fact that levels of poverty are extremely sensitive to even minor changes in the poverty line. While the increase in the number of poor households, as suggested by the Tendulkar Committee, may indeed help expand the coverage of welfare schemes, it would still fall short of including all the needy sections from the ambit of such schemes. One would welcome the newly suggested methodology for arriving at a strictly technical measure of poverty. However, it is important to insist that the new estimates are not mechanically linked to the issue of eligibility to access major welfare schemes. In a country with such mass poverty as India, universalisation remains the most efficient tool for ensuring livelihood security.



© MACROSCAN 2009