29 March 2014

The New Poverty Line: A Methodology Deeply Flawed

The New Poverty Line:
A Methodology Deeply Flawed
Madhura Swaminathan*
This note discusses the Report of the Expert Group to Review the Methodology for Estimation
of Poverty. The Report recommends the use of the existing official urban poverty line as the
poverty line for rural and urban India. While the main reasons cited for this choice are those of
convenience (statistical consistency over time) and ‘general acceptability’, the new poverty line,
it is claimed, also provides for minimum nutritional, health, and educational outcomes. These
justifications do not stand up to scrutiny. In respect of nutrition, the new poverty line is based
on lower calorie norms than used in earlier estimates, and these norms refer to the minimum
intake for persons engaged in light activity. In relation to education and health, the fact that
expenditure at the poverty line on education and health equals the median expenditure does not
ensure adequate outcomes in terms of education and health.
For some years, the Government of India has been under pressure to change the norms
for calculating the official poverty line. The current norms have resulted in a gross
and manifest under-estimation of the numbers of the poor, and, consequently, in the
exclusion of hundreds of millions from development programmes that are based on
the identification of households defined to be below the poverty line (or BPL). The
exclusion of malnourished households from the public distribution system (PDS) has
been the most visible form of such exclusion; exclusion also characterizes a wide range
of development schemes that are based on the principle of targeting BPL households.
The current poverty line is based on a survey of consumer behaviour conducted in
1973-74. Based on this survey, a consumption basket was proposed that would ensure,
on an average, 2100 calories per person per day in urban areas and 2400 calories per
person per day in rural areas. The poverty lines that were then established have been
criticized for being too low, and for focusing exclusively on food consumption norms,
with no allowance being made for expenditure on, inter alia, health, education, and
other basic needs (for example, Saith, 2005).
In November 2009, the Report of the Expert Group to Review the Methodology for
Estimation of Poverty (chaired by Professor Suresh Tendulkar) was submitted to the
Planning Commission (GoI, 2009). This much-awaited report was expected to present
a new and more realistic measure of poverty.
The best-known outcome of the Report of the Tendulkar Committee is that the
poverty line that it has proposed is higher than the current poverty line for rural areas,
Indian Journal of Human Development, Vol. 4, No. 1, 2010
* Professor, Indian Statistical Institute (ISI), Kolkata.
122 Indian Journal of Human Development
and has resulted in a dramatic increase in the proportion of the rural poor in India.
The major contribution of the Report is a total revision in the methodology used to
construct price indices to compare rural and urban prices in different states and the
state-level prices indices with the all-India price indices. ‘New’ or revised poverty
lines are proposed for the rural areas of all the states as well as for the country on the
basis of the use of new price indices with the existing urban poverty line. The revised
all-India rural poverty line is higher than the earlier poverty line. Consequently, at the
all-India level, the Report estimates that 41.8 per cent of the rural households were
below the poverty line in 2004-05, as compared to the current estimate of 28.3 per
cent of rural households. While the main outcome of the Report is thus to raise the
share of the rural population below the official poverty line by about 14 percentage
points, its methodology is deeply flawed. The poverty line that it proposes actually
depends on reduced calorie consumption, and fails to provide for reasonable household
expenditures on schooling and health.
The proposed poverty line is not based on a new consumer survey or revised
method of computation. The new poverty line for rural and urban areas is simply the old
poverty line for urban areas in 2004-05. Reading through the report, the Committee can
be said to justify the choice of the poverty line for urban areas in India in 2004-05 as the
all-India poverty line on three main grounds. First, it is defended on grounds of ‘general
acceptance’: the urban poverty line “is generally accepted as being less controversial
than its rural counterpart.” Secondly, it is defended on grounds of statistical consistency,
mainly with respect to comparability over time. Thirdly, and very significantly, the
Report argues that the proposed poverty line is reasonable because it happens to
simultaneously ensure satisfactory nutrition, health, and education outcomes. Briefly,
the Report says that the old urban poverty line is adequate to meet today’s requirements,
both rural and urban, with respect to nutrition, education, and health.
This claim that the revised poverty line is adequate to meet expenditure
requirements with respect to nutrition, education, and health is invalid. First, the
expert group has actually lowered the calorie intake requirement from 2100 Kcal per
day for urban areas and 2400 Kcal per day for rural areas to a single norm of 1800 Kcal
per day. On calorie requirement, the Report says: “…the revised minimum calorie
norm for India recommended by the Food and Agricultural Organization (FAO) is
currently around 1800 calories per capita per day, which is very close to the average
calorie intake of those near the poverty line in urban areas (1776 calories per capita).”
What the Report does not say is that the standards for energy requirements released
by the FAO [jointly with the United Nations University (UNU) and WHO] are for
‘minimum dietary energy requirements’ or MDER for short (FAO, 2009, technical
annex, emphasis in original). The MDER is defined as the amount of energy needed
for light or sedentary activity. Nutritionists prescribe energy requirements that vary by
age, sex, and activity level (light, moderate, heavy). The proposal that the standard
for light activity be taken as the requirement for an average person with expenditure
around the poverty line is unacceptable. It is a fiction that will result in a gross underestimation
of the population of the poor.
Flawed Methodology of the New Poverty Line 123
According to the FAO, an example of sedentary or light activity is of “a male office
worker in urban areas who only occasionally engage in physically demanding activities
during or outside working hours” (FAO, 2004).1 Not even one poor person struggling
to make a living in the informal sector would fit this description. Can a domestic
worker in urban areas, who scrubs floors and dishes, and washes clothes at work and
home for at least eight hours a day be assumed to be engaging in light activity? Or
can we assume that a head-load worker, who carries heavy sacks throughout the day,
is engaged in light activity? Anyone who has observed how hard the urban poor toil
for their paltry wages will see the absurdity of this assumption. To refer to the FAO
norm of 1800 Kcal, as the Tendulkar Committee Report does, but not to the associated
activity level is disingenuous.
Secondly, the FAO Report warns that in countries with a high prevalence of undernutrition,
“a large proportion of the population consumes dietary energy levels close
to the cut-off point, making MDER a highly sensitive parameter.” Given that we do
have a high level of under-nutrition in India, drawing a poverty line at the MDER is
clearly problematic, since taking a slightly higher cut-off will increase the number of
poor people substantially.
Thirdly, the same FAO Report provides data on the energy intake per capita
for a range of developing countries, classified by income level as well as by the
level of under-nutrition. Now, in all the countries where under-nourishment
affects less than 5 per cent of the population, irrespective of the income level, the
average per capita energy supply is greater than 2800 Kcal per day. The per capita
energy supply was, for example, 3100 Kcal per day in Iran, 3320 in Egypt, 2860 in
Malaysia, and 3030 in Korea. While the calorie intake is not perfectly correlated
with the nutritional outcome, it is nevertheless clear from international experience
that in countries with low malnutrition, the average calorie intake is much higher
than 1800 Kcal.
The Report’s claims about education and health are equally problematic. On
education, the argument of the Report is as follows. First, the Report states that in
2004-05, according to NSS data, 90 per cent of the children in the age group of 5 to 14
years, belonging to households at the poverty line level of expenditure in urban areas,
were in school. This is assumed to be a satisfactory outcome, though it falls short of
the achievement of universal schooling. Secondly, it assumes that the median cost of
sending a child to school, as reported in the NSS employment survey, sets a normative
or desirable level of expenditure on a child in school. Thirdly, according to the Report,
the average expenditure on education per child among households in the poverty line
expenditure class was higher than the median cost of schooling per child.2 From these
observations, it can be concluded that the actual expenditure is adequate to ensure
that children are in school.
The assumption by the Expert Committee that the median cost is adequate to
ensure proper schooling for all children is incorrect, and points to a lack of sensitivity
to the realities of poverty in a committee set up to study the phenomenon.
124 Indian Journal of Human Development
Here, an illustration can serve to highlight this fact. First, given the high degree
of inequality of expenditure on education in urban India, the median cost is likely to
be lower than the mean cost. Thus, if the committee had taken the mean expenditure
as the norm, the actual expenditure may have been inadequate among households at
the poverty line. Secondly, even if all the children of a household at the poverty line
are in school, they may not have all the notebooks or proper uniforms or other study
materials required to imbibe the education. In other words, the fact of school enrolment
or attendance is no assurance of the adequacy of household expenditure on schooling.
There is no discussion of the absolute level of the estimated median cost of schooling,
and whether it can be interpreted as a minimum desirable level of expenditure. Thirdly,
the actual expenditure incurred on education by a household at the poverty line may
be at the cost of rising indebtedness. If a household is borrowing heavily to send its
children to school, the sustainability of educational expenditure is also in question.
The first two points are corroborated by the background study undertaken by Tilak
(2009) for the Expert Group. Tilak shows that the per capita expenditure on education
as well as the share of education expenditure in the total household expenditure
increases with the expenditure class. In 2006-07, among rural households, the per capita
expenditure on education was Rs. 1.91 for those in the lowest expenditure class, and
Rs. 95 for those in the highest expenditure class (ibid., Table 3). Moreover, inequalities
in expenditure on education (by the rural-urban sector and by the expenditure class)
were found to be highest for expenditure on primary education. Tilak argues in favour
of arriving at a normative minimum desirable level of expenditure, or as a second-best
strategy, using the average rather than median expenditure “to enable the poor to get
meaningful and quality education” (ibid.). On the basis of the 63rd Round of the NSS,
Tilak estimates the required monthly per capita expenditure on education (based on
average expenditures) to be Rs. 23 in rural areas and Rs. 96 in urban areas (at 2007-08
prices). (A reading of Table 3 in Tilak suggests that the median value would be much
lower, at around Rs. 7 for rural areas and Rs. 26 for urban areas.)
The Committee offers a similar set of arguments with respect to health expenditure.
Data from the NSS 60th Round are used to calculate the median cost of non-institutional
health care (with a reference period of 15 days) and institutional healthcare or
hospitalization (with a reference period of 365 days). The age-specific incidence of
treatment and hospitalization are also calculated and then multiplied by the age
distribution of the population to arrive at estimates of the incidence of treatment/
hospitalization. The latter figures are then multiplied with the relevant median cost to
arrive at a ‘normative expenditure’ for healthcare. The problems with this approach
are similar to those listed above for education (that is, using median costs rather than
average or normative costs). There is also no guarantee that median costs ensure
appropriate healthcare services. An additional problem in the case of healthcare is that
the estimated incidence of treatment/hospitalization is not the incidence of a particular
illness/disease but is likely to be lower as not all cases may be treated. Therefore, as in
the case of education, there is no justification for assuming that the reported median
expenditure on healthcare services reflects adequate and appropriate treatment.
Flawed Methodology of the New Poverty Line 125
In sum, the Expert Group chaired by Professor Tendulkar chose the urban poverty
line of 2004-05 to serve as the new national poverty line on the grounds that it was
‘less controversial’ than the current rural poverty line and that it also fulfilled the
requirement of statistical consistency over time. This new poverty line was justified
on the grounds that it also provides for minimum nutritional, health, and educational
outcomes. These justifications, however, do not stand up to scrutiny.