Neoliberalism and Rural Poverty in India

Author(s): Utsa Patnaik

Reviewed work(s):
Source: Economic and Political Weekly, Vol. 42, No. 30 (Jul. 28 - Aug. 3, 2007), pp. 3132-3150
Published by: Economic and Political Weekly
Stable URL: http://www.jstor.org/stable/4419844 .
Accessed: 25/01/2012 21:05

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

Economic and Political Weekly is collaborating with JSTOR to digitize, preserve and extend access to
Economic and Political Weekly.

http://www.jstor.org
 and
Neoliberalism Rural Poverty in India
Many economic and social indicators suggest that not only is the level of absolute poverty in
India high, there has also been an adverse impact of neoliberal policies on poverty.
And yet, the poverty estimates by the Planning Commission and many
individual academics, both using a method that renders irrelevant the question of a nutrition
norm, show low levels as well as decline in poverty over the 1990s and beyond. This
article proves that both comparisons over time of the all-India and state-level estimates of
poverty as well as any comparison at a point in time of poverty levels across states, obtained
by this method, are invalid. Using a direct poverty estimation route of inspecting and
calculating from current National Sample Survey data the percentage of persons not able to
satisfy the nutrition norm in calories, the author finds that in 1999-2000 nearly half of
the rural population who are actually poor have been excluded from. the set qf thie officially
poor. For 2004-05, while the official estimate of rural poverty is 28.3 per cent, the
author's direct estimate of persons below the poverty line is 87 per cent. There is clear
evidence of a large and growing divergence over time between the author's direct
estimates of poverty and the official indirect estimates.

UNSA PATNAIK

in widely separated areas in different states (Andhra Pradesh.

Introduction Karnataka, Vidarbha in Maharashtra, Punjab and Kerala) and
these suicides are continuing. All these indicators of general
T he question of poverty levels and trends has become depression, combined with acute distress in specific regions, are
particularly contentious during the last 15 years, owing quite inconsistent with the claims of decline or constancy of
to the repeated claims by the government, by a number poverty.
o' academics associated with the government and by economists Since overwhelming evidence exists for the adverse trends in
associated with the World Bank, that a substantial decline in the ruraleconomy, is it the case that the official method of poverty
poverty in India - rural poverty in particular - has taken place estimation is itself faulty and is failing to capturethe actual trends
in the 1990s, during the period of implementation of neoliberal in poverty? While up to the mid-1990s, poverty estimates were
economic policies and trade liberalisation. The Planning Com- mainly of academic interest, from 1997 the food subsidy has been
mission claims that rural poverty has declined from 37.3 to 27.4 targeted and the population divided into "above poverty line"
per cent of the population comparing the 50th round (1993-94) and "below poverty line". Lower-priced foodgrains from the
and the 55th round (1999-2000) National Sample Survey (NSS) public distribution system are available only to those identified
data on consumer expenditure (though the latter figure has been as poor. How the poverty line expenditure is arrived at and how
recognised as an underestimate). The World Bank's World the poor are actually counted, has acquired an important
Development Report, 2006 presents a figure of 30.2 per cent for policy dimension affecting the lives and welfare of millions of
the latter date. The 61st round 2004-05 data, stated to be com- people in the country. If the counting is incorrect, it will lead
parableto the 50th round, has produced a recent official estimate to the implementation of wrong policy measures lowering
of 28.5 per cent in rural poverty. mass welfare.
On the other hand, the available official data show that, over Poverty has many dimensions, and can be thought of as not
exactly the same period, a number of interrelated indicators of merely material deprivation and a low standardof life, including
ruralwell-being have worsened: ruraldevelopment expenditures poor health indicators, but also deprivation in relation to edu-
have gone down as a share of national product and in real per cation and culture. One particular dimension of material depri-
head terms; all-India crop growth rates have halved in the 1990s vation had been correctly picked out as the most crucial one,
compared to the 1980s and foodgrains output has become stag- namely, the ability to access a minimum nutritionlevel expressed
nant over the last five years; rural employment growth has in terms of a norm of daily energy intake in calories, required
dropped sharply and open unemployment has been growing fast. for working health. This index was simple though it captured
Bank credit to farmers has declined and higher dependence on poverty only partially, and it obtained widespread acceptance.
private usurious credit combined with severe price declines for It was suggested in Dandekar and Rath's pioneering 1971 paper
many crops has led large segments of farmers into a debt-trap. and was taken up by the Planning Commission in India, which
Foodgrains absorption per head has declined sharply to reach set up in 1979 a Task Force on Projection of Minimum Needs
levels prevalent 50 years ago. Rising farm debts have led to loss and Effective Consumption Demand. Its recommendation
of assets reflected in a rise in landlessness, and to the historically was based in turn on the Indian Council of Medical Research
unprecedented situation of many thousands of farmer suicides table of dietary intakes [Gopalan 1992, 1997], which was

3132 Economic and Political Weekly July 28, 2007

applied to the population structureby age and gender. On average Plan period, 3.8 per cent of net national product (NNP) was spent
2400 and 2100 calories per day per capita worked out as the annually on RDE, with well-documented positive effects in
required daily allowance (RDA) for energy intake, for rural and raising non-farm employment and rural wages. From 1991, as
urban areas respectively, and all persons unable to access this contractionary Fund-guided policies started, the share of RDE
throughtheiractually observed expenditure were to be considered was cut sharply to below 2.6 per cent of NNP by 1995-96
as poor. and fell further to 1.9 per cent by 2000-01. Using implicit GDP
This measure using a nutrition norm is an absolute measure deflators, we find an absolute fall in real expenditure per head
of poverty as distinct from the relative measures used in many of rural population.
advanced countries - such as considering all those to be poor, Even though it was the agrarian crisis which led to the fall
who have less than half the average per head income in the of the National Democratic Alliance (NDA) coalition in the May
economy [Anand 1983, 1997; Subramanian 1997]. With a rela- 2004 general elections, the assumption of power by the United
tive measure of poverty, rise in inequality will imply rise in Progressive Alliance (UPA) government saw the deflationary
poverty. The absolute poverty measure adopted in India however hammer being applied once more on agriculture with budget
requires stronger conditions for poverty to show a rise. Increase estimates of RDE for fiscal 2004-05 being much lower than that
in the inequality of income and of expenditure could be quite of preceding years, and with a cut by one-third in funding for
consistent with poverty so defined, showing a decline. Only an the employment generation schemes. The revised estimates for
absolute decline in expenditure for substantial sections of the 2004-05 show a slight rise in RDE to 2.3 per cent of NNP, far
population (not offset by rise for other sections), would lead to short of the required doubling necessary to make an impact on
average poverty rising. rural depression. The simultaneous notifying of the Fiscal
The purpose of this paper is to explore why the poverty Responsibility and Budgetary Management Act, 2004 under-
estimates by the Indian Planning Commission and many indi- scored the strongly deflationist stance of the new government
vidual academics following the same method, show low levels even in the face of rising unemployment. The gross fiscal deficit
as well as decline in poverty over the 1990s, whereas all other as a per cent of GDP has been brought down from 6.1 in 2000-01
economic and social indicators suggest that absolute poverty is to 4.1 by 2005-06; it was slated to be further lowered to 3.8 in
high and there has been an adverse impact of neoliberal policies 2006-07 but has been actually lowered to 3.7.
on poverty. For this purpose, we startby detailing the main content This harsh contractionary fiscal policy has had nothing to do
of neoliberal economic policies guided by the Bretton Woods with any objective resource constraint - indeed with strong
Institutions (BWI) as they have affected rural activities. income shifts towards the already well-to-do, tax receipts have
been buoyant and the tax-GDP ratio has been rising - but has
II simply reflected the government's acceptance of the deflationary
Neoliberalismas an EconomicPolicy Package dogmas of financial interests and in particularof the BWI, which
advise expenditure reduction no matter how high unemployment
Neoliberalism entails a strongly expenditure deflating policy might be, and thereby greatly worsen the problems of unemploy-
package at the macroeconomic level, and India has been no ment and income loss, since the expenditure cuts have multiplier
exception. This proposition may seem strange since India has effects in reducing incomes and employment further. Indeed,
seen 6 to 7 per cent annual gross domestic product (GDP) growth these expenditure - reduction prescriptions are based precisely
rates. The overall growth rate can be misleading however, for on the untenable premise of full employment, for without this
it tells us nothing about the sectoral composition of growth or premise the pre-Keynesian proposition cannot be maintained that
its distributional effects. It is perfectly possible for the material there is a fixed savings pool in the economy such that increase
productivesectors to stagnate or decline while services, including in public expenditure will "crowd out" private investment directly
financial services, are booming, and this has been the case with or via a rise in the interest rate.
India's growth in the 1990s. More rapid structural shifts in the These views have been extensively critiqued [Baker, Epstein
sectoralcontributionto GDP have taken place than in any previous and Pollin 1998]. P Patnaik (2000) presented a critique of the
period; the manufacturing sector's share in GDP ha.,sstagnated "reduce the fiscal deficit" doctrine of the BWI and the theoretical
in the last 15 years while its contribution to employment has premise of full employment on which it is based, and U Patnaik
declined. While the share of agricultural and allied activities in (2003) contained a discussion of the impact on the peasantry,
GDP has fallen sharply, the population dependent on this sector of balanced-budget doctrines of the Great Depression years and
has declined little and faces falling per head real income. the present identical deflationary stance of the international
Agriculture is always a "soft" target for the misguided expen- financial institutions. I have elsewhere argued that this
diture deflating policies which continue to be urged by the BWI, revenantpre-Keynesiantheoryrepresentsthe logical fallacy known
no matter how high unemployment and hunger might be. The as the "converse fallacy of accident", in which from a specific
impact of deflationary policies has been especially severe in rural assumption (full employment) a general inference (expenditure
areas which, already subject to declining public investment, saw deflation) is improperly drawn.1
sharp reduction in public planned development expenditures. In Total capital formation in agriculture continues to stagnate in
rural development expenditures (RDE) I include the five plan real terms, since sharply reducing public investment is not being
heads of (a) agriculture, (b) ruraldevelopment, (c) irrigation and compensated by rising private investment. There is no economic
flood control, (d) special areas programmes, and (e) village and rationale for believing that "public investment crowds out private
small-scale industry. All these expenditures are vital for investment", which is the common deflationist argument for
maintainingruralproductivityand employment. The employment- reducing the state's role in rural development. Precisely the
generation programmes had assumed special importance from contrary has been shown to hold for certain types of investment
the drought year 1987 onwards. During the pre-reform Seventh essential for an irrigation-dependent agriculture like India's.

Economic and Political Weekly July 28, 2007 3133

 Chart 1: Annual Per Capita Output and Availabilityof Foodgrains in Kilograms,TriennialAverage Ending in Specified Years
185
180
175 .
.-
170 "-

165
160 --

155 -
150
145
140
135
1991-92 1994-95 1997-98 2000-01 2002-03
- Per capita Output, Kg 178.8 181.6 176.8 177.7 164.1
--a- Percapita Availability,Kg 177 174.3 174.2 163.2 153

Note: Availability is 0.875 times gross output plus net imports minus net addition to public stocks.

Private tube-well investment is profitable where the water table with over four-fifthsof the fall coming after 1998. This level
remains high owing to seepage from state-built canal irrigation prevailed50 yearsago, in the early 1950s,andis lowerthanthe
systems, and where community integrated watershed manage- 157 kg averageduring1937-41.Fortyyearsof successfuleffort
ment is encouraged with state help. The cutback of public to raise availabilityhas been wiped out in a meredozen years
investment and in RDE has led to a halving of the crop output of economicreforms.TheaverageIndianfamilytodayis absorb-
growth rate in the 1990s, and a collapse of employment growth. ing 115 kg less per year of foodgrainsthan in 1991; average
Both foodgrains and all-crop growth rates nearly halved in the calorieintakehas fallenfromalreadylow levels, andsince data
1990s compared to the pre-reform 1980s, and fell below the show that urbancalorie intakehas risen, it is ruralabsorption
population growth rate leading to declining per capita output, which has fallen much more than the average.
for the first time since the mid-1960s agricultural crisis, which This steep fall in foodgrainsavailabilityper head (Chart1),
however had been short-lived, whereas per head agricultural is a highlyabnormaltrendwhichis notexpectedwhenpercapita
output continues to fall even after a decade. incomeis rising,sincetheincomeelasticityof demandforanimal
The position has since worsened further: the peak-to-peak products,which is indirectuse of foodgrainsas animalfeed, is
foodgrains output has become completely stagnant over the last around1.6 for developingcountries.Foodgrainsavailabilityhas
six years at 112 million tonnes, and per head output is falling to meet not only directconsumptionbut all forms of use like
faster. With increasing use of land for non-agriculturalpurposes, animalfeed,processedfoods,andcommercialuse.Henceecono-
gross sown area has remained static since 1991. It is only through mies with risingper headincomeshow risingavailability,with
yield rise that output growth can be maintained, but yield growth an increasingsharefor indirectuses. The steepfall in percapita
is declining. The agriculturaluniversities had played a majorrole absorptionin India is consistentonly with worseningincome
in developing and disseminating new crop varieties, and the cut distributionof a particulartype, an absolutedecline in incomes
in research funding in these universities, by affecting the search and purchasingpowerfor a majorpartof the population,out-
for better rain-fed crop varieties, has also contributed to yield weighing rise for the minoritywith fast rising incomes. The
growth deceleration. interestedreaderis referredto my earlierpapersfor a more
Falling agricultural growth has produced fast growing open detailedanalysis,whichlocatesthe reasonsfor thedeclinein the
unemploymentcombined with a fall in numberof days employed. severeloss of purchasingpowerinherentin the unemployment-
Assuming constant labourcoefficients, a near halving of employ- raisinganddemand-deflating policiesnotedbrieflyabove,com-
ment growth was to be expected, given the near-halving of crop bined with exposure of our farmersto global price declines
output growth, but the actual job losses have been greater since after1996 as traderestrictionswere removed[Patnaik,U 1996,
mechanisation and use of chemical weedicides has additionally 2003, 2003a, 2004, 2005]. These were also added to by the
reduced labour coefficients. The work participation rate has attemptto cut the food subsidyby raisingissue pricesto final
declined, and open unemployment has been growing at over 5 buyers,morethanprocurementpricesto farmers,whichsimply
per cent annually. The elasticity of employment with respect to resulted in pricing out the poor from the public distribution
output, which was 0.5 during 1983 to 1993-94, fell to zero during system (PDS),andthefinalblow was the misguided"targeting"
1993-94 to 1999-2000. The 61 st round data relating to 2004-05, of the PDS from 1997 underwhich access to cheap food was
shows no improvement: the rural unemployment rate for men no longeruniversaland demand-drivenbut restrictedto those
is stagnant and for females it has risen, but the participation rate arbitrarilydefined as "poor"by the government.The result
shows a slight rise (NSS Report 515). was a massive fall in grain sales from the rationshops, from
I have earlier written extensively on the fact that annual 21 million tonnes in 1991 to only 13 million tonnes by 2001
foodgrains availability per head of total population, has fallen while normallysales shouldhave been risingas the population
steeply from 177 kg in the early 1990s, to only 153 kg by 2003-04, grew [Swaminathan2002].

3134 Economic and Political Weekly July 28, 2007

 Expenditure trends from the thin-sample rounds of NSS con- output - the peak-to-peak growth rate during the six years after
firmed this analysis. The lowest 40 per cent of persons ranked 1998-99 has collapsed to zero. Nothing less than a colonial style
by expenditure levels had absolutely lower per capita total real famine will, it seems, satisfy those whose objective seems to be
expenditure by 2001-02 compared to 1995-96 while the next 40 to turn Indian agriculture once more into a mere supply source
per cent had stagnantreal expenditure [Sen and Himanshu 2004]. for advanced country supermarkets and for retail outlets serving
In fact, the income situation is worse than expenditure because local elites, at the expense of increasing hunger for millions of
asset adjustments have been taking place to maintain consump- its own citizens.
tion flow.
The direct intake part of grain availability also declined in the Ill
1990s in all states except Kerala, West Bengal and rural Orissa. Divergencebetween Directand Official
Gopalan (1992:191) has pointed out that ".... If the habitual IndirectPovertyEstimates
cereal-legume dietaries of poor Asian population groups were
consumed at levels adequate to meet the full caloric needs (and Poverty studies in India since the 1970s, have been based on
here we are talking of caloric needs as conforming to present the use of a "poverty line" expenditure level, defined as that
international recommended mean levels of intake, and not particular observed level of expenditure per capita per month on
of M-2SD levels),2 then protein needs would be automatically
met". The National Nutrition Monitoring Bureau had informed Table 1: Distribution of Persons by Monthly Per Capita
us that "the NNMB has consistently confirmed in successive Expenditure (MPCE) Groups, Average Expenditure and
surveys that the main bottleneck in the dietaries of even the Average Calorie Intake per diem, 1999-2000, All-India
poorest Indians is energy and not protein as was hitherto be- Monthly Average Calorie Per Cent Cumulative
lieved... the data also indicate that the measurement of con- Per Capita MPCE Intake of Per Cent of
Expenditure Per Diem Persons Persons
sumptionof cereals can be used as a proxyfor total energy intake. Per Capita
(Rs) (Rs)
This observation is of considerable significance as it helps to (1) (2) (3) (4) (5)
determine rapid, though approximate, estimates of energy intake
at the household level"3 (emphasis added). The "foodgrains" in Rural
Below225 191 1383 5.1 5.1
this papercomprise cereals plus pulses. It is this strong link which 225-255 242 1609 5.0 10.1
enables us to say that the observed direct foodgrains intake 255-300 279 1733 10.1 20.2
decline given the overall availability decline (which proves that 300-340 321 1868 10.0 30.2
indirect intake is not compensating) means serious nutritional 340-380 361 1957 10.3 40.5
380-420 400 2054 9.7 50.2
decline and rise in poverty, which controverts the official view. 420-470 445 2173 10.2 60.4
To sum up, macroeconomic policies of expenditure deflation 470-525 497 2289 9.3 69.7
is the key to understanding the agrarian depression, and the 525-615 567 2403 10.3 80.0
615-775 686 2581 9.9 89.9
resulting loss of purchasing power or, in Keynesian terms, a 775-950 851 2735 5.0 94.9
severe squeeze on aggregate effective demand of the majority 950 and more 1344 3178 5.0 99.9
of the population, the key to understanding why such abnormal
All 486 2149 99.9
levels of public foodgrains stocks of 64 million tonnes, 40 million
tonnes in excess of buffer norms, had built up by July 2002. These Summary
Rs 470-525 and less; 2289 calories and less - 69.7 per cent
stocks were coming out of more and more empty stomachs. Rs 525-615; 2403 calories - 10.3 per cent
The government and the majority of economists put forward Rs 615-775 and more;2581calories and more- 19.9 per cent
a totally incorrectanalysis of the rising stocks and resulting falling
Urban
availability. They closed their eyes to the declining purchasing Below 300 255.8 1398 5.0 5.0
power broughtabout by public expenditure-deflating policies and 300-350 327.1 1654 5.1 10.1
instead they put the blame on allegedly "too high" minimum 350-425 389.1 1729 9.6 19.7
425-500 463.9 1912 10.1 29.8
supportprice (MSP) which they claimed gave the "wrong signals" 500-575 537.2 1968 9.9 39.7
to the farmers who therefore produced more than the market 575-665 618.6 2091 10.0 49.7
required,and they advocated reduction of MSP. This fallacious 665-775 718.7 2187 10.1 59.8
775-915 840.5 2297 10.0 69.8
argument ignored the fact that foodgrains growth rates had 915-1120 1009.7 2467 10.0 79.8
virtuallyhalved, andthis should have led to compensating imports 1120-1500 1286.2 2536 10.1 89.9
(to the tune of 21 million tonnes by 2001) had demand been 1500-1925 1692.2 2736 5.0 94.9
maintained at the 1998 level. The freeze on procurement price 1925 and more 3074.3 2938 5.0 99.9
which followed, when input prices and credit costs have been All 854.9 2156 99.9
rising, generalised deflation further to more farmers and both Summary
compounded the problem of deficient demand and sent strong Rs 500-575 and less; 1968 calories and less - 39.7 per cent
signals for cutting back output. Rather than restoring lost Rs 575-665; 2091 calories - 10 per cent
purchasing power and boosting aggregate demand by using up Rs 665-775 and more;2187 calories and more- 50.2 per cent
stocks for food-for-work programmes, the government exported Source: National Sample Survey Organisation (55th Round, 1999-2000)
22 million tonnes of grains at a highly subsidised price during ReportNo 471, NutritionalIntake in India, see p 22 for average
2002 and 2003 [Bhalla 2005], which was mainly used as animal calorie intake and average MPCEby expendituregroups. Report
No 454, Household ConsumerExpenditurein India- Key Results,
feed abroad. see pp 17-20 for the distributionof persons and average MPCEby
With its obtuse attack on the viability of farmers, the govern- expendituregroups.Distribution and average MPCEare the same for
ment has succeeded in taking India back to stagnant foodgrains bothreports.30-day recallthroughout.

Economic and Political Weekly July 28, 2007 3135

 Chart2a: Per Cent of Persons below Specified MPCELevels, Chart2b: Per Capita Daily Calorie Intake by MPCE,All-India
All-IndiaRural,55th Round 1999-2000 Rural,55th Round 1999-2000
100 3500

90
3000
80
70 2500

60
a(D 2000
*S 50
o 40 so1500
N- 30
D 1000
20
/ 0
500 ......... --. -.... . .....
10 ?

0 100 200 300 400 500 600 700 800 900 1000 0 200 400 600 800 1000 1200 1400 1600
MonthlyPer CapitaExpenditure
(MPCE),Rs (MPCE),Rs
MonthlyPer CapitaExpenditure
-- Series 1 -- Series 1

all goods and services, whose food expenditure component 28th round NSS data relating to 1973-74, and this expenditure
provided a daily rural energy intake of 2400 calories per capita was defined as the poverty line expenditure. (However, there is
and an urbanintake of 2100 calories per capita. While Dandekar doubt whether the 1973-74 poverty estimates are consistent with
and Rath (1971) had adopted a uniform nutrition norm of 2250 the declared norms, a matter discussed later.) Based on expen-
calories per head, the 1979 Task Force thought a uniform norm diture economists talk of "income poverty," but this is imprecise,
was inadequate, and adopted different norms for ruraland urban for we have no information on income, only on expenditure. It
areas. Using the census data projected to 1982, the population is possible that observed expenditure at or below the poverty line,
was divided into 16 groups defined by age, sex and activity, with is higher than income and is met through borrowing or asset-
energy intakes varying from 300 calories for children below 1 year depletion by some households. Conversely for those spending
to 3600 calories for a young man doing heavy work. The average well above the poverty line, income exceeds spending giving
normwas derived on the basis of this profile, and came to 2435 and rising savings. The latest complete large-sample published
2095 calories per person, ruraland urban, rounded down to 2400 data still remains the 55th round, 1999-2000, from which the
and 2100 calories per person, ruraland urban.Ruralenergy norms relevant information for all-India is given in Table 1. Some of
emerged as higher owing to the unskilled physical labour that the 61st round data for 2004.05 have been recently released, but
more rural workers perform compared to a higher proportion not the energy intake levels. Table 1 differs from other papers
doing lighter work in urban areas. Observed actual calorie intake because it combines, from two different reportsof the 55th round,
in rural India was also higher than intake in urban India from the distribution of persons by expenditure classes and their
the 1950s until the 1990s, after which with rural intake decline average expenditure, with information on the average calorie
and urbanintakerise, the position has been reversedby 1999-2000. intake of the same distributionof persons by the same expenditure
The NSS reports present the distribution of persons and classes.
average expenditure on food and non-food, by monthly per capita Thereby a good idea of the magnitude of headcount poverty
expenditure groups, and they also present the calorie intake per can be obtained easily by the non-specialist without making any
capita per diem by expenditure groups, though the latter tabu- calculations at all, simply by inspecting Table 1. Looking at the
lations are released after a time lag. The quantities of food items first, thirdand the fifth columns, 69.7 per cent or say seven-tenths
actually purchased by sample households are noted as are farm- of the rural population of India, spending less than Rs 525 per
produced food items retained for consumption by farmers.These month per person, was below the average calorie level of 2403
are valued at prevailing prices, and added to expenditure on non- (near-identical to the 2400 norm), which was obtained only by
food items to give the total monthly per capita expenditure. The the next higher spending group of Rs 525-615. Since the lower
different food items have specified calorie equivalents per partof this latter group, roughly half the 10.3 per cent of persons
kilogram, from which the calorie intake per day is derived. Thus in this group or about 5 per cent, also obtained below 2400
the very derivation of per capita expenditure on food is from calories, the actual total percentage of persons in poverty is about
exactly the same data set on physical quantities, which gives the three-quarters. On plotting the data on graphs we obtain 74.5
per capita calorie intake. There is a tight direct association per cent as the exact figure. Yet, the official Planning Commission
between monthly per capita expenditure and daily per capita figure of ruralpoverty from the same data is only 27.4 per cent!
calorie intake (Chart 3). The relation is non-linear as expected, The difference between percentage of population in poverty
with higher than unit elasticity of calorie intake with respect to obtained by direct inspection, 74.5 per cent and the figure as
expenditure at low spending levels. given by the Planning Commission, 27.4 per cent is very large.
Thatparticulartotal monthly per capita expenditure whose food Nearly half of the ruralpopulation - 47.1 per cent or 370 million
expenditure part met the calorie norms, was identified from the people - who are actually poor, are being excluded from the set

3136 Economicand PoliticalWeekly July 28, 2007

 Chart3a: Per Cent of Persons below Specified MPCELevels, Chart3b: Per Capita Daily Calorie Intake by MPCE,All-India
All-IndiaUrban,55th Round 1999-2000 Urban, 55th Round 1999-2000
100 Calorieintake
3500
90

80 3000

70 E 2500
60
C2000 /
a

40 ' 1500 -
'
30
1000
20
500
10
0
0 500 1000 1500 2000 2500 0 500 1000 1500 2000 2500 3000 3500
(MPCE),Rs
MonthlyPer CapitaExpenditure (MPCE),Rs
MonthlyPerCapitaExpenditure
-- Series 1 - Calorie intake

of the officially poor. Again, we see that nearly 40 per cent of Why does the official poverty line come to less than three-
the urbanpopulation spending below Rs 575 per capita per month fifths of the actual cost of accessing the nutrition norm and in
obtained less than 2091 calories (very close to the 2100 norm) what sense therefore is it any longer a "poverty line" at all? It
which was obtained only by the next higher spending group. Since is this unrealistic official "poverty line" - below Rs 11 per day
the lower half of this latter group also obtained less than 2100 for all goods and services - which results in the low poverty
calories, on plotting the graphs, the exact percentage in poverty estimate, excluding 47 per cent of the rural population who are
is 45 per cent. Yet the Planning Commission figure for urban actually poor. Clearly the reason is that the Planning Commission
poverty for the same year using the same data is only about half has not been applying its own original nutrition norm to the
of this at 23.5 per cent. available current data on nutrition by expenditure groups after
We only need to plot two simple graphs to see what is going the initial 1973-04 estimate, which was its first and only direct
on. First, (a) the ogive, or the cumulative distribution of persons estimate, but has been simply bringing forward the ruralpoverty
plotted against the upper-end value of each expenditure class - line for that year by using the Consumer Price Index for
this tells us what percentage of persons is below any given Agricultural Labourers (CPIAL).
expenditure level (col 5 against col 1) shown as Chart 2a and Thus an indirect method of price index adjustment to a base
3a relating to rural and urban India. Second, (b) the per capita year poverty line has been followed, without any reference to
calorie intake plotted against the per capita expenditure (col 3 the current cost of obtaining the nutrition norm, even though
against col 2) shown as Charts 2b and 3b relating to rural and information on this was regularly available from the five yearly
urban India - this enables us to read off the calorie intake at surveys. This amounts to computing a Laspeyres index in which
any given expenditure level. Consider the three variables: (1) the the quantities consumed in that base year are held unchanged
poverty line expenditure, or any other expenditure level, (2) the over time, adjustment being made only for price change. At this
estimated percentage of the population below the poverty line, poverty line however, the current consumption basket is such
or below any other expenditure level, and (3) the calorie norm, that the nutrition norm can no longer be accessed. The crucial
or any specified calorie intake. If we know the value of any one fact which is not mentioned to the public is that at the poverty
of the three variables, the corresponding values of the other two line of Rs 328 for all-India, food giving only 1890 calories daily
can be readoff from the graphs. The relation shown in Chart 2b could be purchased, over 500 calories below the RDA.
and 3b can be also plotted respectively on Charts 2a and 3a by Further, while for all states the official poverty line has been
taking the calorie intake values along the right hand Y axis, since too low and the corresponding nutritional intake well below the
the X axis is common to both. RDA, for a numberof states the use of state-specific price indices
The official ruralpoverty line of Rs 328 for 1999-2000 yields has meant that their official poverty lines have been pushed down
the poverty percentage of 27.4 using the ogive in Chart 2a. We so far below the average all-India level, that by the 55th round
find, using Chart 2b, that only 1890 calories could be obtained the rural consumer could access only 1440 calories to 1600
at this expenditure,over 500 calories per day less than the norm. calories, or a deficit of between 800 calories to nearly 1000
The true poverty line expenditure at which 2400 calories could calories perdiem from the nutritionnorm. These official "poverty
be accessed is Rs 565, and as high as 74.5 per cent of persons lines" have become a travesty of the very idea of poverty line
spent less than this amount - the correct estimate of poverty for and the corresponding poverty estimates - the percentage of
1999-2000. Similarly from Chart 3a we see that Rs 454, the persons below these lines - have lost all meaning.
official urban poverty line allowed only 1875 calories to be Rohini Nayyar(1991), in her careful doctoral study, discussing
accessed. In order to access 2100 calories (the RDA) the urban poverty estimates for the 1960s and 1970s, and Jaya Mehta and
consumer needed to spend Rs 625, and 45 per cent of persons Shanta Venkatraman(2000) discussing the 50th round, 1993-94,
were below this level.4 had already drawn attention to the inability of the price-adjusted

Economic and Political Weekly July 28, 2007 3137

 Chart4a: Declining Calorie Intake at Official All-IndiaRural for successive large-sample years and the same has been shown
Poverty Lines, 1973-74 to 1999-2000 in Chart 4a.
1600 With the nutritional intake accessible at the price-adjusted
- -* * * official poverty line steadily falling over the successive estimates,
1400
E 1200 the poor are being counted not as all those below an invariant
1000 nutrition standard but as all those below a standard which is being
, 800 continuously lowered over time. This very important fact, al-
though it is well known to the estimators, is never mentioned
o 600
by them in their papers. The price index adjustment to a base
0 400
year basket obviously has not only failed to capture the actual
200 current cost of accessing minimum nutrition at each point
of time, additionally the extent of failure has been increasing
1973-74 1977-78 1983 1993-94 1999-2000
fast over time.
- Calories at official poverty line less 1000
- It is not just the case that the particular price index being used
Official calorie norm less 1000
is the problem and there exists some "ideal price index" which
Notes: A constant 1000 calories have been deducted from both sets in lieu of
can capture the changing actual cost of accessing the required
starting the Y-axis values from 1000. See text for argument that actual
energy norm applied for initialofficial poverty line is likely to have been energy intake. Angus Deaton's exercise with alternative price
2200 calories. indices produced even lower poverty estimates than the official
Source: Table 2 line 6. one [Deaton 2003b]. The structural changes in the economy are
such that no price index applied to an invariant consumption
poverty lines to capture the currentcost of reaching the nutrition basket relating to 1973-74, can possibly capture the altered set
norm. The fact is also well known to the Planning Commission of choices that consumers face over time.
and to all the individual estimators following the same method The question is, why use at all, an indirect method of price
(whose papers were published in EPW, January25-31, 2003 and index adjustment to the cost of accessing an increasingly distant
later published in Deaton and Kozel 2005). What these latter base year consumption basket, with all its attendant problems,
economists still do not seem to understand, is that the method- when current data are available which permit the direct estimate
ological basis of their estimates is thereby rendered incorrect, of the poverty line every five years. (At most, the price-index
and the inference they draw regarding change in poverty over adjustment should be confined to the intra-quinquennial period
time or relative poverty across states, has no logical validity. We and thus the base year for the consumption basket should not
propose to show in this paper, the all-India and state estimates
of poverty obtained by the Planning Commission and by indi- Table 2: The Rural Poor as Per Cent of Rural Population
vidual academics who follow the same method, cannot be validly in India
compared over time and statements about rise or decline in RoundNo 28 32 38 50 55 61
poverty cannot be made. Nor at any given point of time, can the 1973- 1977- 1983 1993- 1999- 2004-
states be compared with respect to their extent of poverty. 74 78 94 2000 05
The gap between the official poverty lines and the actual cost Directmethod
of accessing the nutrition norm, was small to begin with but has 1 MPCEgiving
2400 kcal, Rs
been widening fast as the base year of the fixed consumption 56* 67 120 325 565 790
(povertyline)
basket, gets further back in time. The poverty lines derived by 2 Per cent below
bringing forward the 1973-74 rural poverty line of Rs 49 using povertyline 72* 65.5 70 74.5 74.5 87.0
the CPIAL, came to Rs 56 in 1977-78, Rs 86 in 1983, Rs 206 (77.5)
Indirectmethod
in 1993-94 and Rs 328 in 1999-2000, summarised in line 4 of 4 Priceadjusted
Table 2. The official poverty line for 2004-05 is Rs 356.3. The povertyline, Rs
NSS consumption data have been renderedirrelevantfor deriving official 49* 56 86 206 328 356
the official poverty lines. All that is used is the base year direct 5 Per cent of
officially'poor' 56.4 53.1 45.7 37.3 27.4 28.3
poverty line and the price index. (30.4)
These independently derived poverty lines, have been applied 6 Calorieintake
to the ogives from the NSS surveys, to arrive at the poverty at povertyline 2200* 2170 2060 1980 1890 1820
7 Deviationfrom
percentages, shown in line 5 of Table 7. They were 53.1 in RDAof 2400 kcal -200 -230 -340 -420 -510 -580
1977-78, 45.7 in 1983, 37.3 in 1993-94, 27.4 in 1999-2000 and 9 Modifiedprice-
29.5 (20.5) in 2004-05, the last estimatesbeing mine fromAppendix adjustedpoverty
ogives. These alternative estimates emerge from the 61st round, line, Rs, taking
base year MPCE
2004-05 from the uniform recall and mixed recall data. At the 2400 kcal 56 64 98 235 374 414
official poverty lines giving these poverty ratios, the maximum 10 Per cent which
calorieintakeaccessible perdiem was 2170 in 1977-78 (230 calories should be
below RDA), 2060 in 1983 (340 calories below RDA), 1990 officially'poor' 72 63 54 49.2 39 41.5
in 1993-94 (410 calories below RDA) and 1890 in 1999-2000 Note: *See textdiscussionthat2200 calorieswas the actualnormaccessible
(510 calories below RDA). The calorie level accessible at the withRs 49, the 2400 calorienormrequiredRs 56 as the povertyline,
61st round, 2004-05 poverty line is 1820, or a deficit of 580 and about72 per cent of persons was below this.
Source: PlanningCommissionforindirectestimates line3 onwards.Fordirect
calories from RDA. Line 7 of Table 2 shows the steadily increas- estimates, byconstructingCharts2 and3 foreach largesample round
ing deficit from energy intake RDA at the official poverty lines forwhichcaloriedata were available,fromNSS reports.

3138 Economic and Political Weekly July 28, 2007

 Chart 4b: Official and Directly Observed Poverty Line, Chart4c: Official and Directly Estimated Poverty Percentages
All-India Rural Derived by Applying the Respective Poverty Lines,
All-IndiaRural
600 90
80 - -- irect
. 500 > 9
/
70
C
. 60
400
rce 50
300 40
N 30
200 c 20
10 -
10
Q 100
0 1973-74 1977-78 1983 1993-94 1999-2000
1973-74 1977-78 1983 1993-94 1999-2000
-- Official PL -- -4- Official - Direct
Direct PL
Source: Table 7, lines 1 and 4.
Source:Table 7, lines 2 and 5.

be more than four to five years at most, before the next large data set they are using for expenditure, also give the calorie
sample data set become available.) intakes. As already mentioned, the data on physical quantities
This lowering of the nutrition standard over time, inherent in of foods, gives the calorie intakes on applying the standardtable
the official method, is the real reason for the poverty "decline" of calories per kilogram for different foods; and these same
claimed both in official poverty estimates, as well as in the physical quantities are valued and aggregated to give the food
individual estimates published in the EPW, January25-31, 2003 expenditure, which is added to other spending to give the total
issue - which quite clearly is a spurious decline, for no valid expenditure.
comparison over time is possible when the standard is being It is not proper academic procedure to use data selectively -
lowered (or altered in any way). To give an analogy, suppose to use the expenditure data while ignoring and never mentioning
we are watching an Olympic high jump event not directly but the necessarily associatedenergy intake, as is being done by these
mediated through television, where the camera focuses only on estimators. Their poverty numbers would certainly have been
the successive jumps. At the first try the jumperjust barely clears questioned much earlier if this information was known to the
the bar, at the second try she clears the bar by three inches and educated public. The Planning Commission has never officially
at the third try she clears the bar by six inches. It is claimed that given up the nutritionnorms on the basis of which ruraland urban
the performance has improved greatly over the successive tries poverty was defined. The majority of economists in India believe
and everyone believes the claim. However without anyone's that these norms are still being followed. The reality is however
knowledge, the bar has been lowered by six inches for the second that the actual estimation procedure has meant giving up not just
try compared to the first and again by six inches for the third these particular nutrition norms, but has meant giving up any
try compared to the second. The actual situation is that the nutrition norm whatsoever. The question of nutrition has been
performancehas worsened and thejumper isjumping three inches rendered irrelevant in the official method.
lower at the second attempt and six inches lower at the third There is not even any lower bound which is set to the fall in
attempt compared to the first attempt. Obviously the claim of the energy intake corresponding to official poverty lines - for
"improvement"is spurious and moreover it involves suppression
of information since the fact of the lowering of the bar is kept Table 3: MonthlyPer Capita Expenditure in Rupees in 61st
carefully hidden from the public. Round, 2004-05 Compared to 50th Round, 1993-94 by Groups
The "bar"has been lowered by about 100 calories per diem of Persons, All-IndiaRural
for all-India, for every successive five-yearly estimate since Per cent Average (2) Actual MPCE MPCE (6)
1977-78 and by year 2000 it was about 500 calories per diem of Persons, MPCE, Deflated MPCE, on on Adjusted
lower than RDA on average (Chart4a). For some states however 2004-05 Actual by Index 1993-94 Food, Food, by Index
2004-05 to 1993-94 (U 30) Actual Actual to
it had been lowered by 250 calories only and for others by as (U-30) 2004-05 1993-94 2004-05
much as 960 calories per diem, owing to state-specific price (1) (2) (3) (4) (5) (6) (7)
indices being applied (Table 5).
Lowest
Official and individual claims of poverty reduction in the 1990s 30.3 289.90 131.1 150.0 191.49 108.73 240.3
are spurious and arise from this clandestine lowering of the Next
19.6 408.98 185.0 211.0 259.53 149.47 330.3
consumption standard,a lowering which is inherentin the official Next
estimation method itself, which has de-linked estimation from 30.5 552.94 250.2 285.3 295.75 192.34 425.1
the nutritionnorm after 1973-74. The term "clandestine" is used Next
advisedly because unfortunately, neither the Planning Commis- 14.7 853.04 386.0 432.0 446.26 260.05 574.7
sion economists nor a single one of the other academics present- Top
4.9 1956.57 885.3 872.0 659.13 370.49 818.8
ing their poverty estimate using the official price adjustment 100 558.78 252.8 281.0 307.60 177.77 392.8
method, have considered it necessary to mention the crucial fact
Note: U-30 is uniform30-day recallforall goods.
of the lowered calorie intake corresponding to their own poverty
Source: NSS ReportNo 508, 61st Round, Level and Patternof Consumer
estimates for different points of time, when publishing their Expenditurein India,2004-05 ( A-12, A-240 ) and ReportsNos 401,
papers,although they are well aware of it since exactly the same 402, 50th Round.

Economicand PoliticalWeekly July 28, 2007 3139

as many as seven states it has already fallen to 1450 to 1720 Rs 56, and line 10 gives the derived poverty percentages. The
calories by the 55th round (Table 5), and single-digit poverty difference by 2004-05 is quite large - the poverty line should
levels are being claimed for some, although in reality poverty have been Rs 414 and the poverty percentage 41.5 and not 29.5,
is very high. By the 61st round for many states calorie intake even using the faulty official method, if the RDA had been
accessible at official poverty lines will be between 1300 and 1400, actually applied in the base year.
over 1000 calories below RDA. Thus a completely different
measure entailing a different definition of "poverty" is being IV
used, compared to that adhered to theoretically. This definition IncreasingUnderestimation
Cumulatively
will logically lead to further absurd claims of great "success" overTime
in poverty reduction when the official estimates for most states
in India reach single-digit levels as they will soon do. The real If the official procedure has always led to spurious poverty
reason would be that the poverty line is far too low for anyone reduction, why has the extent of such reduction been much greater
except the poorest of the poor tribal people and some unfortunate in more recent years during the 1990s, compared to earlier
destitutes and beggars, to survive below it. decades? From Table 2, during the decade 1973 to 1983 there
The logically correct method of comparison is to count the poor was a decline by about 10 points from 56 to 46; over the next
below a temporally and spatially unchanged consumption norm, decade to 1993-94 there was a decline by 9 points to 37, but
for then the same definition of poverty line is applied for suc- over a mere six years from 1993-94 to 1999-2000, the decline
cessive estimates and for different states. A simple and trans- was by 10 points to 27. It is the large decline by 10 points over
parent measure of changing poverty depth is to take lower-than- only six years in the 1990s, which made people sit up and take
RDA energy cut-offs (say, 2100 and 1800 calories) and note the notice of poverty estimates. Urban official poverty percentages
percentage below these levels, keeping the levels unchanged over too are lower by a massive 15 points during the dozen years 1987-
time. I have applied this direct method to obtain the poverty 88 to 1999-2000 compared to much smaller official declines in
percentages for all large sample rounds (all-India rural, Table 2 the 15 years preceding 1987-88.
line 1) while Table 5 gives the state estimates for the 50th and We get the answer to larger spurious declines over time, if we
55th rounds. We find that actual rural poverty is very high, it remember, first that the distribution of persons by expenditure
has not declined but on the contrary has risen in 10 out of 15 is skewed, with two-thirds spending less than the mean expen-
states, and the depth of poverty has increased during the 1990s diture in both rural and urban India (which is reflected both in
in nine states. the slope of the ogives and the non-linear relation of calorie
Many authors have pointed out that the estimation basis for intake to spending). Both curves rise steeply as we go from
the initial official poverty lines was itself opaque. The relevant very low to medium levels of MPCE, then rise less steeply and
nutritional data for 1973-74 were never published and the thereafter level off at high spending levels. Second, the official
estimate was based on a limited nine-month sample [Mehta and method has been underestimatingthe poverty line in six separate
Venkatraman2000; Rath 2003]. Plotting the NSS data for 1970- and sequential five-yearly episodes of price adjustment over 31
71, for which calorie intakes were derived in R Nayyar (1991), years, resulting in a cumulative large deficit from the true poverty
we find that 72 per cent and 54 per cent of the population was line. The initial official poverty lines in the 1970s were not too
below 2400 and 2200 calories. This suggests that the official distant from the correct poverty lines required to access the
1'73-74 estimate of 56.4 per cent in poverty is not of the right nutrition norm, but became cumulatively distant from the true
order of magnitude to correspond to the official norm of 2400 poverty line over time, intersecting the ogive at its lower and
calories RDA. The period 1970-71 to 1973-74 was of rapid food increasingly steeper segments, and this led to larger spurious
price inflation which gave rise to widespread unrest and to the poverty decline.
price rise resistance movement led by Jaiprakash Narayan. To illustrate: looking at the slope of the 1999-2000 rural
Inflation did not moderate until the draconian laws of the ogive in Chart 2a at the correct poverty line of Rs 565 at which
Emergency period. It is impossible that using the 2400 norm the 2400 calories could be accessed, Rs 50, say, of underestimation
poverty percentage could have declined to such a large extent of the poverty line or PL from this point to Rs 515 reduces
over a mere three years of rapid inflation, from 72 per cent in the poverty percentage only by 5 to 6 points. But at an official
1970-71 to 56 per cent by 1973-74. PL of say, Rs 450 which is already substantially lower than
The official 56.4 per cent figure for 1973-74 is however the true one, (the actual case by the late 1980s) Rs 50
entirely consistent with a 2200 calorie norm. Our hypothesis underestimation to Rs 400, leads to a 10 toll points decline
is that the initial official estimate itself was fudged, perhaps in the percentage of persons below this line given the steeper
because 72 per cent or more of the population in poverty yielded segment of the curve.
by the RDA, appeared far too "alarming". This would explain If the official PL is already at a severely underestimated level
the non-transparency - probably quite deliberate - of the of Rs 380, an additional Rs 50 decline to Rs 330 (actually the
basis of the estimate, that other writers have noted. Another official PL for 1999-2000), leads to a massive 15 points decline
quick check: in 1970-71, the expenditure enabling a ruralperson in poverty, because we are almost at the extreme left hand end
to access 2400 calories was Rs 40, and since the CPIAL rose of the distribution by now where the ogive is steepest. There
by 40 per cent it should have been Rs 56 at least by 1973-74 is a sharper drop in the percentage of people still surviving at
and not Rs 49, the official figure. The same argument applies such low expenditure and calorie intake levels.
to the urban poverty line, which should have been higher This argumentusing a single ogive to talk of change over time,
than stated. of course assumes that the ogives when drawn in terms of real
Table 2, line 9, gives the price-index adjusted poverty lines expenditure are not shifting over time. But even if conditions
appropriatefor a 2400 calorie norm in the base year which cost are actually getting worse, the real ogives are shifting leftwards

3140 Economic and Political Weekly July 28, 2007

and actual poverty is rising, provided these leftward shifts are The entire rural population except the top 5 per cent, shows
still small, clearly we would still get an overall net decline lower real expenditure in 2004-05 when we adjust by this index
in official poverty percentages every five years owing to the and compare with 1993-94 (Table 3). The bottom 80 per cent
cumulatively larger underestimation bias in the latter's poverty of persons needed to spend 14 per cent more than they actually
lines. This is what we do observe up to the 55th round, 1999- could, to maintain the same real spending as a decade earlier.
2000. The real ogives have been shifting leftwards and actual Adjusting the 1993-94 expenditure on food upwards by the index
poverty has been rising slowly since 1977-78 to 1999-2000 shows that average actual expenditure in 2004-05 was lower than
(Chart4c), but the cumulatively larger underestimation bias in adjusted expenditure for all groups. Average expenditure needed
the official poverty lines has led to the actual rise being more to be higher by about one-quarterfrom actual; even the top group
than outweighed, and to a greater extent each time, thus showing spent less than required and the unfavourable gap was relatively
up as spurious poverty decline. more for some of the poorer spending groups. The observed
The official poverty line for 2004-05 is Rs 356 and the poverty average decline over time in the share of food expenditure in
ratio is 28.3. The official poverty percentage has not ceased to total expenditure, in such a specific context where per capita real
be spurious;thatit does not show a furtherlarge decline as earlier, expenditure although initially at low levels declines furtherover
I would argue is because there has been a very much larger time, represents a case Ernst Engel did not explore and indicates
leftward shift than ever before, in the rural ogive during the five exactly the opposite of betterment. It is not surprising that it is
years after 2000, as agrariandepression has intensified and real accompanied by decline in calorie intake and rise in poverty.5
income decline (owing to unemployment) has become more The actual decline of incomes has been greater than that of
pervasive, engulfing larger groups of people - an adverse shift expenditure, as the Situation Assessment Survey of Farmers by
so large, that it has neutralised the built-in large underestimation the NSS show. Table 4 summarises the striking results at the
bias in the official procedure. all-India level, of consumption and net investment expenditures
This is supportedby the nutritiondata which have been recently relative to income from all sources. For over 96 per cent of
released. The direct estimate of the poverty line requiredto access farming households total income from all sources did not cover
2400 calories in 2004-05 is Rs 795 and an all-time record high, 87 consumption expenditure and led to deficit. In many states deficits
percent of the populationis below this level (see Appendix tables). were financed through asset depletion by the majority who
reported negative investment (see Tables A-178 to A-192 of
V Report 497) and for all-India, net investment per household on
InitialFindingsfromthe 61st Round,2004-05 productive assets was a paltry Rs 124 per month. Admittedly
2002-03 was a drought year but even if we arbitrarily reduce
The proposition that (a) there has been a substantial worsening these numbers by 20 per cent for a normal year, over 75 per cent
of income distribution, and (b) that the worsening has been of of all farming households would still be in deficit.
a particulartype, namely, absolute real decline in ruralincomes, The realistic poverty line in 2003 would be about Rs 610 per
is consistent with the 61st round expenditure data. The CPIAL month per head (adjusting the 1999-2000 direct poverty line of
does not capture change in the cost of living adequately, so we Rs 565 for price change) and given the average family size of
use the ruraldirect poverty lines for the 2200 calories level (the 5 members, Rs 3,050 per month is the minimum realistic poverty
base year actual nutrition norm in official estimates) from the line per household. Some 80 per cent of all households in Table 4
50th and 61st rounds to construct an index. The poverty lines spent less than this on consumption and still evidently had to
are Rs 260 and Rs 575, rising by 121 per cent, compared to the finance it through borrowing or asset depletion to the extent of
rise by 76 per cent in CPIAL. The interpretation is as follows the shortfall of income.
- unchanged real expenditure using this index means that nu- Further,the Land and Livestock Surveys of the NSS for 1992
tritionalaccess is preserved at the same specified level as before, and 2003 (Reports 408 and 493) show a large rise from 22 to
without assuming a constant consumption basket or unchanged 32 per cent of households with nil operated land. In Andhra
economic environment. Pradesh households with nil operational holdings rose from

Table 4: All-India Rural Monthly Expenditure from All Sources, Consumption Expenditure and Investment
in Productive Assets (Rs), 2002-03
Area Net IncomeReceipts
Possessed Wages Cultivation Animal NFB Total Consumption Balance Investment Surplus/ Per Cent Cumulative
Ha Income Farming in Productive Deficit of Per Cent
Assets HH of HH
1 2 3 4 5 6 7 8 =(6-7) 9 10(8-9) 11 12
< 0.01 1075 11 64 230 1380 2297 -917 40 -957 11.6 11.6
0.01 - 0.4 973 296 94 270 1633 2390 -757 37 -794 34.0 45.6
0.04 - 1.0 720 784 112 193 1809 2672 -863 96 -959 27.6 73.2
1.0 - 2.0 635 1578 102 178 2493 3148 -655 151 - 806 15.1 88.3
2.0 - 4.0 637 2685 57 210 3589 3685 - 96 387 -483 7.9 96.2
4.0- 10.0 486 4676 12 507 5681 4626 1055 685 370 3.3 99.5
> 10.0 557 8321 113 676 9667 6418 3249 737 2512 0.5 100.0
ALL 819 969 91 236 2115 2770 -655 124 -779 100
Notes: Column8 is (Col6-Col 7) and Column10 is [Col6 - (col 7 + col 9)] and these have been calculatedby the author.Note thatonly the top 3.8 per cent
of all households earned enough to meet consumptionexpenditure.
Source: 59th Round,NSS ReportNo 497, Income, Expenditureand ProductiveAssets of FarmerHouseholds,Table A-192.

Economic and Political Weekly July 28, 2007 3141

37 to 53 per cent of all rural households, in Tamil Nadu from inference "every student knows exactly where she lives" is not
36 to 67 per cent, and in Kerala from 6 to 38.6 per cent. The true. But fallacies of equivocation in economics are more difficult
nil holdings percentage in operational holdings has doubled in to spot. Intelligent non-specialists do not scrutinise arguments
Haryana,Bihar and West Bengal, all from around 14-15 per cent by economists carefully, nor do fellow economists not hitherto
to 28-30 per cent. Effectively much of the gains of past land working in that particulararea, because they trust the specialists
reforms have been reversed by the impact of state expenditure at the intellectual level.6 They take it for grantedthatterms which
deflation and market-orientedreforms in unleashing ruraldepres- express concepts, must be correctly used by these trained pro-
sion and impoverishment, forcing poor and small peasant owners fessional scholars. This is a reasonable expectation but unfor-
to part with livestock and land. tunately it is by no means always realised, as the official method
and the uncritical use of the same method by individual econo-
VI mists following the 1993 Expert Group report, shows.
The Fallacyof Equivocation The official poverty estimation method discussed in the pre-
vious sections provides an excellent example of the fallacy of
The official and individual estimators follow the procedure of equivocation. The issue turns on declaring a particularconcept
the 1993 'Reportof the ExpertGroup on Estimation of Proportion anddefinition of the term"povertyline expenditure"and applying
and Number of Poor'. This had made two main recommendations it in a particular year, but then using a completely different
- first, a long-overdue one, that the earlier practice should be definition of "povertyline expenditure",and improperlydrawing
discontinued, of blowing up the NSS fractile-specific consump- the inference that "poverty"has declined. The fallacy of equivo-
tion figures by using the ratio of the aggregate CSO consumption cation thus arises because the term "poverty line" is used in two
estimate to the NSS consumption estimate. It had also recom- different senses in the course of the same argument, so the
mended that state-specific price indices should be used to inference about change in poverty, is not true. The fallacy has
estimate the state poverty lines. But, unfortunately, the Expert been committed by the Planning Commission in India since
Group did not consider departing from the indirect method of 1973-74, by the 1993 Expert Group which recommended
price-adjustmentin favour of the direct method for all previous continuing with the same fallacious method, and by a number
estimates, nor did it bring the base year for the consumption of individual economists uncritically following the fallacious
basket, forward to 1993-94 as it could have done. This would procedure advised by the 1993 Expert Group.
have meant, taking RDA of 2400 calories, a rural poverty line Some academics try to rescue their erringpeers in an empiricist
for 1993-94 of Rs 325 and not Rs 206, and would have given manner, by saying that the de facto nutrition norm has been
a price adjusted poverty line by 1999-2000 of Rs 517, below lowered a bit from the dejure one, and it is not such an important
which 68 per cent of the rural population is observed to fall. matter to make a fuss about. They point out that bodies like the
While an underestimate it would not have been so grossly off UN Food and Agriculture Organisation have been suggesting of
the mark as current official estimates are. The Expert Group late, lower figures of 2110 calories for south Asia and an even
however recommended continuing with the same method of price lower level of 1810 for India as a minimum. It is indeed a fact
adjustment to a by then two-decade old consumption basket. that having signally failed to reduce poverty itself, all inter-
It is still not clear why so many academics in universities should national bodies which talk of poverty reduction are lowering the
have uncritically followed the Expert Group and treated a mere nutritionnorms instead or applying purchasingpower adjustment
report as the Vedas and the Upanishads, ignoring all critical to deflate the dollar a day poverty lines, thereby sanitising their
voices. The poverty lines calculated according to the Expert global poverty estimates to lower and less embarrassing ones.
Group method, continued to be de-linked from the necessity of But such empiricistrescue efforts simply carryno conviction when
satisfying any nutrition norm at all. This de facto deviation from we see what abandoning the nutrition norm has done to official
the original definition of poverty has had far-reaching method- poverty lines andhence poverty estimates in India:they have been
ological implications, which have not been fully appreciated by reduced to conceptual garbage as Table 5 demonstrates.
the academic community. As we have seen, it renders logically No international body has said, or can dare to say that 1400
invalid every attempt to compare the extent of poverty, both to 1600 calories are acceptable nutrition norms for developing
across states at a given point of time as well as over time both countries (the average intake in advanced countries is around
for individual states and at the all-India level. The precise type 3,000 calories). Not even P V Sukhatme, a most vigorous cam-
of logical fallacy involved is the fallacy of equivocation. paigner for a below 2400 calories norm, would have agreed that
The fallacy of equivocation is a specific type of verbal fallacy, a 1700 calories or less daily intake per capita for any population,
in which the same term is improperly used with two different was reasonable - he used a 2200 calories norm in one of his
meanings in the course of the argument to draw the inference, own estimates [Sukhatme 1971]. Sub-human to very low energy
which therefore is not true. Modern books on logic follow intake levels of 1450 to1700 calories however, by 1999-2000
Aristotle's classification of fallacies (Aristotle's De Sophisticis are associated with the official poverty lines for many states
Elenchis or 'Of Sophistical Refutations') supplemented by recent (Andhra Pradesh, Gujarat, Kerala, Tamil Nadu), while Punjab
analysis [Barnes 1984; Hamblin 1970; Thouless 1974]. and Haryana are very close with 1720 calories or less being
We can construct an example of the fallacy of equivocation accessible at their official poverty lines.
as follows: "The professor has been delivering her address for There is a debate among the academics following the official
one hour to the gathering of students. Therefore every student indirect method, that owing to change in the recall period during
knows exactly where she lives." The term "address"is being used the 55th round, 1999-2000 compared to earlier rounds, actual
in two quite different senses in the premise and in the conclusion expenditure is overstated. Making the required adjustment for
- "address " in the sense of speech, and "address" in the sense comparability alters the ogive slightly and raises the 27 per cent
of place of habitation.There is equivocal use of the term, so the below the Rs 328 official poverty line, by another I per cent

3142 Economic and Political Weekly July 28, 2007

according to Sundaramand Tendulkar (2003), and by 3 per cent relations for each of the 15 large states for the four large sample
according to Deaton (2003a). The NSS report however says that years after 1973-74, for which calorie data were available, and
the 50th and 61st rounds are comparable using uniform 30 day the ogives for the 61st round, 2004-05.7
recall in both while the 55th roundis not comparableat all with the It is our exercise with the state poverty estimates which bring
50th one but is comparable with the 61st round mixed recall. No out starkly, how the official method has led to a most bizarre
doubt therewill be yet anotherrecondite debate on comparability. and arbitraryvariation of the calorie intake levels accessible at
The lack of comparability arising from alteration in the recall the poverty lines. The range of variation in the 50th round, 1993-
period, however, is of relatively small importance, compared to 94 is from 1625 calories in Kerala to 2230 calories in Orissa
the fundamental problem of lack of comparability arising from and Uttar.Pradesh, with the all-India figure standing at 1980
the unstated alteration in the consumption standard inherent in calories. By the 55th round, there is furtherdecline in the calorie
the indirect method all these estimators uncritically use. The main intake at poverty lines in every state (except only Gujarat): the
analytical point being made in this paper focuses on this mistake range now being from 1440 calories in Kerala to 2120 calories
which leaves out half the ruralpopulation which is actually poor, in Orissa with the all-India figure dropping further to 1890
andthis basic problemwith all indirectestimates remainswhatever calories. All southern states have extremely low official poverty
adjustments might be made for recall period. lines, at which the calorie intakes were 1600 in Karnataka, 1590
in Andhra Pradesh and 1510 in Tamil Nadu. Clearly the poverty
VII estimate within any state is not comparable over time - except
Reductionin NutritionAccessible at Poverty for Gujarat,where although the official poverty lines and hence
Lines in ManyStates the poverty percentages are fartoo low giving below 1700 calories
intake in both the 50th and 55th rounds, there is no furtherdecline
The public is never informed, when poverty estimates are in intake over the period.
quoted, of the dilution of the energy intake norm leading to The official estimate of poverty for Orissa was 48 per cent,
spuriousestimates and claims of poverty reduction. Large though over four times higher than that for neighbouring Andhra Pradesh
the dilution is, it does not prepareus for the truly heroic reduction at only 11 per cent. But how can we possibly compare and infer
of the consumption level accessible at official poverty lines in that Orissa was poorer than AP once we know that the officially
many states, owing to the extremely low state-specific poverty poor in AP are all those persons consuming below 1590 calories
lines being applied. while the officially poor in Orissa are all those consuming below
How do we obtain the calorie intakes at the official state poverty 2120 calories? The directly measured poverty in Orissa was lower
lines? The basic data are available in the same format for each than in AP and poverty depth was also substantially less, those
individual state as the all-India data in Table I for each large accessing below 2100 calories being 46 per cent and 62 per cent
sample round barring one. By plotting for each state the same in Orissa and AP. Similarly the 13 per cent official poverty figure
two curves - the ogive and the relation between average per capita for Gujarat cannot be compared with the 44 per cent for Bihar
expenditure and average per capita calorie intake, we can obtain and the former state said to be less poor, when we see that the
the energy intake accessible at the official state poverty lines. calorie intake accessible at its poverty line has been pushed down
In all I have plotted 135 relations - graphs containing the two to 1680 in Gujarat compared to 2010 in Bihar. Actual poverty
Table 5: Planning Commission Poverty Estimates by States and Calorie Intake at Official Poverty Lines Compared to Direct
Poverty Estimates
IndirectOfficialEstimate DirectEstimate
1993-94 1999-2000 1993-94 1999-2000 1993-94 1999-2000
Official Calorie Official Calorie <2400 < 2400 < 2100 < 2100
Poverty Intake Poverty Intake Calories Calories Calories Calories
at PL at PL at PL at PL Poverty Poverty Poverty Poverty
(Per Cent) (Per Cent) (Per Cent) (Per Cent) (Per Cent) (Per Cent)
1 2 3 4 5 6 7 8
All-India 37 1980 27.4 1890 75 74.5 49.2 49.5
East
Assam 45.0 1935 40.0 1790 93 91 62 71.0
Bihar 58.2 2150 44.3 2010 73 78 51 53.5
Orissa 49.7 2230 48.0 2120 70 79 42.5 45.5
West Bengal 40.8 2080 31.9 1900 72 81 42.5 55.0
South
AndhraPradesh 15.9 1650 11.1 1590 84 84 56 62.0
Karnataka 29.9 1815 17.3 1600 75.5 82.5 57 50.0
Kerala 25.8 1625 9.4 1440 84 82.5 64 60.0
TamilNadu 32.5 1650 20.6 1510 87 95 77.5 76.0
West-central
Gujarat 22.2 1660 13.2 1680 83.5 85.0 64 68.5
MadhyaPradesh 40.6 2010 37.1 1850 72.5 78 47.5 57.5
MahaRashtra 37.9 1820 23.7 1760 89.5 92 75 55.0
Rajasthan 26.5 2100 13.7 1925 46 52.5 26.5 27.5
North
Punjab 12.0 1825 6.4 1710 52.5 58.5 30 36.5
Haryana 28.0 1990 8.3 1720 55 47.5 34 30.5
UttarPradesh 48.3 2230 31.2 2040 65.5 61.5 38.5 37.5
Notes: Figuresin bracketsforall-Indiaindicateroughadjustmentforrecall-periodchange; no adjustmentis shown forthe states.
Source: As Table 1 and NSS ReportNos 401,402,405.

Economic and Political Weekly July 28, 2007 3143

incidence in Bihar was less than in Gujarat and poverty depth are celebrating the alleged "disappearance"of poverty in rural
was also less as the last two columns show. Punjabon the basis of such selective use of the NSS data, ignoring
Clearly, the official poverty estimates are not comparable the dimension of nutrition completely. The reality is that land-
across states at any given point of time, and they are not com- lessness has gone up in ruralPunjab by nearly 10 per cent points
parableacross time in any state. They no longer make any sense. between 1992 and 2003 (NSS Report493), ruralpoverty had risen
The deafening silence of all the other economists using the to nearlythree-fifths,andover 36 percent were below 2100 calories
same indirect method, on the declining nutritional intake nec- intake compared to 30 per cent five years earlier. The spurious
essarily associated with their own estimated state-wise poverty official Indian poverty estimates are feeding into and rendering
lines, ignore the basic requirementof academic work that it must equally spurious the World Bank estimates both for India and
follow the principles of logic and of transparency. Academic with respect to its global poverty line, whose estimation basis
work cannot be treated in such a cavalier manner, where data has been usefully explained by Reddy and Pogge (2005). For
are used selectively, important information is suppressed, and the year 2000, a uniform poverty line of $ 1.08 a day was derived
thereby elementary logical principles that the world has known by the World Bank, by taking the existing lowest (hence rural)
for 2,000 years, areopenly flouted. The numbersthese economists local currency official poverty lines of 10 poorest countries out
are producing on Indian poverty at the Planning Commission and of a set of 33 countries, "poorest" as identified after converting
at the World Bank, are feeding directly into the making of policy their local currency monthly poverty lines to dollar and adjusting
which affect the lives of millions of poor people. It is surely for purchasing power by using the 1993 PPP conversion factor
incumbent on the concerned economists that they show a more for consumption. The Bank then took the median value of the
responsible attitudeto their own academic work. They can hardly 10 values. This procedure gave Rs 7.51 per day for India at 1993
expect to retain credibility if they continue to pretend that the PPP, and after updating this with a domestic price index and
methodological criticisms over the last decade do not exist. In applying to the distributionof persons by expenditure levels, 35.3
the matter of logical mistakes, there is no strength to be derived per cent in rural India were stated to be "poor".9The implied
from collecting together in large numbers. The fact that not one rupee poverty line is Rs 357 per month or Rs 11.9 per day.
or two, but 20 or more economists are using a logically incorrect The argument that for international comparison, the already
method, does not renderthe method a correctone. The fact that 20 low dollar a day poverty line should be adjusted downwards to
economists and not one or two, are producing senseless numbers only one-quarterto one-fifth, according to the varying purchasing
in the name of poverty estimates, simply becomes a sad comment power of developing country currencies, makes no economic
on the falling standardsof intellectual work not only in our own sense. The unstated assumption has to be that this is a reasonable
universities but also in institutions abroadand at the World Bank. daily poverty line for the US, but clearly it is not, for it would have
Alreadythe false povertyestimatesandspuriousclaims of decline bought at most either one bottle of water there, or 2.5 Ibs of rice,
have played havoc with food security and increased hunger. just as the PPP adjusted$ 1.08 in India (which deflates its nominal
Millions of very poor people have been priced out from the PDS rupee equivalent to about a quarter), could barely buy either a
by labelling them incorrectlyas APL (above poverty line) and now single bottle of water, or at most 1.1 kg rice. Even the reverse
moves are afoot to exclude the so-called APL completely from adjustmentto the one-dollar poverty line, namely takinga multiple
the PDS whose scope and operationsare being run down. At some according to purchasing power, would not give us anything but
level simple common sense appears to have been abandoned by a travesty of a poverty line for the US. Thus, today, one US dollar
the estimators. Since we are not talking of historical data, the when spent within India buys exactly as much as Rs 44.5 does
currentcost of living should be known to them from their own (which is nearly four times the official poverty line). While the
daily experience. It is strange that any economist can seriously purchasing power of one US dollar is about a quarterin the US,
propose that Rs 10 to 12 per day even in an Indian village today surely it is not the case that $ 4 per day, or less than $ 1,500
can meet one person's expenditure on all food and non-food per year, would be a reasonable per capita poverty line for the
requirements, inclusive of the value of farm-produced output. US. How can it be maintained that one-sixteenth of this level
In reality it would buy one kilogram of the cheapest rice on or 25 cents is an adequate poverty line for India? Of course, the
the open market,and nothing else, or one litre of bottled drinking problem has arisen precisely because the $ 1 a-day (or, at present,
water. The official poverty line was lowest in Andhra Pradesh the $ 1.08 a day) measure itself is derived from the unrealistically
at Rs 263 per month or Rs 8.7 per day. Only 11 per cent of the low national official poverty lines of developing countries. Even
ruralpopulation was below this spending level, at which at most the higher of China's two rural poverty lines, 800 yuan per year
1590 calories was accessible. No doubt they belong to the poorest or 2.2 yuan per day, is absurdly low and is equivalent to Rs 11,
of the poor even among the tribal and dalit groups. We can well exactly the same as India's poverty line, at the prevailing ex-
imagine how much more adverse their morbidity and mortality change rate. China's official rural poverty figures too are gross
rates would be in relation to already adverse average rurallevels. underestimates, for with relentless market reforms and user
These unfortunatepersons would be on their way to early death. charges such a paltry sum spells destitution.
Drastic lowering of the calorie intake associated with extremely
low poverty lines are necessarily also implied in the same pro- VIII
cedure followed by the individual academics. At Deaton's StatewiseVariations
andTrendsin Actually
recalculatedmonthly poverty line for Punjabof Rs 316.5 [Deaton
2003b: 367, Table 5], we find from our charts that only 1480
ExistingPoverty
calories were accessible. No wonder only 2.7 per cent of Punjab's Poverty is officially underestimated to the greatest extent in
ruralpopulation in his estimate were "poor"since 1480 calories south India with 800 to 1000 calories per diem deficit from RDA
is a semi-starvation level, costing Rs 50 less than the very low in every state. Although Andhra Pradesh had 84 per cent of rural
official poverty line giving 1710 calories.8 Yet some economists population in poverty during both 199 4 and 1999-2000, its
1993-94

3144 Economicand PoliticalWeekly July 28, 2007

 Chart5a: Deficit of Calorie Intakefrom RDAat Official PL, Chart5b: Deficit of Calorie Intakefrom RDA at Official PL,
South India West-Central India
I .
Kerala ') . /
Rajasthan

TamilNadu Madhya '.-Z", -2 /,

Pradesh

Karnataka Maharashtra

Andhra Gujarat
Pradesh
0 200 400 600 800 1000 1200 0 200 400 .600 800
Deficit of Calories from RDA at OfficialPoverty Line Deficitof CaloriesfromDiem
0 1993-94 0 1999-2000
*1993-94 E1999-2000

government was congratulated by the Planning Commission on for thechangein recallperiodandthe rise in povertyin these 11
stateswouldbe greaterif this was done. In the remainingstate,
reducing ruralpoverty to 11 per cent. It was not mentioned that
the 1993-94 official poverty line was so low it allowed 1650AndhraPradesh,povertydepthhas increased,while Karnataka
calories only to be accessed and this further declined to 1590
registersmoderately loweredpovertydepthdespiterisein poverty.
calories at the 1999-2000 poverty line. The case is similar inThe only statein thecountrywhichhas reducedpovertydepth
Tamil Nadu and Karnataka. The fact that Kerala has always very substantiallyduringthe economic reformsperioddespite
overallpovertyrisinga bit, interestingly,is Maharashtra
historically shown the lowest average calorie intake but good where
the percentagebelow 2100 calorieshas fallen drasticallyfrom
performanceon vital rates (low death rate, low IMR, low maternal
75 to 55 while the below 1800 caloriespercentage(not shown),
mortality)has led to a greatdeal of complacence in official circles.
has also fallen from 38 to 26. This large reductionin poverty
It is forgotten that equality of access to food which is a function
depthis undoubtedlythe positive resultof Maharashtra's
of a relatively less unequal asset and income distribution and long-
standingemploymentguaranteescheme and is a good augury
a well-functioning PDS, is an important factor, and that "while
the level of dietary inadequacy is undoubtedly the dominant for the currentNational Rural EmploymentGuaranteeAct,
providedit is properlyimplemented.Of course,thdi*e1999-2000
determinantof under- nutrition, the level of primary healthcare
datapredatetheproblemsof cottonfarmersandpervasivesuicides
in the community can significantly modify the severity of its
in the Vidarbharegion of Maharashtra.
clinical manifestations" [Gopalan 1992]. The special character-
The rise in povertyin West Bengal duringthe 1990s might
istics of Kerala however are not to be found in other states, and
surprisesome, given the positive effects of land reformsand
even in Kerala the reform decade has impacted hard on farmers
with agrarian depression and suicides. revivedfunctioningof panchayatsin thatstatesince 1978.In fact,
between1977-78and 1993-94therewas a largedropin poverty
In west-central India poverty is officially underestimatedto the
inWestBengal,thepercentageof personswithintakebelow2400
largest degree in Gujarat and Maharashtra,with calorie deficit
caloriesdecliningfrom84 to 72, andalsoa dropin povertydepth,
from RDA at the official poverty lines of 600 per day or more,
thebelow2100caloriespercentagedecliningfrom67 to43 while,
while in Madhya Pradesh too there is a big deficit of over 500
mostimportantly,
calories. In north India poverty is underestimatedto a substantial thebelow1800caloriespercentagealsodeclined
extent in both Punjab and Haryana. These hitherto most pros-drasticallyfrom40 to 17. (The 1977-78and 1983datafor states
have not beenpresentedheresince it wouldlengthenthis paper
perous states have been experiencing serious problems with the
inordinatelyand will be presentedlateralong with urbanesti-
loss of an internal market in India to the tune of over 20 million
mates.)Thus the nutritiondata are entirelyconsistentwith all
tonnesof foodgrains owing to the sharpfall in percapita foodgrain
absorption in the country following income-deflation. Actualpreviousanalysespointingto the verypositiveresultsof thefirst
poverty affects half the population and in Punjab it has been15 yearsof LeftFrontrulein the state.Withneo-liberalreforms
therewasperforceacutbackindevelopmentexpendituresinWest
rising. In east India poverty is underestimated to the largest
extent in Assam while West Bengal too has a substantial Bengal too as in other states, as the centre,takinga strongly
deflationiststance,reducedtax devolutionand gave loansonly
deficit of 500 calories at its poverty line. The only states where
atexorbitantinterest.Someof theearliergainshavebeenreversed
calorie intake at official poverty lines is 2000 or more and hence
over the 1990s: the below 2100 calories percentagehas risen
underestimationwhile present is not very large, are Bihar, Orissa
and Uttar Pradesh. to 55 from 43, and the below 1800 calories percentageto 22
from 17, which is certainly a disturbing development.
The picture with respect to actual poverty is fully consistent
with the adverse macroeconomic trends in the ruraleconomy inNayyar(1991) had pointedout thatthe rankingof the statesof
Indiaaccordingto theirpovertylevels estimatedusing the two
terms of rising unemployment and falling foodgrains absorption
methods,in the 1970s was highlycorrelatedeven thoughthere
discussed in the first section and is borne out by the recent NSS
was an increasinggap betweenthe directandindirectestimates.
surveys confirming agrariandistress. In only four states out of the
15majorstatesof India(Assam, Kerala,HaryanaandUttarPradesh)Spearman'srankcorrelationcoefficientworkedout to 0.89 and
have directlyestimatedruralpoverty fallen slightly between 0.84 (using the official estimateon the one hand, and direct
1993-94and 1999-2000,while in 11 of the remaining12 states, estimatesusing two differentnorms,2200 and 2000 calories)
povertyhasrisenoverthe period.We aremakingno adjustment and was significantat the 1 per cent level. However,we find

Economic and Political Weekly July 28, 2007 3145

 Chart5c: Deficit of Calorie Intakefrom RDAat Official PL, Chart5d: Deficit of Calorie Intakefrom RDAat Official PL,
East India North India

Assam UP

West Bengal

Punjab
Bihar

Orissa
Haryana
0 100 200 300 400 500 600 700
Deficit of Calories Per Diem 0 200 400 600 800
* 1993-94 ! 1999-2000
Deficitof CaloriesPer Diem
* 1993-94 0 1999-2000
that by 1983 the rank correlation coefficient value had dropped J V Meenakshi and B Viswanathan (2003) have used the
to 0.6 and furtherto 0.2 by 1999-2000, and it may well be negative statistical technique of kernel density functions to estimate the
by now. To be precise, Spearman's rank correlation coefficient distribution of persons by calorie intake and have presented the
taking the poverty ranks of the states by the official method, and resulting ogives. It might at first sight appear that they are
by the direct method for 1999-2000, works out to 0.236 and 0.075 following the direct method of estimating poverty but this is not
(using the same two nutrition norms as she had used) and neither the case. While other authors using the official method have de-
is significant at the 1 per cent level.10 linked poverty level expenditure from any nutrition norm,
It must not be thought that all economists have been following Meenakshi and Viswanathan's procedure is the mirroropposite.
the fallacious method recommended by the 1993 Expert Group, They have conceptually de-linked nutritional levels from con-
which has resulted in the contretemps of drastic underestimation sumer expenditure and given a different name, "calorie depri-
of poverty, and arbitraryvariations in poverty across states. There vation" to their estimates. It is as much selective use of the NSS
are a number of writers critical of the official method who have data however, to talk of energy intake alone without relating it
rightly put nutritionback at the centre of their analysis of poverty. to the associated expenditure, as it is to talk of expenditure alone
Nayyar's early research also contained a thorough discussion of without relating it to the associated energy intake. The result is
nutritionalnorms. Others writing recently have followed a direct to permit the spurious official and World Bank claims of de-
poverty estimation route, though a different one from inspecting clining poverty to go unchallenged even when nutrition is dis-
and calculating from current NSS data - the method I have cussed, for although the official poverty lines and poverty ratios
followed in this paper. They have estimated the minimum cost are given, any mention of what is happening to average nutritional
of accessing the calorie RDA on the basis of current nutrient intake at the official poverty line expenditure over time and across
prices by solving for the classic "diet problem" and thus have states, is avoided.
obtained a normative food expenditure. By comparing with the We have not used the classifications given in the NSS reports
actual expenditure on food in the NSS, they arrive at the per- of persons distributed by their levels of calorie intake, precisely
centage of persons failing to reach this requiredfood expenditure, because these are not directly by expenditure levels. Low calorie
and this comes to 66 per cent at the All-India level for the 55th intake of individuals is a necessary but not a sufficient index
Round [see Coondoo, Majumdar, Lancaster and Ray 2004, Ray of poverty. It is to be expected that even in high income groups
and Lancaster 2005]. there will exist a certain proportion of persons, with unusually
S Subramanian(2005) has analysed the impact of relative food low calorie intake for their age and sex, since these groups
price rise, and the loss of common property resources on the include fashion models, racing jockeys, anorexic youth and
demand for food, using the theory of consumer demand to show, sick persons unable to absorb food. While poverty will nec-
in his own words, that "...(a) at an income level which the official essarily lead to low intake, from low intake alone poverty
methodology equates with the poverty line, it would not be cannot be inferred. For obtaining a sufficient index, expenditure
compatible with optimising behaviour to consume food at its levels must always be factored in, so that we are correctly
calorifically normative level; and (b) the level of income required separating out those with enforcedly low calorie intake because
to induce optimal consumption of the calorific norm will be their expenditure is low, from those persons with high total
greaterthan the officially stipulated poverty line" (p 61). He has expenditure who might restrict their energy intake for reasons
rejected the official procedure of taking an invariant base-year other than purchasing power.
consumption basket, as assigning arbitrarilya normative value As might be expected, some of those involved in the 1993
to the consumption pattern of one particular year. This means Expert Group report are trying to defend their position either
rejecting the entire basis of the official poverty numbers. How- directly or by proxy. But the arguments being put forward are
ever it is not clear why every year is regarded by him as being a total academic embarrassment and would not be worthy of
equally arbitrary. Our existential reality is that we live in the even being mentioned here were it not for the fact that they
present,not in the past or the future; the cost of minimum current have already made it to some official publications. One such
consumption needs must provide the partitioning device for argument (apparently made by the Expert Group itself) is that
dividing the set of all persons into poor and non-poor if such in any poverty ranking the state of Bihar can be expected to come
an exercise is to be at all undertaken. towards the bottom and since applying the nutritionnormdirectly

3146 Economic and Political Weekly July 28, 2007

does not put Bihar towards the bottom, the nutritionnorm should Mehta and Venkatraman(2000) had drawn attention to the fact
not be applied. Those who put forward or defend this gem of that crop-straw, fuel-wood and fodder which was earlier gleaned,
illogicality merely expose their preconceptions regarding Bihar gathered or accessed as common property (only partly valued
and ignore research showing substantial rural real wage rise in in the NSS, or valued at low farm gate prices), now had to be
many districts in the pre-reform phase, in which out-migration purchased at retail rates. Food and cooking fuels are jointly
had a role to play. demanded since no one can eat raw food, and with a real income
Another common argument defending the wrong idea that which is constant or declining, a part of food expenditure has
falling calorie intake is voluntary, is that there has been to be enforcedly reduced to buy fuel. They had established that
mechanisation in agricultureand the energy intake needs of rural the rising non-food monetised expenditure on utilities (fuels,
labourers has reduced. One can scarcely find a clearer example transport,health) meant that food expenditure in real terms was
of apologetics than this argument which involves a double non forced down to a lower level by 1993-94. The present author
sequitur. First, the argument assumes that rural labourers were too had drawn attention in a brief but sharply worded manner
adequately fed before mechanisation and there is scope for to the futility of using the consumption basket of 1973-74 to
reducing intake, which is not the case; second, it assumes that estimate current poverty [Patnaik 2004].
with mechanisation human energy intake necessarily goes down, At the 1993-94 official poverty line, 6 per cent of spending
while the converse is observed to be the case everywhere. Even was on "fuel and light" and 13.1 per cent was on miscellaneous
if we consider Asia alone, the highest levels of energy intake goods and services (medical services, transport, education and
of ruralworkersarein the most agriculturallymechanisedcountries rent) adding up to 19.1 per cent. By 2004-05 for the official
like Japan, Korea and China which have seen rising nutritional poverty line expenditure class, the fuel and light share at 10.2
standards of rural workers as their incomes rise, which is as it and the miscellaneous goods and services share at 23.4 per cent
should be since the aim of raising labour productivity through added up to 33.6 per cent of spending. Since real spending has
mechanisation is precisely to improve the lot of people. been stagnantover 1993-94 to 2004-05 for four-fifthsof all persons,
a higher share spent on these items entails not just a lower food
IX share but absolute decline in spending on food. Only Rs 221.8
Inability of OfficialPovertyLinesto Capture per month per head or Rs 7.4 per day, could be spent on food
ActualCostof Living in 2004-05. This is Rs 126 at 1993-94 prices, less than the Rs 143
actually spent on food at the official poverty line of 1993-94.
In constructing the consumer price index for agricultural Second, the more recent changes affecting poverty are the
labourers, zero or negligible weight is given to many items of outcome of the deflationary policies discussed in the first section
spending which are in practice unavoidably important for even of this paper, which are in principle all reversible. Large cuts
poorer workers such as transport to site of work, coping with in development expenditures reduced the level of rural activity
ill-health, and basic utilities. Altering the weighting diagram of and raised unemployment. Rising input and credit costs combined
the CPIAL to take realistic account of these items would certainly with stagnantor falling output prices, or an adverse price scissors
help a bit, but not all that much. In my judgment the more further reduced incomes. Mass demand deflation in turn led to
importantproblemis the arbitraryprocedureof applying the given a drastic lowering of the inflation rate by the end-1990s, and
price-index to a fixed consumption basket which goes back as even in the severe drought year 2002-03, agricultural prices
far as 34 years. However well constructed the price index itself hardly rose since distress sales ensured easy marketsupplies, and
might be, taking such a distant fixed basket cannot but ignore with lower output demand was further compressed. The rise in
importantand mainly non-reversible structural changes taking the official poverty line which entirely reflects the rise in the
place in the economy over time, which are responsible for altering CPIAL, was 60 per cent between 1993-94 and 1999-2000 but
the choices faced by consumers such that the actual consumption was below 11 per cent between 1999-2000 and 2004-05. Neo-
basket is alteredand there is necessarily a much higher cost today liberal deflation squeezed aggregate demand so severely that it
of accessing the minimum energy intake. eventually resulted in price deflation in agriculture. This set of
The changes in the economy which have altered the set of factors has led to a downward shift in the demand curve for
choices consumers face, fall into two categories: First, long-term necessities for a majority of the rural population.
structuralchanges since the 1960s which are mainly irreversible, The recent moderate revival of inflation during 2006-07 is
and second, changes under neo-liberal reforms over the last 15 partlycost-push owing to rising imported oil prices, but is mainly
years which are in principle reversible. The long-term changes shortage-induced owing to the collapse of grain output growth
have been extensively discussed but ignored by the official in the last few years, brought about by the sustained decade-long
estimators. M H Suryanarayana(1996) in a detailed discussion state attack on farmers' viability. Lower inflation during 2000
of the concepts and methods used for estimating expenditure by to 2005 should have benefited rural net food purchasers if
the NSS, had pointed out that the economic environment for everything else was the same, but because it was the result of
labourersand poorer farmers was changing in a mannernot fully expenditure deflation-induced depression, any benefit was
captured by price indices. Over the previous three decades there swamped out by unemployment rising faster and earnings
had been substantial monetisation of the rural economy. Wages declining more rapidly than the inflation rate was decelerating,
paid in kind as grain or meals, valued at low prices in NSS rounds, pushing more people into poverty.
were now paid in cash which the labourer had to exchange for The remedy is simple - a strongly expansionary fiscal policy
food at higher retailprices embodying lower quantities. Common and genuine commitment to implementing the NREG Act by
property and gleaning rights were disappearing. This rendered funding it properly, supported by large-scale revival of
official poverty lines of dubious value, and he had advocated foodgrains and other crops procurement at realistic prices; and
using direct quantitative indices for measuring poverty. scrapping the iniquitous and senseless APL-BPL divide, would

Economic and Political Weekly July 28, 2007 3147

be enough to lift the agrarian economy out of depression and level. The average calorie intake in rural India has declined
reduce hunger.There is no sign however thatthe country's leaders further from 2153 to 2047 over the decade. Average daily protein
have the wisdom to change course and avert the current slide intake has declined by 3 gms and average fat intake has risen'
to the abyss.
As regards poverty measurement the solution lies in using Chart A-1: Ogive Based on Consumption Expenditure from
Uniform Recall, 61st Round
simple, direct and transparentindices of poverty and the mini-
mum use of complex, indirect and opaque measures, however Ogive, 61st Round2004-05 Based on Uniform30 Day Recall
100
enamoured professional economists might be of the latter. The
calorieintakeby differentexpendituregroupswill become available 90-
soon for 2004-05 from the NSS 61st round and will permit direct 80
estimation of actual poverty and poverty depth. Possession of
C70
tangible assets, food grains absorption per head, whether the
family resides in hard-roofed structures, floor area occupied per 50
family, yardage of textiles consumed, use of electric power -
0
all considered by differing economic levels ratherthan in terms
of overall averages alone - these are some of the simple and I 40
crucial indices which will give a clear idea of poverty and its 30-
trends over time. Poverty estimation should be entrusted to an
20
independent body of academics, not to international financial
institutions or to governments, which are parties interested in 10
claiming success for their policies, and have by now amply
demonstrated their lack of objectivity. 0 200 400 600 800 1000 1200 1400
-- Series 1
Appendix
Indirect and Direct Poverty Estimates, Table A-1: 61st Round Consumer Expenditure with Alternative
Recall Periods (Uniform 30-day and Mixed)
61 st Round, 2004-05
1 2 3 4 5 6 7
As this paper was already in press before the 2004-05 nutrition MPCE Per Cent MPCE Per Cent MonthlyPer Capita Estimated
data became available in May 2007, the discussion of direct Class of (U30)) of Expenditure UpperEnd
poverty estimates using these data is confined to this Appendix (U30) Persons Persons M, 1993-94 M,2004-05 of Exp
Prices Prices Class
and awaits a fuller treatment later.
(Rs) (Rs) (Rs) (Rs)
There are alternative estimates which emerge from the 61st
round depending on whether we take the uniform 30-day recall, 0-235 4.8 199.53 5 137 239.7 275
235-270 5.1 253.80 5 169 295.7 315
which is stated to be comparable with the 50th round estimates, 270-320 9.9 296.64 10 193 337.7 360
or the mixed recall which is stated to be comparable with the 320-365 10.5 342.40 10 220 384.9 410
365-410 10.2 387.72 10 245 428.7 450
55th roundestimates. The 2004-05 ogive, using the mixed recall, 410-455 9.4 432.06 10 271 474.2 500
lies to the right of the ogive using the uniform 30 day recall 455-510 9.9 481.55 10 299 523.2 545
510-580 10.2 543.25 10 333 582.7 620
(Chart A-i) by a mere Rs 40 or so for the poorest four deciles. 580-690 10.4 630.40 10 380 664.9 710
With the former, about 20 per cent of persons are below Rs 365, 690-890 9.8 775.00 10 455 796.1 880
890-1155 5.0 999.94 5 569 995.6 1110
while with the mixed recall, about 30 per cent of persons are 1155 + 5.0 1956.57 5 936 1637.8 2166
below Rs 360. Since Rs 356 is the indirectpoverty line the poverty All 100.0 558.78 100 331 579.17
percentages are 18.5 and 28.5 respectively. The mere Rs 40 Notes: MPCEfor2004-05 undermixedrecallMis given in 1993-94 pricesfor
rightward shift in the mixed recall ogive compared to the U 30 percentileson p 19, Table P 7 and is reproducedin the fifthcolumn
one results in a large drop by 10 percentage points in official above againstthe respectivepercentilesincol 4. The values in 2004-
05 prices are then obtained using CPIAL.Column 7 values are
poverty solely because the indirect poverty line is already such approximateand obtainedby assuming thatthe col 6 values are mid-
a gross underestimate and so low, that it intersects the ogive at pointsof the respectiveclasses.
Source: NSS Report508.
a point where it has the steepest slope. The calorie intake
permittedby the official poverty line has declined furtherto 1820 Table A-2: Poverty Estimate, 1993-94 and 2004-05, All-India Rural
(see Table A-2). DirectEstimate
Using ChartA-2 along with A- 1, the percentage of ruralpersons Levels of calorieintakeper day 2400 2200 2100 1800
not able to access 2400 calories at the all-India level is 87.0, Per cent of persons below specified
level, 2004-05 87.0 69.5 60.5 25.0
up sharply from 74.5 per cent in 1993-94 and from about 77.5 Per cent of persons below specified level
per cent in 1999-2000. The requiredspending to access RDA has in 1993-94 74.5 58.5 49.5 20.0
gone up to Rs 795, a rise by two-fifths over the five years. The Requiredmonthlyper capitaexpenditure
actual rise in poverty is even greaterthan anticipated, and poverty in 2004-05 to access nutritionlevel, Rs 795 575 515 342
OfficialEstimate
depth has increased more than during any previous period, not Officialpovertyline (OPL) 1993-94 2004-05
surprisingly given the pervasive agrariandepression and farmer- Rs 206 356
labourer distress. Millions of persons have been pushed down Per cent of persons below OPL 37.3 28.3
to a lower nutritionalstatus. Every nutritional level shows about Calorieintakeat OPL 1980 1820
10 to 12 per cent more of total population below it and 5 per Source: CalculatedfromNSS Report513, Nutritional
Intakein India,2004-05,
cent more of all persons have sunk below the lowest, 1800 calories A-18, A-90, and Report508.

3148 Economic and Political Weekly July 28, 2007

 ChartA-2: Per Capita Daily Calorie Intake by MPCE, 2004-05, actual official base year nutrition norm. As high as 69.5 per cent
All-IndiaRural
of all persons were below 2200 calories intake by 2004-05
3500 -
compared to 58.5 per cent in 1993-94. In only two of the 15
3000 - major states in India (Assam and Kerala) have directly estimated
poverty, unambiguously declined (for all levels) during 1993-94
2500 to 2004-05, while in 11 states it has unambiguously increased
for all levels. Andhra Pradesh shows marginal improvement by
2000-
2004 compared to decline up to 1999-2000 since both extreme
1500 poverty depth and below-RDA percentages have declined. While
there were nine major states in 1993-94 where one-fifth or more
1000 -
of persons could not access 1800 calories, by 2004-05, as many
500 -
as 12 states out of 15 were in that position. Rural poverty has
increased not marginally but markedly, in a wide belt of states
0 4 spanning the entire country, from Punjab. Haryana and UP to
0 500 1000 1500 2000 2500
Gujarat,Rajasthanand Madhya Pradesh andtfrom Bengal, Bihar
MPCE. Rs
and Orissa to Karnataka and Tamil Nadu.
F-e Calorie Intake At the very low official poverty lines, in 1993-94 the rural
population in 10 major states could not access 2000 calories,
by 5.5 gms, both being small changes relative to the large decline and in four major states could access only 1700 calories or less.
in energy intake. But by 2004-05, the number of states so affected were higher.
The below 2100 calories percentage of persons is substantially The official poverty lines did not permit the rural population of
higher at 60.5 compared to 49.5 in 1993-94 and the below 1800 14 (out of 15) major states to access 2000 calories, while the
calories percentage is 25, while it was 20 a decade earlier population of seven major states could only obtain 1700 calories
(Table A-2). The absolute numbers of rural persons below or less. Their official poverty ratios had thereby gone down. Thus
2100 calories has risen by 150 million, from 289 million in the secret of official "poverty reduction" remains the unstated
1993-94 to 440 million by 2004-05. Those accessing below lowering of the consumption standard.[M"
1800 calories, the poorest of the poor, numbered 182 million by
2004-05 compared to 117 million in 1993-94, an increase by Email: [email protected]
65 million. By taking a very low poverty line at which at most
1820 calories can be obtained, effectively the Planning Com- Notes
mission is designating as "the poor", only the poorest among
the actually poor. [An earlierandshorterversionof this paperwas presentedas the Rao Balhadurl
Kale MemorialLectureat the Gokhale Insituteof Politics and Economics.
In Table A-3 the statewise poverty estimates are given, includ- Pune, February3, 2006. I would like to thank lmrana Qadeer, Vcnka"tesl
ing the 2200 calorie level, bearing in mind that this was the likely Athreya and Akeel Bilgrami for comments.
Table A-3: Calorie Intake at Official Poverty Lines, 2004-05 and Direct Estimates by States
Official Indirect Estimates Direct Estimates
2004-05 1993-94 2004-05 1993-94 2004-05 1993-94 2004-05
Official Poverty Poverty Calorie <2400 <2400 <2200 <2200 <1800 <1800
Line (OPL) Ratio at OPL Intake Calories Calories Calories Calories Calories Calories
(Per Cent) at OPL (Per Cent) (Per Cent) (Per Cent) (Per Cent) (Per Cent) (Per Cent)
1 2 3 4 5 6 7 8 9
All-India 356.3 28.3 1820 74.5 87.0 58.5 69.5 20.0 25.0
East
Assam 387.64 22.3 1810 93.0 87.5 72.5 63.5 30.0 21.5
Bihar 354.36 42.1 1960 73.0 84.0 60.0 68.5 25.0 25.0
Jharkhand 366.56 46.3 2020 na 80.5 n.a 60.0 na 10.0
Orissa 325.79 46.8 2010 70.0 82.5 47.0 67.0 16.0 27.5
West Bengal 382.82 28.6 1855 72.0 83.0 52.5 67.5 17.0 24.3
South
Andhra Pradesh 292.95 11.2 1620 84.0 79.5 67.5 67.5 27.0 25.0
Karnataka 324.17 20.8 1625 75.5 95.0 62.5 86.5 29.0 43.5
Kerala 430.12 9.6 1480 84.0 75.0 71.5 66.0 40.0 34.0
Tamil Nadu 351.86 16.9 1600 87.0 94.0 81.0 86.5 43.0 46.0
West-Central
Gujarat 353.93 19.1 1655 83.5 89.5 71.5 83.5 36.0 41.0
Madhya Pradesh 327.78 36.9 1790 72.5 90.5 55.5 81.5 21.5 37.5
Chhattisgarh 322.41 40.8 1805 na 88.5 na 76.5 na 40.0
Maharashtra 362.25 29.6 1675 89.5 95.0 82.0 78.5* 38.0 41.5
Rajasthan 374.57 18.7 1835 46.0 76.0 33.0 61.0 7.5 15.0
North
Punjab 410.38 9.1 1700 52.5 68.0 35.0 53.5 11.0 15.0
Haryana 414.76 13.6 1735 55.0 63.5 40.0 52.5 11.5 20.0
Uttar Pradesh 365.84 33.4 1965 65.5 72.5 45.0 60.0 11.0 16.5
Uttaranchal 478.02 40.8 2205* na 60.0* na 40.0* na 3.5
Note: * Provisional.
Source: NSS Report Nos, 401, 402, 405 for 50th round, Reports 508, 513 for 61st round.

Economic and Political Weekly July 28, 2007 3149

 1 'TheNatureof Fallaciesin EconomicTheory',SatyendranathSen Lecture and Major States, 1983 to 2004-05', Economic and Political Weekly,
delivered at the Asiatic Society, Kolkata, August 11, 2004. February 15.
2 Gopalanis referringto P V Sukhatme'sargument,which he had earlier Meenakshi,J V and B Viswanathan(2003): 'Calorie Deprivationin Rural
refuted [Gopalan 1983] that mean energy requirementlevel minus two India', Economic and Political Weekly,Vol 38, January25.
standarddeviations should be considered for poor populations. Mehta,J and Venkataraman(2000): 'PovertyStatistics- Bermicide'sFeast',
3 National Nutrition Monitoring Bureau, 25 Years of NNMB, Delhi, Economic and Political Weekly, Vol 35, July 1.
1997. National Sample Survey Organisation:
4 Note thatsince the highest expenditureclass is open-endedfor ruraland Report No 408, Livestock and Implementsin Household Operational
urbanIndia, the last point of the relevant ogives have not been shown. Holdings, 1991-92.
Assuming that the given average expenditure is the mid-point of the Report No 493, LivestockOwnershipacross OperationalLandholding
expenditureclass in each case, we get Rs 1,738 and Rs 4,223.6 as the Classes in India, 2002-03.
estimatedupperend values for ruraland urbanexpenditure.The reader Report No 401, Key Results on Household Consumer Expenditure,
can visualise the ogives approaching 100 at these values. 1993-94.
5 Mosteconomistsincorrectlyinterpreta necessaryconditionas a sufficient Report No 402, Level and Pattern of ConsumerExpenditure,1993-94.
one. Rising real income does imply a falling share of food expenditure Report No 405, Nutritional Intake in India, 1993-94.
in total expenditurebut the converse is not true. A falling shareof food Report No 457, Level and Pattern of Consumer Expenditure in
expenditurein total expenditure does not imply rising income and is India, 1999-2000.
consistent with falling income. Report No 471, Nutritional Intake in India, 1999-2000.
6 I repudiatethe views I expressed on poverty in my paperswrittenbefore ReportNo 497, Income, Expenditureand ProductiveAssets of Farmer
2004 where I uncriticallyreproducedPlanningCommission and World Households, 2003.
Bank estimates.I was not then aware of the fatally flawed methodology ReportNo 504, Household CapitalExpenditurein Indiaduring 1.7.2002
used, andonly contradictionof the claims of these bodies with deepening to 30.6.2003.
agrariandistress, led me to look closely at the official procedure. ReportNo 508, Level and Pattern of ConsumerExpenditure,2004-05.
7 Any inaccuraciesin plotting and reading the graphs are mine, but the ReportNo 515, Part 1 and 11, Employmentand UnemploymentSituation
mistakes if any are likely to be small, 10 calories at most. in India 2004-05, available on www.mospi.nic.in
8 Formanyotherstateslike AndhraPradesh,Deaton's recalculatedpoverty Nayyar, R (1991): Rural Poverty in India, Oxford University Press, Delhi.
lines give higher estimates than the PlanningCommission ones but are Patnaik,P (2000): 'The Humbug of Finance', ChintanMemorial Lecture,
of coursestill farbelow the correctestimatesapplyingthe nutritionnorm. January8, 2000, available on www.macroscan.org,and in P Patnaik
9 See ReddyandPogge (2005); WorldDevelopmentReport,2006, Table Al (2003), The Retreat to Unfreedom,Tulika, Delhi.
on p 278, cols 9 and 10. Patnaik, U (2003): 'Global Capitalism, Deflation and AgrarianCrisis in
10 RamanandRam (2004). Developing Countries', Social Policy and Development Programme,
Paper Number 13, United Nations Research Institute for Social
Development (UNRISD) October.
References - (2003a): 'Food Stocks and Hunger- Causes of AgrarianDistress', Social
Scientist, Vol 31, Nos 7-8, July-August.
Anand,S (1997): 'The Measurementof IncomeInequality'in S Subramanian - (2004): 'The Republic of Hunger', Social Scientist, Vol 32, Nos 9-10,
(ed), Measurementof Inequalityand Poverty, Oxford University Press, September-October.
Delhi. - (2005): 'TheorisingFood Security and Poverty in the Era of Neo-liberal
Aristotle (undated):De Sophisticis Elenchis (Of Sophistical Refutations, Reforms', Social Scientist, Vol 33, Nos 7-8, July-August.
from Organon), Vol 1 in Barnes (ed), Collected Worksof Aristotle. Ram, Ramanand(2004): 'PovertyEstimatesin India:A CriticalAppraisal',
Baker, D, G Epstein and R Pollin (1998): Globalisation and Progressive MPhil Dissertation submitted to JawaharlalNehru University, July.
Economic Policy, Cambridge University Press, Cambridge. Rath,S (2003): 'Povertyby Price Indices', Economicand Political Weekly,
Barnes, J (1984): Collected Worksof Aristotle, Vols 1 and 11, Princeton October 4, Vol 38, No 40.
University Press, Princeton. Ray, R and G Lancaster(2005): 'On Setting the Poverty Line Based on
Bhalla, G S (2005): 'The Challenge of Food Security in India', Tyagi EstimatedNutrientPrices:Conditionof Socially DisadvantagedGroups
Memorial Lecture, July. during the Reform Period', Economic and Political Weekly,Vol XL,
Coondoo, D, A Majumdar,G Lancaster and R Ray (2004): 'Alternative No 1, January 1.
Approachesto MeasuringTemporalChangesin Povertywith Application Reddy, S G and T W Pogge (2005): How Not to Count the Poor,
to India', Working Paper, December. www.socialanalysis.org
Cornia,G A, R Jolly and F Stewart(eds) (1987): Adjustmentwith a Human Sen, A (2000): 'Estimates of Consumer Expenditureand Its Distribution:
Face, Vol 1, ClarendonPress, Oxford. StatisticalPrioritiesafter the NSS 55th Round', Economicand Political
Dandekar,V M and N Rath (1971): Poverty in India, Indian School of Weekly,December 16.
Political Economy, Pune. Sen, A and Himanshu(2004): 'Poverty and Inequalityin India: Widening
Datt, G, V Kozel and M Ravallion (2003): 'A Model Based Assessment Disparities during the 1990s', Economic and Political Weekly,
of India's Progress in Reducing Poverty in the 1990s', Economic and September25.
Political Weekly,January25, pp 355-61. Subramanian,S (1997): 'Introduction'in S Subramanian(ed), Measurement
Deaton, A (2003a): 'Adjusted Indian Poverty Estimates for 1999-2000', of Inequality and Poverty, Oxford University Press, Delhi.
EPW, Vol 38, January25-31. - (2005): 'Unravelling a Conceptual Muddle - India's Poverty Statistics
- (2003b): 'Pricesand Poverty 1987-2000', Economicand Political Weekly, in the Light of Basic DemandTheory', Economicand Political Weekly,
Vol 38, January25. Vol XL, No 1, January 1-7.
Deaton, A and V Kozel (eds) (2005): The Great Indian Poverty Debate, Sukhatme, P V (1977): 'Incidence of Undernutrition',Indian Journal of
Macmillan India, New Delhi. Agricultural Economics, July-September.
EconomicandPolitical Weekly(2003): SpecialIssuetitled'PovertyReduction Sundaram,K and S D Tendulkar(2003): 'PovertyHas Declined in the 1990s
in the 1990s', Vol 38, January25. - A Resolutionof ComparabilityProblemsin NSS ConsumerExpenditure
Gopalan, C (1992): 'Undernutrition:Measurement and Implications' in Data', Economic and Political Weekly, Vol XL, No 1, January 1-7.
S Osmani (ed), Nutritionand Poverty, ClarendonPress, Oxford. Also Suryanarayana, M H (1996): 'Poverty Estimates and Indicators:
reproducedin S Subramanian(ed) (1997), Measurementof Inequality Importance of Data Base', Economic and Political Weekly, Nos 35-
and Poverty, Oxford University Press, Delhi. 37, September.
- (1983): 'Measurement of Undernutrition:Biological Considerations', Swaminathan,M (1999): WeakeningWelfare: The Public Provisioning of
Nutrition Foundation of India Bulletin, 4. Food in India, Leftword Books, Delhi.
Governmentof India,Ministryof Finance(variousyears):EconomicSurvey, - (2002): 'Excludingthe Needy - The Public Provisioningof Food in India',
(annuallyreleased in Februaryeach year) for the years 2000-01, 2001- Social Scientist, Vol 30, Nos 3-4, March-April.
02, 2002-03, 2003-04, 2004-05 and 2005-06. Thouless, R F (1974): Straight and Crooked Thinking,Pan Books.
Hamblin, C L (1970): Fallacies, Methuen, London. World Bank (2006): WorldDevelopmentReport, World Bank and Oxford
MahendraDev, S and C Ravi (2007): 'Poverty and Inequalityfor All-India University Press, New York.

3150 Economic and Political Weekly July 28, 2007.