ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

India's Common People: Who Are They, How Many Are They and How Do They Live?

This paper attempts to define the common people of India in terms of levels of consumption and examines their socio-economic profile in different periods of time since the early 1990s with a view to assessing how the economic growth process has impacted on their lives. The findings should worry everyone. Despite high growth, more than three-fourths of Indians are poor and vulnerable with a level of consumption not more than twice the official poverty line. This proportion of the population which can be categorised as the "common people" is much higher among certain social groups, especially for scheduled castes and scheduled tribes. There is also evidence to suggest that inequality is widening between the common people and the better-off sections of society.

kannan.nce@nic.in

SPECIAL ARTICLEmarch 15, 2008 EPW Economic & Political Weekly501993-94, 1999-2000 and 2004-05, the quantitative results presented here are mainly for the latest year of 2004-05. Readers interested in examining the details of the two previous time points can refer to the larger paper published on the web site of the journal (www.epw.in).Economic Growth during Pre- and Post-Independent IndiaThe years between 1993-94 and 2004-05 were the period of a high rate of economic growth in India. While it is not the purpose of this paper to analyse the nature of this growth, it may be useful to take note of the remarkable improvement in our growth performance in the recent years which may be seen from the decadal growth rates of GDP and per capita income given in Table 1. The Indian economy entered a trajectory of high economic growth from 1980 onwards, compared with the previous nearly three decades. The period 1950-80 disparagingly referred to as the years of “Hindu” rate of growth had an average rate of growth of 3.56 per cent for the entire period. The growth rate decreased from 3.91 in the 1950s to 3.70 per cent in the 1960s and 3.08 per cent between 1970 and 1980. Although the so-called “Hindu” rate of growth from 1950 to 1980 was three times what we would like to call the “Colonial” rate of growth of the previous 30 years, the fact remains that during the first 30 years after India’s independence, the improvement in the per capita income was hardly between 1 and 1.95 per cent due to an increasing population growth that resulted from a decline in mortality. That is to say progress made on two fronts – aggregate economic growth and reduction in mortality – offset each other in a kind of developmental irony that most developing countries go through. On a per capita basis, this meant a slow pace of improvement of welfare of the vast masses of the population. As the growth rate picked up from 1980 onwards, there was a marked acceleration in the growth of per capita income contrib-uted by the double blessing of increasing the growth rate and reducing the population growth. The aggregate economic growth rate accelerated further in the more recent period.1 While the growth performance in recent years has been spectacular, it is equally important to assess whether it has fully reflected in the welfare of the common people so as to allow usto evaluate the success of our growth performance in terms of the basic objective of improving the welfare of the vast masses of people. This objec-tive was explicitly set out in the policy of inclusive growth or increasing the welfare of the common people, of the United Progressive Alliance government, in contrast with the campaign about a “Shining India”, in 2004.2 Data and MethodologyIn order to evaluate the success of our growth performance in terms of welfare characteristics captured by the per capita consumption expenditure of our people, we have worked on thedata sets available from the surveys of National Sample Survey Organisation (NSSO) on employment-unemployment and consumption expenditure. We have classified each sample household as accordingly belonging to the “extremely poor”, “poor”, “marginal”, “vulnerable”, “middle income” and “high income” groups if the monthly per capita expenditure (MPCE) of their households was below or above a specified multiple of the poverty line (PL). Though there is considerable criticism about the determination of the poverty line in India, which systematically underestimates poverty and related deprivation, we have chosen not to engage in that debate here to avoid being deflected from the main thrust of our study. Our aim here is to construct a simple taxonomy and identify groups of people with significant differ-ences in average level of consumption. The crite-ria for our classification, whether in the rural or urban areas, are given in Table 2. The official estimates of poverty are computed by the Planning Commission from the quinquennial consumer expenditure survey (CES). However, the analysis in this paper is based on consumption expenditure data as reported in the employment-unemployment survey (EUS) schedules to enable us to relate consumption expenditure status of the people to their work and activity status. In the case of theEUS, data on consumption are collected by using an abridged block in which some major items and item groups only are listed as against CES in which consumption of each individual item is separately collected. Thus, the average consumption expenditure as computed from EUS schedule is expected to be lower than that obtained from CES schedule. The poverty lines, determined on the basis of CES therefore had to be adjusted to make them conform to the level of consumption expenditure as given by theEUS. This was done by estimating theMPCE from the distribution of persons byMPCE classes based onEUS corresponding to the head count ratios obtained from CES. The difference between the two estimates is insignificant. For example, the official poverty line for 2004-05 is Rs 356.3 for rural areas and Rs 538.6 for urban areas. The poverty lines as applicable to data sets available from the EUS, as perour calculation, turn out to be Rs 346.2 for rural areas and Rs 514.0 for urban areas. The reason for using the abridged consumer expenditure block in theEUS is, as mentioned above, due to our objective to link the expenditure classes (poor/non-poor) with employment and work characteristics of the working members in the households. However, we give in Tables 3 (p 51) and 5 (p 52) the relevant statistics obtained from the detailedCES to show that the difference is not significant from the point of interpretation of our findings.On the basis of the adjusted poverty lines, we have estimated the average per capita consumption expenditure per month and per day for 2004-05 for classes as defined in Table 2 above and these are given in Table 3.Table 1: Decadal Growth Rates of GDP and Per Capita Income Growth Rate (%)Decades GDP Population Per Capita Income1920-47 1.18 1.20* -0.031950-60 3.91 1.96 1.951960-70 3.70 2.22 1.481970-80 3.08 2.20 0.881980-90 5.38 2.14 3.241990-2000 5.58 1.93 3.652000-05 5.99 1.70 4.29(1) GDP growth rate during 1920-47 is based on 1949 prices. (2) GDP growth rates in other decades are based on 1999-2000 prices. * Population growth rate for the period 1921-51. Sources: NAS, CSO for GDP and Statistical Abstract (2003), CSO for population data. Figures for 1920-47 are from Sivasubramonian (2000).Table 2: Criteria for Classification of HouseholdsExtremely Poor if MPCE<=0.75 times poverty line(PL)Poor if 0.75 PL<MPCE<= 1PLMarginal if 1 PL<MPCE<=1.25 PLVulnerable if 1.25 PL<MPCE<=2.0 PLMiddle class if 2.0 PL<MPCE<=4.0 PLHigh Income Group if MPCE>4.0 PL
SPECIAL ARTICLEEconomic & Political Weekly EPW march 15, 200851Inshort,for the year 2004-05, those who have beenclassified as extremely poor had an average daily per capita consumption expenditure (DPCE) of about Rs 9. Similarly, the group of people taken as poor hadDPCE of about Rs 12, the marginal had DPCE of about Rs 15 and the vulnerable a DPCE of about Rs 20. The middle income group is identified as people with an average DPCE of roughly Rs 37 and the high income group as those with an average DPCE of Rs 93. These estimates are nationwide averages including rural and urban areas. In this paperwe are pre-senting only the national average or the overallpic-ture for the coun-try as a whole. The average DPCE of different classes of people in terms of US $ with purchasing powerparity (PPP) has been added in the last column of Table 3 to facilitate comparisons with international norms. In 2004-05, the nominal exchange rate withUS $ was Rs 44.93, but PPP US $ was equivalent to Rs 9.12. It can be seen that the average DPCE of the extremely poor in 2004-05 was approximately $ 1 and for the vulnerable group was little above $ 2 per day. This should facilitate the international comparison where people below PPP $ 1 are classified as extremely poor and those below PPP $ 2 as the poor. A fraction of our vulnerable people will have daily consumption slightly above Rs 20 per day whichistheaverage for the whole group of the vulnerable (i e, those between 1.25PL and 2PL). But if we take the average consumption of all the classes belonging to extremely poor, poor, marginal and vulner-able groups, they would be consuming on an average less than Rs 20 a day and would be below the international definition of poverty line. The similarity between the two sources of data is also seen in the distribution of population and unorganised workers by poverty status and social group in 2004-05 as given in Table 5. The table is based on the EUS to be able to relate the poverty status of the social groups to the status of unorganised workers. This distribution of the social group is hardly different from that in the consumer expenditure survey as given in the brackets. In an exercise of this nature using data sets of three rounds of the National Sample Survey (1993-94, 1999-2000 and 2004-05) and estimating the expenditure classes from the abridged schedule of theEUS and linking them up with different social groups, educational background and informal or unorganised work status, etc, there are a number of technical issues and limitations of data. We have therefore discussed in detail in the Appendix the methodology of classification and estimation spelling out these issues and how we have adjusted or coped with them. 3 Changes in the Magnitude of the Poor and VulnerableTable 4 gives a striking profile of the Indian people in terms of their consumption expenditure. As of 2004-05, 21.8 per cent of the people can be regarded as extremely poor or poor.2 When we compare this information with the estimates derived from the 55th round of 1999-2000 and 50th round of 1993-94, there has been some decline in the percentage of people belonging to these groups of poor in the last 10 years of high economic growth – from 31 to 22 per cent. However, the table also shows that, almost 41 per cent of people, consisting of the extremely poor, poor and marginal groups, survived on an average expenditure of less than Rs 15 per day per capita consumption in 2004-05, which by all counts should be regarded as affording no more than a miserable existence. The vulnerable group, whose average DPCE was Rs 20 Table 3: Average Per Capita Consumption Expenditure (2004-05)Poverty Status MPCE (Rs) DPCE (Rs) DPCE (US $ PPP)1 Extremely poor 269 [268] 9 [9] 1.02 Poor 348 [358] 12 [12] 1.33 Marginal 438 [449] 15[15] 1.64 Vulnerable 609 [631] 20 [21] 2.25 Middle income 1,098 [1137] 37 [38] 4.06 High income 2,776 [2879] 93 [96] 10.27 Extremely poor and poor (1+2) 325 [331] 11 [11] 1.28 Marginal and vulnerable (3+4) 550 [570] 18 [19] 2.09 Poor and vulnerable (7+8) 486 [500] 16 [17] 1.810 Middle and high income (5+6) 1,387 [1448] 46 [48] 5.111 All 696 [734] 23 [25] 2.5(1) MPCE = Monthly per capita consumption expenditure. (2) DPCE = Daily per capita consumption expenditure. (3) 1 US $ PPP = Rs 9.12. (4)Figures in square brackets are based on detailed consumer expenditure survey.Source: Computed from NSS 61st round survey on employment-unemployment, 2004-05. Table 4: Distribution of Population and Average Daily Per Capita Expenditure by Poverty StatusPoverty Status Distribution Population Average DPCE Average DPCE of Population Percentage (in Rs) (in US $ PPP) (inMillion) Distribution 1993-94 1 Extremely poor 102.6 11.5 4 0.7 2 Poor 171.6 19.2 6 1.0 3 Marginal 168.2 18.8 8 1.3 4Vulnerable 289.5 32.4 11 1.8 5Middleincome 138.5 15.5 19 3.1 6 High income 23.9 2.7 47 7.6 7 Extremely poor and poor (1+2) 274.2 30.7 5 0.9 8 Marginal and vulnerable (3+4) 457.7 51.2 10 1.6 9 Poor and vulnerable (7+8) 731.9 81.8 8 1.3 10 Middle and high income (5+6) 162.4 18.2 23 3.7 11All 894.4 100.0 11 1.81999-2000 1 Extremely poor 87.4 8.7 7 0.8 2 Poor 174.1 17.3 10 1.2 3Marginal 200.2 19.9 12 1.5 4Vulnerable 349.0 34.8 17 2.1 5Middleclass 167.3 16.7 31 3.6 6 High income 26.1 2.6 73 8.6 7 Extremely poor and poor (1+2) 261.6 26.1 9 1.1 8 Marginal and vulnerable (3+4) 549.2 54.7 16 1.8 9 Poor and vulnerable (7+8) 810.7 80.7 13 1.6 10 Middle and high income (5+6) 193.4 19.3 36 4.3 11All 1004.1 100.0 18 2.12004-05 1 Extremely poor 69.7 6.4 9 0.9 2 Poor 167.3 15.4 12 1.3 3Marginal 207.1 19.0 15 1.6 4Vulnerable 392.0 36.0 20 2.2 5Middleclass 209.8 19.3 37 4.0 6 High income 43.7 4.0 93 10.1 7 Extremely poor and poor (1+2) 237.0 21.8 11 1.2 8 Marginal and vulnerable (3+4) 599.1 55.0 18 2.0 9 Poor and vulnerable (7+8) 836.1 76.7 16 1.8 10 Middle and high income (5+6) 253.5 23.3 46 5.1 11All 1089.6 100.0 23 2.5Conversion rates for $PPP were computed from the calendar year-wise conversion rates provided inInternational Monetary Fund,World Economic Outlook Database, September 2006. The conversion rates used are Rs 6.20, Rs 8.42 and Rs 9.12 for 1993-94, 1999-2000 and 2004-05.
SPECIAL ARTICLEmarch 15, 2008 EPW Economic & Political Weekly52(or Rs 21 as per CES) could be described as people just above the poverty line with some bare minimal additional consumption over subsistence. They are vulnerable because one exogenous shock of death or accident or major hospitalisation or even a temporary loss of job or earnings can drive them to destitution. They constituted about 36 per cent of the population in 2004-05.These four groups of people – extremely poor, poor, marginal and vulnerable – constituted about 77 per cent of the population in 2004-05 (or 75 per cent according to the detailed consumer expenditure schedule). Most of them are living with an average DPCE below the international level of poverty of $ 2 per day and according to us this is the group which should be identified as the common people in India (the ‘aam aadmi’) who have been bypassed by the high growth performance. The remain-ing23percent of the population (or 25 per cent according to the detailed consumer expenditure schedule) belongs to the better off section of the society. In terms of total population, the poor and vulnerable were 836 million out of a total population of 1,090 million in 2004-05 compared to 732 million out of 894 million in 1993-94. Out of that, 237 million people were below the national poverty line and another 207 million belonged to the marginal group of miserable existence at less than a quarter above the poverty line. The next group of vulnerable people with an average per capita consumption of Rs 20 a day were 392 million. More than the absolute level of consumption of those poor and vulnerable, what is more striking is the disparity between this group and the better off section classified here as the middle and high income. The per capita consumption expenditure of the the middle and high income groups taken together is more than four times that of the extremely poor and poor and more than 2.5 times that of the marginal and vulnerable groups taken together. What is more striking is that the ratio of average DPCE of the middle and high income groups to that of the extremely poor and poor showed little sign of a fall although the percentage of people belonging to the poor and extremely poor declined significantly. In income terms, this disparity is likely to be much higher since it is the middle and high income classes who almost wholly account for household savings in the economy. By this reckoning, the divide between the haves and have-nots in India is indeed startling and would be politically as well as socially unacceptable in a democratic polity that guides the process of economic development. The Divide between the Poor and the Non-PoorWhen we examine the social composition of the poor and non-poor as well as their linkage with informal (or unorganised) work status in the economy, there is a significant divide between the two categories. This is presented in Table 5 for 2004-05. The four groups of population – the extremely poor, poor, marginal and vulnerable identified as the “common people” – contain about 88 per cent of the scheduled caste/scheduled tribe (SC/ST) people in the country, which was a small reduction of about 2 percentage points from that in 1999-2000. They also include about 80 per cent of the other backward classes (OBCs) excluding Muslims as compared to 84.5 per cent in 1999-2000. This means that the proportion of OBCs in the poor and vulnerable groups saw a greater decline in the previous five years than that of the theSC/ST population in the same period. However, the trend among Muslims seems to be more disappointing than even that ofSC/ST population. As of 1993-94, 87.4 per cent of Muslims were poor and vulnerable and it fell to 87 per cent in 1999-2000 and to 84.5 per cent in 2004-05, i e, a reduction of only 2.9 percentage points over a period of 10 years.It is especially interesting to consider the work status of the population groups in terms of organised or unorganised (infor-mal) workers.3 In 1999-2000, when the data permits identifica-tion of unorganised workers, 83 per cent of unorganised workers were poor and vulnerable, which came down to 79 per cent in 2004-05. The share of lower social status groups such asSC/ST, Muslims and OBCs in the total unorganised workers was 79 per cent in 1999-2000, which fell to 76 per cent of the total popula-tion during 2004-05. The problem of poverty and vulnerability and its close association with the unorganised worker status is indeed a systemic problem. However, it should be noted that the incidence of unorganised work status is higher among lower social status groups. In 2004-05, it was 95 per cent among SC/STs as well as Muslims and 94 among OBCs while it was 85 per cent among the remaining population group. At the same time, we also note that there is a stratum of middle income (say, middle class) group in every social group. There is also a thin layer (say, creamy layer) from the high income group as well in each social group. However, the variation across social groups is noteworthy. In 2004-05, 11.2 per cent of SCs/STs and 13.3 per cent of Muslims were in the middle income group though their shares in 1993-94 were 8.1 for SCs/STs and 11.1 for Muslims. In 1999-2000, 14.1 per cent of OBCs were in the middle income category and that increased to 17.8 per cent by 2004-05. In contrast, in the group of “Others” the share increased from 29.4 per cent to 34.2 per cent. As far as the high income group is concerned, it is a very thin layer among the SCs/STs and Muslims increasing from 0.9 per cent to 1 per cent for the former and from 1.5 per cent Table 5: Percentage Distribution of Population and Unorganised Workers by Poverty Status and Social Groups(2004-05)Poverty Status Population Un- Organised TotalSC/STOBCMuslimOthersWorkers1 Extremely poor 6.4 [6.5] 10.9[11.4] 5.1[5.2] 8.2[8.2] 2.1[1.6] 5.82 Poor 15.4[15.5] 21.5[21.8] 15.1[15.3]19.2[19.9] 6.4[6.1] 15.03 Marginal 19.0[17.9] 22.4[21.2]20.4[19.6]22.3[19.9]11.1[10.3] 19.64 Vulnerable 36.0[35.5] 33.0[32.7]39.2[38.2]34.8[34.8]35.3[35.0] 38.45 Middle income 19.3 [20.3] 11.2[11.6] 17.8[18.9] 13.3[14.7] 34.2[35.7] 18.76 High income 4.0 [4.4] 1.0[1.3] 2.4[3.0] 2.2[2.5] 11.0[11.3] 2.77 Extremely poor and poor (1+2) 21.8 [22.0] 32.4[33.3] 20.3[20.4] 27.4[28.1] 8.5[7.7] 20.88 Marginaland vulnerable (3+4) 55.0 [53.3] 55.4[53.8] 59.6[57.8] 57.1[54.7] 46.3[45.4] 57.99 Poor and vulnerable (7+8) 76.7 [75.3] 87.8[87.1] 79.9[78.2] 84.5[82.8] 54.8[53.1] 78.710 Middleand high income (5+6) 23.3 [24.7] 12.2[12.9] 20.2[21.8] 15.5[17.2] 45.2[47.0] 21.311 All 100[100] 100.00[100] 100[100] 100[100] 100[100] 100.0(1) Figures in square brackets indicate the estimates based on the detailed consumer expenditure survey. (2) “Others” consist of all those who donot belong to SCs/STs, OBCs and Muslims. (3) The social status of about 0.66 million people were not reported though included in the total population. (4) Unorganised workers consist of those working in the unorganised sector or households, excluding regular workers with social security benefits provided by theemployers and the workers in the organised sector without any employment or social security benefits provided by the employers. (5) The method of identifying different social groups and other characteristics is explained in the Appendix.
SPECIAL ARTICLEEconomic & Political Weekly EPW march 15, 200853to 2.2 per cent for the latter. In the case of OBCs, it was 1.4 per cent in 1999-2000 and 2.4 per cent in 2004-05 while for ‘Others’ it increased fro m 6.4 to 11 per cent during this period.Similarly, of the unorganised workers, 17 per cent belonged to the middle and high income groups in 1999-2000, which increased to 21 per cent in 2004-05. We also examined the distribution of different social groups in each poverty status. About 83 per cent of those in the SC/ST, Muslims and OBC groups belong to the poor and vulnerable group while their share in the middle and high income groups was only 54 per cent in 2004-05. From the detailed classification of the population into six income/expenditure classes and their linkages with social groups and unorganised workers, we can now move on to compress them into three distinct groups for purposes of compari-son. These are (i) extremely poor and poor, (ii) marginal and vulnerable, and (iii) middle and high income. When we combine the first two, we get the larger group of poor and vulnerable. Wherever warranted we would revert to our six-group classification to bring out the differences as sharply as possible.Table 6 gives the shares of unorganised workers in each poverty status and Table7 gives their distribution by social groups in each poverty status category. In the middle and high income group, we found that about 80 per cent of the total workers are unorganised workers and amongthem about 40.4 per cent come from Others (i e, upper caste Hindus, Christians, Sikhs and other groups other thanSC/ST, Muslims and OBCs). Table 7 brings out clearly how the unorganised workers belonging to the extremely poor and poor groups are mostly accounted for bySC/STs, Muslims and OBCs to the extent of 91.4 per cent. If we include the marginal and vulnerable groups, this share comes down to 84.4 per cent, still much higher than their population share of 76 per cent.4 Educational Dimension of Poverty and VulnerabilityThe level of education is an important determinant of poverty status as 86 per cent of the illiterates and 83.3 per cent of those with educational level up to primary school were in the lower strata of the poor and vulnerable in 2004-05. In the case of those with educational attainment of secondary and above but below graduation, 52 per cent and among graduate and above 30 per cent only were in the poor and vulnerable group. On the other hand, about 70 per cent of the graduate and above population in 2004-05 was in the middle and high income group, which had increased from 62 per cent in 1993-94. Whereas 86 per cent of those below primary level of education were in the poor and vulnerablegroupin1993-94and this declined to only 83 per cent in 2004-05. The percentage distribution of population in differenteducationalstatusby poverty status is given in Table 8. While there has been an overall decline of population in the poverty status category of the poor and vulnerable during the 11-year period, the reduction was just over 2 percentage points among illiterates and about 2.5 percentage points among those with up to only primary level of education. In the case of the secondary and above but below graduation group, the reduc-tion in the percentage share of the poor and vulnerable was over 7.3 percentage points and those with graduation and above was 7.9 per cent.The common people belonging to the poor and vulnerable group are mostly illiterates or with educational level up to primary. A small proportion of them have gone through middle and secondary level of education and a minuscule fraction are graduate and above. In Table 9 (p 54), we give the distribu-tion of the population in different educa-tional levels in different poverty status groups across the different social groups. What this table reveals is that given the same level of education, a person who is a SC/ST or Muslim is likely to be more in the poor and vulnerable category than a person belonging to the higher social status group. For example, 90 per cent of SC/ST with only up to primary level of education in 2004-05 belonged to the poor and vulnerable group compared to 67 per cent for the group “Others” representing upper caste Hindus, Christians and Sikhs. We have also examined the details for the social groups for three different time points. What we find isthataspeople acquire higher levels of education they are able to get out of the trap of poverty and vulnerability, but the speed with which that happens is slower for the lower social status groups than the higher ones. Here one should perhaps note the other enabling or disabling factors, as the case may be, in Table 6: Distribution of Total and Unorganised Workers (2004-05)Poverty Status Total Work Unorganised Share of WorkersParticipation WorkersUnorganised (in Million) Rate (%) (in Million) Workers to Total Workers (%)1 Extremely poor andPoor 89.3 37.7 87.8 98.32 Marginaland vulnerable 255.8 42.7 244.8 95.73 Poorand vulnerable(1+2) 345.1 41.3 332.6 96.44 Middleand highincome 112.4 44.3 90.0 80.15 Total 457.5 42.0 422.6 92.4Work Participation Rate = (Total workers/Total Population)*100. Table 7: Percentage Distribution of Unorganised Workers in Different Poverty Status by Social Groups(2004-05)Poverty Status Total ST / SC OBCs Muslim Others 1 Extremely poor and poor 100.0 44.9 33.6 12.9 8.62 Marginal and vulnerable 100.0 30.5 40.6 10.8 18.13 Poor and vulnerable (1+2) 100.0 34.3 38.7 11.3 15.64 Middle and high income 100.0 16.4 35.6 7.6 40.45 Total 100.0 30.5 38.0 10.5 20.9Each row does not add up to 100 as the distribution by social status does not include non-reporting cases in 2004-05.Table 8: Percentage Distribution of Population (Age 15+) in Different Educational Status by Poverty StatusPoverty Status Illiterate Up to Primary Middle SecondaryGraduate Total and Above and Above butBelow Graduate 1993-94 1 Extremely poor and poor 36.9 32.7 18.3 10.9 4.6 26.8 2 Marginal and vulnerable 51.5 53.4 56.4 48.8 33.0 52.3 3 Poor and vulnerable (1+2) 88.4 86.1 74.7 59.7 37.6 79.1 4 Middle and high income 11.6 14.0 25.3 40.3 62.4 20.95 Total 100.0 100.0 100.0 100.0 100.0 100.02004-05 1 Extremely poor and poor 27.1 24.1 13.7 7.4 2.8 18.2 2 Marginal and vulnerable 59.0 59.2 57.5 45.0 26.9 54.4 3 Poor and vulnerable (1+2) 86.1 83.3 71.2 52.4 29.7 72.6 4 Middle and high income 13.9 16.7 28.8 47.6 70.3 27.45 Total 100.0 100.0 100.0 100.0 100.0 100.0Source: Ibid.
SPECIAL ARTICLEmarch 15, 2008 EPW Economic & Political Weekly54determining the speed of such a transition. For lower social status groups, the other disabling factors could be numerous but the main ones are likely to be the weak social network, access to information especially on employment opportunities, weakfamily support due to the low education of parents and relatives and the constraints in the larger social milieu that is often discriminatory. In the case of Muslims, the congruence between poverty and educational status is not very different from that in the case of theSC/ST population. The percentage in the poor and vulnerable group with low education (up to primary level) at about 87 per cent in 2004-05 is closer to that of theSC/ST population.In the case ofOBCs, 84 per cent of those with low education were poor and vulnerable in 2004-05. And this makes their position closer to SC/ST and Muslims than the Others where it was 67 per cent. Within the group of low educated (up to primary level) we should in fact note the congruence between illiteracy and poverty and vulnerability across social groups as given in Table 9. For example, in 2004-05, 91 per cent of SC/ST illiterates, 89 per cent of Muslim illiterates and 86 per cent of OBC illiterates were in the poor and vulnerable group. This was 71 per cent for the group “Others”. This close congruence between illiteracy and poverty is perhaps not unexpected. What is of relevance here is that the incidence of illiteracy is higher among low social status groups. In 2004-05, these were 51 per cent for SC/ST, 41 per cent for Muslims, 39 per cent for OBCs and 19 for Others. Such a situation results in a formidable initial handicap for all but more so, and in a stark form, for those at the bottom.We have summarised in Table 10 the impact of the improve-ment in educational attainment over a period of five years during 1999-2000 to 2004-05. This was not uniform across different social groups. The decline in the percentage of illiterates among SCs/STs during 1999-2000 to 2004-05 was the highest though they still had the largest percentage of illiterates. There was, however, only a marginal increase of 0.3 percentage points in graduates and above among SCs/STs.In the case of Muslims, the decline in the percentage of illiter-ates between 1999-2000 and 2004-05 was about 4.8 percentage points and the decrease in the poor and vulnerable group among them was 4.5 percentage points. The overall increase in the percentage of graduates and above among Muslims was 0.3 percentage points, while the increase in the poor and vulnerable of the group was marginal at 0.04 percentage points.TheOBCs were able to reduce the percentage share of illiter-ates by about 6.5 percentage points from 45.8 per cent in 1999-2000 to 39.3 per cent in 2004-05 with a corresponding reduction of about 5.8 percentage points in the illiterate poor and vulnerable. The graduate and above among OBCs, on the other hand, increased from 3.1 per cent in 1999-2000 to 4.1 per cent in 2004-05 with a marginal increase of only 0.3 percentage points in the poor and vulnerable among them.In the case of those not belonging to SCs/STs, Muslims and OBCs, the share of illiterates reduced by about 4.5 percentage points from 23.5 per cent to 19 per cent, while those belonging to the poor and vulnerable among illiterates reduced by 3.7 percentage points. The graduates and above, on the other hand, increased byabout 1.9 per-centage points from 12.1per cent to 14 percent with a mar-ginal increase of 0.2 percentage points in the poor and vulnerable group among them.The SC/STs and OBCs achieved a larger reduction in the percentage of illiterates. The impact in reducing poverty and vulnerability was also greater among SCs/STs and Muslims. On the other side of the spectrum, however, the SCs/STs and Muslims were not able to increase the percentage of graduates and above significantly. The Muslims were able to avoid an increase in the percentage share of the poor and vulnerable Table 9: Percentage Distribution of Population in Specific Educational Categories by Poverty Status and Social Groups(2004-05)Social Poverty Status Illiterate Up to Primary Middle Secondary Graduate Total Category and Above butand BelowGraduateAbove SC/ST Poor and vulnerable 91.2 89.7 80.7 65.4 47.8 85.0 Middle and HIG 8.8 10.3 19.3 34.7 52.2 15.0 All 100.0 100.0 100.0 100.0 100.0100.0Muslims Poor and vulnerable 89.5 87.2 76.5 63.1 43.4 81.3 Middle and HIG 10.6 12.8 23.6 36.9 56.6 18.7 All 100.0 100.0 100.0 100.0 100.0100.0OBC Poor and vulnerable 85.6 83.6 74.3 59.9 39.4 76.3 Middle and HIG 14.5 16.4 25.7 40.1 60.6 23.7 All 100.0 100.0 100.0 100.0 100.0100.0Others Poor and vulnerable 70.6 66.9 57.7 39.5 21.6 50.9 Middle and HIG 29.4 33.1 42.3 60.5 78.4 49.1 All 100.0 100.0 100.0 100.0 100.0100.0All Poor and vulnerable 86.1 83.3 71.2 52.4 29.7 72.6 Middle and HIG 13.9 16.7 28.8 47.6 70.3 27.4 All 100.0 100.0 100.0 100.0 100.0100.0Source: ibid. HIG means high income group.Table 10: Changes in Educational Attainment and PovertyReduction by Social Groups between 1999-2000 and 2004-05 (in percentage points)Social Group Reduction in Reduction in Increase in Increase in Poor Illiterates Poor and Graduate and Vulnerable Vulnerable and Above among amongGraduates IlliteratesandAboveSC/ST 6.9 6.5 0.3 0.2Muslims 4.8 4.5 0.3 0.04OBCs 6.5 5.8 1.0 0.3Others 4.5 3.7 1.9 0.2Source: From data in NSS 55th and 61st Round Survey on Employment-Unemployment, 1999-2000 and 2004-05. Computed.Table 11: Percentage Distribution of Unorganised Workers in Specific Educational Levels by Poverty Status in Social Groups(2004-05)Social Poverty Status Illiterate Up to Primary Middle Secondary Graduate Total Category and Above and Above butBelow GraduateSC/ST Poor and vulnerable 91.7 90.4 83.5 76.8 63.5 88.3 Middle and HIG 8.3 9.6 16.5 23.2 36.5 11.7 All 100.0100.0100.0100.0100.0100.0Muslims Poor and vulnerable 90.1 88.2 79.5 71.4 50.3 84.4 Middle and HIG 9.9 11.8 20.5 28.6 49.7 15.6 All 100.0100.0100.0100.0100.0100.0OBC Poor and vulnerable 85.8 84.2 76.5 68.2 50.5 79.9 Middle and HIG 14.2 15.8 23.5 31.8 49.5 20.1 All 100.0100.0100.0100.0100.0100.0Others Poor and vulnerable 72.0 69.4 61.6 48.0 31.4 58.6 Middle and HIG 28.0 30.6 38.4 52.0 68.6 41.4 All 100.0100.0100.0100.0100.0100.0All Poor and vulnerable 87.1 84.8 74.6 61.7 41.0 78.4 Middle and HIG 12.9 15.2 25.4 38.3 59.0 21.6 All 100.0100.0100.0100.0100.0100.0
SPECIAL ARTICLEEconomic & Political Weekly EPW march 15, 200855among the graduates. The OBCs and others were able to increase the percentage shares of graduates and above without any signi-ficant increase in the poor and vulnerable group among them.The linkage between poverty and vulnerability and educa-tional attainment among unorganised workers is similar to that of the total population. About 81.7 per cent of the extremely poor and poor unorganised workers were either illiterates or with education up to primary level. Among the poor and vulnerable unorganised workers 73.1 per cent had educational attainment up to primary level. In Table 11 (p 54), we give the distribution of unorganised workers in each social group by educational levels. This is quite revealing in the sense that despite the same level of education, the incidence of poverty and vulnerability varies depending on the social group one belongs to. For example, 92 per cent of illit-erates belonging toSC/ST were in the poor and vulnerable group whereas only 72 per cent of illiterates belonging to Others were poor and vulnerable. The former is the bottom layer followed by Muslims,OBCs and Others at the top – a pattern that emerges throughout our analysis. The above analysis brings out clearly that the impact of better education in poverty reduction was not uniform in different social groups. The details are summa-rised in Table 12. In the case of both the OBCs and Others, there was a higher growth in graduate and unorganised workers with a slower growth in the segment of the poor and vulnerable group. However, the situation was worse in the case of SC/ST communities.There are two unmistakable characteristics that emerge from the presentation of the status of poor people. First, the groups of extremely poor, poor, marginal and vulnerable constitute the category of common people in accounting for about three-fourths of the population. They also account for the socially underprivi-leged group of SCs/STs, Muslims, andOBCs. Secondly, most of these people also belong to the category of unorganised workers. In other words, our common people of 836 million are largely co-terminus with the category of poor and vulnerable SC/STs, Muslims and OBCs. And in terms of productive activity they overwhelmingly belong to the group of unorganised workers. Unorganised workers are either self-employed or regular wage employees or casual workers. Among the unorganised workers, over 90 per cent of the casual workers and about 75 per cent of the self-employed were in the category of the poor and vulnera-ble. Thus most of the casual and self-employed unorganised workers were in the category of common people or ‘aaam aadmi’. The unorganised workers in the middle and high income groups were largely self-employed at a higher level (given their higher educational levels) and regular salaried. The percentage distri-bution of unorganised workers in each activity by poverty status as well as their composition (in row total) is given in Table 13.The relationship between the type of activity, poverty status and education needs to be emphasised here. While the self-employed among the poor and vulnerable group was 57.2 per cent in 2004-05, about 69 per cent of the poor and vulnerable was either illiterates or with education up to primary level. While about 71 per cent of those in the middle income category were self-employed, about 51 per cent of the unorganised workers had either no educa-tion or up to primary level.The extremely poor and poor unorganised workers consisted of 48 per cent self-employed and 45 per cent casual workers in 2004-05 though their shares in 1999-2000 were 44 per cent and 51 per cent respectively. The poor and vulnerable unorganised workers constituted 57 per cent self-employed and 35 per cent casual workers in 2004-05 as compared to 54 per cent and 40 per cent respectively in 1999-2000. The unorganised workers, who are mostly self-employed and casual workers, thus belong to vulner-able and below poverty categories. (The estimates for previous year may be seen from the full text on the web site of the journal.)5 Consumption, Household Characteristics and Work ParticipationThere are significant associations between consumption and household characteristics and work participation. Low consump-tion is associated with a low asset base (e g, land), bigger house-hold size and a higher dependency ratio. This association is evident from Table 14 (p 56). Though the work participation rate declined for all groups in 1999-2000, it improved in 2004-05. The middle and high income groups experienced the highest increase in work participation rate between 1999-2000 and 2004-05, suggesting greater work opportunities apart from the fact that the jobs they secured in general were better paid with greater social security.It has already been indicated earlier that three quarters of the Indian people have an average daily per capita consumption expenditure of below Rs 20.4 From a social point of view the disparity between different classes often becomes sensitive as people tend to compare their position vis-à-vis other groups. The details of per capita consumption for three time points are given in Table 15 (p 56). While the ratio of the average daily per capita expenditure of the high income group to that of the poor was 7.6 in 1993-94, the same in the case of the middle income Table 12: Changes in Educational Attainment and Poverty Reduction of Unorganised Workers by Social Groups between 1999-2000 and 2004-05 (in percentage points)Social Reduction in Reduction in Increase in Increase in Poor Group Illiterates Poor and Graduate and Vulnerable Vulnerableand Above among amongGraduates IlliteratesandAboveSC/ST 7.5 7.2 0.1 0.1Muslims 4.5 4.7 0.6 0.2OBCs 6.6 6.4 0.9 0.3Others 4.8 4.3 1.3 0.3Table 13: Percentage Distribution of Unorganised Workers in Different Activity Status by Poverty Status(2004-05)Poverty Status Self-employed Regular Wage Casual Workers Total Employees1 Extremely poor 4.3 (44.8) 4.3 (6.8) 9.1 (48.4) 5.8 (100)2 Poor 12.3 (49.5) 11.0 (6.7) 21.4 (43.8) 15.0 (100)3 Marginal 17.8 (54.8) 15.4 (7.2) 24.3 (38.0) 19.6 (100)4 Vulnerable 40.3 (63.3) 36.0 (8.6) 35.2 (28.2) 38.4 (100)5 Middle class 22.0 (71.0) 27.6 (13.5) 9.4 (15.5) 18.7 (100)6 High income 3.3 (74.0) 5.7 (19.7) 0.6 (6.3) 2.7 (100)7 Extremely poor and poor (1+2) 16.6 (48.2) 15.3 (6.7) 30.6 (45.1) 20.8 (100)8 Marginal and vulnerable (3+4) 58.1 (60.4) 51.4 (8.1) 59.4 (31.5) 57.9 (100)9 Poor and vulnerable (7+8) 74.7 (57.2) 66.7 (7.8) 90.0 (35.1) 78.7 (100)10 Middle and high income (5+6) 25.3 (71.4) 33.3 (14.3) 10.0 (14.4) 21.3 (100)11 All 100 (60.2) 100 (9.1) 100 (30.7) 100 (100)Figures in the brackets are the percentage distribution of unorganised workers in different poverty status by activity status.
SPECIAL ARTICLEmarch 15, 2008 EPW Economic & Political Weekly56group was 3.1. These ratios did not change much in 1999-2000 and 2004-05 except for a decrease to 7.3 in the case of the high income group in 1999-2000. Similarly, the ratio of average MPCE of the extremely poor and poor group to that of the middle and high income group was more than 4 in all the years. However, disparity in consumption expenditure may not be an appropriate indicator of income disparity since lower income groups do not have the capacity to save (or save only a small proportion of their income) compared to the higher income groups. As far as the growth in consumption is concerned, the monthly total consumption expenditure has been increasing at an average annual rate of 3.3 per cent during the period from 1993-94 to 2004-05. While the growth rate during the period 1993-94 to 1999-2000 was lower at 2.6 per cent, it accelerated to about 4.2 per cent during 1999-2000 to 2004-05. The growth was largely contributed by the middle and high income groups with annual growth rates of about 4.9 per cent during 1993-94 to 2004-05. The poor and vulnerable taken together, however, registered a positive growth rate of 2.2 per cent during the same period. In terms of MPCE, the overall growth rate during 1993-94 to 2004-05 was about 1.5 per cent per year. The extremely poor and poor registered an average growth rate of 0.8 per cent while the poor and vulnerable taken together registered a growth rate of 1 per cent per annum. The growth rate recorded by the middle and high income groups in the average per capita consumption was 0.7 per cent. This low growth rate in the case of high income group may be because of diminishing marginal rate of consump-tion after a threshold level of income. There is a difference between per capita consumption growth and total consumption growth. This is because the size of each expenditure class has undergone some change. For example, 91 million have been added to the middle and high income group whereas 104 million were added to the group of the poor and vulnerable. Those who have been added in the high income group come with a higher per capita expenditure and thus accompanied by an increase in the average monthly expenditure. This needs to be noted because the per capita consumption expenditure indicates the absolute level of purchasing power and its increase on a per capita basis is directly related to the welfare of the people. The total consumption indicates the total purchasing power of the respective group with a corresponding change in the population size of the group, determining the size of the market of different goods consumed by them.In the general profiling of the common people as well as the higher income people, it is important to bring in the composition of consumption because this has significant implications on the pattern of demand and consequent nature of the process of economic growth. If a significant share of consumption is accounted for by the high income groups, then the pattern of demand for goods and services undergoes a change in favour of non-basics and durable consumer goods, luxury goods and specific high value goods consumed by such classes. There will be correspondingly less income spent on wage goods consumed by the lower income groups with a smaller size of their markets. Alternatively, the new investments will seek to satisfy the demand of the richer classes because of the ready market, high profitability and consequent circulation of income within those classes. The long-term result could be strengthening the division of the economy where a modern or organised sector – often capital, knowledge, skill and service intensive – would cater to the demands of the rich while the demand for wage goods of a low value nature would be met by a less modern or unorganised sector, often traditional (such as artisanal and craft production along with peasant agricul-ture) with very little capital intensity, low level of knowledge and skill and access to such inputs as credit.The composition of the consumption of different expenditure classesis given in Table 16. First, Table14:ComparativeEstimatesofHouseholdCharacteristicsbyPovertyStatus (All India)Poverty Status MPCE (Rs) Land HH Size No of Dependency Workforce Possessed Workers Ratio (%) Participation (ha) Rate1993-94 1 Extremely poor and poor 164 0.61 5.51 2.21 2.50 40.03 2 Marginal and vulnerable 294 0.90 4.80 2.04 2.36 42.42 3 Poor and vulnerable (1+2) 246 0.80 5.04 2.09 2.41 41.53 4 Middle and high income 697 1.03 3.79 1.64 2.31 43.34 5 Total 328 0.85 4.76 1.99 2.39 41.861999-2000 1 Extremely poor and poor 266 0.41 5.81 2.14 2.72 36.81 2 Marginal and vulnerable 466 0.65 4.96 1.99 2.49 40.23 3 Poor and vulnerable (1+2) 402 0.58 5.20 2.04 2.56 39.13 4 Middle and high income 1,091 0.67 3.70 1.51 2.44 40.94 5 Total 534 0.60 4.83 1.90 2.54 39.482004-05 1 Extremely poor and poor 321 0.38 5.74 2.18 2.64 37.68 2 Marginal and vulnerable 550 0.59 4.83 2.07 2.33 42.70 3 Poor and vulnerable (1+2) 485 0.54 5.06 2.10 2.41 41.28 4 Middle and high income 1,388 0.67 3.62 1.61 2.25 44.32 5 Total 695 0.58 4.63 1.95 2.37 41.98Source: Computed from NSS 50th, 55th and 61st round survey on Employment-Unemployment Surveys of 1993-94, 1999-2000 and 2004-05.Table 15: Consumption Expenditure by Poverty StatusPoverty Status Total Monthly Consumption Average Monthly Per Capita (Rs Million at 1993-94 Prices) Consumption(Rsat1993-94Prices) 1993-94 1999-20002004-05 1993-941999-20002004-051 Extremely poor and poor 45,072 44,636 42,603 164.4 170.7 179.72 Marginal and vulnerable 1,34,578 1,64,209 1,85,512 294.0 299.0 309.73 Poor and vulnerable (1+2) 1,79,650 2,08,845 2,28,115 245.5 257.6 272.84 Middle and high income 1,13,224 1,33,464 1,91,340 697.1 690.2 754.85 Total 2,92,874 3,42,308 4,19,453 327.5 340.9 385.0Growth Rates in Consumption ExpenditurePoverty Status Annual Growth Rate in Annual Growth Rate in Monthly Consumption Per Capita Consumption 1999-2000 2004-05 2004-05 1999-2000 2004-05 2004-05 to to to to toto 1993-94 1999-2000 1993-94 1993-941999-20001993-941 Extremely poor and poor -0.2 -0.9 -0.5 0.6 1.0 0.82 Marginal and vulnerable 3.4 2.5 3.0 0.3 0.7 0.53 Poor and vulnerable (1+2) 2.5 1.8 2.2 0.8 1.2 1.04 Middle and high income 2.8 7.5 4.9 -0.2 1.8 0.75 Total 2.6 4.2 3.3 0.7 2.5 1.5Source: Ibid.Table 16: Percentage Share of Total Monthly Expenditure by Poverty StatusPoverty Status Share of Total Monthly Expenditure 1993-941999-20002004-051 Extremely poor and poor 15.4 13.0 10.22 Marginal and vulnerable 45.9 48.0 44.23 Poor and vulnerable (1+2) 61.3 61.0 54.44 Middle and high income 38.7 39.0 45.65 Total 100 100 100Source: Ibid.
SPECIAL ARTICLEEconomic & Political Weekly EPW march 15, 200857thetotalconsumptionexpenditure of the poor and vulnerable constituting a little over three-fourths of Indian people accounts for only 54 per cent while that of the middle and high income groups accounting for a little less than a quarter of the Indian people (23 to 24 per cent as per our estimate) spends the remain-ing 46 per cent in 2004-05. The correspondingfiguresfor1993-94 were that the poor and vulnerableconstituting81 per cent accounted for 61 per cent of total consumption expenditure,while the remaining 19 per cent of the high income group accounted for 39 per cent of consumption. That is to say, in terms of per capita, the consumption ratio worsenedsomewhatfrom 1:2.7 in 1999-2000 to 1:2.8 in 2004-05 between these two classes. If this trend continues, it should be a matter of serious concern to policymakers ingeneral and the politicalleadership in particular.An analysis of item-wise consumption expenditure dur-ing 1999-2000 and 2004-05, given in Table 17, clearly brings out the shift in per-centage share of expenditure from food and beverages to non-food in all classes of people. The maximum de-cline in the share of food consumption was recorded by the middle and high in-come group. However, the group improved the share of expenditure in education by 2.9 percentage points, other non-food items by 2 percent-age points and durable goods by 1.5 percentage points.The annual consumption expenditures of different groups of items at current prices by different poverty groups during 1999-2000 and 2004-05 are given in Table 18.While the overall increase in total consumption expendi-ture between 1999-2000 and 2004-05 was Rs 26,94,769 million, Rs 17,11,851 million (63.5 per cent) was contri-buted by the middle and high income group. About 34.9per cent of the growth in the consumption expend-iture of the group was in othernon-food items and durable goods. In the case of the poor and vulnerable group, 68.6 per cent of the growth in consumption expenditurewasonfood and essential non-food items.6 The Consumption Divide The average MPCE and other household characteristics by poverty status and social groups are given in Tables 19 and 20 (p 58) respectively. These parameters vary significantly across different social groups even within each poverty status. In 1993-94,SC/ST households had the lowestMPCE, the largest average number of workers per household and the highest WPR among all the poverty status categories. However, the highest average house-hold size and dependency ratios (the household size divided by the number of workers) were among Muslims in all the poverty categories. The Muslims also had the lowestWPR and average number of workers per household irrespective of the poverty category. The relative positions of different social groups did not change much by 1999-2000 except that the average number of workers per household in the middle and high income group was the highest among OBCs instead of SC/ST house-holds. There was no change even in 2004-05 except that among marginal and vulner-able households, the average number of workers per household was the highest among OBCs and not SCs/STs.The average MPCE of the middle and high income group was about four times that of the extremely poor and poor group in the case of SC/ST as against 3.8 times in the case of Muslims and 4.2 times in the case of Others in the year 1993-94. The pattern did not change by 2004-05 except that the ratios got reduced in the case of SC/STs and Muslims and it increased in the case ofOBCs and Others. In other words, the disparity among those belonging to the social categoriesofOBCs andOthers got worsened. In Table 17: Percentage Distributionof Total Consumption Expenditure by Specific Groups of Items Poverty Status Food and Essential Education Medical Other Durable BeverageNon-foodNon-foodGoods ItemsItems 1999-2000 1 Extremely poor and poor 67.8 16.8 1.3 3.4 9.4 1.3 2 Marginal and vulnerable 63.0 15.9 2.1 5.0 12.3 1.7 3 Poor and vulnerable 64.1 16.1 1.9 4.6 11.7 1.6 4 Middle and high Iincome 48.7 14.1 4.1 7.3 20.9 4.9 5 All 58.1 15.3 2.8 5.7 15.3 2.92004-05 1 Extremely poor and poor 63.3 19.5 1.9 3.5 10.0 1.8 2 Marginal and vulnerable 58.5 18.2 3.1 4.8 13.0 2.4 3 Poor and vulnerable 59.4 18.4 2.9 4.6 12.4 2.3 4 Middle and high income 41.5 14.8 7.0 7.4 23.0 6.4 5 All 51.2 16.8 4.7 5.9 17.2 4.2Source: Computed from NSS 55th and 61st round survey on Employment-Unemployment, 1999-2000 and 2004-05.Table 18: Annual Consumption Expenditure(Rs million at current prices)Poverty Status Food and Essential Education Medical Other Durable Total Beverage Non-food Non-food Goods Items items 1999-2000 1 Extremely poor and poor 5,74,340 1,41,994 10,627 28,786 79,772 11,056 8,46,575 2 Marginal and vulnerable 19,62,423 4,96,292 66,140 1,54,289 3,81,593 53,278 31,14,013 3 Poor and vulnerable (1+2) 25,36,763 6,38,286 76,766 1,83,075 4,61,365 64,333 39,60,588 4 Middle and high income 12,49,947 3,61,865 1,04,884 1,86,227 5,37,878 1,25,474 25,66,276 5 Total 37,86,710 10,00,150 1,81,651 3,69,3029,99,2431,89,80765,26,8642004-05 1 Extremely poor and poor 5,93,408 1,82,343 18,093 33,165 93,544 16,536 9,37,089 2 Marginal and vulnerable 23,44,262 7,29,615 1,23,088 1,92,252 5,19,382 97,818 40,06,417 3 Poor and vulnerable (1+2) 29,37,670 9,11,958 1,41,181 2,25,417 6,12,926 1,14,353 49,43,506 4 Middle and high income 17,70,914 6,32,626 2,98,232 3,15,993 9,85,263 2,75,099 42,78,127 5 Total 47,08,584 15,44,584 4,39,413 5,41,41015,98,189 3,89,453 92,21,633Source: Ibid.Table 19: Average MPCE (Rs) and Worker Participation Rate (WPR in %) by Social Group and Poverty StatusYear Social Group Extremely Poor Marginal and Middle and All Ratio of and Poor Vulnerable High income Consum- MPCE WPRMPCE WPR MPCEWPR MPCE WPRption1993-94 SC/ST 156 44.1 276 48.2 617 50.5 255 46.6 4.0 Muslims 174 30.2294 33.7 662 37.6295 32.9 3.8 OBCs NA NANA NANA NA NA NANA Others* 168 39.5301 41.7 714 42.8364 41.44.2 All 164 40.0294 42.4697 43.4328 41.9 4.22004-05 SC/ST 303 40.9 516 46.5 1,118 49.0 520 45.0 3.7 Muslims 348 30.4556 34.7 1,279 37.4611 33.9 3.7 OBCs 324 37.7 541 44.21,211 48.1632 43.63.7 Others349 35.6608 39.9 1,611 41.61039 40.34.6 All 321 37.7 550 42.7 1,388 44.3695 42.04.3* -Others include OBCs in 1993-94. NA = Not Available. Ratio = Average, MPCE of Middle & High Income Group/AverageMPCEof Extremely Poor and Poor. Source: Computed fom NSS 50th and 61st round survey on Employment-Unemployment, 1993-94 and 2004-05.
SPECIAL ARTICLEmarch 15, 2008 EPW Economic & Political Weekly58the case ofSC/ST and Muslims,theaverageMPCEs remained lower than the other two groupswithlesserdisparities.While there has been a decline in the overall average house-hold size from 4.8 persons in 1993-94 to 4.6 in 2004-05, those belonging toSC/STs and Muslims could notbenefitfromthe same as given in Table 20. In thecase of SC/STs, the aver-age household size increased from 4.6 in 1993-94 to 4.7 in 1999-2000 and then reduced to 4.6 in 2004-05. In the case of Muslims, the av-erage householdsize increased from5.2 in 1993-94 to 5.5 in 1999-2000 and then to 5.3 in 2004-05. The average house-hold sizes of the ex-tremely poor and poor in all the so-cial groups were considerably higher than those in other poverty categories. Consequently they alsohad higher de-pendency ratios, making theirliving conditions deplor-ably poor.While the average MPCE of the middle and high income group during 2004-05 was Rs1,388, it was as low as Rs 321 in the case of the extremely poor and poor. Even within the poverty category of extremely poor and poor,SC/ST communities were the most poor with anMPCE of Rs 303 in 2004-05. For example, the MPCE of the SC/ST poor is only 87 per cent of the MPCE of Other poor in 2004-05. Thus, the poor are not the same across all social groups. In other words the depth of poverty is the highest for the SC/ST communities followed byOBCs and Muslims. The average annual growth rates in total and per capita consumption expenditure of different social groups belonging to various poverty categories are given in Table 21. While the overall growth rate inMPCE during the period 1993-94 to 2004-05 was merely 1.5 per cent per annum, it was as high as 4 per cent in the case of those not belonging to the SC/ST communities, Muslims and OBCs. The growth rate was the lowest, 1.2 per cent, in the case of Muslims and 1.3 per cent in the case of SCs/STs. During the period of 1999-2000 to 2004-05, the growth rate was relatively higher with an overall average of 2.5 per cent and 4.1 per cent in the case of others. The growth rates of SCs/STs, Muslims and OBCs were 1.7 per cent, 2.2 per cent and 2.5 per cent respectively. The percentage shares of different poverty categories in total monthly consumption expenditure by social group in different years are given in Table 22 (p 59). The share of the extremely poor and poor decreased from 15.4 per cent in 1993-94 to 10.2 per cent in 2004-05. The share of the marginal and vulnerable groups also declined from 45.9 per cent in 1993-94 to 44.2 per centin2004-05. It was mainly contributed by the shift of the persons from poor categories to higher categories. However, the shift has not been uniform across all social groups. The most disadvantaged groups were againSCs/STs, Muslims and OBCs in that order.7 A Summing UpTo sum up, an overwhelming majority of the Indian population, around three quarters, is poor and vulnerable and it is a stagger-ing 836 million as of 2004-05. This includes 70 million or 6.4 per cent who may be characterised as extremely poor with a per capita consumption of less than or three-quarters of the official poverty line. To this should be added 167 million of those who are poor with consumption not more than that fixed as the official poverty line. If this is relaxed to include those with a percapitaconsumption of up to 25 per cent above the poverty line, calledmarginally poor here, then we find another 207 million. These three groups account for 444 million or 40.8 per cent of the population. To this we add those with a per capita consump-tion between 1.25 and two times the poverty line asvul-nerable and this group of poor and vulnerable comes to 836 million of Indians or well over 75 per cent of the population. The next major finding is the close association between poverty and vulnerability with one’s social identity. The two socialgroupswho are at the bottom by this classification are the SCs/STs, who consti-tute the bottom layer, and the Muslims, who are in the next layer. This does not mean that the other groups are far better off. The next group is the OBCs but better than the two bot-tom layers. Even for those who do not belong to any of these groups, the incidence is 55 per cent.The obverse of this is equally important. It says that in all communities there is a class of better-off, called the middle and high income group, which varies with social identity. Therefore economic differentiation across Table 20: Household Size and Dependency Ratios by Poverty Status and Social GroupsSocial Group HH size Dependency Ratio 1993-941999-20002004-051993-941999-20002004-05Extremely poor and poor ST/SC 5.3 5.5 5.5 2.3 2.5 2.4 Muslims 6.1 6.7 6.6 3.3 3.5 3.3 OBC NA 5.8 5.8 NA 2.7 2.7 Others 5.5 5.9 5.7 2.5 3.0 2.8 All 5.5 5.8 5.7 2.5 2.7 2.6Marginal and vulnerable ST/SC 4.4 4.6 4.6 2.1 2.2 2.2 Muslims 5.1 5.5 5.3 3.0 3.1 2.9 OBC NA 4.9 4.8 NA 2.4 2.3 Others 4.9 5.2 5.0 2.4 2.8 2.5 All 4.8 5.0 4.8 2.4 2.5 2.3Poor and vulnerable ST/SC 4.8 4.9 4.9 2.2 2.3 2.3 Muslims 5.5 5.9 5.6 3.1 3.3 3.0 OBC NA 5.2 5.0 NA 2.5 2.4 Others 5.1 5.3 5.1 2.4 2.8 2.6 All 5.0 5.2 5.1 2.4 2.6 2.4Middle and high income ST/SC 3.4 3.4 3.4 2.0 2.2 2.0 Muslims 4.0 3.9 3.9 2.7 2.6 2.7 OBC NA 3.6 3.6 NA 2.3 2.1 Others 3.9 3.8 3.7 2.3 2.6 2.4 All 3.8 3.7 3.6 2.3 2.4 2.3Total ST/SC 4.6 4.7 4.6 2.1 2.3 2.2 Muslims 5.2 5.5 5.3 3.0 3.2 3.0 OBC NA 4.8 4.7 NA 2.4 2.3 Others 4.8 4.7 4.4 2.4 2.7 2.5 All 4.8 4.8 4.6 2.4 2.5 2.4Table 21: Growth Rates in Per Capita Consumption Expenditure by Poverty Status across Social Groups Annual Growth Rate in Per Capita Consumption (in %) 1999-2000 to 2004-05 to 2004-05 to 1993-941999-20001993-94Extremely poor and poor ST/SC 0.8 1.0 0.9 Muslims 0.3 1.3 0.8 OBC 1.1 Others 1.3 1.0 1.2 All 0.6 1.0 0.8Marginal and vulnerable ST/SC 0.5 0.8 0.6 Muslims 0.4 0.6 0.5 OBC 0.9 Others 1.1 1.0 1.0 All 0.3 0.7 0.5Poor and vulnerable ST/SC 1.1 1.3 1.2 Muslims 0.8 1.3 1.0 OBC 1.4 Others 2.2 1.2 1.8 All 0.8 1.2 1.0Middle and high income ST/SC -0.2 0.4 0.1 Muslims -0.6 2.0 0.5 OBC 1.4 Others 0.9 2.7 1.7 All -0.2 1.8 0.7Total ST/SC 1.1 1.7 1.3 Muslims 0.5 2.2 1.2 OBC 2.5 Others 3.9 4.1 4.0 All 0.7 2.5 1.5
SPECIAL ARTICLEEconomic & Political Weekly EPW march 15, 200859social groups is a fact of life in contemporary India, albeit in varying degrees.A much more powerful factor in this differentiation seems to be that of educational endowments. There is no doubt that no or low education is more strongly associated with poverty and vulnera-bility. But the interesting finding is that for the socially considered lower groups the threshold level of education required to cross pov-erty is higher compared to other social groups. What we find here is the close correspondence be-tween social identity and educa-tional attainments.Over time there has been some change but we have characterised the speed of this change as “snail’s pace”. For the two bottom layers of the social category, i e,SC/ST and Muslims, the change is largely from poverty to vulnerability. Perhaps the speed of change is also determined by a combination of social identity and educational endowments. Despite the formidable con-straints, education seems to provide the best hope for overcoming pov-erty and vulnerability. The speed of change has been faster wher-ever the educational attainments are higher.When we examine the informal work status of the earners in the households, there is a close asso-ciation with social identity and low education. Therefore, informal work status is often an outcome of the inherited and acquired endowments. Here again, education acts as a moderator in that a higher level of education ofinformalworkers results in less pov-ertyandvulnerability.Thisalsoworks in a dynamic way in that thoseinformalworkers with higher education get out of poverty and vulnerability faster than others. Therefore, it is hardly neces-sary here to emphasise the role of education and skill develop-ment of those in the informal sector. There is considerable disparity in the consumption expenditure of the poor and vulnerable with that of the middle and high income group and it has been worsening over the years. What is disturb-ing is the trend that shows a slower rate of growth of consump-tion of the poor and vulnerable with a smaller average per capita consumption compared to the middle and high income group with a much higher average consumption that has been growing much faster. This is a clear manifestation of the increasingdisparity between the haves and have-nots with different consumption baskets. These numbers point to the often talked about “two Indias”, a shining and resurgent one and a suffering yet labouring one.What we have found is a strong undercurrent of what we would like to call systemic and hierarchical segmentation based on social identity. It is on top of this that we find an economic differentiation in terms of the level of consumption that gets thinner and thinner as we go down the social ladder. This systemic and hierarchical segmentation along social groups in terms of poverty and vulner-ability, informal work status, and education is also reflected in per capita consumption in different poverty status groups. What this conveys is that anSC/ST poor is not the same as the poor Muslim or poorOBC not to speak of the poor Other. The depth of poverty is not the same across social groups even within the category of poor. TheSC/ST population forms the bottom layer in all respects considered here. This is followed by Muslims, thenOBCs and the top layer by Others. What is striking is the evidence that this systemic and hierarchical segmentation along social groups is also reflected across economic groups represented by consump-tion expenditure classes, low education and informal work status. Within these economic groups, the only change in ranking is within the intermediate category of OBCs and Muslims in terms of per capita consumption. The latter has a marginally higher per capita consumption than the former among the poor and vulnerable but this edge is absent when all economic groups are combined. Table 23 summarises the systemic and hierarchical ordering.8 Conclusions Our estimate that a little more than three-fourths of the Indian people are poor and vulnerable in 2004-05, based on a value that is double the official poverty line, is consistent with other estimates. For example, the World Development Report 2006 of the World Bank reports 35 per cent of the Indian population as living below the extreme poverty line of one PPP $ per day. There has, however, been some dissatisfaction with the use of the consumer expenditure survey (CES) data obtained from national sample surveys because of its difference with the estimates of private final consumption as reported in the National Accounts Statistics (NAS). The problem has been compounded by the increasing disparity between the two sources. Several scholars have addressed this problem and the consensus has been that theCES data is more reliable than theNAS. While there are some genuine sources of difference such as the consumption by non-profit institutions, serious doubts have been raised regarding the estimation of private final consumption in the NAS.5 There is also the argument that the bias in reporting consumption in theCES is more pronounced in the upper income groups than the poor and vulnerable. This is especially true of consumption outside the household, purchase of expensive luxury goods and celebration of marriages and other events and so on. The question Table 23: Systemic and Hierarchical Segmentation of Social Groups across Economic Groups (Ranking)Social Group Share of Incidence of Incidence of Low Low Proportion of UnorganisedExtremelyPoorand EducationEducationMiddleand Workers in Poor and Vulnerable among Poor among High Income Total Workers Poor and Unorganised among Vulnerable Workers Social GroupsOthers 1 1 1 1 1 1OBC 2 2 2 2 2 2Muslim 3 3 3 3 3 3SC/ST 4 4 4 4 4 4Ranking in descending order. 1 indicates the most favourable position and 4 indicate the least favourable position.Table 22: Percentage Share of Total Monthly Expenditure of Different Poverty Categories across Social GroupsSocial Group Share of Total Monthly Expenditure 1993-941999-20002004-05Extremely poor and poor ST/SC 26.5 22.4 18.9 Muslims 22.3 19.3 15.3 OBC NA 14.3 10.4 Others NA 4.6 2.9 All 15.4 13.0 10.2Marginal and vulnerable ST/SC 51.8 55.2 55.4 Muslims 49.5 53.4 52.4 OBC NA 54.7 51.5 Others NA 36.7 27.7 All 45.9 48.0 44.2Poor and vulnerable ST/SC 78.2 77.6 74.3 Muslims 71.7 72.7 67.7 OBC NA 69.0 61.9 Others NA 41.3 30.6 All 61.3 61.0 54.4Middle and high income ST/SC 21.8 22.4 25.7 Muslims 28.3 27.3 32.3 OBC NA 31.0 38.1 Others NA 58.7 69.4 All 38.7 39.0 45.6Total ST/SC 100 100 100 Muslims 100 100 100 OBC NA 100 100 Others NA 100 100 All 100 100 100
SPECIAL ARTICLEmarch 15, 2008 EPW Economic & Political Weekly60of increasing differences has, however, not yet been settled or anywhere near consensus. This subject, therefore, is likely to remain a challenge for experts in the field. If the reported consumption expenditure by the better off sections has a downward bias, then it may be hypothesised that the perceived difference in living standards between the poor and vulnerable and the higher income groups (middle and high income) as revealed by the consumer expenditure data is an underestima-tion of the actual differences. The problem of inequality would then assume far more seriousness for its implications than what is made out on the basis of consumption expenditure. We would also like to comment on the official poverty line on the basis of which estimates of the incidence of poverty are made. It is highly questionable to rely on an unchanging poverty line adjusted only for price changes, in terms of consumer expenditure since the early 1970s (i e, 1973-74) despite an increase in per capita income and substantial changes in consumption pattern over the last 31 years (i e, as of 2004-05). For example, the official poverty line was Rs 49.63 for rural areas (Rs 56.76 for urban) in 1973-74 in terms of MPCE when the per capitaGDP per month was Rs 88.58 and the rural poverty line was thus 56 per cent (64.1 per cent for urban) of per capitaGDP. In 2004-05, the official poverty line was Rs 356.30 (MPCE) for rural areas (Rs 538.60 for urban) when the per capitaGDP stood at Rs 2,183.42 per month, i e, the poverty line was 16.3 per cent (24.7 per cent for urban). The notion of an absolute minimum of a basket of goods yielding a calorie value plus some essential items loses most of its significance in a growing economy relative to per capita income. Poverty should be reckoned in relative terms to capture the inequalities in the system. There is nothing absolute about an absolute minimum for a poverty line when the economy is on a growth path of an unprecedented kind. That this point has not been factored, not just in India but even in some other countries with much faster rates of growth (e g, China), perhaps reflects an eagerness to show a declining trend in poverty or, for that matter, the magic of “trickle down” growth. There is no doubt that the case for revisiting the poverty line would become stronger as the economy continues to grow.6 Use of Official Poverty LineIn this paper, we have not entered into these discussions as we have used the poverty line only as a benchmark for decomposing the total population into six different groups, in terms of poverty status. We have accepted the official definition of the poverty line as the absolute minimum standard of living at different points of time and divided the Indian population into groups with average per capita consumption of up to 0.75 of the poverty line (PL) 0.75 to 1PL, 1 to 1.25, 1.25 to 2PL, 2 to 4PL and above 4PL. This allowed us to go beyond the absolute minimum expenditure norm to look at the relative status of different groups of our population evolv-ing over time, combining the multidimensional elements of deprivation which poverty, in essence, reflects.The need to adopt indicators of deprivation, other than consumer expenditure, is so compelling in the Indian situation that the case for a multidimensional approach is quite strong. In this analysis we have been able to bring in directly only indica-tors on education, social status and activity status (because these are collected in the employment-unemployment surveys) but these are striking in their association with the incidence of poverty and vulnerability. There are other equally compelling indicators of deprivation such as infant mortality, children under nutrition or anaemia among adult women. Our analysis here starting with consumer expenditure and linking them with social, educational and work status dimensions tell us that there is a strong case for building up a multidimensional profile of poverty, however complex the exercise might be. The idea is not to reduce every-thing into a single index but to provide a framework for using appropriate indices of deprivation for specific policies and programmes rather than a single measure of poverty for all programmes. This will also be a formal acknowledgement that development is multidimensional even in its basic sense of a life with dignity. And this is what India should strive for by pursuing an agenda of inclusive development and not just inclusive growth.Meeting such a challenge has to certainly go beyond mere government interventions in terms of schemes and programmes. The entrenched systemic and hierarchical segmentation cutting across poor and non-poor has to be acknowledged as a social problem. It is on top of that one finds economic differentiation of varying degrees. Therefore both social and economic depriva-tions are facts of life in India, but the social deprivation arising out of social identity (i e, caste/community) seems to be so deeply entrenched that it cannot be brushed aside by addressing only economic deprivation. There is a strong case for addressing social and economic deprivation together. In that sense, the existence of a “creamy layer” of middle and high income group in every social category and the need to identify it is extremely important. Our attempt to quantify these would go a long way in giving shape to a coherent and comprehensive macro social policy. There is also an immediate need to act through an employment policy because the majority of the poor and vulnerable are the working poor without any social protection or meaningful promo-tion. They are either unemployed or employed without any job security or earning capacity above the minimum subsistence. These are the informal or unprotected workers and the way out is mainly through adequate employment and its quality – regularity, decent wages and/or productivity (for the self employed), social security and working conditions. Measures to improve the living conditions of the self-employed need to be multifaceted through such means as security of assets (e g, land), access to credit, skill upgradation, marketing and technology. For the wage labour it has tobe through better skills, working conditions and by ensuring decent wages. Both these groups need social security to protect them from such contingencies as sickness and such eventualities as old age and death. It goes without saying that such interventions can bejustified from both a growth/development as well as equity perspective. To pose them as trade offs is not only short-sighted but could be counterproductive even to achieve the limitedobjec-tive of inclusive growth. It is this perspective that ledusandour colleagues at the National Commission for Enterprises in the Unorganised Sector to undertake a detailed analysis of the condi-tions of work and livelihood promotion issues of the informalor unprotected workers in this country and put forward a package of immediate measures for action as a matter of priority.7
SPECIAL ARTICLEEconomic & Political Weekly EPW march 15, 200861Notes 1 We do not intend to deflect our attention here to the continuing debate on the beginning of the growth acceleration in the Indian economy. While some economists point out early 1990s as the turning point (that coincides with the launch of a major pro-market and pro-private sector reform programme) there are others who point out that early 1980s as the turning point [e g, Rodrik and Subramanian 2004]. Recent econometric work in this area points out to two turning points, the first being in the early seventies [e g, Balakrishnan and Parameswaran 2007] implicitly recognising the importance of economic policies and pro-grammes carried out during the first two and half decades of independent India. While such research exercises are likely to continue, we do notthink that economic policy can be delinked from the larger political context and what politi-cal leaders think about India’s national interest. One of us has written on this subject [e g, Sengupta 2001: Chapters 1, 2 and 6]. A notable and early contribution in this area is that by Frankel (2005). 2 This is similar to the estimates of the Planning Commission by using the mixed reference period (MRP) data. This will increase to 28 per cent if we use the uniform reference period (URP) data. We have chosen the former for the sake of compari-son with 1999-2000 and 1993-94. 3 We follow the definition adopted by the National Commission for Enterprises in the Unorganised Sector (NCEUS). According to the NCEUS, “unorganised workers consist of those working in the unorganised enterprises or households, excluding regular workers with social security benefits, and the workers in the formal sector without any employment/social security benefits – (2006): ‘Social, Economic and Educational Status of the Muslim Community in India: A Report’, Prime Minister’s High Level Committee, Cabinet Secretariat, New Delhi.IHD (2007): ‘International Seminar on Revisiting the Poverty Issue: Measurement, Identification and Eradication, Seminar Papers’ (unpublished), In-stitute of Human Development, New Delhi.Minhas, B S (2005): ‘Validation of Large-Scale Sample Survey Data: Case of NSS Estimates of Household Consumption Expenditure’ in Deaton and Kozel (2005).NCEUS (2006): Report on Social Security for Unorganised Workers, National Commission for Enterprises in the Unorganised Sector, Government of India, New Delhi. Also available in web site: www.nceus.gov.in – (2007): Report on Conditions of Work and Promo-tion of Livelihoods in the Unorganised Sector, Na-tional Commission for Enterprises in the Unor-ganised Sector, Government of India, New Delhi. Also available in web site: www.nceus.gov.inRodrik, D and A Subramanian (2004): ‘From “Hindu Growth” to Productivity Surge: The Mystery of the Indian Growth Transition’, NBER working paper 10376,Sengupta, Arjun K (2001): Reforms, Equity and the IMF(See Chapters 1, 2, and 6), Har-Anand Publi-cations, New Delhi.Sivasubramonian, S (2000):The National Income of India in the Twentieth Century, Oxford University Press, New Delhi.Srinivasan, T N (2007): ‘Poverty Lines in India: Reflec-tions after the Patna Conference’ inEconomic & Po-litical Weekly, Vol 42, No 41, October 13-19.Sundaram, K and S D Tendulkar (2005): ‘NAS-NSS Es-timates of Private Consumption for Poverty Esti-mation: A Further Comparative Examination’ in Deaton and Kozel (2005).provided by the employers”.Fordetails,see NCEUS 2007. 4 The average daily per capita expenditure of the vulnerable group was Rs 20 in 2004-05. The other groups like marginal, poor and extremely poor have average monthly per capita expenditure of less than Rs 20. 5 For an informed discussion of the sources of dif-ferences as well as the limitations of NAS and the higher reliability of CES, see Minhas in Deaton and Kozel (2005). Also see the articles by Sundaram and Tendulkar in Deaton and Kozel (2005). 6 Many countries in the world, especially in the European Union, set poverty line entirely on a relative basis. Also see, for example, the papers presented in the International Seminar on Revisiting Poverty held in Patna during 2007 [IHD 2007]. T N Srinivasan, one of the high priests of the neoliberal economic reforms in India, has recently advocated revising the offi-cial poverty line to two times the current PL. See, Srinivasan (2007). 7 See NCEUS (2006 and 2007).ReferencesBalakrishnan, P and M Parameswaran (2007): ‘Un-derstanding Economic Growth in India: A Prerequisite’ inEconomic & Political Weekly, Vol42, Nos 27 and 28, July 14.Deaton, A and Valerie Kozel (ed) (2005): The Great Indian Poverty Debate, MacMillan India, New Delhi.Frankel, F R (2005): India’s Political Economy, 1947-2004: The Gradual Revolution, Oxford Uni-versity Press.Government of India (2007): ‘Poverty Estimates for 2004-05’, Planning Commission, New Delhi. Appendix: A Methodological Note on the Analysis of Poverty 1 The DataSetsThe estimates of incidence of poverty in India are traditionally being computed from the quinquennial consumer expenditure surveys un-dertaken by the National Sample Survey Organisation (NSSO). The last three surveys in this regard were during 1993-94, 1999-2000 and 2004-05. In these surveys, data on consumption were collected from each of the selected sample households in a detailed schedule containing the list of every conceivable item of consumption, classified into the following broad groups and placed in different blocks,(i) food, pan, tobacco and intoxicants,(ii) fuel and light,(iii) clothing, bedding, etc,(iv) footwear,(v) education and medical (institutional) goods and services,(vi) miscellaneous goods and services including medical (non-institutional), and(vii) durable goods including repair and maintenance.In the consumer expenditure survey conducted during 2004-05, con-sumption data in respect of food, tobacco and intoxicants; fuel and light; and miscellaneous goods and services were collected for a period of 30 days prior to the date of survey. In the case of other item groups like clothing, bedding, footwear, education, medical (institutional) and durable goods, data on consumption were collected both for the last 30 days and for the last 365 days and recorded in separate columns against each individual item. The field staff has been specifically instructed to collect both sets of data independently and by separate probing. Thus a reference period of both 30 days and 365 days were used for the collec-tion of consumption data in respect of the above groups of items.In the 50th round consumer expenditure survey conducted during 1993-94 also the data were collected by using the same reference periods for different groups of items, except that the items relating to medical (non-institutional) were included in the reference periods of 30 days and 365days.In the 55th round consumer expenditure survey conducted during 1999-2000, there were variations in the reference periods used for the collection of consumption data. In the case of food, pan, tobacco and intoxicants reference periods of both seven days and 30 days were used for data collection and the field staff was instructed to collect the data based on 30 days reference period first and thereafter the data based on seven days. In the case of fuel and light; and miscellaneous goods and services including medical (non-institutional), data were collected by using a reference period of 30 days as in the case of 61st round. In the case of all other items, a reference period of 365 days was used for the collection of consumption data. A comparative state-ment indicating the reference periods used for each group of items for the collection of consumption data in the three rounds of survey is given below. In the computation of head count ratios of poverty based on 50th Round Survey, the data collected by using the uniform reference period of 30 days only were used though data were also collected Reference Periods Used for the Collection of Consumption Data in the Last Three Rounds of Consumer Expenditure Surveys(in days)Sl No Item Group 50th Round 55th Round 61st Round1 Food, pan, tobacco and intoxicants 30 7 and 30 302 Fuel and light 30 30 303 Clothing and bedding 30 and 365 365 30 and 3654 Footwear 30 and 365 365 30 and 3655 Education and medical (institutional) 30 and 365 365 30 and 3656 Miscellaneous goods and service (a) medical (non-institutional) 30 and 365 30 30 (b) Others 30 30 307 Durable goods 30 and 365 365 30 and 365
SPECIAL ARTICLEmarch 15, 2008 EPW Economic & Political Weekly62forsomeoftheitem groups by using reference periods of both 30 days and 365 days.In 55th round, data were not collected by using the reference period of 30 days for some of the items and as such a uniform reference period could not be used for the computation of Head Count Ratios. Accordingly, the incidence of poverty was estimated by using the data sets of mixed reference periods.The head count ratios based on the 61st round consumer expendi-ture survey were estimated by using both the uniform and mixed reference periods. Thus any comparison of head count ratios across the three rounds can ony be based on mixed reference period, as the ratios based on uniform reference periods cannot be computed for 1999-2000.Though the official estimates of poverty ratios for 1993-94 were based on data sets of uniform reference period of 30 days, Abhijit Sen, has estimated the poverty ratios based on mixed reference period data for comparison purposes.The educational and employment status of individual members of the households are not collected in the consumer expenditure surveys. It is, therefore, not possible to analyse the nexus between the level of poverty and socio-economic variables like educational attainment, employment status, etc, based on data sets available from consumer expenditure surveys. Data on consumer expenditure were, however, collected in the NSS 55th and 61st round employment-unemployment surveys in an abridged block by using mixed reference period. In the case of NSS 50th round, the sample households for consumer expenditure and employment-unemployment surveys were the same. The average monthly per capita consumption expenditure available from the consumer expenditure schedule was thus transferred to the employment-unemployment survey schedule of each household. These data items contained in the employment-unemployment sched-ules were, therefore, used for cross tabulation and analysis of linkages between poverty and various socio-economic characteristics.2 Adjustment of Poverty LinesThe head count ratios of poverty are estimated by the Planning Commis-sion by using the distribution of the households by monthly per capita consumption expenditure classes as available from quinquennial con-sumer expenditure surveys and poverty lines obtained by updating those defined by an Expert Group in 1993 separately for rural and urban areas in major states. These poverty lines are based on consumer expenditure surveys in which data on consumption were collected for each individual item in each group. In the case of employment-unem-ployment surveys, data on consumption are collected by using an abridged block in which only some major items and item groups are listed. Thus, the average consumption expenditure as computed from employment-unemployment survey schedule is expected to be lower than that obtained from consumption expenditure schedule. The poverty lines are, therefore, required to be adjusted to make them conform to the level of consumption expenditure as revealed by the employment-unemployment survey.The consumption expenditure block in both the 55th and 61st round of employment-unemployment surveys used a reference period of 30 days for collecting consumption data on food, pan, to-bacco and intoxicants; fuel and light and miscellaneous goods and services. For all other items, a reference period of 365 days was used for data collection. Therefore, the poverty ratios from employment-unemployment survey can be estimated by using only the mixed reference period.In view of the above, head count ratios as estimated from the 55th and 61st round consumer expenditure surveys by using the mixed refer-ence period were assumed to be the true indicators of poverty. These head-count ratios as estimated from the consumer expenditure surveys were super-imposed to the distribution of persons by per capita monthly consumption expenditure (MPCE) classes as obtained from employment-unemployment surveys and interpolated the corresponding MPCE level. These MPCE levels were taken as the poverty lines conforming to the consumption expenditure pattern revealed by the employment-unemployment survey.In the case of 50th round, the head count ratios as estimated by Abhijit Sen by using mixed reference period were used instead of the official es-timates based on uniform reference period. However, the unit level data sets contained theMPCE estimates based on uniform reference period. Therefore, in this case also the poverty lines were adjusted by using the same methodology.The adjusted poverty lines were thus worked out both for rural and urban areas at the all India level and at the level of each state/UTs based on the respective poverty ratios. The analysis at the national level is done by using the all India poverty lines and the analysis at the state level is done by using state level poverty lines.3 Classification of Households and Persons by Poverty StatusThe members of each household in the sample were classified as extremely poor, poor, marginal, vulnerable, middle class and higher income group based onMPCE of the corresponding households. The specific criteria used for classification both in rural and urban areas are the following.Extremely Poor ifMPCE<=0.75 times Poverty Line (PL)Poor if0.75PL<MPCE<= 1PLMarginal if1PL<MPCE<=1.25PLVulnerable if1.25PL<MPCE<=2.0PLMiddle Class if 2.0 PL<MPCE<=4.0PLHigher Income Group ifMPCE>4.0PLThe limits worked out for each of the years 1993-94, 1999-2000 and 2004-05 in terms of MPCE for classification are given in Tables 1 to 3.By comparing the average MPCE of each sample household with the MPCE limits mentioned above, every surveyed household and the individuals in them were clas-sified into one of the above poverty statuses.4 Estimation and AnalysisThe unit level data sets provided by the NSSO con-tained household level multipliers for obtaining popu-lation estimates. These multipliers were worked out on the basis of the sampling design used for the selec-tion of households for the survey. The population esti-mates obtained by using these multi-pliers were, how-ever, lower than the census population estimates/projections due to various reasons. The given multipliers, therefore, needed further ad-justments to correspond to census population. The ratios of the pro-jectedcensus population to the design estimate in each of the population segments of the rural male, rural female, urban male and urban female Table 1: Monthly Per Capita Consumption Expenditure Limits (1993-94)Poverty Status MPCE Limit in Rupees RuralUrbanExtremely poor MPCE<=145.28 MPCE<=201.47Poor 145.28<MPCE<=193.70 201.47<MPCE<=268.63Marginal 193.70<MPCE<=242.13268.63<MPCE<=335.79Vulnerable 242.13<MPCE<=387.40 335.79<MPCE<=537.26Middle class 387.40<MPCE<=774.80 537.26<MPCE<=1074.52Higher income group MPCE>774.80 MPCE>1074.52Table 2: Monthly Per Capita Consumption Expenditure Limits (1999-2000)Poverty Status MPCE Limit in Rupees RuralUrbanExtremely poor MPCE<=228 MPCE<=314Poor 228<MPCE<=304314<MPCE<=419Marginal 304<MPCE<=380419<MPCE<=524Vulnerable 380<MPCE<=608524<MPCE<=838Middle class 608<MPCE<=1216 838<MPCE<=1676Higher income group MPCE>1216 MPCE>1676Table 3: Monthly Per Capita Consumption Expenditure Limits (2004-05)Poverty Status MPCE Limit in Rupees RuralUrbanExtremely poor MPCE<=259.65 MPCE<=385.50Poor 259.65<MPCE<=346.20 385.50<MPCE<=514.00Marginal 346.20<MPCE<=432.75 514.00<MPCE<=642.50Vulnerable 432.75<MPCE<=692.40 642.50<MPCE<=1028.00Middle class 692.40<MPCE<=1384.80 1028.00<MPCE<=2056.00Higher income group MPCE>1384.80 MPCE>2056.00

15 April 2008

Dear reader,

To continue reading, become a subscriber.

Explore our attractive subscription offers.

Click here

Comments

(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top