1 Background

I

september 17, 2011 vol xlvi no 38

down process thereby taking care of the poor and the problem of poverty.

2 Rationale for a Regional Analysis

For such a large and continental-size country as India, the rationale for an analysis of the profile of poverty in the states is selfevident. Apart from the federal character of the polity that confers some autonomy to the states in policymaking, there are sound differences in terms of the structure of the regional economies, development in infrastructure, industrialisation, urbanisation and, most importantly, human development. In addition, there is the question of governance capacity. Keeping all these in mind and with a view to “learning from each other”, we attempt an analysis and assessment of the regional profile of poverty and vulnerability in India.

As in the case of our earlier exercise, we take the official poverty line1 (PL) as the benchmark for constructing the poverty profile. They are grouped as the “extremely poor” (those at not more than 0.75 of the PL, “poor” (equal to 0.75 to 1.0 PL), “marginally poor” (1 to 1.25 PL), “vulnerable” (1.25 to 2 PL), “middle income” (2 to 4 PL) and “high income” (4 PL and above). The PL used here related to the mixed reference period covered in 2004-05. The distribution of the population in each major state by the above classification is given in Table 1.

In order to give some meaning to the comparative performance of states, we have classified the states as top, middle and bottom level performers. Such a classification basically takes into account the distance between the top performer and the b ottom performer. The distance between the two is divided by three. The first one-third is then added to the top performer (lowest percentage of poor in this case) to obtain the top level and another one-third of this is added to get the middle level and the

Table 1: Percentage Share of Population in Different Poverty Status Groups in 2004-05

Extremely Poor Poor Marginal Vulnerable Middle Income High Income

Source: Computed from unit level data of NSS 61st round. Unless otherwise stated, the findings presented in all tables are based on computations from this data set. Each row adds up to 100.

Economic & Political Weekly

EPW

remaining is grouped in the bottom level. The simple idea here is the ranking of states according to their distance from the top performer. The finding of such an exercise is given in Table 2 for those below the official poverty line (extremely poor and poor) at two time points, viz, 1993-94 and 2004-05, the time period b etween which approximates to the first decade of economic r eform in the country. By working out the annual percentage point reduction in the incidence of poverty, we are also in a position to get an idea of the speed with which the incidence of officially recognised poverty is getting reduced in different states.

Table 2: Distribution of Major States by Incidence of Poverty

There are a number of significant points emerging from this scenario. First of all, in 1993-94, the incidence of poverty in the bottom performer (Bihar) was 5.2 times as much as in the top performer (Punjab); this jumped to 9.3 times in 2004-05 with Jammu and Kashmir as the top performer and Orissa as the bottom performer. Second, the incidence of extremely poor and poor declined both at the all-India level and in all states. While the all-India decline shows an annual reduction of 0.8 percentage points, there are wide variations across the states. The top performers in this case were the north-eastern (NE) states with Assam at the top (1.62 percentage points per annum) followed by Bihar, Tamil Nadu, Jammu and Kashmir, Himachal Pradesh and West Bengal. The worst performers were Rajasthan, Madhya Pradesh, Orissa and Maharashtra. The slow reduction in states with already low incidence of the extremely poor and poor such as Punjab and Gujarat are understandable although there is scope for a faster reduction.

A faster reduction in the incidence of poverty has enabled the NE states including Assam to move to the top level, but Bihar continued to be at the bottom level due to high incidence at the initial period. What is of concern is the slow reduction in states with high incidence of poverty such as Madhya Pradesh, Orissa and Maharashtra. Slow reduction has adversely affected the position of some states such as Rajasthan (which slipped from the top to the middle level) and Madhya Pradesh (from middle to bottom level). The states that distinguished themselves with a faster (than the national average by some margin) reduction who were already at the top level were Jammu and Kashmir, Himachal Pradesh and Kerala.

As we had pointed out in our earlier paper, the problem of poverty in India is not just a matter of crossing a “line”, given the fact that, a substantial segment of the population cluster around the poverty line and hence the categories of “marginally poor” and “vulnerable” become important. By combining the categories of extremely poor, poor, marginally poor and vulnerable, we get the category of poor and vulnerable. The upper limit of this category (equal to two times the official poverty line) is only marginally above the international poverty line of $2 in PPP terms. We estimated that 76.7% of the population in India belonged to this poor and vulnerable category in 2004-05. What is the regional profile of this group of poor and vulnerable? Has there been a significant reduction of this segment in some states although the all-India picture is a mere 5.1 percentage points between 1993-94 and 2004-05?

The regional picture, presented in Table 3, suggest a much less rosy one than in Table 2 where we only examined those below the official poverty line. Here the rate of reduction between 1993-94 and 2004-05 was not only slower in many states with large populations, but was also characterised with greater variation across states.

Table 3: Distribution of Major States by Percentage of Poor and Vulnerable (P and V)

1993-94 2004-05 Percentage State P and V Ratio State P and V Ratio Change

Punjab and Kerala emerged as the top level performers followed by Himachal Pradesh, NE states (including Assam) and Gujarat. Despite this, only Kerala and Himachal Pradesh climbed to the top level from the earlier middle level with Punjab and Haryana retaining their top positions. The NE states excluding Assam have, in fact, moved to the middle level from the earlier bottom level.

On the other hand, the disappointing performances relate to Rajasthan and Madhya Pradesh where the incidence of the poor and vulnerable category in fact increased. The other states where the rate of reduction was so small that it hardly made any difference were Uttar Pradesh, Bihar and Orissa.

When comparing the performance of the officially classified poor, the picture that we get from the point of the poor and vulnerable is more disconcerting. While only two states (out of 17) were at the bottom level in terms of the official incidence of poverty in 1993-94 this becomes seven in terms of poor and vulnerable. In 2004-05, the comparative figures were six and nine, respectively.

The lesson we draw is that in most states of India, escaping from the official poverty line has meant an extremely tenuous process with the poor finding themselves in the category of marginally poor; for those in the marginally poor category, the movement has meant a similar escape from the absolute poverty, but not out of vulnerability.

In an otherwise rather dismal scenario, there are some silver linings. The performance of Kerala and Himachal Pradesh has demonstrated that it is possible to achieve significant reductions not only in absolute poverty, but also in poverty and vulnerability. The case of Punjab, despite its lacklustre performance in growth, was even more remarkable in that it is the only state in the country where the majority of the population are out of the category of poor and vulnerable as we have defined here. But it is not inconceivable that the other three top performers, viz, Haryana, Kerala and Himachal Pradesh might well join Punjab within the next few years or so if this rate of reduction continues for the d ecade starting from 2004-05.

Yet another lesson that needs to be highlighted relates to the context of the neo-liberal dictum of growth translating to poverty reduction, especially in a context of high urbanisation and industrialisation. The top level states are all those who do not belong to the group of a similar top level in either urbanisation or industrialisation of the kind pursued so far. As a matter of fact, states with high urbanisation and modern industrialisation seem to show a rather poor record in poverty reduction. The case of Maharashtra seems so glaring that despite being one of the high per capita income states, with high urbanisation and modern industrialisation, its record in reducing either the officially defined poverty or poverty and vulnerability is quite disappointing. The same is the case with Tamil Nadu although its record in reducing the incidence of poverty (as officially defined) is quite impressive. The fact that it has an incidence of poverty and vulnerability at around 73% is yet another pointer to the clustering of the poor within the band of poor and vulnerable. West Bengal is another case with a similar record. This lacklustre performance of West Bengal in reducing poverty and vulnerability despite

september 17, 2011 vol xlvi no 38

a supposedly pro-poor political alliance in power for the last 28 years (as of 2005) is quite disappointing, to say the least. Gujarat, another fast industrialising state, has a somewhat better record; yet it is not able to match the record of the top four states. In fact, the record of Gujarat is lower than that of Andhra Pradesh.

3 Poverty and Vulnerability across Social Groups

In a country like India, poverty is often measured and portrayed in regional terms; but it is something that has a strong social group orientation as we found in our earlier all-India exercise. Further enquiry into selected indicators of human deprivation such as the availability of a private toilet facility, housing condition and malnutrition among women also revealed the dominance of the social divide over the regional divide (Kannan 2009). This dimension then merits an examination at the regional level to see whether the regional variation has some lessons to offer. We do

Table 4: State-wise Incidence of Poverty and Vulnerability by Social Group (2004-05)

Total Population SC/STs Muslims OBCs Others

India [76.7] India [87.8] India[84.5] India[79.9] India [54.8]

The acronyms used here refer to the following states: AP = Andhra Pradesh, As = Assam, BH = Bihar, CH = Chhattisgarh, GJ = Gujarat, HP = Himachal Pradesh, HR = Haryana, JH = Jharkhand, JK = Jammu and Kashmir, KE = Kerala, KR = Karnataka, MH = Maharashtra, MP = Madhya Pradesh, NEA = north-eastern states excluding Assam, OR = Orissa, PJ = Punjab, RJ = Rajasthan, TN = Tamil Nadu, UK = Uttarakhand, UP = Uttar Pradesh, WB = West Bengal.

Economic & Political Weekly

EPW

by State for All Population

Top level Middle level Bottom level

this in Table 4 where we retain the three-level classification with reference to total population and classify the position of the four broad social groups. The findings are quite striking. Those belonging to the SC/ST group do not find a place at the top level for any state. Except for five states out of the 21, all states belong to the bottom level emphasising the very high incidence of poverty and vulnerability among this social group across the country. Even the relatively better performing states show a very high incidence as in the case of Jammu and Kashmir (67.4%), the NE states excluding Assam (68.9%) and Punjab (71.3%).

The next group from this bottom is that of the Muslims. And here again there are no states at the top level again emphasising the high incidence of poverty and vulnerability, although it is somewhat better than the SC/ST group. Here the best performing state is Punjab, but given the very small share of the Muslim population in this state (less than 1%), the credit should go to Kerala (68.8%) with a population share of around 23%. Note that Jammu and Kashmir occupies only the 8th position in this group ranking. As in the case of the SC/ST group, only five out of 21 states are in the middle level, all the rest come under the bottom level states.

The scenario gets modified to some extent with respect to the OBC group. Four states are at the top level; in fact, one more than that for all population. Seven states are in the middle level and the remaining 10 are at the bottom level. Bihar is at the bottom of all the states for SCs/STs, Muslims and OBCs. The best performance is in Punjab with an incidence of 50.7% which is very close to that of the total population.

When we examine the group Others, the picture that we get is opposite that of the SCs/STs. At the top level, we find 14 states followed by five at the middle level. Only in the two states of Orissa and Uttarakhand, we find them at the bottom level. Although the range of variation is the highest in this group, the incidence is consistently lower than all other social groups in all the states. The best performing state is Tamil Nadu with an

Figure 3: Incidence of Poverty and Vulnerability by Level of Performance

Figure 4: Incidence of Poverty and Vulnerability by Level of Performance by State for Muslims

Bottom level Middle level

incidence of just 23.3% followed by Punjab (26.5%), Gujarat (37.8%), Andhra Pradesh (39.4%), Haryana (40.5%) and Kerala (45.6%). Thus, in these six states, a majority of the social group Others are outside the net of poor and vulnerable as we defined here. This is in sharp contrast to just one state – Punjab – when all the population is considered.

These findings come out sharply when we present them visually in the map of India with state boundaries according to the level of performance. The contrast between Figure 1 (p 63) (where all population groups are considered) and Figures 2 to 5 are quite striking; but more striking is between Figure 2, where the social group consisting of upper caste Hindus, Christians and Sikhs are considered and that of Figures 4 and 5 where Muslims and SC/ST groups, respectively are considered.

Figure 5: Incidence of Poverty and Vulnerability by Level of Performance by State for

SCs and STs Middle level Bottom level

4 Extent of Social Inequality in Poverty and Vulnerability

We now move to examining another social dimension in the incidence of poverty and vulnerability. This is measured by the gap between the incidence in poverty and vulnerability among the most advantaged social groups represented by Others and the least advantaged represented by SCs and STs. Table 5 (p 65) shows the percentage point differences in the incidence between these two social groups. The results are revealing in more than one sense. First of all, the top level states show a relatively high social distance or inequality with Haryana and Punjab occupying the third and fourth ranks. However, the highest inequality is in Tamil Nadu with more than 61 percentage point gap thereby climbing to the top slot in inequality. Gujarat with a nearly 47 percentage point gap occupies the second position. A number of inter pretations

september 17, 2011 vol xlvi no 38

Incidence of P & V among SC/STs and Others

State Social Inequality Gap Rank in consumption. Top levelThe crucial lesson is that

5 Poverty and Vulnerability and Educational Incapability

Punjab 44.5 4

Haryana 46.8 3

Kerala 33.1 7

Middle level

Maharashtra 25.1 14

West Bengal 26.7 13

Bottom level

Orissa 18.1 18

Bihar 20.8 16

has anything to do with low social ine-

EPW

was able to maintain the educational level of SC/ST population on the top among all the states, this group finds itself at the bottom level in terms of incidence of poverty and vulnerability, suggesting factors beyond mere educational incapability are at work in keeping them in their state of poverty despite an impressive performance for all the population.

The states which were able to keep the SC/ST population at the middle level in terms of the percentage of poor and vulnerable, largely maintained their level of education also at the middle level except Punjab and Andhra Pradesh. Among the bottom level states in terms of the percentage of poor and vulnerable SC/ST population, Kerala maintained its supremacy in education at the top level, while Tamil Nadu, Gujarat, Assam, Maharashtra and Uttarakhand could retain the level of education at the middle level. All the remaining 10 states were at the bottom level both in terms of poverty and education with regard to the SC/ST population.

In the case of the Muslim population the middle level states in terms of poverty were divided across all the levels in terms of education. While Kerala and Chhattisgarh maintained the top level, Punjab and Andhra Pradesh were placed at the bottom level. Tamil Nadu remained at the middle level both in terms of poverty and education. Among the bottom level states in terms of poverty, Maharashtra elevated its position to top level; Himachal Pradesh, Gujarat, Jammu and Kashmir, Orissa, Karnataka, NE states excluding Assam and Madhya Pradesh became middle level states in terms of educational level of Muslims. The remaining states were at the bottom level both in terms of level of poverty and level of education.

In the case of the OBC population, the top level states in the incidence of poor and vulnerable were divided equally between the top and middle levels in terms of the level of education. The middle level states in terms of percentage of poor and vulnerable almost retained their position even in the level of education except Andhra Pradesh which was pulled down to the bottom level. The bottom level states in terms of percentage of poor and vulnerable among OBC population were, however, not at the bottom level in terms of education in all cases. Only Rajasthan, Madhya Pradesh, Chhattisgarh and Bihar remained at the bottom level both in terms of poverty and level of education.

In the case of other communities, as many as 11 states were classified as top level states both in terms of level of poverty and level of education. In other words, the socially advanced groups of people enjoyed a higher status both in terms of level of poverty and education. Only the states of Orissa and Uttarakhand were at the bottom level in terms of percentage of poor and vulnerable and among them Uttarakhand was at the top level and Orissa was at the middle level in terms of education.

While it may not be axiomatic that poverty and vulnerability go along with low levels of educational capability, it would indeed be logical to presume such a relationship in the context of a poor developing country like India. Education enhances capabilities that are valued in seeking employment and enhancing remuneration from such employment because of its ability to contribute to productivity and bargaining. Similarly, educational capability for the self-employed enhances the possibility of more economical use and allocation of resources to maximise output and income. It also helps develop an efficient organisation of production. Only in a context of widespread rent-seeking arising out of inherited assets such as absentee landowners, educational capability would indeed be less important in overcoming poverty and associated vulnerability.

6 Educational Incapability and Poverty

The data from the employment-unemployment survey of the NSS enable us to probe into the educational capabilities of individuals and households. We have therefore examined the incidence poverty and vulnerability among those with low levels of educational capability. This has been done by identifying adults (15 years and above) with education of not more than the primary level (that is five years of schooling) that would also include illiterates and then isolating the poor and vulnerable among them. The results are indeed revealing, to say the least, suggesting a close association between educational incapability and poverty and vulnerability. However, given the influence of a number of other factors, the association is not so straightforward and it gets revealed in the regional profile.

There are several dimensions of the association between poverty and vulnerability on the one hand, and the low level of educational capability on the other. Before we examine and comment further on this theme, a few initial observations on the enormity of this educational backwardness across states may be made.

First, a majority of those aged 15 and above but educated only up to the primary level are poor and vulnerable. But such an outcome is a combination of the differential disadvantage experienced by the constituent social groups. The most disadvantaged are those belonging to the SCs and STs followed by Muslims.

Table 7: Educational Incapability and Poverty and Vulnerability: State-wise Incidence of Poor and Vulnerable among Those Aged 15 Years and Above and Educated Only up to the Primary Level in Each Social Group (%)

EPW

Second, a state-wise profile brings out the inter-regional variation quite sharply. The most disadvantaged are the SC/ST poor and vulnerable in Bihar (97.2%), followed by Uttarakhand (96.6%), Madhya Pradesh (95.8%) and Jharkhand (95.8%). These segments are followed by Muslims in Uttarakhand (95.9%), Jharkhand (95.7 %) and Madhya Pradesh (95.4%).

Third, if we take these two relatively more disadvantaged social groups, it is equally important to highlight the better performing states. The best performing segment is the Muslims in Punjab (62%) followed by Muslims in Himachal Pradesh (72.2%) and in Kerala (73.9%). Since the share of Muslims in the former two states is so small (less than 1%), only the performance of Kerala would really stand out.

7 An Econometric Exercise

In order to study the association between the social groups, educational attainments and place of residence, we undertook a logit analysis by defining a group of binary variables. The variables thus defined are the following: d_pov = 1 if the person belongs to the category of poor

and vulnerable, = 0 if otherwise. d_sector = 1 if the place of residence is rural, = 0 if otherwise. d_edu = 1 if the level of education of the person is primary or below, = 0 if otherwise. d_sc/ st = 1 if the person belongs to the social group SC/ST, = 0 if otherwise. d_muslim = 1 if the person belongs to the social group Muslim, = 0 if otherwise. d_obc = 1 if the person belongs to the social group OBC, = 0 if otherwise. d_others = 1 if the person does not belong to any of the above social groups, = 0 if otherwise. d_sc/edu = 1 if the person is both SC/ST and education up to primary, = 0 if otherwise. d_muslim/edu = 1 if the person is both Muslim and education up to primary, = 0 if otherwise. d_obc/edu = 1 if the person is both OBC and education up to primary, = 0 if otherwise. d_others/edu = 1 if the person is both Others and education up to primary, = 0 if otherwise.

By using the above binary variables, a logit analysis was undertaken by taking d_pov as the dependant variable. The independent variables were d_sector, d_edu and one of the social group variables and its interaction variable with education. Thus, there were four sets of regression equations for each state. The state-wise regression coefficients and odds ratios are given in Tables 1 to 4 (pages 71, 72) in the Appendix. All the coefficients

Economic & Political Weekly

EPW

were found to be significant except the constant in the case of Tamil Nadu. The odds ratios reveal that probability of being poor and vulnerable is significantly high if the person is a resident of rural areas in some of the states though it is not so in some other states. The probability ratio is as high as seven in the case of Jharkhand and 6.72 in the case of Assam followed by Other NE states (5.23), Himachal Pradesh (3.32), West Bengal (3.19), Bihar (2.86), Orissa (2.58), Uttarakhand (2.24), Chhattisgarh (1.66), Uttar Pradesh (1.58), Maharashtra (1.56), Gujarat (1.41) and Tamil Nadu (1.41). Punjab, Haryana, Madhya Pradesh, Karnataka and Rajasthan have odds ratios near about 1. However, Kerala, Jammu and Kashmir and Andhra Pradesh have odds ratios much lower than 1 indicating that the probability of being poor and vulnerable is lower in rural areas than in urban areas in respect of these states. The situation remains almost invariant in all the four groups of regressions factoring different social groups.

The odds ratio of being poor and vulnerable due to low education is as high as 4.68 in West Bengal when the social group SC/ST is factored in the model. In other words, the probability of being poor and vulnerable is 4.68 times greater for those belonging to SC/ST in West Bengal, with only education up to primary level as compared to other social groups. The states with odds ratios higher than four in the first set of regressions include West Bengal, Chhattisgarh, Jharkhand and Bihar. Kerala has the lowest odds ratio of 1.93. There are some variations in the odds ratios while factoring different social groups. In the case of regressions factoring OBCs, the odds ratio of education is as high as 6.22 in the case of Madhya Pradesh, 5.73 in the case of Orissa and 5.28 in the case of West Bengal. Kerala has the least odds ratio in all the social groups.

In the case of those belonging to SC/ST social group, the odds ratios of being poor and vulnerable varies between 4.89 in the case of Haryana to 0.4 in the case of other NE states. The odds ratio is more than 3 in the case of Haryana, Madhya Pradesh, Punjab, other NE states, Rajasthan, Uttar Pradesh and Gujarat apart from Haryana. It is less than 1 in the case of Jammu and Kashmir (0.92) and other NE states (0.40). It is between 1.55 and

2.95 in respect of other states. The odds ratio of interaction variable SC/ST social group and low education is greater than 1 in the case of 11 states and it varies between 1.6 in the case of Uttarakhand and 0.68 in the case of Jharkhand. It is important here to report that the logit exercise supports the assumption that social group and the low level of education have significant effects on poverty and vulnerability individually rather than as an interaction variable.

The odds ratios of Muslims vary between 4.66 in Himachal Pradesh and 1.05 in Chhattisgarh. Since Himachal Pradesh has only a very tiny proportion of population in the Muslim group, the higher odds ratio of 3.49 in Jammu and Kashmir should be counted as the significant result. This indeed is puzzling given the fact that Muslims form a majority of the population in this state. The odds ratio is more than three in the case of Haryana (3.54), Jammu and Kashmir (3.49), other NE states (3.48) and Punjab (3.34); in all these states also the share of Muslim population is too small to accord any significance to these results. The odds ratio of interaction variable Muslim and up to primary level education varies from 1.52 in the case of Haryana to 0.32 in the case of Himachal Pradesh. The ratio is more than one in the case of only six states.

In the case of OBCs, the odds ratio of being poor and vulnerable varies between 3.34 in the case of other NE states to 0.71 in the case of Jammu and Kashmir. In fact, the ratio is more than two only in the case of other NE states. The odds ratio of interaction variable with education is greater than one only in the case of Uttarakhand.

In the case of persons belonging to Others category, the odds ratio of being poor and vulnerable is less than one in respect of all the states. In other words, if a person belongs to the social group of upper castes or Christian or Sikh, the probability of being poor and vulnerable is lower than that of a person belonging to SC/ST, Muslim and OBC social groups. However, the interaction variable others with primary education has odds ratios greater than one in the case of nine states. Jammu and Kashmir has the highest odds ratio of 1.43 followed by Chhattisgarh 1.23.

8 Informal Work Status and Poverty Vulnerability

We had reported earlier (Sengupta et al 2008) that 92.3% of the Indian workforce can be classified as informal workers (86% in the informal sector and 6.3% in the formal sector). When we take the population belonging to our classification of poor and vulnerable, the incidence of informal workers goes up to 96.2%. This varies from Bihar with 98.6% to Kerala with 88.9%. However, there is a greater variation in the incidence of poor and vulnerable in the total population with Punjab at the top with only 47.5% to Bihar with 92.5%. What this implies is that in many states a segment of informal workers do not belong to the category of poor and vulnerable. These are people engaged in selfemployment with some capital (such as independent professionals, shopkeepers, owners of small restaurants and owners of workshops for repairs and manufacturing). In addition, there could be a segment of manual workers with or without regular employment enjoying relatively high wages as in Kerala, Punjab and Haryana. What Table 8 presents is the incidence of poor and vulnerable among the informal workers. It varies from 82.1% for the other social group in Kerala to 99% for OBCs in Bihar. In sum, just as the incidence of informal work status is quite high among the working population, so is the incidence of poor and vulnerable among this informal worker group.

9 Average Daily Per Capita Consumption Expenditure

We now examine another dimension of the consumption profile with particular reference to the various poverty status groups in relation to the non-poor groups categorised as middle income and high income.2

While Table 5 presented the social inequality measured by the gap between the incidence in poverty and vulnerability between the SC/STs and the Others, in Table 9, we present the results of another dimension of inequality measured as between the extremely poor and the high income group, the two extreme groups in our classification. This, of course, ignores the social dimension, but draws our attention to the inequality in consumption.

The highest inequality is in Chhattisgarh, while the least inequality is in Jammu and Kashmir. Among those with relatively high inequality, say, a ratio of more than 10, we find the top level states of Kerala and Haryana as well as the middle level and bottom level states in terms of the incidence of poverty and vulnerability. Such a scenario points out to the fact that high social inequality in consumption is present, irrespective of the performance of states in terms of reducing poverty and vulnerability. The worst case scenario relates to the poorly performing states such as Chhattisgarh, Orissa and Uttar Pradesh. Here the situation is

Table 8: State-wise Incidence of Poor and Vulnerable Informal Workers by Social Group (%)

State STs / SCs Muslims OBCs Others

september 17, 2011 vol xlvi no 38

EPW

and nuanced to take care of the varying regional contexts. That the social divide is a well-entrenched one not only in terms of consumption expenditure, but also a combination of measures to constitute a multidimensional poverty index has been brought out recently and published in the Human Development Report

Figure 6: Incidence of Multidimensional Poverty in India (%)

100

75

50

25

0 Scheduled Scheduled Other Backward All-India Upper caste

tribes castes Classes Population Hindus

2010. This multidimensional poverty measure incorporates a number of indicators for standard of living, education and health.3

Although the exercise is confined to the social groups among the Hindu population (constituting around 83% of the total) and for all population in India, the results are quite compatible with our findings based on consumption poverty. We depict the results of this multidimensional poverty in Figure 6 and they speak for themselves. The incidence of multidimensional poverty is not only the highest among the STs, but includes an overwhelming share of their population (81%) followed by the SCs (66%), OBCs (58%) and the upper caste Hindus (33%) who have the least incidence. The incidence among the STs is 250% higher than the upper caste Hindus, while it is 200% higher or twice for the SCs.4

Another important exercise in the measurement of multidimensional deprivation has recently been provided by Jayaraj

81 66 58 55 33

Percentage of Population

and Subramanian (2010). Based on a theoretical formulation for sensitising both the identification and the aggregation problems to the range of deprivation, the paper reports a gradation (or range) of multidimensional deprivation for major states in India.5 We have found that this gradation in multidimensional deprivation and our gradation based on consumption poverty have a high degree of correlation, especially between poverty groups at the lower levels. Moreover, the paper reports that the reductions that have taken place between 1992-93 and 2005-06 are largely in the nature of a movement from being more d eprived to less deprived that are similar to our finding of a movement from being more poor to less poor between 1993-94 and 2004-05. However, the paper did not address the social dimension of deprivation.

On the basis of our earlier exercise at the all-India level and this one at the regional level, we are of the opinion that India’s unresolved poverty question is closely related to its unresolved social question.

It is our position that the findings in our earlier paper as well as in this one have a direct relevance to the ongoing debate on poverty and the initiative to have a national legislation to provide a measure of food security to the poor. Given the range of poverty and other forms of deprivation, the proposal to cover up to three quarters of the population has considerable merit. The point here is to factor the state-wise variation with due emphasis on the social dimension. The decision by the Government of India that a survey to identify poor households along with their social identity is a welcome development and a recognition of India’s social reality. The challenge is to ensure a fair coverage and rightful inclusion and to bring down the size of this universe by a process of progressive realisation of basic socio-economic security. And we have no doubt that this should be treated as one of the primary obligations of the Indian state representing a democratic polity.

Notes that does not have access to “clean” fuels like GoI (2010): Report of the Expert Group to Review the kerosene, liquid petroleum gas, bio-gas, or electri-Methodology of Poverty, Planning Commission,

household that does not have access to a “pucca” Jayaraj, D and S Subramanian (2010): “A Chakraartyhouse; (5) a household that does not have access D’Ambrosio View of Multidimensional Deprivawith a higher expenditure threshold that was to any description of toilet (including a pit latrine); tion: Some Estimates for India” in Economic & higher by around only 22% of the now old poverty (6) a household with members six or more years Political Weekly, Vol XLV, No 6, 6 February.

line. It is still lower than the international extreme poverty line of one PPP dollar per capita per day. Kannan, K P (2009): “Dualism, Informality and Social

old and is illiterate; (7) a household that does not have access to even a bicycle for meeting the re-

For a discussion on this poverty line in the context Inequality in India: An Informal Economy Per

quirements of mobility; and (8) a household that

of the need for identifying the poor and related spective of the Challenge of Inclusive Develop

does not have access to even a radio as a source of

issues see, Indian Journal of Human Development, ment in India” in The Indian Journal of Labour Volume 4, No 1, January-June 2010. Economics, Vol 52, No 1, January-March.

entertainment. In our view, some of the selected indicators may not necessarily fit in with a meas

2 The absolute figures (in rupees) of the DPCE for Sengupta, Arjun, K P Kannan and G Raveendran

urement of basic deprivation and some such as

the different poverty status groups are given in (2008): “India’s Common People: Who Are They,

health, as the authors admit, fall short of adequate

Table A5 (p 73) in the Appendix. How Many Are They and How Do They Live?” in

representation. Yet the rank correlation among 20

Economic & Political Weekly, 15 March.

3 In terms of standard of living it takes into account

major states between our extremely poor and

access to/ownership of electricity, drinking water,

poor group with that of the severely deprived of

sanitation, quality of housing, cooking fuel and

Jayaraj and Subramanian (JS) (2010) showed a

certain assets. In terms of education it takes into

coefficient of 0.85 and that between our margin

account the average years of schooling and child

ally poor and considerably deprived of JS showed

enrolment. For health it takes into account the

0.81. Combining the severely, considerably and

available at

outcomes in terms of child mortality and nutrition.

moderately poor of JS and correlating it with the

For details see, Alkire and Santos 2010.

larger group of poor and vulnerable in our paper 4 Results for major states in India are also reported showed a rank correlation coefficient of 0.82. Life Book House by Alkire and Santos (2010), but it does not incor-

Shop No 7, Masjid Betul

porate the social dimension in terms of the social

groups. Mukarram Subji Mandi Road

References

5 The eight selected deprivational indicators are Bhopal 462 001 those belonging to (1) a household that does not Alkire, Sabina and Maria Emma Santos (2010): “Acute

Madhya Pradesh

have access to a source of drinking water on its Multidimensional Poverty: A New Index for Devepremises; (2) a household that does not have loping Countries”, Oxford Poverty and Human Ph: 2740705 access to electricity for lighting; (3) a household D evelopment Initiative, Working Paper No 38, July.

70 september 17, 2011 vol xlvi no 38

Table A2: Results of Logistic Regression with Muslims Social Group as a Variable

Table A1: Results of Logistic Regression with SC/ST Social Group as a Variable
Sl No	State			Coefficients				Odd Ratios
		Sector	Up to Primary	SC/ST	SC/ST and	Constant	Sector	Up to Primary	SC/ST	SC/ST and
					Primary					Primary
1	Andhra Pradesh	-0.628	1.1215	0.6475	0.1429	0.0868	0.5338	3.0696	1.9108	1.1536
2	Assam	1.905	1.3028	0.4382	0.1169	-1.078	6.7181	3.6795	1.55	1.124
3	Bihar	1.051	1.4028	0.9977	-0.0372	0.5175	2.8595	4.0667	2.7121	0.9635
4	Gujarat	0.313	1.0556	1.1002	-0.0492	-3376	1.4085	2.8736	3.0048	0.952
5	Haryana	0.036	1.0527	1.5873	0.1799	-0.647	1.037	2.8655	4.8906	1.1971
6	Himachal Pradesh	1.2	1.0358	0.7819	-0.0656	-1.428	3.3193	2.8174	2.1856	0.9365
7	Jammu and Kashmir	-0.561	1.0369	-0.083	-0.3096	0.786	0.5707	2.8205	0.9207	0.7337
8	Karnataka	-0.108	1.2025	1.0821	-0.2095	0.2543	0.8974	3.3285	2.9508	0.811
9	Kerala	-0.258	0.6583	0.97	-0.0679	0.2364	0.7726	1.9315	2.6379	0.9352
10	Madhya Pradesh	-0.075	1.2828	1.381	0.2303	0.6106	0.9274	3.6066	3.9791	1.2589
11	Maharashtra	0.447	1.0064	0.7502	0.2697	0.1996	1.5628	2.7359	2.1174	1.3095
12	Orissa	0.948	1.2791	0.9839	0.377	0.3296	2.5801	3.5933	2.6748	1.4578
13	Punjab	0.085	0.9947	1.2959	0.2355	-1.346	1.0889	2.7039	3.6543	1.2655
14	Rajasthan	-0.166	1.2196	1.2352	0.0426	0.3457	0.8472	3.3859	3.439	1.0436
15	Tamil Nadu	0.341	1.1971	0.8032	-0.1694	0.0008	1.4066	3.3104	2.2327	0.8442
16	Uttar Pradesh	0.455	1.2367	1.1628	-0.3253	0.3205	1.5762	3.4442	3.1989	0.7223
17	West Bengal	1.162	1.5433	0.4601	0.1511	-0.705	3.1948	4.6801	1.5842	1.1631
18	Jharkhand	1.945	1.4291	0.8796	-0.3792	-0.307	6.9954	4.175	2.4098	0.6844
19	Chhattisgarh	0.507	1.4943	0.7101	-0.3273	0.2299	1.6606	4.4562	2.0342	0.7208
20	Uttarakhand	0.808	0.998	0.8667	0.4722	0.6141	2.2427	2.7128	2.379	1.6035
21	Other NE states	1.654	0.8814	-0.919	0.169	-0.278	5.2301	2.4144	0.3991	1.1842
22	Other states	-0.159	1.3177	1.2534	-0.0534	-0.828	0.8527	3.7347	3.5023	0.948

Sl No State 1 Andhra Pradesh 2 Assam 3 Bihar 4 Gujarat 5 Haryana 6 Himachal Pradesh 7 Jammu and Kashmir 8 Karnataka 9 Kerala 10 Madhya Pradesh 11 Maharashtra 12 Orissa 13 Punjab 14 Rajasthan 15 Tamil Nadu 16 Uttar Pradesh 17 West Bengal 18 Jharkhand 19 Chhattisgarh 20 Uttarakhand 21 Other NE states 22 Other states Economic & Political Weekly	Coefficients Sector Up to Primary Muslims -0.522 1.1808 0.4925 1.845 1.2412 0.4433 1.124 1.4701 0.6801 0.572 1.087 1.0558 0.117 1.202 1.2643 1.225 1.0902 1.5396 -0.555 0.992 1.2492 0.077 1.2346 0.8706 -0.21 0.7021 0.487 0.233 1.4557 0.5548 0.544 1.09 0.8231 1.041 1.5446 0.2815 0.235 1.1886 1.2065 -0.042 1.3575 0.7021 0.416 1.2154 0.4339 0.593 1.2109 0.2888 1.139 1.5781 0.7982 1.974 1.411 0.4789 0.585 1.4678 0.0496 0.91 1.0778 0.8583 1.493 0.9321 1.2467 -0.122 1.4611 1.7263 september 17, 2011 vol xlvi no 38	Muslims and Primary -0.137 0.1944 -0.3262 -0.3925 0.4195 -1.1391 -0.1776 -0.3208 -0.1785 0.221 -0.1534 -0.8432 -0.8032 -0.5968 -0.4634 0.0291 -0.1651 0.079 -0.7696 0.0901-0.2033 -1.0217	Constant 0.1039 -1.017 0.51 -0.339 -0.512 -1.273 0.1328 0.2339 0.1858 0.5872 0.2156 0.4496 -1.05 0.4582 0.07 0.3397 -0.702 -0.178 0.3829 0.6333 -0.781 -0.724	Sector 0.5932 6.3251 3.0784 1.7712 1.1242 3.4035 0.5741 1.0799 0.8109 1.2626 1.7231 2.8306 1.2654 0.9587 1.5154 1.8086 3.1222 7.1963 1.7955 2.4836 4.4507 0.8848	Odd Ratios Up to Primary 3.2568 3.4598 4.3497 2.9654 3.3269 2.9748 2.6967 3.4371 2.018 4.2874 2.9744 4.686 3.2825 3.8865 3.3716 3.3566 4.8456 4.1 4.34 2.9382 2.5398 4.311	Muslims 1.6364 1.5579 1.9742 2.8742 3.5408 4.6625 3.4876 2.3885 1.6275 1.7415 2.2775 1.3252 3.3419 2.018 1.5433 1.3349 2.2215 1.6144 1.0508 2.3592 3.4788 5.6199	Muslims and Primary 0.8719 1.2146 0.7216 0.6754 1.5212 0.3201 0.8372 0.7255 0.8365 1.2473 0.8578 0.4303 0.4479 0.5506 0.6292 1.0295 0.8478 1.0822 0.4632 1.0943 0.816 0.36 71

EPW

Table A3: Results of Logistic Regression with OBC Social Group as a Variable

Sl No	State			Coefficients				Odd Ratios
		Sector	Up to Primary	OBC	OBC and	Constant	Sector	Up to Primary	OBC	OBC and
					Primary					Primary
1	Andhra Pradesh	-0.58	1.3616	0.609	-0.444	-0.079	0.5598	3.9026	1.8386	0.6414
2	Assam	1.913	1.3922	-0.046	-0.3306	-0.97	6.7764	4.024	0.9546	0.7185
3	Bihar	1.106	1.5988	0.2824	-0.2362	0.4219	3.0207	4.9475	1.3263	0.7896
4	Gujarat	0.471	1.0826	0.6302	-0.1657	-0.375	1.6021	2.9524	1.8781	0.8473
5	Haryana	0.103	1.3372	0.5178	-0.3957	-0.629	1.1082	3.8086	1.6783	0.6732
6	Himachal Pradesh	1.231	1.0799	-0.227	-0.0184	-1.227	3.425	2.9443	0.7966	0.9817
7	Jammu and Kashmir	-0.571	1.0171	-0.336	-0.7086	0.7972	0.5652	2.765	0.7144	0.4923
8	Karnataka	-0.009	1.2925	0.1076	-0.1856	0.3276	0.991	3.642	1.1136	0.8306
9	Kerala	-0.21	0.7673	0.0941	-2279	0.2521	0.8107	2.154	1.0986	0.7961
10	Madhya Pradesh	0.044	1.8279	0.6517	-0.9487	0.4818	1.1101	6.2212	1.9189	0.3872
11	Maharashtra	0.441	1.1619	0.259	-0.2125	0.2541	1.5538	3.1959	1.2956	0.8086
12	Orissa	1.046	1.7463	0.1019	-0.5633	0.409	2.8466	5.7334	1.1073	0.5693
13	Punjab	0.229	1.1914	0.1259	-0.0354	-1.06	1.2567	3.2918	1.1342	0.9652
14	Rajasthan	-0.025	1.5862	0.0771	-0.5779	0.4478	0.9752	4.8851	1.0801	0.561
15	Tamil Nadu	0.409	1.3951	0.0935	-0.2861	0.027	1.505	4.0353	1.098	0.7512
16	Uttar Pradesh	0.498	1.4571	0.5258	-0.5368	0.2334	1.6461	4.2936	1.6917	0.5846
17	West Bengal	1.205	1.6653	0.3208	-0.5472	-0.665	3.3352	5.2874	1.3782	0.5786
18	Jharkhand	1.983	1.4966	0.3646	-0.1569	-0.284	7.2646	4.4663	1.44	0.8548
19	Chhattisgarh	0.566	1.5919	0.36	-0.3533	0.258	1.7619	4.913	1.4334	0.7024
20	Uttarakhand	0.846	1.1206	0.1819	0.0246	0.6855	2.3303	3.0666	1.1995	1.0249
21	Other NE states	1.569	1.0536	1.2059	-0.3422	-1.039	4.8002	2.8679	3.3399	0.7102
22	Other states	-0.259	1.527	1.0623	-0.6943	-0.759	0.772	4.6045	2.8931	0.4994

Table A4: Results of Logistic Regression with ‘Others’ Social Group as a Variable

Sl No State 1 Andhra Pradesh 2 Assam 3 Bihar 4 Gujarat 5 Haryana 6 Himachal Pradesh 7 Jammu and Kashmir 8 Karnataka 9 Kerala 10 Madhya Pradesh 11 Maharashtra 12 Orissa 13 Punjab 14 Rajasthan 15 Tamil Nadu 16 Uttar Pradesh 17 West Bengal 18 Jharkhand 19 Chhattisgarh 20 Uttarakhand 21 Other NE states 22 Other states 72	Sector -0.632 1.789 1.043 0.116 -0.047 1.178 -0.648 -0.078 -0.223 -0.171 0.409 0.879 0.056 -0.237 0.29 0.41 1.007 1.748 0.251 0.881 1.488 -0.159	Up to Primary 1.0517 1.3549 1.4374 0.9642 1.2187 1.0562 0.7931 1.1077 0.5828 1.2948 1.0843 1.4836 1.2247 1.1747 1.1242 1.0328 1.5828 1.3477 1.2786 1.3134 0.9201 0.9974	Coefficients Others -1.308 -0.6550 -0.8220 -0.553 -1.207 -0.5320 -1.353 -0.956 -0.856 -1.339 -0.898 -0.776 -1.1930 -1.0050 -1.959 -1.132 -0.781 -1.112 -1.231 -0.742 -0.117 -1.673	Odd Ratios Others and Constant Sector Up to Primary Others Others and Primary Primary 0.0306 0.6172 0.5316 2.8624 0.2703 1.0311 -0.2673 -0.656 5.9806 3.8765 0.5295 0.7654 -0.7040 0.8891 2.8363 4.2096 0.4395 0.4946 -0.25090 0.6660 1.1227 2.6229 0.2117 0.7781 -0.3061 0.3036 0.9544 3.3826 0.2992 0.7363 -0.0227 -0.909 3.2493 2.8755 0.5874 0.9775 0.3565 1.2547 0.5231 2.2103 0.2586 1.4284 0.0138 0.8028 0.9246 3.0273 0.3844 1.0139 0.1364 0.5615 0.8005 1.791 0.4247 1.1462 -0.2856 1.4672 0.8432 3.6502 0.2622 0.7516 -0.1203 0.7483 1.5053 2.9574 0.4073 0.8866 -0.1606 0.8562 2.4092 4.4087 0.4604 0.8516 -0.4360 -0.3190 1.0577 3.4032 0.3033 0.6466 0.0678 0.9774 0.7891 3.2373 0.3661 1.0702 0.0682 0.2894 1.3368 3.0778 0.1411 1.0706 -0.0957 0.9348 1.5074 2.8090 0.3224 0.9087 -0.2301 -0.069 2.7368 4.8683 0.4578 0.7944 -0.3855 0.3383 5.7428 3.8484 0.3288 0.68012 0.2083 0.9281 1.2857 3.5917 0.2919 1.2315 -0.5081 1.2139 2.4125 3.7190 0.4764 0.6017 0.1666 -0.729 4.426 2.5094 0.8893 1.1813 0.195 0.4544 0.853 2.7113 0.1877 1.2153 september 17, 2011 vol xlvi no 38 Economic & Political Weekly

EPW

Table A5: Average Daily Per Capita Consumption Expenditure (DPCE) by State and Poverty Status

State Extremely Poor Poor Marginal Vulnerable Poor and Vulnerable Middle Income High Income Total 2 3 4 5 6 7 8 9

Andhra Pradesh 8.3 12.2 14.1 18.1 16.0 30.0 79.1 23.6 Assam 7.3 10.9 13.7 19.2 30.5 20.2

16.2

70.4 Bihar 7.6 13.1 13.7 74.4

10.4 30.8

17.8 15.3 Gujarat 9.3 15.8 18.7 93.2

12.3 22.0 27.8

39.2 Haryana 9.7 16.0 19.5 105.8

12.5 22.6 31.6

38.4 Himachal Pradesh 8.9 12.1 15.2 21.2 18.9 35.7 79.8 Jammu and Kashmir 10.6 15.6 17.9 23.1 21.6 36.8 89.5 26.7 Karnataka 10.5 14.4 16.5 98.8

12.3 19.5 35.6 23.7

13.7 23.7 33.4 Kerala 9.5 17.1 20.3 39.8 99.2 Madhya Pradesh 8.8 11.3 13.7 19.0 14.5 34.9 83.4 18.4 Maharashtra 10.9 16.7 18.4 112.1

13.9 23.5 45.0 28.1 Orissa 7.3 12.5 12.5 79.6

10.1 17.7 33.2

15.2 Punjab 8.9 11.7 14.7 19.0 80.2

21.2 35.9 32.7 Rajasthan 10.7 15.8 17.8 93.0

12.9 21.0 36.7 22.5 Tamil Nadu 9.6 13.1 16.4 22.3 18.2 41.2 97.8 27.4 Uttar Pradesh 8.0 11.2 14.1 19.4 15.3 33.3 83.6 19.2 West Bengal 8.2 11.3 14.2 20.0 16.0 35.8 89.2 23.2 Jharkhand 8.2 14.1 14.4 75.3

11.1 19.6 36.1 18.0 Chhattisgarh 8.1 13.3 14.2 95.2

11.1 19.3 35.2 18.3 Uttarakhand 11.4 18.9 19.7 110.6

14.9 26.4 46.4 24.2 Other NE states 8.9 12.8 16.3 22.6 19.1 37.3 87.2 26.4 Total 8.9 11.7 14.7 20.4 16.3 36.8 92.7 23.2