Mean Consumption, Poverty and Inequality in Rural India in the Sixtieth Round of the National Sample Survey

Transcription

1 Revised 27 Oct 2006 Mean Consumption, Poverty and Inequality in Rural India in the Sixtieth Round of the National Sample Survey Raghbendra Jha, Raghav Gaiha, Anurag Sharma Australian National University of Delhi Monash University University ABSTRACT This paper reports on mean consumption, poverty (all three FGT measures) and inequality during January to June 2004 for rural India using National sample Survey (NSS) data for the 60 th Round. Mean consumption at the national level is much higher than the poverty line. However, the Gini coefficient is higher than in recent earlier rounds. The headcount ratio using the thirty day recall is 22.9 per cent and with the seven day recall this stands at 17.9 per cent and, with the combined data, this figure is 20.6 per cent. Mean consumption, all three measures of poverty and the Gini coefficient are computed at the level of 20 states and 63 NSS regions in these 20 states. It is surmised that despite impressive growth rates deprivation is pervasive, pockets of severe poverty persist, and inequality is rampant. All communication to: Prof. Raghbendra Jha, Australia South Asia Research Centre, RSPAS, Australian National University Canberra, Act 0200 Australia Phone: Fax: r.jha@anu.edu.au ASARC Working Paper 2006/11

2 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Mean Consumption, Poverty and Inequality in Rural India in the Sixtieth Round of the National Sample Survey I. Introduction With the recent release of the National Sample Survey (NSS) data for the 60 th round (conducted during January to June 2004) the debate on the impact of the current economic policy stance will, no doubt, be renewed anew. The change in the methodology for collecting NSS data put into effect in the 55 th Round ( ) created controversy about the comparability of the 55 th Round results with the earlier rounds. The 60 th Round data is not entirely comparable with the 55 th Round data yet it does provide an inkling of how the movement of poverty has developed during this period. In this paper we provide estimates of mean consumption, poverty and inequality at three levels: the national level, state level and at the level of 63 NSS regions. II. Poverty During the 1990s A spate of studies (contained in Deaton and Kozel, 2005 and elsewhere) contain widely varying estimates of reduction in poverty during the 1990s a period that witnessed wideranging economic reforms both domestic and external liberalisation and accelerated growth. The reduction in poverty, however, varies depending on whether the estimates are obtained from adjusted or unadjusted 55 th round NSS estimates (see, for example, Sundaram and Tendulkar (2003, 2005), Deaton (2005), Datt, Kozel and Ravallion (2005), Deaton and Dreze (2002), Kijima and Lanjouw (2003) and Planning Commission, Government of India as reported in Sen and Himanshu (2004). To give the flavour of the debate, two sets of evidence are reviewed below. ASARC Working Paper 2006/11 2

3 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Although official estimates of poverty reduction are high (the HCR in rural areas fell from per cent to per cent, and the number of poor from 244 million to million over the period ), the balance of evidence with careful adjustments of the 1999/2000 NSS data point to markedly lower reductions. An admirably comprehensive and by far the most persuasive is the contribution by Sen and Himanshu (2005). It casts doubts over earlier estimates of poverty reduction specifically, that it was faster than that in the previous decade. As it turns out, the headcount ratio declined but at most by 3 percentage points and the absolute number of the poor did not decline over The reason is that both interstate and within-state inequality rose sharply during the 1990s, with the latter being the larger component of total inequality in India. Within-state urban inequality rose sharply in the 1990s, while a declining phase of within-state rural inequality was reversed during this period. So, despite rapid growth, poverty reduction in the 1990s was muted. The characterisation of the 1990s as a relatively lost decade for poverty reduction is thus not an exaggeration. The reasons relate to weak linkages between liberalisation and agricultural growth on the one hand, and between the latter and poverty reduction on the other. First, although the growth rate of agricultural (crop) output was slightly higher during the 1990s (3.08 per cent annually as against 2.94 per cent during the 1980s), it slowed down to 1.6 per cent during and 2000/1. 1 Secondly, whether agricultural growth translates into poverty reduction depends greatly on how unequal is the distribution of land. An implication of highly unequal distribution of land is that the growth of real agricultural wages tends to be sluggish. Independently of agricultural growth, abrupt increases in price of food result in higher 1 For details, see Nayyar (2003). ASARC Working Paper 2006/11 3

4 Raghbendra Jha, Raghav Gaiha & Anurag Sharma poverty. 2 Some additional but more recent evidence further points to slow reduction in poverty. (i) Between , the employment elasticity of agricultural growth was barely 0.01 as compared to 0.70 during (ii) There was a sharp drop in the growth rate of real agricultural wages, from 3.99 per cent per annum during to to 2.34 per cent annually during to (iii) The (wholesale) price index for food rose faster than that for all commodities. With as the base at 100, food prices rose to 433 in , while for all commodities it rose to only In any case, the disadvantaged groups especially the STs and female-headed households- continue to be considerably worse-off than all other rural households. 4 Their social exclusion, and limited access to assets especially land and education.perpetuate their poverty (Gaiha and Imai, 2004). Yet another dimension that needs critical scrutiny is spatial concentration of the poor. An analysis based on the NSS regional data over the period shows that the ranking of the poorest 20 regions has not changed much and that they continue to account for a large share of the rural poor. 5 What these findings point to is that not only do persistently poor require greater access to human and physical assets, markets and financial services but also conditions that would enable them to overcome social exclusion. While a definitive comparison of the spatial concentration over the period covered by the earlier NSS rounds with the 60 th round must await a comparable regional classification, our analysis draws 2 For illustrative results on the causal role of these variables based on the ICRISAT data for the semi-arid region, see Gaiha (1995). 3 For details, see Nayyar (2003). 4 In India, survey data for show that per capita incomes among SCs were lower than the state averages (24 per cent in Andhra Pradesh, and 41 per cent in Kerala). In the eastern state of Orissa, more than 24 per cent of the population consisted of STs, as compared to 9 per cent in the country as a whole. 92 per cent of the households belonging to a ST in rural southern Orissa were poor twice the state poverty rate and three and a half times the national poverty rate (CPRC, 2004). Besides, the overall prevalence of stunting among SC/ST children is much higher than the all-india average (over 63 per cent and about 58 per cent, respectively. For details, see Gaiha and Kulkarni (2005). See also Sundaram and Tendulkar (2005) for illustrations of relatively low rates of poverty reduction among subsets of SC/ST households during This is based on a collaborative study by R. Jha and R. Gaiha, funded by a DFID grant. See Jha and Gaiha (2003) and Jha and Sharma (2003). ASARC Working Paper 2006/11 4

5 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS attention to marked variations within states and across regions as a continuing feature of poverty in India despite spectacular growth rates of income. III. Methodology The poverty measures used in this paper are all drawn from the popular Foster-Greer- Thorbecke (FGT) class of functions. This is written as: Y α = [( z yi ) / z] yi < z α / n where Y is the measure of poverty, y i is the (per capita) consumption expenditure of the ith household or the ith class of household, z is the poverty line, n is the number of observations, and α is a non-negative parameter. The headcount ratio, PG0, given by the percentage of the population who are poor, is obtained when α=0. This measure fails to capture the extent to which individual income (or expenditure) falls below the poverty line. Hence we use our second measure: the poverty gap index (PG1) given by the aggregate income shortfall of the poor as a proportion of the poverty line and normalized by the population size. This is given by setting α=1 in (1). PG1 captures the acuteness of poverty since it measures the total shortfall of the poor from the poverty line. In other words, it measures the total amount of income necessary to remove that poverty. This measure has the drawback that it does not consider the importance of the number of people who are below the poverty line. For this reason, it is important to use both measures of poverty jointly to evaluate the extent of poverty. There are certain policy changes that favor one group of poor and adversely affect another group. In such cases PG0 may not register any change but PG1 may get around this problem to some extent. A further improvement is the Foster-Greer-Thorbecke measure which is obtained by setting α=2 in (1). We abbreviate this as PG2. 6 (1) 6 This is also referred to as a distributionally sensitive measure. More generally, if α >1, the FGT index is distributionally sensitive. ASARC Working Paper 2006/11 5

6 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Also reported in this paper are data on the Gini coefficient, a standard measure of inequality. Typically this is defined as: G = n n i i= 1 j = 1 2n 2 y y µ j (2) Thus G is calculated as the relative mean difference, i.e., the mean of the difference in consumption levels between every possible pair of households divided by the mean size µ. IV. Results Consumption data in the 60 th Round of the NSS has been reported in three formats: (i) schedule 1 uses 30 day recall; (ii) schedule 2 uses 7 day recall; and (iii) consumption data combining the two recall periods are reported. We present the results at three levels: (a) national; (b) state level (20 states) and (c) agroclimatic zone wise (63 NSS regions) for all three data formats. The poverty lines used in the analysis are reported in Table 1. These poverty lines are updates for the official poverty lines used by the Government of India for the year 2004 using CPIAL figures for the nation as a whole (for India) and state level CPIALs. Table 1: Poverty Lines Used in the Analysis: Rs. per capita per month India Assam J&K West Bengal Himachal Pradesh Orissa Punjab Madhya Pradesh Haryana Gujarat Rajasthan Maharashtra UP Andhra Pradesh Bihar Karnataka Manipur Kerala Tripura Tamilnadu Meghalaya ASARC Working Paper 2006/11 6

7 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Computations on mean consumption, poverty and inequality at the national level are reported in Table 2. Table 2: Mean Consumption, Poverty and Inequality in Rural India at the National Level 30-Day recall 7-day recall Combined Mean per capita consumption 7 (Rs. Per month) PG0 l PG PG Gini Coefficient) Although mean per capita consumption is much higher than the poverty line, the Gini coefficient is higher than in the earlier rounds. The Sen Himanshu estimate (2005) for the 55 th round, for example, is for the adjusted 7-day question on food. It may be noted, however, that our Gini estimates do not vary much with the recall period. But the poverty estimates do. These are lower with the 7-day recall. State level results for mean consumption are presented in Table 3. 7 This and other mean consumption magnitudes are weighted means of expenditures. ASARC Working Paper 2006/11 7

8 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Table 3: Mean Consumption in Indian States (Rs. per capita per month) 30-Day recall 7-day recall Combined J&K Himachal Pradesh Punjab Haryana Rajasthan UP Bihar Manipur Tripura Meghalaya Assam West Bengal Orissa Madhya Pradesh Gujarat Maharashtra Andhra Pradesh Karnataka Kerala Tamilnadu In Table 4 we report on state level poverty levels for all three FGT measures for all three recall measures. ASARC Working Paper 2006/11 8

9 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Table 4: State Level FGT Poverty measures 30-day recall 7-day recall Combined 30-day recall 7-day recall Combined J&K Assam PG PG PG PG PG E PG Himachal Pradesh West Bengal PG PG PG PG PG PG Punjab Orissa PG PG PG PG PG PG Haryana Madhya Pradesh PG PG PG PG PG PG Rajasthan Gujarat PG PG PG PG PG PG UP Maharashtra PG PG PG PG PG PG Bihar Andhra Pradesh PG PG PG PG PG PG Manipur Karnataka PG PG PG PG PG2 9.41E E E-05 PG Tripura Kerala PG PG PG PG PG PG Meghalaya Tamilnadu PG PG PG PG PG PG ASARC Working Paper 2006/11 9

10 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Some comments are in order. 8 I. In all cases, the poverty estimates are lower with the weekly recall- in a few cases (e.g. Assam, Orissa and Karnataka), the differences are large. This holds for all poverty indices. II. Head count ratios vary considerably from 3.6 per cent for Punjab to per cent for Orissa. Further there is considerable variation in the other two FGT measures of poverty. Thus there is great heterogeneity in poverty incidence across Indian states. III. What is important to note is that some of the poorest states (viz. Bihar, UP, Orissa, Madhya Pradesh) continued to exhibit high incidence and severity of poverty in the 60 th round. As these states also accounted for a large majority of the poor, it is safe to conclude that large subsets of the rural population continue to remain in abject poverty. IV. Among superior performers, Assam and Maharashtra stand out. Turning now to within-state inequality we report on the Gini coefficients within the 20 Indian states for the three recall periods in Table 5. Table 5: The Gini Coefficient in Select States: 2004 State Gini Gini Gini State Gini Gini Gini 30-day recall 7-day recall Combined 30-day recall 7-day recall Combined J&K Assam Himachal Pradesh West Bengal Punjab Orissa Haryana Madhya Pradesh Rajasthan Gujarat UP Maharashtra Bihar Andhra Pradesh Manipur Karnataka Tripura Kerala Meghalaya Tamilnadu Note that the comparisons are relative to the estimates given for the 30 day-recall with adjustments in Sen and Himanshu (2005), and essentially ordinal in nature. ASARC Working Paper 2006/11 10

11 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS I. Apart from the relatively small north eastern states with low Ginis, in most other states the Ginis are high (e.g. Andhra Pradesh, Kerala, Tamil Nadu, Maharashtra, West Bengal, and UP0, pointing to considerable inequality in consumption expenditure. Other things being equal, and to the extent that consumption expenditure inequality reflects inequality in physical (e.g. land) and human capital (e.g. education), even high rates of growth of income are unlikely to translate into substantial poverty reduction. 9 II. Our second observation is that the Gini and recall period are unrelated in so far as in some cases the 7-day recall Ginis are higher (e.g. Himachal Pradesh, Punjab, Madhya Pradesh, Gujarat, Maharashtra, West Bengal) while in others (e.g.kerala, Karnataka, Andhra Pradesh, Haryana) the 30-day recall Ginis are higher. We now report results for the 63 NSS regions in the 20 states studied in the sample. Since there is considerable heterogeneity in the mean income, poverty and inequality across Indian states we rank the NSS regions in respect of these criteria. Table 6 reports mean consumption across the 63 NSS regions. 9 For poverty reduction, some forms of inequality matter more than others. Important ones include inequality in the distribution of assets, especially land, human capital, financial capital, and access to public assets, such as rural infrastructure. The fast growing economies of East and South East Asia had the advantage of low asset inequality compared to other Asian and Pacific economies in some countries, following land reforms along with a better spread of education. Simulations for selected Asia countries show that even with low historic growth rates observed over continuing over the period , achievement of the MDG of halving poverty by 2015 will depend largely on reduction in income Gini. A case in point is Bangladesh. For details, see Gaiha et al. (2006). ASARC Working Paper 2006/11 11

12 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Region Mean per capita consumption Rs. per month, using 30 day recall Table 6: Mean Consumption in 63 NSS regions. Mean per capita consumption Rs. per month, using 7 day recall Mean per capita consumption Rs. per month, using combined data Orissa - Southern Bihar -Northern Orissa - Southern Bihar - Northern Orissa-Southern Bihar - Northern Vindhya Vindhya South Manipur Hills Manipur Hills Manipur Hills Northern Orissa - Coastal Bihar Central Madhya Pradesh Vindhya Bihar - Central Orissa - Northern Assam- Hills Assam - Hills South Western Orissa - Coastal Central Assam - Hills Orissa - Coastal South Bihar - Central UP-Eastern Northern J&K - Outer Hills Orissa - Northern UP Eastern UP - Southern UP - Eastern Inland Central South Central Rajasthan - Southern Rajasthan - Southern Rajasthan - Southern Tripura Central UP - Southern Maharashtra Eastern Eastern Plains Tripura UP - Southern Orissa - Northern Eastern Rajasthan - Southeastern Inland Central Northern Tripura Eastern Plains Assam - Western Plains Eastern Assam - Western Plains Eastern Plains Assam - Western Plains Himalayan Inland Central Himalayan Himalayan Malwa Madhya Pradesh Malwa Madhya Pradesh Malwa Western Plains Maharashtra J&K - Outer Hills Gujarat - Eastern Plains Gujarat-Eastern Plains Rajasthan - Southeastern Assam - Eastern Plains Gujarat - Dry Areas Western Plains Gujarat - Dry Areas West Bengal Western Plains Gujarat - Eastern Plains Inland Eastern Rajasthan-Southeastern Gujarat - Dry Areas Assam- Eastern Plains Assam - Eastern Plains Inland Eastern Rajasthan-Western Inland Eastern ASARC Working Paper 2006/11 12

13 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Tamilnadu - Southern Meghalaya Rajasthan - Western UP - Central Rajasthan-Northeastern Coastal Coastal Coastal Inland Eastern Rajasthan - Western UP -Western UP - Central J&K - Outer Hills UP -Central Meghalaya Tamilnadu - Coastal Karnataka-Inland Eastern UP - Western Coastal and Ghats Haryana-Western Rajasthan - Northeastern UP - Western Maharashtra-Inland Eastern Tamilnadu - Southern Meghalaya Gujarat Northern Plains Coastal and Ghats Tamilnadu - Inland Karnataka-Coastal and Ghats Tamilnadu - Coastal Rajasthan - Northeastern Tamilnadu-Coastal Haryana - Western Manipur Plains Manipur Plains Tamilnadu - Inland Haryana - Western J&K- Jhelum Valley Manipur Plains J&K - Jhelum Valley Tamilnadu-Southern J&K - Jhelum Valley Gujarat - Northern Plains Tamilnadu-Inland Gujarat - Northern Plains Andhra Pradesh Kerala - Northern Kerala - Northern Kerala - Northern Gujarat - Saurashtra Gujarat - Saurashtra Gujarat - Saurashtra Punjab - Northern Haryana - Eastern Haryana - Eastern Himachal Pradesh Inland Western Himachal Pradesh Central Plains Himachal Pradesh Central Plains Inland Western Central Plains Inland Western Haryana - Eastern Punjab - Northern Punjab - Northern Gujarat - Southern Plains Tamilnadu - Coastal Northern Gujarat - Southern Plains Tamilnadu - Coastal Northern Punjab - Southern Tamilnadu - Coastal Northern Punjab - Southern Gujarat - Southern Plains Punjab - Southern J&K - Mountains J&K - Mountains J&K - Mountains Kerala - Southern Kerala - Southern Kerala - Southern Coastal Coastal Coastal Tables 7, 8 and 9 give results on FGT measures of poverty for 30-day recall, 7-day recall and combined data for the 63 NSS regions. All figures have been arranged in ascending order to emphasise regional variations in the incidence and severity of poverty. ASARC Working Paper 2006/11 13

14 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Table 7: Poverty Incidence and severity in NSS regions: 30-day recall 30 - day recall NSS Region PG0 NSS Region PG1 NSS Region PG2 Manipur Plains 0 Manipur Plains 0 Manipur Plains 0 Haryana - Eastern Haryana - Eastern Meghalaya J&K - Mountains Meghalaya Haryana - Eastern J&K - Jhelum Valley Punjab - Northern Punjab - Northern Coastal Manipur Hills Coastal and Ghats Meghalaya J&K - Mountains J&K - Jhelum Valley Punjab - Northern Coastal Coastal Kerala - Southern J&K - Jhelum Valley J&K - Mountains Punjab - Southern Coastal and Ghats Himachal Pradesh Manipur Hills Kerala - Southern J&K - Outer Hills Gujarat - Saurashtra Gujarat - Saurashtra Gujarat - Saurashtra Himachal Pradesh Himachal Pradesh Kerala - Southern Gujarat - Southern Plains Punjab - Southern Central Plains Haryana - Western Gujarat - Southern Plains Gujarat - Southern Plains Central Plains Central Plains Manipur Hills Kerala - Northern J&K - Outer Hills Inland Western Inland Western Inland Western Rajasthan - Northeastern Rajasthan - Northeastern Inland Eastern Punjab - Southern Rajasthan - Western Haryana - Western Rajasthan - Western Coastal and Ghats Rajasthan - Northeastern Haryana - Western Gujarat Northern Plains Rajasthan - Western Assam - Eastern Plains Rajasthan - Southeastern UP - Western UP - Western UP - Western Assam - Eastern Plains Rajasthan - Southern J&K - Outer Hills Gujarat - Northern Plains Gujarat - Northern Plains Tamilnadu - Southern Tamilnadu - Southern Tamilnadu - Southern Assam - Eastern Plains Tamilnadu - Inland Tamilnadu - Inland Tamilnadu - Inland Kerala - Northern Assam - Western Plains Western Plains Inland Eastern Western Plains Tamilnadu - Coastal Northern Eastern Plains Western Plains Inland Eastern Assam - Western Plains Assam - Western Plains Rajasthan - Southeastern UP - Southern Rajasthan - Southern Gujarat - Eastern Plains Tamilnadu - Coastal Northern Kerala - Northern ASARC Working Paper 2006/11 14

15 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Inland Eastern Coastal Eastern Plains Coastal Tamilnadu - Coastal Coastal Tripura UP - Central UP - Southern Rajasthan - Southeastern Eastern Malwa Tripura Tripura UP - Central Eastern Plains Inland Eastern Assam - Hills Northern Assam - Hills Madhya Pradesh Malwa Assam - Hills UP - Central Bihar - Northern UP - Southern Tamilnadu - Coastal Tamilnadu - Coastal Northern Malwa Inland Central Gujarat - Dry Areas Northern Himalayan Vindhya Gujarat - Eastern Plains Tamilnadu - Coastal Inland Central Himalayan Bihar - Northern Gujarat - Dry Areas Vindhya Orissa - Coastal Inland Central Gujarat - Dry Areas Vindhya Gujarat - Eastern Plains Eastern Rajasthan - Southern Himalayan Eastern Bihar - Central Northern Bihar - Northern Orissa - Coastal Bihar - Central Central Orissa - Coastal Bihar - Central Inland Eastern UP - Eastern Central UP - Eastern Central UP - Eastern South South South Orissa - Northern Orissa - Northern Orissa - Northern Orissa - Southern Orissa - Southern Orissa - Southern ASARC Working Paper 2006/11 15

16 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Table 8: Poverty Incidence and Severity in NSS regions: 7-day recall 7 - day recall PG0 PG1 PG2 J&K - Mountains 0 J&K - Mountains 0 J&K - Mountains 0 Manipur Plains 0 Manipur Plains 0 Manipur Plains 0 Coastal and Ghats 0 Coastal and Ghats 0 Coastal and Ghats Manipur Hills J&K - Jhelum Valley 8.32E - 05 J&K - Jhelum Valley 4.86E - 07 Inland Eastern Manipur Hills Manipur Hills 3.74E - 05 J&K - Jhelum Valley Inland Eastern Inland Eastern 3.98E - 05 Punjab - Northern Haryana - Eastern Haryana - Eastern 9.87E - 05 Coastal Punjab - Northern J&K - Outer Hills Haryana - Eastern Meghalaya Punjab - Northern Kerala - Southern Assam - Hills Meghalaya Gujarat - Saurashtra Coastal Assam - Hills Meghalaya Kerala - Southern Coastal Punjab - Southern J&K - Outer Hills Assam - Eastern Plains Kerala - Northern Assam - Eastern Plains Kerala - Southern Assam - Hills Himachal Pradesh Himachal Pradesh Himachal Pradesh Punjab - Southern Kerala - Northern Assam - Eastern Plains Kerala - Northern Gujarat - Northern Plains Central Plains Gujarat - Saurashtra Punjab - Southern Central Plains Gujarat - Southern Plains Rajasthan - Northeastern Central Plains Tamilnadu - Inland Rajasthan - Northeastern Rajasthan - Western Gujarat - Northern Plains Gujarat - Saurashtra Inland Western Western Plains Tamilnadu - Coastal Haryana - Western Inland Western Western Plains Inland Haryana - Western Western J&K - Outer Hills Tamilnadu - Coastal Gujarat - Dry Areas UP - Western Rajasthan - Western Haryana - Western Rajasthan - Northeastern Tamilnadu - Inland UP - Western Assam - Western Plains UP - Western Rajasthan - Western Tamilnadu - Coastal Tamilnadu - Inland Inland Northern Tripura Gujarat - Eastern Plains Western Plains Gujarat - Dry Areas Tripura Tripura Gujarat - Southern Plains Assam - Western Plains Rajasthan - Southeastern Assam - Western Plains Coastal Gujarat - Southern Plains Tamilnadu - Southern Rajasthan - Southern Rajasthan - Southern ASARC Working Paper 2006/11 16

17 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Gujarat - Dry Areas Gujarat - Eastern Plains Inland Eastern Gujarat - Northern Plains Coastal Eastern Plains Rajasthan - Southern Rajasthan - Southeastern Himalayan Tamilnadu - Coastal Northern UP - Southern UP - Southern Inland Eastern Orissa - Coastal Malwa Orissa - Coastal Coastal Eastern Plains Eastern Plains Inland Eastern UP - Southern UP - Central UP - Central Gujarat - Eastern Plains Himalayan Rajasthan - Southeastern UP - Central Tamilnadu - Southern Bihar - Central Orissa - Coastal Tamilnadu - Coastal Northern Northern Inland Central Malwa Malwa Tamilnadu - Southern Himalayan Bihar - Central Bihar - Northern Bihar - Central Inland Central Inland Tamilnadu Bihar - Northern Southern Coastal Northern Central Northern Vindhya Bihar - Northern UP - Eastern Inland Central Northern Vindhya UP - Eastern UP - Eastern Central Central Vindhya South South Eastern Eastern Eastern South Orissa - Southern Orissa - Southern Orissa - Southern Orissa - Northern Orissa - Northern Orissa - Northern ASARC Working Paper 2006/11 17

18 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Table 9: Poverty Incidence and severity in NSS regions: Combined data Combined NSS Region PG0 NSS Region PG1 NSS Region PG2 Manipur Plains 0 Manipur Plains 0 Manipur Plains 0 J&K - Mountains Haryana - Eastern Manipur Hills J&K - Jhelum Valley Meghalaya Haryana - Eastern Haryana - Eastern Manipur Hills Meghalaya Manipur Hills J&K - Mountains Coastal and Ghats Punjab - Northern Punjab - Northern Punjab - Northern Coastal J&K - Jhelum Valley J&K - Jhelum Valley Kerala - Southern Coastal and Ghats J&K - Mountains Meghalaya Coastal Coastal Punjab - Southern Kerala - Southern Inland Eastern Gujarat - Saurashtra Himachal Pradesh J&K - Outer Hills Himachal Pradesh Inland Eastern Himachal Pradesh Coastal and Ghats Gujarat - Saurashtra Kerala - Southern Kerala - Northern J&K - Outer Hills West Bengal Central Plains Central Plains Punjab - Southern Gujarat - Saurashtra Gujarat - Central Southern Plains Plains Punjab - Southern Haryana - Western Inland Western Rajasthan - Northeastern Rajasthan Inland Western Northeastern Assam - Eastern Plains Rajasthan - Western Assam Eastern Plains Inland Western Assam - Eastern Plains Haryana - Western Gujarat - Northern Plains Rajasthan - Gujarat Northeastern Southern Plains Haryana - Western Inland Eastern Kerala - Northern Rajasthan - Western Tamilnadu - Inland Rajasthan - Western Gujarat - Southern Plains Gujarat Northern Plains UP - Western UP - Western UP - Western J&K - Outer Hills Western Western Plains Plains Assam - Hills Tamilnadu - Inland Tamilnadu - Inland Assam - Western Plains Assam - Hills Kerala - Northern Rajasthan Southeastern Rajasthan - Southern Tamilnadu - Southern Assam - Western Plains Assam - Hills Gujarat - Northern Plains Tamilnadu - Coastal Gujarat - Eastern Plains Rajasthan - Southern Assam - Western Plains Western Plains Tripura Tripura Tamilnadu - Coastal Inland Eastern Coastal Tamilnadu - Southern Inland Eastern ASARC Working Paper 2006/11 18

19 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Inland Rajasthan Eastern Southeastern Tamilnadu - Coastal Tripura Tamilnadu - Southern Tamilnadu Coastal Northern Coastal UP - Southern Coastal Eastern Plains Eastern Plains Gujarat - Dry Areas UP - Southern Gujarat - Dry Areas UP - Central Malwa UP - Southern Inland Southern Rajasthan - Southern Himalayan Eastern Plains Tamilnadu - Coastal Northern UP - Central Rajasthan - Southeastern Gujarat - Dry Areas Northern Gujarat - Eastern Plains Malwa UP - Central Himalayan Malwa Himalayan Bihar - Northern Gujarat - Eastern Plains Orissa - Coastal Orissa - Coastal Tamilnadu Coastal Northern Inland Central Inland Central Bihar - Central Orissa - Coastal Northern Eastern Bihar - Central Inland Central Vindhya Bihar - Northern Bihar - Central Vindhya Vindhya Northern UP - Eastern Bihar - Northern UP - Eastern Central Central Central UP - Eastern Eastern Eastern South South South Orissa - Northern Orissa - Northern Orissa - Northern Orissa - Southern Orissa - Southern Orissa - Southern In Table 10 we report on the Gini coefficients in these regions again arranged in ascending order. ASARC Working Paper 2006/11 19

20 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Table 10: Gini Coefficients in NSS regions 30-day recall 7-day recall Combined data Manipur Hills Manipur Hills Manipur Hills Manipur Plains J&K - Outer Hills Manipur Plains J&K - Jhelum Valley Assam - Hills Assam - Hills Assam - Hills Manipur Plains J&K - Jhelum Valley J&K - Outer Hills Bihar - Northern Meghalaya J&K - Outer Hills Coastal and Ghats J&K - Jhelum Valley Meghalaya Meghalaya Inland Eastern Coastal and Ghats Rajasthan - Southeastern Tripura Bihar - Northern Southestern Coastal and Ghats Tripura Vindhya Bihar - Northern UP - Southern UP - Southern Assam - Western Plains Inland Eastern Punjab - Northern Rajasthan - Southern Haryana - Western Tripura Haryana - Western Assam - Western Plains Rajasthan - Western UP - Southern Vindhya Assam - Western Plains Assam - Eastern Plains Assam - Eastern Plains Northern Rajasthan - Southern Inland Eastern Assam - Eastern Plains Gujarat - Saurashtra Haryana - Western Gujarat - Saurashtra Vindhya Rajasthan - Western Eastern Gujarat - Saurashtra Inland Central Eastern Plains Northern Orissa - Coastal Haryana - Eastern Eastern Plains Rajasthan - Southern Western Plains Western Plains Western Plains Rajasthan - Western Haryana - Eastern Inland Northern Gujarat - Dry Areas Orissa - Coastal Eastern Plains Bihar - Central Gujarat - Dry Areas J&K - Mountains Punjab - Northern Inland Eastern Orissa - Coastal J&K - Mountains Haryana - Eastern Gujarat - Eastern Plains Gujarat - Eastern Plains Gujarat - Dry Areas Rajasthan - Northeastern Bihar - Central Tamilnadu - Southern Rajasthan - Northeastern Central Kerala - Northern UP - Eastern Northern Coastal Eastern Malwa Gujarat - Northern Plains Inland Central ASARC Working Paper 2006/11 20

21 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS Gujarat - Eastern Plains UP - Eastern Central South Punjab - Northern Inland Eastern UP - Eastern Central J&K - Mountains Himalayan Rajasthan - Southeastern Bihar - Central Orissa - Southern Orissa - Southern Inland Southern UP - Western Coastal Rajasthan - Northeastern Tamilnadu - Coastal Malwa Himachal Pradesh Eastern Himalayan Gujarat - Southern Plains Tamilnadu - Inland Gujarat - Northern Plains Orissa - Southern Himachal Pradesh Kerala - Northern Himalayan Inland Central Himachal Pradesh Coastal Inland Eastern Tamilnadu - Southern Coastal Kerala - Southern Tamilnadu - Coastal Gujarat - Northern Plains UP - Western UP - Western Malwa Tamilnadu - Inland Tamilnadu - Coastal Inland Western South Punjab - Southern Tamilnadu - Southern Inland Western Tamilnadu - Inland South Kerala - Southern Inland Western Punjab - Southern UP - Central Punjab - Southern Gujarat - Southern Plains Kerala - Northern Rajasthan - Southeastern Orissa - Northern Coastal UP - Central South Western UP - Central Coastal Central Plains Kerala - Southern Gujarat - Southern Plains Orissa - Northern Orissa - Northern Central Plains Tamilnadu - Coastal Northern Tamilnadu - Coastal Northern Central Plains Tamilnadu - Coastal Northern We explore patterns of regional inequality in respect of mean consumption, the three FGT measures of poverty and inequality in table 11 (for variation across states) and table 12 (for variation across the 63 NSS regions). ASARC Working Paper 2006/11 21

22 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Table 11: Variation of Mean Consumption, Poverty and Inequality across Indian States Mean Consumption (Rs. per capita per month) PG0 PG1 PG2 Gini Coefficient Average Highest Lowest Coefficient of Variation 30 day recall day recall Combined day recall day recall Combined day recall day recall Combined day recall E day recall E Combined E day recall day recall Combined It is interesting to note that the 7-day recall magnitudes are generally more volatile than 30- day magnitudes with the combined results lying in between. Further, the coefficients of variation of the FGT measures are much higher than those for mean consumption and the Gini coefficient. Within the FGT measures PG2 has higher coefficient of variation than PG1 which in turn has a higher coefficient of variation than PG0. Thus poverty, particularly severe poverty, is more volatile. Table 12: Variation of Mean Consumption, Poverty and Inequality across 63 NSS regions Mean Consumption (Rs. per capita per month) PG0 PG1 PG2 Gini Coefficient Average Highest Lowest Coefficient of Variation 30 day recall day recall Combined day recall day recall Combined day recall day recall Combined day recall day recall Combined day recall day recall Combined ASARC Working Paper 2006/11 22

23 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS We offer additional comments by concentrating on the upper and lower tails of the distributions for average, consumption, FGT indices and the consumption Ginis. In order to avoid tedious comparisons, we will first pick the lowest 5 regions in terms of per capita consumption and the highest five, using the 30-day recall. How many of these overlap with those with the lowest or highest FGT indices and Gini coefficients would help understand better spatial dimensions of well-being, deprivation and inequality. I. The regions with the lowest per capita expenditure (in ascending order) are 10 : Orissa-Southern Bihar-Northern Andhra Pradesh-South western Madhya Pradesh-South Orissa-Coastal. Those with the highest per capita expenditure (in ascending order) are: Tamil Nadu-Coastal Northern Punjab-Southern J&K-Mountains Kerala-Southern Maharashtra-Coastal II. It turns out that three out of the five regions with the lowest per capita expenditure are also the ones with the highest incidence of poverty (PG0). These are (in ascending order): Madhya Pradesh-South Andhra Pradesh-South Western Orissa-Southern 10 Manipur Hills is omitted, as it belongs to a relatively small state. ASARC Working Paper 2006/11 23

24 Raghbendra Jha, Raghav Gaiha & Anurag Sharma Let us now examine the overlap with regions with the highest distributionally sensitive poverty (PG2). It turns out that the regions with the highest incidence of poverty are also the ones with the highest PG2. III. There is some overlap between regions with the highest per capita expenditure and lowest incidence of poverty (PG0). These are : J&K-Mountains Maharashtra-Coastal The overlap with regions with the lowest PG2 is smaller, as it comprises just one region viz. Maharashtra Coastal. IV. Let us now turn to the overlap between lowest consumption and lowest Gini. The overlap is small comprising Madhya Pradesh-South. V. The overlap between regions with the highest consumption and Gini is not so small, as it comprises Kerala-Southern, and Tamil Nadu-Coastal Northern. VI.Finally, let us examine the overlap between the Ginis and FGT indices. The regions with the highest Ginis (in ascending order) are: West Bengal-Central Plains Kerala-Southern Karnataka- Andhra Pradesh- Tamil Nadu-Coastal Northern. None of these regions overlaps with those with the highest incidence of poverty (PG0). However, there is some overlap with regions with the highest distributionally sensitive poverty, comprising Karnataka. ASARC Working Paper 2006/11 24

25 Mean Consumption, Poverty & Inequality in Rural India in the 60 th Round of the NSS IV. Concluding Remarks Even though we have not gone beyond ordinal comparisons with the estimates of poverty and inequality from the previous 55 th round of the NSS, some of the findings are noteworthy. First, more than a few of the states (e.g. UP, Madhya Pradesh, Orissa), with high levels of poverty measured using the FGT class of poverty indices, and high concentrations of the poor in rural areas continue to do so in the 60 th round of the NSS. Second, except for the small north eastern states, high levels of inequality in consumption expenditure distribution persisted in many of the states. Third, the overlap between regions with the lowest consumption expenditure and highest poverty, and between those with highest consumption and lowest poverty was far from negligible. Fourth, the overlaps between lowest consumption expenditure and lowest Ginis, and between highest consumption and highest Ginis were negligible or small. Fifth, in general, the 7-day recall magnitudes (average consumption expenditure, poverty and inequality) are more volatile than 30-day magnitudes. Sixth, the CVs of the FGT measures are higher than those for mean consumption and the Gini coefficient. Finally, the CVs of PG2 (or distributionally sensitive poverty) are higher than those of PG1 (intensity of poverty) which in turn are higher than those of PG0 (or the head-count ratio). It is surmised that despite impressive growth rates deprivation is pervasive, pockets of severe poverty persist, and inequality is rampant. ASARC Working Paper 2006/11 25