Chapter4 ESTIMATION OF POVERTY AMONG ELDERLY IN INDIA

Transcription

1 Chapter4 ESTIMATION OF POVERTY AMONG ELDERLY IN INDIA 4.1 Introduction In the last chapter we looked into the profile of the aged population in India and its differentials across different Indian states. Such a profile provides information on the characteristic features of the elderly and helps to draw inferences regarding their disadvantages or ill being if any as against the non-elderly population on one hand, and interpret the same in the context of the households they belong to. However, this does not facilitate an objective assessment of deprivation like poverty or compromised well-being because of age. Hence we make a simplistic attempt at redefining household well-being in the presence and absence of elderly in order to comment on the associated deprivation of elderly. The essential difference that we wish to make is in terms of interpreting household well-being defined on a per-capita scale with incorporation of economies of scale and age-sex composition of the household. Following an assessment of the magnitude of poverty among elderly in different Indian states, its determinants are also discussed; This chapter is organised in the following sections: introduction, concept of poverty, income poverty, issues in the measurement of poverty, adjusting size and composition, estimating dependency adjusted poverty among elderly, characteristics and determinants of poverty among aged. The measurement and analysis of poverty, inequality, and vulnerability are crucial for cognitive purposes (to know what the situation is), for analytical purposes (to understand the factors determining this situation), for policymaking purposes (to design interventions best adapted to the issues), and for monitoring and evaluation purposes (to assess the effectiveness of current policies and to determine whether the situation is changing). Various definitions and concepts exist for the term 'well-being'. First, well-being addresses what is typically referred to as poverty, that is, whether households or individuals possess enough resources or abilities to meet their current needs. This definition is based on a comparison of the individual's income, consumption, education, or other attributes with some defined threshold below which individuals are considered as being poor in that particular attribute. Second, the chapter focuses on inequality in the distribution of income, consumption, or other attributes across the population. This is based on the premise that the relative position of individuals or households in society is an important aspect of their welfare. In addition, the overall level of inequality in a country, region, or population group, in terms of monetary and 60

2 non-n~.onetary dimensions, is in itself also an important summary indicator of the level of welfare in that group. Here we try to map the well-being of the aged in India by looking at the poverty that exists among them. 4.2 Concept of Poverty: Different Approaches As defined by the World Bank (1993), poverty is defined as the inability to achieve a politically acceptable living standard. 'Living standard' is a near synonym for real income, appropriately measured and after adjustments are made for goods and services produced and consumed at home, for income-in-kind and for household size and composition. The setting of an acceptable living standard for use in defining poverty is an important policy matter The issue of poverty has received a lot of attention over the decades. The concern is the identification of the poor. Both poverty research and social policy employ a wide variety of poverty definitions. However, all definitions may be fitted into one of the following categories (Schram, 1993): 1. Poverty is having less than an objectively defined, absolute minimum. 2. Poverty is having less than others in society, in relative terms. 3. Poverty is feeling you do not have enough to get along. Poverty according to the first category of definitions is absolute, it is relative according to the second category, and is a mixture of absolute and relative poverty according to the third category.. Defining the level of poverty according to the different definitions under these categories may result in different estimates of the number of poor people in the population. This part discusses attempts to define poverty and different approaches Absolute Poverty Approach The absolutist approach to defining poverty begins with the concept of "minimum subsistence", that is, some bundles of goods and services that are regarded as essential to the physical needs of an individual or a family. Those who do not possess the economic resources to obtain these goods and services are considered poor. In the most severe conception of the absolute approach, this bundle of economic goods consists of the minimum caloric intake essential to human existence, and perhaps some forms of shelter (Ravallion, 1992). 61

3 This method of analysis was most commonly employed in the early period of the century. With the advancement of society and the enhanced expectations for a better quality of life, researchers nowadays seldom employ the absolutist approach. However, in places where abject poverty is still prevalent, such as Central Africa, the absolutist approach is still being used from time to time to determine minimum living standards (Paul, 1994). The most severe criticism against the absolute approach is that within that approach, human needs are interpreted as being predominantly physical needs - that is, for food, shelter and clothing - rather than as social needs. People are not, it is argued, simply individual - organisms requiring replenishment of physical energy. They are social beings expected to perform socially demanding roles such as parents, workers and citizens. They are not simply consumers of physical goods but producers of those goods and active participants in their societies. They are dependent on collectively provided utilities and facilities. Moreover, the measurement of costs for food is a problematic matter. The amount and cost of the food which is. eaten depends on the social roles played and the dietary habits observed, as well as the kinds of foods available in the market. Specifying the costs of meeting dietary needs is therefore very difficult. The absolute poverty approach involves conceptual and methodological difficulties, including (i) the determination of necessities, and (ii) the quantitative assessment of the demand for necessities (UNDP, 1998). Faced with these complexities, many researchers have suggested an alternative approach to defining poverty, one that depends on relative standards Relati,vist Poverty Approach According to the relativist approach to defining poverty, the mean or median value of national income represents the economic indicator that corresponds to the dominant life style. An individual or a family whose income is less than that value can be defined as being on the poverty threshold, with no means to live in that life style. For example, a person or family with less than one half of the average after-tax income can be said to be poor. By defining poverty in these terms, the relativists can avoid defining "absolute needs" and put emphasis on the equitable distribution of income. The proponents of the relativist approach push the government to formulate policies for redistributing wealth in order to tackle the poverty problem (Townsend, 1982). The relativist approach to defining poverty originated in the 1960s and was put forward by the critics of the welfare policy adopted by the United Kingdom (Townsend, 1996). The poor 62

4 people were those not better off in terms of their position relative to the average standard of living in society. Relative poverty prevented people from participating in activities, which were customary in the societies in which they lived. There are a number of criticisms of the relative measure of poverty. First, it perpetuates poverty in the statistical sense that some fixed proportion of the population is always regarded as poor. Poverty therefore persistently exists. Second, while a relative measure of poverty can give an estimate of the size of the poor, it cannot provide any information on the quality of life of the poor Budget Standards Method The budget standards method to define poverty is based upon the determination of a list of necessities. Persons who cannot obtain all items on the list will be considered poor. It is an absolutist approach in structure but various socially determined essential needs are incorporated to form a set of budgets. In other words, the budget standards definition is based on the notion of a basket of goods and services, having the elements of "subsistence" and "basic minimums for social lives". In operation, determining the absolutely basic ingredients of human subsistence assembles an inventory of budget standards. Such a list may include minimum food, clothing, shelter and fuels. To this list of basic minimums we may add means of transportation, some recreational activities, and certain goods and services we deem appropriate, if not absolutely essential. Those persons who cannot afford all the items will be classified as poor (Tabatatbai, 1995) This method has its critics. First, there is difficulty in producing generally acceptable criteria for selecting the items to be included in the list. Second, the exposition of need also depends on the development of societies. Thus subjective value judgments have to be made in the context of each community's economic circumstances to establish the components. Third, it requires enormous efforts to develop a budget for such expenditure components and to keep each list up-to-date in line with changing social circumstances. 4.3 Measurement of Poverty The choice of variables will affect findings on the composition as well as the size of the poor population, and hence the policy response will differ accordingly. The three issues considered are the choice of poverty indicator, unit of analysis, and tools of measuring poverty, which are to be fixed prior to measuring poverty (ADB, 1993; Walker,2004). 63

5 4.3.1 Choice of Poverty Indicator Income is always taken as a single indicator to measure poverty because total consumption of basic needs (e.g., food, housing and clothing) and essential needs (e.g., transportation and social activities) are difficult to quantify. In other words, income is taken as a proxy for living standards. However, assessments of poverty based on income have to be qualified: assessing the standards of living by income alone may understate or overstate the level of living, if the composition of the families or the market supply situation is not taken into consideration. Some examples are useful to explain the concept. A family that can share in the consumption of others may have higher living standards than the total family income would permit. persons living with their children may benefit from their expenditure. Such considerations would lead one to expect the actual number of poor people to be less than the estimated number. Conversely, income may overstate the level of living when money alone cannot guarantee the ability to buy the necessary goods, for instance, if the required goods are not available in the market. If income is used as a proxy for consumption, the reference period should capture permanent rather.than transitory living standards. Therefore, using permanent income as the indicator would give a closer match than an indicator using current income. By the same token, an indicator derived from annual income should be a better one than the one derived from income in a week or a month Unit of Analysis The defining of units of analysis will affect the findings of the studies. The standard of living is dependent upon the consumption pattern. Scale of economy renders multi-member units more effective in consumption of economic resources than individuals. Three definitions are commonly used for unit of analysis: Household- one or more persons, whether related or unrelated, who share common living quarters. Family - two or more persons living together (sharing common living quarters) who are related by blood or by marriage. Individual - one single person. In considering the choice among different units, it is important to note that the consumption attributes of different goods and services are different. Some items of consumption such as 64

6 food are essentially individualistic, whereas others, such as housing, have some of the attributes of public goods as far as individuals in the household are concerned. For the latter, all members of the household enjoy broadly equal levels of consumption and living, but for the former, there may be inequality within the household. The distribution of benefits within the unit is difficult to observe. In practice, researchers usually use households as the unit of analysis, assuming that all within the unit share the same standard of living. If poverty is measured with an individual as the unit of analysis, and if the analysis is based on an individual's income, the result of the research will be different. As mentioned, if individuals with low income live in relatively well-off households, the living standards of its members are more directly related to the total income of the family than to their individual income. Hence, the number of people in poverty would be smaller if household is used as the unit of analysis Instruments of measuring poverty Counting the number of people or households below a poverty line is also contentious: as this head count alone cannot reveal the extent of poverty of individuals or households. In other words, it does not tell us how far they are below the poverty line. Thus, determination of the poverty gap is necessary in measuring the extent of poverty. This distinction has a fundamental and guiding effect on the policy on poverty. The policies and resources required to lift out of poverty a large number of people just below the poverty line would be different from those required to raise the standard of living of a smaller number of people far below the poverty line. After looking at the issues to be considered in the measurement of poverty, we look into income poverty (the most common indicator of poverty). 4.4 Income poverty The income measure of poverty is the most widely used measure of poverty. Here we fix the measure of income as an indicator of welfare and populations are divided based on minimum income requirement. As such, here income is taken as a measure of the welfare function that determines the end outcome (Alcock, 1997). Lack of income to minimum calorie intake is considered to be below poor in terms of the income poverty concept (Spieker, 1999). Then the poverty band is fixed based on income to fix the intensity of poverty (UNDP, 2000). The measures of USD 1 per day in Purchasing Power Parities is the standard used to monitor poverty trends at the global level but for country level monitoring national authorities are 65

7 encouraged to use indicators based on national poverty lines to measure income poverty (UNDP, 2003). There exist different bands to measure the income poverty in various countries as per country specific factors. One of the major limitations of measuring income poverty is the likeliness of underreporting of income in the developing counh ies (Sundaram and Tendulkar, 1993; Lenvine, 2006). The ability to transfer income to well-being is another serious issue raised (Sen, 1982; Rainwater and Smeeding, 1995). So in general, income poverty is often attributed based on proxy of income. The income proxy method is a behaviorist approach based on people's consumption patterns. Data on household expenditure are used as an income proxy to measure poverty. The theory is based on the fact that the consumption patterns of various family compositions are the result of the relative competition between need and choice. Whether need or choice prevails depends on the amount of income earned by the families. In practical terms, low-income families usually spend a greater proportion of expenditure on necessities, while better-off families will purchase more quality goods and non-necessities. Therefore, if researchers or governments can, from the data on household expenditure, identify a proper proportion of expenditure on necessities against the total expenditure, a poverty threshold can be worked out. This provides an income proxy for poverty, based on the proportion of total expenditure spent on the purchase of necessities. However, the cut-off I point need not be 30 per cent, and the measure need not be restricted to expenditure on food. As societies are going through rapid changes, any standard devised in the past is difficult to justify in new circumstances. Hence, these standards have to be recurrently checked in order to tie in with the development of social trends. However, the measures needed to keep updating the standards would be costly and time-consuming. The concrete definitions of necessities are always controversial (Rainwater and Smeeding, 1995), There is no secondary data source for tracing the trend in income poverty in India and its states. Analysts therefore depend upon the National Sample Survey Organization (NSSO)., quinquennial series of household consumption expenditure surveys to study trends in inequality in per capita expenditure (proxy for income) generally by estimating rural and urban poverty. These surveys started in (27th Round), give detailed consumer expenditure data for 1983 (38th Round), (43rd Round) (50th Round) (55th Round) and (61st Round). 66

8 4.5 Literature Review on various issues on measuring poverty Empirical research by economists on poverty in developing countries has generally been concerned with its measurement in terms of income and consumption. Behind this metric lies the concept of utility or welfare, which people are assumed to derive from income and consumption. Yet there has been little attempt to measure poverty in terms of subjectively perceived welfare. In this chapter, we shall explore the income poverty approach, attempting to gain insights from new research on the economics of poverty and well-being for understanding poverty in developing countries. Economic research on poverty and well-being (or subjective well-being, or satisfaction with life - we use the terms interchangeably) is sparse.and recent but growing rapidly. It is apparent from this literature that there are two important gaps to be filled: first, the one reflecting different aspects of poverty and vulnerability (Diener, 1994), and second, issues in the conventional measures of poverty. Some theoretical research on poverty in developing countries has eschewed income or consumption as the evaluative criterion. Alternative criteria have been put forward, for e.g., the fulfillment of basic needs and the extent of peoples' capabilities to be and to do things of intrinsic worth. Much of the research has involved the estimation of poverty and well-being functions, in which poverty and well-being (subjectively rated on an ordinal or cardinal scale) are the dependent variables and various socio-economic characteristics of the individual, household or community are used as explanatory variables. Some of the research relates to particular countries (generally advanced economies), using either cross-section or panel data sets; and some covers many countries, normally using comparable data sets derived from the World Values Survey. The main findings from the general literature are the following. First, poverty and well-being increase with absolute income, ceteris paribus, but not proportionately and at a diminishing rate (Frey and Stutzer, 2002). Moreover, differences in income explain only a small proportion of the variation in poverty and the importance of income appears to vary among countries: poverty and well-being levels are lowest in the poorest countries, but the relationship between income and poverty and well-being is weak beyond a fairly low international level of income per capita. This is consistent with the argument that poverty and well-being depend in part on the gratification of certain absolute biological and psychologic~! needs (Veenhoven, 1991). 67

9 The limited role of absolute income is further suggested by the fact that income, poverty and well-being are positively related in cross-section but not in time-series studies. It is possible that mean poverty and well-being did not rise over time because aspiration levels adjusted to, and so rose along with, mean incomes in the society, and poverty and well-being varied positively with income but negatively with aspirations (Easterlin, 2001). The second main finding, therefore, is that poverty and well-being depend on relative income, defined by the reference group or the reference time that people have in mind. The concepts of absolute and relative income are not the only economic determinants of poverty and well-being. Other factors in the functional set can affect well-being independently of its effect on income (Clark and Oswald, 1994; Winkelmann and Winkelmann, 1998). The general unemployment rate also has a depressing effect, suggesting that having a higher risk of becoming unemployed reduces poverty and well-being. Another indication of economic insecurity is inflation: countries and periods with higher inflation display lower poverty and welbeing, ceteris paribus (Di Tella et al., 2001). Subjective wellbeing is influenced by several factors that are non-economic or potentially so, such as age, sex, marital status, health status, education, social capital, religion, and social and political institutions (Helliwell, 2002). Sen (1983) introduced the concept of a person's "capabilities" to be and to do things of intrinsic worth, i.e., resources adequate to achieve a specified set of "functioning". He~ argued that absolute deprivation in terms of a person's capabilities could imply relative deprivation in terms of income, resources or commodities, e.g., for taking part in the life of the community, for the avoidance of shame, or for the maintenance of self-respect. He favoured the capability to function as the criterion for assessing the standard of living, and by implication, poverty, rather than the utility that might be derived from using that capability. Thus, Sen eschewed the "Welfarist'' approach to poverty with its underlying assumption that the evaluative criterion is the utility that people derive from goods and services. Atkinson and Bourguignon (1999) use the same framework but from a welfare perspective. They regard poverty as 'inadequate command over economic resources' but view this as an intermediate concern, the ultimate concern being in terms of 'capabilities' in the sense that Sen interpreted the term. Though there are different approaches, which give different aspects of understanding poverty, measuring income poverty still holds the key Here, we take the conventional path of measuring income poverty and look into income poverty among the elderly in India. 68

10 4.6 Estimation of Poverty among the Aged Like most developing countries, India has been experiencing population ageing, attributable to the decline in both fertility and mortality over the past five decades or so. This phenomenon has important implications for the poverty reduction strategies in the country. Although demographic (Visaria, 1998) and other socio-economic and health (Prakash, 1999; Rajan et al., 1999) aspects of ageing in India have been examined by various social scientists, there are no official measures of old age poverty in India as in many other developing countries. Some academic works have attempted to understand poverty among the aged in India (Subbarao et al., 2005; Barrientos et al., 2003). Deaton and Paxon (1995) provi~e estimates of old age poverty in six large Indian states for A more recent study by Pal and Palacious in 2005 tried to estimate poverty among elderly by using data from the 52nd Round of the NSS and found that elderly households are better in terms of poverty; however, the study failed to capture intra-household distribution of welfare with the presence of an aged person in the household. In general, there has been a lack of research into an understanding of the extent, magnitude and nature of old age poverty in the Indian states. Here, we estimate poverty based on household expenditure. We try to adjust the size and composition of the households in the estimation of poverty as they have direct implications on the welfare distribution of the households. This is done by calculating dependency-adjusted poverty which is sensitive to the composition of the household with old age dependency ~ld age dependency-adjusted poverty In most practical work that documents levels of living and poverty, researchers make a look to differences in household size by working with income or consumption per capita. Such measures effectively assume that household resources are allocated according to the need of a member within the household. We propose a simple method for examining the sensitivity of welfare comparisons to assumptions based on welfare distribution within the households. We try to look at two important issues. One is the question of how re- sources are allocated within the household with higher levels of dependency and their implications on overall welfare functions. Deflating total household resources by an equivalence scale, defined as a function of the size of the household and its demographic composition, measure individual living standards. The second issue that we do not consider is the traditional estimation of equivalence scales. There is large and long-standing literature on the measurement of equivalence scales, (see the discussion in Deaton and Paxson 1998). Our approach here follows Buhmann et al. (1988), Coulter et al. (1989, 69

11 1994) and Banks and Johnson (1994) in assuming a parametric form for the equivalence scale, and examining the consequence of changing the parameters Methodology of old age dependency-adjusted poverty We assume that an elderly costs a fraction of what a non elderly costs, and that elasticity of costs with respect to adjusted household size through following redefining of 'N' through a dependency adjustment Here scaling is done by adjusting N to capture effects on per capita consumption N is calculated as follows, N=(A+aB)B Which A is non-elderly population and B is elderly population in a household. The parameter a is obtained by adding old age dependency to unity (1 +ODR) to scale down welfare loss in the household due to old age dependency. The estimation controls the costs of the elderly relative to total population The superscript Beta captures economies of scale that are reflected in a family due to consumption of indivisibilities. The value of 9 is always between 0 and 1. The greater the value of 9, the greater the level of economies of scale within the household (see Deaton and Paxon, 1998). If 9 is 1, there are no economies of scale within the household and thus APCE (Average Per Capita Expenditure) is the same as original per capita distribution and vice versa Adjustment in the Poverty Line After adjusting the per capita welfare to a revised APCE by adjusting the composition, size and economies of scale, we revise the poverty line given by the Planning Commission for the Indian states for different levels of economies of scale. Following Dreze and Srinivasan (1997), we consider the following formula for the adjustment of the different levels of economies of scale to the given poverty line in order to calculate HCR at different levels of economies of scale: Zn (8) =Z (h)1-8 Here 'Zn' is new poverty line, 'Z' is the old poverty line and 'h' denotes average family size of the states adjusted for ODR(an). Here we make adjustments of the old poverty line in confirmation with different values of economies of scale (8) that range from 0 to 0.8, - when the value of 8 is high, this indicates a decline in the overall distributive justice of welfare within the household. So here HCR will be calculated based on the newly adjusted APCE and the Adjusted Poverty Line for each value of 8. Table 1 and 2 give the newly adjusted poverty line for the major Indian states for both urban and rural areas. 70

12 Table 4.1: Poverty line for rural India for different economies of scale State Scale=1 Scale =.2 Scale =.4 Scale =.6 Scale =.8 Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana Himachal Pradesh Jammu& Kashmir Jharkhand Kama taka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh , Uttaranchal : West Bengal Source: Estimated by the author by making adjustments in poverty line of the P1anning CollUlUssion Table 4. 2: Poverty line for urban India for different economies of scale, State Scale=1 Scale =.2 Scale =.4 Scale =.6 Scale =.8 Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana Himachal Pradesh Jammu & Kashmir Jharkhand Kama-taka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: Fstimated by the author by making adjustments in poverty line of the Planning ColllllUSSlon 71

13 The tables 4.1 and 4.2 give the poverty line across different states of India with varying economic of scale in the households. It is clear that urban poverty line is higher than that of the rural one in all the states. As per our adjustments, we consider that the poverty line with no economies of scale of distribution is when Es=l. When Es=l, with the newly defined Dependency adjustment we get the newly adjusted poverty line which has only a slight deviation from the original estimates because of the revision in the welfare distribution within the population. As the economies of scale move higher with revised N, there occurs a drastic change in the welfare function in terms of the APCE (Average Per Capita Expenditure), and levels of APCE move closer toy, and the poverty lines will be revised. Thus we get the poverty lines given in the tables and show that there is a relative increase in the levels of the poverty line with an upward shift in the economies of scale and the component e tends to zero. The poverty line of Indian states after adjustment shows inter-state disparity. The levels of the urban poverty lines are higher than that of the rural levels. This is because of the relative increase in the levels of income to attain minimum calorie needs. States like Utharanchal, Kerala and Haryana have recorded high levels of the poverty line in the rural areas and others like Andhra Pradesh, Chhattisgarh and Madhya Pradesh report low levels of the poverty line. Karnataka records highest poverty line in urban India and the lowest levels of poverty bands are seen in the states of Assam, Bihar and Jharkhand. 4.7 Estimates of Dependency adjusted Poverty Here we try to estimate poverty based on the new APCE by using a conventional method of calculating poverty rates based on the HCR. In India, we find households of different compositions with divergence in the patterns of. the households. The demographic composition gives the distribution of the population in such a w~y that there is occurrence of the households with elderly, without elderly, with single or multiple elderly and with elderly and children, and elderly households without children. Here we calculate the poverty estimates of households with elderly, without elderly, with single or multiple elderly and with elderly and children, and elderly households without children to understand how the presence of elderly make a household poor and thus put the elderly in poverty. This estimation is done over APCE for 5 combinations of the economies of scale (0,0.2,0.4,0.6,0.8). The next section shows how these estimates of poverty vary across the Indian states. 72

14 4.7.1 Reflection on the estimates of Old dependency adjusted poverty in Indian states. The tables given below show various levels of dependency-adjusted poverty within the Indian states. This is shown for the different combinations of the households possible in Indian society, say, the distribution of the population is such that that there are households with elderly, without elderly, with single or multiple elderly (two or three or more), elderly and children, and elderly households without children. This is given across different economies of scales. Table 4.3: Dependency adjusted Poverty estimates with scale adjustment Es=1 for Rural India State elderly more non HHwith elderly than3 elderly elderly elderly elderly out elderly child ern Andhra Pradesh Assam Bihar Chhattisgarh '.29 Guiarat Haryana Himachal Pradesh Jammu & Kashmir Jharkand Karnataka Kerala Madhya Pradesh Maharashtra Orrissa Punjab Rajasthan TamilNadu Uttar Pradesh Utharanchal West Bengal Source: estimated by using NSS 61st Round ( ) Table 4.3 and Table 4.4 give the poverty estimates which assume no economies of scale where ES=1. Here welfare percapita distribution is almost in conformity with APCE after adjusting old age dependency. We have levels of poverty for households with different demographic scenarios, say, households with elderly, without elderly, with single or multiple elderly (two or threeor more), elderly and children, and elderly households without children, it is clear that as the old age presence in a household increases there are higher rates of poverty in terms of HCR in the households. 73

15 Table 4.4: Dependency adjusted Poverty estimates with scale adjustment Es=1 for Urban India State elderly more non HHwith elderly than3 elderly elderly elderly elderly out elderly children Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Hary~na Himachal Pradesh Jammu & Kashmir Jharkand Karnataka Kerala Madhya Pradesh ' Maharashtra Orrissa Punjab Rajasthan TamilNadu Uttar Pradesh Utharanchal West Bengal Source: estimated by using NSS 61st Round ( ) In both Rural and Urban areas, the rates of poverty are higher in the elderly households in comparison to non-elderly households. As the number of elderly increases, there are higher levels of poverty. The elderly households without children in general report less HCR poverty on an average compared to that of elderly households. The Urban areas of the states of Utharanchal, Rajasthan, Olhattisgarh and MP record relatively high rates of poverty among elderly, while states like Assam. Punjab, Haryana, etc., have lower rates of poverty. The Rural areas, of the states of Utharanchal, Olhattisgarh Orrisa, UP and MP show high incidence of poverty. It is lowest in Rajasthan. The gap between the HCR in the elderly and non-elderly are not uniform across the Indian states. It is relatively higher in the states of Himachal Pradesh and Kerala. It is negative in Jharkhand for Urban scenario. In Rural India, the gap is higher in Utharanchal, MP and Kamataka with the difference of more than 11 points, and it is minute in the case Rajasthan, Punjab and Haryana. The gap is relatively high in the states with high levels of demographic and poverty reduction achievements of the general population Now we examine the changing poverty rates across the households with elderly, without elderly, with single or multiple elderly (two or three or more) and with elderly and children, elderly households without children with different levels of economies of scale.(0.1,0.4,0.6, 0.8) 74

16 Table 4.5: Dependency adjusted Poverty estimates with scale adjustment ES= 0.8 for Rural India State With non 1 2 HH 3or HH elderly elderly elderly without more children children Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana HimachalPradesh Jammu &Kashmir Jharkhand Kama taka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu ; Uttar Pradesh Uttaranchal West Bengal Source: estimated by using NSS 61st Round ( ) In both Rural and Urban areas, the rates of poverty are higher in the elderly households in comparison to non-elderly households. As there is an increase in the number of elderly, there are high levels of poverty. The elderly households without children in general report less HCR poverty on an average to that of elderly households. When the economies of the scale is0.8,in Urban areas states of Utharanchal, Rajasthan, Chhattisgarh and MP record relatively high rates of poverty among elderly, while the rates are is low in states like Assam, Punjab, Haryana, etc. In the Rural areas, the states of Utharanchal, Chhattisgarh Orrisa, UP and MP show high incidence of poverty. It is lowest in Rajasthan. The gap between the HCR in elderly and non-elderly is not uniform across the Indian states. It is relatively higher in the states of Himachal Pradesh and Kerala. And it is negative for Jharkhand in the Urban scenario. In Rural India, the gap is higher in Utharanchal, MP and Kamataka, and is very small in the case of Rajasthan, Punjab and Haryana. The gap is relatively higher in the states with high levels of demographic and poverty reduction achievements of the general population. 75

17 Table 4.6: Dependency adjusted Poverty estimates with scale adjustment ES=0.8 for Urban India State non 1 2 With3 or HH HH elderly elderly elderly more children without children Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana HimachalPradesh Jammu &Kashmir Jharkhand Karnataka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by usmg NSS 61 st Round ( ) Table 4.7: Dependency adjusted Poverty estimates with scale adjustment ES= 0.6 for Rural India Urban With3 non 1 2 HH or HH elderly elderly elderly without more children children Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana HimachalPradesh Jammu &KashiPir Jharkhand Kama taka Kerala Madhya Pradesh Maharashtra Orissa punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by usmg NSS 61st Round ( ) 76

18 Table 4.8: Dependency adjusted Poverty estimates with scale adjustment ES= 0.6 for Urban India Urban With non 1 2 HH elderly elderly elderly 3 or HH more children without children Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana HimachalPradesh Jammu &Kashmir Jharkhand Kama taka Kerala Madhya Pradesh Maharashtra Orissa punjab Rajasthan TamilNadu Uttar Pradesh Uttarkhand West Bengal Source: estimated by using NSS 61st Round ( ) In both the Rural and Urban areas, the rates of poverty are higher in the elderly households in comparison to non-elderly households. As there is an increase in the number of elderly, there are high levels of poverty. The elderly households without children in general report less HCR poverty on an average to that of elderly households. When the economies of the scale go up to 0.6, there is relative increase in the poverty rates in comparison with a situation when FS=O.B. The pattern of distribution remains the same. The Urban areas of the states of Utharanchal, Rajasthan, Chhattisgarh and MP records relatively high rates of poverty among elderly, while it is low in the states like Assam, Punjab, Haryana etc. In the Rural areas of the states of Utharanchal, Chhattisgarh Orrisa, UP and MP show high incidence of poverty. It is lowest in Rajasthan. The gap between HCR in elderly and non-elderly is not uniform across the Indian states. It is relatively higher in the states of Himachal Pradesh and Kerala, and it is negative for Jharkhand in the Urban scenario. In Rural India, the gap is higher in Utharanchal, MP and Karnataka and very small in the case Rajasthan, Punjab and Haryana. This shows high variation in the interstate estimates of poverty. 77

19 Table 4.9: Dependency adjusted Poverty estimates with scale adjustment ES= 0.4 for Rural India Urban non 1 2 With3 HH HH elderly elderly elderly or more without children children Andhra Pradesh Assam Bihar Chhattis~arh Gujarat Haryana HimachalPradesh lammu &Kashmir Jharkhand Karnataka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by using NSS 61st Round ( ) Table 4.10: Dependency-adjusted Poverty estimates with scale adjustment ES= 0.4 for Urban India non 1 2 With3 or HH HH Urban elderly elderly elderly more without children children Andhra Pradesh Assam Bihar ~ Chhattis~arh Gujarat Haryana HimachalPradesh Jammu &Kashmir Jharkhand Karnataka Kerala Madhya Pradesh Maharashtra Orissa punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by usmg NSS 61st Round ( ) 78

20 In both Rural and Urban areas, the rates of poverty are higher in the elderly households when compared to non-elderly households. As the number of elderly increases, there are higher levels of poverty. The elderly households without children in general report less HCR poverty on an average to that of elderly households. When the economies of the scale go to 0.4, there is relative increase in the poverty rates in comparison with a situation when FS =0.6. The Urban areas of the states of Utharanchal, Rajasthan, Chhattisgarh and MP record relatively high rates of poverty among the elderly, while it is low in the states like Assam, Punjab, Haryana, etc. while in the Rural areas, the states of Utharanchal, Chhattisgarh Orrisa, UP and MP show high incidence of poverty. It is lowest in Rajasthan. As number of elderly goes up, the estimates go up. The gap between HCR in elderly and non-elderly is not uniform across the Indian states. It is relatively higher in the states of Himachal Pradesh and Kerala and it is negative for Jharkhand for the Urban scenario. In Rural India, the gap is higher in Utharanchal, MP and Karnataka. The gap is small in the case Rajasthan, Punjab and Haryana. The states with high levels have demographic and poverty reduction achievements of the general population show a wider gap. Table 4.11: Dependency adjusted Poverty estimates with scale adjustment ES= 0.2 for Rural India With non 1 2 HH 3or HH State elderly elderly elderly without more children children Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana HimachalPradesh Jammu &Kashmir Jharkhand Kama taka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by using NSS 61st Round ( ) 79

21 Table 4.12: Dependency adjusted Poverty estimates with scale adjustment Es=0.2 for Urban India Urban With3 non 1 2 HH or HH elderly elderly elderly without more children children Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Hary_ana : Himachal Pradesh Jammu &Kashmir Jharkhand Karnataka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by using NSS 61st Round ( ) In general, Tables' give the different rates of poverty in terms of HCR. This is estimated for various levels of economies of scales ranging from (0.6,0.4,0.2). These is an increase in the levels of poverty when the economies scale in the welfare distribution in terms of income moves down towards 0. In general, in both Rural and Urban areas, the rates of poverty are higher in the elderly households when compared to non-elderly households. As the number of elderly increases, so does the level of poverty. The elderly households without children in general report less HCR poverty on an average to that of elderly households. In the Urban areas, the states of Utharanchal, Rajasthan, Chhattisgarh and MP record relatively high rates of poverty among elderly, while it is low in the states like Assam, Punjab, Haryana, etc. In the Rural areas, the states of Utharanchal, Chhattisgarh Orrisa, UP and MP show high inddence of poverty. It is lowest in Rajasthan. The gap between HCR in elderly and non-elderly households are not uniform across the Indian states. It is relatively higher in the states of Himachal Pradesh and Kerala and it is negative for Jharkhand for the Urban scenario. In Rural India, the gap is higher in Utharanchal, MP and Karnataka and the difference is greater than in other states, and it is the least in the case Rajasthan, Punjab and Haryana. 80

22 4.8 Characteristics of poor elderly in Indian states The overall welfare of an individual is conditioned by the social, individual and familial capabilities of the individual. Difference in the characteristics and composition of the household and individuals act as building blocks of overall welfare. Characteristics such as sex, sector, social group, economic condition, and social and familial aspects influence the quality of life (UNDP 1994; World Bank 1998). These individual and social characters reflect on outcome indicators in the form of human development (UNESCO, 2000; UNDP, 2003). Different studies across the globe show that poverty has something to do with the social and individual character of the population though it varies in magnitude depending upon the nature and composition of the system (World Bank, 1994; UNDP, 2000, Johansson 2005). The social and familial system in India is too complex to affect the welfare distribution of the individuals and thus, poverty. There are attributes of how social and individual characters affect poverty. Studies show that various individual and social attributes affect welfare and thus poverty in the population, say, the population with a backward social and economic background is more likely to be poor (Planning Board,2005; Economic Survey, 2007). Among the elderly who are more dependent on others, poverty is likely to be more in the vulnerable groups that become more fragile to demographic and social shocks. Moreover, the characterwise analysis provides evidence for the necessity of different kinds of policy intervention to target various groups in order to improve the overall quality of the population in old age. Here we look at the characteristics of poor elderly in the Indian states. After estimating poverty among different combinations of elderly in Indian households, we here intend to look into the characteristics of the poor elderly. This is to understand how poverty among the aged is distributed across different sections of society, say, gender, household type, caste group, economic status and size of the households. The basic premise of the analysis is that it is the characteristics of the households that determine the levels of welfare of certain individuals (Sen, 1994; Deaton and Sreeninvasan, 1997). These characteristics also influence the well-being of the aged. In a country like India, the living conditions of elderly are conditioned by the dynamics in the socio-economic front (Alam, 2007). So here we look up on the distribution of the poor elderly (elderly in the poor households across different socio-economic factors. This distribution is done at economies scale of 0.6, which is ideal for India. So here we look into gender difference, economic characteristics of the household, economic conditions measured by SLI, and caste and size of 81

23 the household. The following section provides the distribution of poor elderly across different social and economic characteristics. Table 4.13 gives the gender-based distribution of the elderly below the poverty line. Here, the poor elderly are referred to as those come from poor households. It is found that there is not much gap in poverty across the male and female among the elderly. The pattern is not uniform. In general, the majority of the North Indian states record larger proportions of the poor elderly among the males, while in the case of the South it is the reverse. In Kerala, more than 56 per cent of the poor are women. It has the sharpest deviation in the poverty rates. 82

24 Table 4.13: Gender difference among poor elderly Male Female Male Female Andhra Pradesh Kerala Assam Madhya Pradesh Bihar Maharashtra Chhattisgarh Orissa Gujarat Puniab Haryana Rajasthan Himachal Pradesh TamilNadu Jammu &Kashmir Uttar Pradesh Jharkhand Uttaranchal Karnataka West Bengal Source: estimated by using NSS 61st Round ( ) Table 4.14 gives the distribution of the poor elderly across the house type. It is clear that in general, the major chunk of the poor elderly is from agriculture - workers or self employed - who account for almost three-fourths of the total poor in almost every Indian state. Some states like Kerala and Himachal Pradesh show some exceptions, with greater poverty among the other categories of workers. This pattern shows the need for strengthening the social security system in the informal sectors of the economy. Table 4.14: Poverty across household characteristics (type of household) self employed in agriculture other self employed non agriculture labour labour in agriculture Others Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana Himachal Pradesh Jammu &Kashmir Jharkhand Kama taka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by using NSS 61st Round ( ) 83

25 To understand how the poor are distributed across standards of living of the households, we look at the SLI of a household which is a composite index made to measure its general wellbeing on the basis of land owned, source of drinking water, cooking fuel, type of ration card, etc. The distribution of poor elderly shows that as the households SLI goes up, the concentration of the poor in these categories goes down, but the pattern is not similar in the Indian states. Table 4.15: SLI and poverty rates SLiofHH Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana Himachal Pradesh Jammu &Kashmir Jharkhand Karnataka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Bengal Source: estimated by using NSS 61st Round ( ) In the South Indian states, there is high concentration of poor in the first two SLI brackets. In Jammu and Himachal, it is concentrated in the second category while it is almost evenly distributed in states like Rajasthan, Maharashtra, Madhya Pradesh, etc. 84

26 Table 4.16: Poverty rates across social groups ST sc OBC Others Andhra Pradesh Assam Bihar Chhattisgarh Gujarat Haryana Himachal Pradesh Jammu &Kashmir Jharkhand Kama taka Kerala Madhya Pradesh Maharashtra Orissa Punjab Rajasthan TamilNadu Uttar Pradesh Uttaranchal West Beng_al Source: estimated by using NSS 61st Round ( ) Table 4.16 gives the distribution of poor elderly across the social groups of India. Some states like Jharkhand, Chhattisgarh, Rajasthan and Madhya Pradesh record a higher proportion of elderly poor in the ST population. The states of Punjab, Haryana, and Himachal Pradesh give a low concentration of elderly poor in the SC groups. In all other states except Jammu & Kashmir, West Bengal and Uttaranchal. we can see a distribution concentrated in the OBC groups. Uttarkhand, West Bengal and Jammu &Kashmir record a high proportion of poor in the general population. / 4.9. Determinants of Poverty Among In the last section, we look at the distribution of the elderly across different social attributes of the population. In this section, we examine what determines or makes an elderly poor visa-vis the non-poor. We take different indicators that can affect poverty directly or indirectly. Various studies in India have shown there exist sectoral variation in poverty estimates (Tendulkar, 1993; Planning Board, 2003) and it also reflected in the well-being of the elderly (Alam, 2005). The social and demographic development of various religions has different implications for the well-being of the elderly (Bose, 2000; Alam, 2005) and can also reflect upon poverty. Different social groups play a crucial role in individual well-being and thus on poverty estimates. Here also, we take the level of economic development of the state as a 85