-A A +A

Multidimensional Poverty and Vulnerability

The poverty of an entity is manifest in her deprivation not only in income but also in other dimensions such as health, nutrition and sanitation. Hence, a comprehensive measure of poverty must ideally take into account the performance of an individual across multiple dimensions. Vulnerability to poverty captures the likelihood of an entity falling into poverty, given the current status of the household. Unlike poverty, which describes the status of an entity at a point of time, vulnerability is predictive in nature. This study ranks 15 major states in terms of multidimensional poverty and vulnerability to multidimensional poverty at two time points in the 1990s. The results show that both multidimensional poverty and vulnerability provide additional insights for prescriptive policies.

SPECIAL ARTICLEEconomic & Political Weekly EPW may 17, 200877Multidimensional Poverty and Vulnerability Rosa A Abraham, K S Kavi KumarRosa A Abraham (rosa@mse.ac.in) and K S Kavi Kumar are at the Madras School of Economics, Chennai. The poverty of an entity is manifest in her deprivation not only in income but also in other dimensions such as health, nutrition and sanitation. Hence, a comprehensive measure of poverty must ideally take into account the performance of an individual across multiple dimensions. Vulnerabilityto poverty captures the likelihood of an entity falling into poverty, given the current status of the household. Unlike poverty, which describes the status of an entity at a point of time, vulnerability is predictive in nature. This study ranks 15 major states in terms of multidimensional poverty and vulnerability to multidimensional poverty at two time points in the 1990s. The results show that both multidimensional poverty and vulnerability provide additional insights for prescriptive policies.The World Development Report [WDR 2000-01] describes poverty as “pronounced deprivation in well-being”. Pov-erty describes the current status of an entity with regard to the attainment of a critical level in a dimension (like income or nutrition). So poverty is essentially capturing the outcome or end result of an entity. However, what would perhaps be a better guide for preventive policy actions would be to know how likely an entity is to fall into poverty, i e, the vulnerability to poverty of an entity. Vulnerability describes the likelihood of a household falling into poverty, given the current status of the household. Vulnerability describes the outcome of a household’s decision-making process while faced with a variety of risks, ex ante. It tells us what the potential outcome would be whereas poverty descri-bes the actual outcome. In this sense, poverty can be taken to be a “snap-shot” view of the current status of the household, where-as vulnerability being forward-looking provides a “crystal ball” view of the status of a household in the future [Dercon 2001]. As Prowse (2003) describes it, when a “vulnerable” population meets with some “hazard” or risk, the result is a “disaster” like poverty and deprivation. Here, vulnerability of a household is viewed as being manifested in a low level of achievement in one or more dimension – a level that is not low enough to qualify them as being poor but is still very close to this level. The higher the deprivation of the household, the more susceptible or vulner-able they are to being classified as poor. The dynamic, ex ante nature of vulnerability measures make them an important input in implementation of preventive policy. Those households identified as vulnerable need not always be labelled as poor and hence may be overlooked by policymakers. However, not being poor does not warrant ignoring these house-holds. Households identified as vulnerable must come under the lens of preventive policy measures. The need for incorporating the concept of vulnerability along-side measures and descriptions of poverty can be motivated from the following example [cited in Ligon et al 2003]. Consider a household with very low consumption expenditure but faced with minimal or no risks. Such a household would definitely not want to trade places with another household with a higher ex-pected consumption but experiencing a high consumption risk and possibility of starvation. A measure of poverty would nor-mally overlook the element of risk or vulnerability that the house-holds are faced with and paint a rather misleading picture about the comparative states of the two households. Hence, it is neces-sary to explicitly include the extent of vulnerability of households to various risks and the adverse outcomes accompanying them. Not only is the study of vulnerability to poverty important, it is also important to ensure that the outcome of the vulnerability that is being studied (in this case, poverty) is one that is
Exchange AD

q

z cP1

α

=

N

z

i=1

SPECIAL ARTICLEmay 17, 2008 EPW Economic & Political Weekly80Though there is no unique set of dimensions that are endorsed by all, it may be possible to arrive at a set of dimensions that can be deemed to be “core” in nature. In the case of more extreme situations of poverty, as is the case for developing countries, Martinetti (2000) points out that “it seems relatively easier to find a consensus about a minimum set of basic functionings such as to escape avoidable morbidity and premature mortality, to be nour-ished and sheltered, to receive a primary education and so on”. Empirical attempts at compiling multidimensional measures have done so with the aid of primary household level surveys. In an attempt to apply Sen’s capability approach in South Africa, Klasen [as quoted in Clark and Qizilbash 2003] uses information from a primary survey to narrow down on 14 dimensions or com-ponents of poverty. The survey revealed those needs and capa-bilities that ordinary South Africans think are basic (core and non-core dimensions) and where they draw the line between the poor and non-poor (critical thresholds). Similarly, Miceli (1998) uses consumption survey data to infer important dimensions and indicators of well-being in Switzerland. In a study conducted across three Indian states, Anirudh Krish-na (2005) finds that there was virtually no difference in the local understanding of poverty. When villagers were asked to demar-cate the stages of progress of a household out of poverty, the first three stages were invariably the same across all villages. Food was the highest priority item followed by shelter and basic cloth-ing. Repayment of debts also figured as an important indicator of movement out of poverty. Such studies can prove useful in identi-fying relevant core and non-core dimensions. It is crucial to note that each of the dimensions chosen to rep-resent poverty is relevant in the absolute sense. They are impor-tant in their own right. There is some absolute level of achieve-ment in each of these dimensions that are indicative of well- being. One must be careful not to view these dimensions as important only because they indirectly contribute to an enhance-mentin income or are indicative of high income. For example, a good education has its own intrinsic value and represents a basic level of capability for an individual. Hence education, by itself, directly contributes to the enhancement of an individual’s well-being. The educational achievement mustnot be valued for its contribution to earning capacity and the consequent improve-ment in well-being (i e, instrumental value). Such a logic can lead to a sort of circular reference and prompt restricting the poverty measurement to income alone because “achievements” in all the other dimensions would eventually reflect themselves in higher income. Therefore, each dimension must be seen as having an intrinsic value directly contributing to well-being. Indeed, this is not a dif-ficult proposition to make, as demonstrated through the choice of dimensions in the present study: a decent level of consumption to meet basic material and nutritional needs; access to safe and po-table water supply and proper sanitation facilities; a decent level of educational achievement; and a clean and healthy energy source for cooking are all important aspects of human well- being, each aspect having its own intrinsic value. Under the methodology used here, the concept of HCR and the extent of deprivation are used to measure both poverty and vulnerability. The highest possible level of achievement and the worst level of achievement in a given dimension are chosen from the actually observed achievements. The definitely poor are iden-tified as those who are farthest away from the highest possible level of achievement (or represent the worst level of achieve-ment). Unlike the HCR andFGT measures, which require an exter-nal minimum level to be identified (represented as z in the equa-tion), the measure of poverty used here takes the highest ob-served level of achievement in a dimension as the benchmark. So once those persons who are farthest away from the highest level are identified, the ratio of the number of such people to the total population (i e, the headcount ratio) is taken to represent the level of poverty. 1.3 Measuring VulnerabilityThe concept of vulnerability is interpreted in different ways by various disciplines. The vulnerability to poverty literature, as de-scribed by Alwang et al (2001) is conceptually weak and empiri-cally strong. Empirical studies attempting to quantify and meas-ure vulnerability to poverty are manifold. However, the concep-tual analysis of the notion has been inadequate and there exists no one systematic framework for analysis. Economists have inter-preted the concept of vulnerability to poverty from different per-spectives and consequently, several different frameworks have arisen to study the concept. A distinction in the studies on vulnerability can be made based on the approach they adopt. Vulnerability may be viewed as an individual’s exposure to being harmed or an unfavourable situa-tion. On the one hand, this exposure can be studied in terms of the extent of risk or the shock faced or it may be studied in terms of the outcome or end result. Empirical studies on vulnerability are carried out using either panel data models or error process models. Outcome based meas-ures of vulnerability are forward-looking and aim at constructing distributions of outcomes in various states of the world. Under such an approach, the vulnerability measure is often an expected value of some measure of poverty. Chaudhuri et al (2002) used cross section data and an error process model to estimate pre-dicted consumption values and thus, infer vulnerability. Pritchett et al (2000) define vulnerability as the probability that a house-hold will experience at least one episode of poverty in the near future and calculate a headcount rate to arrive at the proportion of households vulnerable to poverty. Using two panel data sets, they are able to arrive at a measure of vulnerability across gender and income groups over two periods. The above mentioned measures are outcome based in the sense that they view vulnerability through the lens of the out-come of vulnerability. The sustainable livelihoods literature and food security literature adopt a different approach. They focus on the ownership of assets and social security in affecting vul-nerability. Asset values are taken as proxies for vulnerabilities. The covariate risks attached with these assets are studied to fully understand their value in times of crisis. By their very nature, these indicators are highly contextual, require extensive survey information andhence, it is difficult to arrive at a generalised measure [Dercon 2001].
SPECIAL ARTICLEEconomic & Political Weekly EPW may 17, 200881Qizilbash uses the concept of vagueness in the measurement of poverty to motivate the measurement of vulnerability. In reality, there will always be individuals or households who cannot be unambiguously called “not poor” or “poor”. It is this ambiguity that Qizilbash (2002) uses to capture the concept of vulnerability. Those entities that are neither poor nor non-poor may be seen in terms of their closeness to being poor. The closer an entity is to being classified as unambiguously poor, the more vulnerable it is. So the degree of belonging to the set of poor can be used as a measure of vulnerability. It is in determining this degree of belonging that fuzzy logic proves useful. By translating linguistic statements such as “close to poor”, “moderately poor”, “moderately well to do”, degrees of belonging to the set of poor can be found. This degree is inter-preted as the measure of vulnerability. The higher the degree of belonging or membership to the set of poor, the greater the vul-nerability of that entity. Qizilbash (2002) uses such an interpretation of vulnerability to measure poverty and vulnerability across South African provinc-es using fuzzy logic. The uniqueness in his study is that he inter-prets the fuzzy membership value as being indicative of vulnera-bility to poverty. The numerical values of membership degrees are interpreted as being indicative of the extent of vulnerability. The present study adopts such an approach for measuring vul-nerability to multidimensional poverty and the next section de-scribes the exact methodology in detail along with the data used for the analysis. 2 Data and MethodologyAs mentioned earlier, the purpose of the present study is twofold – first, to arrive at a multidimensional measure of poverty and secondly, to analyse vulnerability to multidimensional poverty. The analysis is carried out with the states of India as units of analysis and the data used corresponds to two time points in the early and late 1990s. To capture multidimensional poverty and the vulnerability of entities to multidimensional poverty, a set of core dimensions are identified and representative indicators for each dimension are assembled for the two time points of analysis across 15 major states of India. For vulnerability measurement, fuzzy logic based approach suggested by Qizilbash (2002) is adopted. 2.1 Choice of Dimensions and IndicatorsFirst, the important dimensions of poverty are identified. These represent the basic capabilities that any individual requires. The dimensions chosen are consumption, education, sanitation, ac-cess to water, source of energy for cooking and dwelling. In the absence of a comprehensive primary survey to identify the basic needs of the individuals in different parts of India, the dimen-sions are chosen primarily to reflect the broad consensus from literature on what constitutes the well-being of people.The choice of indicators to represent the identified dimensions is often governed by the availability of data. As Qizilbash (2004) admits, “often the best defence of the actual indicators selected relates purely to the limits of data availability” and indeed this is the case for many indicators chosen in this study. Ideally, when studying poverty and vulnerability, the data used should be individual level information. NSS data, being sourced at the household level are ideal for poverty and vulnera-bility studies. Accordingly, the study uses published data from NSS reports for two time points 1993-94 and 1999-2000 for rural and urban India.1 Also, the indicators are chosen based on the criteria that they should be represented in terms of HCR. Besides representing the extent of deprivation, the other advantage of theHCR is that it facilitates comparison across states. By compar-ingHCR at various levels of achievement across states, a relative position of a state can be inferred.2 So for each indicator, the in-formation on the percentage of the population at different levels of achievement is collected.The identified indicators for chosen dimensions are as follows: consumption (monthly per capita consumption expenditure), ed-ucation (level of educational attainment), sanitation (type of la-trine used), energy use (primary source of cooking fuel), water source (primary source of drinking water) and dwelling (per capita floor area). All these indicators have categories, which represent levels of achievement in the given dimension. These categories are ordered in terms of increasing levels of well-being. For education, the chosen indicator, levels of educational at-tainment represent different levels of well-being. Those with no education constitute the worst-off category, up to primary school were the next best and so on. In the case of energy use, a household’s use of cleaner energy for its cooking purposes ensures an improved well-being of the members of the household, largely in the form of a cleaner and healthier environment. As a household moves up the energy lad-der, the well-being is improved due to lesser pollution load. The categories have been ranked in accordance with the energy lad-der ranking. The worst-off category is leaves/straw/firewood in 1993-94 and the firewood and chips in 1999-2000. Undeniably, these energy sources are far from clean and also require greater time and effort to collect. Thus, their use has an adverse impact on an individual or household’s well-being. Therefore, house-holds using firewood can justifiably be categorised as being the worst-off in the energy dimension. To represent the sanitation dimension, the type of toilet facility that households have access to is considered as an appropriate indicator. The category “no latrine” is identified as the worst-off and this is fairly commonsensical.In the case of drinking water, the source of drinking water is taken to be the indicator of the level of achievement. However, this indicator was available only for the first time period 1993-94. The distance from the principal source of drinking water is consid-ered as alternative indicator. But, it is observed that almost 90 per cent of the households had drinking water facilities well within their premises or at most 0.2 kms away. So the headcounts for this indicator were clustered largely around the upper two categories, forcing the dimension to be dropped from analysis. Per capita floor area is the indicator chosen to represent achievements with regard to shelter and personal space. Since in most developed countries, the average floor area per person is close to 20 square metres, households having per capita floor area below this norm are categorised as the worst-off category.

vulnerable population. The ratio of vulnerable to total population represents the HCR of vulnerability to poverty.

Once poverty and extreme vulnerability HCRs are obtained for each dimension, the next task is to provide a comprehensive overall multidimensional poverty ranking and vulnerability to multidimensional poverty ranking. The HCR of poverty and vulnerability poverty in each dimension is used to obtain relative ranking of the states. Aggregation over the dimensions is carried out with the help of simple Borda ranking in each state. The interpretation of the multidimensional poverty ranking is fairly simple to understand. The Borda ranking process employed implies that every dimension is given an equal weight and seen as contributing equally to overall well-being. If a state fares poorly in the ranking, it means that there is a relatively high percentage of its population living in a situation of extreme deprivation in various dimensions. In case of vulnerability measure similar interpretation is possible.

3 Results and Discussion

Based on the HCR of population that is identified as definitely poor and vulnerable in each dimension, the states are ranked in terms of poverty and vulnerability. The overall poverty and vulnerability ranking across states is arrived at using a simple Borda ranking procedure. The final rankings for multidimensional poverty (Pmd) and vulnerability to multidimensional poverty (Vmd) are compared with other similar indicators so as to check for their robustness. Available data on HDI of Indian states is used for such a comparison. Similarly, focusing on consumption dimension alone, income poverty (P) and vulnerability to income poverty

y

(V) are culled out to enable comparison with existing evidence

y

on income based poverty measures across Indian states.

3.1 Dimension Specific Poverty

Looking at the performance of each state across various dimensions can provide an in-

SPECIAL ARTICLE

proportion of poor higher in rural areas, it falls only marginally over the two time periods. Based on their performance across the dimensions and the time points, the states are classified into groups (Figure 2).

One well performing dimension, significant improvement in at least one dimension over time: Andhra Pradesh, Gujarat, Haryana, Karnataka, Maharashtra, Punjab, Rajasthan and Tamil Nadu witnessed an improvement over time with a fall in the population in the definitely poor category in consumption and education. The improvement is most pronounced in the education dimension.

One critical dimension having poor performance with no improvement, uniform improvement across other dimensions: Assam and Kerala exhibited dismal performance, especially in the energy dimension, with only a marginal improvement across the two periods. The performance in all the other dimensions is much better and it improved over time.

All dimensions critical/worsening in performance: Rural Orissa witnessed deterioration in performance in the consumption dimension whereas in Madhya Pradesh, performance remained stagnant in that dimension. In the remaining states of Uttar Pradesh, West Bengal and Bihar, all the dimensions remained critical with a high proportion of the population still in the definitely poor category.

3.1.2 Urban India

Urban areas fared relatively better than rural areas across all dimensions, and also exhibited significant improvement over time. Based on their performance across the dimensions at the two points in time, the states are classified into groups (Figure 3, p 84).

Uniform improvement in all dimensions across time: For these states, the HCR of the population in each of the dimensions reduced uniformly over time, with a very small proportion of the population in absolute poverty. Graphically, these are the states

Figure 2: Poverty Over Time, Across Dimensions: Rural

sight into what dimen-

Group 1 Group 2 Group 3

sions are contributing to

Andhra Pradesh S Gujarat S Assam S Bihar S Orissa S

1

the overall well-being.

Headcount ratios of the definitely poor are compared across dimensions for a given state at the two points in time. The headcount ratios are represented as a proportion of the population and hence will range between 1 and 0.

3.1.1 Rural India

In rural areas, all states perform poorly in the sanitation and energy use dimensions, with the exception of Kerala and Assam (in the sanitation dimension). Not only is the

C

Haryana

C

Karnataka

C

Punjab

C

1.00

1.00

1.00

1

0.50

0.50

0.50

0.5

0.5 Ed 0

C 00.00

Ed C 0.00 Ed C 0.00

Ed

C Ed

En En En En En

S Rajasthan S

1.00

1.00

0.50

0.50

Kerala S

1.00 Madhya Pradesh Uttar Pradesh

Ed C 0.00

Ed S S

0.00

1 1

0.50

0.5 0

C 0.00 Ed 0.5

Ed C

Ed S S

En En C 0

Maharashtra

1.00 En

1.00

0.50

0.50 En

En

0.00

0.00

EdC Ed

Key 1993-94

West Bengal S

1

En En

S Tamil Nadu S

1.00 0.5

1999-2000

1.00

0C Ed0.50

0.50

0.00

Ed C 0.00

Ed C – Consumption dimension S – Sanitation dimension En Ed – Education dimension En – Energy dimension

En En

Source: Author’s own calculations.

Economic & Political Weekly EPW may 17, 2008

SPECIAL ARTICLE
having a sort of “shrink-Figure 3: Poverty Over Time, Across Dimensions: Urban

Group 1 Group 2 Group 3

ing diamond” where the

Andhra Pradesh Gujarat S Assam S Madhya Pradesh S Haryana 0.50 S

0.50 0.50

0.50

lighter shaded area (1999-0.50 S

0.25

0.25

0.25

2000) is wholly encom-0.25 0.25

0.00

0.00 C Ed C

Ed C

Ed C Ed

0.00 0.00

C Ed

passed within the darker 0.00 area and the shift inwards is seen across all the axes.

En En En En En

Karnataka Uttar Pradesh Kerala West Bengal S Rajasthan S

S 0.50

0.50

0.50

The states that witnessed S S

0.50

0.50

0.25

0.25

0.25

such an across-the-board 0.25

0.25

C

0.00

0.00

0.00

EdEd C 0.00 Ed C

C

Ed

improvement over time C 0.00

Ed

are Andhra Pradesh, Gujarat, Karnataka, Punjab, En

En En En En

Punjab S Taml Nadu S Bihar S Maharashtra S

Tamil Nadu and Uttar 0.50

0.5

0.5

0.50

Pradesh. 0.25

0.25

0.25

0.25

0.00

Ed C Ed C

Ed C 0 Ed

0

Poor performance in C 0.00

one dimension coupled

Key 1993-94

with low improvement En En

En En

Orissa

over time in that dimen-Source: Author’s own calculations. S 1999-2000

0.5

sion: Graphically, such

0.25 states are C

Ed C – Consumption dimension by a “shrinking comet” S – Sanitation dimension Ed – Education dimension

represented 0

shape. Kerala and Assam

En – Energy dimension En

exhibit such a pattern. These states have a very low proportion in definite poverty in all human development index (P) for the years 1993-94 and

HDI

the dimensions, except in the energy use where they have a rela-1999-2000. tively high proportion in definite poverty. This is similar to the The results are discussed through several comparisons of the trend observed for these states in rural areas. rankings and these include: (a) multidimensional poverty versus

No change/worsening in a few or all dimensions: For Madhya income poverty; (b) multidimensional poverty versus HDI; Pradesh, Haryana, Rajasthan and West Bengal, there was no (c) vulnerability to multidimensional poverty versus vulnerabilichange in the performance in at least one dimension. In the case ty to income poverty; and (d) vulnerability to multidimensional of Madhya Pradesh and Haryana, there is no change in the per-poverty (at present) versusmultidimensional poverty (in future). formance in the sanitation dimension, whereas for Rajasthan and West Bengal, stagnancy is observed in the consumption dimen-Multidimensional Poverty versus Income Poverty: The first set sion. Bihar, Orissa and Maharashtra witnessed an adverse trend of results relates to the multidimensional poverty rankings (Pmd). with an increase in the definitely poor population. For Orissa and These rankings are compared with the rankings obtained using Maharashtra, this is seen with respect to the consumption dimen-the conventional income based approach. The comparison reveals sion, whereas for Bihar, both the sanitation and energy dimen-that there are considerable disparities. Most of the conventionsions saw a deterioration in performance. ally top-ranked states fall in their relative rankings when meas

ured using Pmd. This is especially so in the case of rural Andhra

3.2 Multidimensional Poverty

Table 2: Relative Ranks of States in Rural Areas Pradesh, which falls from an income

and Vulnerability States 1993-94 1999-2000based ranking of 2 to a Pmd ranking P P P (1991) V V P P V V

ymdHDIymdymdymd

While the dimension-specific poverty of 4 in 1993-94 and from 4 to 11 in

Andhra Pradesh 2 7 9 14 11 2 11 10 8

assessment is indicative of well- 1999-2000. Deterioration is also seen

Assam 12 11 15 5 15 10

1510 11

being, it may not provide the neces-in the case of Rajasthan. However, for

Bihar 15 8 1 13 12 12

11 1314

sary overview. For comparison, the Kerala, the reverse is observed. Once

Gujarat 3 6 6 5 9 3 3 4 6 overall poverty and vulnerability Haryana 5 2 4 1 6 5 4 2 4 multiple dimensions are introduced, results are presented in the form of Karnataka 7 4 8 12 13 7 11 9 1 Kerala’s ranking improves consider

relative ranking across 15 states. Kerala 8 10 1 7 14 4 1 3 2 ably except in rural Kerala in 1993- Table 2 and Table 3 (p 85) show the Madhya Pradesh 10 14 15 13 12 13 14 13 9 94. This is in keeping with what was Maharashtra 9 1 5 3 2 9 7 6 7

overall results for rural and urban ar-observed when the HDI, also a multi-

Orissa 14 12 10 15 10 15 15 7 11

eas respec tively. Each table shows the dimensional measure, was used. Ma-

Punjab 1 3 2 2 4 1 2 1 3

relative ranks of states as per income harashtra and Assam fare much better

Rajasthan 4 13 12 4 5 6 9 5 13

poverty (P), vulnerability to income when poverty is extended to include

y

Tamil Nadu 6 4 3 11 7 8 8 8 5

pov erty (V), multidimensional pov-dimensions other than income.

y

Uttar Pradesh 11 7 14 6 3 12 10 11 14

erty (Pmd), vulnerability to multi-However, at the bottom end of the

West Bengal 13 9 7 9 8 10 6 14 15

dimensional poverty (Vmd), and Source: Author’s own calculations. rankings, the comparison shows that

84 may 17, 2008 EPW Economic & Political Weekly

SPECIAL ARTICLEEconomic & Political Weekly EPW may 17, 200885there is not much difference in the list of states that are identified as income poor and multi-dimensionally poor. In rural India, Bi-har, Orissa and Madhya Pradesh are consistently among the worst performing states. In urban India, Orissa, Madhya Pradesh, Uttar Pradesh and Andhra Pradesh are identified as income as well as multidimensionally poor. This result is in keeping with other similar observations [Kapur et al 2002]. The similarity in the rankings among the poorer states under both the measures suggests that be-ing income poor implies being multi-dimensionally poor but not being in-come poor does not imply being multidimensionally poor.If income is viewed as a means to an end, then not having income implies not having the means of achieving ba-sic capabilities or at least being severe-ly restricted in the extent of achieve-ment possible. At the same time, an individual can still have income, and not achieve basic capabilities because he/she is constrained in terms of ac-cess, opportunities, quality of the achievement and other factors. So merely having income does not take a household/individual out of poverty. Income is only the means and it must be supplemented by other factors such as access, opportunities, etc. The rank correlation between the two measures moves up from 0.50 in 1993-94 to 0.68 in 1999-2000 in rural India, where-as it drops from 0.90 in 1993-94 to 0.66 in 1999-2000 in urban India. The increase in correlation between two measures in rural India is perhaps indicative of increased integration of the rural economy with the market and the consequent monetisation of the economy. On the other hand, the decline in correlation be-tween the two measures in urban India could be a consequence of increased government intervention in the urban economy un-dermining the role of money. Another possible explanation is the increased role of subjective perceptions and values, aspects which cannot be captured by income. Multidimensional Poverty versus HDI: For the purpose of comparison and in order to check for the robustness of the multidimensional poverty rankings, the HDI rankings for the states were also considered. However, for 1999-2000, the dis-aggregate HDI rankings, i e,HDI ranks for urban and rural India, were not available and hence, this period has not been considered. In 1993-94, for both rural and urban India, the cor-relation between theHDI and multidimensional poverty rankings is reasonably high at around 0.60. This correlation is higher than that between income poverty and HDI rankings. This is as would be expected – a unidimensional measure is further away from theHDI than a multidimensional measure. However, the fact that the correlation is not perfect (i e, exactly equal to 1) implies that a multidimensional measure may indeed capture information that would have been overlooked by a measure such as the HDI. Thus, this warrants the need for an alternative multidimensional measure of poverty. Vulnerability to Multidimensional Poverty versus Vulner-ability to Income Poverty: If poverty can be viewed in the in-come/consumption space alone, then vulnerability to poverty can also be limited to this space. A dimension-specific vulnerabil-ity measure would reflect the likelihood of that entity falling into poverty in that dimension at a later period. Looking at an entity’s vulner-ability to poverty in the income space vis-à-vis her overall multidimension-al vulnerability, one can understand whether an entity’s income vulner-ability is indicative of their overall vulnerability. While in rural India the two vulnerability ranks had a stable correlation of approximately 0.60 in both periods, the correlation in urban India was fairly unstable with very low correlation in 1993-94 and rela-tively high correlation in 1999-2000. Thus, it may be inferred that vulner-ability to income poverty may not be indicative of an entity’s vulnerability to multidimensional poverty – states that are not vulnerable to income poverty may still be vulnerable to poverty in other dimensions. Vulnerability to Multidimensional Poverty (Present) versus Multidimensional Poverty (Future): Since vulnerability is an in-dicator of the likelihood of falling into poverty, states that are iden-tified as vulnerable currently could be expected to fall into poverty unless preventive measures are implemented. However, the time taken for vulnerability to manifest in poverty is not known and will depend on a number of factors. Keeping this caveat in mind, a comparison between the vulnerability rankings in the initial period (19993-94) with the multidimensional poverty rankings in the subsequent period (1999-2000) would be insightful. Did those states that were identified as highly vulnerable fall into poverty over time? This can be seen by comparing Vmd rankings in 1993-94 with Pmd rankings in 1999-2000.Figures 4 and 5 (p 86) show the comparison between the vul-nerability rankings and multidimensional poverty rankings for rural and urban areas respectively. Taking vulnerability rankings in 1993-94 on the x-axis and poverty rankings in the later period (1999-2000) on the y-axis, these figures showcase whether states identified as vulnerable indeed turn out to be poor. It may be noted that in the following discussion, states with a rank that is above eight are labelled as worst performing states. The first quadrant represents states that were identified as vulnerable and did fall into poverty in the second period (true positives). The second quadrant represents all states which were identified as not vul-nerable but fell into poverty in the next period (false negatives). Quadrant three represents the true negatives – states that were identified as not vulnerable and did not fall into poverty in the Table 3: Relative Ranks of States in Urban AreasStates 1993-941999-2000 Py Pmd PHDI (1991) Vy Vmd Py Pmd Vy VmdAndhra Pradesh 15 13 12 7 14 6 13 9 6Assam 1 2 5 14 5 1 1 1310Bihar 8 8 14151211141013Gujarat 7 4 7 6 3 5 2 5 7Haryana 4 7 3 9 2 3 5 3 4Karnataka 1111 8 2 1 8 126 5Kerala 5 5 1 12137 4 4 2Madhya Pradesh 14 13 11 5 15 14 7 8 12Maharashtra 9 5 6 1 6 12151 1Orissa 1315134815111414Punjab 2 1 2 10 4 2 5 2 3Rajasthan 6 9 10 13 7 10 9 1511Tamil Nadu 12 12 4 3 10 9 7 7 8Uttar Pradesh 10 10 15 11 11 13 10 11 15WestBengal 3 3 9 8 9 4 3 129Source: Author’s own calculations.
DRANT 2 DRAN

Orissa

Maharashtra

Madhya Pradesh

Bihar Uttar Pradesh Rajasthan Tam Madhya Pradesh il Nadu Karnataka Andhra Pradesh
Maharashtra
West Bengal
Haryana
Gujarat
Punjab Kerala
Karnataka Raj asthan Orissa Andhra PradUttar Pradesh Bihar esh
Punjab Haryana Gujarat West Bengal Kerala Tamil Nadu
SPECIAL ARTICLEEconomic & Political Weekly EPW may 17, 200887Madhya Pradesh/Orissa, rural Uttar Pradesh/Andhra Pradesh), there were several states that were not poor but were at the same time highly vulnerable to poverty (for example, urban Andhra Pradesh/Karnataka/Gujarat and rural Kerala/West Bengal). It is the existence of such states that warrants the need for a measure of vulnerability. These states, by their non-poor status, would not be the focus for policymakers. At the same time, given their vulnerable status, ideally such states should be given particular attention to ensure that they do not slip into poverty. Measures to prevent poverty would be the priority for policymakers in such states. Therefore, the measures of poverty and vulnerability are in-deed insightful because of the additional information they provide, information that would have been overlooked if traditional one dimensional measures of poverty had been used. Though these measures are in terms of ranks and relative in nature, they still go a long way in painting a fuller and more comprehensive picture of poverty and the vulnerability of poverty of a set of entities. Notes1 The NSS reports used were sourced from the web-page of the Ministry of Statistics and Programme Implementation, http://mospi.nic.in/stat_act_t14.htm 2 The use of the HCR is not meant to be an endorse-ment of this measure over all others such as the income-gap ratio and the squared income poverty gap mentioned previously. 3 The choice of 0.7 as a threshold is purely arbitrary and another appropriate cut-off may also be con-sidered.ReferencesAlwang, J, P B Siegel and S L Jorgensen(2001): ‘Vul-nerability: A View from Different Disciplines’, Social Protection Discussion Paper, No 0115, World Bank.Clarke, D and D Hulme (2005): ‘Towards a Unified Framework for Understanding the Depth, Breadth and Duration of Poverty’,Global Poverty Research Group and Institute for Development Policy and Management, University of Manchester, UK.Clark, D A and M Qizilbash (2003): ‘Core Poverty and Extreme Vulnerability in South Africa’, Discus-sion Paper 2002-03 Economics Research Centre, University of East Anglia.Chaudhuri, S, J Jalan and A Suryahadi (2002): ‘As-sessing Household Vulnerability to Poverty: A Methodology and Estimates for Indonesia’, Dis-cussion Paper 0102–52, Department of Econo-mics, Columbia University.Dercon, S(2001): ‘Assessing Vulnerability to Poverty’, Jesus College and CSAE, Oxford University. Foster, J, J Greer and E Thorbecke (1984): ‘A Class of Decomposable Poverty Measures’,Econometrica, Vol 52, No 3, pp 761-66.Hulme, D and A Mckay(2005): ‘Identifying and Meas-uring Chronic Poverty: Beyond Monetary Meas-ures’, paper presented at CPRC-IIPA seminar on ‘Chronic Poverty: Emerging Policy Options and Issues’, draft. Kapur, Aasha, Ramakrushna Panigrahi and Sashi Sivramkrishna (2002): ‘Operationalising Multidimensional Concepts of Chronic Pov-erty: An Exploratory Spatial Analysis’, paper presented at the Research Design Workshop for Exploring Appropriate Solutions to Chronic Poverty, IIPA.Krishna, Anirudh (2005): ‘Poverty Knowledge and Poverty Action, Evidence from Three States in India’.Ligon, E and L Schechter (2003): ‘Measuring Vulner-ability: The Director’s Cut’, Economic Journal, Vol 113, No 486, pp 95-102. Martinetti, C (2000): ‘A Multidimensional Assess-ment of Well-being Based on Sen’s Functioning Approach’, University of Pavia, Italy. – (2004): ‘Complexity and Vagueness in the Capability Approach: Strengths or Weakness?’ University of Pavia, Italy. Miceli, D (1998): ‘Measuring Poverty Using Fuzzy Sets’, Discussion Paper No 38, National Centre for Social and Economic Modelling (NATSEM), Uni-versity of Canberra. Nandy, Ashis (2004): ‘The Beautiful Expanding Future of Poverty – Popular Economics as a Psy-chological Defence’,Economic & Political Weekly, Vol 39, No 31, pp 94-99.Pritchett, L, A Suryahadi and S Sumarto (2000): ‘Quantifying Vulnerability to Poverty: A Proposed Measure, with Application to Indonesia’, Social Monitoring and Early Response Unit (SMERU), Working Paper.Prowse, M (2003): ‘Towards a Clearer Understanding of ‘Vulnerability’ in Relation to Chronic Poverty’, CPRC Working Paper No 24, Chronic Poverty Re-search Centre,University of Manchester. Qizilbash, M (2002): ‘A Note on the Measurement of Poverty and Vulnerability inthe South African Con-text’,Journal of International Development, Vol 14, No 6, pp 757-72. – (2004): ‘On the Arbitrariness and Robustness of Multidimenstional Poverty Rankings’, Research Paper No 2004/37, WIDER. Sen, A (2000):Development as Freedom, Oxford Uni-versity Press.WDR (World Development Report) (2000-01): ‘Attacking Poverty’, World Bank, Oxford University Press.REVIEW OF WOMEN’S STUDIESApril 26, 2008Between Public and Private Morality – Maithreyi KrishnarajThe 'Virtuous Woman': Law, Language and Activism – Kanchana MahadevanBody as Space, Body as Site: Bodily Integrity and Women's Empowerment in India – Kanchan MathurEthnographies of the Global Information Economy: Research Strategies and Methods – Carol UpadhyaGender Budgeting in Disaster Relief: Need for a New Methodology – Meenakshi ThoratMainstreaming Gender, Engendering Development: Reflections on a Case Study – Padmini Swaminathan, J JeyaranjanTheReview of Women’s Studies appears twice yearly as a supplement to the last issues of April and October. Earlier issues have focused on: Gender in Medical Education (April 2005); Gender Budgeting(October 2004); Women, Work and Family (April 2004); New Challenges (October 2003); Women, Work, Markets (October 2002); Women and Education (April2002); Women: Security and Well-Being (October 2001); Women and Philosophy (April 2001); Reservations and Women’s Movement (October 2000); Women, Censorship and Silence (April 2000); Women and Ageing (October 1999); Gender Inequities: Focus on Tamil Nadu (April 1999).For copies write to Circulation ManagerEconomic and Political Weekly320-321, A to Z Industrial Estate, Ganpatrao Kadam Marg, Lower Parel, Mumbai 400 013.email: circulation@epw.in

To read the full text Login

Get instant access

New 3 Month Subscription
to Digital Archives at

₹826for India

$50for overseas users

Comments

(-) Hide

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

The Indian labour market is characterised by a high level of informality, with large numbers of workers in poorly paid “lower tier” informal jobs...

Geoffrey Bawa (1919–2003) was Sri Lanka’s most celebrated architect in the 20th century and his half-a-century long career shaped the nation’s...

Whether the “practising Adivasi” or the practitioners of traditional knowledge are subjects of different rationality is examined here. Through a...

The Indian tea economy is undergoing acute transformations, with the divestment of tea companies from plantations leaving thousands of plantation...

Firms can avoid taxes legally, even though it is well understood that tax payment is a fundamental and measurable behaviour towards society. In...

How the pattern of inequality in maternal healthcare service utilisation has evolved after the adoption of the National Rural Health Mission in...

Following the announcement of demonetisation on 8 November 2016, India saw the withdrawal of nearly 86% of the cash in circulation. This caused...

The evolving COVID-19 pandemic requires that data and operational responses be examined from a public health perspective. While there exist deep...

Although there are multiple vulnerabilities that may prevent access to maternal and child health services in India, the literature has so far...

The implications of the recommendations of the Fourteenth Finance Commission on finances of Bihar as a result of changes in the tied, untied and...

Back to Top