Multidimensional poverty indices must consider crucial insights inherent in the capabilities approach propounded by Amartya Sen.
According to the UN Development Programme Multidimensional Poverty Index published in 2019, in collaboration with the Oxford Poverty and Human Development Initiative, India is to home to about 369 million poor people, the largest number of people in a country. The report also mentions that India managed to lift about 270 million people out of poverty between 2006 and 2016. However, as a consequence of the COVID crisis, India runs the risk of 260 million of its people, who are currently at the brink of the poverty line, being pushed into the state of poverty. As such, India would need a strategy comprising anti-poverty interventions. Fundamental to such interventions is a clear and nuanced understanding of the phenomenon on the ground. There is a need to comprehend the character of poverty as it plays out at the granular level. This demands a methodology that can measure poverty in a manner that approximates ground realities. Does the most recent method applied by Indian policymakers satisfy this criterion?
The methodology proposed by Dr. C. Rangarajan-led Expert Group endorsed the importance of determining the normative requirements not just of calorie consumption and along with it of protein and fat intake, but also of the essential non-food components such as clothing, house rent, conveyance and education. The poverty line is an aggregate of the prices of all of these normative levels as well as the prices of the behaviourally determined levels of other non-food expenses in the year 2011-12. The data utilised for this was the NSSO household consumption expenditure of 2011-12.
To define the nutritional benchmarks in terms of calorie, protein and fat intake, the Expert Group worked with the ICMR norms categorised by age, sex and activity for all-India rural and urban areas. It was found that these norms were achieved in the sixth fractile (25-30 percent) of the class distribution of nutrient intake in the rural areas, and in the fourth fractile (15-20 percent) in the urban areas. The expenditure norm of food consumption is the average monthly per capita expenditure on food in these fractile classes of INR 554 in the rural areas and INR 656 in urban areas.
However, there are limitations of this method. It is very much possible that people might want to subtract from their nutritional needs to address some other pressing issues. Although the money is spent, the nutritional adequacy is not acquired. Another issue relates to the absence of a protocol to make substitutions between calories, proteins and fats. As such, one might be spending less than the requisite amount, signalling deficiency. This would lead the calculation of the headcount ratios of poverty haywire.
As far as the essential non-food components are concerned, the Expert Group concedes that it is difficult to demarcate the normative benchmarks for these components. The solution espoused by the Group is to treat observed expenditures by households on these items in the median fractile (45-50 percentile) as the norm. What could be the rationale underlying this step? The median fractile represents the most frequent or recurring levels of expenditure which would be incurred by households. Hence, it would appear to reasonably represent normative expenditure levels of a large number of people. But, in a poor economy like India, the observed expenditure of the median fractile might very well be an underestimate. The ideal way of determining normative levels of expenditure would be to commission government agencies to undertake a comprehensive scientific study or survey that would draw up reliable estimates of these expenditures.
Is there a credible alternative to the above-mentioned methodology of poverty measurement? Although there is a decent consensus on the need of the multidimensional character of poverty measurement, such a consensus is absent when it comes to considering a unique poverty index. Several multidimensional indices have been developed. For example, the Alkire-Foster Index, the Datt Index, the Chakravarty, Mukherjee and Ranade Index; the Chakravarty, Deutsch and Silber Index; the Tsui Index, the Bourguignon and Chakravarty Index and so on. These indices have been appraised based on their adherence to certain axioms. This axiomatic framework ensures cogency and consistency of such indices. Although indices obeying the axioms end up being sound, do they reliably approximate the behaviour of poverty as it plays out on the ground? There is no straight answer to this question nor even a well-developed set of conditions that enable us to answer this question. Nevertheless, there are certain nuances implicit in Sen’s capabilities approach, which, if operationalised by poverty indices, will definitely improve their correspondence with ground realities. Unfortunately, these nuances have seldom been incorporated in the design of the various multidimensional poverty indices mentioned above. It is the aim of the discussion that follows to highlight these nuances.
Sen delineates two types of freedoms. The first type of freedoms is referred to as the constituent component of development, which are significant because of their intrinsic value i.e. their own value to the process of development. Some of these freedoms are as fundamental as good health and education, sanitation and clean drinking water. The second type of freedoms are those instrumental to development such as political freedoms, social opportunities, economic facilities, transparency guarantees and protective security. These freedoms are instrumental to the attainment of other significant freedoms. For Sen, poverty is an ‘unfreedom’ and is reflected in the deprivation of both these forms of freedoms. Multidimensional poverty indices mentioned above are capable of measuring deprivation only in terms of their intrinsic value. The instrumental role of freedoms which represents the achievement of other freedoms is not accounted for by the design of the any of the above-mentioned poverty indices.
The capabilities approach has coined two terms which are fundamental to it: ‘capabilities’ and ‘functionings’. While functionings refer to the conditions of being and doing, and representing actual attainments for example, being well-educated, owning a house and a vehicle, etc. Capabilities, on the other hand, refers to the ‘ability’ to be someone or do something. It refers to the set of functionings to which one has effective access to, but one might ‘choose’ not to access the functioning. A girl belongs to a family which has the capability to educate her, but chooses not to do so. This is an example of having a capability, but not converting it into a functioning.
A measure of poverty must be fungible or flexible in terms of what is sought to be evaluated, the functioning or the capability. The poverty indices mentioned above assume that the what is evaluated is quantitative, can be calculated numerically and that a threshold for the same can be well defined. As such, these measures always end up assessing functionings when in fact capabilities is more appropriate for being assessed. Why is fungibility of the poverty measure desirable? For example, despite being able to afford a vehicle, one might choose not to purchase one and travel by public transport. Here, the capability should be considered. In certain cases, functionings become more relevant than capabilities. For instance, the decision to realise one’s capability may go against social norms. For instance, one may choose not to build a toilet in their home even though it is socially desirable to do so. Here, the functioning must be considered. The poverty index must enjoy this fungibility of what to measure.
The capabilities approach advocates that the evaluation of poverty and deprivation must be a social choice exercise. Such evaluation is expected to be a product of reasoned consensus following from ‘public debate and discussion’ and ‘democratic understanding and acceptance.’ When poverty measurement is attempted as a social choice exercise, it can yield several advantages. For example, the author, in her attempt to build a new multidimensional poverty index discovered four such advantages. The need to determine arbitrary poverty line becomes redundant. Since there is no global or even national consensus on what poverty line is the most appropriate, arbitrary poverty lines is a major problem solved. Another arbitrary element of the poverty indices is the weights assigned to different attributes in terms of which deprivation is measured. Making poverty measurement a social choice exercise allows weights to be constructed in a logical manner rather than determining them arbitrarily. It also allows categorical variables to be considered in poverty estimation.
By making poverty a social choice exercise, the poverty index can take into consideration not just the command over resources of different individuals, but also their ability to convert the same into real achievements. For example, a billionaire who has contracted an incurable disease cannot be cured despite his resources because no cure exists, and he will eventually die.
The key recommendation of the above discussion is that multidimensional poverty indices must consider all of the above-mentioned insights inherent in the capabilities approach propounded by Amartya Sen. It is incumbent to devise an index that makes all forms of arbitrariness redundant and employs well-reasoned methods of assessment to arrive at results. Such an index is likely to generate the much-needed consensus on the methodology of poverty appraisal and eventually help India to prepare a nuanced strategy to control any sudden spike in poverty levels on account of the pandemic.
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.
Renita DSouza is a PhD in Economics and a Fellow at Observer Research Foundation Mumbai under the Inclusive Growth and SDGs programme. Her research focus ...Read More +