A recent discussion paper released by NITI Aayog studies the decline in the number of multidimensionally poor in India and states that the headcount poverty ratio (proportion of population that is poor) reduced from 55.3% in 2005-06 to 24.9% in 2015-16 and further down to 15% in 2019-21.
Using a compounded annual rate of change, the change in the previous period, i.e. 2005-06 to 2015-16, was slower (-7.69%) than that during 2015-16 to 2019-21 (-10.66%).
The discussion paper extrapolates the multidimensional poverty index (MPI) for 2014-15 and 2022-23 and argues that almost 25 crore people have escaped poverty during this period.
Although these projections have no basis, especially considering that during this period, the country has faced major health and economic shocks because of the COVID-19 pandemic, there is no doubt that there has been an improvement in many of the indicators included in the MPI over the last 20 years. But the question remains: is this sufficient to claim not only that poverty has declined but that it has done so at a faster pace during the two terms of the present government?
Understanding MPI
Inspired by Amartya Sen’s theory of capabilities and functioning, the HDI (Human Development Index) and MPI expand focus away from the ‘means’ (income) to ‘ends’ (expanding choices/capabilities). This has been an important development as it enables policy discussion to also move from a single-minded focus on increasing GDP to looking at the outcomes for people and their lives.
GDP growth or higher per capita incomes do not necessarily always translate into better education, nutrition or health outcomes. This is best seen in the case of India compared to neighbouring Bangladesh or Nepal, which have lower per capita incomes, but better human development indicators.
Even within India, the ranking of states by per capita income does not exactly match the ranking by human development indicators. Southern states, particularly Kerala and Tamil Nadu, have been known to have achieved relatively better health and education outcomes through progressive social policies.
Also read: ‘Hunger, Undernutrition Stalking India; Placed Worse Than Least Developed Nations’: Prabhat Patnaik
The MPI measures deprivations across health, education and living standards with equal weightage being given to these three dimensions. Health includes nutrition and child and adolescent mortality; education includes years of schooling and school attendance; and standard of living includes cooking fuel, sanitation, drinking water, electricity, housing, assets, and bank accounts. The indicators within each dimension are given the same weight. Therefore, for example, the education dimension has two indicators of deprivation, each of which has a weight of 1/3*1/2 = 1/6.
Every household is given a deprivation score based on whether the household is deprived on the basis of these chosen indicators (1 if deprived and 0 if not), and then the weighted total in percentage terms is the deprivation score. If the deprivation score of a household is more than 33%, then this household is identified as being multidimensionally poor. The proportion of multidimensionally poor (weighted by household size) gives the head count ratio or the proportion of poor in the country.
Simply put, since each household is given a score based on their deprivations across dimensions, this methodology also allows us to understand the intensity of poverty at the household level.
India’s National MPI (NMPI) uses the same methodology, but has added two further indicators – maternal health in the ‘health’ dimension and financial inclusion in the ‘standard of living’ dimension. As with any multidimensional index, there can be a valid debate on which indicators are included and what the method for aggregating them is.
In the case of the NMPI, for instance, the method of aggregation results in child and adolescent mortality being given a weightage of 1/12, whereas school attendance has the weightage of 1/6 and having a bank account is weighted 1/21. Without going into the merits or demerits of the weighting matrix, the issue is that the limits of what the MPI indicates must be understood clearly.
Similarly, in the case of choice of indicators, as has been shown by others, the MPI includes both input and outcome indicators. Therefore, while we have a household being deprived if there was a child or an adolescent death in the last five years (outcome), another indicator is on whether the family has electricity supply (input). If one looks at other inputs, such as the supplementary nutrition or school meals provided to children, then there is a deterioration given the decline in Union budgets for these programmes over the last ten years.
Further, even along a single dimension, it needs to be understood that the MPI does not give a full picture. For example, by the nutrition indicator, a household is deprived if a child under five or an adult man or a woman is undernourished as defined by anthropometric indicators. Therefore, even though the prevalence of anaemia has increased between NFHS-4 and 5 (2015-16 to 2019-21), the nutrition indicator shows an improvement. The MPI also does not tell us anything about the quality of improvements. So, we can have the education indicator improving rapidly, while the learning outcomes as shown by repeated ASER data remain appalling.
There are a number of other issues related to the methodology that have been actively discussed in the literature around measurement of poverty and development. Without going into those details, it is broadly understood that the MPI adds value to our understanding of poverty and development by taking attention away solely from money-based indicators such as per capita incomes or consumption expenditure to actual achievements, such as being educated or being healthy.
In recent years, MPI has been included as one of the indicators of development in the UNDP’s Human Development Reports (HDR) and the NMPI is the indicator for one of the targets of the SDG Goal One on ending poverty. Based on this, it is being claimed that India is on the path to meeting this goal. SDG Goal One also includes targets on the proportion of population living below the national and international poverty lines, for which we have no data.
While MPI takes us forward as it draws attention to a multidimensional understanding of poverty – poverty not just in terms of incomes or consumption expenditure but across different dimensions, tracking individual indicators is critical for policy purposes and the limits of any multidimensional indicator should be understood.
Also read: Economists Say NITI Aayog Claims on Poverty Reduction Distort the Truth
MPI and consumption expenditure: complementary and necessary
Moreover, by giving up on income and consumption indicators entirely and only focussing on MPI to measure poverty is also not appropriate. Household incomes are difficult to measure accurately in a developing country like India, where there is a large informal sector with irregular and multiple sources of income. Consumption expenditure has been used as a proxy for income.
Consumption expenditure gives an estimate of how the purchasing power of households is changing and can indicate the immediate short-term impact of macroeconomic trends. An increase in or slow decrease in poverty based on consumption expenditure, along with increasing national income, indicates that the growth is unevenly distributed and is not reaching the poorer sections. At the household level, income is still an important determinant (although not the only one) of access to most goods and services, and hence human development outcomes as well.
In India, poverty was usually measured based on households that had a monthly per capita consumption expenditure (MPCE) below an officially designated poverty line. The level at which the poverty line is set has been a matter of great public debate, with various committees having been set up to update the poverty line. The most recent was the Rangarajan committee which submitted its report in 2014. It is, however, not clear whether India now even has an official poverty line.
The last available official consumption expenditure data on which poverty estimates are based is for 2011-12. Although NSS conducted a consumption expenditure survey in 2017-18, this report was junked by the government without adequate explanation. A leaked version of the report showed that the consumption expenditure of the poor had fallen, probably indicating an increase in poverty in the period immediately following demonetisation.
Consumption expenditure data can also throw light on the current discussion on whether the revival in growth is V-shaped or K-shaped as well as inequality by class and social group, change in patterns of food and non-food consumption, the effect of expanded public distribution system, or PDS, on food consumption and many more important issues. Further, including consumption expenditure-based poverty ratios in our analysis allows for a longer-term analysis of poverty in the country, whereas the MPI, which is based on the NFHS data, is only possible for the last couple of decades.
Also read: ‘Over 74% Indians Unable to Afford Healthy Diet’: UN Report
More data, better analysis
Based on the MPI alone, claiming that all the recent initiatives of the government, ranging from Anaemia Mukt Bharat (anaemia is not even included in the MPI!) to free grains in the PDS (introduced only in 2020 after half the NFHS survey was completed), as is being done in various reports and press releases is a bit far-fetched. These programmes have to be analysed in their entirety, taking into account the fact that budgets for Poshan Abhiyan and other schemes have actually been slashed and PDS coverage has remained stagnant, based on Census 2011 population numbers.
On the other hand, improvements in toilet access and bank accounts are easier to link to direct government schemes. Reviving India’s data ecosystem by conducting the Census, consumption expenditure survey and others would be the first step towards an honest analysis of what worked and what didn’t.
Dipa Sinha is faculty at Dr. B.R. Ambedkar University, Delhi.