National Family Health Survey-5 the state level data: Let’s play with some indices: Part 1

NFHS-5 data at the state level is now available for all states. Taking various ideas and indices altogether, let’s talk in the more easy way, who progressed and who didn’t.

Whenever one talks about development, the first index that comes to mind is Human Development Index. The human development index has three main indicators: health (Life expectancy index), education (expected years of schooling and mean years of schooling), and income (per capita gross national income (PPP)). This covers the main important pillars. While the Human development index measures development, there are other indices like the multi-dimensional poverty index which measures poverty. Global Multidimensional Poverty Index, given by Oxford Poverty and Human Development Initiative, has 10 indicators. Table no.1 gives more details.

Table no.1

Source: https://ophi.org.uk/multidimensional-poverty-index/  

While Human Development Index can easily be constructed at the state level, the multidimensional poverty index is difficult to construct at the state level. The indicator values in MPI is constructed for household and then if it is more than a threshold level then a household can be identified as a poor household. And then a total number of poor households are calculated and then state-level analysis can be done. But since household level NFHS data is still not available, in this article, I am trying to adjust these indices to construct an index at the state level.

I am trying to incorporate all indicators from MPI and HDI:

1)     Life expectancy rate

2)     Expected years of schooling

3)     Mean years of schooling

4)     Per capita Gross National Income (PPP)

5)     Malnutrition (Children under 5 years who are severely wasted (weight-for-height))

6)     Child mortality (Under-five mortality rate (U5MR))

7)     Cooking (Households using clean fuel for cooking)

8)     Sanitation (Population living in households that use an improved sanitation facility)

9)     Drinking water (Households using clean fuel for cooking)

      Electricity (Population living in households with electricity)

In the above 10 indicators, the direction of Malnutrition and Child mortality is opposite. In other words, having a higher number of child mortality represents backwardness or a low level of development. Same for malnourishment. Therefore, we have to change them in a positive direction. Under-five mortality rate (U5MR) is per 1000 while Children under 5 years who are severely wasted (weight-for-height) is in percentage. Therefore, the Under-five mortality rate (U5MR) can be subtracted from 1000, and Children under 5 years who are severely wasted (weight-for-height) can be subtracted from 100.

Nourished children= 100- Malnourished Children

Child Survival = 1000 - Child mortality (Under-five mortality rate (U5MR))

Normalization is done by using the following steps:

Where x: life expectancy rate, Nourished Children, Cooking, Sanitation, Drinking water, Electricity (as 100 can be a maximum value which can be achieved)

Where y: Child Survival (as 1000 can be the maximum value which can be achieved)

Where z: Expected years of schooling, mean years of schooling, per capita GNI.

After normalizing, weights are assigned to each indicator. As we can see in table 1, different weights are assigned. I am trying to keep these weights the same. Therefore, to keep these weights unchanged, 1/16 weightage should be assigned to cooking, electricity, sanitation, water, and 2/16 weightage should be assigned to life expectancy, survival, nourished children, mean years of schooling, expected years of schooling, per capita GNI.  

Therefore, a newly formed index will be

1/16*(Indicator_cooking + Indicator_electricity + Indicator_sanitation + Indicator_water) + 2/16*(Indicator_{per capita GNI} + Indicator_{mean years of schooling} + Indicator_{life expectancy} + Indicator_{expected years of schooling}+ Indicator_survival + Indicator_{nourished children})

The data for mean years of schooling, expected years of schooling, per capita GNI, life expectancy is extracted from the global data lab (https://globaldatalab.org/). The data for other indicators are taken from NFHS data.

Further states are divided into the larger and smallest states. This categorization is done based on NITI Ayog's categorization (http://social.niti.gov.in/uploads/sample/health_index_report.pdf).

Index Score and Ranking: SS for the Smaller States, LS for the Larger States and UT for Union Territories

Kerala has maintained the rank in 2014 as well as in 2019’s NFHS time period.  Surprisingly, poor states like Bihar, Jharkhand, and Uttar Pradesh and rich states like Gujarat and Maharashtra have seen a decline in this development index score. Although comparing the index score might be a problem as in normalization of Expected years of schooling, mean years of schooling, per capita GNI, the maximum and minimum values from the given sample. Therefore the decline in the score might happen due to changes in minimum and maximum values. But even after fixing the minimum and maximum values, these states have witnessed a fall in index scores. Gujarat and Maharashtra have also seen a decline in index ranking. Why these rich states are witnessing a fall in the index score and ranking? This decline comes from one indicator i.e. malnutrition. These states have witnessed an increase in malnutrition.

Figure no.1 highlights some interesting points.

This is a scatterplot between per capita GNI (PPP) in 1000 US dollars.

Yellow dots are for the 2019 year. We can see easily that yellows dots are right and below of blue dots. This suggests that in terms of income, states are moving ahead but in terms of malnutrition, states are not doing well.

(I have just taken one indicator for malnutrition i.e. Children under 5 years who are severely wasted (weight-for-height), if we take stunted then the development index score will go down further as states are doing very bad as far as malnutrition in terms of children who are stunted.).

Continued...