Construction of the Sub-National Human Development Index
The Subnational Human Development Index (SHDI) presented at this website is a translation of the UNDP’s official HDI (hdr.undp.org) to the subnational level. As such, it is an average of the subnational values of three dimensions: education, health and standard of living. In its official version defined at the national level, these dimensions are measured with the following indicators: Education measured with the variables ‘Mean years of schooling of adults aged 25+’ and ‘Expected years of schooling of children aged 6’; health measured with ‘Life expectancy at birth’ and standard of living measured with ‘Gross National Income per capita (PPP, 2011 US$)’. Details about the estimation procedure can be found here
To construct the SHDI for the period 1990-2017, we computed the subnational variation in these indicators and applied it to their national values derived from the UNDP website (hdr.undp.org/data). For computing the subnational values, basically two different data sources were used: indicators constructed on the basis of survey and census datasets by the Global Data Lab (www.globaldatalab.org) and indicator data obtained from statistical offices. The indicators derived from survey and census datasets are the major source of data for low and middle income countries (LMICs) and those obtained from statistical offices for high-income countries (HICs).
Because life expectancy and Gross National Income per capita (GNIc) are not readily available in household surveys and census datasets, their subnational values for LMICs had to be estimated. For life expectancy, this was done on the basis of data on under-five mortality and for GNIc on the basis of household wealth. To measure household wealth, the International Wealth Index (IWI) was used, an indicator of household’s standard of living based on asset ownership, housing quality and access to public services (Smits & Steendijk, 2015).
For years for which no subnational data for one or more indicators was available, the values of these variables were estimated by linear interpolation between the preceding and succeeding year for which this information was available. If interpolation was not possible, the subnational values of the nearest year were used. The subnational variation obtained in this way was subsequently applied to the UNDP national values, so that for each year in the period 1990-2017 the (population weighted) mean of the subnational values is in line with the UNDP values.
Further information on the construction of the SHDI Database can be found in:
Smits & Permanyer (2019), The Subnational Human Development Database. (Nature) Scientific Data, 6, 190038. Download paper here.
Jeroen Smits & Iñaki Permanyer
Smits, J. & R. Steendijk (2015). The International Wealth Index (IWI). Social Indicators Research, 122(1), 65-85. Download here