HOWTHE STATES WERE RANKED
mentary education per 6 to 14-yearold, ratio of boys to girls in elementary school and teacher-pupil ratio); consumer market (households owning TV
sets, number of affluent households, per capita deposits in banks and per capita ownership of two-wheelers); agriculture (percentage of cultivated area, agriculture GSDP per rural population, food grain yield, loans to farmers and net irrigated area); infrastructure (percentage of homes with electricity, per capita road length, percentage of villages connected with pucca roads, bank branches, LPG connections, post offices and telephones); and investment (per capita capital expenditure, commercial bank credit and gross capital formation in manufacturing, ratio of factories to number of disputes, ratio of industrial work- ers to urban 15-59 age population, and percentage of sick Small Scale Industries). Data on all the parameters was normalised.
To obtain a state’s performance under any one head, variables have to be aggregated. We used principal components analysis ( PCA), which churns out weights in the estimation process. Accordingly, for each head, we have scores for each state. Hence, the eight heads are aggregated into an overall performance index for each state. For this aggregation, we are reporting equal weight aggregation, since in this case there is little difference between equal and PCA weights. All the data is publicly available and is sourced from government and related organisations. However, in understanding the development of a state, what is important is not the base level score, but its increment. And this year we move to emphasising the changes in the scores. The base level depends on various historical factors and incremental improvements provide a lot of impetus for growth. Using the increments in scores, states are ranked for each head on the basis of their improvement.
We acknowledge the efforts of Minakshi Chakraborty, Swati Gupta, and Sunil Bhatt of Indicus Analytics in sourcing the information as well as conducting the basic analysis.
BIBEK DEBROY(L) AND