Aggregation underestimates growth: Case study in Population Modelling for India

摘要

Jensen's inequality implies that aggregation underestimates growth and overestimates decay compared to an uncoupled disaggregated approach. As a study in disaggregation, this paper extends previous work in India's top down population modelling. Top down estimates of vital parameters used in country level population modelling hide the wide dispersion and interplay between these parameters that sub-national population cohorts have. Differences in total fertility rates, mortality rates and the demographics between national level averages and sub-national actuals, for large and diverse countries such as India, lead to a significant gap in population growth forecasts at the two levels. Population projections from International agencies, that are based on top down estimates, should be read in the context of these limitations. Higher, country level, starting data points conceal population groupings that demand more policy attention. Disaggregated modelling, used in this paper provides a plug and play alternative to construct sub national population projections that are used to derive country level estimates. As an example, state level population data, based on India's latest, 2011 census, have been used to create a population projection for India. Disaggregated population modelling provides another insight: geographically focused policy planning for fertility programs.