Poverty Index With Time Varying Consumption and Income Distributions

TL;DR

This paper develops a dynamic poverty index based on time-varying income and consumption distributions, providing a more accurate and predictive measure of poverty trends in developing economies.

Contribution

It introduces a unified theoretical framework incorporating endogenous time-dependent consumption and income distributions, improving upon previous models that relied on data extrapolation.

Findings

CD-dynamics lower the basic necessity consumption threshold

The new index aligns better with recent poverty trends

It enables probabilistic future poverty predictions

Abstract

In a recent work (Chattopadhyay, A. K. et al, Europhys. Lett. {\bf 91}, 58003, 2010) based on food consumption statistics, we showed how a stochastic agent based model could represent the time variation of the income distribution statistics in a developing economy, thereby defining an alternative \enquote{poverty index} (PI) that largely agreed with poverty gap index data. This PI used two variables, the probability density function of the income statistics and a consumption deprivation (CD) function, representing the shortfall in the minimum consumption needed for survival. Since the time dependence of the CD function was introduced there through data extrapolation only and not through an endogenous time dependent series, this model left unexplained how the minimum consumption needed for survival varies with time. The present article overcomes these limitations and arrives at a new unified theoretical structure through time varying consumption and income distributions where trade is only allowed when the income exceeds consumption deprivation (CD). Our results reveal that such CD-dynamics reduces the threshold level of consumption of basic necessities, suggesting a possible dietary transition in terms of lower saturation level of food-grain consumption. The new poverty index conforms to recently observed trends more closely than conventional measures of poverty and allows probabilistic prediction of PI for future times.