Four Papers on Structural Household Welfare Dynamics
Despite recent progress, abject poverty remains pervasive in many countries around the world. Achieving further sustained reductions in poverty will require more effective poverty eradication policies. The effectiveness of these policies, in turn, depends on how well we understand the structural dynamics of households moving in and out of poverty. The four papers in this dissertation explore several issues in the modeling and measurement of structural household welfare dynamics that to date have received little attention in the academic literature but that are directly relevant for the design of poverty reduction policies. The first paper examines quantitative methods for modeling household welfare dynamics and identifying long-run welfare equilibria and poverty traps. It proposes a new semiparametric panel data estimator that has several advantages over methods used in the extant literature. The empirical application to data from three Indian villages shows deep structural immobility. Structural poverty traps loom large, as rural Indian households who start out asset-poor are likely to remain poor. The second paper proposes and applies a statistical test to examine whether high estimates of economic mobility and transitory poverty in the existing literature are partially driven by stochastic one-off income flows. It finds that these estimates are inversely correlated to the length of the interval between panel observations suggesting that estimates based on short panel spells represent (high) upper bounds of underlying structural economic mobility and (low) lower bounds of chronic poverty. The third paper introduces several new classes of intertemporal poverty measures that can incorporate the variability of household welfare and the distribution of poverty across households over time. Accounting for these intertemporal factors in rural Pakistan leads to greater estimates of poverty than using existing, static poverty indicators. The fourth paper uses a regression-based technique to explore the household characteristics that determine income inequality in rural Pakistan. The level of inequality is determined primarily by land ownership and location. These structural variables are difficult to change by policy, in contrast to the factors that reduced inequality over time, such as access to secondary education and lower dependency ratios.
welfare dynamics; poverty; inequality; development economics
dissertation or thesis