On the Measurement of Poverty Dynamics

This paper introduces a family of multi-period poverty measures derived from commonly used static poverty measures. Our measures trade-off poverty levels and changes (gains and losses) over time, and are consistent with loss aversion. We characterize the partial ranking over income dynamics induced by these measures and use it in two empirical applications with longitudinal household level data. Comparing two decades of income dynamics in the United States we find that the income dynamics of the 1990s -post Welfare reform- dominates the income dynamics of the 1980s -pre Welfare reform. Next, we compare the contemporary income dynamics of three industrialized countries and conclude that United Kingdom dominates Germany and United States, and Germany dominates the United States if poverty stocks are given more importance than poverty flows. The differences between our ranking and those obtained using other welfare criteria such as social mobility suggest that our measures capture critical information about the evolution of poverty.