Normative Choices and Tradeoffs when Measuring Poverty over Time

This paper examines the aggregation of an indicator of wellbeing over time and across people to measure poverty. We characterise the general form of an intertemporal poverty measure under mild normative principles and show that it must embody an unambiguous ordering of possible trajectories of an individual’s wellbeing. We motivate further normative principles and examine their consequences for the form of the measure, showing that some measures suggested in the literature are not consistent with these principles. We discuss additional stronger properties that may be argued to be desirable for an inter-temporal or chronic poverty measure. We identify compatibilities and tradeoffs among certain of these properties. For example, a poverty measure cannot simultaneously capture chronicity of poverty and sensitivity to fluctuations. We argue that a poverty analyst should choose among these properties according to context and the particular conception of poverty she seeks to measure.