SMU SOE Seminar (Aug 6, 2018): Is Poverty Persistent?

We use data collected for a randomized control trial of an asset transfer program to the ultra-poor where data was collected in four waves between 2007 and 2011 in Bangladesh to track their economic condition to differentiate between a “convergence” vs a “poverty trap” view of individual-level economic change. We find evidence that is consistent with there being a poverty trap. We identify a threshold level of initial capital needed for day laborers to take on a new and more remunerative occupational activity (livestock rearing). Individuals close to this threshold are pushed above it by the asset transfer and escape poverty whereas those further away fail to do this and fall back into poverty. We use data from control villages to show that: (i) there is a missing mass around the threshold; (ii) differences in individual productivity or random shocks to productive assets cannot explain the response to the program. We test three possible mechanisms underlying the trap - nutritional, behavioral, and technological and find evidence only for the last. We discuss the implications of our results for the design of anti-poverty policies to reach the extreme poor who live in risky environments.