An empirically driven theory of poverty reduction

Programme of work

Increasing evidence transferability

Principal investigator(s)

Sudhanshu Handa

Host institution

Carolina Population Center, University of North Carolina at Chapel Hill

Other institutions

Centre for Social Research, University of Malawi
School of Education and Leadership, University of Ghana
Department of Agricultural Economics and Extension, University of Zambia

Dates

January 2020 to January 2022 (TBC)

Project type

Secondary data analysis

Country/ies

Malawi, Ghana, Zambia

Research question

This study will develop a middle-range theory of economic growth to un-derstand how cash transfers (an external liquidity injection) affect psychological states and household behavioural responses to produce different effects among households with different characteristics.

Research design

The study will use recent developments in machine learning to identify groups of households with different treatment effects.

Using household consumption as the key outcome, machine-learning tools will identify the key subset from a very large set of pre-treatment variables that enables some households to realise large gains in consumption. The specific actions of this set of households will then be analysed to understand what they did to achieve those large gains in consumption.

Data source

The study will draw on secondary data from impact evaluations of the four countries’ national unconditional cash transfer programmes merged with a large set of secondary variables on the microenvironment (i.e. market access, climate, topology and land cover).

Policy relevance

The results will provide insights on which households are candidates to graduate out of poverty, and the pathways for doing so. It will help inform the design and targeting of future cash transfer programmes.