This study assesses whether the protective and productive impacts of the Zambian government's Child Grant Programme (CGP) persist two years after households became ineligible for the programme.
The Zambian Child Grant Programme (CGP) provided an unconditional cash transfer to households with a child under age five in three rural districts. A reform and consolidation of the grants programmes in 2014 resulted in some beneficiaries no longer being eligible for the grant. An initial evaluation of the CGP showed strong productive and protective effects. A key policy question in Zambia and other countries is whether households can eventually graduate out of poverty with a simple cash transfer without any complementary services. This study seeks to respond to this question.
- Do the impacts of a cash transfer among the ultra-poor persist after beneficiaries leave the programme? The main indicator is consumption, and the secondary indicators are measures of economic productivity (value of harvest, non-farm enterprise) and assets (livestock, productive tools).
- How quickly do new cash transfer beneficiaries catch-up to those that have been receiving cash for four years? The main indicator is consumption, and the secondary indicators are measures of economic productivity (value of harvest, non-farm enterprise) and assets (livestock, productive tools).
The CGP provided an unconditional flat cash transfer of approximately US$12 per month, paid bimonthly in cash in three rural districts. Beneficiaries are households with a child under age five at the time of programme initiation in 2010.
Theory of change
An unconditional cash transfer to the ultra-poor should have an immediate effect on consumption and food security, assuming food is available in the local market and there are no inflationary effects. After initial consumption needs are met, there could be productive effects if households are credit constrained.
a) A RCT was implemented in 2010-14 to evaluate the impact of the CGP. In 2015, households were-targeted as part of a policy reform, and households in the original treatment and control groups were no longer eligible for the new programme. This study compares these two groups of households, who are similar due to the original random assignment.
b) The randomisation was at the village-cluster level, stratified by the 3 districts. In each district, 15 village clusters were assigned to treatment or control status, for a total of 90 clusters in the sample, and 2519 households. All eligible households in treatment districts were served by the programme. A household survey was conducted at baseline in 2010, and in 2012, 2013 and 2014. As part of this study an additional survey was conducted in 2017, after the reform.
c) This is a multi-site RCT, and treatment and control village clusters come from the same district. Spillover can occur to non-eligible households in treatment villages, and this was studied as part of the original evaluation. Contamination is a function of the distance between village clusters across the two arms.
d) The sample is ultra-poor households.
The main finding is that consumption among the original treatment group is no longer statistically different from that of the original control group, two years after they were removed from the cash transfer. Similarly, there are no longer any statistical differences in productive indicators such as value of harvest, area cultivated, and livestock holdings. This suggests there is fade-out of original treatment effects two years after removal from the programme.
The secondary research question is whether new entrants into the cash transfer catch-up to those who have been in the programme for 4+ years. The study finds there is immediate catch-up in consumption and productive indicators.
Implications for policy and practice
CGP households live in extremely harsh environments, with limited access to markets, electricity, water, and other infrastructure, and few complementary productive services. Initial impacts of the CGP were extremely large and positive across a range of indicators, but were not sustained after households left the programme. The results suggest that without an enabling environment that allow households to diversify or make structural changes in their livelihoods strategies, permanent increases in consumption through a short-term (four years) cash transfer programme to the ultra-poor is unlikely.
Implications for further research
This is the first study that looks at effects of a government-run cash transfer programme after beneficiaries have left the programme. As governments are unlikely to purposefully remove households from a programme for research purposes, further evidence on the question of graduation from poverty will need to rely on policy reforms and other types of natural experiments in order to build the thin evidence based in developing countries on the persistence of impacts of poverty programmes.