Cash and Change: a replication and further analysis of a cash transfer experiment in Malawi (replication data)

Replication files for Cash and Change (Reimao 2025), which is a replication and measurement and estimation analysis of Cash or Condition (Baird et al. 2011). These are the files for replication, with data from 3 rounds appended. The original data can be found in The World Bank's Microdata Library (Household Surveys from WB M…). Cash or Condition is one of the few papers using a randomized control trial to directly compare the effects of conditional and unconditional cash transfers (CCTs and UCTs). Limited to the data made publicly available, I first conduct a narrow replication and find that the results that can be reproduced are largely similar to the original study. However, these results alone would lead us to the opposite conclusion from that in the paper, and would indicate that UCTs are more effective at increasing school enrollment and attendance, as well as pregnancy and marriage, implying that UCTs are generally superior to CCTs for these outcomes. Widening the analysis to consider primary and secondary school girls separately, I find that while the UCT decreases the likelihood of early marriage for girls in both levels, its effect on early pregnancy is concentrated on primary school girls – even though households with secondary school girls received larger transfers.