Publication Title: International Initiative for Impact Evaluation (3ie)

Abstract: This report evaluates the five year impacts of the Zomba Cash Transfer Programme (ZCTP) in Zomba, Malawi. The ZCTP took place for two years during 2008-2009, and involved giving cash transfers, both conditional on schooling and unconditionally, to initially never-married 13-22 year old young women. The Schooling, Income and Health Risk (SIHR) study was designed to evaluate the impacts of the cash transfer programme on a variety of outcomes ranging from education to health to sexual behavior. The SIHR study is a randomised control trial where young women were randomly assigned to one of three groups: control, unconditional cash transfer (UCT), and conditional cash transfer (CCT). Baseline data was collected in 2007, with follow-up data collected in 2008-2009 (Round 2), during the programme, in 2010 (Round 3), immediately upon the conclusion of the programme and in 2012-2013 (Round 4), two years after the programme ended.

The strong and significant short-term impacts of the ZCTP (using data collected in 2008 and 2010) have been documented elsewhere. This report focuses on impacts two years after the programme ended, in 2012-2013, to try and understand whether this relatively short (two-year) intervention of cash transfers – introduced at a particularly important period of transition from adolescence to adulthood – can have lasting effects on this cohort of young females and their future families. The analysis focuses on four key domains for the recipients of the cash transfer programme: education; marriage and fertility; health and nutrition; and sexual behavior. The analysis focuses on whether results found in the short term were sustained two years after the programme ended. Results are analyzed separately for young women who were in school at baseline (baseline schoolgirls) and those that were out of school at baseline (baseline dropouts), an oft-overlooked group. The analysis for baseline schoolgirls focuses on differential impacts between the CCT and UCT arm, while the analysis for baseline dropouts focuses on the difference between the CCT and the control (no UCT experiment was conducted for this group). Overall, results suggest that the substantial benefits conferred by unconditional treatment while the programme was in place in the domains that we investigate here were almost completely transient. Even the conditional programme, when implemented among those in school at baseline (and therefore likely to continue with schooling even in the absence of a CCT) had few detectable long-term impacts. The programme that provided conditional cash transfers to girls who had already dropped out of school at baseline, on the other hand, had large and durable impacts on a wide range of outcomes – including primary school completion, years of education, marriage rates, likelihood of having started childbirth, and desired fertility. Our results suggest that long-term impacts are sustained only when a cash transfer programme achieves substantial improvements in the stock of a durable form of capital, such as human capital. The results of the evaluation of the ZCTP, both in the short and long term, provide important lessons for policy makers thinking of designing cash transfer programmes as part of their social protection policy.