What Happens Once the Intervention Ends? The Medium-Term Impacts of a Cash Transfer Experiment in Malawi

Publication Details

Baird, S, Chirwa, E, McIntosh, C, and Özler, B, 2015. What happens once the intervention ends? The medium-term impacts of a cash transfer programme in Malawi, 3ie Impact Evaluation Report 27.New Delhi: International Initiative for Impact Evaluation (3ie)

Link to Source
Author
Sarah Baird, Ephraim Chirwa, Craig McIntosh, Berk Ozler
Country
Malawi
Region
Sub-Saharan Africa (includes East and West Africa)
Sector
Social Protection
Subsector
None specified
Gender analysis
Yes
Equity Focus
Gender
Evaluation design
Randomised Control Trials (RCT)
Status
3ie Series Report
3ie Funding Window
Open Window Round 2

Synopsis

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 behaviour.

Context

The question whether behavioural conditions, such as vaccinating children or sending them to school, should be attached to cash transfers remains a highly debated policy topic.  

According to the Malawi Integrated Household Survey-2 (2004-2005), households reported financial constraint as the primary reason for dropout from school. The Zomba district in the southern region of Malawi was chosen as the site for this study given the large percentage of school-aged girls who are at risk of dropping out of school early, becoming pregnant, and marrying early. Hence, cash transfers were expected to have sizeable effects in Zomba.

Research questions

This study reports outcomes over a five-year timespan for a cash transfer experiment that ran for two years, from 2008-2009.

The study is a randomised controlled trial of Zomba cash transfer programme (ZCTP), which was designed to test the importance of key design parameters that contribute to the success of a cash transfer programme. Specifically, SIHR assessed the effects of offering cash transfers to the families of school-aged girls for a duration of two years – while randomly assigning key policy design parameters, such as conditionality of transfers on school attendance, transfer amount, and the recipient of the transfers within households (parents or adolescent girl).

The one- and two-year impacts suggest that cash transfers had significant effects on outcomes ranging from education to early marriage and pregnancy to mental health. This paper presents the trajectory for the same set of outcomes – from baseline in 2007 until the five-year follow-up in 2012 – separately for girls who received cash transfers conditional on school attendance (CCT), who received transfers unconditionally (UCT), and those who did not receive them (control).

Methodology

Intervention design

The ZCTP is a randomised cash transfer intervention targeting young women in Malawi providing incentives (in the form of school fees and cash transfers) to current schoolgirls and young women who have recently dropped out of school to stay in or return to school.

Theory of change

Adolescent girls in Sub-Saharan Africa face a multitude of hazards during their transition from childhood to adulthood. Net primary school enrollment for females in the region is lower than 75 per cent -- with lower completion rates and much lower transition rates into secondary school, resulting in net enrollment rates in secondary schools of around 25 per cent in countries like Malawi (World Bank 2013). Short-term evidence from the cash transfer experiment under examination in this paper shows that cash transfers can also significantly delay marriage and pregnancy among school-aged girls. 

If the aim of these programmes is to not only increase current welfare of adolescents, but actually improve their lives in the long-run by making investments in their human and physical capital during an important period of transition in their lives, then it is important to find out whether the short-term improvements were sustained or not.

Evaluation design

The evaluation design has two treatment arms: conditional cash transfers (CCT) and unconditional cash transfers (UCT).

Treatment status was assigned at the enumeration area (EA) level. Zomba district is divided into 550 EAs, which are defined by the National Statistical Office of Malawi and contain an average of 250 households spanning several villages.  

A sample of 176 EAs selected from three different strata: Zomba city (urban, 29 EAs), near rural (within a 16 kilometre radius of Zomba city, 119 EAs), and far rural (28 EAs) were randomly divided into two groups of equal size: treatment and control.  

The sample of 88 treatment EAs was further divided into two arms based on the treatment status of baseline schoolgirls: (i) CCT arm (46 EAs), and (ii) UCT arm (27 EAs). In the remaining fifteen treatment EAs, no baseline schoolgirls were made offers to receive cash transfers. All baseline dropouts residing in the 88 treatment EAs received CCT offers regardless of the assigned treatment status of baseline schoolgirls in their EAs. No EA in the sample had a similar cash transfer programme before or during the study.

Main findings

The findings of this study may serve to temper some of the recent enthusiasm for unconditional transfers by showing that the substantial effects observed in critical outcomes like teen pregnancy and early marriage during the course of the programme were almost entirely transient. As soon as the programme ended the UCT arm engaged in ‘catch-up’ behaviour that caused the trajectory of these outcomes to return to the same level that they would have reached had the programme never been put in place.

Implications for implementers

The current study re-focuses attention on the importance of bolstering human capital among the most vulnerable segments of the adolescent population.

The findings also suggest that the two-year cash transfer programme aimed mainly at girls eligible to attend the final years of primary school or higher may have been too short and too late. Many programmes target children earlier and last longer. Cash transfers that lasted a few more years might have caused more permanent effects on health, nutrition, and fertility and prevented dropouts and pregnancies before it was too late.

Implications for policy and practice

Understanding target populations and key outcomes of interest is critical for designing both the right type of policy, as well as the optimal length of the programme. In this particular case, there were important differences in the impacts of CCTs and UCTs in the short run, as well as in the impacts of CCTs between baseline schoolgirls and baseline dropouts. The intervention that will have the largest desired impact will ultimately depend on the outcomes that the policymaker cares most about, as well as the individuals they are hoping to reach. There are likely to be important trade-offs between policy options.

Implications for further research

The findings also suggest that researchers analysing short-term effects should be careful in interpreting them. Short-term findings (two years after baseline) from this study suggested that while CCTs were more successful in improving schooling outcomes than UCTs, UCTs were more successful in reducing teen pregnancies and early marriages. Five years from baseline, these significant effects disappear and the outcomes are indistinguishable between the control, CCT, and the UCT groups. The trends in marriage and fertility in the UCT group between Round 3 and Round 4 (i.e. trends after the cash transfers stopped) are particularly telling given that all of the significant gains obtained during the programme were wiped out by trends in the other direction. In other words, UCTs simply delayed pregnancies and marriages rather than preventing them.

Policy influence

Findings have informed two World Bank-supported government programmes , a school bursary programme and a pilot cash transfer programme for needy students. DFID, in collaboration with the Ministry of Education Science and Technology in Malawi, is introducing a national cash transfer programme for parents of girls, in order to improve transition rates from primary to secondary schools and address adolescent pregnancy and marriage. In addition, the International Rescue Committee is using findings to design the Girl Empower programme in Ethiopia and Liberia. The Bill & Melinda Gates Foundation acknowledged the study’s contribution to changing the discourse on HIV prevention and incubating the design of other programmes and studies, including the HIV Prevention Trials Network in South Africa, the STRIVE project in rural Tanzania, and the Johns Hopkins study on HIV combination prevention interventions.

Read the journal articles (published and forthcoming): 

Baird, S., de Hoop, J., Özler B. “Income Shocks and Adolescent Mental Health,” Journal of Human Resources, forthcoming. 

Baird, S., McIntosh C., Özler B. “Cash or Condition: Evidence from a Cash Transfer Experiment. The Quarterly Journal of Economics, Oxford University Press, vol. 126(4), pages 1709-1753. 

Baird, S. and Özler B. “Examining the Reliability of Self-Reported Data on School Participation,” Journal of Development Economics, Vol. 98(1), 2012, pp. 89-93. 

Baird, S., Garfein, R., McIntosh C., Özler B. “Effect of a cash transfer programme for schooling on prevalence of HIV and herpes simplex type 2 in Malawi: a cluster randomized trial” The Lancet, Volume 379, Issue 9823, Pages 1320 - 1329, 7 April 2012. 

Baird, S., Garfein, R., McIntosh C., Özler B. “Cash transfer scheme for reducing HIV and herpes simplex type 2 Author’s reply,” correspondence in The Lancet, Vol. 380(9844), 2012, pp. 802. 

Read working papers (published and forthcoming): 

Baird, S., Gong, E., McIntosh, C., & Özler, B., 2013. The Heterogeneous Effects of HIV Testing. University of Otaga Economics Discussion Paper No. 1310 

Baird, S., Chirwa, E., de Hoop, J., and Özler, B. “Girl Power: Cash Transfers and Adolescent Welfare. Evidence from a Cluster-Randomized Experiment in Malawi,” in S. Edwards, S. Johnson, D. Weil (editors), NBER Volume on African Economic Successes, University of Chicago Press, 2014, forthcoming. Baird, S., Bohren, A., McIntosh, C., & Özler, B. “Designing Experiments to Measure Spillover Effects,” Submitted.

Additional publications

Media coverage

Blogs

Videos

Scroll to Top