81208
                  enGender Impact: The World Bank’s Gender Impact Evaluation Database



                   Cash or Condition Evidence from a Cash Transfer Experiment

      Author(s)            Sarah Baird, Craig McIntosh, Berk Ozler
      Contact              bozler@worldbank.org

      Country              Malawi

  Organizing Theme         Education and Skills, Voice and Agency

       Status              Completed

Intervention Category      Cash Transfer

       Sector              Social Protection
                           This article assesses the role of conditionality in cash transfer programs using a unique
                           experiment targeted at adolescent girls in Malawi. The program featured two distinct
                           interventions: unconditional transfers (UCT arm) and transfers conditional on school
                           attendance (CCT arm). Although there was a modest decline in the dropout rate in the
      Abstract             UCT arm in comparison with the control group, it was only 43% as large as the impact in
                           the CCT arm at the end of the 2-year program. The CCT arm also outperformed the UCT
                           arm in tests of English reading comprehension. However, teenage pregnancy and marriage
                           rates were substantially lower in the UCT than the CCT arm, entirely due to the impact of
                           UCTs on these outcomes among girls who dropped out of school.
 Gender Connection         Gender Focused Intervention

  Gender Outcomes          Female school attainment, cognitive skills, decision making over family formation

     IE Design             Clustered Randomized Control Trial (Clustered at Enumeration Area level)
                           There are two treatment arms and one control group. The first treatment arm is a standard
                           CCT. The size of the CCT was determined by lottery and ranged from $1-5 per month.
    Intervention           Monthly school attendance must have been above 80%. Also the school fees would be paid
                           in full. The UCT arm offered identical transfers but had no school-attendance requirement.
                           The cash transfer was focused only on girls.
 Intervention Period       Treated girls received monthly payments for two years from 2008 to the end of 2009
                           The sample comes from 176 Enumeration areas in the Zomba district in Malawi. The
  Sample population        target population was never-married females aged 13-22. There was a total sample size of
                           2907 schoolgirls.
Comparison conditions      There are 3 treatment arms, CCT, UCT and a control group.

   Unit of analysis        Individual Level
                           There was s a follow up during the intervention and an additional survey 9 months after the
  Evaluation Period
                           intervention ended.
                           Dropout rates declined in both treatment arms, however the UCT arm is 43% as large as
       Results
                           the CCT arm. The fraction of days attended in the CCT arm is significantly higher than the

   Last updated: 14 August 2013                                                                     1
                   enGender Impact: The World Bank’s Gender Impact Evaluation Database


                            UCT. Cognitive ability, math and English scores significantly improve for the CCT but not
                            for the UCT. Marriage and Pregnancy Rates tell a different story. Probability of pregnancy
                            and marriage were much lower in the UCT arm. The delays in marriage and fertility in the
                            UCT arm are found entirely among adolescent girls who dropped out of school after the
                            start of the intervention. Improvements in human capital for the recipients of the CCT are
                            achieved at the cost of worse outcomes for people who drop out of school.
Primary study limitations   The paper does not discuss limitations
                            Global Development Network, Bill and Melinda Gates Foundation, NBER Africa Project,
                            World Bank Research Support Budget Grant, World Bank Knowledge for Change Trust
    Funding Source
                            Fund, World Development Report 2007 Small Grants Fund, Spanish Impact Evaluation
                            Fund, Gender Action Plan Trust Fund
                            Baird, S., McIntosh, C., & Özler, B. (2011). Cash or condition? Evidence from a cash
      Reference(s)
                            transfer experiment. The Quarterly Journal of Economics, 126(4), 1709-1753.
    Link to Studies
                            http://qje.oxfordjournals.org/content/126/4/1709.short
       Microdata




    Last updated: 14 August 2013                                                                     2