Program

Aug 04, 2016

OVERVIEW

Conditional Cash Transfer (CCT) projects are an increasingly common tool used throughout the developing world to give vulnerable populations cash resources that are conditioned on certain desired behaviors.  Often, they condition the transfer of funds on children’s school attendance or visits to the doctor for regular check-ups.  This summary reviews two evaluations of a CCT program in Malawi.  In Malawi, a CCT was combined with an unconditional cash transfer arm (UCT) in order to increase school attendance and decrease risky sexual behaviors.  Girls in both groups were less likely to be HIV positive at the end of the intervention and were less likely to have a sexual partner who was older than 25. Additionally, individuals in the intervention groups were more likely to be enrolled in school than those in the control group while the only difference between the two control groups was that participants in the UCT arm were less likely to be pregnant or married.

DESCRIPTION OF PROGRAM

Target population: Never-married women ages 13 to 22 at risk of HIV acquisition

This program in Malawi gives cash to never-married female adolescents and young adults in order to reduce HIV and herpes-2 acquisition. The theoretical rationale is that attending school will reduce risky behavior but also that having enough cash for school will reduce the need for young women to sleep with older men who are willing to pay for school fees and uniforms.  Young women in the conditional treatment group are given one to five dollars, and their parents are given four to 10 dollars if they participate, conditional on going to school on 80 percent of school days. A second treatment group received unconditional payments.

EVALUATION OF PROGRAM

Baird, S. J., Garfein, R. S., McIntosh, C. T., & Ozler, B.  (2012). Effect of a cash transfer program for schooling on prevalence of HIV and herpes simplex type 2 in Malawi: a cluster randomised trial. The Lancet, 379, 1320-29.

Baird, S., McIntosh, C., & Ozler, B.  (2011). Cash or condition?  Evidence from a cash transfer experiment. The Quarterly Journal of Economics, 126, 1709-1753.

Evaluated population: Young women who lived in the Zomba district in southern Malawi (an area characterized by high poverty, low enrollment in school, and high rates of HIV) were participants in this study.  Girls were included from a random selection of 176 of the 550 enumeration areas in this southern region of Malawi.  Twenty-nine enumeration areas were chosen from an urban area, 119 from near rural areas (less than 16 kilometers from the city), and 28 from far rural areas (more than 16 kilometers from the city).  In these areas, a census of all girls in households was conducted and all girls who had dropped out of school were included while girls who were still in school were chosen randomly.  This resulted in study sample of 4,051 individuals, in which 3,796 completed the baseline survey.  Of these, 889 girls were school dropouts (baseline dropouts) and 2,907 were in school (baseline schoolgirls).  Of the baseline schoolgirls, between 19 and 22 percent of respondents had ever had sexual intercourse, 97 percent had never been pregnant, the average age was approximately 15, and the average age of sexual debut was just under 16.  Most had completed seventh grade, approximately 15 percent of the girls were maternal orphans, and 25 percent were paternal orphans. Of the baseline dropouts, between 68 percent of respondents had ever had sexual intercourse, 44 percent had ever been pregnant, the average age was 17.6 and the average age of sexual debut was approximately 16.4.  Most had completed sixth grade, approximately 22 percent of the girls were maternal orphans, and 34 percent were paternal orphans.

Approach: Of the 176 enumeration areas, 88 were randomly assigned to be control areas (no cash transfers) and 88 were assigned to be treatment areas.  In these areas, all of the baseline dropouts were given conditional cash transfers (conditioned on going to school for 80 percent of the days) and some or all of the baseline schoolgirls were assigned to no cash transfer, conditional cash transfers, or unconditional cash transfers in order to study spillover effects. The conditional cash transfer was conditioned the same as for the baseline dropouts (80 percent of days) while the unconditional cash transfer just required that the participant come to the cash transfer point.  In the areas where baseline schoolgirls were offered transfers, approximately one-third of the areas gave cash transfers to 33 percent of the baseline schoolgirls, another third gave cash transfers to 66 percent of the baseline schoolgirls, and a final third gave cash transfers to 100 percent of the baseline schoolgirls.  The amount given to the girls was randomly assigned monthly and ranged from one to five dollars.  Guardians were also given money between four and 10 dollars (in two-dollar increments).  There were no statistically significant differences in the baseline characteristics of the different groups within the different groups of baseline schoolgirls.  Data on both behavioral and knowledge outcomes was measured and biologic specimens were collected six and 18 months after the intervention began for both HIV and HSV-2 from a random sample of areas.  Seven percent of the baseline schoolgirls and 10 percent of the baseline dropouts were lost to follow-up after the first year.  Logistic regressions were conducted to estimate unadjusted and adjusted odds ratios.

Results:

Study 1:  The primary outcomes were HIV and HSV-2 prevalence, but school enrolment, marriage, pregnancy, sexual behavior, knowledge about HIV/AIDS, and syphilis were secondary outcomes.  At 18 months, the prevalence of HIV in the combined intervention groups of baseline schoolgirls was 1.2 percent and in the control group it was 3.0 percent, which was a statistically significant difference in the adjusted models.  Additionally, sexual activity was lower as was the likelihood of having a sexual partner who was 25 or older in the intervention groups.  Finally, individuals in the intervention groups were more likely to be enrolled in school than those in the control group.  There were no differences in marriage, pregnancy, consistent use of condoms, sexual debut, syphilis, HIV knowledge, HIV testing, or health training around HIV. There was only one statistically significant difference between participants in the CCT and UCT groups in these outcomes of interest: four percent of those in the CCT group were pregnant at the time of the follow-up, compared with just one percent of those in the UCT group. For baseline dropouts, re-enrollment in school was higher in the intervention groups, and they were also less likely to have gotten married or be sexually active. However, the prevalence of HIV and HSV-2 was similar at approximately 9 percent and 8 percent of the sample respectively. When the intervention groups were disaggregated by conditionality status, the directions of most of the odds ratios stayed the same, although significance changed slightly.

Study 2: This study focused primarily on the differences in outcomes between those assigned to the conditional versus unconditional cash transfer arms.  The primary outcomes in this study were school enrollment, attendance, performance on exams in cognitive ability, mathematics, and English reading comprehension, marriage and child-bearing.  Dropout rates declined in both the CCT and UCT arms; however, the dropout rates in the UCT arm remained significantly higher than in the CCT arm.  For some but not all school terms, attendance was found to be significantly higher for participants in the CCT arm than in both the control and UCT groups, which were statistically similar.  Additionally, those in the CCT groups performed significantly better than those in the other intervention groups on tests of English and cognitive ability, and marginally significantly better for math tests.

In terms of the other outcomes – marriage and child-bearing – participants in the UCT arm fared better.  Those in the UCT arm were 27 percent less likely than the control group to be pregnant or married at the end of the study, while the CCT arm had small and non-significant declines relative to the control.  These drops in pregnancy and marriage were found almost exclusively amongst girls who had dropped out while girls who were still in school had very low rates of pregnancy and marriage.

SOURCES FOR MORE INFORMATION

References

Baird, S. J., Garfein, R. S., McIntosh, C.T., & Ozler, B.  (2012). Effect of a cash transfer program for schooling on prevalence of HIV and herpes simplex type 2 in Malawi: A cluster randomised trial. The Lancet, 379, 1320-29.

Baird, S., McIntosh, C., & Ozler, B.  (2011). Cash or Condition?  Evidence from a cash transfer experiment. The Quarterly Journal of Economics, 126, 1709-1753.

KEYWORDS: adolescents, youth, young adults, female only, urban, rural and/or small towns, community-based, high school completion/dropout, births, teen pregnancy, STD/HIV/AIDS, condom use and contraception, sexual activity

Program information last updated on 8/4/2016.