Smoothing the Cost of Education: Primary School Saving in Uganda

How to cite

Karlan, Dean; Linden, Leigh. 2014. "Smoothing the Cost of Education: Primary School Saving in Uganda." AEA RCT Registry. AEARCTR-0000081. https://doi.org/10.1257/rct.81-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Dean Karlan* , Leigh Linden

* Primary investigator

Registry ID
AEARCTR-0000081
DOI
https://doi.org/10.1257/rct.81-1.0
First registered
Last updated
Status
Completed
Trial dates
2009-09-01 – 2013-12-01
Countries
Uganda ()
Registry record
http://www.socialscienceregistry.org/trials/81

Abstract

Can a financial service be used to improve educational outcomes? To answer this question, Innovations for Poverty Action (IPA) partnered with the Private Education Development Network (PEDN) to design and implement the Super Savers Program. A school based savings program, the Super Savers Program aimed to connect savings and education in Uganda. Along with program implementation, IPA directed a research study to assess the program’s impact.

The study used a randomized control design to measure the impact of different variations of the Super Savers Program. In one variation, students who participated received their savings in cash at school, with some encouragement to invest their savings towards school expenditures. In the second variation, the savings payout was provided in a more restrictive form: vouchers requiring an education-related purchase. Students at a third group of schools served as a comparison.

The program variation which included a savings payout in the form of cash generated higher savings in the program accounts. When combined with a parent outreach program, the program led to an increase in the percentage of children in the classroom with basic scholastic materials and an 8% increase in test scores on an exam covering language and math skills. We find no effect for the fully-committed account, and we find no effect for either account on attendance, enrollment, or non-cognitive skills.

The results from this study contribute to existing research in two ways: First, while we know that savings programs are particularly effective when tied to a specific goal—such as education—putting too many restrictions on payouts can be counteractive. Second, reaching out to parents and households can be an important way to improve primary education.

Intervention

Working with a local nonprofit organization Private Education Development Network (PEDN) in the Busoga sub-region of the Eastern region of Uganda, we randomly assigned 136 primary schools to one of three groups: a strong commitment savings account, a weak commitment savings account, or control. For both treatments Innovations for Poverty Action (IPA) developed a brief training component for the teachers, and also coordinated the transfer of money from a savings box held at the school to a local bank for safekeeping. One year into the implementation, we implemented one sub-treatment, a parental involvement workshop, in half of the treatment schools. We measured school inputs, test scores, and attendance, both before and after the intervention.

Experimental design

We selected 136 primary schools from the Jinja, Iganga, Mayuge and Luuka districts of the Busoga Region because they predominantly comprised poor rural and peri-urban schools. We then administered a baseline survey and test during the final trimester of 2009. We then randomly assigned schools to receive either the cash treatment, voucher treatment, or no treatment. Each treatment group comprised 39 schools, and the remaining 58 schools became the control group.

Following the first randomization, school outreach began. It took two trimesters to recruit the majority of schools, but by the beginning of the third trimester of 2010, 95 percent of the treatment schools had agreed to participate.

In 2011, we conducted a second randomization for the parent sensitization program, to isolate the effect of the program while still treating all of the schools. Half of the treatment schools were assigned to receive the treatment in the first term of 2011 while the rest received the treatment during the second term.

Finally, we conducted the follow-up survey and exam during the beginning of the third trimester of 2011.

Design details

Randomization method
In office by a computer. We stratified by the total normalized score on the baseline exam and by geographic regions call sub-counties. In 2011, we conducted a second randomization for the parent sensitization program. Half of the schools in each treatment were assigned to each group. We stratified assignment by the schools’ initial treatment group and sub-county, and checked for balance using the demeaned savings rates from 2010.
Randomization unit
school
Clusters
136 schools
Observations
4,716 students from the 136 schools
Arms
39 cash treatment, 39 voucher treatment, 58 control schools
Primary outcomes
Savings levels (by school and by student), familiarity with the program, use of the program, possession of school supplies, parental involvement, savings attitudes, payment of school fees, test scores

Cited by

No citation data available for this registration yet.

Keywords