How Rwanda Tied Teacher Pay to Student Learning and is Making it Work
Contracts that reward teachers for improving student learning, tracked by a comprehensive teacher and student performance monitoring system, are transforming how the government implements policy to achieve education goals. A collaboration between IPA, the Ministry of Education, the National Examination and School Inspection Authority (NESA), the Rwandan Basic Education Board, and the Georgetown University Initiative on Innovation, Development and Evaluation (gui²de), the program is now part of official policy, reaching 3,548 schools and 107,411 teachers nationally.
The Gaps in Rwanda's Education System
Rwanda has expanded access to education. Turning that access into learning has proven harder. About 35 percent of the entire population is enrolled at some level from pre-primary through higher education (Rwanda 2023/2024 Education Statistical Yearbook), a figure the Ministry of Education uses to explain why so much of its budget goes to getting students into classrooms rather than to the quality of what happens inside them. Nationally representative assessments conducted before the COVID-19 pandemic captured the gap that remained. Among Primary 4 students, 43 percent were performing below Primary 3 expectations in numeracy, and 37 percent fell below them in literacy.
Aline Dukuzimana could see that gap in her own school but had no way to measure it. As head teacher of EP Karama, a primary school in Gasabo District in the northeast of Kigali, she relied on term exams and teacher reports that told her how students scored without telling her whether they had understood the lesson.
"We used to rely on term exams and teacher reports," Dukuzimana recalled. "It was hard to know if students were actually understanding the content, or the component of the curriculum, or the lesson the teacher was applying."
Behind the learning gap stood a workforce that was hard to build and harder to keep. A 2019 analysis by IPA and Georgetown University found that one in five teachers left their posts each year, with about half moving to other schools and about half leaving the profession altogether. The churn left schools covering classes with teachers working outside their subjects. That strain grew sharper after Rwanda adopted English as its language of instruction, according to Andrew Zeitlin, an associate professor at Georgetown University's McCourt School of Public Policy who helped design and evaluate Rwanda's pay-for-performance contracts. The government's task was to attract teachers with the right skills and hold on to the strongest among them.
Rwanda already had an instrument built for exactly this, and it had stopped working. Every civil servant in the country signs an annual performance contract called imihigo, a Kinyarwanda word for target-setting, with a bonus of up to 5 percent of salary attached to the result. In most sectors, the contracts carried weight. In education they had become dormant. Nearly every teacher earned top marks and the full bonus regardless of results, and no two schools judged performance by the same measure. One head teacher might score teachers on whether the classroom was swept and the desks were straight. Not one contract asked whether students were learning.
A contract that paid out no matter what could not motivate anyone, and this one did not. Teachers carrying large classes and heavy workloads saw no link between their effort and their reward. Head teachers, handed subjective criteria, had no fair way to tell a strong teacher from a weak one. Nothing in the system measured what happened in the classroom or tied a teacher's pay to whether children learned.
One of the key gaps in the education system is that classroom data reporting was lagging behind, leaving accountability for teacher monitoring wide open. In 2020, schools reported teacher vacancies inconsistently, leaving the central system without a clear view of how many teachers were needed, in which subjects they were needed, and where they needed to be hired. When teachers left, districts could only provide replacements 77 percent of the time.
Without a comprehensive, standardized system in place that could effectively recruit, maintain, and reward teachers who were moving the needle in improving student learning, Rwanda was going to have difficulties reaching its educational goals.
“The policy instrument was there, the need was there, and we had even seen work from the health sector in Rwanda that had tied payments to health outcomes,” explained Zeitlin. “That made it seem like there was appetite for these ideas.”
From Research Question to Policy Opportunity
The question Zeitlin and his colleagues set out to answer borrowed its logic from the health sector. If tying payment to outcomes had improved health results in Rwanda, could tying teacher pay to student learning do the same in schools? Between 2015 and 2018, Zeitlin and his co-researchers Clare Leaver, Owen Ozier, and Pieter Serneels worked with IPA Rwanda and the Rwanda Education Board (REB) to find out. They ran a randomized evaluation across 164 primary schools in six districts. Some schools offered teachers contracts that rewarded teachers for student learning gains and overall teaching quality; others kept the standard fixed-wage contracts.
The study revealed that the performance-based contracts raised student learning by the equivalent of an extra year of learning for each year spent in school by the second year of implementation, effectively doubling the rate of learning growth.
There was now evidence that tying teachers’ bonuses to student achievement was an effective instrument. Carrying that result into national policy took a combination of changes inside both IPA and the government over the next several years.
The first change came in 2022 when the government of Rwanda raised salaries by 40 percent for secondary school teachers and 88 percent for primary school teachers (Ministry of Education, 2022). The increase created a natural test. If teachers were already better paid, would a performance bonus still change how they worked?
A second piece had been in place since 2019, when IPA and gui²de established an Embedded Evidence Lab inside the Ministry of Education. Working alongside officials from the Rwanda Basic Education Board and the National Examination and School Inspection Authority (NESA), the lab used evidence from the pay-for-performance evaluation to inform policy decisions around performance-based contracts and teacher monitoring.
The third was infrastructure. Through the Embedded Evidence Lab, the Ministry developed two national data systems that feed into the teacher contracts. In 2019-2020, the Ministry of Education (MINEDUC) launched the Comprehensive Assessment, a nationwide examination initiative that aims to standardize the evaluation process and tests all primary and secondary students across subjects at the end of each academic year. This was complemented by the
Comprehensive Assessment Management Information System (CAMIS), which was also launched in the same time frame and rolled out nationally with IPA’s support in 2022-2023. CAMIS serves as the digital backbone for student learning data—such as test scores—which would be regularly monitored and used to judge teachers’ performance. The other data system is the Teacher Management Information System (TMIS). Created after the centralization of teacher recruitment by the national government in January 2019 and supported by IPA from 2021 onward, TMIS tracks teacher recruitment and retention.
With the evidence, the institution, and the data in place, the work shifted from proving a point to designing a policy. In the 2022-23 school year, IPA started working with the Georgetown University Initiative on Innovation, Development and Evaluation (gui²de) and a government task force. Together they designed and refined contracts under an approach called Supporting Teacher Achievement in Rwandan Schools (STARS).
The performance contracts would now evaluate teachers on a scale in five key areas: presence, preparation, pedagogy, participation in CAMIS, and pupil learning gains.
Presence
Classroom attendance, verified through unannounced visits
Preparation
Quality of lesson planning, assessed during unannounced school visits
Pedagogy
Teaching quality, evaluated through classroom observations
Participation in CAMIS
Accurate and consistent entry of student assessment data into the national system
Pupil Learning Gains
Student progress measured against previous performance
The bonuses paid to teachers were also now standardized based on this evaluation scale.
Teachers who scored below 70 would receive no bonus.
Those scoring between 70 and 79 receive a 3 percent bonus.
Teachers who scored 80 or higher receive a 5 percent bonus.
Where the old contract had paid nearly everyone the same regardless of what happened in the classroom, the new one rewarded teachers against a common, visible measure. Teachers welcomed this change, owing in large part to the extensive communication and support that accompanied the rollout.
"The initial response of teachers has been 'finally we know exactly what we have to do.'"
Adia Umulisa, Head of Education Sector Planning, Monitoring & Evaluation Department at the Ministry of Education, Rwanda
Why Performance Contracts Were Effective
The original evaluation had shown two things, and the second mattered more than the first. The contracts drew slightly stronger teachers into the profession, which gave them value as a recruitment tool. But their larger effect was on the teachers already in post, who showed up more reliably and taught better once their pay was tied to how their students did. The incentive changed the behavior of the existing workforce, not only the makeup of the incoming one.
Performance contracts were yielding visible changes in how all teachers—not just the ones who were already putting in high effort—were teaching and how students were actually learning better.
"In Primary 5, the teacher had a struggle with reading comprehension. After the teacher started using weekly assessment data, the teacher was using a small group for helping the weak learners. Now the students are confident and participate actively.”
Aline Dukuzimana, Head Teacher of EP Karama Primary School
With the evaluation results in hand as a proof of concept, IPA brought the case to senior government officials, among them the Minister of Finance, the Minister of Education, and representatives from the Office of the President. Those conversations set up the next phase of the work. The contracts had been shown to raise learning; the open question now was whether the government could run them itself, across a national system.
Evidence Becomes Institutionalized
There was now strong evidence that the pay-for-performance contracts from the evaluation worked and a government committed to adopting them as the new imihigo contract.
With that, the project’s scope shifted explicitly. It was no longer whether pay-for-performance raised learning, but which version of the contract raised it most, and whether the government could run that version on its own.
This is where the Embedded Evidence Lab that IPA and gui²de established with the Ministry of Education in 2019 stepped in to answer these questions. The Lab goes beyond research; it seeks to institutionalize the evidence, to ensure that it is being used and acted upon.
Through the CAMIS and TMIS systems, which the Lab co-built and helped to strengthen, data from schools and districts on students and teachers were now centralized. With a civil service required and trained to use it, more informed decisions could be made about how to boost student learning and teacher performance.
The Lab’s activities have even extended to fostering institutional change within the government. This includes providing technical input on policies, legal frameworks, teacher management guidelines, and inspection processes. By doing so, the pay-for-performance contracts would thus become well integrated into national education policy.
Yet this process did not happen in a short timeframe but rather over the course of three years, from the 2022-23 school year through the 2024-2025 school year. During this time, gui²de and IPA designed and tested eight teacher contracts across 345 schools to identify the most effective version while the Embedded Lab worked on increasing the government’s ability to manage the program sustainably
"The multi-year approach actually allowed the entire STARS team, with gui²de and IPA, to iteratively refine both the contract design, as well as develop the supporting systems, and based on what was working and where the gaps remained."
Arnesh Chowdhury, Program Director, STARS, gui²de
Year 1 (2023)
- 248 schools, 4,762 teachers, 207,523 students across 10 districts and 62 sectors
- 5 contract models tested
- Teacher performance improved: better lesson plan preparation, instructional quality, and CAMIS data entry
- No measurable gain in student learning yet, consistent with existing pay-for-performance literature
Year 2 (2024)
- 345 schools, 6,664 teachers, approximately 304,000 students across 10 districts and 100 sectors
- Dropped one variation from the previous year and added one more contract model into the current mix: introducing head teacher imihigo contracts
- Teacher preparation and classroom practices remained strong; CAMIS participation rose 5 percentage points
- Models involving headteacher evaluations, headteacher imihigo, and student re-testing began to show early learning gains. Less effective models were dropped.
Year 3 (2025)
- 345 schools, 6,485 teachers, 279,047 students
- Six contract models refined and tested
- All models made teachers 11 to 15 percentage points more likely to have a lesson plan ready for class.
- The strongest model, which rewarded both teachers and headteachers based on student learning, produced 0.6 years of additional learning compared to similar students in other schools.
After three school years of designing, testing, and iterating pay-for-performance contracts and working to build and maintain effective student and teacher data systems and institutionalizing the performance contracts for policy, IPA & gui²de through the Embedded Lab accomplished the following:
- Completed an infrastructure redesign and a new version of CAMIS to support the following scale-up, with a continuous strategy to improve data quality, system adoption, and use for decision-making
- Expanded the system to include secondary school students, identification of at-risk learners, and support for remedial students
- Tracked learning outcomes for 89 percent of students
- Integrated TMIS with Integrated Personnel and Payroll Information System (IPPIS) salary validation by the Ministry of Finance and Economic Planning
- Finalized the contract model that would become the new imihigo: Rewarding both teachers and headteachers based on student learning
- Completed series of capacity building on learning on imihigo learning gains calculation and overall imihigo score calculation
- Supported NESA in issuing national repetition guidelines and REB in launching a remedial tutoring program informed by CAMIS data
- Upgraded TMIS to include more data on teacher management, enabling education data systems integration for more effective policy decisions to be made
- Developed a maternity leave module to manage teacher leave and replacements
- Developed a teacher scholarship module to support teacher development, including study leave and efforts to address gender gaps in leadership
- Developed a module to manage Technical and Vocational Education and Training (TVET) teachers
- Improved the teacher placement process by incorporating teacher preferences and matching algorithms
The imihigo contracts were now ready to be scaled nationally.
Performance Contracts as National Policy
In the 2025-2026 school year, the performance contracts became national policy. The Ministry of Education rolled out the performance contracts across Rwanda, reaching 3,548 schools and 107,411 teachers at the primary and secondary school level.
The landmark national Teacher Development and Management Policy that was approved in November 2024 officially introduced the new imihigo contract. Under the contract, head teachers and teachers would be evaluated for their performance based on their students’ learning. The imihigo contract is one of many interventions in a broad initiative to support and motivate teachers between 2025 and 2030.
Making the contracts work at national scale meant building them into the systems the government already ran. The Ministry of Education digitized the imihigo contract and integrated it with CAMIS and TMIS. By doing so, head teacher and teacher performance monitoring is more transparent and driven by data while accountability increases.
Moreover, the government rolled out training on how to use this fully digitized system to district education staff, Sector Education Inspectors, head teachers, and teachers. By covering all levels of civil servants, this builds comprehensive capacity to use and sustain the CAMIS, TMIS, and imihigo systems, which will in turn strengthen the potential for policies that generate stronger impact.
Because the logistical complexities of testing and refining the contracts and building up the CAMIS and TMIS systems were resolved over the previous three years, the national implementation is occurring more smoothly than it may have otherwise been. Still, during this scale-up, some work remains, including extending the contracts to vocational schools.
The Ministry is currently working with IPA and gui²de to develop performance indicators for teachers and students in vocational schools, and the new contracts will be unveiled at the start of the 2026-2027 school year. It’s also exploring how teachers’ value-add—how well teachers are improving student learning—can inform promotion decisions at schools.
IPA’s role through this phase is to continue to ensure that the incentives in the performance contract are aligned with the broader policy objective of improving student outcomes and driving educational development.
The trajectory of pay for performance contracts, from a policy window, to a study evaluating their effectiveness, to a multi-year research project as STARS, to national policy, showcases the impact of involving governments from the beginning. Not only does evidence transfer into the hands of a structure that can implement a program at this level, but it fosters systemic, long-term ownership.
As the contracts are embedded into official policy and impacts materialize, challenges and opportunities will naturally form that will compel the government to prioritize to further improve the country’s educational outcomes.
“As implementation may drift, how will that be picked up on and corrected quickly? How can this become a living instrument of policy and not just a kind of static object that was tied to priorities at one moment in time?”
Andrew Zeitlin, Associate Professor, Georgetown University's McCourt School of Public Policy
Given the government’s continued active investment in the STARS evidence and its comprehensive steps to maximize the civil service’s capacity to sustain the program, those closely involved in the program are optimistic about its potential as an official education policy.
If the past decade of extensive evidence generation, refinement, system strengthening, and collaboration serves as a barometer for the future, there is good reason to be confident that Rwanda’s 4 million students will receive a stronger education.
Written by Michael Podesta and Florence Dzame | Videos Edited by Oswald Suonbai-kyi | Designed by Cara Vu
Special thanks to Andrew Zeitlin and Arnesh Chowdhury, the Georgetown University Initiative on Innovation, Development and Evaluation (gui²de), the Ministry of Education of Rwanda, Karla Petersen, Juan Hernández-Agramonte, and IPA Rwanda (Jean Damascene Nsengiyumva, Sophie Mushimiyimana, and Cassien Havugimana) for their review and contributions.
All photos courtesy of gui²de and the IPA Rwanda team.
Built with Shorthand