By Tyler Hook, University of Wisconsin
In September 2016, the Ministry of Education of Liberia officially launched a public private partnership called the Partnership Schools for Liberia (PSL) pilot, with 8 providers operating 93 schools. Promoted with the aim of dramatically improving learning outcomes for children, in an equitable, cost-effective, and sustainable manner, the partnership drew strong responses from critics and supporters in the international education and development world.
Amid this contentious environment, the midline report by the randomized control trial (RCT) team was released earlier this week.
Conducted by researchers at the Center for Global Development in partnership with Innovations for Poverty Actions, the midline report looks at 185 randomly assigned schools. These are divided into treatment-partnership schools and control-traditional public schools, drawing from a total of 2,619 primary public schools located in 13/15 counties in Liberia.
The report shows significant learning increases for PSL students, but at an unsustainable cost and with potential negative impacts on non-PSL public schools. It highlights the necessity of further research that engages in a variety of questions, methods and frameworks to evaluate the full impact of the intervention should it be continued.
Unsurprisingly, considering the large increases in spending, resources, training and accountability/attention under PSL, the midline report shows improvement in student learning between treatment (partnership) and control (traditional public schools), raising student learning an additional 60%, or the equivalent of 0.6 years of schooling over the first year.
The report further finds increased teacher attendance and quality of instruction, increased percentage of students having a textbook and being in school, and increases in parent and student satisfaction with their school. The RCT midline report also found that enrollment increased overall across providers, though it decreased in schools and classes with large enrollments, particularly for one provider: Bridge International Academies (BIA). 
Questions and Concerns
Despite noted improvements, the midline report reinforces some of the concerns raised by critics, including large spending differences amongst providers, discouraging cost-effectiveness results, and evidence of some providers pushing out students and teachers, potentially negatively impacting students, teachers, and surrounding public schools.
According to the midline report, providers are spending extremely divergent amounts of money in year 1, from just 57 USD per student by Youth Movement for Collective Action, to an estimated 663 USD by BIA (cost per-pupil). BIA’s budget per-pupil is significantly higher, at an estimated 1052 USD, over 20 times more than the average traditional public school.
Per-pupil cost estimates were self-reported and included both recurring and start-up costs. These costs are far beyond the planned 100 USD per child discussed in previous PSL documents, and come despite the fact that schools in year 1 were not a representative sample of schools in Liberia, with the report stating that “most contractors have schools with better infrastructure than the average public school in the country” (p.42).
According to BIA’s Memorandum of Understanding with the Ministry of Education , BIA also had several other stipulations, requesting only schools located near main roads, with good mobile access, a sufficient number of desks, enclosed classrooms, pit latrines, and recently built and refurbished classrooms, clustered to “ensure cost-effective oversight” ( ).
While contractors claim that costs will decrease substantially as they scale-up, whether or not this will be the case remains to be seen, especially as providers move into harder to reach and more rural parts of the country in year 2.
The report also delivers rather discouraging results in the area of cost-effectiveness, finding that the first year of PSL was not as cost-effective as other programs in developing countries evaluated by randomized controlled trials (p.49). This despite the fact that the researchers calculated costs assuming a cost of $50 per pupil, even though some providers spent significantly higher amounts (individual contractor cost-effective numbers were not given).
There is also reason to believe that the cost of the program may be worse than it appears, as costs within the trial were self-reported, and thus may not reflect the full costs of the provider, who may push various expenses onto workers.
Cost and learning outcomes may also be complicated by expenses made through partnerships with other outside organizations, for example Akon’s solar light partnership with BIA schools. As discussed in previous reports, in order to generate robust cost-effectiveness data, it is essential that research goes beyond reported numbers to examine the experiential cost of various PSL activities.
Overall, issues with cost-effectiveness clearly impact PSL’s long-term sustainability, as the report notes that the program “risks becoming too expensive for the government to contemplate, even in optimistic scenarios”.
The two most disturbing findings in the midline report are the exiting of teachers and the pushing out of students from BIA schools. The report notes that in BIA operated schools, 74 percent of teachers were dismissed or exited during the first year. The report also notes that BIA’s class caps of 45-55, along with the shutting down of second shifts, resulted in the “removing of thousands of students from schools where class sizes were large” . Both of these findings are particularly troubling given that BIA will more than double its number of schools in year 2 of the pilot, bringing its total to 68 schools. Furthermore, if we assume, as the RCT team does, that exiting students and teachers are being pushed onto neighboring schools, serious questions arise as to how PSL is impacting Liberia’s overall school ecology.
Indeed, there have already been reports of overcrowding and other negative impacts on surrounding public schools. Considering that PSL schools receive enormous increases in spending and inspection, and, as the midline report notes, had “first pick of better trained” recently graduated teachers, this situation has the potential to create a two-tiered education system in Liberia.
Questions Remain, Further Research Needed
The midline report shows what people on both sides of the PSL debate predicted: large increases in learning, but at an enormous cost, impacting on surrounding schools and environments. While the report attributes half of all learning increases to the composition of teachers , further research surrounding what is behind these learning increases, contractor spending, and the overall impact on surrounding schools and communities is needed.
As our research team has repeatedly noted, this research must look at how neighboring schools are being impacted, and how teachers, students, and community members are experiencing and understanding PSL. Amid all this data, there remain questions and concerns that demand further investigation before any definitive conclusions can be made.
 Supporters viewed PSL as providing quick and abundant change, transforming Liberia’s education system from a “mess to the best,” while critics questioned the sustainability, equitability, cost-effectiveness, and legality of outsourcing school management to private providers.
 The midline report follows the publishing of a baseline report conducted 2-8 weeks after the school year began. The baseline report attributes differences between the samples to early treatment effect, and finds small, positive short-term impacts on teacher and principal behavior and student enrollment. This baseline report, along with recently released midline reports by 4 of the 8 service providers, has been repeatedly used by service providers and the MoE as evidence of PSL’s success and justification for its expansion.
 Measured by whether or not students think going to school is fun or not (p.35)
 This provider was originally called Liberian Youth Network or LIYONET. Youth Movement for Collective Action is one of two Liberian owned and operated providers. Stella Maris is the other and did not fully operate in year 1 of the pilot.
 The plan was a 50 USD subsidy by the government, plus an additional 50 USD per student coming from a philanthropic subsidy.
 Funded by the Global Partnership for Education
 Summary, p.3
 Qualitative research that looked at both PSL schools and neighboring schools was rejected by the MoE in May 2017.