Addressing challenges from lack of data in disbursement of aid to education to low-income countries Evaluation of primary completion rates estimates

Aid to education may impact the country's progress towards achieving universal primary education. Yet, the data which donor institutions use to allocate aid is often missing or miscalculated. That is likely to result in misallocation of aid and impact country's ability to achieve universal primary education. This paper addresses a challenge of the missing Primary Completion Rate data, which Global Partnership for Education (GPE) uses to estimate aid to education allocations. It compares three methods of estimating missing Primary Completion Rate: imputation from average PCR levels (Method 1); inference from the past level of PCR (Method 2), and imputation from country characteristics (Method 3). The efficiency of the methods is evaluated by comparing the magnitude of the estimated error term obtained using each of the methods and reported PCR levels for 44 low income counties. Results show that inference from the past level of PCR outperforms the other two methods. Simulations using the estimated results also reveal that the average deviation from the true allocation may reach 86 percent and 71 percent for Methods 1 and 3, and 33 percent for Method 2.