This week, the Strauss Center’s CCAPS Program and AidDatapublished a first-of-its-kind geocoded dataset from the Government of Malawi's Aid Management Platform. These data are also available through the CCAPS dashboard, which enables users to overlay and visualize data on climate change vulnerability, conflict, and aid, and analyze how these issues intersect in Africa. The geocoded aid dataset was created in partnership with Malawi's Ministry of Finance (MoF). It includes sub-national geocodes for approximately 550 aid projects undertaken in Malawi since 2000, representing nearly $5.3 billion in total commitments or roughly 80% of all aid reported to the MoF during that time. The dataset is the first effort of the kind envisioned by the Open Aid Partnership, an initiative spearheaded by the World Bank that seeks to make local information on donor activities more accessible to improve the targeting, coordination, and delivery of aid.
AidData and CCAPSresearchers have begun using these data to explore the sub-national drivers of aid allocation in Malawi. We are still working to create robust models, but initial results suggest that population density may be the most important determinant of allocation patterns, more so than poverty, access to electricity, or other measures of need (see graphs below prepared by BYU faculty member and AidData/CCAPS researcher Mike Findley).
This pattern can be seen in a map as well where it’s clear the less-densely populated Northern region has comparatively fewer projects, despite relatively high levels of poverty. Conversely, the main population centers—Lilongwe (the capital city) and Blantyre (in the Southern region)—seem to attract dense clusters of aid projects.
These represent very preliminary findings, but they should give readers an idea of how geocoded aid information can be put to productive use. We will continue to update The First Tranche with research on aid distribution and impact in Malawi and are excited to see what others will do with these data now that they are available to the public. To download the full dataset, go here.