Geospatial Impact Evaluation Case Studies

Spatial Impact Evaluation: Land Rights and Deforestation Rates

AidData is collaborating with the German Development Bank (KFW) on a large-scale evaluation of land rights and enforcement on deforestation rates in the Brazilian Amazon. The collaboration employed 30 years of remotely-sensed landcover outcome data to provide a rigorous estimate of program impact. Between 1995 and 2008 (see image to the right), the project demarcated 106 indigenous lands in the Brazilian Amazon. As part of the collaboration, KfW provided detailed boundaries of community lands, administrative data on the criteria and timing for treatment, and additional project documents. AidData processed high-resolution satellite imagery and employed quasi-experimental matching and panel methods to estimate program impacts on forest cover. We tested the effect of demarcation on deforestation rates using propensity score matching and fixed effect techniques.


Value for Money for Foreign Aid Impacts on Carbon Sequestration and Biodiversity

Our interdisciplinary research team along with the World Bank’s IEG developed a framework for a value-for-money (VFM) assessment of the standing tropical forests. By combining contextual information about the World Bank-funded development projects with (a) data on the costs of historic projects, and (b) existing research on the value of tree coverage in terms of carbon sequestration and biodiversity, we provided an analysis of VFM for historic projects, and a recommended methodology for VFM assessments based on the proposed location(s) of future World Bank projects. The final dataset included 19,940 project locations and our analysis also used relevant covariates and outcome variables at each World Bank project site sourced from datasets with global coverage - including population, distance to urban areas, nighttime lights, distance to roads and rivers, a variety of environmental characteristics: climate, slope, and elevation as covariates, and in-situ measurements of biodiversity and satellite estimates of vegetation as outcomes.