Chinese vs. World Bank Development Projects: Insights from Earth Observation and Computer Vision on Wealth Gains in Africa, 2002-2013
🤖 AI Summary
Disagreement persists regarding the welfare impacts of development projects on African populations, primarily due to observational bias and the scarcity of community-level wealth data. This study leverages sectoral projects implemented by China and the World Bank across 9,899 communities in 36 African countries (covering 88% of the continent’s population) from 2002–2013. We construct a novel machine-learning-based Wealth Index (IWI) integrating nighttime light and daytime satellite imagery. Employing inverse-probability weighting and high-dimensional fixed-effects panel models, we achieve the first continental-scale causal identification combining multi-source remote sensing and tabular data. Key contributions: (1) Chinese projects exhibit greater political unpredictability in site selection, whereas World Bank projects are more predictable from geographic–economic covariates; (2) both project types significantly increase community-level wealth, with Chinese projects yielding larger average effects (e.g., +14.32 IWI for emergency-response projects); (3) incorporating satellite imagery substantially attenuates estimated treatment effects, indicating that conventional approaches likely overestimate true impacts.
Application Category
📝 Abstract
Debates about whether development projects improve living conditions persist, partly because observational estimates can be biased by incomplete adjustment and because reliable outcome data are scarce at the neighborhood level. We address both issues in a continent-scale, sector-specific evaluation of Chinese and World Bank projects across 9,899 neighborhoods in 36 African countries (2002 to 2013), representative of 88% of the population. First, we use a recent dataset that measures living conditions with a machine-learned wealth index derived from contemporaneous satellite imagery, yielding a consistent panel of 6.7 km square mosaics. Second, to strengthen identification, we proxy officials' map-based placement criteria using pre-treatment daytime satellite images and fuse these with rich tabular covariates to estimate funder- and sector-specific ATEs via inverse-probability weighting. Incorporating imagery systematically shrinks effects relative to tabular-only models, indicating prior work likely overstated benefits. On average, both donors raise wealth, with larger gains for China; sector extremes in our sample include Trade and Tourism for the World Bank (+6.27 IWI points), and Emergency Response for China (+14.32). Assignment-mechanism analyses show World Bank placement is generally more predictable from imagery alone, as well as from tabular covariates. This suggests that Chinese project placements are more driven by non-visible, political, or event-driven factors than World Bank placements. To probe residual concerns about selection on observables, we also estimate within-neighborhood (unit) fixed-effects models at a spatial resolution about 450 times finer than prior fixed effects analyses, leveraging the computer-vision-imputed IWI panels; these deliver smaller but directionally consistent effects.
Problem
Research questions and friction points this paper is trying to address.
Evaluating wealth impacts of Chinese vs World Bank development projects in Africa
Using satellite imagery and machine learning to measure neighborhood living conditions
Comparing project placement predictability and treatment effects across funders and sectors
Innovation
Methods, ideas, or system contributions that make the work stand out.
Machine-learned wealth index from satellite imagery
Inverse-probability weighting with fused satellite-covariate data
High-resolution fixed-effects models using computer vision panels
🔎 Similar Papers
No similar papers found.
