Power Plants, Roads, Clinics: Which Aid Projects Actually Change Lives?

The comparison between China and the World Bank is highly relevant. The World Bank is often associated with social sectors, public systems, and human-capital investments. China’s development finance in Africa is frequently linked to infrastructure, energy, transport, and utilities. Policy debates often treat these as competing development models. The study suggests that this framing is too simplistic. What matters is not just who finances a project, but what the project does, where it is placed, and whether its benefits reach households.

A new study of development projects across Africa reveals that foreign aid can improve local wealth, but not uniformly, not automatically, and not always in the sectors policymakers might expect.

The research, Temporal Dynamics of Development Aid in Africa: Evidence from a Staggered Difference-in-Differences Study of China and World Bank Projects, examines whether World Bank and Chinese development projects improved material living standards in African communities between 2002 and 2013. Authored by Mattias Antar, Adel Daoud, and Connor Jerzak, the paper analyzes 2,166 Demographic and Health Survey clusters across 35 African countries, linking geocoded aid projects from AidData to satellite-imputed estimates of the International Wealth Index.

The study findings suggest that neither China nor the World Bank consistently raises local wealth across all sectors. Instead, positive results are concentrated in specific donor-sector combinations.

From grand aid debates to neighborhood realities

The study reflects a major shift in aid research: from asking whether foreign aid boosts national growth to asking whether projects improve conditions where they are actually implemented. Aid is experienced locally. For instance, a clinic can improve healthcare access for nearby households, a school can expand educational opportunities within a community, a road may shift trade patterns while benefiting some areas more than others, and a water project can transform daily life in ways that national growth statistics often fail to capture.

The researchers also move beyond nighttime lights, a widely used proxy for local economic activity. Nighttime lights can capture electrification, urban expansion, and commercial activity, but they may miss improvements in housing, sanitation, assets, health access, or household services. This is particularly important in parts of Africa where welfare gains may occur without dramatic increases in luminosity.

Instead, the study uses the International Wealth Index, which measures material living standards through household assets and living conditions, including electricity, sanitation, water access, housing quality, consumer durables, and crowding. The index is satellite-imputed from a wealth surface developed using household survey data and satellite imagery. This gives the authors a way to track local material welfare over time, rather than relying only on visible economic activity.

The comparison between China and the World Bank is highly relevant. The World Bank is often associated with social sectors, public systems, and human-capital investments. China's development finance in Africa is frequently linked to infrastructure, energy, transport, and utilities. Policy debates often treat these as competing development models. The study suggests that this framing is too simplistic. What matters is not just who finances a project, but what the project does, where it is placed, and whether its benefits reach households.

The strongest gains appear in health, water, and social infrastructure

Among World Bank projects, health shows the clearest positive evidence. Under the authors' preferred method, World Bank health projects are associated with delayed but meaningful gains in local wealth, reaching an estimated 4.42 points on the International Wealth Index at the final observed horizon. The result is especially important because the diagnostic evidence for this sector is relatively clean, meaning treated and comparison areas do not show strong signs of diverging before project onset.

Health projects rarely transform household wealth immediately. Their benefits may emerge through lower medical costs, better labor productivity, improved child health, and stronger household resilience. Over time, these channels can support asset accumulation and better living conditions.

World Bank education also shows positive estimated gains, reaching about 3.69 IWI points under the preferred estimator. However, the authors interpret this result more cautiously because treated areas already showed different wealth dynamics before projects began. In practical terms, education may be beneficial, but the study cannot isolate its impact as cleanly as it can for health.

For Chinese projects, the most promising findings appear in water supply and sanitation and other social infrastructure and services. Chinese water and sanitation projects reach an estimated 7.08 IWI points at the final horizon, while other social infrastructure reaches around 4.44 points. These sectors have direct links to household welfare. Better water access, sanitation, and community infrastructure can affect daily living conditions in ways that a household wealth index is designed to capture.

Both Chinese water and social infrastructure projects show significant pre-treatment imbalances. This means projects may have been placed in areas already experiencing weaker conditions or different trends. The observed gains are therefore best understood as promising associations rather than definitive proof of causal impact.

Why big infrastructure does not automatically mean local wealth

Under a conventional two-way fixed effects model, Chinese energy generation and supply appears to deliver large local wealth gains, reaching 7.29 IWI points. However, under the preferred switcher-stayer estimator, the estimate drops to about 0.16 and is not statistically meaningful. This finding does not imply that energy projects lack economic value. Such investments may still support firms, industrial activity, grid expansion, public facilities, or broader regional growth; however, within the study period and measured through a household-centered welfare lens, the results do not show clear local wealth gains for nearby communities.

Chinese transport and storage projects also become less favorable under the preferred design. A conventional model suggests a positive impact, but the preferred estimator turns negative at the medium-run horizon. The authors do not claim this proves harm. Possible explanations include construction disruption, benefits flowing to corridor nodes rather than nearby households, local activity shifting elsewhere, or gains appearing beyond the study's time window.

The finding is important for infrastructure-led development strategies. Roads, power plants, and transport corridors may be essential for growth, but their benefits are not automatically inclusive. Without complementary policies such as local connections, affordable access, safeguards, local employment, market integration, and community planning, large infrastructure may generate macroeconomic value without quickly improving household welfare nearby.

The policy lesson: measure impact, not just delivery

The study shows that the way researchers estimate aid impact can change the policy story. Aid projects are not randomly placed. Donors and governments select locations based on need, feasibility, politics, strategic value, access, or administrative capacity. If treated areas were already poorer or on different trajectories before projects began, simple comparisons can misread recovery or pre-existing trends as aid impact.

To address this, the authors compare a conventional two-way fixed effects event-study model with a switcher-stayer estimator designed for staggered treatment timing. The preferred method avoids some contaminated comparisons that arise when locations receive projects at different times. It does not solve every problem, but it provides a more cautious and credible reading.

The diagnostic results are sobering. Across 23 donor-sector panels, many treated areas show significant pre-treatment differences. This means that selection bias is not a minor technical issue; it is central to interpreting aid effectiveness.

Governments must evaluate development finance by sector-specific welfare pathways, not donor identity or project visibility. For multilateral institutions, the findings support continued attention to health and service delivery, where household welfare channels may be clearer. For China and recipient governments, the results suggest that water and social infrastructure can support local welfare, while energy and transport projects may need stronger complementary policies to ensure benefits reach communities.

For development agencies and NGOs, the study reinforces the importance of monitoring lived outcomes, not only project completion. For investors, it is a reminder that infrastructure returns and household welfare gains may follow different timelines and reach different beneficiaries.