How Does the Quality of Aid Project Management Affect Violent Conflict in Africa?

This study investigates whether the quality of aid project management affects violent conflict in Africa. Using geo-coded data on all World Bank projects from 1995–2014, the researcher links project outcomes to the quality of assigned project leaders, measured by performance scores. Combining this with conflict data from ACLED, the study finds that better-managed aid projects significantly reduce violent conflict, while poorly-managed ones increase it. A one-point improvement in project performance score lowers the probability of conflict by 1.8–2.4 percent; moving from worst to best project quality reduces conflict likelihood by 10–12 percent. Project supervision is especially important.

Full Citation and Link to Article

Moscona, Jacob. “The Management of Aid and Conflict in Africa.” American Economic Journal: Economic Policy(Forthcoming). Retrieved from AEA website

Extended Summary

Central Research Question

This article investigates a longstanding and contested issue in international development: does the management quality of development aid projects influence violent conflict outcomes in aid-recipient countries? Specifically, the study examines the causal relationship between World Bank project management and the incidence of violent conflict in sub-Saharan Africa. The central question is whether differences in how aid is managed—especially at the level of project leaders—can explain the mixed findings in prior research on whether aid increases or reduces violent conflict.

Previous Literature

The relationship between foreign aid and conflict has produced mixed empirical results. Some studies suggest that aid can exacerbate conflict (Crost et al., 2014; Nunn and Qian, 2014; Dube and Naidu, 2015), while others find aid has no effect or even reduces violence (Collier and Hoeffler, 2002; De Ree and Nillesen, 2009; Beath et al., 2017). Much of this variation is attributed to differences in recipient characteristics or aid types. However, less attention has been paid to the role of how aid is implemented. Qualitative accounts—such as those by Prendergast (1996) and Anderson (1999)—suggest that the management and monitoring of aid projects can critically shape whether aid becomes a trigger for violence or a force for peace. Despite these insights, the role of aid management quality has remained largely unexamined in rigorous empirical work.

This paper fills that gap by focusing on management quality, particularly at the level of project leaders assigned by the World Bank. The author argues that variation in management, especially in project monitoring and oversight, can critically alter the outcomes of aid delivery in fragile environments.

Data

The study combines several rich datasets:

1. World Bank Project Data: The core data come from geo-coded World Bank development projects in Africa from 1995–2014, compiled by AidData. Each project includes detailed information on its sector, start and end dates, and the identity of the Task Team Leader (TTL), the official in charge of implementation.

2. Performance Scores: Project quality is measured using the World Bank Independent Evaluation Group’s (IEG) performance scores, which rate projects on a six-point scale. These scores capture overall quality, supervision quality, and performance at entry.

3. Conflict Data: Violent conflict outcomes are drawn from the Armed Conflict Location and Event Data Project (ACLED), which provides geo-coded conflict event information, including types of violence (e.g., battles, violence against civilians) and whether government forces were involved.

4. Other Controls: The analysis includes spatial grid-cell identifiers, year and sector fixed effects, and controls for resource endowments and previous conflict exposure.

Methods

The study exploits variation in the quality of World Bank project leaders—measured using performance scores and leader identity—to estimate the causal effect of aid management on conflict. The empirical strategy hinges on two core techniques:

1. Instrumental Variables (IV) Strategy: Project leader identity is used as an instrument for project quality. Since TTLs are assigned prior to project design and location decisions, and based largely on availability and sector expertise, their assignment is plausibly exogenous to local conflict trends. This method identifies the causal impact of project management quality on conflict outcomes.

2. Split-Sample Value-Added Approach: As an alternative, the paper estimates “leader effects” on conflict directly using a value-added approach. By isolating leader-specific contributions to conflict across projects and locations, the author demonstrates that some leaders consistently manage projects in ways that reduce conflict, while others do the opposite.

The core outcome variable is whether a violent conflict event occurred in a given one-by-one-degree grid cell in a given year. The author focuses both on the extensive margin (whether conflict occurs) and the intensive margin (how many conflict events occur), as well as disaggregating effects by conflict type and actor.

Findings / Size Effects

The study’s central finding is that better-managed aid projects reduce violent conflict, while poorly-managed projects increase it. Several key results support this conclusion:

• A one-point increase in the IEG project performance score reduces the likelihood of violent conflict by 1.8–2.4 percentage points. Since the baseline conflict probability is 17%, this represents a 10–12% relative reduction in conflict likelihood.

• Moving from a project with the lowest possible score to the highest reduces the probability of violent conflict by as much as 10–12%.

• Conversely, poorly-managed projects increase the probability of conflict compared to regions with no projects at all.

• The impact is especially large in high-risk settings: regions with a recent history of conflict and projects that involve large, appropriable resources (e.g., energy or infrastructure projects) are more sensitive to management quality.

• Project supervision is the most important component of management quality. The effect of the supervision score on conflict is substantially larger than the effect of performance at entry (design phase), which is statistically insignificant.

• Conflict intensity is also affected: well-managed projects reduce not just the incidence but the scale of conflict. The effects are strongest for conflicts involving fatalities and violence against civilians.

• Effects are driven by violent, confrontational conflict—especially battles and government-rebel clashes—not by non-violent strategic activity such as rebel base-building or land transfers.

• Effects are most pronounced in later project phases, after most funds have been disbursed. This suggests that violence is more about predation of aid resources than strategic disruption of implementation.

• Leader-specific “value added” estimates confirm that individual TTLs matter. Leaders in the 75th percentile of the conflict-reduction distribution are associated with a 15% lower conflict probability than those in the 25th percentile.

• Leader effects explain more variation in conflict than lending sector or country-year variation. This highlights the substantial influence of aid administrators, rather than just policy type or recipient context.

Conclusion

This article provides compelling empirical evidence that aid management—especially the quality of project supervision—plays a pivotal role in shaping whether development aid fuels or mitigates violent conflict in Africa. By leveraging novel data on World Bank project leaders and conflict outcomes, the author offers a robust and policy-relevant answer to a longstanding puzzle in the aid literature.

The implications are significant. First, they help reconcile contradictory findings in past work: the same type of aid can have opposite effects on conflict depending on how well it is implemented. Second, they point to actionable levers for reform. Improving bureaucratic quality, investing in project monitoring, and professionalizing aid management may be as important as targeting the right sectors or countries.

The study also opens new avenues for research. The bureaucratic channels through which aid affects governance and violence merit further exploration. Moreover, while the focus here is on Africa and the World Bank, the insights likely extend to other multilateral and bilateral donors.

In sum, the article makes a strong case that better aid management is not merely an administrative concern—it is a peacebuilding strategy.