Is There Proof That Conflict Prevention Really Works?

Arguments made in support of shutting down USAID relied heavily on a distrust in the effectiveness and relevance of humanitarian aid. It’s critical we debunk these myths and show how humanitarian programs can, in fact, show credible reductions in violence. Investment in peace and conflict prevention is still needed and we have ways to show why, if we build on research designs that prove causality.

According to the Institute for Economics and Peace (IEP), countries spent $47.2 billion on peacebuilding and peacekeeping in 2024.¹ This expense may seem large, but in comparison to the costs the world suffers due to violence, which the IEP estimated at close to $20 trillion, these investments in peace are shamefully small. This imbalance has led to persistent calls on governments and other donors to increase spending on peacebuilding. That need has only increased in the last year after the cancellation of most funding for peacebuilding by the United States Agency for International Development (USAID), and reductions in funding by many other governments.

Although there is broad agreement on the need for more resources to promote peace, we lack clear evidence on how to spend them most effectively. Scientific research on the effects of interventions designed to reduce deaths from violent conflict remains scarce. In large part, this problem reflects the reality that conducting high-quality, ethically administered research on peacebuilding efforts is extraordinarily difficult. Surprisingly, though, few governments, NGOs, or scholars have even attempted this kind of research. Most interventions designed to promote peace are not optimally structured to evaluate their effects on violence. Indeed, many do not even attempt to measure their effects on violence at all, instead focusing only on measuring intermediate outcomes such as the number of individuals who received aid or training, or the number of combatants who were demobilized.

Scientifically demonstrating that an intervention has reduced violence is challenging, but progress is possible. The field of development economics, for example, faced similar challenges to measuring the effectiveness of economic aid programs. Beginning in the 1990s, economists began to push for the use of randomized controlled trials and other scientific methods in impact evaluations of aid initiatives. Many of these studies revealed that large and widely supported aid programs had little or no effect on economic well-being. Fortunately, some studies also began to provide evidence for interventions that did work, like those from Poverty Action Lab.² Today these methods are widely used in academic research and they are demanded by governmental and NGO donors to international development programs.³ Bringing about this change required overcoming resistance and skepticism from scholars and practitioners who had spent decades advocating for and administering aid in traditional ways without the benefit of insights from scientific impact evaluations.

I believe the peacebuilding community needs a similar revolution to spur scholars, practitioners, and funders to establish a foundation of evidence upon which future peacebuilding efforts can progress. To do so, however, we need examples of how this kind of research can be accomplished.

In this brief essay, I describe one such effort, which I conducted along with several colleagues in partnership with Mercy Corps, a global nongovernmental humanitarian aid organization. The project, called PARTNER (Peace Action for Rapid and Transformative Nigerian Early Response), was a USAID-funded initiative launched in 2021 as a five-year effort to reduce political violence in Nigeria’s Middle Belt — specifically Kaduna and Plateau states — where cycles of communal conflict and armed violence have persisted for at least two decades. This region has long been characterized by repeated episodes of localized insecurity that can intensify quickly, overwhelm conventional conflict-management mechanisms, and spill across community and administrative boundaries.

The drivers of conflict in Kaduna and Plateau are diverse, but two mechanisms are especially central. The first is recurrent violence between herders and farmers, often overlapping with ethnic and livelihood identities. These disputes frequently arise from competition over land and water, shifting patterns of access and migration, and the breakdown (or politicization) of local governance arrangements that once helped manage resource-sharing. The second mechanism is predation by organized armed groups — commonly described as “bandits” — engaged in cattle rustling, kidnapping, extortion, and violent attacks on rural communities and transport routes. Such violence both reflects and deepens local insecurity, raising the costs of everyday economic activity and contributing to displacement and communal polarization.

Crucially, these two forms of violence are not independent. Armed groups sometimes emerge from earlier disputes over land and access, recruit within the same social networks implicated in herder–farmer tensions, and exploit grievances generated by prior clashes and state responses. Conversely, banditry can intensify resource competition by displacing communities, disrupting agricultural cycles, and prompting cycles of vigilante mobilization that further erode trust and local dispute-resolution capacity.

PARTNER was organized around a set of linked activities designed to strengthen early warning and early response (EW/ER) capacity at multiple levels of governance and across diverse local constituencies.

At the community level, the program organized community, ward, and district peace officers into structured reporting groups responsible for gathering and transmitting information relevant to emerging or ongoing violent conflict. A central design feature was deliberate inclusion: reporting groups were assembled with a concerted effort to incorporate members of all locally salient groups, both to improve information coverage and to reduce the risk that the system would be perceived as biased toward any single constituency.

To convert locally generated information into actionable insight, PARTNER established a “situation room” responsible for monitoring, aggregating, and analyzing reports from all participating sites.

PARTNER paired these information and analytic capacities with investments in response readiness. The program trained local response stakeholders and relevant security agencies to interpret early-warning reports and to act promptly on them, with the aim of shortening the lag between detection and response. In parallel, PARTNER developed communication and engagement channels that enabled two-way flows between communities and the early-warning architecture. It established a social media presence through which community members could submit reports and receive warnings, complemented by SMS messaging to expand reach and improve accessibility in contexts where internet-based communication is uneven.

Finally, in a subset of communities, PARTNER implemented social cohesion programming and conflict-management training. These activities were intended to complement the EW/ER system by strengthening local capacities for de-escalation and dispute resolution and by fostering communication between ethnic groups that might defuse retaliatory dynamics when conflict occurs.

The conception linking these activities is that timely information, rapid response to violence, and increased understanding between groups can interrupt escalation dynamics, reduce the likelihood of retaliatory cycles, and mitigate the broader political consequences of recurring violence.

Similar peacebuilding interventions have been tried before in Nigeria and elsewhere over the past two decades. What made PARTNER different was the commitment of Mercy Corps and our local partners to scientifically assess the impact of the intervention.

To accomplish this, we designed a randomized controlled trial (RCT) to estimate PARTNER’s effect in communities with a history of interethnic and interreligious conflict. The study covered 46 communities in Plateau and Kaduna and fielded in-person surveys in two waves (2022 and 2024), interviewing about 60 randomly selected respondents per community via a random-walk procedure — roughly 2,700 respondents per wave.

Communities were randomly assigned (stratified by pre-existing violence) to one of three groups: 35 communities received some combination of PARTNER’s early warning/early response (EWER) system and social cohesion training, while 15 received no intervention during the study period. By comparing changes across the treated and untreated groups between 2022 and 2024, the RCT was designed to identify whether the intervention improved outcomes in the groups that received PARTNER’s programming.

The results we found after less than two years of PARTNER were exciting. Across four sets of indicators — warning, response, violence, and social cohesion — outcomes improved in the PARTNER communities relative to the control group. Respondents in PARTNER communities were more aware of the availability of early warning resources and more likely to say that they heard about violence before it actually occurred. People in PARTNER communities were also more likely to say that the police responded quickly to violence, and they were more satisfied with the response of security forces in preventing violence and resolving violent conflicts. People in PARTNER communities were more likely to report that they felt they had the skills to resolve conflicts themselves and were more likely to trust members of other ethnic groups.

Most importantly, communities that received the intervention experienced measurably lower levels of violence than control communities. Communities that received PARTNER programming saw a decrease in fatalities due to violence of over 60 percent between 2022 and 2024. In the control communities, in contrast, fatalities increased by more than 13 percent.

The PARTNER project offers several important lessons about peacebuilding. First, it shows that peacebuilding can work, and it offers a promising framework for other peacebuilding interventions both elsewhere in Nigeria and in other countries plagued by communal violence.

Second, it demonstrates that it is possible to scientifically evaluate the effectiveness of peacebuilding interventions, building evidence that should be strong enough to convince even skeptics that we can make a difference in places where violence has been endemic for decades.

Finally, though, PARTNER’s story also comes with a gloomier lesson. Despite the success we documented in 2024, funding for the project, which was supported by USAID, was abruptly cut off in February 2025, along with funding for hundreds of other peacebuilding, public health, and economic development projects around the world. In the case of PARTNER, this meant that we could not fulfill our promises to communities in the control group to offer them the same programming that effectively reduced violence in the treatment communities. Even in the communities that had already received PARTNER programs, the sudden collapse of funding meant that Mercy Corps was unable to complete the training of local civil society groups who were meant to carry on PARTNER’s work.

This outcome was bitter, to say the least, since much of the criticism leveled at USAID centered on arguments that the money it spent on humanitarian projects was ineffective and wasteful. There was undoubtedly some truth to those critiques. Some foreign aid projects did fail to deliver on their promises. The logical response to that concern, however, is not to end support for all foreign aid, but to increase support to help organizations better demonstrate what works and what doesn’t. With dramatically less money flowing into peacebuilding today, the need to spend that money effectively has never been more important.

Footnotes

1. Institute for Economics and Peace, Global Peace Index 2025: Identifying and Measuring the Factors That Drive Peace (Sydney: Institute for Economics and Peace, June 2025), http://visionofhumanity.org/resources. ↩

2. Abdul Latif Jameel Poverty Action Lab (J-PAL), “Evidence to Effect,” Poverty Action Lab, https://www.povertyactionlab.org/evidence-effect. ↩

3. Abhijit Vinayak Banerjee, Esther Duflo, and Michael Kremer, “The Influence of Randomized Controlled Trials on Development Economics Research and on Development Policy,” in The State of Economics, the State of the World, ed. Kaushik Basu, David Rosenblatt, and Claudia Sepúlveda (Cambridge, MA: MIT Press), 439–88. ↩