Many wars are fought over distinct phases, where some underlying capabilities in one portion may have little or no bearing on others. However, existing models of bargaining while fighting treat battles as identical. We relax that assumption and examine what happens when a source of uncertainty only affects later phases. Under such conditions, learning becomes more difficult. More specifically, we find weaker types do not separate equilibria in the current period if types have identical likelihoods of winning the upcoming battle, even if later battles do depend on type. Thus, we have a natural information-based explanation for prolonged fighting. A case study of the Shining Path in Peru illuminates the mechanism.

Logical fallacies are often embedded in misinformation to mislead the public and manipulate opinion. How can we help individuals resist such tactics? Existing research offers a variety of interventions to counter misinformation, yet the asymmetry between misinformation production and correction, the underrepresentation of authoritarian and non-English speaking contexts, and motivated reasoning continue to limit their efficacy and external validity. We introduce a refined intervention: a series of short videos that train individuals to recognize and resist three logical fallacies commonly found in misinformation—false causality, hasty generalization, and personal attack. We conducted a preregistered survey experiment (N = 2,416) in China in August 2025 to evaluate the efficacy of the intervention. Results show that participants exposed to the logical fallacy training were less likely to endorse related misinformation overall, with particularly strong effects among more vulnerable groups, including nationalist participants and those with lower cognitive reflection ability. Our study contributes to the literature on misinformation correction by extending external validity to an understudied authoritarian context and by designing an intervention that bypasses potential censorship through non-political, reasoning-based, fable-like narratives. In addition, using short-form videos substantially lowers the costs of replicating, implementing, and receiving the intervention for all parties involved in misinformation correction. Finally, our intervention strengthens cognitive immunity among those most susceptible to misinformation.

International donors frequently rely on foreign aid to support refugee-hosting countries, yet the security implications of such aid remain contested. While some studies argue that aid projects targeting refugees can be exploited by non-state armed groups to fuel violence, recent research highlights the potential of aid to foster peace by improving the well-being of both refugees and local communities. I argue that whether foreign aid exacerbates or dampens conflict depends on the nature of policies governing refugees rights. Host state policies that allow refugees to work and move freely create positive externalities for local community, which can alleviate tensions between locals and refugees, thereby reducing individual grievances that increase the risk of violence. In contrast, restrictive refugee policies limit these spillovers, increasing the risk of refugee militarization and conflict. Analyzing the impact of World Bank aid in 1753 refugee-hosting provinces across 107 countries in the Global South, I find that aid allocation in countries with inclusive refugee policies is associated with a significant decrease in conflict in refugee-hosting provinces. The results from the subnational analyses hold using multi-level models and instrumental variable analyses, which help alleviate inferential problems associated with observational data. The findings indicate that international donors should consider the domestic refugee policies of aid recipients to enhance the effect of aid for both improving the welfare of refugees and locals and dampening the risk of violence in refugee hosting communities.

Refugee inflows are often linked to increases in reported violence in host communities. We argue that this association may be partly attributable to changes in observability. Refugee inflows draw humanitarian actors, security forces, and media attention into peripheral areas, increasing the likelihood that incidents are detected even when underlying conflict dynamics shift little. This reporting-induced bias should be more pronounced for low-scale non-state violence than state-based violence and in areas with large refugee settlements. We evaluate these claims by combining global subnational analyses with micro-level evidence from Afghanistan and using novel reporting-quality proxies. Areas hosting refugees appear to experience more communal violence and protests, but much of this pattern coincides with increased media presence and higher reporting quality. While we cannot fully recover true violence levels or adjust for this bias, our findings underscore how ignoring detection dynamics can yield misleading inferences about the effects of refugee hosting on violence.

Although refugee return is widely promoted by host states and international organizations as a durable solution to displacement, its implications for security and postwar order remain poorly understood. In particular, little is known about how return reshapes communal relations between returnees and stayees in conflict-affected areas. This paper advances a theory of post-return violence that emphasizes the role of displacement experiences. Contrary to the common assumption that exposure to violence deepens trauma, fosters resentment, and fuels new cycles of conflict, I argue that such exposure can instead facilitate reconciliation. When both returnees and stayees have endured wartime violence, shared traumatic experiences foster empathy, promote ingroup re-categorization, and reduce tensions, thereby lowering the risk of communal violence. I test this argument with granular settlement-level data from the International Organization for Migration’s Integrated Location Assessments (2019–2022) in post-ISIS Iraq, using a shift-share instrumental variable design to address the endogeneity of refugee returns. The results provide causal evidence that shared exposure to violence decreases communal conflict in returnee settlements, and suggest that this effect is even stronger in more fractionalized communities. These findings speak to broader debates on displacement, reconciliation, and postwar peacebuilding. They show that returned refugees need not destabilize communities, but can peacefully coexist with the stayees when underpinned by shared suffering. For policymakers, the results highlight the importance of coupling material assistance with reintegration initiatives that acknowledge collective experiences of violence and strengthen community institutions to sustain peace.

The migration patterns of civilians are a crucial determinant of battles on the frontline and conflict spillover to the backline. Civilians in conflict zones serve as a recruitment pool for rebels, while relocation and internal migration to the backline can export instability to previously peaceful regions. Governments therefore face a tradeoff between weakening rebels in contested areas and preventing conflict from spreading, and rebels must balance conscription and governance to ensure that enough civilians stay. Current research provides insights on civil war actors' regulation of migration through violence, yet nonviolent tools such as wartime governance, displacement aid, and protection remain understudied. This paper develops a formal model with a government, a rebel group, and a continuous distribution of civilians to examine when wartime actors provide aid and governance to civilians in contested areas, and how effective migration-management aid shapes wartime internal migration. The equilibrium shows that when civilians’ cost of war is high, governments and rebels may collude in migration control by increasing aid on the frontline. Moreover, when the backline is highly unstable, the government has an incentive to aid rebel supporters to offset the contagion of conflict, even if this strengthens the rebel. Using the International Organization for Migration’s camp assessment panel surveys (Rounds 24–32) in Ethiopia during the Tigray War (2020–2023), I provide empirical evidence consistent with the model’s predictions.