Highlights

The identification and analysis of fragility factors and conflict drivers, relevant for achieving development effectiveness, have improved over the evaluation period. This is due to International Development Association fragility, conflict, and violence policy commitments; the development of the fragility, conflict, and violence strategy; country management commitments; and the elevation of the Risk and Resilience Assessments to a core diagnostic to inform lending. Compared with those of the first half of the evaluation period (2010–15), more recent conflict analyses are twice as likely to identify relevant factors of fragility and to articulate specifically how these factors influence conflict and violence.

However, the client-facing nature and the potentially broad distribution of conflict analyses in the World Bank have sometimes prevented frank assessments of fragility and conflict drivers, limiting the transmission of conflict considerations into portfolio and operational decisions. Although these issues may be understood by World Bank country managers, the limited availability to task teams of information on the political contributors to conflict undermines efforts to tailor operations to conflict drivers. Also, the quality of the diagnostic, or hard-to-operationalize or missing recommendations in conflict analyses, have sometimes limited the transmission of conflict considerations into strategy and operations.

Conflict-informed sector advisory services and analytics prepared in the wake of conflict events have helped navigate conflict-related and political economy issues and inform future engagement. These have involved analysis to understand local dynamics to inform World Bank responses.

However, few sector advisory services and analytics conducted before major warring activities discussed conflict or political economy–related factors. Virtually all sectoral advisory services and analytics conducted before conflict—and easily accessible by staff (many political economy analyses remain confidential)—were not conflict sensitive.

Country teams are increasingly innovating with real-time conflict risk identification and monitoring. Critical to these efforts is the use of local knowledge gleaned from social media, newspapers, and word of mouth, and the ability to interpret these events in relation to real-time decisions to adapt the World Bank’s country engagement.

The identification and analysis of fragility and conflict drivers relevant for achieving development effectiveness have improved over the evaluation period. The World Bank’s primary tool for identifying and analyzing key drivers and risks of fragility and conflict is the RRA (box 2.1). There are also country social and economic impact assessments, political economy analyses, and fragility assessments (the predecessors to RRAs), which assess factors of fragility and conflict drivers.1 Compared with those of the first half of the evaluation period (2010–15), more recent conflict risk analyses are twice as likely to identify and articulate specific factors of fragility and associated drivers of conflict relevant for achieving development effectiveness and addressing risks. Conflict analyses conducted during the first half of the evaluation period tended to include a description of many factors of fragility but were imprecise in articulating how these factors influence what is more specifically driving conflict and violence. Examples of this earlier trend include assessments in South Sudan (before 2015), Madagascar (2014), Mali (2015), Cameroon (2016), and the Democratic Republic of Congo (2016). For example, the Democratic Republic of Congo assessment does not distinguish between causes, drivers, symptoms, consequences, and indicators of fragility; it further offers little guidance on how to focus conflict-reduction efforts at the local level or how to differentiate engagement in conflict and non-conflict-affected areas. Good examples of robust analysis from the second half of the evaluation include the Niger RRA (2016; see box 2.2), the regional Sahel RRA (2020), and the Iraq Country Social Assessment (2020). For example, the regional Sahel RRA has been instrumental in building a common understanding of conflict drivers in the Sahel; identifying areas at risk of violence; analyzing the dynamics at play; and proposing priorities, integrated or coordinated actions, and prevention strategies across the countries and donors of the Sahel Alliance. Also, the Iraq conflict sensitivity analysis includes a thorough analysis of conflict considerations per sector and provides clear guidance on prioritizing and sequencing.

Reasons for this improvement include several mutually reinforcing factors introduced through IDA18 and IDA19 policy commitments; the development of the FCV strategy; the elevation of the RRA to a core diagnostic that must inform country engagement; and the increase of Country Management Unit (CMU) engagement in the RRA production. First, RRAs were made a mandatory part of the country engagement cycle for IDA FCS during IDA18, at which point the World Bank made a commitment to ensure that CPFs in these countries were informed by an analysis of fragility and conflict risks. Second, in parallel, CMUs began to finance RRAs out of their own budget, which strengthened CMU ownership, engagement, and use of the diagnostic tool. Third, the RRA was made a core diagnostic, mandated to be conducted before lending programs are begun in countries with FCV. Additionally, it is expected that the new set of RRA guidelines, developed with the input of staff with extensive FCV experience, will further improve the quality of RRAs.2

The upgrading of the RRA to a core diagnostic—to be conducted before the development of lending programs—has addressed a critical sequencing challenge. Of the 23 countries covered by this evaluation, 18 had both an RRA (or, if earlier, a fragility assessment) and a Country Strategy or Engagement Note. Yet, the RRA (or fragility assessment) was made available in a timely way in only 9 cases, during the latter part of the evaluation period. Because RRAs are currently being updated (or new ones are being written)—for example, those for Cameroon, Central Asia, and Chad—they are increasingly sequenced just in time in relation to CPFs.

IDA19 policy commitments have also helped the World Bank effectively transcend its state-centric model and develop cross-border analyses and targeted regional operations. Spurred by the deteriorating situation in the Sahel, IDA19 deputies and clients requested that the World Bank better consider the cross-border nature of fragility and conflict and reflect this in its analytics and the design of its operations. As a result, the IDA19 policy commitments regarding developing at least three regional programs are helping to change the World Bank’s approach: it has finalized a Sahel RRA—in collaboration with the UN and the Sahel Alliance—that is informing both World Bank and UN cross-border operations, with a focus on preventing violent extremism, youth engagement, land rights, and management of transhumance linked with the growing risks brought about by climate change. The broader analysis allowed for a consideration of risks that were left unaddressed in the past (box 2.2). Three other regional analyses are also under way (a Central Asia regional RRA, being conducted jointly with the United Kingdom and UN, that focuses on the specific stresses and history of the Ferghana Valley and border issues with Afghanistan; another for the Lake Chad region; and a third for the Horn of Africa).

However, the client-facing nature and the potentially broad distribution of conflict analyses in the World Bank have sometimes prevented frank assessments of fragility and conflict drivers, limiting the transmission of conflict considerations into portfolio and operational decisions. The FCV strategy indicates that the Bank Group’s approach and interventions should be tailored to FCV drivers, but the analysis and redress of these drivers are more difficult when the root causes of conflict are more political in nature (that is, geopolitics, elite capture, corruption, and pervasive governance challenges). In several instances, RRAs have been found to omit geopolitical factors relevant for development decision-making (for example, in Iraq, Sudan, and the Republic of Yemen); to have only partial coverage of drivers or limited discussion on the political nature of conflict (for example, in Burundi, Cameroon, and Iraq); or to omit regional risks resulting from geopolitical shifts or border issues (which are now being addressed through regional RRAs, for example in Central Asia, the Horn of Africa, Lake Chad, and the Sahel before 2020; see also box 2.2). Although these issues may be understood by World Bank country managers, the limited availability to task teams of information on the political contributors to conflict undermines efforts to tailor operations to conflict drivers.3

In Iraq, the overtly political nature of the conflict and the client’s sensitivity in relation to discussion of conflict drivers meant that, even though the RRA effectively identified and analyzed drivers of conflict, the World Bank’s subsequent Systematic Country Diagnostic (SCD) and country strategy did not take up this analysis. The 2016 RRA effectively identified and analyzed drivers of conflict.4 However, this analysis was not well leveraged in the subsequent SCD, which portrayed conflict as an explanatory variable for current challenges rather than as an obstacle to current and planned development. Although conflict-specific projects (for instance, the Iraq Social Fund for Development project) were naturally conflict sensitive, a conflict lens was not present in non–Social Development projects. Many projects approved after the SCD did not, for instance, analyze their anticipated impact on existing social inequalities or cleavages, including how cash-for-work and reconstruction activities would differentially affect Iraq’s social, ethnic, or sectarian cleavages (key conflict drivers identified in the RRA). To be sure, task teams interviewed noted the pressure to omit an emphasis on conflict to satisfy both the client (which, buoyed by high oil prices, had declined earlier support a few years earlier) and the World Bank Board of Executive Directors.

In Cameroon, the analysis and treatment of conflict drivers has been partial due to political sensitivities regarding the anglophone crisis. There, the RRA and SCD adequately covered risks associated with the country’s highest profile security threat, Boko Haram, but not its secondary conflict emerging in anglophone areas of the country. Consequently, although the CPF points to specific ways to adapt the country program in response to the risks posed by Boko Haram in the north, it does little to address the issues associated with the burgeoning anglophone conflict. World Bank staff were hard pressed to engage on the topic—many reported having insights but not feeling comfortable talking openly about the growing anglophone conflict—because the World Bank’s client (the government) was reticent to acknowledge the conflict. The situation was particularly difficult and volatile, since, in this case, the root causes of conflict were political; further complicating the matter were the attitudes and behaviors of some external players who publicly shamed the government.

The quality of the conflict analysis, or hard-to-operationalize or missing recommendations, have sometimes limited the transmission of conflict considerations into strategy and operational decisions. Almost half of all RRAs (including both older and newer analytics) did not provide operationally relevant recommendations, and therefore most of the associated CPFs lacked clear articulation of policy actions to address conflict drivers identified by the RRA. For example, the Sudan RRA (no date) recommendations called for rapid advisory services and analytics (ASA) but bypassed quality and relevance issues that plagued World Bank ASA efforts at the time, thus not reflecting the sequence of actions required to get out of the cycle of economic deterioration. Consequently, the 2020 Country Engagement Note does not reflect Sudan’s fragility and does not provide sufficient clarity on concrete sequencing.5 Another example is the Iraq RRA, whose recommendations were vague or lay outside the purview of the World Bank (for example, “adopting a granular and nimble development approach,” or “achieving political settlements between warring actors”). The Guinea-Bissau fragility assessment’s recommendations were too high level: it recommended increasing staffing, applying for the IDA TAR, and accelerating the country program’s delivery but included only one midlevel actionable recommendation. Consequently, the CPF was informed by a subset of conflict recommendations from the SCD instead. The resulting CPF omitted the most important fragility drivers and policy actions needed to address them, primarily security sector reform (which may or may not be in the Bank Group’s comparative advantage to address, but it should be acknowledged). In other cases, such as Chad, the RRA was backward looking, lacked prioritization, and had no recommendations, apart from a statement that the World Bank should act to break Chad’s cycle of fragility. This led to a CPF that addressed the cost of conflict rather than the causes of conflict.

In other analytics, such as Damage and Needs Assessments (DNAs) conducted in postconflict situations, the omission of a conflict lens has prevented the World Bank from ensuring that the implementation of assessment recommendations does not create new grievances or exacerbate those that already exist. Although DNAs focus on core reconstruction and recovery needs and, as such, are not intended to serve as catch-all postconflict recovery assessment tools, they do still need to consider shifting conflict dynamics or how the World Bank should navigate them in conflict-affected countries.6 Several DNAs have been conducted in conflict-affected countries, including Lebanon, the Syrian Arab Republic, and the Republic of Yemen, but many have had mixed success in identifying and analyzing conflict dynamics. Currently, conflict sensitivity is not integrated into DNAs or the priorities that the recommendations establish; rather, the World Bank relies on other analytical tools to do this. However, this omission of a conflict lens from what is often the main tool used to guide reconstruction prevents the World Bank from ensuring that the implementation of assessment recommendations does not create new grievances or exacerbate those that already exist. For example, the Republic of Yemen Dynamic DNAs (I–III) did not analyze conflict drivers, perhaps because the internationally recognized government of the Republic of Yemen—an actor in the country’s conflict—was a partner to the assessments. These analyses focus on mitigating the consequences of active conflict, including the effects on different conflict-affected populations, but do not do enough to ensure that their implementation does not exacerbate tensions. To be sure, it appears that recent DNAs may be becoming more integrated with other conflict and postconflict assessment tools, leading to greater conflict sensitivity. For example, in Lebanon the World Bank is currently using the Beirut Rapid DNA to reconsider its priorities using a whole-of-government approach that brings together government, civil society, the private sector, activist groups, youth groups, think tanks, and academia around a shared vision; the same is true for the most recent Republic of Yemen DNA.

Although DNAs conducted in conflict-affected countries seek to deliver much-needed “quick wins,” this can come at the expense of correctly identifying the conflict drivers and structural issues that need to be addressed to prevent future conflict. In a conflict-affected context, “quick wins” could be perceived as resources allocated to one side in a conflict that reinforce divisions in society. The Iraq DNA, for example, focused on the reconstruction of cities destroyed by the Islamic State, an area also prioritized by other actors, leading to a perception that certain social groups were benefiting at the expense of others. Furthermore, the problem of basing a reconstruction plan on an assessment of what has been destroyed is problematic in war-torn countries like the Syrian Arab Republic or the Republic of Yemen, where long-lasting conflict has changed the situation to the extent that an in-depth, conflict-sensitive assessment is needed to avoid prioritizing investments that could exacerbate conflict, including by focusing on social, institutional, and governance impacts, in addition to reconstruction needs (box 2.3).

Leveraging Knowledge to Inform Engagement and Management Decisions at the Country Level

Conflict-informed ASAs conducted during and after the onset or escalation of conflict events have helped navigate conflict-related and political economy issues and inform future engagement. This has generally involved analysis to understand local dynamics to inform the World Bank’s response. For instance, after Madagascar’s 2009 coup and during its political transition, the World Bank suspended and significantly cut back its lending but ramped up its nonlending activities. It engaged in extensive analysis to supplement its understanding of technical issues across several sectors with conflict- and political economy–related factors, in light of heightened political instability, the country’s main conflict driver. The results were a series of high-level policy notes that became a critical reference for other development partners and nonstate actors to understand conflict dynamics from a sectoral level and to support future lending. Much the same was true in Burundi (2014–15), Iraq (2014–16), Myanmar (2017), and South Sudan (2013–present).7

However, almost none of the sector ASAs conducted before major warring activities in conflict-affected countries—and made widely accessible—included conflict and political economy–related factors to supplement an understanding of technical issues. Although predicting conflict is hard, conflict risk drivers do not come out of nowhere. Yet virtually all sectoral ASAs that were conducted before the onset of conflict—and that were available in the Operations Portal (many political economy analyses remain confidential)—were generally not conflict sensitive. These analyses generally recommended “best practice” technical solutions rather than “best fit” alternatives informed by conflict analysis, adjusted to local institutions and operating conditions (see also Woolcock 2014; World Bank 2011). A notable exception was in the Republic of Yemen, where conflict sensitivity analyses and political economy analyses predated the 2012 crisis. Another was Somalia, where sector ASA analyzed drivers and prioritized efforts to address conflict, making the ASA a strong informant of, and complement to, project financing. More recently, with the incentive of the TAR (now the TAA), Madagascar (post-2016) and the Central African Republic (post-2018) modified their portfolio to include conflict sensitivity in most ASAs, including multisector programmatic ones to ensure portfoliowide strategy and policy actions.

Country management is increasingly innovating with real-time conflict risk monitoring as a form of adaptive management. Although these monitoring efforts have been piloted in reaction to major conflict-related events that posed significant risks to the World Bank and country portfolios, many have been sustained as a portfolio monitoring tool. Although it is still too early to assess the effectiveness and impact of these pilots, it appears that the use of local knowledge gleaned from social media, newspapers, and word of mouth is critical to these efforts, as is the ability to interpret these events in relation to the World Bank’s country engagement. For example, in Myanmar, after the Rohingya crisis (2015), the CMU made project approval contingent on a task team’s application of a “peace and inclusion lens”—which was used to inform several conflict-sensitive project adaptations—and partnered with a local think tank, which thereafter has acted as a conflict observatory. In Lebanon, after the Beirut explosion in 2020, a new FCV risk monitoring framework was prepared, offering more frequent and proactive assessments to strengthen preparedness, prevention, and decision-making at various levels. In Somalia, country management routinely seeks local political economy insights on the affiliation and reputation of local stakeholders in terms of their perceived legitimacy and relationship to government. Having observers in the field has proven critical for designing urban interventions that can be successfully implemented while considering different clan interests.