Highlights

The fragility, conflict, and violence strategy aimed to step up partnerships with a more diverse set of humanitarian, peace building, security, and private sector actors at both the strategic and operational levels. At the operational level, third-party implementation (TPI) has become a key operational modality in fragile and conflict-affected situations (FCS) and has been applied with increasing frequency in recent years as it enables continued engagement in countries with active crises and conflicts or where direct World Bank implementation of projects with the government is not possible.

TPI projects tend to be more output oriented than World Bank–implemented FCS projects—one aspect that led to the perception of blurred lines between the World Bank’s development engagement and humanitarian support. To prevent this perception, TPI-based engagements could define clear short-term time limits and define phase-out strategies at project approval, as well as highlighting their long-term development impact.

The majority of the World Bank’s TPI projects included development components—such as long-term strengthening of institutions and capacity components—yet not systematically so. The long-term development impact of TPI could therefore be enhanced by systematizing support (and indicators) for institutions, capacity building, and human capital.

The portfolio supported by TPI has performed well compared with other World Bank–implemented projects in FCS, despite the higher risks.

There is no established practice or guidance for the selection, contracting, and cost parameters of TPI arrangements and their exit strategies. The World Bank needs to improve its practices and guidance to enhance the consistency, transparency, and value for money of TPI arrangements.

The FCV strategy aimed to step up partnerships with a more diverse set of humanitarian, peace building, security, and private sector actors. The goal was to maximize impact based on the complementarities and comparative advantages of different institutions. The strategy aimed to improve the leveraging of partnerships and extend it to nontraditional partnerships to enhance operational impact and ensure effective implementation in FCS settings. The strategy also emphasized collaboration with multilateral development banks for joint diagnostics and training and with regional organizations to address cross-border challenges. Engagement with civil society organizations would be scaled up, particularly those that promote the empowerment of women and girls. Additionally, the strategy sought to develop the private sector in FCV settings and foster partnerships with international organizations such as the International Monetary Fund, the Organisation for Economic Co-operation and Development, and the European Union, as well as bilateral agencies, to ensure complementarity in operations.

The Bank Group engaged with partners at both the strategic and operational levels. At the global strategic level, the Bank Group sought to engage with partners on coordinated approaches to engage on common objectives, including through convening, partnership frameworks such as between the UN and the Bank Group (2017), joint analytical and data work, and strategic collaboration at the country level (for example, on policy advice). At the operational and transactional levels, the Bank Group has partnered through pooled financing (such as multidonor trust funds); common approaches and design in the context of the IDA FCV Envelope, especially the PRA; and joint implementation through TPI, a key engagement modality during the strategy period discussed in detail in the following section.

Partnerships have supported the World Bank in effectively navigating and responding to diverse challenges across FCS. Across the country cases reviewed for this evaluation, partnerships were essential for the Bank Group to operate effectively in FCS, with their form and intensity shaped significantly by the severity of the conflict and institutional fragility.

In high-intensity conflict contexts, TPI through partnerships with UN agencies and NGOs was the primary operational modality. These partnerships effectively leveraged the UN’s extensive field presence and humanitarian access alongside the Bank Group’s financial and technical oversight, enabling essential service delivery and maintaining the World Bank’s engagement in environments where direct government collaboration was not feasible. In the Republic of Yemen, collaboration with established national institutions supported by UN agencies allowed the Bank Group to preserve critical local service delivery capacities. However, a common challenge across these contexts was that such emergency-oriented partnerships often bypassed national institutions, limiting longer-term capacity building and weakening the potential to restore and strengthen the social contract.

In fragile environments such as Lebanon and Mali, partnerships allowed for more strategic engagement. As IEG’s case studies indicate, the Bank Group played critical roles as a convener and intermediary through initiatives such as Lebanon’s Reform, Recovery, and Reconstruction Framework, co-led with the UN and the European Union, and Mali’s Sahel Alliance, which fostered structured policy dialogue and cooperation amid political instability. While these platforms provided continuity of operations and enabled targeted interventions, they also highlighted persistent constraints related to elite capture, inconsistent government engagement, and limited political will for deep structural reforms.

The FCV strategy has significantly enhanced the Bank Group’s credibility and commitment in addressing FCV, notably building on the foundational Pathways for Peace initiative. Over the past 15 years, the Bank Group has deepened its engagement with international NGOs, international organizations, and multilateral development banks, reflecting a notable shift toward collaborative and inclusive approaches in addressing FCV issues. This shift has included strengthened policy dialogues, joint strategic planning, and shared analytical work aimed at identifying and mitigating the root causes of fragility and conflict. Enhanced coordination and structured collaboration with international partners have allowed for more nuanced, tailored interventions that are sensitive to local contexts and dynamics.

A key outcome of this collaborative approach has been the exponential growth of partnerships with the UN, which is particularly evident in TPI arrangements and interinstitutional engagement as part of the IDA PRA and TAA, as evidenced in this evaluation. These arrangements have significantly enhanced the reach and effectiveness of World Bank projects among marginalized and otherwise inaccessible populations, enabling crucial service delivery during active crises and conflicts.

Engagement with civil society organizations has also intensified and become more structured, with systematic and regular interactions both within and beyond FCS. These interactions include ongoing consultations; joint monitoring of project outcomes; capacity-building initiatives; and coordinated advocacy efforts aimed at policy reform, governance improvements, and accountability mechanisms. Civil society organizations play an increasingly crucial role in ensuring that interventions are grounded in local realities and responsive to community needs.

Strategic collaboration has proven most effective when it transcends transactional relationships, placing a strong emphasis on analytical work, robust data generation, comprehensive knowledge synthesis, and shared learning platforms. Such collaborative efforts enable partners to collectively navigate complex operational environments and respond to evolving contexts. A notable example of such effective strategic collaboration is the Bank Group’s partnership with UNHCR, which covers multiple thematic areas (for example, private sector, joint data, building evidence) and has successfully combined operational capacities and technical expertise to deliver improved outcomes for vulnerable populations, particularly in displacement contexts. However, significant and persistent challenges remain, such as uneven funding predictability, entrenched institutional weaknesses that limit implementation capacity, and insufficient mutual understanding and alignment among partnering agencies. Collaboration at the country level relies heavily on individual initiative and personal commitment, rather than being consistently supported by robust institutional frameworks and mechanisms. This underscores considerable potential for the World Bank to further institutionalize and strengthen strategic cooperation, particularly in complex areas such as conflict-sensitive interventions, comprehensive approaches to peace building, resilience, and sustainability of outcomes in FCV contexts.

Third-Party Implementation and Its Operational Relevance During the FCV Strategy Period

A key area of partnerships during the strategy period is TPI. This evaluation focused on analysis of partnering in project implementation, a core modality scaled up in the context of the FCV strategy’s priority of remaining engaged. TPI involves the use of external agencies—such as the UN agencies, NGOs, and private sector entities—to deliver development assistance on behalf of the Bank Group. TPI enables the World Bank to maintain engagement where direct implementation is constrained by security risks, institutional limitations, or political sensitivities (including, for instance, in countries under OP 7.30).

TPI has become essential for maintaining World Bank engagement in countries experiencing active crises and conflicts. Between FY15 and FY24, TPI was used in 228 out of 1,169 FCS projects, accounting for 20 percent of the FCS projects and representing $10.3 billion (or 10 percent) of total commitments. TPI is particularly concentrated in countries such as the Federal Republic of Somalia, South Sudan, and the Republic of Yemen, where it plays a crucial role in delivering services amid complex political and security challenges. The Republic of Yemen alone accounted for $3.7 billion of direct TPI, highlighting TPI’s strategic importance in maintaining engagement in conflict-affected areas.

TPI in FCS has relied on borrower contracting (or indirect TPI) and direct financing (or direct TPI). The distinction between these approaches is the entity responsible for procurement and contract management. Borrower contracting, or indirect TPI, places the responsibility on the borrower (typically the government or a designated agency), which selects and manages third-party providers while adhering to World Bank procurement guidelines, and is not limited to FCS. Direct TPI, or direct financing, by contrast, enables the World Bank to engage in procurement by selecting and managing a third-party implementer directly, independent of the government. This approach is used in exceptional circumstances, such as when there is no functioning government, the government lacks control over project areas, technical capacity is severely constrained, or World Bank staff cannot access sites due to insecurity. In such cases, the World Bank may finance TPI through its own resources or trust funds. According to the FCV strategy, direct TPI should aim to support system rebuilding, institutional strengthening, capacity development, and long-term sustainability.

The use of TPI has grown significantly over the past decade, and it has become a key operational modality in FCS. The use of TPI has accelerated markedly since FY17, driven by escalating conflicts in countries such as Afghanistan, South Sudan, and the Republic of Yemen, alongside global crises such as the COVID-19 pandemic. During this period, both direct and indirect TPI modalities expanded. Direct TPI, once limited to smaller projects, increased sharply, while indirect TPI also rose from an already broad base. Direct TPI accounted for $5.9 billion, more than half of all TPI, and the largest recipients were the Republic of Yemen, Afghanistan, and South Sudan.

TPI use peaked during the FCV strategy period (in FY22) and has since declined, yet its growth over time reflects the World Bank’s increasing reliance on third-party actors. This reliance happens in particular in contexts where government systems are weak or access is severely restricted. In 82 percent of FCS (27 out of 33), the share of the portfolio disbursed through TPI increased, indicating the growing institutional dependence on this modality (figure 6.1).1

Figure 6.1. Evolution of Third-Party Implementation in FCS by Contract Amount

Image

Share Tweet Facebook WhatsApp Linkedin Email

Figure 6.1. Evolution of Third-Party Implementation in FCS by Contract Amount

Source: Independent Evaluation Group based on World Bank Operations Policy and Country Services data.

The World Bank’s ambition to remain engaged during active conflict has contributed significantly to the expansion of TPI. The use of TPI is largely driven by FCV contexts, which account for 76 percent of TPI spending. The goal of remaining engaged in certain countries that experience active conflict, such as the Federal Republic of Somalia and South Sudan, has driven a large share of that spending. Direct TPI is also crucial to remaining engaged in countries under OP 7.30, as has been demonstrated in Afghanistan, Myanmar, and the Republic of Yemen, which together account for 41 percent of FY15–24 overall TPI spending. The rise in conflict has also contributed to the use of TPI, as countries experiencing conflict rely significantly more on TPI than do institutionally fragile countries. Over the same period, Afghanistan, South Sudan, and the Republic of Yemen have received almost half of TPI spending, with the Republic of Yemen—where almost all World Bank financing is provided through TPI—accounting for 27 percent. The top 10 recipient countries account for 70 percent of TPI.

Direct TPI financing is mostly spent on health and social protection, delivering essential services such as health interventions or cash transfers in inaccessible or volatile areas where government agencies no longer provide such services.2 Health and social protection–related contracts account for two-thirds of direct TPI spending. Direct TPI has been used extensively in cash and social transfer programs, particularly in displacement-affected settings. Almost one-third of TPI direct financing contracts relate to health interventions such as emergency response, vaccines, medical supplies, health infrastructure, and technical assistance. Another 19 percent has gone toward social protection, including cash transfers, public works, equipment, and services. The large expansion of TPI during the FY20–24 period was mainly driven by large COVID-19 pandemic operations focused on the procurement of medical and personal protective equipment, public health initiatives, and technical assistance for pandemic response. The majority of both health and social protection TPI spending was in the Republic of Yemen, which accounts for 55 percent of direct TPI financing.

For indirect TPI (borrower contracting), a large share goes toward external technical advice. Borrower contracting accounts for 38 percent of overall FCS TPI spending. This spending can be divided into four procurement categories, with consulting services being the largest, accounting for 41 percent of borrower contracting. Consulting services include any external technical advice, usually presented in the form of a report, and spending in this category has been especially high in the Federal Republic of Somalia, which accounted for 34 percent of all spending on consulting services in the period. The other procurement categories of goods and nonconsulting services both account for slightly less than 30 percent of spending. Goods include all tangible items and physical assets and are governed by the corporate procurement directive. Spending on goods has been particularly high in the Democratic Republic of Congo (29 percent) and Mozambique (21 percent). Nonconsulting services are services that are not primarily intellectual or advisory in nature. These services include maintenance, operation of facilities or plants, logistics, and other support activities. Spending on nonconsulting services is largest in Lebanon, which accounts for 36 percent of such spending. The final procurement category, works, refers to construction-related activities and accounts for only 9 percent of indirect TPI spending.

While the majority of TPI projects include outcome indicators and development components, TPI projects tend to be more output oriented than other FCS projects. In an FCS, especially during active conflicts or crises, there is a continuum of development and humanitarian or emergency interventions assistance. In these volatile or conflict-affected environments, assistance for service delivery is developmental—with many seeking to protect human capital or broader development gains—given the lack of government structures or ability to provide essential services during conflict. An analysis of project indicators during the period identified that 4.6 percent of all FCS projects can be characterized as predominantly output focused. For FCS projects that included TPI contracts, this share was twice as high (10.7 percent). This difference indicates that while a larger share of TPI projects are output focused (capturing more humanitarian activities), the majority (73.2 percent) include at least some outcome-focused indicators and development components. In South Sudan and the Republic of Yemen, TPI projects emphasized institution building through local NGO partnerships or semiautonomous parastatal institutions and demonstrated a path toward sustaining developmental objectives even in the absence of formal government engagement. However, such arrangements were not systematic across the TPI project portfolio.

More than half of TPI projects’ commitments is spent on short-term-oriented interventions. About 40 percent of the TPI project components are short-term oriented, accounting for more than 50 percent of the total commitment amount (table 6.1). Most of these short-term components support service delivery in emergency situations, such as emergency short-term cash transfers, health services, nutrition, or water supply. The provision of social safety nets, including long-term social protection, also accounts for a significant share (17.1 percent) of the total commitment amount. Additionally, TPI project components related to the COVID-19 pandemic response, including vaccine procurement and delivery, make up a substantial share of total commitments. In contrast, long-term-oriented components represent a larger share of the total number of components (44.4 percent) but receive a smaller share of commitments (32.9 percent), indicating a relatively stronger focus of TPI project components on emergency relief over structural transformation.

Source: Independent Evaluation Group.

Note: This analysis is based on the 220 TPI projects with complete component data, out of 228 TPI projects. FCS = fragile and conflict-affected situations; TPI = third-party implementation.

The larger output focus and short-term nature of TPI projects is driven by the fact that TPI is often used in emergency projects. Half of TPI FCS projects are emergency projects, making up more than twice the share of non-TPI FCS projects. In emergency contexts, output-focused short-term interventions are often more relevant to address immediate needs to save lives and preserve human capital and institutions. Additionally, in low-capacity FCS, interventions resembling humanitarian aid can serve as critical building blocks for longer-term development.

Effectiveness of Third-Party Implementation

The portfolio supported by TPI has performed well compared with regular World Bank–implemented projects in FCS, despite perceptions of higher risks associated with the use of TPI. Evaluated projects involving TPI were rated high in achieving their development outcomes (89 percent rated moderately satisfactory or above), compared with 83 percent of evaluated FCS projects that did not involve TPI (figure 6.2).

The efficiency of TPI arrangements is unclear. Information on the cost of TPI arrangements is not available systematically, so this evaluation could not assess efficiency. Several CMUs noted that the negotiation of overhead costs and service packages with UN agencies is inconsistent because there are no centralized benchmarks or OPCS support to CMUs or task teams. Staff indicated that UN overhead costs vary widely and can be as high as 40 percent. While some variation is justified given that TPI partners assume substantial security risks in many cases to implement projects in environments where the World Bank cannot operate, limited consistency in selection and a lack of information and transparency in certain contexts can contribute to higher costs. Attempts to compete with multiple viable TPI options, even from within the UN system, have resulted in substantially reduced costs. However, without systematic guidance or comparative costing data, such negotiations rely heavily on individual team capacity.

TPI costs need to be benchmarked against the cost of direct World Bank implementation. The average share of project management or implementation costs of FCS projects was 11 percent of total lending commitments.

Figure 6.2. Bank Performance and Development Outcome Ratings

Image

Share Tweet Facebook WhatsApp Linkedin Email

Figure 6.2. Bank Performance and Development Outcome Ratings

Source: Independent Evaluation Group project evaluation data.

Note: TPI = third-party implementation.

Challenges Related to Third-Party Implementation

The concentration of partners for TPI raises concerns about accountability and sustainability. The vast majority of TPI contracts are with UN agencies, accounting for more than 95 percent of TPI funding.3 Among those agencies, the UN Children’s Fund receives 35 percent of funding, the UN Office for Project Services 22 percent, and the World Food Programme 16 percent. The share of international NGOs receiving TPI funding is small. This concentration among UN agencies marginalizes NGOs and local actors. The FCV strategy MTR raised concerns about accountability, sustainability, and the potential for moral hazard, as some implementing partners become financially dependent on Bank Group contracts. Such concerns are shared by Bank Group staff and were raised in an FCV Group internal review of TPI arrangements.

Operational challenges surfaced in several conflict situations where TPI became the preferred operational approach or where the institutional risk stance did not permit government-led project implementation. When a UN agency, through TPI, takes on the responsibility for fiduciary, compliance, and security risks, it enables the Bank Group to operate in regions that would otherwise be beyond its tolerance for risks to its reputation, staff safety, and operational safety. Although this TPI facilitated operational continuity in the absence of explicit risk tolerance guidelines, it also revealed a challenge: the World Bank required full fiduciary accountability from implementers while offering limited flexibility for their constraints. In practice, third-party actors were tasked with operating in high-risk environments while adhering to audit-level compliance, which led to friction and operational challenges.

There is no established practice for TPI selection and contracting arrangements—country teams often need to determine TPI arrangements on an ad hoc basis, without centralized guidance on standards, cost negotiation, or partner selection. OPCS has negotiated framework agreements with third-party implementers that manage these engagements, ensuring compliance with the World Bank’s policies and procedures. For certain partners, existing corporate agreements with OPCS also set indirect costs. However, negotiating individual TPI contracts, including the direct support and operating costs (that is, overhead), is within the purview of the CMU. As there is no established practice for TPI selection, some CMUs take systematic and robust approaches, while others are inconsistent. Some CMUs feel supported by OPCS in negotiating direct costs, but others do not.

The lack of an exit strategy can provide perverse incentives for TPI arrangements to become a long-term plan rather than an exceptional modality for the World Bank to engage in FCS, contributing to potentially high costs. Such long-term arrangements can create the potential for moral hazard and create incentives that run contrary to development imperatives, as some implementing partners become financially dependent on Bank Group financing and contracts. A lack of an exit strategy can allow TPI to become a costly long-term default rather than an exceptional modality.

Most importantly, prolonged direct TPI poses risks to client capacity building and ownership. When humanitarian partners deliver services, they should support system or institution building and sustainability. For example, in South Sudan and the Republic of Yemen, projects that emphasized institution building through local NGO partnerships or parastatal institutions demonstrated a path toward sustaining developmental objectives even in the absence of formal government engagement. However, such focus on institution and systems building often did not consider sustainability beyond the TPI arrangements and any transition arrangements or exit strategy. Beyond a perceived lack of sustainability, reliance on TPI can reduce accountability, as TPI substitutes, rather than supports, the development of government-led checks and balances. The lack of government influence or participation in projects implemented through direct TPI can contribute to a perception among governments that these are UN projects, reducing opportunities for ownership and alignment with national development plans. Some countries have explicitly asked the World Bank to avoid further engagement with specific UN agencies due to dissatisfaction with reporting or perceived overreach. This underscores the importance of having clear exit strategies and transitioning to indirect or government-led models when feasible.

Experience in countries where both modalities are used has shown higher client ownership when projects are implemented through indirect TPI, or borrower financing. Additionally, projects implemented through third parties without government involvement do not typically build government capacity and systems and can drain government capacity as UN implementers or subcontracted local NGOs attract government staff with higher salaries. In those instances, TPI may contribute less to building institutions and government systems typically associated with development interventions. However, direct TPI projects, especially in the Republic of Yemen, have supported valuable parastatal institutions, thus preserving government capacity.

The World Bank’s development mandate calls for TPI to function not only as a means of delivering assistance but also as a tool for fostering sustainable institution building. This entails designing third-party arrangements that strengthen local capacity, allow for a shift toward government or community-led delivery when conditions permit, and prevent the creation of long-term parallel systems. Country evaluations show cases, such as in the Federal Republic of Somalia, when TPI programs integrated systems and institution building more proactively in their design, pointing to possible examples of leveraging partners’ strengths.