Highlights

The World Bank’s engagement was well adapted to the capacity and financial constraints of the fragility, conflict, and violence (FCV) situation in Somalia. The phased and iterative approach of the portfolio was well suited to overcoming Somalia’s limited capacity. However, rapid program expansion could overwhelm existing capacity and approaches.

The Multi-Partner Fund for Somalia facilitated coordination among development partners and financed capacity building, enhanced supervision budgets, and third-party monitoring critical to the program’s success. The use of United Nations–supported project implementation has expanded the World Bank’s reach. However, the World Bank’s operating model does not adequately reflect the high cost of doing business in FCV and relies on Multi-Partner Fund financing for operating in Somalia.

The World Bank has achieved above-average project ratings and disbursement rates and shorter project preparation times. However, a lack of FCV-related indicators limits insights into the FCV impact of the portfolio.

Adaptations in Programming and Approach

Early financing limitations forced the World Bank to take a phased approach, moving from core governance reform to sector interventions. Before IDA resources became available in 2020, the World Bank, guided by the FY14–18 ISN, focused on building the foundations for larger-scale financing. Public sector investment projects described in chapter 2 focused on recurrent costs and payroll management, PFM reform, and civil service reform. With these early investments that helped put in place the basic governance systems, the portfolio could gradually expand under the FY19–23 CPF to include citizen-centric engagements such as water and rural resilience, urban resilience, and social safety nets.

The phased approach was suited to overcoming Somalia’s capacity limitations. Limited availability of financing forced teams to strategically use analytics and small trust-funded pilots before scaling up initiatives. Early engagement in water, education, social protection, and urban resilience was initiated through analytic work designed to lay the groundwork for increased investments. This approach allowed for adaptation to the highly dynamic context. Although the phased approach requires significantly more support from task teams, it is worthwhile to continue, as has been recognized by the PLR and Completion and Learning Review (World Bank 2022c, 2024f).

Recurrent Cost and Reform Financing used capacity and systems-building activities with financing based on disbursement-linked indicators with eligibility criteria and a sliding scale. This nimble approach supported discrete steps for building systems and state capacity that was paired with financing and tailored to the Somalia context. It also proved to be a vehicle for pragmatic and effective decentralized engagement with the federal member states (World Bank 2023e). Thus, the Recurrent Cost and Reform Financing series became an effective instrument for the Somalia context (see chapter 3).

The World Bank has increasingly been able to attract and retain staff based in Nairobi, while deploying more regular, flexible support to clients in Somalia. Attracting more (GH-level) staff with FCV experience remains important, as Somalia’s complexity requires staff willing to make a long commitment. Several World Bank projects depend on a small number of highly capable staff with a long record and deep understanding of the Somali context, making the program vulnerable to turnover.

Enhancing the Reach of the World Bank Portfolio

The World Bank has increased its geographic coverage to reach less secure and more challenging locations over time. World Bank engagement initially focused on the relatively safe, higher-capacity areas of Mogadishu, Somaliland, and Puntland. Those areas received two-thirds of government-centric project engagements and accounted for half of citizen-centric project sites in FY14–18. In the FY19–23 CPF period, as the program shifted to investments in human development and national infrastructure, geographic coverage also changed. Engagement in the higher-risk central and southern regions grew, and the first projects with near-national coverage took place. With available financing increasing, the number of individual project sites of citizen-centric projects rose by 265 percent from FY14–18 to FY19–22.1

As the portfolio expanded, the World Bank was able to operate where needs were highest but was also more exposed to conflict. During FY14–18, 52 project sites were within 20 kilometers of a battle, terrorist attack, or other conflict event. During FY19–22, this increased almost threefold to 141, which is half of all project sites. Almost 80 percent of the sites of projects targeting food vulnerability were in areas of crisis-level food insecurity (figure 5.1).

Figure 5.1. Citizen-Centric Projects Targeting Food Vulnerability, FY14–22

Image

Share Tweet Facebook WhatsApp Linkedin Email

Figure 5.1. Citizen-Centric Projects Targeting Food Vulnerability, FY14–22

Sources: Independent Evaluation Group; World Bank project documents (entire portfolio).

Note: Food insecurity phases were derived based on the calculation of the average between IPC’s data for the following periods: (i) October 2022 to December 2022 (current), and (ii) August 2023 to September 2023 (current). The determination of the corresponding food insecurity phase was determined by applying the same decision rules established by IPC (refer to IPC Technical Manual version 3.1). This map has been cleared by the World Bank Group cartography unit. IPC = Integrated Food Security Phase Classification.

Conflict and Fragility Sensitivity in Project Design

Few projects benefited from FCV analyses beyond that provided by the fragility assessments. Although almost half (45 percent, 22/49) of financing projects cited some source of FCV insights that had informed project design, this was largely limited to World Development Reports and lessons from other countries or previous projects. Positive exceptions include the Somali Urban Investment Planning Project, which benefited from cross-cutting studies on conflict and the socioeconomic impact of infrastructure components; the Somalia Emergency Drought Response and Recovery Project, which was informed by a World Bank internal Rapid Drought Needs Assessment; and SURP, which conducted institutional and social economy assessments for every city in which it operated.

Such analyses are particularly important given the lack of presence in the field and the absence of portfolio-level FCV risk monitoring, which would track the evolving FCV context, identify implications for the portfolio, and determine what course corrections may be needed.2 For example, some sectoral analysis has been conducted by the Governance Global Practice on the role of the state in the education sector (World Bank 2024l). The Multi-Partner Fund makes available dedicated funding to the World Bank’s country unit to support the implementation and operationalization of the fragility assessment, including FCV risk monitoring, FCV project lens, FCV diagnostic and political economy analysis, and FCV Enhanced Supervision Support.

The Somalia portfolio was designed to address conflict drivers. The 2023 fragility assessment identified three interlinked and reinforcing fragility drivers. Addressing the state-state conflict driver requires strengthening the stability, coherence, and capacity of the state. A third of the portfolio (15/49) primarily targeted this driver, mostly through PFM interventions (6/15). This was complemented by an ASA portfolio that provided direct analytic and research support, with a quarter of ASAs (16/66) covering PFM (World Bank 2015).

Overcoming the state-society conflict driver requires the state to provide more services and benefits to its people to gain authority. Over half of the portfolio (28/49) primarily targets this driver, with a focus on capacity injection and service delivery, the topic of half (37/66) of ASAs. The World Bank has set up civil service training institutions and developed training programs.3 Through the Capacity Injection Project, the staffing of line ministries and central agencies was strengthened, and operational plans for all key ministries and monitoring and evaluation frameworks were set up to track their performance (World Bank 2024d).

Inclusive access to resources and mechanisms to manage intercommunal disputes are key to addressing the society-society conflict driver. This driver was the least addressed by the World Bank program. The Biyoole project addressed population displacement because of insufficient water resources and weak government authority, strengthening local intercommunity trust (World Bank 2024k).

Source: Independent Evaluation Group portfolio analysis.

Note: CPF = Country Partnership Framework; ISN = Interim Strategy Note; UN = United Nations.

Overall, 67 percent of lending projects were designed with adaptations to the FCV context. These adaptations were included to mitigate conflict risks to the projects, their staff, or property, or to overcome conflict-related risks to achieving the project objective while not exacerbating conflict risks (table 5.1). Adaptations were most common in citizen-centric projects, as they are more exposed to FCV risks than government-centric projects. Other design adaptations included being sensitive to clan issues when hiring consultants and awarding contracts or using a two-track project rollout (World Bank 2014b).4 Projects least frequently included local partnerships and FCV indicators in their design.

Implementation Modalities

To overcome government capacity constraints, the portfolio remains reliant on UN agencies and NGOs to support implementation. The Bank Group used three project delivery arrangements: government-led implementation, third-party implementation, and hybrid models. The portfolio’s expansion and sectoral diversification added significant complexity and increased reliance on the implementation capacity of state and municipal governments. Their persistent access constraints challenged implementation and monitoring. To overcome this, the World Bank has continued to rely on hybrid arrangements, in which the World Bank concludes an agreement with the government that signs service contracts with UN agencies or NGOs to deliver specific components, reducing the burden on the government.5 In both the ISN and CPF periods, a quarter of projects had hybrid implementation arrangements. As of January 1, 2023, government-managed service contract commitments with UN agencies totaled $342 million, of which $314 million has been disbursed, representing 52 percent of the active portfolio.6

IEG found no evidence that hybrid-implemented projects operate more quickly or in riskier environments. Hybrid implementation of projects reduces the government’s project implementation burden. The fragility assessment also argued that it allowed projects to operate faster and further than a government PIU could achieve. No evidence of this has been found. For example, the fully third party–implemented Somalia Emergency Drought Response and Recovery Project was able to achieve, on average, 76 percent of its direct support targets within the first six months of effectiveness.7 Other projects that combined direct support with longer-term investments do not show consistent differences in achievement rates between conventional government-led projects and the hybrid approach. There is also no evidence that hybrid-implemented projects can operate in more insecure environments than government-led projects. During the CPF period, government-led citizen-centric project sites were almost twice (1.78 times) as likely to be within 20 kilometers of a conflict event as were hybrid-approach project sites. Yet hybrid implementation can limit implementation pressure on the government.

The 2023 fragility assessment reported challenges in mobilizing and delivering financing under the Crisis Response Window or Contingent Emergency Response Components (World Bank 2024g).8 Even when funds are disbursed quickly by the government to an implementing partner, they can still take months to reach beneficiaries. For example, the Shock Responsive SNLRP took less than eight months to go from concept to first disbursement in December 2020.9 Yet over 80 percent of disbursements were made to implementing partners after April 2021, a year after the start of the locust outbreak. The Somalia COVID-19 Emergency Vaccination Project did not disburse in its first year of effectiveness, with the first disbursement taking place in October 2022 (World Bank 2024g). The Bank Group’s comparative advantage is in providing predictable financing for systems strengthening and investments that address the underlying drivers of vulnerability and fragility, as demonstrated in the SURP, Biyoole, and Baxnaano projects.

There are trade-offs between short-term delivery needs and longer-term statebuilding objectives. Relying on UN agencies and NGOs for project delivery can undermine efforts to strengthen country systems and build state legitimacy (World Bank 2022e). Some projects mitigated this by providing dedicated systems-building contracts to UN partners to ensure that in addition to service delivery, government capacities and systems are built. Even when the government is not implementing, its role, whether coordination or facilitation, should be communicated to the public to support state authority.

Inclusion, Beneficiary Selection, and Feedback

Most citizen-centric projects took measures to ensure inclusion and avoid doing harm. As the World Bank portfolio expands and reaches ever more vulnerable areas, its risk profile increases, as does its potential to do harm. The 2020 FCV Strategy emphasizes that development programming should follow “do no harm” principles (World Bank 2020g). An important component is ensuring the inclusion of underresourced groups among beneficiaries. Three-quarters (76 percent; 19/25) of citizen-oriented projects included measures to reach vulnerable groups such as agro-pastoralists, victims of gender-based violence, elder citizens, and those affected by droughts, food insecurity, and locust infestations. As projects expand support for service delivery, the need to focus on fostering social cohesion increases.10

An increasing number of projects include beneficiary feedback in their results frameworks, yet consultation with local stakeholders remains modest. Sixteen percent of FY14–18 ISN projects (3/19) included indicators to track beneficiary feedback. The share rose to 57 percent (17/30) in the FY19–23 CPF period. These indicators, on satisfaction, feedback, and perception of project success, help track beneficiary perceptions and reduce the potential of doing harm (World Bank 2014a). However, only 8 percent (4/49) of projects involved partnerships with Somali organizations. These beneficiary engagements were considered key to achieving beneficiary satisfaction and preventing local conflict (World Bank 2019e).

Monitoring and Supervision of World Bank Operations

A lack of FCV-related indicators limits insights into the direct FCV impact of the portfolio. The portfolio explicitly aims “to address the structural drivers of fragility” (World Bank 2018a, 6), yet very few projects measured direct FCV impacts. The Biyoole project was unable to complete an intended monitoring system for localized resources-based conflict but did record perceptions of reduced conflict and tracked community stakeholders who attributed project benefits to the government (World Bank 2019e). The Special Financing Facility for Local Development similarly tracked perceptions of Regional Administrations or federal government performance and benefits for both IDPs and host communities, as did SURP (World Bank 2016a). Yet the vast majority of projects did not include relevant indicators, without which the portfolio’s impact on stabilization efforts cannot be assessed.

The innovative use of a third-party monitoring agent financed by the Multi-Partner Fund was critical to collect data on project indicators. The fund has financed enhanced supervision budgets and a portfolio-level third-party monitoring agent. Especially in southern Somalia, the World Bank is fully dependent on third-party monitoring. The monitoring agent has also provided capacity-building support to PIUs and government counterparts. This is complemented by a professional services contract allowing task teams to hire consultant-type assets in the field beyond the security duty of care of the World Bank (World Bank 2022b). The success of this and other mitigation measures is shown by the fact that, despite Somalia’s significant security and capacity challenges, only 1 closed and evaluated project (out of 15) did not collect data on all its indicators.

The experience with the Geo-Enabling Initiative for Monitoring and Supervision (GEMS) has been mixed, and its rollout would require appropriate support. GEMS can help track risks and monitor project outputs in real time, enabling more strategic oversight.11 However, there are limitations, as PIUs need to supply sufficiently granular data, which only six projects do so far. An attempt to track the World Bank’s drought response could not be completed because of a lack of data. Many PIUs need support to take this on. Finally, GEMS is a picture-based platform and does not deliver the human intelligence of site visits. The Biyoole project mitigated this by working with CARE International. This yielded not just pictures but also the benefits of a strong local network at a lower cost than the third-party monitoring agent.

Partnerships and Donor Coordination

The Multi-Partner Fund for Somalia fostered an environment of strong donor coordination. The Multi-Partner Fund coordinates donor financing and allows the use of country systems to finance investment projects and increasingly cofinance IDA investments. This financing created an operating environment where collaboration and partnerships became the norm. Although the portfolio has been transitioning from Multi-Partner Fund–only to hybrid IDA–Multi-Partner Fund financing, the Multi-Partner Fund continues to serve as a platform for dialogue. As a result, development partners and counterparts perceive the World Bank in Somalia as very open to cooperation and coordination. The World Bank also engages frequently with other organizations, including the European Union,12 the International Organization for Migration,13 the African Development Bank, and key bilateral donors such as Germany, Norway, and the United Kingdom. Donor coordination and harmonization are key to avoid overstretching the authorities and donor overlap (World Bank 2022c). Continuing good cooperation is vital even as the relative importance of the Multi-Partner Fund declines and IDA funding becomes increasingly available.

The aid architecture supported by the Multi-Partner Fund was a catalyst for a strong and productive UN–World Bank partnership in Somalia. The UN–World Bank Partnership Trust Fund financed early joint work on the Multi-Partner Fund aid architecture. Recognizing the value of this, both institutions devoted more internal resources to the partnership, for example by financing a UN–World Bank liaison officer to deepen cooperation and advance joint priorities and activities.14 Partnership with the UN Assistance mission in Somalia also allowed the World Bank to engage in the security and justice sectors while staying within its mandate and comparative advantage. Whereas United Nations Somalia brought its political mandate, network, technical expertise, and experience in the Somalia security and justice sectors, the World Bank offered technical expertise in PFM and human resources reforms.

Flexibility in Policies and Processes

The pragmatic application of Operational Policy (OP) 7.30 avoided a complete disengagement in 2021–22. In 2021 and 2022, Somalia faced a political impasse. Parliamentary elections were postponed multiple times because of disagreements. This situation resulted in the contentious extension of President Farmaajo’s term beyond the constitutionally mandated end date. In response, the World Bank conducted a review under OP 7.30 in April 2021. OP 7.30 lays out the conditions under which the World Bank can suspend disbursements for existing country financing instruments because of the accession to power by a de facto government by means not provided for in a country’s constitution. The assessment found that despite political turmoil, Somalia had succeeded in ensuring smooth implementation of the portfolio, preparing new operations, and honoring all financial commitments with the World Bank. Moreover, neither the incumbent government nor any opposition leaders challenged or objected to the government’s ability to secure development partner assistance.

The OP 7.30 assessment concluded that the preparation, appraisal, negotiations, and approval of IDA19 pipeline operations should continue, with a focus on supporting Somalia’s resilient and inclusive recovery. This meant prioritizing projects with a clear focus on crisis response and human capital while postponing more politically sensitive projects (such as a proposed DPF operation and a currency reform project) until after a transition of power. The Country Management Unit made clear that HIPC could not continue without a resolution of the impasse and progressively curtailed the options for engagements while maintaining ongoing projects. It also warned of the potential complete withdrawal of World Bank funding, creating an incentive to resolve the impasse while not undercutting ongoing projects. Thus, the World Bank avoided the cessation of activities, which in other countries had disrupted programs (World Bank 2021f; World Bank 2022g). The application of OP 7.30 was lifted after the election of President Hassan Sheikh Mohamud in May 2022.

Engaging in Somaliland required additional adaptation and flexibility. As Somalia moved from Multi-Partner Fund to IDA financing, the World Bank had to sign financing agreements with the Federal Government of Somalia, which in turn had to sign subsidiary agreements with the federal member states. This became a challenge in the case of Somaliland, which does not recognize Federal Government of Somalia authority and has no official channels of communication and fiscal transfers with the Federal Government of Somalia. Through careful negotiations led by the country team with support from the Legal Vice Presidential Unit, the first subsidiary agreements were signed in 2021, having so far allowed 11 IDA-financed projects to operate in Somaliland.

Financial and procurement controls remain weak, as does the capacity to manage contracts, both of which increase fiduciary risks, especially as the portfolio grows and becomes more complex (World Bank 2024f). Project investments are now divested to regional and local governments and spread across larger areas, constraining the ability to apply standard methods for project supervision and implementation support. The World Bank requires 100 percent prior review of procurement transactions and operational costs, frequent reviews, and reporting on key transactions (World Bank 2022c). As a result, World Bank staff are required to provide “no objections” for most project decisions and expenditure requests above $500. The World Bank has established External Assistance Fiduciary Section units at the federal government and the federal member states to handle the financial management responsibilities for all externally financed projects (World Bank 2024b). This strengthened and consolidated national capacities while reducing the risk of overstretching limited human resources (World Bank 2022c). However, as the portfolio grew, financial and procurement control challenges increased, putting significant demands on project teams’ time and contributing to implementation delays (World Bank 2023e, 2024g, 2024j). Staff perceive this as a lack of operational flexibility and see a need for more operational support for project teams. Similar constraints apply to the Bank Group’s ESF, where client capacity for application is limited (World Bank 2024g).

Security and Cost of Doing Business

The rapidly expanding portfolio is stretching the logistics and security arrangements needed for project appraisal and supervision. The establishment of centralized platforms for mission travel early in the Bank Group’s reengagement enabled more flexible and frequent travel to Mogadishu, Garowe, and Hargeisa. However, when Bank Group staff visit Mogadishu, their engagements are mostly confined to a small compound at the airport, which is insufficient to meet the increasing demand. Travel to other cities, including the other federal member states, such as Kismayo and Baidoa, and especially rural areas, including in Puntland, remains challenging and a constraint on project site selection (World Bank 2018c and project documents).15 There, teams remain reliant on the UN for security, accommodation, and logistics, while areas outside of secondary cities in South Central are still off-limits (World Bank 2022f).

The Bank Group’s operating model does not adequately reflect the high cost of doing business in Somalia’s FCV environment and depends on Multi-Partner Fund financing for continued operation. Engagement in Somalia requires intensive client engagement, as reflected in high appraisal and supervision costs. Presence in the field is crucial, but security and logistics costs are high.16 Field visits are prohibitively expensive. Reportedly, a supervision visit to one or two Biyoole sand dams was budgeted at $10,000 against a yearly supervision budget of $20,000. Task team leaders perceive their supervision budgets as too low, limiting the quality of supervision and learning. The Multi-Partner Fund financed much of the additional cost for the portfolio’s enhanced risk management, including project appraisal and supervision budgets, security and logistics costs for mission travel, and the portfolio-level third-party monitoring agent, and will continue to at least until 2028 (World Bank 2024g). However, this dependence on trust fund resources for essential project costs indicates that the Bank Group’s business model is insufficiently adapted to the high cost of doing business in FCV.

Project Preparation and Implementation Times

The World Bank has achieved above-average project ratings and disbursement rates and short project preparation times, but government absorptive capacity may be declining. In FY22, projects in the Somalia portfolio took just over 17 months from Concept Note to first disbursement, in contrast to the 22-month average for FCV countries and the 25-month average in non-FCV countries. Disbursement rates have been high. In FY24, the disbursement ratio for the portfolio stood at 44 percent, slightly above the FCV average of 41 percent (World Bank 2024e). However, a decrease in investment project financing disbursement rates in recent years, from 58.7 percent in FY20 to 34.6 percent in FY21 and 30 percent in FY22, indicates that the growth of project sizes is exceeding government capacity, according to the 2023 fragility assessment.

Outlook

The expanding program poses challenging questions about the balance between personal safety, cost, and in-person supervision. Staff value in-person site visits over reliance on third-party monitoring or GEMS, but staff interviewed for this evaluation expressed that Corporate Security limitations had become more restrictive, limiting access to project sites. More missions at an acceptable risk level could be made possible by spending more on security. Corporate Security indicated that greater flexibility could be achieved by expanding their team, making them less reliant on local vendors, and by expanding the security budget to allow for more armored cars, secure accommodation, and intelligence.17 Corporate Security saw a potential for greater and earlier cooperation with task teams by increasing their involvement in project design to allow more security-informed site selection and accurate projection of supervision costs.

Improvements in core government capacity have not kept pace with the rapid growth of the portfolio. World Bank support has been channeled through country systems to strengthen government service delivery capacity and accountability to its citizens. However, improvements in core government capacity and client institutional capacity to implement projects, particularly in procurement, financial management, and safeguards, notably ESFs, have not kept pace with the increasing size and complexity of the portfolio (World Bank 2024f, 2024g). Projects are reliant on core government systems, which are often overstretched and not always interoperable across federal and state levels. This limits the application of systemwide workflows and controls. Few projects have established effective supervision and governance arrangements from their parent ministries, and PIUs are hindered by a lack of institutional support from broader government institutions, according to the 2023 fragility assessment. There is demand among teams for common approaches, similar to the External Assistance Fiduciary Section units, to address government capacity, especially on ESF implementation.