Mid-term Evaluation of the IMF’s Data for Decisions Fund (D4D)

Background
At the request of the International Monetary Fund (IMF), SEO Amsterdam Economics conducted the mid-term evaluation of the IMF’s Data for Decisions (D4D) Fund. This thematic trust fund was established by the IMF in 2017 with the aim of putting more and better data in the hands of decision-makers, thereby enhancing evidence-based macroeconomic policies and supporting the achievement of the Sustainable Development Goals (SDGs). The Fund finances Capacity Development (CD) in low-income and lower-middle-income countries through four modules: (1) Addressing Data Needs and Quality Concerns, (2) Financial Access Survey, (3) Online Learning, and (4) Statistical Information Management. The evaluation covered Phase I (June 2018–April 2021) of the D4D Fund and assessed its performance against OECD-DAC criteria. The main purpose of the evaluation was to provide inputs for Phase II of the D4D Fund and to support institutional learning.  

Methods
SEO applied the IMF’s Updated Common Evaluation Framework, combining a Fund-wide portfolio assessment with in-depth case studies. The team analysed the IMF’s Results-Based Management (RBM) data, progress reports and budget information; carried out a large stakeholder survey among CD recipients, providers and donors; and conducted interviews with IMF staff, authorities and development partners. Five country case studies under Module 1, combined with dedicated assessments of Module 2 and Module 3, provided bottom-up evidence on results and good practices.  

Key findings 

• Relevance: D4D-funded CD was found to align well with country needs, IMF priorities and SDG-related data gaps, and primarily targeted countries that needed support most. Stakeholders generally considered CD design responsive to national priorities, though country authorities were not always sufficiently involved in work planning, which can limit ownership.  

• Effectiveness: CD projects largely delivered improved data compilation and dissemination and strengthened staff skills. While around 80 percent of rated milestones were largely or fully achieved, COVID-19 and limited national capacity constrained results in some cases; lack of medium-term country project plans and limited post-mission follow-up also reduced effectiveness.  

• Impact and sustainability: Impact was strongest where new or better data were actually used by authorities, IMF country teams or third parties (for example in debt sustainability analysis or Eurobond issuance). Sustainability was at risk in cases where senior management support, domestic financial and human resources, or institutional arrangements to retain knowledge were weak.  

• Coherence: Internal coherence with IMF surveillance, lending and other CD was generally good, especially where area departments and regional centres were closely involved. External coherence with World Bank and other CD providers was generally good, but coordination remained uneven across modules and countries.  

• Efficiency:Despite limited project-level cost data, case studies suggested that the Fund offered high value for money, with strong knowledge transfer and sensible cost-saving measures (such as online delivery). Budget execution was low in the first phase, partly due to COVID-19, but stakeholders were broadly satisfied with the timeliness, number and duration of CD activities.  

Key recommendations 

• Move from a pure “CD delivery” model towards a structured “change management” approach that actively supports organisational change, resource allocation, and data use by policymakers. 

• Require explicit needs assessments for data use in project design, covering national authorities, IMF staff and third-party users, and involve country authorities more in defining objectives and work plans. 

• Systematically assess and document country ownership and absorption capacity, and use this in prioritising and sequencing CD support. 

• Strengthen monitoring and evaluation at the level of data users by adding user-focused indicators, clarifying responsibilities for follow-up, and allocating budget for user-level monitoring. 

• Enhance coordination and synergies with other development partners and between D4D modules (especially between online learning and in-country technical assistance), while further combining remote and in-person CD.