From safeguard policies to outcomes: evidence from development finance institutions

Purpose This paper aims to analyze environmental and social (E&S) safeguard systems across major multilateral development banks using complaint data, evaluations and case studies. Despite increasingly comprehensive policies, project outcomes – especially in resettlement, fragile contexts and consultation – remain weak. The study finds strong policy convergence but uneven implementation, driven by limited borrower capacity, weak supervision, poor monitoring and accountability gaps. It concludes that effectiveness depends more on implementation culture than policy design. Priorities include shifting to results-based E&S management, strengthening accountability mechanisms, improving early integration, building institutional capacity and requiring disclosure of net development impacts. Design/methodology/approach This study uses a mixed qualitative-comparative design with four components. First, institutional mapping compares MDBs and bilateral DFIs using matrices of E&S frameworks, accountability systems and evaluation practices, selected by ODA share. Second, a systematic review of 13,543 World Bank, Inter-American Development Bank and ADB evaluations applies qualitative comparative analysis and keyword frequency to identify E&S patterns across sectors and regions. Third, an E&S scoring framework assesses 30 International Finance Corporation projects across Performance Standards, incorporating CAO penalties and expert-derived estimates. Fourth, caseload analysis of major accountability mechanisms examines complaints, harms and remedies. An additional AI-assisted ESPI measures directional E&S performance change. Findings The paper finds a persistent gap between increasingly sophisticated environmental and social safeguard policies and weak project-level outcomes across multilateral development banks. Despite strong policy convergence, implementation remains inconsistent, particularly in resettlement, consultation and fragile contexts. Evidence from evaluations and over 950 complaints shows recurring failures driven by limited borrower capacity, weak supervision – especially for medium-risk projects – and inadequate monitoring. Accountability mechanisms reveal rising complaint volumes and insufficient remedies, highlighting an “accountability gap.” Ultimately, safeguard effectiveness depends less on policy design than on implementation culture, requiring a shift toward results-based management, stronger oversight and transparent disclosure of both positive and negative impacts. Originality/value The paper’s originality lies in combining large-scale accountability mechanism data (over 950 complaints) with independent evaluation evidence and structured case studies to produce a cross-institutional, outcome-focused analysis of environmental and social safeguards. It moves beyond policy comparison to systematically link governance design, implementation practice and community-level impacts. Its value is in identifying consistent, cross-MDB failure patterns – particularly in resettlement, supervision and fragile contexts – and reframing the problem as one of implementation culture rather than policy gaps. By highlighting the lack of net impact disclosure and the limits of current accountability systems, it offers actionable, system-level recommendations for improving real-world safeguard performance.