Abstract
Corruption remains a persistent governance challenge in South Africa, where comprehensive anti-corruption policies and institutions have achieved limited impact. This article addresses the gap between policy design and implementation by proposing a developmental evaluation framework for continuous assessment and learning. Drawing on principal-agent, collective action, institutional, and political settlements theories, the study explains why conventional reforms often fail to alter entrenched incentives and behaviors. Using a qualitative approach, it analyses legislation, the National Anti-Corruption Strategy (2020-2030), oversight reports, and scholarly sources, supported by insights from key informant interviews. The findings reveal four systemic barriers: elite resistance that undermines enforcement; institutional fragmentation and capacity deficits across watchdog bodies; data and transparency weaknesses that hinder risk monitoring; and the dynamic, sector-specific nature of corruption. These factors explain the persistence of implementation gaps despite strong formal commitments. The article advances an integrated framework anchored in the OECD Development Assistance Committee (DAC) evaluation criteria and a theory of change, emphasizing adaptive feedback and iterative learning. Embedding such evaluation within governance practice can strengthen institutional resilience, promote accountability, and rebuild public trust in South Africa's anti-corruption architecture.