Abstract
Public service organisations operate within layered and often competing institutional logics shaped by successive waves of reform, including Public Administration, New Public Management, New Public Governance, and Public Value. While prior scholarship has examined these paradigms as sequential or co-existing governance traditions, limited attention has been paid to how their cumulative effects shape public service operating models and their capacity for reconfiguration. This article advances a dynamic capabilities perspective on public service systems. We argue that successive reform logics sediment into operating models, imposing distinct and frequently conflicting operational capability demands, ranging from service reliability and efficiency optimisation to network coordination and ecosystem orchestration. By introducing capability layering, we extend dynamic capabilities (DC) theory into public service ecosystems and reposition reform coexistence as cumulative capability conditioning. Under conditions of institutional plurality, public service organisations must continually reconfigure governance structures, resource configurations, and delivery routines to sustain value creation. To explain this process, we introduce a capability layering framework that conceptualises how reform legacies condition both reconfiguration capacity and performance stability. We develop a typology of capability configurations across public service operating models, demonstrating how variation in institutional dominance shapes the scope and limits of transformation. In doing so, the paper extends dynamic capabilities theory into public service ecosystems and contributes to the study of public service operating models by theorising governance–operations coupling under institutional complexity. In doing so, the paper extends dynamic capabilities theory into public service contexts and contributes to public service operations scholarship by theorising how governance logics become embedded in capability layers that shape operating model reconfiguration.