Coherency management by architecture? : adapting enterprise architecture framework for the Finnish government ecosystem

Abstract

Public sector in the West has been exposed to government reforms for a long time. On-going public reforms do not report very high success rates, e.g. concerning digitalization. Public administration (PA) consists of administrative units at national, regional or local levels with various cross-dependencies. Government coherency refers here to a logical, orderly and consistent relation of the parts of the government to the whole. According to Zachman it is impossible to change a complex entity without descriptions of the current state to create different views of the entirety - to facilitate this, enterprise architecture has been approached. Enterprise architecture (EA) has been used to describe and design the business operations, information systems, data structures, and technology platforms of the enterprises. The knowledge on EA best practices is scattered and evolving, though. Since in PA differs from private businesses, e.g., by the mission and business logic, there is a need to study the adaptation and application of the EA in PA. In the study, we explored how the EA framework (EAF) should be adapted in the government ecosystem for coherency management. The study consists of two consecutive case studies in Finland, a national EA method development project and EA development in Kouvola City. In the first, the adaption model of the Finnish national EA framework was constructed. The model was evaluated in the second case study, where it was applied and developed by abductive logic of reasoning. We propose that EAF should be further developed conceptually and practically in order to present the current state EA information and descriptions in the frames of the prevailing management structures beyond other categorizations of the EA framework. In this way, EAF of the current state information would serve as data model of the current state management structures in the government ecosystem, and allow analyses of EA contents in relation to many more aspects than currently. To realize the practical benefits this should be instantiated as a common real-time EA information system. This might require a common contextual ontology in PA as the basis for the analyses of the as-is government ecosystem. The study follows abductive logic of reasoning and presents the ‘next best explanation' based on the available data, and therefore the results require further studies.