Collaborative Governance by Design: What Formation Reveals About Performance

• In Event: Evidence-based Govenrance and Policy for Health, Energy, and Environment in South Korea

Abstract

This study examines how the structural formation of collaborative governance arrangements shapes their perceived performance. Traditional performance metrics often focus on policy outcomes, but these outcomes are frequently influenced by external conditions or unfold over long periods. Therefore, this study adopts an expanded view of performance that includes procedural dimensions such as trust-building, transparency, coordination, and stakeholder support.

Using data from an international case database (Douglas et al., 2020), 60 collaborative governance cases were analyzed based on five variables related to their formation: initiation type, stakeholder diversity, initial trust, decision-making transparency, and administrative support. A cluster analysis using Gower distance and PAM algorithm was conducted to identify latent types of formation, accommodating both categorical and continuous variables. The optimal number of clusters was determined to be three.

The first cluster, labeled "Procedural," includes cases initiated by third-party facilitators. These collaborations exhibit moderate trust, high transparency, and structured coordination. The second cluster, "Mandated Coordination," includes collaborations established through legal or formal directives. These show balanced characteristics with moderate trust and support, and relatively strong citizen legitimacy. The third cluster, "Relational-Voluntary," involves cases initiated internally with high levels of trust and transparency, but lower administrative support and weaker external outcomes.

Each cluster demonstrated distinct patterns in perceived performance across five dimensions: outputs and outcomes, procedural formalization, innovation, unintended results, and stakeholder support. While Cluster 1 excelled in structured processes, Cluster 2 showed strength in external legitimacy. Cluster 3 emphasized internal cohesion but was less effective in generating broader impacts.

Future research will classify these clusters into broader structural types and examine whether formation conditions produce path-dependent performance over time. The study also plans to assess within-cluster variation to explore how different mechanisms operate under similar formation logics. This research contributes to the literature by linking formation design to nuanced patterns of collaborative performance and offers practical implications for tailoring support strategies to different types of collaboration.

* This paper will be included in a submitted panel (Session ID: 2258871).

Author

• Yeobin Yoon, Pennsylvania State University
  Presenting Author