Governments worldwide are intensifying efforts to address escalating flood risks exacerbated by climate change. Central to this endeavor is the implementation of institutional frameworks, such as public policies, aimed at mitigating, planning for, responding to, and recovering from flood events. However, the effectiveness of these institutions relies heavily on their practical application. This study delves into the institutional landscape of flood risk management (FRM) through a comprehensive case study in Sint Maarten, a Caribbean island. Specifically, we scrutinize the degree of institutional compliance, focusing on the alignment between formally advised policies (institutions-in-form) and their informal adoption in practice (institutions-in-use). Employing Institutional Network Analysis (INA), we explore discrepancies between these two dimensions across the various phases of FRM (response, recovery, mitigation, and preparation). Our findings reveal that institutional compliance varies significantly across the FRM phases, with mitigation presenting the most pronounced challenges. Notably, the discrepancies are more prevalent among those tasked with implementing the policies rather than the targeted property owners. Generally speaking, the transition of institutions from mere forms to actionable rules is often hindered by established or emerging practices diverging from prescribed directives.

Climate adaptation measures are shaped and implemented through processes of governance, where the interactions and decision-making among actors lead to the creation and reinforcement of institutions. Institutions in this respect are the rules that shape the interactions of actors in different phases of climate adaptation. Currently there is no comprehensive method to systematically identify and map dependencies between institutions. This study proposes the Institutional Network Analysis (INA) approach that is based on the Institutional Grammar as a systematic and comprehensive tool to (1) visualise institutional dependencies, (2) identify points of concern in the institutional landscape such as conflicts and voids, and (3) provide quantitative insights into the centrality of actors, embeddedness of institutional outcomes, and dependencies between institutions. The approach is applied to the case of climate adaptation of transport infrastructures surrounding the Port of Rotterdam, the Netherlands. The analysis reveals a conflict in the use of risk assessment criteria, as parties in the Port may follow their own matrices in the presence of a shared decision-making framework. Nonetheless, the network metric analysis reveals that the criteria, whether shared or individual, build on the same source of risk analysis, suggesting that this duality may not be detrimental for climate adaptation efforts. Additionally, an institutional void is identified for financial responsibilities in locations where infrastructures overlap. Finally, the network metrics show high dependency between institutions in the risk dialogue phase, and reveal the centrality of infrastructure owners ProRail and RWS in the institutional landscape instead of local or regional governmental bodies.