Experiment-based policy change over time: Learning from experiences in the Dutch fen landscape.

• We explain potential of experiment-based policy to address environmental challenges. • Distinguishing incremental and synoptic experiments helps understand policy change. • Over time, incremental experiments were able to change practices albeit small. • Synoptic experiments tended to downsize experiments over time. • Capacity to navigate stakeholder dynamics and multi-level settings is key to change. This study provides insight into the potential of experiment-based policy making to address complex environmental challenges. The investigation shows the usefulness of making an analytical distinction between incremental and synoptic policy experiments to understand successful change over time. Incremental refers here to gradual adjustment of policies in a specific context, synoptic to radical change of policies on a large scale. Success is understood in terms of drawing lessons and realising policy goals, enabling diffusion and improving environmental status. This unique study, including both the initial investigation and a follow-up 10 years after, analyses the dynamics that emerged from seven empirical cases to identify factors affecting the design and success of these two types of experiments. We studied three incremental experiments and four synoptic experiments in the Dutch fen landscape to learn from their practical experiences. This densely populated and commercially used landscape is vulnerable to environmental pressure such as precipitation peaks, sea level rise and droughts. Initiators of the incremental experiments were able to change land-use management practices, albeit staying relatively close to the status quo. In contrast, while initiators of the synoptic experiments expressed higher ambitions in terms of change, these experiments tended to experience stagnation and difficulties in implementation and diffusion. Observations reveal that experiment-based policy change may crucially depend on capacity to deal with stakeholder dynamics and to embed experiments within multi-level institutional settings. [ABSTRACT FROM AUTHOR]