ABSTRACT
Policy forums are lightly institutionalized and stable forms of governance networks that include administrative authorities, interest groups, and scientists. They are said to produce different types of outputs, from simple actor coordination to position papers and implementation documents, but their productivity has also been questioned. Metagovernance strategies can improve the capability of policy forums to produce outputs. To determine how different metagovernance strategies influence the capability of forums to produce joint position papers, 29 policy forums in the Swiss environmental sector are compared through a qualitative comparative analysis. Results indicate that metagovernance strategies such as state actors as forum members or majority decision rules need to be combined with small forum size or low actor heterogeneity. Furthermore, forum foundation by the state complicates the production of position papers.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. Concepts like bridging organization (Hahn et al. Citation2006, Crona and Parker Citation2012, Kowalski and Jenkins Citation2015), hybrid advisory committee (Krick Citation2015), boundary organization (Crona and Parker Citation2011), collaborative institution (Lubell et al. Citation2010), or partnership (Leach et al. Citation2002, Selsky and Parker Citation2005, Bauer and Steurer Citation2014) are used in the literature to describe policy forums.
2. Note that while we frame these factors as metagovernance strategies in this analysis, not all of them are in the reach of actors who might want to metagovern a forum. While state actors might be able to join a policy forum in order to try to improve its functioning through brokerage or a shadow of hierarchy, forum heterogeneity is inherently more difficult to influence, as it not only depends on state actors but also on other actors which are members of the forum.
3. Others were excluded because they were not policy forums in our definition but interest groups or implementation organizations. The second wave of surveys revealed no additional policy forums that were not already included in the first wave, indicating that we have covered a reasonable population of policy forums in Swiss environmental policy.
4. The consistency score provides information on the extent to which the empirical observation supports the postulate of a perfect relationship between the conditions and the outcome, or how well the solution formula describes the cases. The coverage score is an indicator of the empirical importance of a relationship. It indicates what proportion of the fuzzy-membership values of the cases in the set of the outcome can be explained by the solution (Ragin Citation2008).
5. We acknowledge that policy forums might be more or less productive, that is, the number of position papers produced by policy forums might strongly differ between forums. Our analysis focuses on the difference between forums that do produce (any number of) position papers, and those who do not produce position papers.
6. Regional state actors (Swiss cantons) are members in some forums. Given that we concentrate on policy forums with a national focus, we do not take regional state actors into account for operationalizing state membership.
7. This is admittedly only a rough proxy to account for the potential heterogeneity of interests within a policy forum. Another possibility would have been to take into account also private firms, which are strongly represented in some forums. However, by definition, interest groups represent broader interests than single, private firms. Accordingly, operationalizing heterogeneity with the ratio of both interest groups and private firms results in a solution with lower coverage and corresponds less well to theoretical expectations.
8. Alternative calibrations rely on (a) 95% instead of 71% (for HETERO) and 75 (for SIZE) as upper thresholds, (b) observed gaps in the data as crossover points (0.44 for hetero, 33 for size, see Appendix 3), or (c) observable gaps in the data distribution for HETERO (0 for values between 0 and 0.1, 0.33 for values between 0.11 and 0.3, 0.66 for values between 0.31 and 0.45, 1 for values above 0.45, see Appendix 3). Results are identical to the one presented in the paper, with the only exception of the fourth solution path being size*hetero*MEMBER*MAJOR instead of size*TDOWN*MEMBER*MAJOR for calibrations (a) and (b). This indicates that results are robust to different possible types of calibrations based on the data distribution.
9. This threshold is clearly above the 0.75 minimum threshold recommended by Schneider and Wagemann (Citation2010) and corresponds to the major gap in consistency values in the truth table.
10. Analyses are conducted with fsQCA (Ragin et al. Citation2009) and the QCA package in R (Thiem and Dusa Citation2013). Both reveal the same results.
11. In the tables, * stands for logical ‘and,’ and + represents logical ‘or.’ Conditions and outcomes written with capital letters stand for their presence; those in lower case letters indicate absence of the phenomenon.