Governance and Coercion Fritz Sager University of Bern The present article departs from the assumption often found in literature concerning governance, which is that coercion is the quintessence of government and that, therefore, the growing importance of new forms of governance in policy formulation and implementation will lead to the adoption of softer policy instruments. This hypothesis will first be discussed in the wider context of the instrument choice literature, whereby an opposing view is derived. The two competing hypotheses are then tested in a comparison of the alcohol control policy designs of the Swiss member states, i.e. the cantons. The results of a multivariate regression analysis show that strong governance structures understood as networks embracing both public and private actors lead to the adoption of restrictive policy designs that must be enforced by public authority and as such cannot be employed by non-public governance actors. It is concluded that in their evaluation of policy instruments, governance actors follow a logic of consequen- tiality rather than a logic of appropriateness. Although the governance literature displays a great interest in the question of policy design choice (Hill and Lynn, 2005, p. 179), there has been surprisingly little theorising work based on empirical evidence concerning the question of the relationship between modes of governance understood as networks embracing both public and private actors and instrument choice. I use the term governance structure when I speak of modes of interaction among actors and institutions in the policy process, and the term policy instrument is used to denominate actual interventions addressed towards target groups in order to resolve a societal problem. Gerry Stoker (1998) argues that while governance is ultimately con- cerned with creating the conditions for ordered rule and collective action, its outputs are supposedly no different from those of government. Accordingly, David Richards and Martin Smith (2002, p. 279) state that it ‘is not what the state does that is different, but how it does it’ (emphasis added). Thus, if we want to assess the effects of this ‘difference in processes’, which Catherine Lyall and Joyce Tait (2005, p. 4) see as the factually significant feature of governance, we have to focus on the output of these differently structured processes. Hans Bressers and Laurence O’Toole (1998) state that the two concepts of policy networks (understood as governance structures) and policy instruments have for a long time been considered in isolation from each other in the relevant literature. Interesting insights have been offered in this regard by Andrew Jordan et al. (2005, pp. 478–9) who, based on an overview of the relevant literature,‘start from the proposition that the deployment of so-called “command-and-control” regulatory policy instruments is the quintessence of government (Pierre, 2000, p. 242). By doi: 10.1111/j.1467-9248.2008.00743.x POLITICAL STUDIES: 2009 VOL 57, 537–558 © 2008 The Author. Journal compilation © 2008 Political Studies Association