Introduction

The explanations of public policy networks are not a new phenomenon. As governments grow in complexity, networks appear to be a more relevant subject of study. Since the 1970s, an explanation of how they work, change, and influence public decisions is strongly needed (Heclo, 1977, 1978; McCool, 1995; Theodolou and Cahn, 1997; Sabatier, 1999; Bulcourf and Cardozo, 2008; Picard, 2011, Reynoso, 2011; Estevez, 2014a, 2014b).

In 1968, Dror drew attention to polycentric structures in policy making, at the same time that complex problems were increasing in relevance and visibility. Dror (1968, 1971) believed that it was necessary to look for “optimal” solutions to these problems and that structures (future networks) influenced how issues were defined. In his words (1968, p 206): “Polycentric structures with a few autonomous units have a basic operational rationale that is radically different from that of polycentric structures with many autonomous units … the latest, operate in terms of partisan mutual adjustment mainly by bargaining and forming coalitions with one another”. The increasing complexity of problems for governments is an issue that has grown since the 1970s (Burrell and Morgan, 1979; Denhardt, 1990; Baumgartner and Jones, 2009; Koppenjan and Klijn, 2015; Estevez et al., 2018). Concerning complexity, there would be three types (Koppenjan and Klijn, 2015):

• 1. Cognitive complexity: there are wicked problems, technologically very demanding and involving core values issues that are almost insoluble.

• 2. Strategic complexity: when several interdependent actors are involved with different perceptions and strategies.

• 3. Institutional complexity: when different institutions with different institutional structures and rules participate in a network.

According to Koppenjan and Klijn (2015), public policy networks having to deal with complex problems have to develop a different and adapted theory of governance. They call this “Governance Network Theory”, which has the following characteristics:

• • They explain a highly dynamic game between multiple actors, with different perceptions and strategies.

• • Good decisions are those that satisfy the widest possible actor's range in the network.

• • Success depends on cognitive, strategic, and institutional criteria of members participating in the network.

• • The manager tries to encourage interaction among the members, selfreflection, and designs procedures and rules to favor the dynamics of his network.

On the other hand, Rice (2015) points out that reflection on networks has its tradition in public policy analysis, but studies on management skills to conduct networks do not show the same development.