A footnote to Jonathan Lewis's interesting piece: the main example of international policy networks in the literature is the stuff on "epistemic communities" in the environmental field.
A phrase like "a network would have to be created" is a bit misleading. Policy networks evolve gradually as actors come into contact with one another. And the examples given--which I like--leave an implication that the lack of an existing network will inhibit the government from doing something (the super-collider case). The normal assumption, though it does by no means always work, is that a developed network will protect the status quo.
If anyone is interested I wrote a piece on Japanese policy communities in comparative perspective in a journal called Governance, Vol 2 # 1, January 1989.