Decentralised collective action is no longer on trial in the current academic debate on natural resources management.2 What still needs to be worked out are the conditions for its success and its optimal nesting in or relationship with other institutions. A vast body of research has addressed the question of under what (external) conditions decentralised collective action will succeed (for a recent review, see Agarwal 200 1, specifically in the context of forests, see Ostrom 1998). But there seems much less debate, let alone agreement, on the question of relationships between local-level collective action institutions and other institutions at different levels. From the perspective of those interested in proving that community-level institutions of resource management can and do work, often the only role the state can (and should) play is "non-interference" or at most "legitimisation"? If the need for "nested enterprises" has been recognised, it is only in the case of common-pool resources that are part of a larger system (Ostrom 1990:90). Even here, the precise meaning and form of "nestedness" are yet to be elaborated.4 The focus has remained on the conditions of success for local-level collective action in which the state implicitly has a minimal role.