Technology is a complex adaptive system that is both shaped by, and shapes, institutional arrangements. This critical insight, developed in depth by W. Brian Arthur – the father of complexity economics – is relevant to researchers interested in institutions. Arthur provides a method for capturing the underlying dynamics. He offers conceptual tools centred on the concept of increasing returns to make sense of some crucial challenges. He also suggests technical tools, including agent-based modelling, to tackle ill-defined economic, legal, and institutional problems. This article explores his body of work and derives some institutional insights from it.

Under the Cross-Licensing system (CL), firms are allowed to trade non cooperatively the results of R&D efforts, and compete in the innovation and production stages. First, the paper proposes a simple modeling of this system. Second, a relevant comparison is made with the Cartelized Research Joint Venture (RJV), the form of R&D cooperation recognized to be the best one.

We prove that the Cross-Licensing system may be socially better than the Cartelized RJV. In terms of antitrust policy, for firms, the most favorable collusion mode is Joint Exploitation. We prove that firms are equally tempted by this mode, whether they are under CL or under Cartelized RJV.

This paper investigates the effectiveness of antitrust policy in fighting horizontal collusion under different regimes of policy implementation. We consider two regimes: a public agency regime, where an antitrust authority is in charge of competition policy, and a “delegation” regime, where the policy is chosen by consumers. In both regimes the policy is implemented under discretion. The analysis shows that delegation dominates, both in case of complete and imperfect information about production costs, the public agency regime because consumers credibly start off an higher level of investigation activity than the public agency. This result implies that the public agency will fight the cases involving “relevant” anti-competitive activities, while consumers will act also against “minor” violations. The combination of the two regimes yields an higher welfare than having only a public agency in charge of competition policy, because consumers can partially relax the agency's limited resources constraint.