Coasian Solutions to the Externality Problem in Experimental Markets

Abstract

This paper evaluates Coasian solutions to the externality problem in a series of experimental markets. In an important and innovative study Plott (I983) establishes that there is indeed a behavioural externality problem to solve, in the sense that competitive markets with externalities do converge to a private equilibrium that ignores social costs. Plott (I983) then considers a series of alternative corrective policies designed to remedy this problem. The policies considered are 'Pigouvian taxes', 'standards', and 'marketable licences'. In each case the regulator is assumed to have sufficient information to apply the optimal policy. This informational restriction is readily justified by the exploratory nature of the enquiry.1 We extend the study of possible remedies to include decentralised Coasian bargains, where the only role for the government (or the legal system) is to define transferable property rights. Experimental studies by Harrison and McKee (I985) and Hoffman and Spitzer (i 982; I 986) have demonstrated that the Coase Theorem (i 960) is behaviourally 'alive and well' in relatively sterile and abstract bargaining environments. In these experiments subjects were given payoff tables which told them that if the group decision were a particular number they, individually, would receive those payoffs. In addition, one subject (or group of subjects) was allocated the right to make the decision without consulting the other subject (or subjects). This subject was designated the controller. Those not allocated that right could offer to transfer some of their earnings in order to persuade the controller to enter into a joint agreement. Following Regan (1972), we take the Coasian approach as implying two distinct behavioural outcomes: (i) that the parties will agree on a Pareto optimal level for the externality-generating activity, and (ii) that any such agreement will be attained by means of a mutually advantageous bargain between the parties. Harrison and McKee (i 985) and Hoffman and Spitzer (I985; I986) present experimental evidence for small and large bargaining groups that overwhelmingly support both outcomes. Hoffman and Spitzer (I982) support the first but not the second outcome. Regarding the second outcome, both Harrison and McKee (i 985) and