Regulation of Nonpoint Emissions under Limited Information:A Stress Experimental Test of the Ambient Tax Mechanism

We provide a stress experimental test of the ability of (a damaged based version) the ambient tax mechanism to induce socially optimal outcomes in a non-point pollution context. To mirror the features of naturally occurring environments, we consider a convex damage function, uncertainty in measuring the ambient level of pollution, polluters with heterogeneous profit functions competing against the same opponents for the duration of the experiment (which runsfor an indeterminate length), and in half of our treatments polluters do not know others profit functions. In almost all implemented conditions, the observed total pollution level is not significantly different from the socially optimal level whereas, in none of the conditions, complianceat the individual level is observed. The efficiency performance of the ambient tax mechanism is higher (though not significantly) under limited information than under complete information as subjects comply more with the socially optimal level the less information about the profitfunctions of others they have.