Improving on the Steady State in the State-Dependent Enforcement of Pollution Control

In a model with measurement error, optimal state-dependent enforcement of pollution control was shown by Harford (J. Environ. Econom. Management21, 67-81 (1991)) to reduce steady state expected annual social cost below the level attainable by a static enforcement approach. It is shown here that the expected present value of social cost can be reduced below the level attainable in a purely stationary model by choosing a value for the initial proportions of firms in the two groups that is different from the long-run equilibrium values. How this changes the optimal values of the monitoring probabilities and pollution standards for the two groups is addressed by considering small deviations from the purely steady-state optimum.