Resource balance, limited information and public policy

The U.S. Federal Government through its expenditure programs is having a major impact on all forms of pollution abatement. Total expenditures on Federal environmental programs amounted to $3.3 billion in 1972. A program of this magnitude can easily lead to waste and expenditures on conflicting goals unless carefully managed.

In a recent article in the Am. Econ. Rev. William Baumol examined the theoretical justification for and attacks on the Pigouvian tax and subsidy approach to controlling externalities. He concluded that while it is theoretically possible to control externalities through Pigou's procedure, the existence of multiple equilibrium and information requirements make it impractical. He proposes an alternate approach which changes the policy goal from maximization of social welfare to generation of acceptable levels of externalities. His goals differ drastically from those of the resource balance model presented by Kneese, Ayres and D'Arge of maximization of social welfare. This paper will combine the concept of resource balance with the goal of obtaining acceptable level of externalities at minimum cost in the presentation of a theoretical model that can be operationalized.

This paper proposes a general equilibrium method, utilizing the decomposition principle of linear programming, that will permit expenditures and regulations only on non-conflicting goals and includes a feedback mechanism for determining if a program is wasteful in relation to other programs. The model has five advantages over earlier proposals: (1) it is a general equilibrium rather than partial equilibrium model; (2) it is dynamic rather than static; (3) limited information rather than complete information is required for its implementation; (4) the model is heuristic rather than optimizing in the sense that policy decisions always increase the efficiency of pollution control, but because of the existence of uncertainty, they cannot be interpreted as maximizing social welfare; (5) the model incorporates multiple rather than a single policy tool.