Public Finance, 2nd Edition

considers that the number of real-world distortions is far greater than simply

the need to use distorting taxes, the prospects for achieving a uniWed norma-

tive theory of the public sector are indeed discouraging.

THE COASE THEOREM, BARGAINING, AND

PRIVATE INFORMATION

The Coase Theorem holds out hope that government intervention might not

always be required to solve market failures such as externalities. Its premise is

that in the absence of private information and transaction costs, private

agents have an incentive to bargain with one another and write whatever

contracts are necessary to extract all pareto-superior gains and thereby reach

the pareto optimum. The only prior requirement is the assignment of prop-

erty rights, so that ownership of the activities giving rise to or receiving the

externalities is clearly established.

No one expected the Coase Theorem to apply to instances of widespread

externalities because the transaction costs of bringing large numbers of

agents into a bargaining setting were likely to be formidable. But the hope

was that externalities involving only a few agents could be settled eYciently

through bargaining rather than by government intervention.

Coase published his Theorem in 1960. Since that time, developments in

bargaining theory have pretty much dashed the hopes of the Theorem even in

the case of small numbers of agents. There are two main problems. One is

that the formation of coalitions through cooperative bargaining requires that

certain conditions be satisWed for the coalitions to be stable. Unfortunately,

the set of stable coalitions may not include the pareto-optimal allocation, and

even if it does the bargaining process may not settle on the pareto-optimal

coalition. The other problem is that private information may limit the pay-

ments that agents are willing to make to other parties as part of a bargain.

The acceptable range of payments may preclude the payment that is neces-

sary to achieve the pareto optimum.

The Coase Theorem can get around these problems in principle by

assuming them away. Rational agents may be seen as rejecting any pareto-

inferior coalitions simply because they are pareto inferior. Or, rational agents

may be presumed to reveal their private information willingly if doing so

would lead to pareto-superior allocations. Assumptions such as these eVect-

(20.11) or (20.15): 1). the sources of tax revenues, particularly whether there is the possibility of a

lump-sum tax equal for everyone; 2). whether taxes are set optimally; 3). the distributional

characteristic of the public good, deWned in terms of the covariance between individuals' MRS

and their marginal utility of income; and 4). the eVect of changes in the public good on tax

revenues. Sandmo argues that (1) and (2) should be considered in calculating a marginal cost of

public funds but not (3) and (4), since they are speciWc to particular public goods.

668 THE COASE THEOREM, BARGAINING, AND PRIVATE INFORMATION