Public Finance, 2nd Edition

nonmarginal price changes, in general. Consequently, incidence analysis

using the change-in-relative-prices measure often concludes the formal an-

alysis with the price changes. The link to consumer's welfare is then simply

presented in heuristic terms, in the form of general statements about who

gains and who loses (with many consumers).

In summary, then, the theory of tax incidence presents a quandary even

for simple one-consumer-equivalent economies. Despite the obvious motiv-

ation for developing empirical measures of tax incidence, there appear to be

no obvious candidates for the task unless production technology is linear.

With general technologies, unambiguous measures of welfare loss involve

compensated equilibria that cannot be observed in practice, and observed tax

and price vectors oVer at best only intuitive guidance to welfare losses. As a

practical matter, economists may have to be content with measures of price

changes in response to diVerent sets of taxes that leave the government

budget surplus unchanged, especially given that production technologies

are general and not linear.

The only Wrm conclusion one can draw is that if the incidence of a given

set of taxes is to have any meaning in a general equilibrium context, tax

incidence must be deWned in such a way as to render the impact of a tax only

indirectly relevant to the incidence measure. Tax revenues (or the resulting

budget surplus) from distorting taxes must be returned lump sum to the

consumer to have a well-deWned problem focusing on a single tax. The actual

tax payment can aVect incidence only through its inXuence on the amount

that market prices change in response to the tax. Regardless of whether one

chooses the income compensation or change-in-relative-price approach, the

Wnal incidence measure is fully determined by the resulting changes in the

general equilibrium price vectors.

THE EQUIVALENCE OF GENERAL TAXES

Although the income compensation and change-in-actual-price measures of

incidence approach the problem from diVerent perspectives, they each imply

the following important result: In a perfectly competitive, proWtless economy,

in which tax revenues (or the budget surplus) are always returned lump sum,

any two sets of taxes have identical incidence if they generate the same

changes in relative prices.

Consider, Wrst, the relative price measure of incidence. If production is

proWtless and tax revenues are returned to the consumers, actual consumer

demands (factors suppliers) are functions only of relative prices. Producers'

supply (input demand) relationships are also functions only of relative prices.

Therefore, two sets of taxes that generate the same vector of relative prices

generate the same with-tax general equilibrium. Consequently, they must

have the same incidence by the relative price criterion.

540 THE EQUIVALENCE OF GENERAL TAXES