Public Finance, 2nd Edition

17

EXPENDITURE INCIDENCE AND

ECONOMY-WIDE

INCIDENCE STUDIES

THE INCIDENCE OF GOVERNMENT TRANSFER PAYMENTS

TAX AND EXPENDITURE INCIDENCE WITH DECREASING-

COST SERVICES

SAMUELSONIAN NONEXCLUSIVE GOODS

The Incidence of Nonexclusive Goods: Empirical Evidence

ECONOMY-WIDE INCIDENCE STUDIES

THE SOURCES AND USES APPROACH

Annual Incidence Studies

The Pechman±Okner Studies

Mixing Annual and Lifetime Incidence

Pure Lifetime Tax Incidence

Lorenz±Gini Measures of Tax Incidence

COMPUTABLE GENERAL EQUILIBRIUM MODELS OF

TAX INCIDENCE

DYNAMIC TAX INCIDENCE

The Auerbach±Kotlikoff OLG Model

The Fullerton±Rogers Lifetime CGE Model

Taxes are most often raised to Wnance government expenditure pro-

grams, not just to substitute for other taxes. Once this obvious point is

conceded, it is no longer as compelling to speak only of the incidence of the

tax revenues. The policy-relevant incidence measure is clearly balanced-

budget incidence, the entire tax-and-expenditure package. One might still

argue that tax incidence itself remains relevant since diVerent sets of taxes

could have Wnanced the given expenditure program. Still, ignoring the ex-

penditure side is always dangerous since the very existence of a new expend-

iture program aVects the evaluation of the single tax and diVerential

incidence measures discussed in Chapter 16. Government inputs and outputs

573

enter into the market clearance and government budget equations, thereby

inXuencing the price responses to any change in tax rates. Also, the distribu-

tional consequences of expenditure programs are likely to be as important as

the distributional consequences of the tax revenues raised to Wnance them.

Thus, to the extent incidence analysis is an aid to governmental distributional

policies, considering the incidence of an entire tax-and-expenditure package

would appear to be the most useful strategy.

This is bound to be a diYcult assignment, however, even in theory,

because balanced-budget incidence theory is fraught with the same diYculties

as the theory of tax incidence, plus some other problems as well. At the very

least, an analysis of various balanced-budget alternatives must confront these

issues at the outset:

1. What measure of incidence will be employed? The three most likely

candidates are income compensation or welfare loss measures applied to

individuals, the change-in-relative-prices measure in the Harberger tradition,

or the change in a many-person social welfare function, the same as for the

theory of tax incidence.

2. For any given set of taxes, what expenditure programs are being

Wnanced? The obvious candidates are transfer payments, Samuelsonian non-

exclusive public goods or other externality-generating goods, and govern-

ment-operated decreasing-cost services, although the government might be

buying goods and services that could have been supplied by a perfectly

competitive private sector.

3. Will the analysis consider marginal, balanced-budget changes in taxes

and expenditures, or must it focus on a total package of Wnite taxes and

expenditures? Marginal analysis might make sense for transfer payments but

surely not for decreasing cost services.

4. For any given expenditure program, how are the taxes being raised?

The point that the choice of expenditures aVects the measurement of tax

incidence is reversible. The method of Wnancing the expenditures dictates the

approach to the measurement of expenditure incidence. Are the expenditures

assumed to be Wnanced with lump-sum taxes or with a set of distorting taxes?

If resource-using government expenditure programs are assumed to be

Wnanced with lump-sum taxes, then the incidence analysis could take place

within a Wrst-best context, so long as other appropriate assumptions are

made, such as perfectly competitive private production and marginal-cost

pricing of government services. Lump-sum Wnancing would also provide an

unambiguous method for considering the incidence of a single (set of ) gov-

ernment programs (s), or a separate theory of expenditure incidence, analo-

gous to the incidence of a single-tax program when the revenues are returned

lump sum. This is an important consideration, since Wrst-best expenditure

incidence is more compelling than Wrst-best tax incidence. Resource-using

expenditures are undertaken solely for eYciency reasons in a W rst-best

574 EXPENDITURE INCIDENCE AND ECONOMY-WIDE INCIDENCE STUDIES