Provision of State and local goods

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

effectively included in the rent. Now the water department decides to

install separate meters for each apartment and to bill each separately

rather than the landlord (so the rent is reduced by $X per person for all

tenants). The city justifies the cost of the extra meters and billings on the

grounds that the city's scarce water resources will be used more efficiently.

What is the price to a tenant or apartment per gallon of water before and

after the new meters are installed? Do you think the new procedure will

reduce water use? If so, how might the student tenants of these apartments

act to conserve water? Will there be a gain in economic efficiency?

SELECTED READING

"Costing and Pricing Local Government Services." Governmental Finance, 11 (March 1982): 3-27.

Gramlich, Edward, "Let's Hear It for User Fees." Governing (January 1993): 54-5. Muskin, Selma, ed. Public Prices for Public Products. Washington, D.C.: The Urban Institute,

1972.

Netzer, Dick, "Differences in Reliance on User Charges by American State and Local Governments." Public Finance Quarterly, 20 (October 1992): 499-511.

CHAPTER 9

INTERGOVERNMENTAL GRANTS

The basic economic justification for federal functional grants-in-aid is provided by the widespread, and ever- increasing, spillover of benefits from some of the most

important state and local expenditure programs.' —GEORGE F. BREAK

"THE FEDERAL GOVERNMENT' SPENT $1.9 TRILLION IN THE STATES, THE DISTRICT OF COLUMBIA,

PUERTO RICO, AND OUTLYING AREAS DURING 2002, ACCORDING TO TWO REPORTS RELEASED

TODAY BY THE COMMERCE DEPARTMENT'S CENSUS BUREAU. THIS WAS AN 8 PERCENT INCREASE

OVER 2001.

CALIFORNIANS BENEFITTED THE MOST, RECEIVING $206 BILLION, FOLLOWED BY THE

PEOPLE OF NEW YORK ($129 BILLION), TEXAS ($123 BILLION), FLORIDA ($105 BILLION),

AND PENNSYLVANIA ($86 BILLION). ONE-THIRD OF ALL FEDERAL EXPENDITURES WENT TO

PEOPLE LIVING IN THESE FIVE STATES, WHICH ACCOUNT FOR 36 PERCENT OF THE TOTAL U.S.

POPULATION.

ALTOGETHER, SOCIAL SECURITY, MEDICARE AND MEDICAID ACCOUNTED FOR $890 BIL-

LION (47 PERCENT) OF THE U.S. GOVERNMENT'S 2002 DOMESTIC SPENDING.

GRANT AWARDS CLIMBED TO $412 BILLION, AN 11.6 PERCENT INCREASE OVER 2001, WIT

MEDICAID, THE LARGEST, AMOUNTING TO $148 BILLION, UP 11.1 PERCENT. 2 "

196

HEADLINES

hthisnvernmental Relations in the United Siate-s. W ashington, D.C2 The Brookings InstitUtiOn, 1987, 105:

'U.S. Department of Commerce. Press release entitled "Federal Domestic Spending Up 8 Percent in 2002, Census

Bureau Reports," June 4, 2003.

197

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

198

Intergovernmental grants, sometimes called grants-in-aid, are transfers of funds from one government to another, most often from a higher-level government in the federal system to a set of lower-level governments. These grants are of many different types and are intended to improve the operation of a federal system of government finance. In this chapter, the purposes for grants, the economic effects of the different types of grants, and then an appropriate policy of grant use are considered.

GRANTS IN THE U.S. FISCAL SYSTEM

In 2002, the federal government transferred more than $360 billion of aid to state-local governments, which represented about $.27 for every $1 raised by state- local governments from their own sources. By 2003, federal grants had increased to nearly $386 billion. The Census Bureau report for 2002 shows 663 different federal grant programs applying to state or local governments. Similarly, state governments transferred more than $355 billion to local governments, or about $.60 for every $1 collected by local governments from their own sources. As reflected by the data in Figures 9.1 and 9.2, intergovernmental grants, both from the federal government to states and localities and from the states to localities, have been a dominant feature of the federal fiscal system in the United States for more than 30 years.

Although the absolute magnitude of these grants generally has increased annually over these years (1982 and 1987 being exceptions for federal aid), the purchasing power of grants has not. After adjusting for price increases (measured

Figure 9.

Federal grants

to state-local

governments

400000

350000

300000

25

(11

n 250000 -o 46 200000

0 150000

100000

50000

0 0

1962 1967 1972 1977 1982 1987 1992 1997 1999 2000 2001 2002

Year

Real federal grants -II- Federal grant share

CHAPTER NINE ■ INTERGOVERNMENTAL GRANTS

1 99

Figure 9.2

40 State grants to

local governments 35

30

4 200000 0

150000 15

100000 10

50000

0 1962 1967 1972 1977 1982 1987 1992 1997 2002

Year

I ON State grants -U- State grant share

by the implicit GDP deflator), the real value of federal grants (measured in 2002 dollars) declined in 1973 and substantially from 1978 to 1982, despite increases in the nominal amounts, as shown in Figure 9.1. Essentially, grants increased during these periods at a slower rate than prices rose. The real value of federal grants essentially then remained the same until 1990 and has increased substantially since. In contrast, the real value of state grants to local governments never really declined, although it was nearly constant from 1979 to 1984 and from 1989 to 1991, and has increased in the past decade.

The relative importance of intergovernmental grants increased in the 1960s and early 1970s, peaked in the late 1970s, and has increased gradually since 1987. As shown in Figure 9.1, federal aid increased from about 13 percent of state-local gen- eral revenue in 1962 to about 22 percent by 1977. After a decline, it had risen back to about 22 percent of general revenue for 2002. Over that same period, federal aid increased from less than 9 percent of total federal government outlays to 17 per- cent by 1978. Federal aid's share of the federal government budget then decreased to less than it was in 1969, only to begin rising again since 1990. Although state and local governments, on average, received $.32 of federal aid for every local dollar collected in 1978, and that amount fell to $.18 in 1989, it rose to $.27 by 2002. Sim- ilar decreases in the relative importance of state aid for local governments also occurred in the 1980s, at least partly because states were receiving less federal aid to pass along to localities, as shown in Figure 9.2.

Intergovernmental grants are an important source of revenue for nearly all state-local governments, as confirmed by the data in Table 9.1. In 2002, state

350000

300000

250000

5

0

25

20

P e

rc e

n ta

g e

o f g

e n e ra

l r e

ve n u e

200

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

Table 9.1

Intergovernmental Grants as a Percent ot General Revenue, by Type

of Government, Various years, 1962-2002

Local Governments

Total

2.0% 28.4 30.4

4.3 33.4 37.7

9.2 33.7 42.9

7.6 33.9 41.5

4.8 33.3 38.1

3.5 34.2 37.6

3.9 34.6 38.4

2.9 35.7 38.6

Counties

0.7% 36.3 38.6

1.7 39.1 42.1

9.0 34.5 45.3

6.5 34.1 42.0

3.6 31.7 36.8

2.2 33.5 37.3

2.6 33.5 37.5

2.9 33.8 38.4

Municipalities

25% 16.3 20.4

7.3 24.1 32.9

14.7 23.2 39.7

12.0 20.8 34.6

6.5 20.3 29.0

4.6 21.3 28.3

5.3 20.7 283

5.3 21.9 29.8

Townships

0.8% 20.6 22.5

1.3 19.6 22.0

7.5 20.4 29.7

5.8 22.6 30.1

3.4 22.5 28.1

1.2 21.0 24.3

1.3 19.7 23.3

1.2 19.7 23.3

School Districts

1.4% 37.3 40.8

1.9 42.0 45.0

1.5 47.3 50.2

1.0 51.7 54.3

0.9 52.8 55.3

0.7 52.0 54.0

0.7 53.3 55.2

1.0 54.5 57.4

Special Districts

8.9% 3.2

21.1

15.5 3.9

29.6

21.7 7.4

38.2

18.5

34.7 7.6

16.0 5.3

29.4

14.2 6.8

29.4

14.9 8.7

32.3

17.3 9.2

31.7

'Includes grants from local governments.

SOURCE: U.S. Department of Commerce. Governmental Finances, 1962, 1972, 1977, 1982, 1987, 1992, 1997, 2002.

governments received essentially 30 percent of their revenue through intergov- ernmental grants, and local governments received more than 38 percent, with coun- ties and school districts being the types of local governments most reliant on grants, at least on average. Although state aid is substantially more important than direct federal aid for all types of local governments except special districts, some of that state aid arises from federal grants to the states, which effectively are "passed on" to localities. The particularly high reliance on state aid by school districts reflects a

Year

States

1962 Federal

22.8% State

Total a

24.0

1972 Federal

27.2 State

Total a

28.4

1977 Federal

27.1 State

Totala

28.8

1982 Federal

24.0

State -

Totala

25.1

1987

Federal

22.8

State -

Totala

24.4

1992

Federal

26.1

State -

Totala

27.9

1997

Federal

26.5

State -

Totala

28.3 2002

Federal

29.9 State

Totala

31.6

growing role for state governments in financing local education, a topic discussed more comprehensively in Chapter 19. In fact, the reliance on aid by local govern- meets has not changed dramatically in the past 25 years, except for school districts,

increased from about 45 percent to 57 percent of revenue. whose grants About 75 percent of federal aid to states and localities is directed nominally

toward the three budget categories of education, transportation, and public wel- fare, the last category alone represents more than 57 percent, as shown in Table 9.2 and Figure 9.3. In contrast, education is the dominant category of state aid to localities, accounting for about 64 percent of state aid. As a result of this aid, the federal government finances more than half of public welfare expenditures (61 percent), although state governments finance more than half of expenditures on education (54 percent), even though welfare service is provided directly by states and education by localities.

201

Table 9 . 2

federal and State Aid by Budget Category, 1 997

Federal Aid

Category Federal Aid

State Aid

Category State Aid

Category Aid as Percentage of Category Aid as Percentage

as Percentage of State-Local Category as Percentage of of Local Category

Total Aid Expenditure' Total Aid Expenditure"

14.7% 8.6% 63.7% 53.4% 8.1 24.2 4.1 32.0

50.8 61.0 10.9 742

52 11.3 85 52.4

14.3 8.1 12.8 10.9

100.0 19.6 100.0 35.7

aExpenditure measured as direct general expenditure in category.

SOURCE: U.S. Department of Commerce, Bureau of the Census, Compendium of Government Finances,1997.

PURPOSES OF GRANTS

Traditionally, four potential roles for intergovernmental grants in a federal fiscal system are identified. Grants may be used to correct for externalities-service benefits or tax costs that cross jurisdiction boundaries-and thus can improve the efficiency of fiscal decisions. Grants can be used for explicit redistribution of resources among regions or localities. Grants may be used to substitute one tax structure for another, for instance to take advantage of scale economies in tax collection. Finally, grants have been considered as a macroeconomic stabilizing mechanism for the subnational government sector.

Recall from Chapter 2 that interjurisdictional externalities, or spillovers, can cause service decisions by individual subnational governments to be inefficient from society's viewpoint. If nonresidents benefit from a state or local service, but those nonresident benefits are not considered in the decision about the amount of

Education Highways Public Welfare Health and Hospitals

Other Total

rrE rt NINE • INTERGOVERNMENTAL GRANTS

PART III • PROVISION OF STATE AND LOCAL GOODS AND SERVICES

202

Figure 9.3

Distribution of

federal grants by

program area

Education $29.2 billion

7.6%

Other agencies $35.1 billion

9.1%

Transportation $38.9 billion

10.1%

Housing and urban development

$39.4 billion 10.2%

Agriculture $22.1 billion 5.7%

Health and human services $221.0 billion 57.3%

ail Office of special education programs 58.7 billion

2.3%

Child nutrition programs

$10.4 billion 2.7%

Temporary assistance for needy families

$19.1 billion 5.0%

Housing certificate program

$20.9 billion 5.4%

Highway trust fund $28.6 billion

7.4%

Education - title 1 programs $8.5 billion 2.2%

Children and family services (Head Start) $8.1 billion 2.1%

Centers for Medicare and Medicaid services $164.3 billion 42.6%

Other programs $117.1 billion 30.3%

CHAPTER NINE ■ INTERGOVERNMENTAL GRANTS

Recall from Chapter 5 that individual migration among local communities also may involve a type of externality, if that migration imposes costs on the other res- idents. Individuals may move to avoid subnational taxes or to gain services. If the new residents pay less than the average cost of services they consume, however, existing residents face either receiving less service with the game amount of taxes or paying higher taxes to maintain the services. The potential migrants have no incentive to include those costs imposed on other residents in their decision about whether to relocate, so the distribution of population among localities may become inefficient. Again, intergovernmental grants may be used to resolve this difficulty. Grants to high-tax or low-service localities may forestall some of the migration in search of lower taxes or more services and contribute to a more efficient structure of local government.

Intergovernmental grants effectively substitute the granting government's tax revenue for that of the recipient government. If the taxes used by the granting gov- ernment are more efficient than the ones they replace, this tax substitution is another way that grants may improve the efficiency of the federal system. Because mobility is so much greater among subnational jurisdictions than among nations, a tax levied nationally may generate fewer inefficiencies than a set of similar sub- national taxes. The revenue can be generated nationally but spent locally, with a system of intergovernmental grants. This is at least part of the rationale for revenue-sharing programs.

Also, intergovernmental grants sometimes are suggested as a method of explicit income redistribution for equity reasons. Taxes collected by the federal government or a state may be used to fund grants to lower-level governments that are allocated inversely proportional to income or property value, resulting in an implicit transfer from governments in higher-income jurisdictions to governments in lower-income jurisdictions. The effects of this type of income redistribution are not always clear, however, because jurisdictions seldom are completely homogeneous in income and because the local government determines how the grant funds are to be spent. Even jurisdictions that are low-income on average may have high-income residents, in some cases, a substantial number. If the objective is to assist low-income individu- als and families, it seems preferable in most cases to give grants directly to those individuals, rather than the state or local government where they reside.

203

Note: Total federal aid to states and local governments in fiscal year 2003 was $385.7 billion (100.0%).

SOURCE: U.S. Census Bureau, Federal aid to states for fiscal year 2003. TYPES OF GRANTS

the service to provide, social marginal benefits will be underestimated and too little of the service will be provided. In such a case, an intergovernmental grant can be used to induce the subnational government to provide more of that spe- cific service, as efficiency requires. Moreover, because the grant funds are gener- ated from taxes collected by the granting government, those nonresidents who benefit from the service end up paying for part of the service through their state or federal taxes.

As depicted in Figure 9.4, intergovernmental grants usually are characterized by four factors: (1) whether use of the grant is intended for a specific service or may be used generally, (2) whether grants automatically are allocated by a formula or require an application associated with a specific project, (3) whether the grant funds must be matched by recipient government funds, and (4) whether the poten- tial size of the grant is limited.

Specific, or categorical, grants are the dominant type, both by number and amount of funds, offered by the federal and state governments. As shown in

Close-ended (grant amount

limited)

Open-ended (no limit on

grant amount)

Figure 9.y

204

Types of

intergovernmental

grants

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

Conditions on use:

(no use restrictions) General grants

Allocation method: Formula Project Formula

(no spending

Lump-sum Matching

required)

(no spending Lump-sum

required)

Limit on grant size:

Table 9.3, in 1993, the federal government had 578 different categorical grant programs, representing more than 97 percent of the number of federal grant pro- grams and more than 88 percent of federal aid dollars. The dominant state program provides specific grants for local education.

If the amount of these specific grants does not change as a recipient government changes its taxes or expenditures, then they are called

lump-sum, or nonmatching grants. The amount of the grant cannot be altered by fiscal decisions of the recipi-

ent government. In 1993, 280, or about 48 percent, of federal categorical grants were nonmatching (see Table 9.3). Matching grants,

on the other hand, require recipient government taxes or spending, with the size of the grant depending on the amount of those taxes or that spending. Typically, a specific matching aid pro- gram offers to match each dollar of recipient tax or expenditure on that specific ser-vice with R

grant dollars, intended to be spent on that service. R is called the matching rate. If R = 1,

then each local dollar generates $1.00 in grant money, so that the grant finances half of the expenditure. If

R = .5, then each local dollar gen- erates $.50 in grant funds, and the grant finances one-third of the expenditure ($.50/$1.50). Generally, then, the share financed by the grant (denoted by

M), is M R/(1 + R)

For predicting the effects of matching grants, note that through this matching rate, the grant reduces the price of additional amounts of the aided service to the recipient government. If R = 1,

the grant finances one-half of expenditures, so the cost in local taxes of increasing spending by $1 is only $.50. In general, the local tax price (denoted by P)

of an additional dollar of service (the local marginal cost) is

P 1 — M

= 1 — [R/(1 + R)] = 1/(1 + R)

Categorical grants (use intended for specific activity)

Matching:

Revenue sharing

(tax-effort matters— variable match)

ER N INE ■ INTERGOVERNMENTAL GRANTS

deral Grants, by Type, 1993 (Billions ob Dollars)

comm unity

purpose'

lock grants (for health [5], neral

atyo sdpeovretlaorpi omuein21 [s20] ,c i a i

muntri grou nd services [2], income security [2], and one each for education and employment and training)

Categorical grants Matching

Formula Open-Ended

Project Nonmatching

Formula Open-Ended

Project

Total

Amount

$ 2.4

21.8

206.4

182.2

Percentage of Amount

1.1

10.6

100.0

88.3

Number

578 298

87 12

211 280

72 5

208

593

15

Percentage of Number

2.5

100.0

48.4"

97.5 51.6 b

."General-purpose" grants include a few small payments by federal departments and payments to Puerto Rico and the District of Columbia.

"Percentage of categorical grants only.

SOURCE: ACIR, (January 1994).

If R = 1, each additional dollar of service costs local residents $.50 in local taxes. If R = .5, the local tax price of $1's worth of additional service is $.67. If R = .25, the local tax price is $.80; local residents pay $.80 for each additional dollar of expenditure on the specific aided service.

Both matching and nonmatching categorical grants may be allocated either by for- mula or a project-by-project basis and may be either open-ended (no limit on the grant amount) or closed-ended (the grant amount is somewhat limited because the funds appropriated for the grant program are fixed). Project categorical grants have outnumbered formula grants by more than two to one, while less than 3 percent of federal categorical grants have been open-ended. The class of open-ended, formula, nonmatching, categorical grants (of which there are very few) should be clarified. In these cases, the formula allocating subnational government grants implies a fixed payment for factors outside of the recipient government's control, such as population or population characteristics, but there is no limit on the amount of aid. Programs in this class include unemployment compensation and some child nutrition grants.

General grants, those without use restrictions (or with very loose restrictions), are rare among federal government grants, although somewhat more common among state grants. These grants, which are sometimes said to provide general fis- cal assistance, almost always are allocated by formula. If the formula includes fac- tors outside of the government's direct control, such as population or per-capita income, the grant is a pure lump sum to the government. On the other hand, if the

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

Figure 9.5

Tax price

($)

1

.67 JJ

Di (Lump-sum grant = Go)

Do (No lump-sum grant)

E0 2 Government expenditure

EI

Income and price

effects of a grant

causes the demand curve for government services to shift out (assuming that gov- ernment services are normal goods, as supported by empirical evidence). With the marginal cost of an additional dollar of expenditure remaining at $1, desired expen- diture increases from E0 to E l . On the other hand, a matching grant reduces the marginal cost (or price) of additional expenditure, which causes an increase in the amount of government service demanded, for instance from E0 to E2. In economic parlance, lump-sum (nonmatching) grants increase demand via an income effect, whereas matching grants increase the desired amount of service due to a price effect. Given the characteristics of the grant program and the local political choice system, the economic effects of the grant can be predicted. Several general results follow.

Matching Grants Are More Stimulative Than Lump-Sum Grants Perhaps the most fundamental result of microeconomics is that a decrease in price has a greater effect on consumption than an increase in income, even if that increase is large enough to give a consumer the same choices as the price decrease. When the price of a product decreases, whether for hamburgers or education, con- sumers are influenced by two separate factors. The product whose price has fallen now is relatively less expensive compared to other goods than before the price change, and the consumer's purchasing power has increased—even with constant income, more of all goods can be afforded. The first is called the substitution, or price, effect because it is an incentive for consumers to substitute more of the now relatively less expensive commodity. The second is the income effect. For normal goods, both of these influences are an incentive for consumers to consume more of the product whose price has decreased.

CHAPTER NINE ■ INTERGOVERNMENTAL GRANTS

2 06 207

formula includes factors controlled by the recipient government, such as tax col- lections or tax effort, then the amount of the grant can be altered by recipient gov- ernment decisions. This method, used for the federal and some state revenue- sharing grants, creates a type of matching grant, although the total amount of grant dollars is fixed and the matching rate varies, as discussed later in this chapter. Note that matching, open-ended, general-purpose grants are not a good idea because by redefining all consumption as part of government, all of consumption could be matched. For instance, if residents of a city agreed to buy all food and clothing through the city government, all those expenditures would be matched. It obviously is impossible for this to happen generally.

The best-known general-purpose grant was the U.S. General Revenue Sharing Program, begun in 1972, which initially provided grants totaling about $6 billion annually to state-local governments. The funds first were divided among the states by a formula that included population, per-capita income, and tax effort, with one- third of a state's funds allocated to the state government and the remaining two- thirds distributed to local governments in that state, again by formula. The size of the grant fund was increased slightly in 1976, although states were removed from receiving revenue-sharing grants and the fund decreased proportionately in 1984. The federal revenue sharing program for local governments expired in 1987.

There is also a class identified as block grants. Block grants are specific grants in categories that are broadly or loosely defined. For instance, two separate block grants are available for community development. Correspondingly, there is a long list of approved activities that can be financed with these funds in that general category. The number and size of block grants has been growing in recent years as individual categorical grants have been combined into new block grants. The idea is that these fall in some intermediate area between narrowly defined categorical grants and grants with no use restrictions at all. As you will see later in this chapter, in most cases, these block grants effectively are general grants because the categories are broad enough to allow most recipient governments leeway for reallocating other funds.

ECONOMIC EFFECTS: THEORY

Intergovernmental grants may affect recipient government fiscal decisions either by increasing the resources available to provide government services, called an income effect, or by increasing resources and reducing the marginal costs of addi- tional services, called a price effect. Either effect may influence the amount of gov- ernment service demanded, although in different ways. In taking this approach to analyzing intergovernmental grants, economists retain the notion of individual demands for government services, as discussed in Chapter 4, which must be coor- dinated by a political choice system. If political decisions are made by voting, then the effect of the grant on a government's decisions is determined by the effect of the grant on the decisive voter.

Accordingly, most economic analyses of the expected effects of intergovernmen- tal grants start with the effects of the grants on individual demands, as shown in Fig- ure 9.5. An increase in available resources, which arises from a lump-sum grant,

208

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES clIAPTER NINE ■ INTERGOVERNMENTAL GRANTS

209

Table 9.4 When consumers receive an increase in income, purchasing power rises, but the relative price or cost of different products does not change. Therefore, if the income effect that arises from a price decrease is the same magnitude as the income effect from an increase in income, the price decrease should affect consumption more. The income effects are the same, but the price decrease has an additional substitu- tion effect. In essence, price changes are expected to stimulate greater changes in consumption than equivalent changes in income because price changes alter pur- chasing power and relative costs, whereas income changes alter only purchasing power (and the two changes in purchasing power are the same size).

The implication of this microeconomic principle is that an open-ended matching grant is expected to increase government expenditure on the aided service by a greater amount than an "equal size" lump-sum grant, where "equal-size" is de- fined to mean a lump-sum grant large enough to allow the government the same expenditure as selected with the matching grant. Although the government could select the same expenditure in both cases, it does not because of the price incentive. The change depicted earlier in Figure 9.5 represents this principle. A matching grant that provides $.50 for each $1 of locally financed expenditure reduces the local tax price per dollar of expenditure to $.67, thus inducing an increase in gov- ernment expenditure on the specific service from E 0 to E2. If a lump-sum grant equal to G o were offered instead, which would be large enough to allow the recip- ient government to select expenditure E2, the theory argues that the actual expen- diture selected would be smaller, for instance, equal to E l .

The same principle is shown by the numerical illustration in Table 9.4. Assum- ing initial spending and taxes of $1,000 per capita and a price elasticity of demand for government expenditure equal to -.5, a matching grant providing $.50 for each $1 of local tax reduces the tax price to $.67, a 33 percent decrease, and induces a 16.5 percent increase in spending to $1,165.00. As a result, the jurisdiction receives a matching grant of $388.33 (one-third of total spending). If this jurisdiction received a lump-sum grant equal to $388.33 per capita and assuming per-capita income of $5,000 and an income elasticity of 0.5, income rises by 7.76 percent and spending rises by 3.88 percent to $1,038.80. The matching grant has stimulated a greater increase and level of spending than the equal-size lump-sum grant.

This analysis applies directly to open-ended matching grants but must be mod- ified for closed-ended matching grants. Suppose, for example, that a matching grant is offered of $.50 for each $1 of locally financed expenditure up to a maximum local expenditure of $1,000 per capita. The maximum grant is $500 per capita. The local tax price is $.67 as long as local per-capita expenditure is less than $1,000; above $1,000, the local tax price is $1. In other words, this is initially a matching grant for recipient governments that spend less than $1,000 per capita before the grant pro- gram begins, but it is a lump-sum grant for governments that spend $1,000 per capita or more. Equivalently, this is a matching grant for governments that spend less than $1,500 per capita, including the grant. For instance, a government spend- ing $1,350 per capita on the specific aided function (composed of $900 in local money and $450 of grant) can increase per-capita expenditure by $1 with an extra $.67 of local money. When total per-capita expenditure reaches $1,500, the grant is at its maximum and is, therefore, a lump-sum grant.

expenditure Meets ofj Matching and Lump-Sum Grants

Initial Fiscal Circumstances

Per-capita expenditure Per-capita local tax Price elasticity of demand Income elasticity of demand Per-capita income

Grant Conditions and Effects

Matching Grants

Matching rate

0.50 ($.50 for each $1.00 of each tax)

$0.67 Tax price with grant ($1.00/$1.00 + $.50)

Percentage decrease 33% in price

$1165.00 38833

16.5%

Percentage increase in per-capita expenditure Per-capita expenditure with grant Per-capita grant Per-capita local tax

776.67 Increase in per-capita expenditure

165.00 Decrease in local tax 223.33 Sum = grant amount 388.33

The closed-ended nature of the grant complicates the analysis because (1) it is not possible to determine whether the grant is effectively matching or lump sum with- out knowing the recipient government's position, and (2) a recipient government's reaction to the grant can move its per-capita expenditure across the boundary, transforming an apparent matching grant into a lump-sum grant, or vice versa. For governments "near" the expenditure cap on the grant, the full price effect of the grant may never apply. One expects, therefore, that closed-ended matching grants are more stimulative, in aggregate, than pure lump-sum grants (because some gov- ernments feel some price effect), but less stimulative than open-ended matching grants (because some governments reach the maximum).

Matching Grants Provide Tax Relief,. The preceding analysis argues that matching grants induce an increase in spending on the aided category, but the increase is not as large as the grant. As a result, the matching grant also can increase government spending in other budget categories or allow for local tax relief. As long as the demand for government service is price inelastic, a matching grant increases expenditure by less than the amount of the grant, thus freeing local funds to be spent in other ways. Because the evidence, reported in Chapter 4, shows that demand for most state-local services is price

$1,000 $1,000

-0.5 0.5

$5,000

Percentage increase in per-capita income

7.76% ($388.33/$5000)

3.88% $1038.80

38833 650.47

38.80 349.53 38833

Lump-Sum Grants Per-capita grant amount $388.33

Figure 9.6

Comparison of

alternative lump-

sum grants

Expenditure on other goods

E EG E2 E3 E4 V W Expenditure on government service

ZY= Grant

ZX= Grant/Tax price

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES CHAPTER NINE ■ INTERGOVERNMENTAL GRANTS

210 211

inelastic, matching grants are expected to be used for tax relief in part. The expen- diture and tax effects of matching grants are demonstrated numerically in Table 9.4.

Focusing again on the illustration in Table 9.4, the local tax price falls from $1 to $.67, a decrease of 33 percent. If the price elasticity of demand for the aided service is less than one (inelastic), then expenditures will increase by less than 33 percent and local taxes can decline. In the illustration, spending rises by 16.5 percent to $1,165.00, to be financed by $776.67 of local money and $388.33 of grant money. The matching grant increases total expenditure but decreases the amount of local funds spent on the category by $223.33. This $223.33 can be spent by the govern- ment on other services or on local tax relief.

If demand is price inelastic, matching grants do stimulate increases in total expenditure but do not stimulate increases in locally raised money spent on the service. This led to some confusion as to whether matching grants are stimulative, the confusion resulting from just what "stimulative" means.

Specific Lump-Sum Grants May Be No Different Than General Grants

A lump-sum grant of $G that is restricted for use in a specific category may be no different, from the viewpoint of the recipient government, than a grant of $G with no use restrictions. That is, the two grants may have the same effect on a recipient government's fiscal behavior. This issue depends on whether the government can and does reallocate locally raised funds from the specific budget category to others because of the grant.

The possibilities are depicted in Figure 9.6, which shows the budget options for a community (or individual) between government expenditures on the aided

category and expenditures on all other (government and private) goods. With no grant, this community can spend a maximum of $Z on other goods or a maximum $17 on the specific service or any combination on the budget line between those two points. A lump-sum grant expands the set of affordable options, that is, it shifts the budget line out. A general lump-sum grant equal to ZY shifts the bud- get line to YW; the government receives ZY dollars that can be spent on anything, including entirely on "Other Goods." A lump-sum grant of the same size that must be spent on the aided category shifts the budget line to ZXW; the recipient government must buy ZX units of the aided service, the amount that can be pur- chased using all the grant funds. Thus, all the grant funds are spent on the intended service.

Two implications follow. First, the restriction on use of the grant "matters" to the recipient only if intended expenditures on the aided category are less than what the grant will buy, that is, less than EG. If the recipient government would have spent more than EG anyway, local funds equal to the amount of the grant can be shifted to other uses. Local funds are fungible within the entire budget. Put another way, beyond EG, the budget choices from the two grant programs are identical. Second, a lump-sum categorical grant does not guarantee that expenditures on the aided cat- egory will increase by the full amount of the grant. A government initially spending E2 on the specific service already spends an amount equal to the grant. Rather than increasing to E4, the change from increasing expenditures by the full amount of the grant, the government more likely would increase expenditures to some intermedi- ate level like £3, freeing funds for increased spending in other areas as well.

In the illustration reported in Table 9.4, a lump-sum grant equal to $388.33 per capita is provided. Suppose that per-capita income is $5,000 and the income elas- ticity of demand is .5. The per-capita grant of $388.33 increases income by about 7.76 percent, causing an expenditure increase of 3.88 percent. As a result of the grant, per-capita expenditure increases by $38.80 to $1,038.80, which is financed with $650.47 of locally raised money and the $388.33 grant. 3 Accordingly, the amount of local funds spent on the category decreases by $349.53, which can be spent on other services or tax relief.

Students often are experts on fungibility. Suppose your parents visit you at school and as they are leaving give you a gift of $20, which they insist must be spent on pizza. Even if you always obey your parents, does this mean you will spend $20 more on pizza this week than you usually do? Not necessarily. If you normally spend $25 per week on pizza, you might increase your pizza consumption to $30, including the $20 gift, and shift your own $15 you would have spent on pizza to some other necessity, perhaps books. You satisfied the restriction without having to increase consumption by the amount of the gift. In effect, you behaved in the same way as you would have if the gift came with no use limitation.

The potential for specific-purpose, lump-sum grant funds to be shifted to other uses in this manner has led some to consider other types of use restrictions,

'Another way to think of the result that arises from this income level and the income elasticity of demand is that $.10 of each additional grant dollar is used to increase government expenditure, so the lump-sum grant of $388.33 increases spending by about $38.83.

Table 9.5

[sample Revenue Sharing Program

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

212

particularly requiring maintenance of local effort. This restriction requires not only that the grant funds be spent on the aided category, but also that local funds spent on the category not be reduced. Even this restriction may not be as severe as it seems, however, because expenditure normally would increase annually without the grant. If a government spends $100 on a specific service in one year and plans to spend $110 in the following year, a $10 lump-sum grant with an effort mainte- nance restriction is the same as a $10 grant with no restriction. The grant can be spent on the specified service, and the additional $10 the government would have spent on that service can be reallocated to other uses. In general, the effort mainte- nance restriction is binding only if the grant is larger than the increase in expendi- ture that would be selected without the grant (which is not observed).

Tax Enroll Grants Are Matching

One common factor in the allocation formula for revenue-sharing grants, once used for the U.S. Federal Revenue Sharing Program and still for about one-quarter of state revenue-sharing funds, is tax effort. Tax effort is usually measured either by tax revenue as a fraction of income or, for many local governments, property tax as a fraction of taxable value. In these revenue-sharing programs, a higher tax effort generates a larger grant, given no change in any other allocation factor. A high tax effort can reflect either a great demand for government service in a jurisdiction, a relatively low tax base, or a high production cost for government service. Because a subnational government chooses its tax effort, those recipient governments can affect the size of the revenue-sharing grant (similar to matching grants).

The operation of a representative state revenue-sharing program is demon- strated in Table 9.5, simplified with two equal-size recipient local governments. The state revenue-sharing program divides a fixed amount of state tax collections ($100) between the two localities based on population (POP,) and tax effort, here defined as the effective property tax rate (T,/V,). Both jurisdictions initially collect equal property taxes, but because jurisdiction A's property value is lower, its tax effort is twice as great as jurisdiction B's. Because they have equal populations, A receives 66.7 percent ($66.70) of the revenue-sharing funds, and B receives the remaining 33.3 percent ($33.30).

What happens if one of these governments (B) increases property taxes by 20 percent to $600 while A holds taxes constant? Jurisdiction B's relative tax effort rises, and therefore its share of the revenue-sharing funds also rises. In this exam- ple, because B gains $4.20 in revenue-sharing funds from the $100 increase in taxes, the new local tax price is $.96. Jurisdiction A loses the $4.20 of revenue- sharing funds, a 6.3-percent decrease, even though it made no fiscal changes.

Several implications follow. A recipient jurisdiction can increase its revenue- sharing grant by increasing taxes at a greater rate than its competitor jurisdictions. Even if a jurisdiction does not seek a larger revenue sharing-grant, it must increase taxes just to avoid losing grant funds if any other recipient jurisdiction raises its taxes. Each jurisdiction is in competition with all others for the limited revenue- sharing funds. Because all jurisdictions face these same opportunities and each is

CHAPTER NINE ■ INTERGOVERNMENTAL GRANTS

Feature Population Property tax Taxable value Effective tax rate—tax effort Relative tax effort

TN(

Grant share RTE, x POP;

(RTE; x POP;) Grant (fund = $100)

New property tax New relative tax effort New grant share New grant amount Change in grant Percentage change in grant Price of tax increase

uncertain about the behavior of its competitors, a general incentive exists for an increase in government expenditures. This program is different from a standard open-ended matching grant because the total amount of grant funds is fixed and because the rate at which local taxes are matched by increased grants changes as all the recipient jurisdictions react to the grant. As one special case, if all the recip- ient jurisdictions increase taxes at the same rate, no one's revenue-sharing grant changes, although all increase government spending.

HEALTH CARE AND THE CHANGING COMPOSITION OF GRANTS

Federal grants to state-local governments federal spending. State-local governments

increased at relatively high rates since the went from receiving about $.22 in grants for

early 1990s. Federal grants in the years 1992- each dollar of own-source revenue to about

2002 increased annually on average of more $.27 per dollar of revenue by 2002.

than 10 percent and thus went from repre- One might think that the relatively rapid

senting about 2.8 percent of GDP and about growth of federal aid might ease the fiscal

14 percent of federal government outlays to problems of states and localities, but instead,

3.4 percent of GDP and nearly 18 percent of much of the growth in grants seems to have

Jurisdiction A

Jurisdiction B 50

50 $500

$500 $5,000

$10,000 10%

5%

150

0.75

66.7%

33.3%

$66.70 $33.30

Effect of Property Tax Change

$500

$600

1.36

0.82

62.5%

37.5%

$6250

$37.50

-$ 4.20

+$ 4.20

-6.3%

+ 12.6%

na

$0.96

213

Application 9

Transportation

50

Education

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES HAPTER NINE ■ INTERGOVERNMENTAL GRANTS

214 215

Application 9.1 — Health Care and the Changing Composition of Grants Application 9.1 — Health Care and the Changing Composition of Grants

by 2002.1n contrast, grants for road building

and maintenance from the Highway Trust

Fund grew by only about $15 billion,

although grants for' cash assistance payments

for the poor (TANF) actually declined.

Because federal grants to states for Medic-

aid are open-ended matching grants, the

substantial increases in the amount of

those grants resulted from increases in state

been caused by rapid increases in costs faced

by states, particularly for health care. About

half (actually 46 percent) of the increased fed-

eral grant amount since 1992 was an increase

in grants to finance Medicaid, which pays

health care expenses for low-income individ-

uals and families. Grants for state Medicaid

expenditures rose by about 123 percent from

1992 to 2002, whereas the amounts for all

other federal grants rose about 88 percent. As

a consequence of the relatively larger growth

of Medicaid grants compared to others, fed-

eral grants for health care (almost all of which

is for Medicaid) represented nearly 41 per-

cent of total grant dollars in 1992, but 56 per-

cent in 2002.

The relatively rapid growth of grants for

health services (and especially Medicaid) is

reflected in Figures 9.7 and 9.8. Figure 9.7

shows the dramatic growth in grants from the

Department of Health and Human Services

since 1997 from about $125 billion to almost

$220 billion. When the analysis is by program

(rather than granting agency), as shown in

Figure 9.8, the story is the same. Grants for

medical assistance programs grew from

about $70 billion in 1992 to nearly $180 billion

Figure 9. 7

30

40

35 -

Health and human services

Federal aid

state and liic governrnen

annual ant

by major

area fiscal

1981-2003:

Federal aid to state

and local govern-

ments, annual

amounts by

Major agency:

fiscal years

1981-2003 150

Housing and urban development

200

175

125

C 0

0 0 c 100

.0 c 20

Other agencies

Agriculture

.0 C

75

Medical assistance programs

15

Highway trust fund 10

25

5 , I I . I • 41 0 40 4,1 4 41• 4 4P ciN 4;5 Ci „,4 4 „ C) C) NC) C) NC) C) O) O) O) NC) C) ef)

SOURCE: U.S. Census Bureau, Federal aid to states for fiscal year 2003.

Family support payments (TANF)

0 I I. . :

.1,1' 93) ' ,1? it) c/;\ cP c;\ 41' 07' c)°` 01" ci" 40 o° ,;1 4F) ,■5 4F) 4?) 4?) F F) 4F) ?) ■5 ;) E)

DOL DOS

SOURCE: U.S. Census Bureau, Federal aid to states for fiscal year 2003.

Application 9.1 — Health Care and the Changing Composition of Grants

spending on Medicaid. State spending on

Medicaid can increase either because states

expanded their programs by easing eligibility

or raising benefits, or because the costs of

providing a given set of health-care services

to the eligible population increased substan-

tially. Not surprisingly, the correct explana-

tion is mostly the latter one. The substantial

increase in the demand for and costs of pro-

viding health care greatly increased state

expenditures for Medicaid. Only a portion of

the increased state expenditures are paid by

additional federal grant amounts. As a result,

rather than easing state budget problems,

these increased federal grants arose from the

increased demand for state services, which

caused additional budget pressures for both

states and the federal government. For more

on the issues concerning financing of Medic-

aid, see Chapter 21.

216

ECONOMIC EFFECTS: EVIDENCE

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES cRAPTER NINE ■ INTERGOVERNMENTAL GRANTS

It is difficult and somewhat dangerous to make generalizations about the estimated effects of intergovernmental grants because there seems to be substantial varia- tion in how different governments respond to different grants and because the results of different economic studies often vary greatly even for the same grant program. Nevertheless, some conclusions are broadly supported about the general direction and relative magnitude of effects caused by different grants.

First, open-ended, categorical matching grants seem to increase expenditures on the aided category by a larger amount than equal-size specific lump-sum grants, as predicted by theory. Because the estimated price elasticities for most subnational government services are less than one (in absolute value), the expenditure increase from a matching grant is smaller than the grant, allowing funds to be diverted to other expenditure categories or to tax relief. The numerical example in Table 9.4 generally is representative, therefore, of the statistical evidence.

Although open-ended matching grants are not the most common type of feder- al grant, as previously noted, they have been used for two well-known programs: Aid to Families with Dependent Children (AFDC) until 1996 and Medicaid. Robert Moffitt's (1984) analysis of state government responses to federal AFDC grants supports the general conclusions noted previously. Through grants to states, the federal government paid a percentage of state AFDC benefits (the exact percentage differs by state). Using 1970 data, Moffitt estimated that the elasticity of a state's per-capita AFDC benefit with respect to the national subsidy rate is .15; a 10 per- cent increase in the subsidy rate increases per-capita benefits by 1.5 percent. At that time, the average per-capita AFDC benefit was $45, with the federal government paying about 60 percent of the marginal cost (an additional $1 of benefit costs the state $.40). If the subsidy rate were increased to 70 percent, about a 16 percent increase, the per-capita benefit would increase by about 2.4 percent (16 X .15), or about $1. The average state would have received approximately $1.20 more in per- capita grant, with about $1 going for increased AFDC benefits. More discussion of these types of welfare grants is presented in Chapter 21.

second, there is some evidence that closed-ended categorical matching grants sometimes have greater expenditure effects than open-ended matching grants, which seems contrary to theory. Closed- and open-ended grants are not used for the same services, however, so the different expenditure effects most likely result from differences in demand for the services. For instance, the closed-ended categoricals, which are the most common type of federal grant, may be used for services that state-local governments were not substantially providing or may include effort maintenance provisions. In either case, the opportunity to use grant funds to shift resources to other budget categories is limited, forcing a larger increase in spend- ing on the aided category. Also, the demands for the services aided by closed-ended grants might be more price elastic than those for which open-ended grants are used. 4

Third, lump-sum grants also cause an increase in government expenditures, which seems in most cases to be smaller than the grant. The estimated expenditure effects of lump-sum grants vary widely, however, from an expenditure increase of $.20 up to $1 per dollar of grant received. Again, two reasons for this difference are differences in initial spending on the category by subnational governments and different use restrictions among the grants. The majority of the estimates fall in the range of a $.25 to $.50 increase in expenditure per dollar of grant. If those results are representative, then $1 of lump-sum grant provides between $.50 and $.75 for expenditures in other budget areas or for local tax relief.

The evidence that a substantial portion of both matching grants and lump-sum grants effectively is diverted to uses other than those nominally intended raises the issue of which other budget categories benefit. This "leakage" of grant funds may occur both among different services and different local governments, which over- lap in tax authority. As an example of the latter, aid to municipalities is expected to increase municipal expenditures and decrease local municipal taxes. The lower municipal taxes may, therefore, allow local school districts to also increase expen- ditures (by reducing opposition to increased local school taxes). In fact, there is evi- dence of just this sort of cross-government general-equilibrium effect; aid to either municipalities or independent school districts appears to cause increased spending by both.

The possibility of grant substitution among different budget categories for a sin- gle government was examined in detail by Steven Craig and Robert Inman (1985), who studied state government expenditure responses to federal welfare and edu- cation grants. Craig and Inman concluded that although federal welfare and edu- cation grants to states do increase state expenditures in those categories, both influ- ence expenditures in other areas by a larger amount. For instance, they estimate that an additional $1.21 from open-ended federal welfare grants to states would generate $.34 more in welfare spending, $54 less in state education expenditures, $.63 less in state taxes, and thus $.78 more on other state services (1.21 - .34 + .54 - .63 = .78). Similarly, they found that $1 of additional lump-sum federal

'It does appear that demand for state-local welfare expenditures is less price elastic than the demand for state-local services generally.

where

y, = income for person i .,.-- private consumption by person i

t, = the local tax share for person i

T = total tax collected by person i's local government

E = expenditures by person i's government

G = the lump-sum grant to person i's local government

The left-hand side of the budget equation represents the resources available to be spent on either private consumption or government services. The individual's price for government services is the tax share, ti . An individual voter's implicit share of lump-sum grants received by the government is the voter's tax share mul- tiplied by the amount of the grant; this is the amount of local taxes the individual would have to pay to generate the same amount of revenue as the grant. Equiva- lently, if all the grant were used to lower local taxes, this represents the tax savings to that voter. With this view, it should not matter whether resources arise from an increase in Yi or an increase in G; because both expand the individual's budget and should increase demand for normal goods. 5 The same idea is illustrated by Figure 9.6. An increase of ZY in private income shifts the budget line in exactly the same way as a lump-sum grant equal to ZX.

The key to the argument, of course, is whether individuals have the option or desire to convert lump-sum grants received by the government into private income through tax reductions. If individuals suffer from some type of fiscal illu- sion or if budget-maximizing, monopoly government officials create such an illu- sion, then the grant funds may be treated differently than income. One possible type of illusion occurs because lump-sum grants reduce the average cost to residents of recipient government spending. A jurisdiction that spends $100 per capita and receives a $30 per-capita grant pays only 70 percent of the cost, on average. If individuals believe that this average cost is the price, then it appears that the lump-sum grant has reduced the price of government service similar to a match- ing grant. As a result, the expenditure effect would be greater than from the income effect alone. This is an illusion because the grant is a lump sum (constant). An increase in spending of $1 would cost the local jurisdiction $1; the marginal cost has not been reduced.

From the budget equation, a $1 increase in Yi should be precisely equivalent to an increase of lit, in G.

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES copTER NINE • INTERGOVERNMENTAL GRANTS

219

education aid to states increases state education expenditure by $.43, increases state welfare expenditures by $.23 (only $.09 of which is state money due to match- ing federal welfare aid), decreases state taxes by $.39, and thus allows $.09 to be spent on other state services. Although the specific magnitude of these estimates surely is not precise, it seems clear that intergovernmental grants do have some substantial unintended or unexpected effects on recipient government budgets.

Finally, evidence suggests that an additional $1 of lump-sum grant money has a greater government expenditure effect than a $1 increase in residents' incomes. The results of a number of studies show that although $1 of increased income is expected to increase subnational government expenditure by about $.05 to $.10, $1 in lump-sum grant money appears to increase expenditure by $.25 to $.50. This result has become known as the flypaper effect, reflecting the notion that money paid to a government tends to "stick" in the public sector. If true, this means that a $1 grant has very different allocation effects than a $1 tax decrease by the grant- ing government (which increases income by $1). These results have generated some controversy about whether they reflect important characteristics of political behavior or are illusory and caused by incorrect or imprecise economic analysis. That debate is presented next.

la Grant Money Different Than Tax Money?

Do increases in lump-sum grants and increases in private personal incomes affect subnational government expenditures equally? If not, why not? These two issues have received increasing amounts of attention as a result of the empirical results mentioned previously. The answers seem to fall into two categories. One posi- tion is that no flypaper effect really exists—that the empirical results arise from incorrect statistical work or misinterpretation of those results. The other position, that the flypaper effect is real, is then divided on the cause—whether it reflects political power and control by government officials or behavior actually desired by voters, who may be misinformed.

First, why would economists think that grants and income should influence expenditures equally, anyway? That view arises from the belief that the public- choice process (voting) works to reflect perfectly the desires of various voters, or at least the decisive voter. The majority-voting/median-voter model so favored by economists is in this class; government selects the expenditures desired by the median voter, and if government does not select the desired expenditures, political competition will arise to move the government in that direction. For an individual voter, increases in income or grants to the voter's government are the same because both increase the resources available for consumption. An indi- vidual can convert grant funds into personal income through decreased local taxes.

The idea can be demonstrated through an individual's budget, which leads to that individual's demand for government services, as presented in Chapter 4. The budget is

Y, = C, + ti(T)

Because local taxes must make up the difference between expenditures and grant

furidS,

• = + ti(E – G)

• = + tiE – tiG

• + tiG = +

Figure 9.9

Price ($)

so(Yo)

5 1 (n)

Increases in income can affect both demand and cost for government service

D i ( n)

WY() )

Government service

E2

220

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES Ap TER NINE ■

INTERGOVERNMENTAL GRANTS

221

The flypaper effect also could result from the nature of the political process rather than incorrect perceptions by voters. By controlling the set of options from which voters choose, budget-maximizing officials may be able to get voters to approve taxes to finance desired expenditures and then also spend the grant funds. The grant funds therefore would cause increased spending rather than tax relief. This will work only if voters do not give grant funds the same careful consideration they do taxes, and political competitors do not arise to give voters a different set of choices.

The competing position holds that the flypaper effect really does not occur, with the apparent evidence caused by statistical and analytical error. One possibility is that in studying grants, analysts may make mistakes in classifying grants as lump sum or matching. Howard Chernick (1979) has argued, for example, that in choos- ing among competing projects applying for closed-ended lump-sum funds, offi- cials of the granting government may favor those projects where the recipient government agrees to spend the largest amount of local funds. This converts a nominally lump-sum grant effectively into a matching one. If an analyst considers the grant a lump sum when it is in fact a matching grant, it is not surprising to find an unexpectedly large expenditure effect. With some 600 different federal grant programs plus state grants to consider, many of these types of errors are possible.

Another possibility, suggested by Bruce Hamilton (1983), is that residents' income may affect the cost of providing government services as well as demand. For instance, it may require less government spending to bring students up to a given test-score level in a higher-income community than a lower-income one, due perhaps to nursery school or other educational services purchased privately by the families. If income does affect cost, then increases in income cannot be compared directly to increases in grants. In Figure 9.9, an increase in income increases the demand for service and reduces the cost of providing that service. The increase in

p from Es to El is due to the income effect on demand, although the

increase from El to E2 reflects the cost reduction. Studies that ignore this possibi-

lity underestimate the expenditure effect of income increases, which can be part of

the reason for the flypaper effect results. AAA-tether the flypaper effect is a political fact or a figment of imprecise analysis

is, as yet, unresolved. In general, those who believe that substantial political corn- petition between potential officials and economic competition among jurisdictions

rate prevalent, tend to believe that the flypaper effect must be small or weak. Those

who believe government officials can maintain monopoly power and manipulate public opinion tend to believe that the flypaper effect is real and strong.

INTERGOVERNMENTAL GRANT POLICY

Economic theory and evidence about the effects of alternative types of intergovern- mental grants lead to three major conclusions about grant policy. First, open-ended categorical matching grants are best if the objective is to increase recipient govern- ment expenditures on a specific function. A matching grant with a matching rate equal to the nonresident share of benefits offsets the effects of interjurisdictional externalities by reducing the local tax price. The lower price induces the increase in expenditures necessary for efficiency. For instance, if the marginal social benefit of additional highway spending is half of the total, a matching grant to states that pays $1 for each $1 of state money reduces the state's cost by half and restores efficiency. Although other grants also could be used to increase expenditures, an open-ended matching grant induces the desired expenditure response with the smallest possible grant; matching grants provide the largest expenditure effect per dollar of grant.

Second, general lump-sum grants are a better mechanism than matching grants to redistribute resources among subnational jurisdictions. No economic reason exists for such grants to go to all jurisdictions; of course, they should be targeted to low- income or high-cost jurisdictions. These grants should be lump sums so as not to alter the relative price of government compared to private consumption. Although substantial tax relief is expected to result from such a program, these grants are not equivalent to federal tax reductions if the flypaper effect results are correct.

Third, categorical lump-sum and closed-ended matching grants should generally be avoided. Closed-ended matching grants become lump-sum grants after the maxi- mum grant is reached, and categorical restrictions do not alter grant effects unless the grant is large compared to recipient government expenditures in the category. Open- ended matching grants are preferred, however, if the objective is to increase expendi- tures or to induce recipient governments to begin spending on a specific function.

As we have seen already, the actual intergovernmental grant system in the United States departs substantially from these rules. Categorical closed-ended grants are the most common form of federal grant (both in number and dollars). When matching grants are used, the matching rates often do not seem to correspond to the share of benefits that go to nonresidents. Revenue-sharing grants, the basic general-purpose grants, were given to all general-purpose local governments and included match- ing-grant effects due to tax effort allocation. The specified categories for block

cii•PTER NINE ■ INTERGOVERNMENTAL GRANTS

223

Figure 9.10

Alaska

Wyoming

New York

New Mexico Vermont

North Dakota

Maine

West Virginia

South Dakota

Rhode Island

Mississippi

Montana

Massachusetts

Kentucky

Louisiana

Arkansas

Tennessee

Pennsylvania

Oklahoma

Oregon " Alabama MfflEMNIE I

U.S. Average

California

Connecticut

Arizona

Delaware

Missouri

Wisconsin _

Washington

Idaho

Hawaii

North Carolina

Nebraska

South Carolina

New Jersey

Ohio

Iowa

Maryland

Michigan

Minnesota

New Hampshire

Illinois

Texas

Georgia

Indiana

Florida _

Kansas M=M1.----i Utah -Ammo!

Colorado

Nevada W=I

Virginia 0.USIN

Medical assistance programs

Family support payments (TANF)

El Highway trust fund ❑ Other programs

Federal aid to state and local governments, per-capita amounts by state, by major program area: fiscal year 2003

500 1,000 1,500 2,000 Dollars

2,500 3,000 3,500 4,000

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

222

grants are so broad that they effectively are general grants. Consequently, "reform- ing" the federal grant system is continually discussed.

One reform option advanced by a number of economists is to substitute open- ended grants for closed-ended or lump-sum grants and to set matching rates to cor- respond to nonresident benefits. Gramlich (1985b) has suggested that in many cases this would require reducing matching rates for current open-ended matching grants. The reduction in many matching rates for those programs would then free up resources that could be used to fund larger grants for those programs where no matching exists currently or where there are low caps on matching provisions. As a result of such a policy change, federal grant programs would become more stimula- tive across a broader set of functional areas. As noted, this policy makes sense if the primary objective of grants is to offset spillovers and establish economic efficiency.

However, research by Robert Inman (1988) suggests that offsetting spillovers may not be an important objective of federal grant policy in practice, even if econ- omists argue it should be. Inman compares the interstate distribution of federal grants to variables that might capture the potential for spillovers—these variables include a measure of out migration by residents, the number of new housing starts (reflecting immigration), and the number of local governments per square mile— and finds either no relationship between grants and these measures or the wrong relationship. He concludes that ". . . the spillover rationale for aid does little to help us understand the actual distribution of federal assistance" (Inman, 1988, p. 49).

In contrast, Inman does find support for the idea that a main purpose of federal grants is to further economic equity, that is to bring about a more equitable distri- bution of resources and thus, perhaps, a more equitable distribution of public goods. After correcting for other factors, he reports that "federal aid is almost always inversely related to the level of state income," and that "almost all federal aid is equalizing" (1988, p. 51). The state-by-state distribution of per-capita federal grants for 2003 is shown in Figure 9.10. Even without holding constant other fac- tors that influence the interstate distribution of grants, a negative correlation remains between per-capita grants and per-capita income, although perhaps a bit weaker than that found by Inman. For instance, the state with the smallest amount of per-capita grants, Virginia, has a per-capita income that was 7 percent above the national average in 2002, while the state with the largest per-capita grant (exclud- ing Alaska) is Wyoming, with a per-capita income about equal to the national aver- age. 6 If federal grants are intended to redistribute resources among states, then it is not surprising that they do not seem related to spillovers. If the intention is to redistribute resources only, then general block grants may make more sense than matching grants.

'The lowest-income state, Mississippi, receives per-capita federal grants of about $1700 (in 2003), whereas the highest- income state, Connecticut, received about $1,300. A regression of the log of per-capita grants (G) on the log of per-capita income (Y) is

InG = 9.06 — 0.26 InY,

(4.06) (-1.13)

suggesting that each 1 percent increase in per-capita income is associated with about a quarter of 1 percent decline in per-capita grants.

Note: Chart excludes separate entry for the District of Columbia, which received $6,449 total per capita federal aid.

SOURCE: U.S. Census Bureau, Federal aid to states for fiscal year 2003.

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES C HAPTER NINE ■ INTERGOVERNMENTAL GRANTS

224 225

Another often-suggested reform is for the federal government to reduce grants to states and localities and simultaneously reduce or eliminate some federal tax. The theory of such "revenue turnbacks," as they are often called, is that states would be free to raise a state tax to replace the federal tax and the federal grants. Of course, states would have the option not to do that as well, essentially provid- ing residents fewer public services but more private consumption. President Reagan offered just such a proposal in 1982 (see Chapter 21 for details).

If the objective of federal grants is to redistribute resources to poorer states, then a simultaneous reduction in federal grants and taxes would not achieve the objective, unless the federal tax was collected disproportionately from poorer states, which seems unlikely. We have seen previously that federal grants go disproportionately to poorer states. Similarly, to the extent that states have responded to the price- incentives of matching grants, state-local spending on the aided categories would fall if the grants are removed, even if the states received the same resources in the form of lower federal taxes. Finally, the cost of collecting taxes might be lower at the federal than state level, and such economies of scale alone might justify a federal grant struc- ture. In short, the idea of "revenue turnbacks" makes sense only if no economic reasons exist for the federal grants in the first place or if these reasons no longer apply.

INTERNATIONAL COMPARISON

Grants in Major Federal Nations Intergovernmental grants are a common feature of almost all nations regardless of their intergovernmental structure, but they are particularly important and poten- tially more complicated in federal nations, where at least three separate levels of government exist. Despite the widespread use of grants, the magnitude, main pur- pose, and structure of intergovernmental grants vary substantially, even among federal nations.

As shown in Table 9.6, among these four nations the magnitude of intergovern- mental grants is much greater in Australia and Canada when compared to Germany

Table .6

Use ot Grants in Federal

Grants Received as a

Nations, 2002

Grants Received as a Nation Level Percentage of GDP Percentage of Revenue Australia State 7.1 49.5

Local 0.4 17.2

Canada State 3.3 16.0 Local 2.8 38.8

Germany State 2.1 17.0 Local 2.5 35.1

United States State 2.8 26.0 Local 3.7 40.1

and the United States. Total spending for grants is a larger share of the economy in those nations, and grants provide a larger share of revenue for state and local

governments there, on average. Grants are especially important for state govern- ments in Australia (more than 49 percent of revenue) and local governments in the United States (40 percent), Canada (nearly 39 percent), and Germany (35 percent). In Australia, states are the main providers of direct services but have limited tax authority; thus, they depend on federal government grants for a large portion of their revenue. In Canada, local governments provide a number of social services that are selected and mandated by the provinces (states), which are then funded by provincial grantS to localities. In the United States, the dominant grants received by local governments are to school districts for K-12 education.

What these data do not show is the difference in the main purposes and struc- ture of grants in these countries. As you have already seen, grants in the United States are mainly narrow categorical grants intended to affect spending in particu- lar service categories. The main purpose of grants in the other three countries, however, is for regional redistribution and equalization, accomplished by broad- purpose, revenue sharing type grants. In Australia, the Australian Grants Commission—an independent authority established by the federal government— recommends a distribution of grants among the states and territories to equalize the states' capability to provide a standard set of services, given their costs and local revenue base. In Canada, the federal government provides general equaliza- tion grants to the states based on the states' capability to generate revenue; a state receives an equalization grant if its per-capita revenue from a fixed average set of tax rates is less than the national average revenue yield.

Despite substantial regional economic differences in the United States, general grants with an explicit equalizing objective have been relatively unimportant. The federal government did operate General Revenue Sharing for a few years in the mid-1970s and early 1980s, but its magnitude was always very small. Thus, ACIR (1981, p. 97) noted that "Fiscal equalization is less accepted as a goal and conse- quently is pursued to a lesser extent in the United States than in any of the other three federal nations. . . ." However, even if regional redistribution or equalization has not been an explicit objective of U.S. grant policy, it is certainly true that redis- tribution has occurred and perhaps was implicitly intended.

Intergovernmental grants, sometimes called grants-in-aid, are transfers of funds from one government to another, most often from a higher-level government in the federal system to a set of lower-level governments.

In 2003, the federal government transferred nearly $386 billion of aid to state-local governments, about $.27 for every $1 raised by state-local governments from their own sources. Similarly, state governments transferred about $355 billion to local governments, or about $.60 cents for every $1 collected by local govern- ments from their own sources. The importance of intergovernmental grants rela- tive to the revenue of recipient governments peaked in the late 1970s; after declin- ing, it has been increasing since.

SUMMARY

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES copTER NINE ■ INTERGOVERNMENTAL GRANTS

227 226

Seventy-four percent of federal aid to states and localities nominally is directed toward the three budget categories of education, highways, and public welfare, the last representing more than 50 percent. In contrast, education is the dominant cat- egory of state aid to localities, accounting for about two-thirds of state aid.

Grants may be used to correct for externalities that arise from the structure of subnational governments and thus can improve the efficiency of fiscal decisions. Grants also can be used for explicit redistribution of resources among regions or localities. Grants also have been considered as a macroeconomic stabilizing mech- anism for the subnational government sector.

An open-ended matching grant is expected to increase government expenditure on the aided service by a greater amount than an equal size lump-sum grant, where "equal size" is defined to mean a lump-sum grant large enough to allow the government the same expenditure as selected with the matching grant. If the de- mand for government service is price inelastic, a matching grant increases expen- diture by less than the amount of the grant, thus freeing local funds to be spent in other ways.

A restriction on the use of a lump-sum grant "matters" to the recipient only if intended expenditures on the aided category are less than what the grant will buy. Effort maintenance restrictions are binding only if the grant is larger than the increase in expenditure that would be selected without the grant (which is not observed).

Lump-sum grants cause an increase in government expenditures, usually in the range of $.25 to $.50 increase in expenditure per dollar of grant. A lump-sum grant of $1 thus provides between $.50 and $.75 for expenditures in other budget areas or for local tax relief.

Economic theory and evidence about the effects of alternative types of intergov- ernmental grants led to three major conclusions about grant policy. A matching grant with a matching rate equal to the nonresident share of benefits is best if the objective is to offset the effects of interjurisdictional externalities. General lump- sum grants are a better mechanism than matching grants to redistribute resources among subnational jurisdictions. Categorical lump-sum and closed-ended match- ing grants generally should be avoided in favor of the other two.

DISCUSSION QUESTIONS

1. Because nonresidents benefit from local government public-safety services, suppose that the federal government offers localities a public-safety grant equal to $1 for $1 of local tax money spent on that service.

a. What is the effect of this grant on the price of public-safety spending to these localities? How might the grant correct for the spillover problem?

b. Suppose that Central City currently levies a property tax for public safety at a rate of $10 per $1,000 of taxable value on a base of $10 million of taxable property. If the price elasticity of demand for public safety in Central City is 0.2, calculate and explain the expected effect of the grant on public-safety spending, public-safety taxes, and tax rates in Central City.

2. Instead of the matching grant in the first problem, suppose Central City received a lump-sum grant of $55,000 that must be spent on public safety. If the total income of Central City residents is $22 million and the income elasticity of demand for public safety is 0.8, what is the expected effect of this grant on public-safety spending and taxes? Why does the matching grant increase spending more than the lump-sum grant?

3. Suppose that Central City received a lump-sum grant of $55,000 with no restrictions as to how that money must be spent. Do you think the effect on public-safety spending would be different from the specific lump-sum grant in problem 2? Why or why not?

4. Periodically, it is proposed that the federal government should reduce its role in intergovernmental fiscal relations by eliminating a number of smaller matching intergovernmental grants and simultaneously reducing federal taxes by an equal amount, particularly any that directly finance these grants. This concept is sometimes referred to as "revenue tumbacks," which is the idea that individuals will retain the resources and the option to tax those resources to continue the programs now financed by the grants. If such a change were made, how do you expect states would respond? Do you expect that state spending on the aided categories could rise or fall if states had to finance that spending from the additional private resources?

SELECTED READING

Break, George. Financing Government Expenditures in a Federal System. Washington, D.C.: The Brookings Institution, 1980. See Chapter 3, "The Economics of Intergovernmental Grants," and Chapter 4, "The U.S. Grant System."

Fisher, Ronald and Leslie Papke. "Local Government Responses to Education Grants." National Tax Journal, March 2000,153-168.

Gramlich, Edward M. "Intergovernmental Grants: A Review of the Empirical Literature." In The Political Economy of Fiscal Federalism, edited by Wallace Oates, 219-39. Lexington, Mass.: Lexington Books, 1977.

APPENDIX

Inditlyerence-Curve Analysis ot Grants One also can demonstrate the effects of different types of grants using the tra- ditional consumer-theory tools of indifference curves and budget lines, contin- uing the presentation in the appendix to Chapter 3. In Figure 9A.1, an individ- ual faces budget constraint Al in choosing between governmentally provided good G and a composite good X, representing consumption on all other goods. The slope of the budget line represents this individual's tax price. At the utility maximizing bundle, this individual consumes G o units of good G and spends X0 dollars on all other goods.