Economic Impact of Government Spending:

A 50-State Analysis

by
Richard Vedder
Ohio University

National Center for Policy Analysis
12770 Coit Rd.,
Suite 800
Dallas, Texas 75251

Executive Summary

The total cost to the nation of excess spending by state and local governments was more than $353 billion by 1990 - an average loss of more than $1,400 for every man, woman and child in the United States.

• Personal income was lowered by almost $292 billion by 1990 - $280 billion of it because of excess compensation to public employees.
• Direct costs in the form of added taxes amounted to another $61 billion, of which $47 billion can be traced to excess compensation.

Only in Massachusetts and Hawaii was state and local government spending growth held at a level that imposed no additional burden on residents. Both states had healthy per capita income growth during the 1980s, and Massachusetts' was the second highest in the nation. By contrast, Alaska and Wyoming, with the highest excess burdens from state and local government spending, both had negative per capita income growth.

Econometric studies cast serious doubt on the benefit of most government spending. They show little relationship between most government spending - including education and highways- and economic growth. Specifically:

• There is a strong negative relationship between spending on public assistance and economic growth.
• The excess pay of public employees is the equivalent of a major income transfer program and has had a particularly debilitating impact on growth.
• Data from a study in Ohio show that spending on education has some positive (but diminishing) effects on learning when the money is spent for actual instruction, but administrative expenditures, which have been growing in relative importance, tend to have negative effects on learning.

. State and local government accounted for almost two-thirds of total government spending at the turn of the century. By 1990, the proportion had fallen to less than half, and part of that was financed through federal revenues.2 Yet state and local governments have expanded enormously, both in actual size and relative to the nation's productive capacity.

"State and local government spending more than doubled during the 1980s."

A State and Local Spending Spree. Although the private sector grew dramatically during the 1980s, state and local governments grew even faster.

• State and local government spending more than doubled during the 1980s, from $434.1 billion in fiscal year 1980 to $975.9 billion in 1990.3
• Adjusting for inflation, state and local government spending in real terms rose by 41.7 percent.4
• Adjusting for population growth, real spending per capita rose 29.1 percent, compared with a 17.8 percent growth in the nation's real per capita gross domestic product.

In other words, spending by state and local governments increased, on the average, more than half again as fast as the nation's output of goods and services.5 [See Figure I.]

Where the Money Went. As Table I shows, spending in some traditional service areas grew far less than overall spending.6

• While overall spending increased by 41.7 percent in real terms, spending on highways grew only 15.6 percent.
• Spending for police and fire protection rose only 15.3 percent.
• Spending for interest on government debt and public welfare, on the other hand, grew far more than the average.
• Spending for interest went up 112.6 percent.
• Public welfare spending rose 53.0 percent.

While these numbers show the increases by spending categories, they combine amounts spent both for real services and for government employees' compensation. Thus they conceal the primary reason why state and local government spending increased faster than per capita income: increased compensation of public employees.

Rents for Public Employees. Rents obtained by public employees played a much larger role in the growth of government in the 1980s than those obtained by other special interests. This is because, when other special interests succeed, they often create rents for the government employees who administer the expanded programs.

Rents for Other Special Interests. Special interest groups try to extract money from government in the form of cash grants, subsidies, tax credits, loan guarantees, in-kind payments and so forth. Many special interests invest heavily in efforts to obtain such rents- hiring lobbyists, making political contributions, wining and dining politicians and staff. When government redistributes income, it almost by definition creates economic rents for various groups- the poor (via welfare programs), college students (via scholarships and loans), disabled people (via disability payments), corporations (via subsidized loans) and others. In most cases, it also creates economic rents for public employees.

"While overall spending increased by 47 percent, spending on highways grew only 15.6 percent and spending on police and fire protection rose only 15.3 percent."

Some have argued that expanded government programs employ people who would otherwise remain unemployed, and that by putting people to work these programs expand the total income of all Americans. This argument is not persuasive. If anything, the reverse holds. For example:8

• From 1900 to 1929, the median annual unemployment rate was
• 4.3 percent at a time when governments absorbed, in most years, around 10 or 12 percent of total output.
• From 1970 to 1991, the median unemployment rate was
• 6.45 percent, half again as large - at a time when governments were spending about three times the proportion of the nation's output as in the earlier period.

"Because government grew faster, by 1990 it took away about $245 more per person- almost $1,000 for a family of four."

Excess Compensation of Public Employees. The recent media attention given to New York City school janitors making almost $60,000 a year without being required to keep buildings clean illustrated a national problem. Compensation of state and local government employees has been increasing faster than that of private sector employees- with no evidence of increased productivity- and that trend accelerated markedly in the 1980s:10

• From 1980 to 1990, state and local employees received
• $47.3 billion more than they would have if their compensation had increased at the same rate as private sector pay.
• The phenomenon was widespread- in all but two states, public employee pay rose faster than private sector pay, often by substantial margins.
• In 1990, this excess compensation took, on the average, an additional $190 from every man, woman and child in the country.

The redistribution of income from the taxpaying public to public employees was sizable:

• In 14 states, the excess pay to public employees cost each man, woman and child more than $250.
• In Alaska, the redistribution of income to state employees was $1,038 per person, as falling oil prices depressed private wages but state pay continued to rise in real terms.
• In Louisiana, also pummeled by falling energy prices, the real average annual wage of private sector employees fell by $2,772 in the 1980s, but the average state government employee wage rose $2,472.12

• In Arizona, where rapid population growth masked a below-average increase in per capita income, other excess spending cost $389 per person by 1990 in addition to the $212 cost of excess compensation.
• In Texas, another state with high population growth and low per capita income growth, other excess spending cost $203 per capita in addition to $134 for excess compensation.
• Florida and Wyoming, both with high costs for excess compensation, had even higher costs for other excess spending.

"In 1990, excess compensation of public employees required an additional $190 from every man, woman and child."

The Total Excess Burden. When both types of excess spending are considered, we find that:

• In 29 states, the rise in the size of state and local government relative to income growth was largely or entirely a consequence of giving government employees larger pay increases than private sector employees received.13
• Total spending in seven states actually declined, but public employee compensation still grew faster than personal income.
• In Hawaii and Massachusetts, the relative size of government grew strictly because of spending other than excess compensation.

"In 29 states, the rise in the relative size of government was largely or entirely the result of overpayment of public employees."

The Growth Tax: Excess Compensation. Appendix A describes a statistical model that explains 80 percent of the variation in the rate of economic growth among states. The model shows a statistically significant negative relationship between increased government spending and economic growth, and the dramatic impact of excess employee compensation. Other things being equal, a 1 percent increase in the proportion of a state's per capita personal income going to public employee compensation lowers per capitaincome by more than 6 percent. This impact substantially hampered economic growth in the 1980s.

• By 1990, the average state had transferred 1.23 percent of per capita personal income to public employees in excess compensation since 1980.
• As a result, 1990 per capita income in the average state was
• 7.42 percent lower because of excess compensation alone.
• Since the average state had a per capita income of $15,108 (1990 dollars), the result was to lower per capita income by about $1,121.14
• This means that personal income for the nation was lower by
• $280 billion by 1990 as a consequence of slower growth due to excess compensation of public employees in the 1980s.

"Because excess compensation caused slower economic growth, personal income for the nation was lower by $280 billion in 1990."

Such a finding is extraordinary. Income in the typical state, which grew slightly less than 18 percent in the 1980s, would have grown by more than 25 percent- an increase of 40 percent- had public employee pay stayed in line with private wages. Table III details, state-by-state, the loss in economic growth. As the table shows:

• In 31 states, the excessive pay premium to state and local employees is estimated to have cost each person in the state more than $1,000 in income growth.
• The only two states that experienced a net loss of per capita personal income in the 1980s, Alaska and Wyoming, had the highest per capita burdens from excess compensation, $6,259 and $3,660 respectively, and both states would have enjoyed robust growth without the massive redistribution to public employees.15

• By 1990, the average state had transferred 0.4 percent of per capita personal income to other excess spending.
• This reduced per capita income in the average state by .32 percentage points, or $48.16
• This amounts to an additional $11.9 billion in reduced personal income by 1990 because of other excess public spending in the 1980s.

"The total growth tax from excess public spending was almost $292 billion in 1990."

The Total Growth Tax. Taken together, the loss of almost $280 billion from excess compensation and almost $12 billion from other excess public spending means that personal income in the United States was lower by almost $292 billion by 1990 because of excess public spending in the 1980s. The average per capita loss of income across the nation was $1,169- $1,121 because compensation to public employees exceeded private sector compensation growth and $48 because of other excess public spending.

The Impact on the Largest States. Figure II demonstrates the negative effects of excess spending on the eight largest states. It shows how much personal income would have grown if state and local government spending had risen at the same rate as income, compared to the actual increase. As the figure shows:

• The increase in per capita income in every large state except Illinois would have been at least 20 percent higher.
• In Texas, the growth of income would have been almost twice as high as it was.
• In Michigan and Indiana, income growth would have been more than 70 percent higher.

"The average per capita income loss was $1,169."

Case Studies. Over the very long run, economic forces encourage low-income states to grow faster than high-income ones.17 Businesses seeking to maximize profits move capital from high- to low-wage states. Workers seeking to maximize their wages move in the opposite direction. These movements reduce the differences in the amount of capital available for each worker among the states, which in turn reduces differences in productivity and in wages. Other things being equal, then, there is a natural tendency for per capita income in the states to converge. Yet in the 1980s, high-income states grew as much as low-income ones.

"Illinois grew faster than neighboring Indiana because of the difference in excess compensation."

One reason may be the large pay increases that many public employees extracted. The equalizing effect of competitive markets was offset by the income-reducing effects of redistributing income to government workers. Consider Illinois and neighboring Indiana:

• Although Illinois had a significantly higher per capita income than Indiana in 1980, it grew 19.1 percent in the 1980s compared to Indiana's 15.5 percent.
• The excess compensation of Indiana government employees absorbed 1.7 percent more of 1990 per capita income, compared with only 0.5 percent more in Illinois.
• This difference in excess compensation was enough to explain the difference in economic growth in the two states.
• Two other sets of neighboring states provide similar examples:
• Texas, with relatively low rent payments, outdistanced Louisiana, where excess compensation was twice as high.
• Idaho, with less than half the excess compensation of Montana, outperformed its high-rent neighbor.

By 1990, as Table IV shows, taxpayers had to pay more than $61 bil-lion in taxes to fund excess spending by state and local governments. They also had almost $292 billion less in pretax income than they would otherwise have had. The total cost to the nation of excess government spending was more than $353 billion. On the average, each person lost more than $1,400 in income in 1990. Table V shows these costs state-by-state:

• In more than three-quarters of the states, including seven of the eight largest, the total burden on the average citizen from excess government spending was more than $1,000 by 1990.
• In 12 states, the burden exceeded $2,000.
• In Alaska and Wyoming the burden exceeded $5,000.

High spending burdens were observed in a number of oil-producing states, notably Alaska, Wyoming, Oklahoma and Louisiana. Only in two states did spending restraint result in citizens' actually being better off, and in both cases the gain was well under $500. In Kentucky, Maryland, Michigan and West Virginia, pay increases to public employees were largely offset by reductions in the proportion of income going for other spending. Even in these states, however, income redistribution to public employees caused a significant net burden.

A large body of scholarly literature suggests that a negative relationship exists between the size of government and economic progress. Most of that literature focuses on the effect of taxes on either economic growth or production inputs. Several studies have concluded that tax increases reduce economic growth rates.18 Others have found that increased taxes reduce employment opportunities19 or encourage migration.20 Still others have found that taxes can have an adverse impact on plant location decisions.21

"On the average, every man, woman and child in America was $1,400 worse off in 1990 because of the growth of state and local government in the 1980s."

The literature on the effect of government expenditures on economic growth is much less extensive. Two studies have observed a relatively strong negative relationship between public assistance expenditures and the economic growth rate.22 This is consistent with literature on the labor supply effects of public assistance.23 The author and his colleagues also have documented a negative relationship between public aid benefit levels (relative to wages) and the labor supply.24

"There is no relationship between education spending and economic growth."

Education also has been examined extensively. Some studies purport to show a positive relationship between some forms of government spending on education and economic growth.25 The author's reading of one study, however, suggests that the observed statistical relationship is weak

. Education. There is a voluminous literature on the relationship between student learning and education expenditures. Two surveys have concluded that, in general, the studies show little or no relationship between learning and expenditures.26 Certainly the most careful and extensive of the studies reaches this conclusion.27 Even if spending does have some positive impact on learning, it seems unlikely that the overall effect on economic growth is positive, since taxes levied to finance the spending almost certainly depress economic activity. Using data on specific types of state and local government expenditures and a statistical model similar to the one used to obtain the above findings, we analyzed the relationship between changes in types of spending and economic growth in the 1980s. We concluded that there was no systematic relationship between education spending and economic growth. In fact, the evidence suggests that, for higher education, increases in spending were associated with reductions in economic growth.

"Spending on actual instruction improved student performance, but the impact diminished as the level of spending rose."

Because education is such an important issue, the author and some colleagues examined the role of government spending in learning, using detailed data on 610 Ohio school districts. [See Appendix B.]28 Among the conclusions:

• The relationship between overall spending and student achievement was very weak and statistically insignificant, consistent with dozens of other studies.
• Spending on actual instruction had a positive impact on student performance.
• Even that spending, however, was subject to diminishing returns: at low levels of spending, added expenditures brought increased learning, but at high levels of spending, such expenditures brought little or no more learning. [See Figure III.]
• There was either no relationship or a statistically significant negative relationship between school spending outside the classroom and student achievement.

"Welfare spending lowers economic growth."

Public Assistance. We observed an expected negative relationship between public assistance spending and economic growth. An increase of 1.0 percent in the proportion of personal income devoted to public aid lowered the economic growth rate by nearly 4 percentage points (e.g., from 16 to 12 percent). Thus the negative effects of welfare spending appear to be both real and relatively powerful, consistent with other studies and with economic theory.

In Indiana and Ohio, for example, the public welfare spending burden rose in the 1980s, whereas in neighboring Illinois and Pennsylvania it fell. Economic growth was greater in the states with a falling burden than in the states where welfare spending grew faster than personal income.

Highways. No category of state and local spending so purely fits the definition of infrastructure as highways, and one might expect a positive relationship between increased spending for highway construction and maintenance and the rate of economic growth. The experience of the 1980s, however, shows a slight negative relationship between increases in highway spending as a percent of personal income and the growth in real per capita income. This experience appears to contradict the conventional wisdom

. Other Spending. We also tested the effect of state and local spending on such things as hospitals, police protection, fire protection, parks and recreation, sewers and nonsewer-related sanitation services. With one exception, these forms of spending had a negative impact on economic growth. The exception was parks and recreation spending - in fiscal year 1990, less than 2 percent of total state and local direct general expenditures - where there was a consistently observed positive relationship between changes in spending and economic growth.

"Most states can substantially increase personal income growth by curtailing the growth of government spending. "

Education spending has not had a measurable payoff in terms of economic growth, either. There is some evidence that this reflects in part the growing proportion of resources going into activities other than conventional instruction (e.g., administration and support services). There is clear evidence that public assistance spending has an adverse impact on the economic growth rate. Evidence on other spending is mixed, but on the whole it suggests a negative relationship between spending and economic growth.

NOTE: Nothing written here should be construed as necessarily reflecting the views of the National Center for Policy Analysis or as an attempt to aid or hinder the passage of any bill before any state legislature.