homework

This seemingly straightforward task, however, is fraught with difficulties for many of the services provided by government. What, for instance, is the appropriate measure of service provided by local schools or a city police department? Although amounts of money spent on those functions—expenditures—are the most readily available and commonly used measure of the quantity of service, that measure often is not informa- tive. Additional expenditures that do not translate into more educated students or a safer environment may not represent more "service." Throughout this section of the book, and particularly in Chapter 7, the problems of appropriately measuring service and the limitations of using expenditures as that measure are emphasized.

Finally, governments implement their decisions about providing goods and services through budgets and budget policy. Even before the tax-limit movement, state-local governments had experimented with a variety of budget structures and restrictions. Understanding those constraints and their effects on fiscal decisions may help clarify why some states respond to economic and fiscal changes differently from others and may suggest the advantages and disadvantages of similar budget policies for the federal government.

CHAPTER 7

COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

. . . Rising unit costs have been a major (probably the single most important) source of recent increases

in local public budgets. 1 —DAVID BRADFORD, R.A. MALT, AND WALLACE OATES

"LIKE MANY STATES, NEW JERSEY CONTRACTS WITH AN OUTSIDE VENDOR TO HANDLE. THE BACK-

OFFICE SIDE OF ITS WELFARE AND FOOD STAMP PROGRAMS. PART OF THAT OPERATION, RUN BY AN

ARIZONA - BASED COMPANY CALLED EFUNDS, IS A SMALL CALL CENTER THAT HANDLES TELEPHONE

INQUIRIES FROM BENEFICIARIES. STATE OFFICIALS CAUGHT FLACK LAST YEAR WHEN THE VENDOR

MOVED THE OUTFIT FROM GREEN BAY, WISCONSIN TO BOMBAY, WHERE SALARIES FOR ANSWER-

ING CALLS ARE IN THE RANGE OF $2 TO $4 AN HOUR. IN THE WELFARE-TO-WORK AGE, CRITICS

SAY, IT'S UNFA1R—AND HUGELY IRONIC—TO SHIP ENTRY-LEVEL JOBS SUCH AS THESE OVERSEAS.

SO NEW JERSEY NEGOTIATED WITH EFUNDS TO BRING THE JOBS BACK. IN MAY, A NEW

CALL CENTER OPENED IN ECONOMICALLY DEPRESSED CAMDEN, WITH SEVERAL NEW EMPLOYEES

HIRED FROM THE WELFARE ROLLS. . . . NEW JERSEY AGREED TO PAY EFUNDS AN ADDITIONAL

$888,000 A YEAR AS COMPENSATION FOR THE HIGHER COST OF DOING BUSINESS IN CAMDEN.

... MORE AND MORE U.S. CORPORATIONS ARE MOVING CALL CENTERS, DATA PROCESSING,

AND OTHER BACK-OFFICE WORK TO NATIONS SUCH AS INDIA, ISRAEL, AND THE PHILIPPINES..

WITH THEIR BUDGETS IN SHAMBLES, STATE AND LOCAL GOVERNMENTS MAY BE

INCLINED TO TURN TO CHEAP OVERSEAS LABOR. 2 "

HEADLINES

I sThe Rising Cost of Local Public Services: Some Evidence and Reflections." islational Tax purruil, 22 (June 1969): 201.

2 "Answering the Call Center's Call." Governing Magazine. July, 2003.

141

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

In economics, analyzing supply is essentially analyzing production cost. The cost of producing alternative amounts of output, combined with the structure of the market, determines how producers behave. Similarly, the costs of producing services provided by state-local governments and the factors that alter those costs are crucial for understanding and comparing the fiscal behavior of subnational governments.

Before discussing production technology and cost, one needs to define and mea- sure the good or service produced, which is not straightforward for many services, including those provided by state-local governments. For example, education is the dominant subnational government service in the United States; the question is whether education output should be measured by dollars spent per pupil, by the number of graduating students, by student test scores, or by some other measure. The action required to increase each of these alternative measures of education may be different so that the cost of producing "more" of each may vary and even depend on different factors. The first task in this chapter is to consider alternative ways to characterize the output of state and local government services so that "cost" is defined properly and to investigate the factors that affect cost (and thus supply).

MEASUREMENT AND PRODUCTION OF GOVERNMENT SERVICES

Production Functions 3

To produce services, state-local governments purchase inputs such as labor ser- vices, capital goods, materials, and supplies and combine them to provide public facilities, or what can be called directly produced output, such as police patrols or classrooms with teachers and books. The ways inputs can be combined to produce this type of output are together referred to as technology and can be represented mathematically by a production function. For instance, the directly produced education output is a function of the number of teachers and administrators, the number of buildings and classrooms, and the number of books, desks, and other equipment provided. Mathematically

Q = q(L, K, X) where

Q = directly produced output

L = labor input

K = capital input

X = the set of other inputs such as materials and supplies

The q() function represents production technology. Any given amount of di- rectly produced output usually can be produced by different combinations of

3The discussion in this section follows that in David Bradford, It A. Malt, and Wallace Oates (1969).

ER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

„ as —that is, there is usually more than one way to combine inputs to produce

sicrvice. In other words, the production function q( ) does not specify a unique

°is') tit combination for each output but rather the possible input combinations to :hprimsdpuerccodeosetti

aoctchfiolnevte

technology producing

oouloantpgyyutam. amount of directly produced output depends both on and the prices of the required inputs. In defining

production cost, economists usually assume that for each possible level of output,

producers select the combination of inputs that will produce the chosen output at

lowest cost. 4 For instance, if K1, and X1 are the amounts of inputs that will

produce output Qi at lowest cost, then the

Cost of Qi = tvLi + rKi + pX1

where

w = the price of labor r = the price of capital

p = the set of prices for the other inputs.

Of course, this cost of the directly produced output is also the expenditure of the government on this service.

These public facilities or directly produced outputs provided by state-local gov- ernments may not reflect the services consumers desire, however. One can argue that citizens are more concerned about results than production; for instance, the education output of interest is knowledge and skills acquired rather than merely the number of classroom hours per year. The service result, which is what individuals consume or use, depends both on the directly produced output by the government and on the characteristics of the community and the population. For example, an equal number of classroom hours, teachers, and books will not necessarily produce an equal amount of learning in districts with different num- bers and types of students. It is useful, therefore, to distinguish consumer output, or the final result for consumers, from the directly produced output or facilities.

Mathematically,

G = g(Q, X, N, E)

where

G = consumer output X = private goods purchased directly by individuals

N = population to be served E = environment, a set of community and population characteristics

g( ) = transformation function from output to results

40f course, governments might not always select the minimum cost input mix. For instance, it has been argued that due to patronage consideration or public-employee unionism, state-local governments may choose to use more

labor than is cost minimizing.

► 43

0009 SEVEN • COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

Sa n plc Output Measures for Selected State—Local Services

as

Table 7.r

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

The "cost" of producing more directly produced output Q is different from the "cost" of producing more consumer output G. The latter depends on private con- sumption by residents and on community characteristics E and N, which are often outside the direct control of the state or local government. Private consumption may raise G if individuals purchase goods or services that contribute to the public service, such as private education or locks or smoke detectors; private consump- tion might reduce G if consumption imposes greater burdens on the public service, such as with consumption of alcohol and drunken driving. Changes in population or the environment may require a larger Q, just to keep G constant. For instance, to reduce class size from 25 to 20 students requires 25 percent more teachers and classrooms (assuming teacher workload and school operating hours are to remain the same), but such a change may not provide a 25-percent increase in the desired result of "learning" per student; indeed, it may not increase "learning" at all!

This discussion suggests that the output of state-local governments can be mea- sured in at least three different, broad ways. Output can be measured by the amount of money spent by a government on a service, which is referred to as expenditure; however, expenditure is really a measure of the inputs used by the government in the production process. Alternatively, government service may be measured by the amount of directly produced output provided by the govern- ment. Finally, government service may be measured by results, by the level of con- sumption enjoyed by citizens.

Examples of how these three different measurement concepts can be applied to specific state-local government services are shown in Table 7.1. Fire protection ser- vices, for instance, may be measured by the amount of money spent on firefight- ers, stations, trucks, and other inputs; by the number of hydrants and stations per square mile; or by some mix of the number of fires (prevention) and damage per fire (suppression). Similarly, police protection services may be measured by expen- ditures on officers, vehicles, jails, and other inputs; by the number of police patrols per square mile; or by the number of arrests made and crimes solved. Similar mea- sures can be devised for every service function or responsibility of state-local gov- ernments. But which measure is best? Or perhaps more appropriately, how do the measures differ in the information they provide?

Expenditures Compared to Produced Output Directly produced output on a service can fall even though expenditures are con- stant or even increasing. Similarly, two different subnational jurisdictions with equal per-capita expenditures on a particular function can provide different pro- duced outputs for that service.

Expenditures equal costs, and costs depend both on the amount of inputs used and the prices of those inputs. If the prices of inputs rise, then it will cost govern- ments more to provide the same produced output. Of course, governments may select a different production technology if relative input prices change—using rel- atively less of those inputs whose prices increase the most—but even then, total cost for every amount of directly produced output will increase, although perhaps

Patrol officers, supervisory officers, stations, radios, vehicles, jails, weapons

Teachers, books, buildings, desks, classrooms, computers, and other equipment

by less than if the government did not alter production methods. It follows that if input prices differ for different subnational jurisdictions, equal expenditures do not necessarily translate into equal produced output. Simply put, if teachers of the same quality cost more in one state than in another (and all other inputs cost the same), equal per-pupil expenditures in the two states translate into larger class sizes in the higher-cost state or less of some other input (books, for example).

These implications are very important because expenditures are the most com- monly used measure of subnational government output, at least for comparisons over time and among different jurisdictions. Over time, however, increases in Input prices require increased expenditures unless directly produced output falls or unless new ways (technologies), which require fewer inputs, for producing those services can be found. M with consumer expenditures, one can attempt to

144

Service Fire Protection

Police Protection

Education

Inputs Firefighters, inspectors,

stations, trucks, equipment, water supply

Direct Outputs Consumption Stations per sq. mile, Fire Prevention and

firefighters per Suppression: No. of station, trucks per fires per household station, hydrants per or employer, sq. mile damage ($) per fire,

civilian fire deaths per fire, fire insurance rates

Stations per sq. mile, Crime Prevention no. of patrols (or and Punishment: patrol officers) per Crimes per capita sq. mile, no. of (perhaps by type), intersections with civilian deaths traffic control, no. and/or injuries of jail cells per capita from crime, amount

($) of stolen merchandise, arrests per crime, crimes solved per reported crime

Teachers per student, Knowledge and books per student, Skills: Average classroom hours per &/or variance year, class size, of test scores, no. of subjects percent graduating taught on time," percent

attending college, percent employed after x years, added earnings

Table 7.2

Special District

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES CHAPTER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

146 1 47

allow for changing input prices over time by deflating government-expenditure data with a price index, usually the GNP implicit price deflator, which is separately available for federal and state-local government expenditures. No such general correction is available for comparisons among different jurisdictions, although evidence exists of substantial variation of some input prices among different state-local governments. Particularly, land prices and labor prices appear to vary widely at different locations, and both inputs are purchased in substantial amounts by state-local governments.

Produced Output Compared to Consumed Output

The consumed output or result for a particular service could decline or worsen even though a government provides constant or even increasing direct output. In fact, even if two governments provide equal directly produced output, citizens in those jurisdictions may receive different amounts of consumed output-that is, get different results or benefits as consumers.

The consumer output, which results from a given amount of directly produced output, depends on private consumption and on the environmental characteristics of the community and population. Between two cities with identical fire depart- ments, one might expect more fires and more serious fires in the city with fewer smoke detectors or with older buildings or with more wooden (as opposed to metal or brick) buildings. Equal fire protection in both cities may require more directly produced output in such a city-perhaps fire stations closer together, more pumper trucks per capita, or a more aggressive fire-inspection program. Similarly, as environmental conditions change over time, directly produced out- puts must change if consumer results are to remain the same. Of course, the envi- ronment can change in a positive way over time as well, requiring less produced output to maintain consumer results. For instance, if building materials and tech- nology mean that newer buildings are at lesser risk from fire or if individuals more commonly keep fire extinguishers at hand, then the amount of directly pro- duced fire-protection output consistent with constant fire protection could decline.

There are four reasons government expenditures may not be very good mea- sures of the ultimate benefits received by consumers from government production. Differences among jurisdictions or changes over time in (1) production technology, (2) input prices, (3) community environmental characteristics, and (4) private con- sumption patterns all can intervene in that relationship. For instance, rising expen- ditures may be sufficient to maintain constant produced output, given rising input prices, whereas a deteriorating environment may require increased produced out- put to maintain results. Thus, rising expenditures may not be inconsistent with falling consumed output or declining service quality. The opposite also may be true. In some cases, decreasing expenditures can be consistent with rising service results or quality if input prices decrease, the production environment improves, and/or individuals substitute private consumption for public service. Therefore, at the very least, these four factors must be considered and evaluated when using government expenditures for comparison purposes.

Among government policy makers, the idea of focusing on results rather than spending is referred to as benchmarking. States and localities evaluate their pro- grams by a series of benchmarks or performance measures comparing that juris- diction to others. For instance, Governing reports that "Mississippi plans to shift away from old-style line-item budgeting that merely measures inputs-what the

e

is spending on specific programs-to one that measures outcomes-what,

isast that, o the

in Mississippi contend, "If government begins to measure the effects

actually, is the effect of all that government spending" (Walters, 1994, 33). The hope e focus on outcomes will allow governments to better allocate resources.

O of its activity rather than merely what it spends on those activities, those effects- 'results'-will begin to drive the budget process" (Walters, 1994, 34).

employment and Labor Coots

When expenditures are used as the measure of the amount of government service supplied, output is actually being measured by the government's costs, and the major component of state and local government costs is for labor. As shown in Table 7.2, about 34 percent of state-local government direct noncapital expenditures in 2002 went to cover compensation of employees. Labor costs represented 45 percent of those expenditures by local governments, on average, but 63 percent of direct expenditures in school districts. In comparison, labor costs were only about 13 per- cent of federal government noncapital direct expenditures in 2002. The reason for the much greater importance of labor costs to states and localities compared to the federal government is the difference in the nature of services provided by those gov- ernments. State and local governments mostly provide goods and services to indi- viduals and businesses, which requires a substantial amount of labor to produce. The federal government mostly transfers money either to people (such as with

Wages and Salaries as a Percent ob Noncapital Direct Expenditure, by Type off Government, Selected years

Level of Government

Year Federal State-Local State Total Local County Municipal Township School District 1967 27.5

1977 203 1972 28.7

1987 1

1992 12.0

1982 7 13.8 16.1 9

2002 10.2

'2001

SOURCE: U.S. Department of Commerce, Compendium of Government Finances, table entitled Governmental Expenditure by Character and Object, various years. U.S. Department of Commerce, Governmental Finances, various years.

1997 10.4

54.5 51.2 45.2 42.0 40.7 37.7 36.5 33.6

42.7 39.7 31.5 30.1 29.3 25.1 23.8 20.4

60.5 57.9 54.5 50 50.3 48.6 47.6 46.6 44.8

48.0 44.9 47.5 45.0 42.8 40.8 395 40.2a

56.2 523 47.5 42.9 41.6 38.4 38.7 na

52.4 54.7 54.6 47.9 45.6 44.7 45.6 na

74.1 73.4 68.8 66.7 65.7 65.1 64.5 63.1 a

41.7 41.6 38.4 33.0 29.1 31.4 30.3 27.7 a

CHAPTER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

1 49

Table 7.3 [ state and Local Government employment and Earningo

Full-time Equivalent

-•

Year

State-Local Employment (thousands)

Percent of Total

Employments

State-Local Full-time

Equivalent Employmentb

I, State-Local Employment as Percentage of All State-Local Employment

State-Local Ave. Annual

Earnings per FIE`

1965 7,696 12.7% 6,937 90% S 5,616 1970 9,823 13.9 8,528 87 7,818 1975 11,937 15.5 10,111 85 10,900 1980 13,375 14.8 11,047 83 15,142 1982 13,098 14.6 10,829 83 17,826 1986 13,794 13.9 11,852 85 21,631 1990 15,219 13.9 13,080 86 27,732 1992 15,117 13.9 13,182 87 29,529 1997 16,733 13.6 14,214 85 33,273 2002 18,349 14.1 15,602 85 40,235

'State-local employment as a °A, of total nonagricultural payroll employment.

bFull-time equivalent employment adjusts for the number of part-time employees.

'Average annual compensation per full-time equivalent employee.

SOURCES: Bureau of Labor Statistics Web site; U.S. Department of Commerce, Public Employment, various years.

the average state-local salary. During that same period (1990-2002), per capita per- sonal income in the United States increased about 82 percent and consumer prices (measured by the CPI) increased about 38 percent. It is very difficult to compare salaries or wage costs of state-local governments to those of the federal govern- ment or private business because of substantial differences in the work activities provided in those different components of the economy. One study by Bradley Braden and Stephanie Hyland (1993) showed that comparing aggregate data, the cost per employee for wages, salaries, and benefits appears to be much higher for state-local governments than for private industry. Much of this difference disap- pears, however, when accounting for the differences in the mix of work activities. Braden and Hyland note that "Compensation costs were similar for industry activ- ities common to government and the private sector" (1993, 15).

PRODUCTIVITY AND COSTS

Input price increases lead to increased costs of providing state-local government services unless the input price increases are matched by increases in productivity. Further, because of the substantial importance of labor costs for state-local governments, changes in wages and worker productivity should be particularly important. But the market for state and local government workers is not isolated from the rest of the economy. Changes in the demand for and supply of labor throughout the economy can have important implications for the costs of

social security and Medicare) or to state-local governments (through grants). The federal government produces few services directly, so its labor cost share is lower.

If comparison is limited to expenditures for current operations, labor costs are obviously an even larger share. In 2002, employee compensation was 41 percent of current operation expenditures for state-local governments together, about 27 per- cent for states, and more than 50 percent for all local goverrunents. 5

For the 35 years represented in Table 7.2, the labor-cost share of direct expendi- tures for all levels of government in the United States decreased substantially. From 1967-2002, labor costs decreased from 42.7 to 20.4 percent of direct expendi- tures for states and from 60.5 percent to 44.8 percent for local governments. Simi- larly, the labor-cost share of direct expenditures for the federal government fell from 27.5 percent to 10.2 percent. This decline in the labor-cost share partly reflects some external factors (interest-cost shares are greater now), but also reflects changes in what state-local governments do and how they do it. As governments make relatively more transfer payments, for instance, the labor-cost share of spending falls because the government is spending the money on direct payments rather than on labor. Similarly, if government substitutes capital for labor in pro- ducing some services-automated trucks for sanitation workers, for example-the labor-cost share also will fall.

In 2002, state and local governments employed more than 14 percent-that is, one in every seven-of all payroll employees in the United States, as shown in Table 7.3. The share of total employees working for state-local governments declined from 1975 to 1997, but has risen a bit since. Similarly, although the num- ber of state-local government employees was about constant from 1980 through 1986, it grew by more than 20 percent from 1990 to 2002. Given the importance of labor costs to state and local governments, it is not surprising that this pattern mir- rors the course of state-local government expenditures relative to income noted in Chapter 1-the state-local government sector grew compared to the rest of the economy until the mid-1970s, changed little during the 1980s, but has grown at a faster rate than the economy since.

A substantial number of state and local government employees work part time rather than full time. Although states and localities had more than 18.3 million employees in 2002, only about 13.8 million or 75 percent were full-time employees. Full-time equivalent employment adjusts for part-time workers by computing the number of full-time workers needed to replace part-time workers (so two employ- ees each working half time are equivalent to one full-time employee). Since the mid-1980s, there has been little change in the ratio of full-time equivalent to total employment, suggesting that the use of part-time employees by states and locali- ties has not changed much.

In 2002, state-local governments paid an average salary of about $40,200 to full- time employees. Compared to 1990, this represents about a 45 percent increase in

5 Direct expenditures are total expenditures excluding intergovernmental transfers. Expenditures for current operations are direct expenditures excluding expenditures for capital, assistance and subsidies, interest, and insurance benefits. Expenditures for current operations represent money spent for current goods and services.

PART Ill • PROVISION OF STATE AND LOCAL GOODS AND SERVICES

1 4 8

Supply

Supply g

- - r

MRP 8 = Demand for labor

SuPPIYA

MRPP

Demand for labor

MRPE,

Col LA Labor in A

(a) Sector with productivity gain

WageA

W A

WA

Wages

PART Ill ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES CHAPTER SEVEN al COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

150

151

Productivity gains

cause wage

increases

providing state-local government services. This relationship among worker productivity, wages, and production costs between the state-local sector and the rest of the economy is the basis for one well-known theory about state-local government costs. This perspective is valuable in understanding the growth of state-local government spending.

The Baumol Hypothesia

In a now well-known 1967 article, William Baumol argued that productivity increases in some sectors of the economy would force wage increases throughout the economy, increasing the production costs in those sectors where productivity improvements do not occur. Professor Baumol further argued that the nature of some services, including many of those provided by state-local governments, effectively precludes productivity gains because the essence of the service is the labor itself. Higher wages simply cannot be offset by substituting other inputs for labor. For those services, unit production costs would certainly increase, and the choice for consumers is either to reduce consumption of the service substantially or to spend ever-increasing amounts to continue consuming current levels.

The first part of Baumol's argument is represented in Figure 7.1. The economy is divided, somewhat artificially, into two sectors, one where productivity gains occur relatively easily and regularly (Figure 7.1a) and one where productivity gains are difficult to achieve (Figure 7.1b). For this second sector, Baumol has in mind labor-intensive services with little opportunity for capital/labor substitution. In his words (1967, 416),

"There are a number of services in which the labor is an end in itself, in which qual- ity is judged directly in terms of amount of labor. Teaching is a clear-cut example. . . . Here, despite the invention of teaching machines and the use of closed-circuit

Lq Lg Labor in B

(b) Sector without productivity gain

television and a variety of other innovations, there still seem to be fairly firm limits to class size. . . . An even more extreme example is one I have offered in another context: live performance. A half-hour horn quintet calls for the expenditure of 2 1/2 man hours in its performance, and any attempt to increase productivity here is likely to be viewed with concern by critics and audience alike. 6

Which services and to what degree this characterization applies is debatable. The point here is that productivity gains for some state-local services are more dif- ficult to achieve than in some other industries. Accordingly, the demand for labor

in both sectors is shown in Figure 7.1, with demand less elastic in that sector where

substitution for labor is more difficult. Note that the demand for labor is labeled

the marginal revenue product of labor (MRP), which is the extra revenue a firm receives from hiring one additional unit of labor. The marginal revenue product is marginal revenue times the marginal product of labor and thus depends both on labor productivity and the value of the product produced. From microeconomic principles, a profit-maximizing firm will employ additional labor as long as the marginal revenue product is greater than the marginal cost of another worker, which is the wage in a competitive labor market. The demand for labor, then, rep- resents the benefit to a firm from more labor, which must be compared to the cost of hiring another unit of labor.

An increase in labor productivity in sector A is represented by an increase (a shift up) in the demand curve for labor; marginal revenue product is greater for every amount of labor because workers now produce more. The increase in labor productivity brings an increase in wage, at least in a competitive labor market. Pre- sumably, the same occurs in a controlled labor market as unions recognize the increased productivity of their members and bargain accordingly. The increase in wage in labor market A means that workers in sector A are now earning a relatively higher wage compared to those in market B than before the productivity improve- ment. The relatively higher wages in A attract workers from market B, causing a reduction (a leftward shift) in the supply curve of workers to market B and an increase in the wage of workers in B. In essence, employers in market B must match the wage increase in market A to retain employees.

These wage increases have very different effects in these two sectors. For sector A, workers are earning and producing more so that cost per unit of output need not increase. For sector B, the higher wages have been forced by changes in the other market and are not matched by productivity gains; remember that the premise of sector B is that substantial productivity gains are not possible. Therefore, the cost of producing a unit of sector B output rises. If B represents the position of state-local governments (and other industries as well), productivity gains in the industrial sector of the economy cause cost increases in the production of state-local government goods and services.

'Perhaps Baumol did not foresee the advent of computer-based music synthesizers, so that one programmer- performer could produce the horn quintet. But one might suspect that Baumol, and others, would see this option as substantially changing quality. In essence, the performance by the synthesizer is a different good (or bad) completely when compared to the quintet.

Figure 7.1

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

152

CH APTER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

153

BI Bo

Quantity of B

B 1 Bo Quantity of B

(a) Inelastic demand

(b) Elastic demand

10220111 How increases in costs of govern- ment services affect spending on services

Price

(S)

P1

Price

($)

P 1

Po

Po

The effect of these cost increases on consumption of sector B's output is repre- sented by Figure 7.2. If the demand for output B is price inelastic (Figure 7.2a), then the increased cost results in a higher price but only a small decrease in quantity. As a result, total expenditures on service B rise. If demand for output B is price elas- tic (Figure 7.2b), then the cost increase causes only a small increase in price but a large decrease in consumption. As discussed in Chapter 4, the evidence suggests that the demand for the services provided by state-local governments is very price inelastic. Therefore, the implication of the Baumol hypothesis is that productivity gains in some sectors of the economy will force increasing amounts to be spent on state-local government services. This is consistent with state-local government expenditures representing a larger and larger share of income. This problem will remain as long as private-sector productivity gains continue and public-sector pro- ductivity gains are difficult to achieve.

This story, although simplified, seems applicable to many actual circumstances. As wages in manufacturing and the business-service sector rise, fewer students may be attracted to teaching, a phenomenon that can be particularly evident for sci- ence, math, or business teachers who may find an attractive private market for their general knowledge and skills. As improved technology becomes a more important factor in manufacturing and demand for engineers and computer specialists rises, it becomes more and more difficult and expensive for universities to staff engineer- ing schools. Of course, as noted by Baumol, the process applies to many other ser- vices such as the arts, restaurant me als, fine hand-crafted furniture, and clothes.

Evidence: Government Productivity It is difficult to measure directly productivity change in the production of govern- ment services precisely because it is difficult to measure the output being pro- duced. One study (Hulten, 1984) attempted to measure productivity change in state and local governments indirectly, however, by using the difference between

directly produced output and consumed output. Households can be thought of as nroducing all final services by purchasing and combining different directly pro- rduced outputs, some provided by the private sector and some by government. (For instance, a household combines a privately produced recreational vehicle with a

publicly provided park to produce a service called camping.) In that case, the share of public-to-private expenditures depends on the relative prices of the products and relative change in productivity and environmental factors for the sectors. The combined change in productivity and environmental factors for the state-local sec- tor can be inferred from observed data on the share of state-local expenditures in GNP and relatiire prices.

Using quarterly data for the 1959-1979 period, Charles Hulten estimated the annual rate of change of the combined productivity /environmental factor to be -0.50 percent, although the estimate was not significantly different from zero. Hulten reported that one "cannot reject the hypothesis of zero productivity growth for the state-local sector" (p. 261). Perhaps more accurately, if there had been productivity growth over this period, it was not sufficient to offset a deteriorating production environment. After noting that private-sector productivity had increased at an average 1.45-percent annual rate over this period and that the state-local share in GNP had risen substantially in this time, Hulten concluded that "the results of this paper are thus consistent with the Baumol hypothesis on unbalanced growth. . . ." (p. 263).

Evidence: Government Coats Other studies have directly examined the costs of producing state-local government services and changes in those costs over time. In one such study, the changes in the prices of inputs and workloads from 1962-1972 for different state-local government services were computed and compared to changes in expenditures for those ser- vices over the period (Sunley 1976, reporting work by Robert Reischauer). For instance, local school input prices include teacher salaries, book prices, and trans- portation costs, whereas workload is the number of school-age children. If expen- ditures increased more than required by increases in input prices and workloads, the remainder is assumed to represent increases in amount or quality of service.

The result of this study was that 52 percent of the increase in total state-local expenditures over this 10-year period was due to increases in input prices and that 13 percent resulted from increased workloads. Thus, only about 35 percent of the increase in state-local government spending in that decade represented increased quality or new service. There were, however, substantial variations for different types of service. Workload and price increases were particularly important for highways and parking, health and hospitals, and police and fire protection. The increase in input prices and workloads alone were sufficient to increase total state- local government expenditures from 11.4 percent of GNP in 1962 to 12.0 percent of GNP by 1972. As shown in Table 7.3, this was a period when state-local govern- ment wages were increasing rapidly in an attempt to catch up with private-sector wages. As suggested previously, state-local sector costs are influenced by changes In the rest of the economy, and increases in state-local expenditures do not neces- sarily represent increases in output or service.

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

Percentage Change in Private-Sector Productivity and State-Local Coto

TECHNOLOGY AND THE PRODUCTION OF PUBLIC SAFETY

The essence of the Baumol hypothesis is that it is difficult for service providers, including state and local governments, to increase labor productivity by using more capital-intensive production technologies. However, in the case of one traditional state-local service- police protection and public safety-some new technologies involving electronic inputs-computers and information databas- es, computer analyses of physical and biolog- ical evidence, electronic devices for gathering data, new weaponry-are being used by

various jurisdictions. These methods hold the promise of producing public safety service more efficiently and perhaps lowering costs, but sometimes also raise difficult questions about the role of government and whether capital technologies change the meaning of "public safety."

The use of cameras always has been com- mon in private security work, including for surveillance in banks, retail stores, apartment building entrances, and in recent years at automated teller machines. However, only

since 1985, the prices of state-local goods have increased relative to those in the private sector. In the 1985-1995 period, productivity growth slowed to about 1 .5 percent annually, so the increases in labor earnings (70.2 percent) resulted in increases in business unit-labor costs (25.7 percent). The fact that unit-labor costs in business did not increase even more may be due to increased international labor market competition. Earnings in the state-local sector increased even more, so prices of state-local goods increased more than those of private consumer goods. Beginning around 1995, productivity began to increase substantially at a rate of about 4 percent annually. Earnings continued to grow at about 7 percent annually, but growth of unit-labor costs slowed to only about 1 percent yearly as a result of the productivity change. Although earnings in the state-local sector did not increase as much as in private business, they did grow faster than private unit- labor costs, and thus prices of state-local goods and services rose about twice as much as those of private consumer goods.

The fact that the relative costs of state-local goods continued to increase due to a relative increase in labor costs suggests that states and localities have not discovered ways of increasing labor productivity as much as in the private sector. If govern- ments cannot find these new technologies or production arrangements, then the fundamental assumption of the Baumol hypothesis continues to apply. The result- ing fiscal pressure for state-local governments was somewhat hidden in the 1990s, as the state-local tax structure continued to generate substantial revenue as a result of fast national economic growth. As that ended with the national recession in 2001, however, the resulting enormous fiscal pressure for states and localities returned. Two possibilities for government productivity change are considered next: the use of new technology to produce old services in Application 7.1 and the possibility of substituting private production of government goods in the following section.

Application 7.t

CHAPTER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

1 5 5

Implicit Price Deflator Business

Earnings State-local

Period Productivity

Unit Labor Cost

All State-local Consumption

Expenditure

1963-1973 27.4 46.7 73.7 81.9 38.6 71.3 1973-1982 4.6 109.1 102.4 89.1 96.1 109.2 1982-1985 6.1 6.5 13.3 17.5 12.6 14.5 1985-1995 16.5 25.7 70.2 86.4 27.8 35.6 1995-2004 32.6 11.7 57.2 46.1 13.0 27.8

SOURCES: Economic Report of the President, 1986 and 1993. US. Department of Commerce, Survey of Current Business, July issues, various years. U.S. Department of Commerce, Bureau of Economic Analysis Web site.

From the latter half of the 1970s through the first half of the 1980s, state-local government expenditures did not rise relative to income, however, suggesting that some aspect of the story changed. Possible explanations are that large producti- vity gains have, in fact, been made in producing state-local services or that the demand for state-local services has become more price elastic. The evidence reported in Table 7.4, however, suggests that the inverse of the Baumol hypothesis was operating from the mid-1970s through the early 1980s-low productivity growth in private industry helped to hold down relative state-local sector costs.

From 1973-1982, average annual earnings for full-time employees in all indus- tries rose by 102.4 percent, although business productivity, measured as output per unit of labor, rose by only 4.6 percent; consequently, unit labor costs for business rose by 109.1 percent. Over these years, increases in private-sector wages were not matched by productivity gains, so business labor costs rose substantially, presum- ably inducing business to demand less labor. Fewer private-sector jobs created some slack in the labor market, allowing state-local governments to hold down wages. Over these years, average annual full-time employee earnings in state-local government rose 89.1 percent, losing ground to private-sector earnings. As mea- sured by the GNP implicit price deflator, the prices of consumer goods rose 96.1 per- cent over these years while the price of state-local services rose 109.2 percent. Although the price of state-local services rose slightly compared to private con- sumer goods, the difference was much smaller than in the other two periods, when private productivity gains were large. Therefore, state-local government expendi- tures decreased from about 11.2 percent of GDP to 10.6 percent over these years.

The short 1982-1985 period is similar to the 1963-1973 period when substantial business productivity gains allowed earnings to rise with only modest increases in labor costs. The increase in state-local earnings during the 1982-1985 period is much greater than the increase in business unit-labor costs, and therefore the price of state-local services is again increasing much faster than the average price of private consumer goods. Thus, as expected, the share of GDP represented by state-local expenditures increased slightly between 1982 and 1985.

Pepper ... and Salt THE WALL STREET JOURNAL

0 7;

01. "This is Officer Holloway. You are exceeding

the speed limit by 6 mph. A ticket is being faxed to you."

Source: Reprinted from The Wall Street Journal—Permission, Cartoon Features Syndicate. August 27, 1993, p. A9.

pTER PART PROVISION OF STATE AND LOCAL GOODS AND SERVICES SEVEN • COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

157

Application 7.1 — Technology and the Production of Public Safety 'Application 7.1 — Technology and the Production of Public Safety

acceptance often cause serious problems. In

the case of photo-radar, the camera identifies

the vehicle and not the driver, so the penalty

must be against the registered owner of the

car, who may not be the user. If the owner is

required to "prove" that he or she was not dri-

ving the vehicle at the time of the infraction

to avoid penalty, is the presumption of inno-

cence lost? Technological advances to fight

crime also often result in technological

advances to defeat the new enforcement

technology, as illustrated by the ad in Fig-

ure 7.4 for"Photo B locker" to prevent enforce-

ment cameras from reading license plates.

is Also, The Wall Street Journal reported that

manufacturers of radar detectors were work-

ing on new devices to detect photo-radar

(which is shot across rather than along the

road). Finally, thinking about other uses of

cameras, one can reasonably ask whether

people would feel "better off" or even "safer"

if they were being watched all the time.

In short, technology can be used to

improve efficiency, increase worker pro-

ductivity, and reduce costs in providing

public safety service; however, these tech-

nologies also change the nature of the "pub-

lic safety" service. This issue raises the possi-

bility of an interesting economic choice to

be faced by voters. Voters can accept the

new technological methods of producing

public safety and enjoy lower costs (and

taxes) but suffer a loss of privacy, or they

can retain privacy by continuing to pay

higher and higher costs for producing

public safety with less invasive technology.

In essence, individuals might be asked

to put a value on the privacy that might be

lost in adopting these new technologies.

recently state and local police agencies have

adopted and expanded the use of cameras

and other electronic equipment. For instance,

The Wall Street Journal (Patterson, 1988)

reported about photo-radar, a high-speed

camera attached through a computer to a

radar gun, commonly used in Europe and

being used in California and some other

states. If the radar detects a speeding vehicle,

a photograph is taken, the vehicle is identified

by the license number, and the registered

owner is sent a summons (requiring the

owner to pay the fine or appear in court). In

Australia, similar types of camera-detectors

are used to monitor vehicle stops at traffic

lights or signs. Owners of vehicles that run the

lights (or signs) are mailed the evidence along

with the equivalent of a traffic ticket. The pos-

sibilities for this type of enforcement seem

limited only by imagination, as reflected by

the accompanying Pepper . . . and Salt cartoon (Figure 7.3).

*sim il a rly, computerized information data-

bases hold out the possibility of providing

information about individuals, things, or

events broadly to public safety officials

quickly and at low cost. Information about

individuals is perhaps most controversial. In

theory, it would be possible for public safety

agencies to access extremely detailed per-

sonal information about any person that

could be used in solving specific crimes or

even predicting potential criminal activity. Of

course, to be useful this information must be

available widely, which increases the danger

that it might be misused.

Finally, electronic monitoring now

being used to track or restrict persons who

are under investigation or who have been

arrested for or convicted of crimes. An "elec-

tronic tether" that emits an electronic signal

can be attached to an individual's body

(usually the ankle), allowing officials to

monitor the signal and know the location of

the individual. Such a system might be used

to prevent flight by someone waiting for

trial, as a means of partial confinement

(nonwork hours, for instance) for someone

who has been convicted, or to monitor the

behavior of someone on parole. One can

envision other types of electronic aids in

enforcing laws, promoting safe behavior,

and apprehending violators.

Obviously, some electronic public safety

activities might violate various provisions of

the U.S. and state constitutions, especially

concerning such topics as privacy, unrea-

sonable search and seizure, and the pre-

sumption of innocence. Even when these

measures are constitutional, however, imple-

menting the measures and gaining public

158 PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES pTEIR SEVEN ■ COSTS AND SUPPLY OF STATE

AND LOCAL GOODS AND SERVICES

have you been

caught by a photo

cop yet?

1/4wesecwe' yarcdalaRASH! Case

2 3 4

Degrees ob Public and Private Involvement in Provision of Service&

Choice of Quality/Quantity Public Public Public Private (Perhaps with Deregulation)

Financing Public Public Private Private

Production

Public Private Private Private

Tested by Or Police Dapwtmalit and FOX 'Wurprisingly effective..." FaxA irS

PhantomPlate products Ago, tr ram ft have been featured on EM61.,

What Is Privatization?

One idea that has been proposed to increase the productivity of government and thus reduce costs is to transfer production of government services to private firms, which is called privatization. The term privatization has been applied, however, to several different ways of increasing the activity of the private sector in providing public services, as outlined in Table 7.5. The traditional public service situation is case 1: public-sector choice, financing, and production of a service. The other cases represent various degrees of privatization: private-sector choice, financing, and production of a service, perhaps involving deregulation of private firms providing services; public-sector choice with private-sector financing and production; and public-sector choice and financing with only private-sector production of the ser- vice selected. The first simply means that all responsibility for a service is trans- ferred from the public sector to individual consumers who select the amount of ser- vice they desire and purchase that service from private suppliers. As an example, solid-waste collection is provided and produced by some local governments but left to private choice and private collection firms in other communities. This essentially can be characterized as "let the private sector do it alone."

The second and third versions of privatization refer, however, to joint activity of the public and private sectors in providing services. In case 2, the notion is that

consumers collectively select and pay for the amount and type of service desired through government, which then contracts with private firms to produce the ser- vice. The only difference in case 3 is that consumers pay privately for the service selected publicly. As discussed in Chapter 6, some local governments often con- tract with other governments to produce services in order to take advantage of economies of scale. Contracting with private firms to produce goods and services also may reduce costs. For the example of solid-waste collection, the idea is that the community selects a level of collection service and the government contracts with a private firm to do the collection and disposal. The service might be financed by government taxes and fees or by prices charged by the private producer. The gov- ernment provides for the service, although a private firm produces the service. Privatization has been an issue in recent years as some states and localities have experimented with or at least considered privatization for services usually both provided and produced by government in the past. These concepts of privatization are focused on in this chapter.

Private production of publicly selected and financed goods and services can be applied to intermediate goods used by government in producing services (such as cars and trucks, paper, machines, and materials), to services consumed by government in carrying out its responsibilities (such as maintenance and repair, construction, data processing, and management and financial services), and for the final services consumed directly by taxpayers (such as education, police and fire protection, and transportation). In the first instance, privatization is nearly universal. Few, if any, governments or government agencies produce their own furniture, forms, buses, or computers—all are purchased by the government from private producers. Concerning the other possibilities, in a review of priva- tization experience, Robert Poole and Philip Fixler (1987, 617) noted that "most privatization at state and local levels of government has been applied to either routine housekeeping services in which government itself is the customer (main- tenance of public buildings, vehicles, and infrastructure) or public services with well-defined tangible outputs (garbage collection or recreation, for example)." Increasingly, however, government is considering or experimenting with pri- vate production of traditional public goods, including public safety services and education.

2000-2005 photoblocker.com . All rights reserved.

PRIVATE PROVISION OF PUBLIC SERVICES

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

CHAPTER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

161

How Might Privatization Reduce Coats?

In its simplest form, the argument is that government producers have no in- centive to hold down production costs, whereas private producers who contract with the government to provide service do. Suppose, for example, that a private firm contracts with a local government to pick up six bags of garbage per house per week in the community for a fee of $100 per house per year. Obvi- ously, the lower the cost incurred by the firm in satisfying the contract, the greater profit it makes. Competition among potential private suppliers for this contract (for a limited period, after which government can change contractors) is expected to bring government the lowest possible cost for the specified level of service. As summarized by Janet Rothenberg Pack (1987, 527), ". . . competi- tive bidding by profit-maximizing firms for a well-specified output guarantees that the product will be produced at the lowest cost. The absence of compe- tition and profit incentives in the public sector is not likely to result in cost minimization."

The simple notion that government has no incentive to hold production costs down may be too strong, at least in the local government context, because local officials face competition from potential candidates and communities face competition from other communities for residents and businesses. If gov- ernment production costs for a service in one community are higher than they need to be, then taxes in that community also are higher than they need to be. As a result, households or businesses might move, as in the Tiebout process (see Chapter 5), to those communities with lower production costs for a given level of service. Similarly, candidates for public office could make the production inefficiency an issue in the local election. Therefore, it may be more accurate to argue that the incentive to hold cost down is greater for a profit-maximizing firm than it is for a government but not completely lacking in the latter. Essentially, the contention is that economic competition is more effective than political competition.

The three potential sources of lower production costs for private firms most often cited are lower labor costs, better management, and more research/develop- ment and faster innovation of the results. Lower labor costs may arise either from lower wages (which means that the government was paying wages higher than necessary for a given skill) or from less labor input (which means that government was hiring unnecessary workers or that fewer workers are needed with an alter- native production method). A private firm may more readily try out different pro- duction approaches, whereas government may tend to stick with the current approach, given that change often creates substantial political difficulties for local officials. Better management or experimentation and innovation with different production methods may be the reason a given level of service can be produced with fewer workers. In addition, private firms may use retained earnings to finance research or to purchase new capital equipment, which lowers unit production costs, whereas government may not be able to allocate tax revenues to those purposes as easily, given the many competing demands for a share of the government's budget.

When Might Privatization Not Work Well?

The three most often cited potential problems with private provision of govern- ment services arise from the bidding process, the precise specification of the con- tract, and monitoring and enforcing the contract. First, comvetitive bidding may not provide the service at lowest cost to the contracting government if there are

only a few (or even one) potential suppliers and the government has a limited idea about the level of costs. This might be the case especially in rural areas or when the production technology is relatively new. In addition, potential suppliers initially might offer a price to the government that is less than actual production costs to induce the government to adopt privatization or to win the contract. Subsequently, the contractor then would demand a higher price after the government has eliminated or dismantled its own production system. The chance of such "low- balling" in the bidding process may be reduced if the local government requires relatively long-term contracts.

The second potential difficulty with privatization concerns specifying the ser- vice to be provided in the contract. Earlier in this chapter, you learned that the out- put of a government service can be characterized by the inputs used or by alterna- tive measures of the produced output or final result, none of which are unique for a particular service. Characterizing output for some services is particularly diffi- cult when the government has multiple objectives. If society and the government are not certain what "good" education is and how to measure it, for instance, how can government contract for it? In the discussion about producing education in Chapter 19, the distinction between the average student test score and the variance of scores is emphasized. Getting the highest average test score may require apply- ing more educational resources to the better students with the effect of reducing the scores for the students at the bottom. As a result, the variation in test scores increases, which might contradict the distributional objective of government pro- vided education. It is difficult to think about how one would begin to specify the contracted output for police protection (a specified percentage of different types of crimes must be "solved"?) or fire protection (fires must be responded to in X min- utes with average damages limited to $Y?).

The third potential problem with private provision concerns monitoring the ser- vice quality provided by the private supplier and enforcing the contract when problems arise. Monitoring the performance of the private contractor itself creates costs, which may be substantial, for the government. In some cases, new data may have to be collected and analyzed. As one example, consider the costs of the U.S. Department of Defense in testing and evaluating weapons produced by private contractors to ensure they meet the contract standards. In addition, there must be a reasonable remedy if the supplier does not provide or stops the expected service. Suppose that the contractor underestimates the cost of production so that the price charged the government is not sufficient to cover all production costs, resulting in losses for the firm. If the firm simply stops providing the service, the implications could be serious in the case of many services such as police and fire protection.

David Sappington and Joseph Stiglitz (1987) have termed these contractual issues "the need for and costs of intervention in the private production process."

Figure 7.5

Privatization of

services to public

prisons C 0 0 p N15 $

c «. *.o

to o

IR VI

fa a/ C 1.# 2 2 >. c ca.

1

a. 7

76 1:1 7 ... to u 0

al ul

cc 0 0 • -0 9;

C

0

P e

rc en

ta g

e p

ri v

a ti

zi n

g c

e rt

a in

s e

rv ic

es

100

90

80

70

60 -

50

40

30

20

10-

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES CHAPTER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

162

X 63

They suggest that government should consider both the probability that interven- tion will be necessary and the costs of intervening if necessary. They conclude that "two important elements of this calculation include the complexity of the task under consideration and the need for rapid adaptation to unforeseen contingen- cies. When the task is particularly novel and complex, unforeseen contingencies are likely to arise. If rapid adaptation to these events is crucial, . . . public provision is more likely to be the preferred mode of organization" (p. 581).

Experience with and Proapecta for Privatization The available evidence shows that nearly all state—local governments contract with private firms to provide some final services to consumers or intermediate services to the government, but which services are contracted for varies greatly among gov- ernments. Donald Kettl (1993) reports that only two services (vehicle towing and legal services) were contracted out by at least half the local governments surveyed by the International City /County Management Association, and only 15 of the 75 service categories were contracted for by at least one-third of localities. According to this survey and others, the other most common examples of government con- tracting with private firms include legal services; hazardous waste disposal and solid waste collection; vehicle leasing and maintenance; vehicle towing; street light operation; street repair; landscaping and grounds maintenance; management of public facilities such as stadiums/arenas and convention centers; architectural, engineering, and management consulting; ambulance and EMT services; and some public-health services, especially for mental health and drug treatment. (The detailed list is in Table 6.4.) In addition, government continues to purchase most intermediate goods from private producers.

Kettl notes that government contracting is more likely, as with the preceding examples, when that service already is commonly provided in the private market. In such cases, government provision of these services may not be necessary when private provision is already readily available. Conversely, only a very small frac- tion of state and local governments contract for traditional programs central to those governments, such as prisons, police service, fire protection, traffic enforce- ment, libraries, or water and sewage treatment. As a result, privatization attempts in these areas are among the most dramatic and controversial.

One area where privatization has been tried but remains very controversial is public safety. For many years, fire-protection service in Scottsdale, Arizona, has been provided by a private contract service. Although this case has received substantial attention and at least one study shows it to be less costly than public protection, pri- vate fire-protection service is mostly restricted to specialized cases such as airports. A 1997 survey of local governments by the International City/County Management Association (VVemtz, 1999) showed that only about 2.5 percent of localities reported providing fire prevention or suppression service through contracts with private firms. Public safety areas in which contracting with private firms was used sub- stantially included ambulance and emergency medical services (a quarter to a third of localities) and vehicle towing and storage (80 percent of local governments).

Even more attention has been focused on the private ownership or operation of prisons, of which there are currently about 50 in operation spanning 14 states

Note: Data based on survey respondents from 25 corrections agencies in 22 states.

SOURCE: Reprinted with permission, Governing magazine, copyright 1993, 1994, and 1995. May 1993, p. 47.

(Lemov, May 1993). So far, most of these contracts have been for detention centers and minimum-security facilities, although there is some movement toward expanding the trend to higher-security facilities. Short of actually operating pris- ons, it is even more common for states to contract for private provision of some services in prisons, particularly health care, food service, and rehabilitation and education, as shown in Figure 7.5.

Many of the potential advantages and problems about privatization discussed previously are illustrated by the case of public safety. So far, any lower costs from private operation seem to have come mostly from lower wages or benefits paid to workers, but some firms are developing new educational or work programs for inmates. In establishing contracts, governments often try to specify both cost lim- its (often as a percentage of the cost at a state prison or a limited rate of growth) as well as performance measures to ensure that lower costs do not arise simply by providing worse conditions for prisoners. Governments that have privatized pris- ons believe that monitoring the actions of the contractor is crucial and thus often maintain a monitoring state inspector at the prison. Yet, serious worries about lia- bility remain. If an employee was seriously negligent, it seems possible that both the private contractor and the state could be sued. As a result of the damages, the private contractor could go bankrupt, leaving the state to pay damages and to absorb the costs of running the prison.

Poole and Fixler (1987, 619) argue that "four other functions generally carried out by government today are likely candidates for privatization in the next decade: tran- sit, highways and freeways, water supply, and education." Private provision already is used in all four of these areas to some degree. For instance, a number of private bridge firms have been employed, including the firm operating the tunnel under

PART 111 ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES ER SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

165 1

and the bridge over the Detroit River between Detroit and Windsor, Ontario, and the company operating the private toll bridge between Fargo, North Dakota and Moor- head, Minnesota. Both California and Virginia are experimenting with private toll roads that will compete with "free" public roads. Chicago, New York, and other cities have contracted out the collection of parking fees and some traffic fines.

Privatization of education may turn out to be most controversial, however. Edu- cation already involves a mixture of public and private provision in not only private elementary and secondary schools but also private day care, nursery schools, tutor- ial services, and extracurricular activities (for example, music and sports). But the notion of additional privatization in education as usually envisioned by its propo- nents involves either private provision entirely or direct competition between pri- vate and public schools. Individuals could receive education vouchers from the gov- ernment that could be "spent" for any school desired. Thus, government would continue to finance a substantial portion of education through taxes, but the educa- tion service would be produced by private schools. Proponents argue that the result- ing competition would reduce education costs and/or improve education results, partly because students would select schools most appropriate for them. Opponents of more privatization in education usually cite concerns about the distributional effects if students become more sorted by ability or other characteristics than they are with the current system. Indeed, some have argued that a diverse student mix is essential to the socialization objective of education and are concerned that there would be less diversity with private education than with the current public schools' structure. If education involves such important externalities, then public provision may be called for. These issues are considered in further detail in Chapter 19.

Substantial and increasing evidence also supports that the use of private firms to produce services has resulted in lower costs, especially for the more typical types of privatization. Kettl (1993) reports that 80 percent of localities that tried contract- ing out had cost savings of at least 10 percent (suggesting that localities are select- ing the correct services to contract for in aggregate). For instance, studies by E. S. Savas and Barbara Stevens (1977) and by James McDavid (1985) have found that public solid waste-collection services are 50 to 70 percent more expensive than equivalent private collection services. (This may not be too surprising because waste collection is entirely a private-sector activity in many communities, suggest- ing that there may be specific reasons why collection remains a public service in cer- tain areas). These cost savings almost all are labor savings arising from more flexi- ble methods of organizing and using workers or from paying lower wages than government or, most importantly, from providing fewer or lower-level benefits.

From a different perspective, John Donahue (1989) argues that any cost savings and other benefits from privatization arise not so much from contracting out per se, but rather from competition. Greene (2002) stresses this point as well. He notes that the research about costs of residential solid waste collection shows that gov- ernment contracting with private firms is least costly, but that an entirely open or private system is most costly. This may suggest that substantial monopoly power may arise in an entirely private system. Having government compete with private producers—as well as the private producers competing with each other—may help maintain competition and keep costs lower.

overnrnents may be able to enjoy the benefits of economic competition without ntracting out by encouraging public agencies or divisions to bid for projects airist private firms or other public entities. In Rochester, New York for example,

refuse collection workers changed their methods to keep costs below those of a .vate firm the city considered, contracting with. On the other hand, contracting

y riot generate lower costs or better quality if there are few private suppliers who

erefore have monopoly power (especially if competition from public provision

is eliminated). For instance, Kettl reports that only two main firms are involved in hazardous waste disposal, and that requests for mental health contracting in Massachusetts drew only 1.7 proposals from private producers, on average. With

few private suppliers and little competition in bidding, the government has few options if the contractor does not carry out the contract satisfactorily. From this perspective, the main focus of government officials should be on developing and maintaining competition among producers of government-provided services.

Application 7.2

PRODUCING CITY FIRE PROTECTION SERVICES ?

cost, is the large variation in the number of fire

hydrants; on average, there were 85.6 hydrants

per square mile, although the standard devia-

tion was 50 and the range from 14 to 302!

Input prices also varied substantially. Compen-

sation cost per full-time employee (a weighted

average of salaries and fringes for a first-class

firefighter and a department captain) varied

from $27,000 in Springdale, Arkansas, to

$119,000 in Washington, D.C. An index of the

cost of building and operating a fire station in

these cities varied from 76 to 114, with an aver-

age of all the city values equal to 96.

Economists would suspect that the amount

and type of inputs selected by these depart-

ments would be influenced by input prices.

Getz reported that the elasticity of labor per

square mile with respect to the wage was

-36; cities with higher wages used fewer fire-

fighters per square mile, although demand is

relatively inelastic. Interestingly, cities with

higher wages also used fewer stations and

trucks per square mile. Apparently, these fire

'See Getz (1979).

In a classic example of detailed examination of

a single service, Malcolm Getz (1975) surveyed

371 central-city fire departments about inputs

and production methods, costs, city character-

istics, and results. The survey resulted in usable

data from 187 different cities covering 44

states plus the District of Columbia. Many

other state-local government services have

been studied also (transportation and educa-

tion are discussed in subsequent chapters),

but this detailed examination of fire protection

provides an especially interesting example of

many issues discussed in this chapter.

Getz discovered great diversity among

these city fire departments in the amount and

types of inputs used to produce fire protec-

tion. On average, each fire station served an

area covering three and one-half square miles,

although the range was from one station for

one square mile to one for nine square miles.

Similarly, there were 6.8 firefighters per station

on average, with a range from 3.3 to 11.6. Per-

haps even more interesting, given their low

Application 7.2 — Producing City Fire Protection Services

departments attempted to keep the amount

of firefighters and trucks per station constant

and responded to higher wages by decreasing

use of all three. Getz also found that the

amount and mix of inputs depended on city

characteristics. Cities with older housing tend-

ed to use more of all inputs, cities with more

manufacturing used relatively more aerial

trucks compared to pumpers, whereas cities

that had more business than residential activ-

ity also used more of all inputs.

Getz attempted to measure how variations

in inputs influenced the effectiveness of the

fire department but found very little statisti-

cal relationship between additional inputs

and improved output. Fire-department out-

put was measured by number of fires per

1,000 houses and per 1,000 commercial and

residential employees, by the dollars of dam-

age per fire, and by the number of civilian fire

deaths per million population. Two results

that did appear were that more fire-code

inspectors decreased the number of multi-

family house fires and that more stations per

square mile decreased the amount of dam-

age per industrial fire. In the statistical work,

both the number of and damage from fires

were mostly related to the age of structures in

the city—cities with older structures had

more fires and more serious fires.

The premise of the Baumol hypothesis is that

productivity improvement is difficult to achieve

for some services, state-local government ser-

vices included. But Getz did find some major

technological changes in the methods and

equipment used in fire fighting. Among meth-

ods, upon arriving at a fire, a department must

choose whether to first run water-supply hoses

from the nearest water supply or to immedi-

ately attack the fire using a relatively small

amount of water carried in a pumper truck.The

latter method, called a "booster attack," was

introduced around 1922 and is now routinely

used by more than half of the departments.

Technological changes involving equipment

include use of breathing apparatus (first used

in 1940); power saws for quick access (1958);

chemicals added to water for fighting flam-

mable-liquid fires, called "light water" (1956);

and a quick-connect hose coupling (1964).

One other major attempt to increase pro-

ductivity and lower costs of fire (and police)

protection has been the creation of consoli-

dated public safety departments to provide

both fire and police functions. 8 Some depart-

ments are fully consolidated with all duties

performed by public safety officers, others

perform dual duty only for some services or in

limited geographic areas, and still others are

consolidated only at the administrative level.

Crank (1990) reported that only 1 to 2 percent

of local police and fire services are provided

through consolidated departments, with the

middle 1970s the most common period for

initial consolidation. Not surprisingly, consoli-

dated service tends to be more common

among smaller localities where there are

potential cost savings from economies of

scale. Because of quality concerns about

training of dual-service personnel and eco-

nomic concerns about appropriate wages

and benefits, some states and localities have

forbidden such a production arrangement.

Therefore, it appears that some productivity

improvement has occurred in producing fire

protection, but it is not clear that the gain has

been sufficient to prevent cost increases nor

that all such possibilities will be embraced.

Application 7.3

PRIVATIZING ENTIRE STATE DEPARTMENTS 9

Governing Magazine reported about an

attempt in Texas to take privatization in a new

and much more expansive direction than has

been used in the past. As previous discussion

in this chapter suggests, most privatization by

the state-local sector is applied to purchases of

specific inputs used by the government (such

as motor vehicles) or to very specific, direct

services (such as health care in prisons or com-

puter systems programming). But Texas is now

exploring the option of consolidating state

departments and contracting as much of the

operation of those departments as possible.

A recent law directs the Texas Health and

Human Services Commission to consolidate 12

health and human service agencies into 4 and

then to seek private sector operation of as

much of the work of those agencies as possible.

If implemented, the Commission would serve

the continuing political or traditional role of

state government, contracting with private

firms to undertake the day-to-day work of the

agencies. The initial intent is to contract with a

private firm to establish a call center staffed

by private employees who would evaluate the

circumstances of individuals and determine

whether and what benefits they might be eligi-

ble for. So, rather than visiting the local office of

the state human services department to apply

for specific benefits, the individuals would tele-

phone the center and have those decisions

made by the contractor. As this approach is

expanded, the state is on the verge of transfer-

ring substantial operational decisions—and

perhaps even policy decisions—to individuals

who are not state employees but contractors.

The goal, it seems, is to be more efficient—

to deliver more service at constant cost or

constant service at lower cost.

Expanding privatization in this manner

clearly is controversial and will be watched

closely. It may or may not turn out to be a

mechanism to reduce costs. Imagine, if you

will, not going to your local school for class

taught by a teacher employed by the school

district. Instead, students go to Language, Inc.

for Spanish class, Math R Us for math class,

and so on—each provided by a private con-

tractor hired by the school system. That is the

nature of the issue.

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

"See Crank (1990).

CHAPTER SEVEN • COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

167

SUMMARY

The output of state–local governments can be measured in at least three different, broad ways. Output can be measured (1) by the amount of money spent by a gov- ernment on a service, referred to as expenditures; (2) by the amount of directly produced output provided by the government; and (3) by results—the level of consumption enjoyed by citizens.

State–local governments purchase inputs such as labor services, capital goods, and materials and supplies and combine then in some way to provide public facil- ities or what is called directly produced output. The cost of the directly produced

'See Walters, Jonathan. "Going Outside." Governing Magazine, May 2004, pp. 23-29.

PART III ■ PROVISION OF STATE AND LOCAL GOODS AND SERVICES

168

output, which depends on the production technology and the prices of the inputs, is the expenditure of the government on this service.

The service result, called the consumer output, depends on the directly pro- duced output provided by the government, on the private consumption decisions of individuals, and on the characteristics of the community and the population.

If the prices of inputs rise, then it will cost governments more to provide the same produced output. And if input prices differ for different subnational juris- dictions, equal expenditures by different jurisdictions do not necessarily translate into equal produced output.

Expenditures for direct-labor services represent about half of the expenditures by state-local governments on average. State-local governments are also one of the largest employers in the economy, employing about one of every seven employees.

Baumol argued that productivity increases in some sectors of the economy would force wage increases throughout the economy, increasing the production costs in those sectors where productivity improvements do not occur. The nature of some state-local government services precludes productivity gains because the essence of the service is the labor itself. For those services, unit production costs would certainly increase, and the choice for consumers is either to substantially reduce consumption of the service or to spend ever-increasing amounts to contin- ue consuming current levels.

DISCUSSION QUESTIONS

1. "If one city spends more on police-protection services per capita than does another, one expects less crime in the first city than in the second." True, false, or uncertain? Explain.

2. At a public-budget hearing, a citizen once argued, "Education expenditures have increased 5 percent in each of the past three years even though student enrollment has been declining. Where is the extra money going? It seems to me that if the number of students declines, expenditures should also decline." Is the citizen right or wrong?

3. "If the Baumol hypothesis is correct concerning local government finances and if the price elasticity of demand for local services is inelastic, then we are in trouble—eventually, spending for education, police and fire protection, and sanitation will require half of our incomes." Evaluate this concern. What changes could occur to prevent this from happening?

4. Competing with private-sector salaries is a common problem for some academic departments in universities, particularly in engineering, accounting, other business fields, and biological science. If universities do not match the salaries, they may be unable to hire professors, or at least the better candidates; if they do match the salaries, then the cost of operating those programs (and eventually tuition) will increase. How might universities change the production of engineering or business education to

cHA g SEVEN ■ COSTS AND SUPPLY OF STATE AND LOCAL GOODS AND SERVICES

169

problem—that is, how could professors be substituted for or

made aoudid

ietYin

this re productive? Do you think those changes would affect the „

or nature of education in these fields? Does this problem apply to

as well as public universities?

SELECTED READINGS

Baunio, William. "Macroeconomics of Unbalanced Growth: The Anatomy of the Urban

Crisis." American Economic Review, 62 (June 1967): 415-26.

Bradford, David F., R. A. Malt, and Wallace E. Oates. "The Rising Cost of Local Public Services: Some Evidence and Reflections." National Tax Journal, 22 (June 1969):

185-202.

Greene, Jeffrey D. Cities and Privatization. Upper Saddle River, N.J.: Pearson Education,

2002.

Hirsch, Werner. "State and Local Government Production." In The Economics of State and

Local Government. New York: McGraw-Hill, 1970,147-65.

Kettl, Donald F. Sharing Power: Public Governance and Private Markets. Washington: The

Brookings Institution, 1993. Pack, Janet Rothenberg. "Privatization of Public-Sector Services in Theory and Practice."

Journal of Policy Analysis and Management, 6 (Summer 1987): 523 -40.