View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.



A review from the
Federal Reserve Bank
of Chicago

The 199 0 C lean A ir A ct:
a tougher regulatory challenge
facin g M idw est industry
Producer services:
trends and prosp ects
for the Seventh D istrict



The 199 0 C le a n A ir A c t: a tougher
regulatory challeng e facin g
M idw est in d u stry................................................................................................ 2
Donald A . Hanson

A market incentive system for reducing sulfur
dioxide emissions is one of the highlights of the
1990 Clean Air Act Amendments. The author
discusses this and other issues in an overview of
the 1990 CAA Amendments and illustrates the
economic impact of environmental regulations
using a simple economic model.

Producer service s: trends and
p rosp ects for the Seventh D is t r ic t ............................................................19
W illiam A . Testa

Does the economic future of the Midwest depend on
selling advertising and management consulting rather
than wheat and widgets? The author discusses the
evidence for and against the idea that growth in
producer services can independently spur regional
economic growth.


M ay/Ju ne, 1992 Volum e XVI, Issue 3

Karl A. Scheld, Senior Vice President and
Director of Research

the Research Department of the Federal Reserve
Bank of Chicago. The views expressed are the
authors’ and do not necessarily reflect the views
o f the management of the Federal Reserve Bank.
Single-copy subscriptions are available free of
charge. Please send requests for single- and
multiple-copy subscriptions, back issues, and
address changes to Public Information Center,
Federal Reserve Bank o f Chicago, P.O. Box 834,
Chicago, Illinois 60690-0834, or telephone
(312) 322-5111.
Articles may be reprinted provided source is
credited and The Public Information Center is
provided with a copy o f the published material.

Editorial direction

Carolyn McMullen, editor, David R. Allardice, regional
studies, Herbert Baer, financial structure and regulation,
Steven Strongin, monetary policy,
Anne Weaver, administration

Nancy Ahlstrom, typesetting coordinator,
Rita Molloy, Yvonne Peeples, typesetters,
Kathleen Solotroff, graphics coordinator
Roger Thryselius, Thomas O’Connell,
Lynn Busby-Ward, John Dixon, graphics
Kathryn Moran, assistant editor

ISSN 0164-0682

The 19 9 0 Clean A ir A ct: a tougher
regulatory challenge facing
M idw est industry

Donald A . Hanson

Protecting the environment
and human health is very
important. Toward these
goals, several new programs
have been initiated under the
1990 Amendments to the Clean Air Act (CAA).
This article provides an overview of these pro­
grams and their potentially extensive impacts
and discusses the attendant challenges facing
management and workers in U.S. industry.
According to preliminary U.S. Environ­
mental Protection Agency (EPA) estimates, the
annual cost of these regulatory programs will
be around $20 billion [U.S. EPA (1990a)], but
other estimates are considerably higher [Fumento (1992)]. The potential for a greater
burden is due to a number of considerations
which are difficult to quantify: 1) a greater
need for compliance planning, including con­
tingency planning; 2) administrative costs,
potential delays, and a loss of operating flexi­
bility for industrial facilities because of a com­
plex, new permitting program; 3) direct compli­
ance costs from more stringent regulations
requiring higher pollutant removal rates and the
potential for even tighter regulations in the
future, as provided in the 1990 Amendments;
4) increased monitoring and reporting require­
ments; 5) new, stringent civil and criminal
enforcement penalties; and 6) various rigidities
which may aggravate economy-wide effects.
The term “rigidities” refers to elements in labor
markets or in markets for goods and services
which may lead to deviations from full employ­
ment and efficient growth.
Increased protection of the environment
and health will require significant changes in


management, worker, and consumer attitudes
and behavior. In attempting to lessen the costs
of more stringent regulations, consumers and
producers will act more judiciously in choosing
among products purchased and the processes
and materials used to manufacture them. But
even with these adjustments, national and re­
gional incomes as traditionally measured,
which exclude many of the benefits of environ­
mental improvements, may still be lowered by
environmental regulations.
Because of its manufacturing orientation,
the Midwest economy faces a greater challenge
in striving for improvements in products and
manufacturing to reduce environmental residu­
als. If the region can master these challenges, it
may be possible not only to mitigate much of
the potential cost of environmental legislation,
but even to transform a regulatory burden into
enhanced growth and welfare. By developing
expertise in the design and manufacture of
clean processes and environmental controls,
some businesses could potentially cultivate a
new source of income. The challenge will not
be easily mastered, however, because many of
the requirements (or alternative options) for
CAA compliance are ambiguous or have yet to
be determined. Businesses will have to work
with multiple federal, state, and local govemDonald A. Hanson is a visiting scholar at the Feder­
al Reserve Bank of Chicago, manager of the energy
policy section at Argonne National Laboratory, and
adjunct professor of economics at the University of
Illinois at Chicago and DePaul University. The
author w ould like to thank W illiam A. Testa fo r his
extensive and insightful com m ents on several
versions of this article.


ment agencies in a heretofore untested partner­
ship. Further, scientists do not fully understand
the physical mechanisms leading to environ­
mental effects. These uncertainties will require
flexible business decision making and contin­
gency planning.
In addition to providing an overview of the
CAA, a second objective of this article is to
present a simple economic model which illus­
trates more rigorously the nature of the costs
associated with environmental regulations. In
the model, environmental expenditures are
placed in the context of the larger economy in
order to investigate the macroeconomic, indus­
trial, consumer, and regional impacts. The
model identifies one of the potential reasons for
macroeconomic cost impacts: workers may not
be willing to accept the lower wages (reflecting
clean air costs) which would be necessary to
maintain full employment. Such so-called
“sticky wages” can magnify the economic costs
of the CAA program substantially; a numerical
example shows costs under such a scenario to
be higher by a factor of two compared with
direct abatement expenditures. Addressing
labor compensation issues in the face of in­
creased real abatement costs and, in many
cases, strong international competition, is a
great challenge for industry and labor alike.
From fairness and political points of view, it is
probably undesirable for environmental regula­
tory costs to fall h eavily on labor incom e; how­
ever, if the costs are borne by capital income,
there may be less incentive for investment and
future economic growth. This model addresses
the cost side of environmental regulation but
does not attempt to estimate the future benefits
from a cleaner, more healthful, and productive
The second section of this article provides
an overview of the 1990 Amendments. The
third section describes the acid rain regulations
and highlights compliance within the Midwest
region. The fourth section reviews the other
major CAA areas: ambient air quality stan­
dards and nonattainment, prevention of signifi­
cant deterioration, visibility and other air quali­
ty values, industrial air toxics, and stratospheric
ozone depletion. Some of the burdens and
challenges to the economy under the CAA
programs are illustrated using a simple model
presented in the fifth section. Conclusions are
presented in the final section.



O rigins and o v e rv ie w o f th e 1 9 9 0
C A A A m e n d m e n ts

Following a rising tide of environmental
awareness in the United States during the 1980s
and a decade of research and congressional de­
bate on acid deposition controls, the Bush Ad­
ministration took a lead in the reauthorization of
the Clean Air Act. Compromise legislation was
finally crafted among the Administration’s pro­
posed bill, a House version, and a Senate ver­
sion, which resulted in the 1990 CAA Amend­
ments, signed by President Bush in November of
A system of tradable sulfur dioxide (SO,)
allowances was adopted as the central approach
to reducing acid deposition. This approach was
recommended in a December 1988 study spon­
sored by Senator Wirth, Colorado, and Senator
Heinz, Pennsylvania, entitled Project 88, Har­
nessing Market Forces to Protect Our Environ­
ment: Initiatives for the New President. The
study participants included not only academic
economists but also business leaders and repre­
sentatives of environmental groups, such as the
Environmental Defense Fund. The adoption of
this market based approach helped achieve the
passage of the CAA legislation in Congress and
hopefully may foreshadow considerably more
reliance on market incentive approaches to other
environmental regulations in the future. Recent­
ly, round two of Project 88 has been published,
further assessing the applicability of market
incentives for environmental improvements.
Overview of the main titles

Three new innovative titles have been add­
ed: Title IV on acid rain control, Title V on com­
prehensive source permitting, and Title VI on
stratospheric ozone protection. Many modifica­
tions and additions to existing titles were also
made. For example, the hazardous air pollutant
section of the 1970 CAA, Section 112, which
had been an unworkable framework for control­
ling toxic air emissions, has been totally replaced
with a new program; and the difficult urban
ozone nonattainment problems have been tackled
with new stringent requirements under Title I
(see Box l).1
A cleaner environment is no free lunch

Acid rain control was a contentious issue
during the 1980s because of the sharp divergence
of regional interests and the perceived high con­
trol costs, estimated to be about $4 billion per


year [NAPAP (1991)]. The 1990 CAA Amend­
ments authorize two significantly larger pro­
grams than acid rain control—urban ozone
regulation and industrial air toxic pollution
control—which are each likely to be two or

three times as expensive to the nation and the
Midwest economy as acid rain controls [see
Portney (1990) and U.S. EPA (1990a)].
Other provisions of the CAA may also
impose significant regulatory and compliance



The major titles under the new Clean Air Act and the relationship
to the titles in the 1990 Amendments
Structure of new Clean Air Act

Comments and relation
to 1990 Amendments

Title I

Central title o f the CAA

Air pollution standards and controls

Part A
Sec. 109

Provides for setting National Ambient Air
Quality Standards (NAAQS) for common
SIPs and NSPS continue to be
required after the 1990


Requires states to develop State
Implementation Plans (SIPs) to attain
NAAQS in nonattainment areas and to
maintain air quality in attainment areas.
SIPs require federal approval.


New Source Performance Standards (NSPS)
must be met as a minimum level o f control
for new sources or existing sources undergoing
a major retrofit.


Control of hazardous air pollutants

Sec. 112 of the 1970 CAA was
replaced with a new program,
Title III of the Amendments.


Federal enforcement

This section o f the 1970 CAA was
replaced with a much stronger
provision in the Amendments
Title VII.


Prevention o f Significant Deterioration applies
to maintaining air quality in areas already in
attainment with NAAQS.

Part B o f 1970 CAA was replaced
with a more stringent ozone
control program under Part D.

Part D

Nonattainment areas

Additional provisions were added
for ozone nonattainment controls,
carbon monoxide, particulate
matter, sulfur dioxides, nitrogen
oxides, and lead.

Title II

Mobile Source Emissions

Title III

General Administrative/Misc.

Title IV

Acid Deposition Control

New title

Title V


New complex permitting program

Title VI

Stratospheric Ozone Protection

New title

Additional Titles VIII - XI under the Amendments call for more research, a visibility impairment assessment,
and other matters.



costs. The new Title V comprehensive permit­
ting program could be a sleeping giant. If each
significant piece of industrial equipment emit­
ting at least one regulated pollutant, such as an
industrial furnace, is defined by the EPA as
requiring a permit, the paperwork and attendant
delays could be overwhelming. Operating
flexibility of industrial facilities will be reduced
and continuous monitoring, record keeping, and
periodic reporting will increase. Even though
the 1990 CAA Amendments require that each
state have its own permit program (with the
federal government running the permit pro­
gram, if a state refuses), each source permit
may be reviewed by the EPA and the designat­
ed Federal Land Manager, such as the National
Park Service. Further, after the federal review
there still can be intervention by affected par­
ties in the state or adjacent states. For example,
sulfates affect the atmospheric process of light
scattering and can cause visibility impairment.
Because of the long range transport of sulfates,
it is possible that sulfur dioxide (SO,) reduc­
tions needed to obtain a permit will be much
more stringent than would be necessary under
the acid rain control program, Title IV of the
1990 CAA Amendments. Restrictive permits
may also interfere with the effectiveness of the
tradable SO, allowance program, which is de­
scribed later in this article.
Advantages of market approaches to
emission reduction

The CAA pioneers the heretofore under­
used market based approach to reducing emis­
sions. The market based approach grants firms
pollutant emissions allowances while ensuring
that total pollution does not exceed national or
regional limits. The overall limit ensures that
society’s goals for clean air are met. The mar­
ket based approach allows firms to trade their
emissions allowances so that pollution reduc­
tions are made by those firms that can make
them most cheaply.
Market approaches, such as tradable emis­
sion allowances, can lower society’s costs of
pollution reduction by providing a level playing
field for emission reductions among potential
sources while, over time, stimulating the devel­
opment of new abatement methods and technol­
ogy. These characteristics of market approach­
es differ markedly from technology or regulato­
ry standards. For example, if the law requires
industry to install the “best available control



technology,” then there may be little or no incen­
tive to innovate in ways that reduce pollution
because the government would only proceed to
make emission standards more stringent. Also,
more stringent standards on new sources than on
existing sources encourage the continued use of
existing equipment beyond the point at which it
would otherwise have been economical to mod­
ernize and replace the equipment. Hence, some
critics argue that new source performance stan­
dards lower investment and reduce economic
growth. In contrast, the level playing field under
a market approach encourages new investments,
reducing emissions and the need to purchase
emission permits. In addition to the national
market for SO, allowances, a tradable permit
program is being proposed to address the region­
al smog in the southern California air basin
around Los Angeles.
Interaction among the requirements

Urban ozone is one of the pollutants for
which National Ambient Air Quality Standards
(NAAQS) have been set to protect human health
and other environmental values under Title I of
the CAA. Those airsheds in which the pollutant
concentration exceeds the NAAQS are classified
by the EPA as nonattainment areas for that pol­
lutant. The six common pollutants which are
associated with NAAQS are shown in Box 2,
along with a summary of the reductions in these
pollutants since the mid-1970s achieved so far
under the CAA.
The interaction and overlap between air
programs can lead to uncertainty about compli­
ance strategies. A single source may be regulat­
ed under more than one program. Many volatile
organic compound (VOC) sources will be regu­
lated under both urban ozone nonattainment and
industrial air toxic provisions. Nitrogen oxides
(NOx will be regulated under Title I in nonat­
tainment areas and under Title IV on acid rain
control. Sulfur dioxide is regulated under Title I,
Title IV, and may be further controlled in the
future to improve regional visibility. Any new
source not only must go through New Source
Review and meet New Source Performance
Standards (NSPS), but also must meet possibly
more stringent regulations due to being in nonat­
tainment areas or being subject to prevention of
significant deterioration applicable to cleaner air
areas. Sources of hazardous air pollutants re­
ferred to as air toxics must also meet Maximum
Achievable Control Technology (MACT).


BOX 2

Trends since the mid-1970s following the original CAA
Regulations implementing the original CAA,
such as State Implementation Plans (SIPs), have
taken effect over various lengths o f time. Mean­
while, the composition o f economic activity has
been shifting away from material intensive process­
ing, a trend working in the direction of lowering the
amount o f pollution per dollar o f GNP. For exam­

1975 to 1990
% change
Pollutants affecting am bient
air quality standards

Sulfur dioxide (S 0 2)
Oxides of nitrogen (NOx)
Volatile organic
com pounds (VOC)


Carbon m onoxide (CO)




Particulate matter


Energy use







Economic activity

Real GNP

ple, the share o f manufacturing in GNP has de­
creased from 24 percent to 19 percent from 1977 to
1989. In spite o f these complexities, it is interesting
to compare the changes which have taken place over

Existing sources are also subject to the New
Source Review if operating changes increase
emissions or if a major capital outlay is re­
quired to refurbish a unit. While the new state
permit program is designed to consolidate over­
lapping requirements on a source, obtaining
permits may involve lengthy delays and the
permit might require even stronger controls to
protect what the CAA refers to as “air quality
related values.”
With the attention focused on urban ozone
nonattainment, any changes in operation of a
facility will be scrutinized carefully, and per­
mits for sources of VOCs or NOx will be grant­
ed only with stringent controls. The cleanup
of existing sources will need to be extended to
the multitude of small sources and small busi­
nesses such as dry cleaners and auto repair



Em ployment


the 16 year period from 1975 to 1990. The time
trends for the three precursors o f acid deposition are
shown in Figure 1.
Hence, emissions of major pollutants regulated
under the CAA have decreased significantly while
measures of economic value o f output and employ­
ment have risen dramatically. Electricity and coal
use have grown at about the same rate as GNP,
whereas oil and gas use are nearly unchanged.

garages using surface coatings, solvents, and
cleaning fluids.
Emission sources in the U.S. and Midwest

The sources of three of the major pollut­
ants regulated under Title I on nonattainment
areas, Title IV on acid deposition control, and
under other regulations are shown in Figure 1a
for the U.S. and compared with the Midwest in
Figure 1b. For the purpose of this Figure, the
Midwest is defined as the East North Central
Census Division which includes Illinois, Wis­
consin, Indiana, Michigan, and Ohio, plus the
states of Iowa, Minnesota, and Missouri.
Figure 1 shows that the major sources are
different for each pollutant. Most of the S 02,
less than half of the NOx, and essentially none
of the VOCs are emitted by electric utilities.


Highway transportation emits NOx
and reactive hydrocarbons (VOCs)
Share of emissions and economic activity
but not SO,, although in the past,
in the Midwest in 1985
diesel ships on the Great Lakes
Pollutant or
Share in M id w e s t by sector (%)
have often used high sulfur oil. A
significant amount of VOC emis­
and other
U tility
sions comes from wood burning,
which is included in Figure 1 in the
industrial and other category.
Table 1 presents the shares of
these pollutants in the Midwest and
Gross state
compares these shares with the
product (GSP) 21.5
share of economic activity in the
Midwest, that is, the 21.5 percent of
gross state product that was gener­
utilities. This is due to the Midwest’s heavy
ated in the Midwest in 1985. For example, 42.9
reliance on locally mined, high sulfur coal for
percent of SO, emissions from electric utilities
are emitted in the Midwest. The share of emis­
electricity generation.
sions in the Midwest approximately mirrors its
A cid rain c o n tro l o f c o a l-fire d e le c tric
share of economic activities except for the
u tilitie s
higher amount of SO, emitted from electric
The acid deposition title, Title IV, of the
1990 Clean Air Act Amendments is innovative
in its approach to environmental protection
policy; it creates a market incentive system
based on SO, “emission allowances.” An al­
lowance must be obtained for each ton of SO,
emitted, as described below. Once allocated by
the EPA, allowances can be traded among
companies or be reserved for future use or to
hedge against higher emission allowance pric­
es. Allowances are tradable between years, a
concept called “emission banking.” A two
phase approach in Title IV is also innovative, as
are the use of incentives to encourage flue gas
desulfurization (FGD), also known as scrub­
bing, and the adoption of clean coal technology
The acid rain title is scheduled to cut sulfur
dioxide (SO,) emissions from electric utility
power plants from 16 million tons in 1985 (see
Figure 1) to about 9 million tons in the year
2000. The ambient air quality standards for
SO, had already reduced SO, emissions from
their peak national level of 29 million tons for
all sectors in 1977 to about 23 million tons in
1990 (see Figure 1 in Box 2).2 However, the
previous standards did not achieve the level of
reduction that was sought for acid deposition
control. If the high emitting power plants,
many of which bum high sulfur midwestem
coal, were to retire at age 30, then SO, emis­
sions would rapidly decline in the 1990s and
perhaps no Title IV would have been necessary,




because the New Source Performance Stan­
dards (NSPS) regulations would have assured
very low emission rates for replacement units.
However, the general trend in the electric utili­
ty industry has been to refurbish, life extend, or
even repower existing power plants with new
clean coal technology (CCT) combustors [see
U.S. Department of Energy (1989)]. The
choice to maintain the existing higher emitting
plants rather than retire and replace them was
the result of a desire to avoid more stringent
regulation of new sources as well as the diffi­
culties which have beset the utility industry,
such as lower than expected load growth, ex­
cess capacity, prudence reviews by state public
utility commissions, the high cost of capital,
rising construction costs, and high debt for
those utilities that embarked on large nuclear
energy programs. W ith no end in sight for the

cleanup of the existing high S 02 emitting
plants, many of which are in the Midwest, those
concerned with the acid rain issue wanted a
newly focused approach to S 02 controls.
Marketable emissions allowances

Emissions trading and banking provide
cost savings over mandatory technologies in
achieving long run environmental goals. These
gains are achieved because, rather than manda­
tory control technology, firms gain the flexibili­
ty to reduce pollution by choosing among the
cheapest technologies, alternative fuels, or
alternative schedules in lowering emissions.
Factors which will affect the least cost choice
are: plant design suitability for retrofit, land
availability, economies of scale in abatement
technology, access to alternative fuels including
differences in competition in transportation to
power plants at different locations, and alterna­
tive local air quality requirements.
Gains from trade can be illustrated with a
simple example. Suppose 100 tons of emission
reduction are needed to meet the environmental
objective. Suppose plant A has a marginal
abatement cost (MAC) of $300 per ton and
plant B has a MAC of $500 per ton. Under a
uniform rollback policy each plant would re­
duce emissions 50 tons at a total cost of
$40,000. However, suppose each plant is is­
sued 50 tons of tradable emission allowances.
Then plant A, which has a lower MAC, can sell
its allowances to plant B for, say, $400 per ton.
Plant A then reduces emissions 100 tons for
$30,000, and gains $20,000 in revenue from the


sale of allowances. Thus, the net cost to A of the
reduction in emission is $10,000. Plant B’s
$20,000 cost of purchasing allowances is less
than the $25,000 it would have had to pay to
reduce emissions by 50 tons. The total cost of
the 100 ton emission reduction is $30,000—
$10,000 for A and $20,000 for B—rather than
the $40,000 cost for the uniform rollback policy.
Under the new CAA, allowances are issued
gratis to existing polluting utility units based on
their “baseline” fuel use as measured by the
annual average of 1985, 1986, and 1987 British
thermal units (Btu) consumption. The basic
Phase I allowances are calculated as 2.5 lb. SO,
per 106Btu times the unit’s baseline, and the
basic Phase II allowances for the larger, dirtier
units are calculated as 1.2 lb. per 106Btu times
the unit’s baseline, though allowance allocations
are generally not larger than those required to
meet historical emission rates.
Table 2 shows the utility generating units
affected in Phase I. All but the smallest units are
affected in Phase II. The Table illustrates that
although the Midwest receives a disproportionate
share of the emission allowances, it also is likely
to have greater control costs because greater
emission reductions are needed in the Midwest to
meet the allocated allowances (alternatively, the
Midwest can buy or sell allowances and reduce
emissions less or more, respectively.) The esti­
mates in Table 2 indicate that in Phase I, the
Midwest receives 46 percent of the allowances
but still must reduce emissions 43 percent (based
on these allowances), compared with an average
U.S. reduction requirement of 33 percent. In
Phase II, the Midwest receives 28 percent of the
allowances and must reduce emissions 65 per­
cent from the 1989 level, compared with a 41
percent U.S. average Phase II required reduction.
Additional Phase I and II allowances are
also distributed based on other considerations. In
Phase I, a maximum of 3.5 million tons of SO,
allowances are to be awarded to units installing
scrubbers by 1997. These units can maintain
their existing emissions for the first two years of
Phase I and then after 1997, also receive ‘2-for1’ bonus allowances for emission reductions
beyond those required by the 1.2 lb. per 106Btu
As the CAA plays out over time, it is ex­
pected that utilities will in fact choose to bank
Phase I allowances for use in Phase II. This is
partially connected with the relative concentra-




Phase 1affected units1


Phase II2
% o f U.S.

% o f U.S.

Num ber of
generating units





























(th o u s a n d m eg a w a tts)

S02 emissions
(m illio n to n s )

(m illio n tons)

Reduction needed
to meet allowances
(p e rc e n tf

'Phase 1affected units are listed by name in Table A of the 1990 Amendments; these are the larger, higher emitting
units in the country.
2 few small units are not included.
Illin o is , Michigan, Wisconsin, Indiana, Ohio, Minnesota, Iowa, and Missouri.
"Required reduction based on no net trading of allowances outside the region.
SOURCE: Argonne U tility Simulation Model for 1989.

tion of coal production in the Midwest (see
Figure 2). The incentives for installing FGD
under the 1990 CAA Amendments along with
pressure by mining interests in the Midwestern
high sulfur coal producing states to scrub rather
than switch to low sulfur coal, will result in
banked allowances for use in Phase II. Another
reason is that, assuming low sulfur coal prices
are not bid up too high in Phase I, a unit may be
able to switch fuels and achieve an emissions
rate of less than 2.5 lb. per
106Btus. The banked emissions
will lower the cost of complying
with the more stringent rate effec­
tive in Phase II. For example, a
utility could scrub those of its units
that are the easiest to retrofit FGD
and then bum low or medium
sulfur coal in the remainder of its
units, thereby banking allowances
to cover any excess emissions in
Phase II.
Bonus allowances of 0.53
million tons per year are also pro­
vided in Phase II to be awarded to
units with low capacity factors in
the baseline years and to units


which would be otherwise penalized because
they were already low emitting units as of
1985. Any excess allowances can be traded or
used in conjunction with new growth in coalfired generation. Utilities which contract for
approved CCT may be awarded a four year
Phase II extension.
Figure 3 shows qualitatively the anticipat­
ed paths for emissions and allowances. Allow­
ance awards are the highest in 1995 and 1996


due to extensions for Phase I FGD.
The allowances in 1997-1999 are
Anticipated S 0 2emissions and allowances
based primarily on an allowed 2.5
lb. per 106Btu emission rate ap­
q u a n tity o f S O 2
plied to the baseline fuel use for
110 affected plants in Phase I as
defined in Table A of the 1990
CAA Amendments. The allowanc­
es in Phase II are based on 1.2 lb.
per 106Btu or less, as applicable,
with a four year extension for
approved CCT. Hence, as illustrat­
ed in Figure 3, allowances are
issued at a much higher rate early
in the program. Although actual
emissions will be decreasing over
time, they will also decrease at a
slower rate than allowances, which
thereby implies an accumulation of
allowance prices by facilitating the entry of
banked allowances in Phase I and the using up
speculators who are willing to bear some of the
or depletion of these banked allowances in
risks of risk averse utilities.
Phase II (see Figure 4). The time at which
Major uncertainties affecting allowance
banked allowances are eventually used up (that
prices include: future gas supply and deliveris, when the market regime switches to one of
ability, success of renewable energy, effective­
annual market clearing) is denoted by T* in
ness of demand side management programs
Figures 3, 4, and 5.
(DSM), recovery of the nuclear industry, CCT
Hedging risks
performance and future penetration, the extent
The market price of allowances is expected
of low sulfur coal reserves, future electricity
to rise steadily over the course of Phase I and II
demand growth, and regulatory risk. These
through the middle of the next decade (see
uncertainties all affect required coal-fired gen­
Figure 5). This is because (as illustrated in
eration during Phase I and Phase II and hence
Figure 4), there is expected to be an excess
they effect the demand for allowances [see
stock of allowances held and the only advantag­
Hanson (1991a and 1991b)]. The holding of
es to holding allowances instead of acquiring
allowances can be used to hedge against these
them in the future as needed would
be capital gains derived from an
allowance price expected to rise or
the holding of allowances for
hedging against uncertainty in the
escalation rate of allowance prices.
The actual time path of prices will
depend not only on technical eco­
nomic factors such as fuel switch­
ing costs, but also on the motiva­
tions of market participants. Risk
aversion provides a possible mo­
tive for electric utilities to bank
allowances, thereby increasing the
current price of allowances. But
forward contracts and futures mar­
kets for SO, allowances, such as
those proposed by the Chicago
Board of Trade, may also influence




Anticipated allowance price path with and
without utility risk aversion
Expected S 0 2
allowance price path

NOTE: T*=year banked allowances are used up.

uncertainties. By lowering costs and risks
everyone can gain, including rate payers and
utility shareholders.
Electric utilities may consider hoarding
more allowances than is prudent or installing
more scrubbers than is cost effective when
this behavior is sanctioned by their regulatory
bodies (public utility commissions). However,
if a formal market exists for allowances, it
may be more difficult for public utility com­
missions to make regulations that inhibit elec­
tric utilities from making least cost abatement
choices, since the existence of market prices
makes the alternatives clear to the public and
to all involved.

Abatement cost functions and
emission reduction

The extent of emission re­
ductions in Phase I and Phase II
will be a function of the market
price for allowances. A firm can
either reduce emissions using fuel
switching or scrubbing, or use its
allowances. The rule of thumb in
economics is that it is cost effec­
tive to reduce emissions up to the
point where marginal abatement
costs equal the price of allowanc­
es. For example, if the marginal
abatement cost is greater than the
price of allowances at the Phase
II basic emission rate of 1.2 lb.
per 106Btu, it would be cheaper
for the firm to increase emissions
and buy allowances. Figure 6
illustrates the total abatement costs for SO,
reduction by switching to lower sulfur coal or
retrofitting a scrubber for a typical Midwest
coal-fired power plant. As Figure 6 illustrates,
scrubbing is typically more economical at high­
er reduction percentages. The slope of the total
abatement cost curve is the marginal abatement
cost (MAC). The firm can observe current
allowance prices, PA, but it must forecast fu­
ture values for PA. The value of PA used by
the firm for planning purposes is the slope of
the tangent line (see Figure 6). As shown in the
Figure, when the price of allowances is low
(represented by the flatter line labeled PAa), the
solution for emission reduction is shown as
point (a) and fuel switching is used. When the
price of allowances is higher in Phase II (repre­
sented by the steeper line labeled PAb), the
solution is shown as point (b). In this case, the
firm finds it economical to install a scrubber
in Phase II.
Allowances and coal market price
path interaction

Interestingly, marginal abatement costs
(MAC) are, in theory, proportional to the low
sulfur coal price premium, the additional
amount paid for coal per unit reduction in sul­
fur content. Let PC(S) denote coal prices as a
function of sulfur, where we define S in terms
of the resulting lb. SO, per 106Btu emission
rate. The MAC is just the extra price paid for
lower sulfur coal from which it follows (adjust­
ing for a change in units):



(1) MAC = (-2,000 lb/ton) APC/AS.
Since the condition PA = MAC provides the
cost minimizing compliance strategy and emis­
sions reduction as a function of the allowance
price, PA, then
(2) PA =-2,000 APC/AS.
Therefore, market equilibrium low sulfur coal
prices are closely connected with market equi­
librium allowance prices. Bidding up allow­
ance prices is equivalent to bidding up the price
premium on low sulfur coal. Hence, observing
the sulfur price premiums in the coal market is
a proxy for emission allowance prices.
The o th e r title s o f th e 1 9 9 0
A m e n d m e n ts

The EPA has already issued many of the
detailed regulations implementing the various
titles of the 1990 CAA Amendments, but many
more regulations are still scheduled to be pro­
mulgated in the future. Interested parties are
encouraged to comment on notices of prelimi­
nary regulations. This section provides a little
more background on some of the other impor­
tant titles and issues in the new CAA.
Ozone nonattainment areas

Studies show that ozone damages materials
and plants, contributes to urban smog, and is
not healthy to breathe. Ozone is one of six
common pollutants for which National Ambient
Air Quality Standards (NAAQS) have been set
under the authority of the earlier 1970 Clean
Air Act. The other five common air pollutants
are sulfur dioxide (SO,), nitrogen oxides (NOx),
carbon monoxide (CO), particulate matter, and
lead. Under the CAA, the U.S. is divided into
Air Quality Control Regions, or airsheds, which
are monitored and are deemed to be either in
attainment or in nonattainment of the NAAQS
for each of the six air pollutants. Considerable
progress has been made in the last twenty years
in bringing nonattainment areas into compli­
ance with NAAQS for all of the six air pollut­
ants except ozone.
The 1990 CAA Amendments contain a
tough new program designed to bring ozone
nonattainment areas into attainment. In the
1990 Amendments, ozone nonattainment areas
are to be classified as either extreme, severe,
serious, moderate, or marginal. Los Angeles is


the only extreme nonattainment area. The
areas likely to be designated as severe in the
Midwest are the Chicago and Milwaukee met­
ropolitan statistical areas. The classification of
an ozone nonattainment area determines the
actions required under the 1990 CAA Amend­
ments and the schedule.
Each of the states is required to submit a
plan to bring nonattainment areas into compli­
ance over a scheduled number of years. These
are called State Implementation Plans (SIPs).
Any new sources resulting from economic
growth in a nonattainment area, or the replace­
ment or modification of existing sources, re­
quires New Source Review. The emission rates
for new sources in nonattainment areas are even
more stringent than the usual new source per­
formance standards. Further, “offsets” must be
obtained for new growth. An offset is a reduc­
tion in emissions, sometimes greater than l-for1, from another source. Various mobile source
controls also apply in nonattainment areas.
Prevention of significant deterioration

The areas already in attainment, that is the
clean air regions, are classified as Class I, II, or
III, with Class I areas being the most deserving
of clean air. National parks are all classified as
Class I areas because of their scenic beauty.
Air quality may be allowed to worsen in exist­
ing clean air areas, but only by a very small
increment, as set out by EPA regulations. This
is called the PSD increment. The State Imple­
mentation Plans (SIPs) are also required to
maintain existing air quality. New sources in
the region of a Class I area may require lower
emission rates than the usual NSPS.
Visibility and permitting

The Acid Deposition Title IV may not
reduce SO, sufficiently to achieve goals of
improved visibility in national parks, such as
the Grand Canyon in Arizona, Shenandoah
National Park in Virginia, and other scenic
areas. Title VIII of the 1990 Amendments
requires that the federal government undertake
a study to identify and evaluate possible sourc­
es of regional haze. Unfortunately, it appears
that, because sulfates are carried over long
distances rather than deposited locally, only a
small percentage of the contribution to visibili­
ty impairment comes from local sources.
Based on Argonne National Laboratory’s Ad­
vanced Statistical Trajectory Regional Air


Pollution model (ASTRAP), relative contribu­
tions arriving at Great Smoky Mountain Na­
tional Park by source state are illustrated in
Figure 7. The role and frequency with which
sulfates contribute to visibility impairment are
now being studied statistically. If more con­
trols are deemed to be needed, then consider­
ably greater SO, reduction costs will be in­
curred beyond the compliance costs already
under Title IV.
The reduction in visibility impairment
cross cuts several of the regulatory controls in
the CAA. The 1977 CAA Amendments set a
national goal for no man-made visibility im­
pairment, with reasonable progress to be made
toward this goal over time. The PSD program
can be used to enforce further controls on emis­
sion sources. Alternatively, air quality related
values, such as visibility and sensitive ecologi­
cal areas, can be protected through federal
intervention in the state permitting process.
Based on recent trends in rejecting new source
permits and new regulations promulgated by
the Department of the Interior, intervention by
the Federal Land Manager to require more
stringent emissions caps on existing sources is
expected under the new permit program of the
1990 CAA Amendments.
Industrial emissions of hazardous air pollutants

The 1990 Amendments include a list of
189 potentially toxic industrial chemicals to be
targeted for regulation. The EPA also has the
authority to add to the list or modify it.
A best technological approach has been
adopted to regulate air toxic releases under the


1990 CAA Amendments, in
contrast to the risk assessment
requirements under Section 112
of the original CAA, which
proved to be unworkable. The
technological approach is called
Maximum Achievable Control
Technology (MACT). However,
a firm may postpone the stringen­
cy of the MACT controls if it
opts for early compliance.
Hence, planning is required to
make the best decision because if
a firm waits and a stringent
MACT is required, the cost may
be much greater. After this pro­
gram is implemented, another
wave of regulation is possible.
Cost benefit studies will be commissioned by
the EPA to see if future controls are warranted.
Stratospheric ozone depletion and
global warming

Chlorofluorocarbons (CFCs) and similar
bromine compounds are destroying ozone in the
stratosphere and are also so-called “green­
house” gases. Recent scientific studies indicate
a more rapid thinning of the stratospheric ozone
layer than previously thought, which will, for
example, increase the ultraviolet radiation from
the sun that reaches the earth’s surface.1
Hence, more skin cancer cases are projected.
The Montreal Protocol on ozone depleting
substances calls for remedial steps to be taken
by industrial countries. Title VI of the 1990
Amendments complies with the Montreal Pro­
tocol which calls for controlling both the chem­
icals themselves and products containing the
chemicals. The production and sale of a list of
chemicals and yet to be determined substitutes
is to be regulated by the EPA. Hence, manufac­
turers will need to develop substitute chemicals
and products including chemicals for use in
automobile air conditioners and for cleaning
fluids for electronic and photographic equip­
The Administration’s position on control­
ling greenhouse gases is to reduce these emis­
sions when it is beneficial to do so based on
criteria other than the effect on global warning.
As a result, cost effective energy conservation
is also encouraged to reduce greenhouse gas


Burdens and challenges: a general
e q u ilib riu m m o d el

Many of the burdens to individual firms
complying with the new Clean Air Act Amend­
ments have been discussed here: increased
expenditures for emission abatement or for
related planning and administrative activities,
loss of flexibility in industrial facility opera­
tions, the need to monitor emission releases,
and the resulting burden of higher product
prices or lower profits or wage payments. But
the challenge facing industry is one that econo­
mists have trouble describing succinctly be­
cause it goes beyond the aforementioned bur­
dens. Emission control is just one of many
challenges facing management and labor, who
must also be concerned about marketing, prod­
uct quality, reliability, worker health and safe­
ty, labor productivity, cooperative manage­
ment, worker morale, supplier relationships,
new product development, and shareholder
profits. Industry must reduce emissions to
comply with new regulations and at the same
time increase product quality and lower costs.
How this can all be done is perhaps a topic in
management and organization. From a public
standpoint, analysis must focus on how product
prices and wage increases are impacted by an
environmental regulation. At least in the short
run, this reflects the notion that society can
enjoy and benefit from a cleaner environment
only at the expense of reduced income.
In this section, it is suggested that the ex­
tent of these costs on the macroeconomy will
depend heavily on the behavior of the real wage
rate, that is, a wage adjusted for its power to
purchase market goods. The economy may be
aptly characterized by sticky wages if workers
are reluctant to recognize that lower wages are
the cost of a cleaner environment and, for ex­
ample, continue to negotiate for constant pur­
chasing power. A falling real wage can partly
offset the abatement expenditures and full em­
ployment can be maintained. But if the real
wage is sticky in the downward direction, then
the economy will adjust to the new regulations
at a greater total cost. Moreover, in a dynamic
context, much of this magnified cost will fall
on profits which could lower investment spend­
ing and new product development. Reasons to
anticipate sticky nominal and/or real wages
have recently been assessed by Robert Gordon
(1990). Descriptions of economic behavior in
terms of dynamic wage price spirals have been


used in numerous analyses such as the DRI quar­
terly model of the U.S. economy. Recent empir­
ical evidence has been presented by Mehra
A simple model has been constructed to
illustrate these relationships based on microeco­
nomic principles [see, for example, Varian
(1984)]. For ease of exposition, the economy is
divided into two industry sectors: Sector X,
which creates pollution, and Sector Y which is
assumed not to pollute. To obtain numerical
results for illustration we assume that 1/3 of the
labor force is employed in Sector X and 2/3 in
Sector Y. The time horizon is short to medium
term which means that technology and produc­
tion facilities are fixed but that sectorial output
can be raised or lowered by hiring more or fewer
workers. A marginal product of labor is assumed
in which an n percent increase (decrease) in
labor gives rise to a 1/2 n percent increase (de­
crease) in sectoral output. This marginal product
of labor assumption defines the production possi­
bility frontier at full employment, as labor is
shifted from one sector into the other.
Under the environmental regulation, some
output is used for real abatement expenditures.
Abatement expenditures are assumed to be divid­
ed between heavy manufacturing equipment and
materials purchases from Sector X and service
and light equipment purchases from Sector Y.
These abatement purchases are represented as
intermediate goods described by interindustry
flows. Substituting between these two abatement
factors is allowed (that is, to minimize costs of
regulation to the firm) and the feasible substitu­
tions are described by a Cobb-Douglas technolo­
gy. Abatement costs are taken to rise more than
proportionally with reductions in emissions per
unit of output in Sector X. Total abatement
costs (TAC) using the price in Sector Y as the
numeraire, is given by:
(3) TAC = (P IP )X + Y ;
where P and P are the prices of goods X and Y
respectively, and Xa and Ya are the outputs of
sector X and Y respectively, sold to sector X
firms for pollution abatement. Hence, the sub­
script “a” refers to real resources used in emis­
sion control. It is often suggested that environ­
mental regulations create income and jobs and
hence presumably have positive impacts on the
economy. I agree with this position only in part.
It is true that abatement expenditures, modeled


here as TAC, also represent income earned by
some business and that the production of abate­
ment goods, Xa and Ya, will employ workers.
(In fact any output from sectors X or Y will
employ workers.) The reason, however, that
abatement expenditures (TAC) are a cost to the
economy is that the skilled workers and other
resources producing Xa and Y have an opportu­
nity cost; they could produce other valuable
output for society.
In a numerical example, we consider a new
regulatory program with TAC equal to 20,
which corresponds to the preliminary cost esti­
mate for the 1990 CAA Amendments cited
earlier of $20 billion. This is 0.42 percent of
national income, /, taken for illustration to be
(4) I/Py = (PJPJX + Y\
= (.8)2,000 + 3,200 = 4,800;
where X and Y are the original equilibrium
outputs assumed to be 2,000 and 3,200, respec­
tively (see Figure 8). The numerical values
shown for income, 4,800, and the relative price
ratio, 0.8, correspond to the consumer equilibri­
um point described by the consumers’ demand
curve. The price elasticity of demand is as­
sumed to be -0.5, so that a 10 percent increase
in the relative price of X, that is, PJPV reduces
the relative demand for X by 5 percent. The
equilibrium point is illustrated in Figure 8.

Another behavioral assumption of the

model is profit maximization. In addition,
firms in each sector are assumed to be price
takers. Firms hire labor and produce output up
to the point where the value of the marginal
product of labor (MPL) equals the wage rate,
W. For Sector Y we express this formally as:
(5) MPl (L) = W/Py .
In Sector X, expenditures for pollution
abatement (TAC) per unit of output must be
subtracted to obtain the net value marginal
product of labor:
(6) (PJP -TAC/X)MPl (LJ = W/P .
These three relationships: 1) the produc­
tion function, 2) profit maximization, and 3)
consumer equilibrium, define the model. It is a
general equilibrium model in which sectoral
prices, wages, outputs, and employment are all
determined endogenously. The idea is to com­
pare the initial equilibrium point of the model
(that is, without the environmental regulations)
to the new general equilibrium solution which
is reached subsequent to the environmental
The distribution of income between labor
and capital at the initial equilibrium can be
calculated from the above assumptions, yield­
(7) (IIPJ = WL/P + n/Pv .
We wish to examine changes in real in­
come, real labor income and real capital in­
come. The real values are the nominal values /,
WL, and k deflated by a price deflator consis­
tent with consumer preferences. The rate of
change in the price deflator is given by:
(8) dP/P =1/3 dP IP + 2/3 dP IP ,
where the appropriate weights 1/3, 2/3 have
been derived from the model.
The “general equilibrium” results from the
model are calculated for three policy scenarios
which are based on differing assumptions about
how flexibly the economy reacts to environ­
mental regulation: 1) full employment, in
which the wage rate, W /P, decreases in re­
sponse to environmental controls, 2) a constant
real wage based on the price deflator shown




Illustrated direct and total effects of environmental policy

change in
real income*

Change in real
labor income
D o lla rs

Change in
real profits

D o lla rs

D o lla rs

Full e m p lo y m e n t,
lo w e r w age


-2 0 .0


-14 .0

-0.4 2


-0.4 2


C o n sta nt W /Py


-2 3 .3


-16 .3

-0.4 9


-0 .4 8


C o n sta nt
real w age




-14 .0

-0.4 2

-2 6

-1.8 0



D o lla rs


Addition to

in c lu d e s direct abatement cost.

investment spending and delay the adoption of
above, and 3) an intermediate case in which
new competitive technology, thereby dampen­
W/Py is constant. In this intermediate case
output in Sector Y is unchanged but some un­
ing economic growth. In a world characterized
by stiff competition in most products, it may
employment will arise from reductions in out­
put for Sector X. However, because the price
not be possible for firms to simply raise prices
of good X increases due to new abatement
to partly offset abatement costs.
Some observations can be made regarding
expenditures, the constant value W/P will
actually represent a decreasing real wage based
regional impacts. It would appear that a region
on the price deflator. In the case of a constant
is better off if it produces more of good Y, such
as services or light industrial products whose
real wage, the result is even higher unemploy­
ment and a magnification of cost due to lost
output under full employment expands, than a
region which produces more of good X, which
economic output.
can be thought of as heavy manufacturing,
Using these assumptions and numerical
chemicals, refining, coal-fired power plants,
values under the constant real wage case in
mining, and materials processing.
which labor attempts to maintain its real in­
come, the loss in national income
is about double the loss in the full
employment case. Under the full
Illustration of the effects of regulation
employment case the loss in na­
under a constant real wage
tional income just equals the direct
abatement cost (see Table 3).
real output
Further, with a constant real wage,
0.4 percentage points are added to
the unemployment rate and real
profits decrease about 2 percent,
which represents an amplification
effect: a direct cost of $20 reduces
real profits by $26 and wages by
$14 so that the constant real wage
shifts the burden more onto capital
income (see Table 3 and Figure 9).
Of course, this framework
accounts only for a “snapshot”
estimate of costs to the economy.
real output
As a matter of conjecture, in a
1 Production point with full employment.
dynamic context, a more signifi­
2 Production point with constant Y output.
3 Production point with constant real wage.
cant burden would be expected to
4 Consumption point with constant real wage.
develop if lower profits reduce



There are also regional implications regard­
ing goods consumed versus goods produced.
Final goods consumed decline to the extent that
they are displaced by the production of abate­
ment goods. And even if the national level of X
does not decrease very much, there may be
regional reallocations of production as some of
the output from X goes into the abatement ex­
penditure activity. A regional goal to preserve
regional income in heavy manufacturing areas
like the Midwest will be to capture a large part
of the market share for pollution abatement sales
rather than importing these goods from other
regions or from other countries.
However, it should be noted that according
to the model, even if a region maintains its
share of abatement business in proportion to its
original output from sector X, total national
output from X is expected to decrease. A region
would have to gain share in order to maintain its
output. At the present time, the prospects for the
U.S. or its industrial regions gaining share would
be difficult in view of the strong market posi­
tions of other nations, particularly Germany and
Japan, in the abatement equipment market.
C onclusions

It is unlikely that the announced costs of
around $20 billion per year for the 1990 Clean
Air Act Amendments will actually come to pass.
Rather, costs are likely to be either higher or
lower depending on how the economy adjusts;
that is, either how it rises to the challenge, or on
the other hand, how the burdens become magni­
fied. In the best case scenario, industry success­
fully plans to meet the requirements and tech­
nology is developed or adopted to meet emis­
sions requirements so that with continued in­
creases in labor productivity and economic
growth and adaptation of thinking on the need to
control emissions, there may be little public
attention regarding an emissions control burden.

This optimistic scenario largely reflects previ­
ous experiences with environmental legislation;
many environmental programs have proven to
be less costly than originally feared.
The alternative, less positive scenario is
one where planning is difficult because regula­
tory requirements remain uncertain, partly as a
result of uncertainty as to environmental effects
and partly due to the unpredictable outcomes of
political wrangling over CAA implementations.
The permitting process for industrial sources is
long, detailed, inflexible, and uncertain. The
goals of reducing urban ozone and cleaning up
industrial waste and emissions remain as elu­
sive as they have been over the first twenty
years of the Clean Air Act’s history. U.S. in­
dustry falls behind in developing and marketing
new, more competitive, cleaner processes.
Corporate funds available for investment in
new or retrofitted facilities are low. In this
negative scenario of lower economic growth,
the distribution of income becomes more con­
tentious. Labor, not meeting its expectations
for real standard of living growth, attempts to
increase its real wage at a time when real envi­
ronmental control costs are rising. These mac­
roeconomic effects have been observed histori­
cally in the 1970s and 1980s, following the oil
price shocks of 1974 and 1979 and the rising
environmental clean up expenditures during
these decades [for example, see Wilcoxen and
Jorgenson (1990) and Hickman, Huntington,
and Sweeney (1987)]. This, as was shown in
our simple economic model, magnifies the
direct abatement costs and thereby gives rise to
additional lost output and slightly more unem­
ployment. The challenge is to promote com­
munication and cooperation among business
leaders, scientists, engineers, and labor so that
the worst case scenario involving high costs of
a cleaner environment can be mitigated.

'The titles and sections under the original 1970 CAA and
the 1977 Amendments are distinguished from the titles of
the 1990 Amendments. For example, Title III of the 1970
CAA deals with administrative and miscellaneous matters,
whereas Title III of the 1990 Amendments provides lan­
guage to replace Section 112 of the original CAA dealing
with hazardous air pollutants from industry (see Box 1 for a
listing of the major titles).

authority of Title I of the original 1970 CAA in order to
protect health and environmental values.
3Silver (1990), see Chapter 9.
4A paper describing the model in more detail is available
from the author.

2National Ambient Air Quality Standards have been set for
the six common pollutants shown in Box 2 under the




Fumento, M., “The hidden cost of regulation,”
Investor’s Business Daily, March 9, 1992.

gram as Amended in 1990, Morgan, Lewis &
Bockius, 1990.

Gordon, R.J., “What is new Keynesian eco­
nomics,” Journal of Economic Literature, Sep­
tember 1990, pp. 1115-1171.

Silver, C., One Earth, One Future: OutChanging Global Environment, National Acad­
emy of Science, 1990.

Hanson, D.A., “Forecasting the market for SO,
emission allowances under uncertainty,” pre­
sented at EPRI’s Eighth Electric Utility Fore­
casting Symposium, Baltimore, October 1991a.

U.S. Department of Energy, Clean Coal Tech­
nology: The New Coal Era, DOE/FE-0149,
November 1989.

_________________, “Two-period model of
emission abatement and allowance banking
under uncertainty,” Illinois Economic Associa­
tion, Chicago, October 1991b.
Hickman, B.G., H.G. Huntington, and J.L.
Sweeney, Macroeconomic Impacts of Energy
Shocks, North-Holland, 1987.
Kohout, E.J., et al., “Current emission trends
for nitrogen oxides, sulfur dioxide, and volatile
organic compounds by month and state: meth­
odology and results,” Argonne National Labo­
ratory Report, ANL/EAIS/TM-25, August
Mehra, Y.P., “Wage growth and the inflation
process: an empirical note,” The American
Economic Review, September 1991, pp. 931 937.
National Acid Precipitation Assessment Pro­
gram (NAPAP), 1990 Integrated Assessment
Report, November 1991.
Portney, P.R., “Policy watch: economics and
the Clean Air Act,” Journal of Economic Per­
spectives, Vol. 4, No. 4, Fall 1990, pp. 173-181.
Quarles J., and W.H. Lewis, Jr., The NEW
Clean Air Act: A Guide to the Clean Air Pro­


U.S. Environmental Protection Agency,
Clean Air Act Amendments: Cost Compari­
sons, Office of Air and Radiation, January
Clean Air Act Amendments of 1990, Summary
Materials, November 1990b.
______________________________, National
Air Pollutant Emission Estimates 1940 - 1990,
EPA-450/4-91-026, November 1991.
U. S. Congress, Senate, Project 88, Harness­
ing Market Forces to Protect our Environment:
Initiatives for the New President, sponsored by
Senator T.E. Wirth, Colorado, and Senator J.
Heinz, Pennsylvania, December 1988.
______________________________, Project
88, Round II - Incentives for Action: Designing
Market-Based Environmental Strategies, spon­
sored by Senator T.E. Wirth, Colorado, and
Senator J. Heinz, Pennsylvnia, May 1991.
Varian, H.R., Microeconomic Analysis, W.W.
Norton, second edition, 1984.
Wilcoxsen, P.J., and D.W. Jorgenson, “Envi­
ronmental regulation and U.S. economic
growth,” Rand Journal of Economics, Vol 21,
No. 22, 1990, pp. 314-340.


Producer services: trends
and prospects for the
Seventh D istrict

W illiam A . Testa

Economic development strate­
gists in the Seventh District
have tried to hitch their wag­
ons to the booming service
sector as a means of replacing
disappearing paychecks in manufacturing and
agriculture. But not all services are capable of
driving regional growth. Service sector jobs
satisfying local consumer demand, such as dry
cleaning and most retail sales, do not usually
generate additional personal income. Accord­
ing to the so-called “export base” theory, goods
and services sold afar generate income which,
in turn, finances local income and spending on
nonexport goods and services. In this way, the
portion of regional growth that is generated by
changes in external demand for the region’s
exports is usefully identified for purposes of
development policies. Accordingly, the re­
gion’s perspective correctly focuses on those
service sectors that are driven by external de­
mand—that is, so-called “export sales.”1 In
particular, producer services industries have
been associated both with strong growth of late,
and with external rather than local markets.
Producer services are services sold to firms
rather than to consumers and typically include
accounting, management consulting, financial
services, real estate, insurance, engineering,
architecture, and credit reporting.2 The produc­
er services sector represents one of the most
rapidly growing sectors as measured by the rate
of job growth (see Table 1). Moreover, its
growth has been consistently high over the past



two decades, growing at an annual rate of 4.5
percent per year from 1969 to 1979, and 4.8
percent from 1979 to 1989.
Given this robust growth, producer services
appear to be a rich target for local and regional
development policies. But the richness of the
target may be to little avail without a deeper
understanding of where and why producer ser­
vices are growing. Surely, not all regions will
be attractive to all producer services industries;
and not all industries will be attractive to re­
gions. What institutions and amenities are
attractive to producer services industries? Are
both urban and rural locales attractive? Do
producer services industries sell beyond the
boundaries of the local economy and, if so,
which industries do so?
In this article, I review current studies in
order to shed light on these issues. Studies are
reviewed which claim to measure a regional
economy’s propensity to sell services beyond
its own boundaries. Further, the article exam­
ines the tendency for producer services to favor
locating in more urbanized areas rather than in
less urbanized or rural areas, and how techno­
logical changes in personal computers and
telecommunications technology may be chang­
ing the locational tendencies of producer serW illiam A. Testa is a senior regional econom ist and
research officer at the Federal Reserve Bank of
Chicago. The author thanks David R. Allardice,
Andrew J. Krmenec, Richard H. M attoon, and Car­
olyn M cMullen for com ments, and gratefully ac­
knowledges the research assistance of Sonia
Chung and David D. Weiss.


question to a greater degree, or
because it produces the service
with more labor and less capital
Em ploym ent
Annual grow th
and land in comparison to other
regions. Alternatively, its topog­
th o u s a n d s
( ------- p e r c e n t -------- )
raphy or climate may require
greater local production and con­
Producer services
sumption of the service. For
Business services
example, pest control services for
buildings are in greater demand in
Legal services
hot humid climates such as Hous­
M em bership org.
ton or New Orleans, while snow
M iscellaneous*
removal services are demanded in
Com m unication
Buffalo. These problems aside,
Electric, gas, and sanitary
differences in service employment
Retail trade
across regions can be used to
Wholesale trade
determine which regions export
services, to a reasonable order of
The most commonly used
measure for service export has
All services (nongoods)
been a simple index of employ­
includes engineering, architecture, accounting, management, and
ment concentration which looks at
public relations services.
SOURCE: Bureau of Economic Analysis.
an industry’s share of total em­
ployment in a region. This share
is put into index form by dividing
vice firms. The final section summarizes the
by the industry’s share of total employment in
the nation:
findings and discusses the overall prospects for
producer services industries in the 1990s.
(1) Index = (E../E t o t a,iy. )' /v( £ .S/ £ total,U S'*
v 7
v IJ

Can service e xp o rts spur a
reg io n al econom y?

Two types of studies support the idea that
growth in services can independently spur a
region’s economy. Both do so by suggesting
that a significant proportion of a region’s ser­
vices are sold afar rather than locally, which
implies that services growth need not be de­
rived from the growth in local goods produc­
tion. The first set of studies looks at the pro­
portion of a region’s services employment
across industries. The observation that a re­
gion’s employment concentrates in a particular
service industry, that is, that there is surplus
labor employed in producing a service, sug­
gests that the region produces more than it
needs and therefore exports the surplus.3 As
pointed out by the authors themselves, this
method and its variants are subject to error in
their purported measurement. A region may
display greater employment concentration, not
because it exports the service afar, but because
its population prefers to consume the service in


for service industry i and region j. For exam­
ple, if advertising employment in Chicago
accounted for .5 percent of Chicago’s total
employment, but only .25 percent of total em­
ployment in the nation, then advertising in
Chicago would have an index number of 2. An
index number greater than 1 suggests that the
region produces a surplus in the service which
is exported elsewhere. Accordingly, an index
number of 1 would indicate little or no trade
while an index number less than 1 would sug­
gest that the region imports the service.
The idea of observing service specializa­
tion and export from employment concentration
has been extended to service production within
an urban hierarchy and, in particular, to the
question of whether some services specialize in
urban centers of roughly equal sizes and are
exported downward to smaller areas [Gilmer,
Keil, and Mack (1989)]. This technique has
also been used to look at services trade within
systems of large cities in a region in order to


determine whether particular urban areas spe­
cialize in particular services, apart from and in
addition to the export question [see Gilmer
The employment concentrations of a sam­
pling of producer service industries are dis­
played in Table 2 for the metropolitan areas in
Seventh District states. The strong propensity
for producer service firms to favor large metro­
politan areas is evident; producer services tend
to concentrate in the Chicago area, which ranks
near the top both regionally and nationally.
The largest metropolitan areas in the Seventh
District—Chicago, Detroit, Indianapolis, Des
Moines, and Milwaukee—display a tendency to
export services.5

While an urban hierarchy is evident in the
Seventh District, with services being exported
from urban centers to hinterlands within the re­
gion, other observations also suggest that metro­
politan areas specialize in particular services,
quite aside from the urban hierarchy. For exam­
ple, although Milwaukee is located only 90 miles
from Chicago, a city with more than 3 times as
many people, Milwaukee serves as an indepen­
dent purveyor and specialist in certain urban
services such as advertising, consumer credit
reporting, and accounting. Moreover, many
small metropolitan areas rank close to or above
the larger areas in particular services: Peoria and
Cedar Rapids in advertising, Lansing and South
Bend in consumer credit reporting, Sheboygan in


Index of employment concentration in business service industries
(Top ranked Seventh District MS As, 1987)



Chicago, IL PMSA
Cedar Rapids, IA MSA
Milwaukee, Wl PMSA
Peoria, IL MSA
Detroit, Ml PMSA
Elkhart-Goshen, IN MSA
Ann Arbor, Ml PMSA
Madison, Wl MSA
Waterloo-Cedar Falls, IA MSA
Kalamazoo, Ml MSA



Computer programming
and data processing



Ann Arbor, Ml PMSA
Des Moines, IA MSA
Madison, Wl MSA
Cedar Rapids, IA MSA
Lafayette-West Lafayette, IN MSA
Chicago, IL PMSA
Detroit, Ml PMSA
Janesville-Beloit, Wl MSA
Champaign-Urbana-Rantoul,IL MSA 1.06
Milwaukee, Wl PMSA
Engineering, architecture.
and surveying
Ann Arbor, Ml PMSA
Detroit, Ml PMSA
Sheboygan, Wl MSA
Cedar Rapids, IA MSA
Madison, Wl MSA
Green Bay, Wl MSA
Jackson, Ml MSA
Iowa City, IA MSA
Chicago, IL PMSA
Indianapolis, IN MSA




credit reporting







Lake County, IL PMSA
Indianapolis, IN MSA
Battle Creek, Ml MSA
Chicago, IL PMSA
Ann Arbor, Ml PMSA
Grand Rapids, Ml MSA
Cleveland, OH PMSA
Fort Wayne, IN MSA
Detroit, Ml PMSA
Green Bay, Wl MSA




Accounting, auditing.
and bookkeeping




Madison, Wl MSA
Chicago, IL PMSA
South Bend-Mishawaka, IN MSA
Des Moines, IA MSA
Grand Rapids, Ml MSA
Milwaukee, Wl PMSA
Aurora-Elgin, IL PMSA
Kalamazoo, Ml MSA
Indianapolis, IN MSA
Detroit, Ml PMSA




Des Moines, IA MSA
Lansing-East Lansing, Ml MSA
Chicago, IL PMSA
South Bend-Mishawaka, IN MSA
Milwaukee, Wl PMSA
Indianapolis, IN MSA
Fort Wayne, IN MSA
Green Bay, Wl MSA
Kankakee, IL MSA
Champaign-Urbana-Rantoul, IL MSA
Management and
public relations


engineering and architecture, Grand Rapids in
accounting, and Battle Creek in management
and public relations. Those smaller metropoli­
tan areas hosting major state universities such
as Ann Arbor, Madison, and Champaign-Urbana figure prominently as service exporters.
Computer programming, engineering, research,
and testing labs draw heavily on university
skilled labor and institutional capital.
Direct observation of service exports

A second group of analytical studies direct­
ly observes the sales transactions of service
firms as exported outside the region. These
studies have been conducted in particular re­
gions, with the assumption that the results could
be applied to other regions. The pioneering
work was conducted for the Puget Sound area
of Washington state by William B. Beyers,
Michael J Alvine, and Erik G. Johnsen (1985).
The authors interviewed 2,000 firms in the
services, utility, communication, FIRE (fi­
nance, insurance, and real estate), business, and
professional services industries. The central
finding was that a heretofore unrecognized
proportion of service firms (55 percent) export­
ed more than 10 percent of their sales outside
the area. All firms taken together exported
more than 36 percent of sales outside the Puget
Sound region, with healthy sales outside of the
state as well. Other findings indicated a wide
variation among industry types in the propensi­
ty to export services. Most individual indus­
tries tended to export between one-third and
two-thirds outside the region; real estate and
accounting tended to depend more heavily on
local sales while R&D labs and transport ser­
vices tended to export more.6
Additional studies have been conducted as
case studies of areas within the industrial Mid­
west [Goe (1990) and Porterfield and Pulver
(1991)]. Such studies have corroborated the
findings of the Beyers, et. al. study that produc­
er services are often exported. In addition, the
subsequent studies offer many additional in­
sights into the geography of producer services
sales such as the preferences of producer servic­
es export firms to locate in large metropolitan
areas, rural areas, or in small to medium size
metropolitan areas.
U rb an ten d en c ie s

Producer services have long displayed a
marked preference for locating in urbanized


areas. By one recent study, 93.7 percent of
producer service employment located in Metro­
politan Statistical Area (MSA) counties in the
U.S., and 65.5 percent located in the largest 39
MSAs [O hUallachain and Reid (1991)]. This
urban tendency also holds true with respect to
the scale of establishment in larger versus
smaller areas; that is, firm size increases with
the size of the MSA [Krmenec and Cohn
( 1991 )].7
The reasons for more urbanized concentra­
tion may include gains from larger scale than is
possible in urban locations. Demand for a
service such as public relations or environmen­
tal law can be met at lower cost when the ser­
vice is provided to a large number of local
clients. In addition, producer services inherent­
ly involve the transfer of information, either in
person or through electronic transmission. The
delivery of services through face-to-face com­
munication from a central place to a surround­
ing area economically provides such an ar­
rangement. Transportation from a central place
outward through a spoke-like travel grid can
minimize time. For example, recent studies
have illustrated the importance of air travel
access to service firms locating in large metro­
politan areas [Beyers, et. al. (1985)]. At the
same time, a transportation grid can also effi­
ciently bring in service customers for face-toface meetings. These may be corporate cus­
tomers, wholesale buyers, or people attending
conventions and similar meetings.
A second set of reasons for urban concen­
tration are the gains of close proximity enjoyed
by individual service industries, many of whom
sell to one another and who presumably can
shop for and/or deliver services at lower cost if
they are located in the same urban area. Re­
portedly, producer services often sell to other
service industries or the administrative arms of
goods producing industries in a region [Goe
Similar to the transportation benefits
achieved by manufacturing industries from
close proximity in years past (for example,
apparel and textiles in New York City or more
recently steel, machine tools, and autos in the
Great Lakes region), service transactions in­
volving face-to-face contact may also benefit
from location and interaction. Others have
emphasized that “higher order” or “information
intensive” service industries do benefit from
strategic location, but the benefits do not neces­


sarily derive from direct interaction [O hUallachain and Reid (1991)]. Rather, a shared
labor pool of creative and flexible workers may
enable such firms to solve unforeseen and com­
plex problems as they arise.
Rural prospects

The revolution in electronic communica­
tion and information transmission has stirred
the hopes and interests of rural areas in attract­
ing producer services firms. Indeed, advances
in satellite communications, fiber optics, fac­
simile machines, and microwave technology
have dramatically lowered costs and opened
new vistas for communications in more remote
locations. Anecdotal information such as Cit­
icorp’s siting of a credit card processing facility
in South Dakota has created expectations in
rural and smaller metropolitan areas of a decon­
centration of producer jobs. More general
evidence of the developments in producer jobs
is furnished by the recent study by O hUallachain and Reid, which reports that employ­
ment share of the nation’s producer services
(SIC 73 and 89) increased from 3.6 to 6.3 per­
cent from 1976 to 1986. Gains were also re­
ported for small and intermediate metropolitan
areas at the expense of the 39 largest MSAs.
Despite this interest and evidence, the
fundamental forces affecting the rural versus
urban location decision are highly ambiguous;
opportunities for rural development in producer
services industries have possibly been oversold.
While it is true that communications advances
can allow some services to be delivered from
less costly (and possibly more amenable) rural
locations, the opposite is also sometimes true;
communications advances also allow services
to originate from urban sites and be delivered to
rural locales more cheaply, as has been suggest­
ed by Kim, Conway, and Beyers (1990). These
authors report little or no deconcentration of
producer services down the urban hierarchy
from 1974 to 1984. In another study conducted
in rural Washington, the authors found that
information technologies created nine jobs
but eliminated eleven [Dillman, Beck, and
Some analysts also believe that many ser­
vice activities, especially those accounting for
rapid growth in producer services industries
such as management consulting and financial
analysis, are becoming more sophisticated and
skill intensive [Stanback and Noyelle (1982)].



But these types of activities—which involve
information analysis and intepretation—require
face-to-face meetings and interaction with
associated industries. That is, communications
are complementary to these activities rather
than substitutes. Accordingly, rural locations
will not tend to be a drawing card for such
Air transportation represents another com­
plement rather than substitute for “higher or­
der” service activities. Airline deregulation has
strengthened the producer services standing of
those large metropolitan areas hosting major
hub airports. Surveys attest to the importance
of frequent and expansive air connections for
both producer services and administrative es­
tablishments [see Testa (1992), Kim, Conway,
and Beyer (1990), and Drennan (1989)]. In
contrast, the near completion of the U.S. inter­
state highway system (along with supporting
road networks) has tended to deconcentrate
overland and distribution services.
Analysts also observe a widening spatial
“division of labor” or dichotomy arising from
these relative advantages for centrally rather
than rurally located service industries. Similar
to the branching of routinized manufacturing
production activities which has occurred in the
U.S. from North to South and from urban to
rural, service firms are unbundling the more
routinized information processing activities
such as data entry and claims processing to
remote areas. Service firms do this in order to
capitalize on lower cost labor where possible
while retaining skilled service functions in
larger metropolitan areas. Standardized and
routinized service activities, so-called “back
office” or “lower order” activities, are seen as
most amenable to innovations in communica­
tions technologies.
Information on the geography of services
by type of activity tends to be scarce. One
example has emerged from a recent survey of
mortgage servicing firms and thrifts in the
United States reported in American Banker.
The results of the survey suggest that rural and
small metropolitan areas have captured an
inordinate share of routinized service activity.
Grand Rapids, Michigan, and Gainesville,
Georgia, ranked among the top locations in
mortgage servicing. Troy, Michigan, Pasadena,
Texas, and Fargo, North Dakota, were among
top-ranking thrift locales.8


A promising research avenue to
identify a rural/urban spatial divi­
sion of labor has been recently
undertaken by Kassab and Porter­
field (1991). In examining business
service industries and distribution
industries in metropolitan and non­
metropolitan counties, the authors
identify workers by their occupa­
tion as categorized by low skill or
high skill using the Current Popula­
tion Survey. While results are
preliminary at this time, the authors
do find significant skill and occupa­
tional differences for the same
industries in urban versus rural
If accurate, this rural/urban
dichotomy between higher order
and lower order producer services activities and
occupations holds at least two potential draw­
backs for less urbanized areas. First, less
skilled jobs tend to be lower paying so that, as a
potential avenue for development, these jobs
may be less desirable in their overall economic
impact. The tendency has been for lower pay­
ing jobs to locate in less urbanized areas. Fig­
ure 1 shows average annual payroll per employ­
ee for four size classes of counties within the
Seventh District.1 For producer services in
aggregate, and for each individual category as
well, a distinct and marked difference in payroll
per employee can be seen for 1989. For servic­
es overall, compensation per em­
ployee in rural counties is approxi­
mately one-half that of the largest
class of MS As in 1989 (see Figure
1). Perhaps even more convincing,
a continuum of average payroll
from higher paying (in large coun­
ties) to lower paying (in smaller
counties) can be seen for each of
the subcategories comprising pro­
ducer services.
Over time, the same data
source suggests that the payroll
disparity between rural counties and
MSAs has become more pro­
nounced in the Seventh District (see
Figure 2). All of the subcategories
of producer services industries,
except auxiliaries, contributed to
the total effect in which, for aggre­
gate services, the ratio of average


payroll per employee for the largest MSAs to
the average for rural counties climbed from
1.56 in 1975 to 1.96 in 1989.
Previous research has been conducted on
payroll disparities between rural and urban
counties in the upper Midwest states of Minne­
sota, Wisconsin, Illinois, Michigan, and Iowa
by Porterfield and Pulver (1991). The results
are notable in that, according to their survey of
18 producer services industries, few systematic
differences in average annual payroll emerged
between metropolitan areas, nonmetropolitan
areas, and nonmetropolitan areas not adjacent
to metropolitan areas. One explanation for the


area economies, urban and rural alike, regardless
apparent conflict in findings is that the Porterof their skill base, must struggle to re-invent
field-Pulver study examined industries defined
themselves and to encourage a flow of new jobs
much more narrowly (for example, at the threeand activities to replace those that are fleeing. In
digit SIC code level).1 This suggests that,
this light, back office service jobs can be viewed
when industry mix is taken into account, pay­
as a challenge rather than as a problem. One
roll disparities may not be so profound. Either
possibility would be to institute programs to
way, many of the implications are the same for
upgrade worker skills in order to preserve exist­
policy purposes. Whether it is the higher pay­
ing producer service jobs or to develop new
ing industries or the higher paying activities of
any industry that prefer urban locales, or both,
industry activities.
lower paying and lower skill jobs in rural areas
Rural and urban e vid e n ce in th e
are the result.
S even th D is tric t
A second concern about rural areas special­
Within the Seventh District states of Illinois,
izing in back office operations is that standard­
Indiana, Michigan, Iowa, and Wisconsin, it ap­
ized jobs are often those that are the most tran­
pears that any change in the tendency for produc­
sitory. In contrast to nonstandard jobs involv­
er services to favor larger metropolitan areas is
ing interpretative skills and face-to-face interac­
evolutionary rather revolutionary (see Table 3).
tion, routinized jobs are most
amenable to automation, thereby
leading to job base shrinkage.
For example, there is a concern
Concentration index of employment
for the future of data entry jobs;
(Seventh District)
technological advances may
All producer
replace data entry workers with
establishm ents
data scanning devices. Even if
mechanization is not the cause
of worker displacement, low
> 1 m illion
skill jobs can be lost through
250,000 - 1 m illion
migration. As many rural areas
< 250,000
have discovered from their ex­
Rural counties
perience with branch plant man­
ufacturing jobs, the next stop in
Finance, insurance,
the geographic product cycle is
real estate
services (SIC 73)
often cheaper overseas locations.
Recent news reports document
the flight of routinized service
> 1 m illion
jobs from the U.S. to other coun­
250,000 - 1 m illion
tries. Insurance companies have
< 250,000
set up claims offices in Ireland;
Rural counties
airlines are processing tickets in
Barbados; and book publishers
services (SIC 8 9 )*
and data entry firms have moved
keyboarding activities to Asia.
But of course, this concern
does not mean that lower order
> 1 m illion
producer service jobs are unde­
250,000 - 1 m illion
sirable as targets for rural eco­
< 250,000
nomic development. The transi­
Rural counties
tory nature of employment is
hardly confined to routinized
‘ Includes engineering, architecture, accounting, management, and public
relations services.
producer service employment.
NOTE: Membership organizations (SIC 86) not shown.
In today’s global economy, it
SOURCE: U.S. Department of Commerce, Bureau of the Census, C ounty
Business Patterns.
may be easily argued that all





Concentration index of employment in producer services industries
(Selected MSAs, Seventh District)

Des Moines



M ilw aukee


All prod, services










































Business (73)






















Membership org.






















•Includes engineering, architecture, accounting, management, and public relations services.
SOURCE: U.S. Department of Commerce, Bureau of the Census, County Business Patterns.

As measured by employment, producer
services in aggregate remained some 25 percent
more concentrated in the largest metropolitan
areas (population greater than one million) in
1989. This concentration continued to diminish
throughout metropolitan areas of smaller size
until, for rural (nonMSA) counties, producer
services employment share was 50 percent below
the corresponding share of the five state region
A modest tendency for rural areas to become
less concentrated in producer services can also
be observed from 1974 to 1989, while small and
medium sized metropolitan areas tended to gain
concentration (see Table 3). Some of the gains
in small and medium sized area specialization
came at the expense of the largest tier of metro­
politan areas—those with populations greater
than one million. Detroit and Chicago both
reported a lower relative specialization in pro­
ducer services (see Table 4).
C onclusions and o u tlo o k

Producer services industries are becoming a
more important part of the regional economic
base. As producer services become increasingly
specialized, so too have regions become more
specialized in particular producer services. As a
result, services are often sold from afar and can
therefore correctly be considered as part of the
regional “export base.” Technological and
organizational advances in both communication
and personal travel, including fiber optics net­
works and mega-hub airports, have facilitated the
sale and delivery of specialized services from
particular regions.


The successful development of producer
services has become a much coveted prize for
regions. For one reason, overall growth in
producer services industries has been extremely
robust throughout the 1970s and 1980s. How­
ever, during the recent U.S. economic slow­
down, overall employment in producer services
also slowed owing to weakness in financial
and real estate industries.1 So too, some ana­
lysts have speculated that long awaited gains
in service sector productivity are now material­
izing (following capital investments during the
early 1980s). If so, these productivity advances
may be labor saving. But despite the recent
slowdown in employment growth, not all pro­
ducer services industries are pulling back.
For example, business services such as account­
ing and computer programming continue to
expand, and producer services remains an
active target of economic development in
many regions.
Not all regions have been successful in
gamering growth in producer services but rath­
er, these industries remain concentrated in large
urban areas. While existing evidence is not yet
conclusive, it appears that advances in telecom­
munications and personal travel have only
strengthened the hand of urban areas in hosting
producer services and subsequently delivering
producer services to peripheral areas. In the
Seventh District states, payroll premiums for
urban workers in producer services have appar­
ently increased in comparison to rural counties.
As a counter trend, remote areas have succeed­
ed in attracting back office or lowered skilled


producer services activities in many instances.
As a matter of policy, such a trend should not
necessarily be discouraged by rural areas. Such
jobs may be quite valuable to rural labor mar­

kets and may lead to higher skilled jobs through
either an active development policy such as
worker training, or a natural evolution of sevice
industry growth.

'The view that only goods and services that are sold afar
are growth generating is not strictly true. Rather, it has
been a useful paradigm in viewing the process of regional
development. However, there is surely a remainder of
growth and welfare which is internally generated. En­
hanced specialization and trade within regional boundaries
can also increase the region’s welfare. Secondly, greater
efficiency in providing local goods and services can in­
crease the region’s competitiveness and capacity to supply
exports to other regions so that, indirectly, related nontraded job activities are also “growth-generating.”
2The definition of which industries comprised producer
services varies from study to study. In fact, all of the
industries make sales to both firms and to household
consumers so that the classification “producer services” is
somewhat of a misnomer. For a review of industry tenden­
cies to sell to both household and to business sectors over
time, see Duchin (1988).
3See Mack and Keil (1986) for information on Indianapolis;
Groshen (1987) on Ohio and neighboring states; Gilmer,
Keil, and Mack (1989) on rural southeastern U.S.; Gilmer
(1990) on major Texas cities; and Austrian and Zlatoper on
Cleveland and large MSAs.
4Another possible reason for thinking that this measure is
misleading is that the classification of service industries
may be inexact. For example, a region may display an
index of 1.0 for the advertising industry, suggesting no
external trade in services. Nonetheless, the region may
actually specialize in a particular ty p e of advertising which
is exported while, in turn, another specific ty p e of advertis­
ing service is imported. The failure to reveal such trade and
specialization lies with the insufficiently disaggregated
classification scheme of the data.
5In constructing these indexes, we are comparing service
employment share in the region to the overall nation. It
should be noted that, because the overall region specializes
in goods production—especially manufacturing—the
metropolitan areas of the region will not rank as highly
among the nation’s metropolitan areas.

6These figures are drawn from Table III-11 of the study,
which refers to those 1,100 surveyed establishments report­
ing more than 10 percent export sales.
7See Goe (1990) and Porterfield and Pulver (1991). An
earlier study by Stephen M. Smith (1984) surveyed 350
nonmanufacturing firms in nonmetropolitan areas of
Wisconsin. The author reported 28.1 percent of sales
outside of the town boundaries.
8Data provided by ADP Data Service division, New York,
from its data base of the T h rift F in a n c ia l R e p o r t, as report­
ed in A m e ric a n B a n k e r, October 21, 1991.
9Kassab and Porterfield (1991).
10Data are drawn from C o u n ty B u s in e ss P a tte r n s , U.S.
Bureau of the Census. Payroll employment reflects both
part-time, full-time, and part-year employees. Employees
are recorded for a single week in March of the year report­
ed. Only counties for which no disclosure problem existed
were sampled. County classification reflects the population
size of the MSA in 1988, rather than the population size of
the county itself.
Recent evidence by Kassab and Porterfield (1991) also
finds sharp propensities for business service workers in
rural areas to be employed part-time (often involuntarily),
thereby lowering observed payrolls per employee.
"The Porterfield-Pulver survey (1991) covered only a
small subset of industries so that their results are not neces­
sarily in conflict with our own. Also, a methodological
difference is that the Porterfield-Pulver study reported from
data that was imputed or estimated for county areas. We
prefer to sample only from rural counties in which there
were no disclosure problems with the data.
12Strongin (1990).

Austrian, Ziona, and Thomas J. Zlatoper,
“The role of export services,” REI Review, Fall
1988, pp. 24-29.
Beyers, William B., Michael J. Alvine, and
Erik G. Johnsen, The Service Economy: Ex­
port of Services in the Central Puget Sound



Area, Central Puget Sound Economic Develop­
ment District, Seattle, 1985.
Dillman, Don A., Donald M. Beck, and John
Callen, “Rural barriers to job creation remain
even in today’s information age,” Rural Devel­
opment Perspectives, G: 1, 1989, pp. 21-27.


Drennan, Matthew P., “Information intensive
industries in metropolitan areas of the U.S.,”
Environment and Planning A, Vol. 21, 1989,
pp. 1603-1618.
Duchin, Faye, “Role of services in the U.S.
economy,” in Bruce R. Guile and James Brian
Quinn, eds., Technology in Services: Policies
for Growth, Trade and Employment, National
Academy Press, Washington D.C., 1988, pp.
Gilmer, Robert W., “Identifying service-sector
exports from major Texas cities,” Economic
Review, Federal Reserve Bank of Dallas, July
1990, pp. 1-15.
Gilmer, Robert W., Stanley R. Keil and
Richard S. Mack, “The service sector in a
hierarchy of rural places: potential for export
activity,” Land Economics, Vol. 65, No. 3,
August 1989, pp. 217-227.
Goe, W. R„ “Producer services, trade and the
social division of labor,” Regional Studies, Vol.
24, No. 4, 1990, pp. 327-342.
Groshen, Erica L., “Can services be a source
of export-led growth? Evidence from the Fourth
District,” Economic Review, Federal Reserve
Bank of Cleveland, No. 3, 1987, pp. 2-15.
Kassab, Cathy, and Shirley L. Porterfield,
“The spatial allocation of labor in producer
service industries,” paper presented at the Re­
gional Science Association Meetings, New
Orleans, November 8, 1991.
Keil, Stanley R., and Richard S. Mack,
“Identifying growth in the service sector,”
Growth and Change, Vol. 17, No. 4, 1986,
pp. 1-10.
Kirn, Thomas J., Richard S. Conway, Jr.,
and William B. Beyers, “Producer services
development and the role of telecommunica­
tions: a case study in rural Washington,”
Growth and Change, No. 4, 1990, pp. 33-50.
Krmenec, Andrew J., and Roger Cohn,
“Business services within a system of cities,”
paper presented at the 22nd annual meeting of


the Mid-continent Regional Science Associa­
tion, Chicago, May 31, 1991.
O hUallachain, Breandan, and Neil Reid,
“The location and growth of business and pro­
fessional services in American metropolitan
areas, 1976-86,” Annals of the Association of
American Geographers, Vol. 81, No. 2, 1991,
pp. 254-270.
Porterfield, Shirley L., and Glen C. Pulver,
“Service-producers: exports, impacts and loca­
tion,” International Regional Science Review,
Vol. 14. No. 1, 1991, pp. 1-14.
Smith, Stephen M., “Export orientation of
nonmanufacturing businesses in nonmetropoli­
tan communities,” American Journal of Agri­
cultural Economics, May 1984, pp. 145-155.
Stanback, Thomas M., Jr., and Thierry J.
Noyelle, Cities in Transition, Allanheld, Osmun
and Co., Totowa, N.J., 1982.
Strongin, Steven, “Services also slow down,”
Chicago Fed Letter, August 1990.
Testa, William A., “Job flight and the airline
industry: the economic impact of airports on
Chicago and other metro areas,” Regional Eco­
nomic Issues, Federal Reserve Bank of Chica­
go, WP-92-1, 1992.
Testa William A., and David D. Weiss, “Ser­
vice establishments of manufacturing compa­
nies, their role in understanding regional manu­
facturing output,” Regional Science Perspec­
tives, Vol. 21, No. 1, 1991, pp. 3-17.
U.S. Department of Commerce, Bureau of
the Census, 1981 Census of Service Industries,
U.S. G.P.O., Washington, D.C., 1990.

County Business Patterns (data base) Washing­
ton, D.C., various years.
U.S. Department of Commerce, Bureau of
Economic Analysis, Regional Employment
Data Base, Washington, D.C., various years.



Public Information Center
Federal Reserve Bank of Chicago
P.O. Box 834
Chicago, Illinois 60690-0834


D o N o t F orw ard
A d d ress C o rrec tio n Requested
R eturn P o stag e G uaranteed