EPA-600/5-74-012
May 1974
Socioeconomic Environmental Studies Series
The Economic Damages of
Air Pollution
I
55
\
5322
LU
CD
Washington Environmental Research Centei
Office of Research and Developmsnt
U.S. Environmental Protection Agency
Washington. D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and appli-
cation of environmental technology. Elimination of traditional grouping
was consciously planned to foster technology transfer and a maximum inter-
face in related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the SOCIOECONOMIC ENVIRONMENTAL STUDIES
series. This series includes research that will assist EPA in implement-
ing its environmental protection responsibilities. This includes examining
alternative approaches to environmental protection; supporting social and
economic research; identifying new pollution control needs and alternate
control strategies; and estimating direct social, physical, and economic
cost impacts of environmental pollution.
EPA REVIEW NOTICE
This report has been reviewed by the Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
-------
EPA-600/5-7k-012
May 197 b
THE ECONOMIC DAMAGES OF AIR POLLUTION
by
Thomas E. Waddell
Economic Analysis Branch
Washington Environmental Research Center
Washington, D.C.
Program Element 1AA004
Washington Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
For sale by the Superintendent of Document*, U.S. Government Printing Office, Washington, D.0.20402 - Price $1.95
-------
ABSTRACT
Air pollution is a problem because it endangers man's health and the
environment in which he lives. The information researched in this report
indicates that the cost of air pollution damage in 1970 (for measured
effects only) falls within a range of $6.1 to $18.5 billion, with a "best"
estimate of $12.3 billion. These estimates are based on: (1) a survey
of the literature on environmental economics; (2) a critical review of
completed studies that have attempted to estimate air pollution damages;
and (3) prevailing air quality levels in 1970. Such information on air
pollution damages provides policy-makers with some understanding of the
seriousness of the air pollution problem, and with some knowledge of
the potential benefits of abating air pollutant emissions.
A benefit-cost analytical framework for environmental decision-makinq is
outlined. The methods that have been or can be used to estimate the damages
of air pollution are identified. These methods are: (1) technical coeffi-
cients of production and consumption; (2) market studies; (3) opinion sur-
veys of air pollution sufferers; (4) litigation surveys: (5) political
expressions of social choice; and (6) the del phi method. The strengths
and weaknesses of each method are discussed.
The technical coefficients method is utilized in estimating the value of
air pollution damage to human health, to man-made materials, and to vege-
tation. The "best" estimates of damages for these effect categories for
1970 are $4.6 billion for health, $1.7 billion (adjusted for double-counting)
for materials, and $.2 billion for vegetation. A particular market study
method, the site differential or property value approach, yielded a "best"
•
damage estimate of $5.8 billion (adjusted for double-counting) for aesthetic
and soil ing-related costs. Economic losses associated with air pollution
effects on domestic animals and wildlife and the natural environment are
not estimated because of data limitations.
11
-------
Estimates of damages are allocated by major pollutant and source cate-
gories. The utility and limitations of gross damage estimates are dis-
cussed, and comparison with other such estimates is made. One of the
major informational gaps identified is the economic effects of automobile
and related air pollutants on human health and welfare.
111
-------
CONTENTS
Abstract
List of Figures
List of Tables
Acknowledgements
Sections
I Summary and Conclusions 1
II Introduction. . . .' 5
Clean Air: A Scarce Natural Resource 5
Why Should Government Act? 5
A Framework for Analysis 8
What Information 1s Needed? 8
A Hypothetical Example - Decision-Making 13
How Can Gross Damage Estimates be Used? . . . 19
III Methods of Assessing Air Pollution Damage 22
Overview 22
Individual Methods 23
Technical Coefficients of Production and
Consumption 23
Market Studies 25
Opinion Surveys of Air Pollution Sufferers 29
Litigation Surveys 32
Political Expressions of Social Choice 34
Delphi Method 34
What Method is Best? 35
IV Assessment of Non-Specific Air Pollution Damages 38
Overview of the Problem 38
Opinion Surveys 38
Property Value Studies 43
Ridker and Henning - St. Louis 43
Zerbe - Toronto 44
Jaksch and Stoevener - Toledo, Oregon 45
Anderson and Crocker T St: Louis, Kansas City,
Washington, D.C 45
Crocker - Chicago I . 47
Peckham - Philadelphia . 47
Spore - Pittsburgh . 48
Weiand - St. Louis 48
Crocker - Polk County, Florida 48
Flesh and Weddell - Southern California 49
Summary 49
National Estimate of Aesthetic and Soiling Costs 50
iv
-------
Page
V Assessment of Air Pollution Damage to Human Health. . . 56
Overview of the Problem 56
Individual Studies 58
Ridker - Morbidity and Mortality, Respiratory
Diseases 59
Riggan - Human Health and Motor Vehicle Pollution . 61
Jaksch - Morbidity and Air Pollution in Portland,
Oregon 62
EPA - Morbidity Selected Respiratory Diseases ... 63
Lave and Seskin - Mortality . 64
National Estimate of Health Losses 67
VI Assessment of Air Pollution Damage to Man-Made
Materials 73
Overview of the Problem 73
Individual Studies 73
Robbins - Electrical Contacts 73
ITT - Electrical Components 74
Salmon - General Materials 75
Mueller and Stickney - Rubber Products 78
Spence and Haynie - Paints 81
Fink - Corrosion 83
Gillette - So Effects on General Materials .... 87
Salvin - Dye-Fading 88
National Estimate Estimate of Materials Losses 90
VII Assessment of Air Pollution Damage to Vegetation. ... 92
Overview of Plant Surveys 92
Individual Studies 93
Middleton and Paul us - California, 1955 93
Lacasse - Pennsylvania, 1969 and 1970 93
Feliciano - New Jersey, 1971 95
Pell - New Jersey, 1972 96
Naegele - New England, 1971-72 96
Millecan - California, 1970 96
Benedict - Nationwide Survey, 1969 and 1971 .... 97
National Estimate of Plant Damages 99
VIII Assessment of the Effects of Air Pollution on
Aesthetic Properties 100
Overview of the Problem 100
Odors 101
Visibility 102
Works of Art 105
Ornamental Plantings 106
Conclusions 107
-------
IX Assessment of the Effects of Soiling 108
Overview of the Problem 108
Individual Studies 108
Mellon Institute - Pittsburgh Nuisance 108
Beaver Report - London Smog Episode 108
Michel son and Tourin - Household Costs 109
Ridker - Urban Soiling ..... Ill
Booz, Allen and Hamilton - Philadelphia Survey ... 112
Conclusions 115
X Effects of Air Pollution on Animals 117
Overview of the Problem 117
Domestic Animals 117
Wildlife 118
Conclusions 119
XI Effects of Air Pollution on the Natural
Environment 120
Overview of the Problem 120
Conclusions 122
XII Estimation and Allocation of National Gross Damages. . . 124
Gross Damage Estimation 124
Source Emissions 126
Assignment of Damage Costs by Pollutant and Source ... 128
XIII Discussion 132
Some Limitations of Gross Damage Estimates 132
Comparisons Among Gross Damage Estimates 134
Some Concluding Remarks 136
Recommendations 139
XIV References 140
XV Bibliography of Literature on the Assessment of Air
Pollution Damages 152
VI
-------
FIGURES
No. Page
1 Total Air Pollution Costs and Benefits 10
For Hypothetical Region 2
2 Marginal Air Pollution Costs and Benefits 14
For Hypothetical Region Z
3 Underestimation of Property Value Losses 54
4 Overestimation of Property Value Losses 54
5 Human Response to Pollutant Exposure 57
vn
-------
TABLES
No. Page
1 Cost of Cleaning Up After Boiler Malfunction 41
2 Amount of Money Respondents Would Pay Annually to 42
Reduce Air Pollution in Morgantown, West Virginia
3 Anderson-Crocker Property Value Regression Equations 46
4 Summary of Property Value Studies 51
5 Costs of Diseases Associated with Air Pollution 60
6 Health Cost of Air Pollution, 1963 64
7 Factors Affecting the Mortality Rate in U.S. Cities 65
8 Rough Estimates of Some Health Benefits that Can be 68
Realized by the Control of Sulfur Dioxide, Suspended
Sulfates and Suspended Particulates
9 Ranking of Material Economic Losses Caused by 77
Deterioration
10 Estimated Costs of Air Pollution Resistant Materials 79
11 Costs of Shortened Life of Rubber Products 80
12 Interval for Exterior Repainting as a Function of 82
Particulate Concentrations
13 Frequency for Exterior Wall Painting in Philadelphia 82
as a Function of Particulate Concentration
14 Economic Assessment of Air Pollution Deterioration of 84
Exterior Paints
15 Annual Cost of Corrosion by Air Pollution Damage of 86
External Metal Structures,.1970
16 Estimated Costs of Dye Fading in Textiles 89
17 Plant Losses Due to Air Pollution 99
18 Relationship of Cleaning and Maintenance Operations 114
to Air Particulate Levels
Vlll
-------
No. Page
19 National Estimates of Air Pollution Damages 125
(Unadjusted), 1970
20 Estimates of Nationwide Emissions, 1970 127
21 National Costs of Air Pollution Damage, by Pollutant 130
and Effect, 1970
22 National Costs of Pollution Damage, by Source and 131
Effect, 1970
IX
-------
ACKNOWLEDGEMENTS
The author wishes to acknowledge that some of the information contained
herein, has been extracted from the earlier report prepared by Larry B.
Barrett and myself. Thus, some of of credit, if any is due, should go
to my former coauthor.
Dr. William Watson of Resources for the Future, Inc., has been a substantial
contributor to this report, Section I in particular. His counsel and
participation throughout the course of the project have been especially
helpful and are appreciated.
Other significant input was provided by Dr. Donald Gillette, Dr. Fred Abel,
and Allen Basala, also of EPA, and by Professors Thomas Crocker of the
University of California, Riverside and Robert Anderson of the Pennsylvania
State University. Appreciation for their contribution is also expressed
to Dr. William Nelson, Dr. Allen Heagle, Fred Haynie, and Frank Blair, all
of EPA. All of the above are, of course, exculpated from any errors that
remain.
-------
SECTION I
SUMMARY AND CONCLUSIONS
The cost of air pollution damage in the United States in 1970 is estimated
to fall within the range of $6.1 billion and $18.5 billion. The "best"
estimate for measured effects for that year is determined to be $12.3
billion. These estimates are based on: (1) a survey of the literature
on environmental economics; (2) an extrapolation of studies that have
attempted to estimate air pollution damages and that passed a critical
review; and (3) prevailing air quality levels in 1970.
An evaluation is also made of the methods that can be employed to estimate
the damages of air pollution. These methods are: (1) technical coefficients
of production and consumption; (2) market studies; (3) opinion surveys of
air pollution sufferers; (4) litigation surveys; (5) political expressions
of social choice; and (6) the delphi method. It is concluded from such a
review that some combination of the methods surveyed will ensure the most
accurate assessment of the economic damages resulting from air pollution
insults. Such damages, in turn, when properly translated, become the
benefits of abating air pollutant emissions.
It is shown in this report that only the technical coefficients and
market study approaches have been used with measurable success in assessing
the benefits of controlling air pollution. The technical coefficients
method was utilized in estimating air pollution damages to human health,
man-made materials, and vegetation. The "best" (unadjusted) estimates for
these effect categories for 1970 are $4.6 billion for health, $2.2 billion
for materials, and $.2 billion for vegetation, and total to $7.0 billion.
A market study method, the site value differential or property value
approach, yielded a "best" (unadjusted) estimate of $5.9 billion. This
figure represents the value in 1970 of the negative insults of air pol-
lution that are capitalized in the residential, urban property market.
It is argued in this report that capitalized in this estimate are primarily
those costs associated with aesthetics and household soiling.
1
-------
Since it is likely that there is some overlap in the $7.0 billion and
$5.9 billion estimates, they can be considered additive only with minor
adjustments. By making such adjustments, any double-counting will be
minimized. With such adjustments, the $7.0 billion determined via the
technical coefficients method becomes $6.5 billion and the $5.9 billion
determined via the property value method becomes $5.8 billion.
The estimate of $12.3 billion for 1970 developed here, differs from the
1968 estimate of $16.1 billion developed by Barrett and Waddell because
of the following reasons: (1) the 1970 estimate is based on information
that wasn't available in the 1968 study; (2) the levels of air pollutants
being worked with in the 1970 study are generally lower than the levels
for those same pollutants in 1968; (3) a re-evaluation of the available
data has forced the modification of certain assumptions in this report.
The information surveyed in this report establishes that $12.3 billion
is the "best" estimate for 1970. Given the lack of conclusive information
to indicate that what is estimated in the $5.9 billion does not signifi-
cantly overlap with what is estimated by the $7.0 figure, the option is
left for the reader to use the $7.0 billion as a measure of air pollution
damages in 1970. While the evidence is far from clear, it is reasoned
that as interpreted in this study, the estimates determined via the site
differential and technical coefficients methods should be considered
additive, with only minor adjustment for obvious areas of overlap.
While it is known that air pollution causes losses of domestic animals
and wildlife, such losses were not quantified in this report because of
data limitations. Air pollution is also believed to cause pervasive
effects in the biosphere and on geophysical and social processes. These
effects are not without some economic consequences, but until the relation-
ships can be more clearly identified, large-system economic analysis is
somewhat premature.
-------
The cost estimates for aesthetics and soiling, health, materials, and
vegetation are distributed among the several pollutants considered respon-
sible for the effect. The pollutants considered are sulfur oxides (SO ),
X
particulates, and oxidants (Ov). Damages in 1970 attributable to SO are
X X
estimated to fall within the range $2.8 - $8.0 billion, with a "best"
estimate of $5.4 billion. Particulate damages are estimated to fall
between $2.7 and $8.9 billion, with a "best" estimate of $5.8 billion.
Oxidant-related damages are estimated to fall in the range $0.6 - $1.6
billion, with a "best" estimate of $1.1 billion. Every attempt is made
in this attribution process to identify where data deficiencies precluded
the generation of estimates. For example, health costs associated with
oxidant-related air pollutants are not estimated because of the lack of
data.
The same costs are distributed among sources on the basis of the relative
level of pollutant emissions. Damages of $6.1 billion in 1970 are attri-
buted to the general source category, fuel combustion in stationary sources.
Damages of $4.0 billion are attributed to industrial process losses, $1.1
billion to transportation, $0.4 billion to both the agricultural burning
and the miscellaneous categories and $0.3 billion to solid waste disposal.
Although estimates are obtained and presented, the reader is cautioned
concerning their use. The estimate of air pollution damages of $12.3
billion is not to be taken as absolute, but is to be considered as indi-
cative of the seriousness of the air pollution problem. The range of $6.1
to $18.5 represents the significant uncertainty in which the "best" esti-
mate of $12.3 billion should be couched. Limitations of gross damage
estimates are spelled out in greater detail in the paper. There is
certainly at least one significant limitation: many benefits to be gained
from air pollution control are not yet amenable to quantification in dollars
and cents. Thus, the decision framework set up in the paper is designed
to take this limitation into consideration.
-------
While these estimates provide some basic justification for environmental
policies and programs, aggregate point estimate offer little policy
information for setting environmental standards. The research iden-
-*
tified in this report needs to be extended to determine more accurate
dose-response relationships, i.e. damage functions, and the economic value
of the receptor response over a range of pollution levels. Such information
would be very useful for decision-making in matters relating to environ-
mental management.
-------
SECTION II
INTRODUCTION
CLEAN AIR: A SCARCE NATURAL RESOURCE
Only recently has clean air been perceived as a resource which is limited,
and sometimes scarce, so that society must become involved in deciding
how it is to be used. This is the fundamental argument: air pollution
damages human health and welfare. But it is also true that the abatement
of pollution will necessitate the use of natural, human, and capital
resource—all of which may be scarce.
In other terms, air pollution results in (external) costs that must be
borne by the community (i.e., increased medical and cleaning expenditures,
etc.); and in the same sense, the abatement of air pollution necessitates
the use of resources that could be used for other competing social goals
(improved education, urban renewal, etc.). These are the two significant
aspects of the environmental pollution problem. And it seems that it is
here that the question of "How clean is clean enough?" is relevant. In
a world where knowledge of the costs and benefits of air pollution control,
implementation costs, and income redistribution or burden considerations
are not known with precision, it is a very difficult task to determine
envi ronmental poli cy.
WHY SHOULD GOVERNMENT ACT?
The concept of "externalities" is becoming well recognized and understood.
Externalities of pollution are the adverse (negative externalities) or
favorable (positive externalities) effects of residuals produced by con-
sumption, production, and distribution activities. Typically, full pay-
ment for positive externalities or full compensation for negative external-
ities to the party affected is not made. These compensations are not
-------
required under existing economic and legal mechanisms. Society is
interested because it believes that under alternative economic and legal
arrangements, positive externalities could be increased and negative
externalities could be decreased until, ideally, net gains are maximized.
Air pollution represents a classic example of a negative externality. The
atmosphere is a common property resource providing two services: (1)
removal of waste residuals discharged by firms and individuals, and "life
support" for individuals, including "support11 of aesthetics and concern
of individuals for future generations; and (2) "support" of material
objects that individuals own. As long as there is no conflict between
the two services, then there is in effect no pollution. But when the two
are incompatible, negative externalities can exist.
There are many actions which are not externalities even though they possess
some of the characteristics of an externality. For example, goods production
and trade—positive actions—have been internalized by the market system.
Traffic congestion—a negative action—has been internalized to some
extent, by traffic lights and the willingness of individuals to obey them.
Assigning "responsibility" or "ownership" and devising enforcement mechanisms
for the purpose of achieving net gains are means of internalizing exter-
nalities. Unfortunately, comparable, "naturally evolved" institutional
frameworks for internalizing air pollution costs, frequently do not exist.
Such institutions do not exist because: (1) ownership of air cannot be
easily defined; (2) air "congestion", for large segments of the population,
has risen to the "peril" point (in a physical sense) only in recent times;
and (3) only in recent times have individuals perceived air pollution as
"perilous" (in both a physical and a metaphysical sence) relative to their
•
other wants.
-------
At first glance, one might think that traditional adaptive institutions,
like the market and common law legal systems, would eventually provide a
means of internalizing air pollution. But this is not likely to happen.
Because the atmosphere is a common property resource and because ownership
rights cannot be exclusively defined, air rights cannot, in a traditional
sense, be bought and sold in a market. Moreover, costs of common law
legal settlements are often prohibitive. Many individuals are affected
by air pollution. This pollution often comes from a number of sources.
It is extremely difficult and costly to reach agreement, within a group of
affected individuals, on the extent and nature of their air pollution damages
and then show in a court of law, the source of these damages.
On these bases, government regulation of air pollution is necessary and
desirable. It is desirable that these regulations be designed to repli-
cate the workings of traditional market institutions. In other words, an
efficient allocation of resources will be attained when polluters act as
if the costs their activities impose upon others are their own costs. A
government acting to ensure efficiency should establish mechanisms such
that net gains from pollution control are maximized. This government
action will, ideally and as a first approximation, require standards
set at a level where marginal costs of control including implementation
costs equal marginal control benefits. To achieve such standards, govern-
ment might provide regional planning and assign emission reductions (perhaps,
for example, through the use of effluent charges) so as to minimize the costs
of achieving established air quality standards in an affected region, and/or
establish mechanisms for minimizing implementation costs. These are costs,
for example, of setting, administering, and enforcing environmental standards,
In effect, a government operating in this fashion—setting and enforcing
environmental standards so as to maximize net gains—would be internalizing
air pollution externalities. Through this internalization government would
be ensuring an economically efficient allocation of scarce air resources.
-------
A FRAMEWORK FOR ANALYSIS
In a wider decision framework, however, income redistribution or burden
considerations must be coupled with efficiency considerations. Who pays
the cost of pollution control? Who benefits? Is the resulting income
redistribution a socially "fair" one? These latter equity considerations
might very well temper allocative considerations and result in smaller net
efficiency gains. Political processes, in trading off efficiency gains
against income redistribution or burden considerations would, in the main,
determine the outcome. Nonetheless, whatever the result, it is possible
in principle (as shown later) to determine the costs of various income
redistribution or burden outcomes in terms of foregone efficiency gains.
To sum up, the decision-maker setting environmental standards should be
fully aware of all of the consequence of his actions. For various alter-
native enforcement schemes and for the region under study, he should be
provided with comprehensive estimates of pollution control costs including
direct, indirect, and implementation costs, and with comprehensive estimates
of benefits and "burden" impacts. Some benefits can reasonably be measured
in dollars with confidence bands; other classes of benefits can only be
described in physical terms. But both costs and benefits (however quan-
tified) and "burden" impacts should be estimated over a range of pollution
control levels.
What Information is Needed?
Four lists of information—what economists call functions—are needed
for presentation to environmental managers. The first list or function
would display all of the appropriate costs of pollution control which would
be incurred in meeting a range of pollution control levels. This list
would include: (1) the direct costs of installing and operating pollution
control equipment, or the extra, direct costs of undertaking process
changes which result in less pollution; (2) indirect costs such as the
8
-------
differential costs of retraining and relocating workers when plants are
retired earlier than otherwise because of the enforcement of environ-
mental policies; and (3) costs of setting, administering, and enforcing
environmental policies—so-called implementation costs.
Actually, there would be a series of these cost functions, each corresponding
to a specific way of allocating emission reductions among the affected
emission sources. For example, reductions could be allocated proportionately
with each emission source being required to reduce its emissions by the same
percentage; or, alternatively, reductions could be allocated so that for each
overall level of control, specific sources incur the same cost of control
on the last unit of pollution controlled. Strictly in terms of direct,
indirect, and implementation costs, proportionate reduction--!.e., a
9Q% reduction in emissions by all polluters—is likely to be more expen-
sive than marginally allocated reduction which treats emitters individually.
The proportionate reduction approach may be chosen, however, if political
considerations outweigh the extra control costs which are incurred.
In order to further clarify these concepts of tradeoffs, two cost curves
or functions, Cl and C2, are illustrated in Figure 1. Control costs—
including direct, indirect, and implementation costs—are measured in
dollars-per-year ($/year) along the ordinate; degree of pollution control
(tons/year) is measured along the abscissa. As the control level increases,
o
empirical studies have shown that the costs of control are likely to rise
at an increasing rate, thus the upward bow of curves Cl and C2. Curve Cl
is an example of a cost function associated with a relatively expensive
way of allocating emission reductions (say, proportionate reduction) while
C2 exemplifies a cost function associated with a less costly way of allo-
cating emission reductions (say, marginally allocated reduction).
-------
KEY: Cl = Expensive Control Instrument
C2 = Inexpensive Control Instrument
DBU = Dollar Benefits (Upper Limits)
DB = Dollar.Benefits
DBL = Dollar Benefits (Lower Limits)
Cl
(O
*«»•
•t
to
•r-
q-
0)
CO
CO
o
o
Emission Control of Pollutant X, tons/year
Figure 1. Total Air Pollution Costs and Benefits
for Hypothetical Region. Z
10
-------
The second list of information or function needed for policy-making,
would display, over a range of pollution control levels, those benefits
resulting from pollution control which can reasonably be measured in
dollars—a benefit function. Such benefits might include, for example,
avoided out-of-pocket costs of soiling incurred by individuals living in
a polluted environment. But these out-of-pocket costs would not cover
all of the "true" costs in the soiling category. For example, individuals
probably adjust to a dirty environment, in part, by undertaking extra
cleaning and, in part, by relaxing their cleanliness standards. A dollar
quantification of extra out-of-pocket soiling costs would probably not
consider many of these "adjustment" costs. At present, there is no well-
defined method of measuring in dollars, "psychic costs" resulting from a
relaxation of cleanliness standards. To take another example, health
benefits of pollution control could be partially quantified by measuring
the health care costs which individuals would incur in a dirty environment.
However, out-of-pocket costs for health care are not necessarily adequate
measures of willingness to pay for such care. Therefore, given the current
state-of-art, it is extremely difficult to quantify dollar benefits in some
categories. Methods that have been used to quantify damages—which become
benefits with effective abatement—will be reviewed in Section III.
s
It is useful, however, to measure those dollar benefits that can be quan-
tified in a way which reflects the uncertainty of our benefit measures.
For example, empirical studies indicate that it is extremely difficult to
isolate the extra out-of-pocket health care costs associated with living
in a polluted environment. Pollution is only one of a number of factors
which influence health care expenditures. One solution is to use confi-
dence bands to reflect uncertainty in benefit measures by indicating upper
and lower benefit numbers within which there is, say, a 90% chance that
the true benefit number lies. These confidence bands can be based upon statis-
tical procedures, if, for example, benefit functions are quantified using
statistical procedures such as regression analysis. Even .in those cases
where benefit measures are judgmental, benefit analysts-should be asked
to provide upper and lower bounds on their benefit estimates.
11
-------
A dollar benefit curve or function, DB, has been drawn in Figure 1 to
clarify these concepts. Curve DB represents the measured benefits
in dollars over a range of control levels. It includes only benefits
which can reasonbly be measured in dollars such as avoided out-of-
pocket cleaning costs and avoided out-of-pocket health care costs. The
function DB could be derived by estimating avoided dollar costs (i.e.,
dollar benefits) in a cleaner air environment where the air environ-
ment is characterized by ambient air quality and then relating these
benefits to actual reductions in emissions from specific sources using
an atmospheric dispersion model. In practice it may be very difficult to
make this latter transformation. The curves DBL and DBU are lower and
upper confidence limits on dollar benefits, respectively. They are meant
to display uncertain knowledge of dollar benefits and are drawn to cover
some specified range of confidence, say, for example, 90%. This range
is to reflect uncertainty in dollar quantification of benefits relative to
ambient air quality and uncertainty in transforming ambient air quality
into specific emission reductions.
#
Most of our dollar benefit measurements, such as the ones reviewed in this
report, have been taken in relatively dirty environments employing devices
and methodologies which are tuned to these more severe conditions. Rela-
tively less is known about dollar benefits in a cleaner environment. Hence
the upper arid lower hypothetical confidence limits on measured dollar benefits,
DBU and DBL, have been drawn in Figure 1 with a widening spread to reflect
this increased uncertainty. The bowed-over shapes of the dollar benefit
functions reflect the assumption that benefits from increasing levels of
pollution control, increase at a decreasing rate.
The third list of information needed for the setting of standards is a
tabulation of all of the benefits from pollution control which cannot
reasonably be measured in dollars. The previously mentioned psychic benefits
from improved health and higher cleanliness standards, are examples. These
non-dollar benefits should be fully described in physical terms over a
range of control levels. Information should be provided on the numbers
and characteristics of the human, plant, and animal populations and inanimate
objects which are impacted by these non-dollar benefits.
12
-------
The fourth list of information needed for policy-making is a description
of income redistribution or burden impacts. Who are the gainers and who
are the losers? Who are affected, and to what degree, by residual pollu-
tants after standards are implemented? This information should be provided
over a range of pollution control levels.
All of this cost,-benefit and "burden" information should be related to
specific pollutants (or groups of pollutants when effects are synergistic)
and to specific regions. Weather conditions, topography, climate, the mix
of emission sources, and sensitivity of the exposed population vary over
time and from location to location. Furthermore, income redistribution or
burden considerations may be important for particular regions and for spe-
cific sources of particular pollutants and may point to politically attractive
enforcement schemes for these regions and these pollutants. In view of
these temporal and spatial considerations, cost-benefit analysis can be done
for different time periods on a regional basis, either for individual
pollutants or mixes of pollutants. For example, in dealing with mixes of
pollutants, cost savings on control systems and reduction of damage function
problems are possible by solving problems of individual (but related) pol-
lutants with a package approach.
A Hypothetical Example - Decision-Making
A hypothetical example should aid in putting into better perspective the
previous discussion on what information is needed by the decision-maker.
Let's now assume: (1) that a decision-maker is interested in controlling
pollutant X in region Z; (2) that relevant estimates of total dollar costs,
and total benefits (dollar and non-dollar), and "burden" impacts for all
politically acceptable alternative enforcement schemes (say, there are only
two: proportionate reduction and marginally allocated reduction) are
o
available ; and (3) that this information (except for non-dollar benefits and
"burden" impacts) is summarized in the cost and benefit curves drawn in
Figure 1. The decision-maker has responsibility for establishing an ambient
air quality level for pollutant X and would like to know the cost-benefit
implications of a range of levels.
13
-------
KEY: MCI = Marginal Control Cost (Expensive)
MC2 = Marginal Control Cost (-Inexpensive)
MBU' = Marginal Benefit-Upper (Includes1
Dollar and Non-Dollar Benefits)
MBU = Marginal Benefit-Upper (Includes
Only Dollar Benefits)
MB = Marginal Benefit Function
MBL1 = Marginal Benefit-Lower (Includes
Dollar and Non-Dollar Benefits)
MBL = Marginal Benefit-Lower (Includes
Only Dollar Benefits)
0)
-------
A useful way in which to present dollar cost-benefit information to the
decision-maker is in the form of marginal cost and benefit curves. These
are drawn in Figure 2. The units on the vertical axis are dollars per
ton ($/ton); units on the horizontal axis are tons removed per year (tons/year),
Curves MB, MBU, and MBL are marginal benefit, and upper and lower (confidence)
marginal benefit curves—the first derivatives of curves DB, DBU, and DBL in
Figure 1--respectively. MCI is the marginal cost curve for proportionate
reduction and MC2 is the marginal cost curve for marginally allocated reduction-
the first derivatives of the Cl and C2 curves in Figure 1—respectively.
Implicit evaluation of non-dollar benefits must be introduced into the
analysis. Let's say that after information on the character and impacts
of non-dollar benefits is given to the decision-maker, he decides that these
non-dollar benefits are half again as large (in dollars) as dollar benefits
and that they have the same degree of uncertainty as dollar benefits. This
would increase MBL and MBU at every pollutant reduction level by 50%, pro-
viding the marginal benefit schedules MBU1 and MBL1 respectively (see
Figure 2). These marginal benefit schedules can be compared with marginal
cost schedules to determine the range of control levels within which mar-
ginal costs equal marginal benefits with a confidence level of 90%, i.e.,
if the hypothesis is true, there is only a ten percent probability of making
the mistake of rejecting it.
For example, consider MCI in Figure 2. Marginal costs equal the lower 90%
confidence level of marginal benefits at level A. At levels to the left
of A, MBL1 would be greater than MCI implying that additional benefits
can be gained by increasing the pollution control level. At levels to
the right of A, MBL1 is less than MCI implying that net benefits are
declining. Hence the best control level for maximizing net efficiency
gains (i.e., total benefits minus total costs) at the lower 90% confidence
level is to remove A tons of pollutant X per year. Similarly, removal
15
-------
level B would maximize net efficiency gains at the upper 90% confidence
level. Assuming that emission source controls are implemented propor-
tionately, this provides the control range AB within which the decision-
maker can be 90% confident that net efficiency gains are maximized. Range
CD is a comparable 90% confidence control range, assuming emission control
is achieved by having every source incur the same marginal cost of control.
As drawn in Figure 2, these hypothetical control ranges have relatively
wide spreads meaning that the decision-maker can have little confidence
that a specific control level within these ranges is, in fact, that level
A
where net efficiency gains are maximized.
Assume now that the decision-maker would like to set a level such that
there is roughly a 90% chance that marginal benefits from pollution con-
trol will be covered by marginal costs. We can say that such a decision-
maker has an aversion to risking excess pollution damage (relative to
efficient damage levels) and that he is, perhaps, averse to legal challenges
by environmental groups. Assume also that political realities favor pro-
portionate control of pollution sources. Under these circumstances, the
decision-maker would choose emission level B, promulgate the appropriate
ambient air quality standard, and implement this standard by proportionate
reduction in source emissions. Given a preference for marginally allocated
reduction of pollutant X, the same decision-maker would instead, choose
control level D and reap additional implicit efficiency gains of EB'D'F.
But since he does not do this, income redistribution or burden considerations
implied by using proportionate reduction must be at least equal to the
foregone efficiency gain, EB'D'F.
Similarly, a decision-maker who has an unwillingness to risk excess pol-
lution control costs (relative to efficient cost levels) would pick a
control level near the other end of the control range. If the decision-
maker wanted to set a level such that there were roughly a 90% chance that
marginal costs would be covered by marginal benefits (say he was averse,to
legal challenges by industries having to control pollution), he would choose
control level A, assuming proportionate reduction of emissions, or he would
choose control level C, assuming marginally allocated reduction of emissions.
16
-------
Choosing level A rather than C implies foregoing an implicit efficiency gain
of EA'C'F in exchange for the income redistribution and burden "gains"
implied by proportionate reduction. In other words, it is the relative
price structure the decision-maker himself faces that matters.
An alternative way of presenting the information summarized in Figure 2 would
be to: (1) consider, one-by-one, a number of control levels, each control
level coupled with one of the two ways of reducing source emissions; and
(2) characterize for each control level-source reduction pair the type of
decision-maker who would choose a particular control scheme. For example,
a cost-benefit analyst would describe a decision-maker, choosing control
level B with proportionate reduction, in this way:
1. This decision-maker wants marginal benefits to equal or exceed
marginal costs with a confidence level of 90%; he is likely to
be adverse to legal challenges by environmental groups.
2. This decision-maker values non-dollar marginal benefits half
again as much as dollar marginal benefits and believes that
dollar and non-dollar benefits have the same degree of un-
certainty.
3. This decision-maker is willing to forego an efficiency gain
of EB'D'F in exchange for the income redistribution and
burden "gains" implicit in proportionate reduction.
Information like this, for each control level reduction pair, could be
presented to an actual decision-maker who would then have to ask himself:
Which of these alternatives captures my (or my constituents') concerns
and my (or my constituents') preferred tradeoffs?
So far, dynamic considerations have been glossed over in this presen-
tation. Cost-benefit analysis should actually be carried out over some
planning horizon. For a particular region and for particular pollutant
17
-------
sources, costs of pollution control are likely to decline over time as
shifts are made from add-on control devices to process changes, and as
advances in technology provide new, less costly control options. Like-
wise, for a particular region and for particular pollutants, pollution
control benefits (i.e., avoided damages) might increase over time as
population grows and as individual willingness-to-pay increases as
incomes increase.
These considerations can be incorporated into cost-benefit analysis by
modelling regional development and by estimating costs and dollar benefits
over time and then discounting these to present values. This would pro-
vide discounted marginal cost and dollar benefit curves which could be
analyzed exactly as the marginal curves of Figure 2 have been analyzed.
Non-dollar benefits, as before, can only be,described in physical terms
although implicit values (relative to a chosen standard) can be assigned
to them. Dynamic cost-benefit analysis is needed because different regional
development policies will point to: different regional growth rates;
different associated mixes of environmental standards and impacts; and
different implementation times to meet such standards. These development
tradeoffs should be made apparent to the decision maker.
This description of policy-making is not meant to be an absolute analytical
framework. Events in the real world are highly uncertain; in many cases,
little is known about the beneficial and harmful effects associated with
pollution levels and about their "burden" impacts; it is not easy to
model regional development over time. Nonetheless, cost-benefit analysis
should at the very least, attempt to clearly spell out the implied and
direct values which are involved in choosinq alternative levels of environ-
mental quality.
18
-------
HOW CAN GROSS BENEFIT ESTIMATES BE USED?
An attempt is made in this paper to review methodologies for estimating
potential air pollution control benefits and to present a systematic tabu-
lation of existing benefit estimates for the U. S. Most of these estimates
generally relate to reductions in U. S. air pollution to levels required by
existing federal ambient air quality standards.
There is some danger that these estimates will be misinterpreted with
respect to making precise policy decisions. To prevent this, questions
about some of the numbers presented in this monograph will be posed and
then answered using the preceding policy-making framework as point of
reference. It is hoped that this excercise will clarify the meaning which
can be attached to these benefit estimates.
Estimates for the U. S. of air pollution damages which can reasonably be
measured in dollars range from a low of $6.1 billion to a high of $18.5
billion per year with a "best" estimate of $12.3 billion for 1970. Are
current national primary standards too stringest if the costs, for the
U. S. of meeting national primary standards, are $40 billion per year?
What if, instead, the costs are only $5 billion per year?
On the basis of the information provided, it is not possible to answer
these questions. Here are several reasons:
1. Only single alternative cost estimates and a single dollar benefit
estimate with high and low spreads are provided in this example. Economic
analysis of environmental tradeoffs requires cost and benefit functions.
In other words, it is necessary to find the range of control levels at
which marginal costs equal marginal benefits. Information on total costs
4 and dollar benefits for a single control level are not adquate to judge
whether or not costs and benefits are reasonably equal at the margin.
19
-------
2. No information is provided here on the characteristics and the distri-
bution of those benefits which cannot reasonably be measured in dollars.
These non-dollar benefits—such as psychic benefits—do exist and they
are likely to be quite large. Even if control costs are substantially
greater than dollar benefits, non-dollar benefits could be sufficiently
large at' the margin to justify quite stringent policies.
*
3. No information is probided on "burden" impacts. Who pays? Who bene-
fits? Who suffers residual damage, i.e. that damage remaining even after
desired levels are achieved? This information is needed to measure the
reasonableness of the "burden" tradeoffs implicit in national policies.
4. No information is provided on regional costs and benefits and regional
burdens. Without this it is impossible to determine whether or not the
costs exceed the benefits in any region at the nationally determined
levels.
5. No information is provided on pollutant-specific costs and benefits
and burdens. Without this, it is impossible to determine whether or not
pollutant tradeoffs, relative to nationally determined levels for
specific pollutants, are reasonable.
6. No information is provided on what is happening to regional costs and
benefits and burdens and to regional development over time. Without
this information, it is impossible to know whether or not nationally
determined levels are dynamically efficient.
If air pollution damages in the U.S. are estimated to be $12.3 billion for
1970, what is an appropriate use of this estimate? This estimate tells us
that in an aggregate sense, air pollution is a relatively serious problem
and that we should probably attempt to reduce air pollution emissions. The
level to which emissions should be reduced, however, can only be determined
by undertaking a series of quite complicated regional cost-benefit or trade-
off analyses, each in concept like the one described previously.
20
-------
There are several recent applications of gross estimates of air pollution
damage to cost-benefit analyses that are worthy of mention. One is the
application made in the Economics of Clean Air where it is shown that
the present government program of promulgating air quality standards is
justified on the basis of gross damage estimates generated by Barrett
and Waddell (1973). Another recent application is in the report prepared
for the Office of Science and Technology, Cumulative Regulatory Effects on
the Cost of Automotive Transportation (RECAT) ,in which it is concluded that
present estimates of air pollution damage (again, taken from the Barrett-
Waddell report) raise serious questions about the justification of the
stringency of the legislated mobile source emission standards. It is quite
evident that such gross damage estimates can only be used meaningfully when
the user has a thorough understanding of their limitations.
In summary, gross benefit estimates will be useful to environmental managers,
but only in a limited way. They will provide some measure of the seriousness
of the air pollution problem. Yet such estimates will not provide the environ-
mental manager with the kind of information that is needed to establish environ-
mental quality standards. If the manager is to consider realistically the
tradeoffs in setting such standards, information in the form of functions
that relate costs to varying levels of emission control and damages to
varying levels of air quality are needed. The damages-air pollution relation-
ships—the damage functions—can be constructed using one of several methods.
These methods will be discussed in the next section.
21
-------
SECTION III
METHODS OF ASSESSING AIR POLLUTION DAMAGE
OVERVIEW
As discussed earlier, it is difficult to relate economic damages to
varying levels of air quality. Such functions that relate economic damages
to varying levels of air quality—damage functions—can be viewed as
society's demand schedule for pollution abatement. The demand curve for
cleaner air is a schedule of what people would be willing to pay for
various levels of air quality if the world, except for air pollution,
were efficiently organized.
The various methods that economists use to estimate damages from air
8
pollution have been discussed to some degree or another by Kneese,
9 10 11 "• 12
Ridker, Crocker, Lave, and Anderson and Crocker. There are six
general methods that have been used with different degrees of success.
These methods are: (1) technical coefficients of production and con-
sumption; (2) market studies; (3) opinion surveys-of air pollution suf-
ferers; (4) litigation surveys; (5) political expressions of social choice;
and (6) the delphi method. The six methods are not necessarily mutually
exclusive, but each is distinct enough to justify individual treatment
in this section.
The strengths and limitations of each method, and how they have been
employed are also discussed in this section. Applications of the different
approaches will be discussed in detail in later sections.
22
-------
INDIVIDUAL METHODS
Technical Coefficients of Production and Consumption
Copious literature exists concerning the physical and biological effects
of air pollution upon artifacts and organisms. For example, there is
well documented evidence that: particulates and oxides of sulfur exacer-
bate respiratory diseases in humans (Lave and Seskin, 1970); oxidants
severely inhibit the growth rate and yield of citrus and grapes (Benedict
et al., 1971); and oxides of sulfur cause excessive corrosion of metals
(Fink, et al., 1971). In general, the method is developed by: (1) deriva-
tion from experimental data by the observations on objects in conditions
simulating their natural environment; (2) estimation of the physical or
biological damage function which relates damage to pollution levels; (3)
translation of the physical damage function into economic terms; and (4)
extrapolation of the function to the population, using appropriate coeffi-
cients, if an aggregate damage estimate is desired.
Because of the lack of adequate dose-response functions, a variation of
the basic method outlined above is followed. In what might be termed a
"damage factor approach," the investigator will estimate what proportion
of a damage category can be identified as being related to or caused by
air pollution. Then by applying this proportionality factor to the damage
category, estimates of air pollution damage can be determined. Good examples
of this damage factor approach are given in studies by Lave and Seskin (1970)
and Benedict, et al. (1971). These as well as other applications of this
method will be discussed in more detail in Sections V, VI, and VII.
In many cases, the magnitudes of these physical and biological damages can
be predicted with some degree of accuracy because the forms of damage under
restricted conditions are known. Attempts to translate these physical and
biological damages into meaningful economic relationships have been less
successful in identifying economic damages over a range of pollution expo-
sures. Success in this method has been obtained onlv within narrowly
23
-------
circumscribed limits. Why? Because controlled laboratory conditions
usually have little semblance to real world conditions. To minimize the
confounding effects of other causal factors in the real world, the normal
scientific method holds everything constant except one factor (in this
case, a single pollutant or mix of pollutants). For purposes of generating
damage estimates, the extrapolation of laboratory results to the true world
is risky. Such a process ignores possible nonconstant marginal products,
factor proportions, nonlinearities, jointness, etc.
13
Other problems are those of aggregation and substitution. Crocker has
argued that to obtain anything vaguely resembling a market estimate of
collective damages, some means of making individual receptors (i.e. those
who suffer damage) commensurate must be found; and then, only rarely will
the aggregation process involve a straightforward arithmetic summation over
all individual properties. Anderson and Crocker go on to say, "However, the
collection of receptors cannot simply be treated as some arithmetic sum
of individual receptors, for the prices of the substitution possibilities the
single receptor views as fixed are not necessarily fixed for the collection
14
of receptors." That is, the substitution of one input by an individual
will not normally affect relative prices, but if the same substitution is
carried out by all receptors, then relative input prices will often be
changed. The problems in employing the technical coefficients approach
are those of: (1) extrapolation from controlled research environments
to real world conditions; (2) aggregation of damages; and (3) enumeration
of the technically feasible and then the technically efficient possibilities
because of substitution.
Even given these limitations, this is the method that has been most widely
used. And given the adaptability of the method of focusing on a single
receptor and effect, the studies are quite amenable to the development
of gross damage estimates.
24
-------
Market Studies
In this approach, air pollution damages are measured through the explicit
use of market valuations. The consideration, here, is the impact of air
pollution dosages on human behavior as reflected in markets. This approach
completely circumvents the need to know the physical or biological damage
function—the basic dose-response relationship. The investigator applies
statistical tools and econometric models to market data to isolate the
incremental adverse effect of air pollution on a particular activity or
behavior as expressed in the market place.
One particular type of market study of interest is the indirect effect of
air pollution on expenditures for a particular product or activity. A
good example of such a study is one by Vars and Sorenson (1972). In their
study, they attempted to explain the relationship between air quality and
consumer behavior, or more specifically, the consumption of recreation-
related activities.
Another type of market study is the use of property values to estimate air
pollution damages. One of the significant features of air pollution is its
locational nature. Fortunately, then, there do exist markets in which the
services and/or disservices of air pollution can be measured. As Ridker
said, "If the land market were to work perfectly, the price of a plot of
land would equal the sum of the present discounted stream of benefits and
costs derivable from it...Since air pollution is specific to locations
and the supply of locations is fixed, there is less likelihood that the
negative effects of pollution can be significantly shifted onto other
markets. We should, therefore, expect to find the majority of effects
reflected in this market, and can measure them by observing associated
15
changes in property values."
25
-------
Thus, given that people are willing to pay to avoid the effects of air
pollution, property values and air pollution concentrations must vary
inversely. Measures of this relationship, obtained through common multi-
variate estimation techniques, should yield rough estimates of the air
pollution damages. Good examples of this approach are given in docu-
mentation by Ridker and Henning (1967), Anderson and Crocker (1970),
Peckham (1970), Crocker (1971), and Spore (1972).
The investigator will face a very significant problem in using the market
study approach: he must account for all the factors that explain consumer
preferences and behavior. Such an explanation is, of course, a monumental
task, both theoretically and empirically. And then, when robust statis-
tical relationships have been compiled, there is the difficulty of inter-
preting the causality and relative importance of those pollutant measures
accounted for in the study. The investigator must be sensitive to the
possibilities of spurious relationships.
With respect to the causality problem, if there is a high degree of inter-
correlation between two pollution measures, too much significance should
not be attached to the magnitude of the coefficients of the individual
pollution variables. Pollution tends to be a composite phenomenon. That
is, the presence of one pollutant is frequently a reliable hint that others
are also present. Thus, it is possible that the pollutants measured by
these variables are not the causative agents, but are simply surrogates for
others that are producing the undesirable effects.
A common criticism of the property value method is that for the method to
have validity, buyers must know that pollution differs at various sites.
Actually, buyers need only know that they prefer some properties to others,
and other things equal, are willing to pay more for the preferred properties,
If the non-preferred properties contain attributes or effects of pollution,
this is sufficient for a differential property value to result. The notion
of cause and effect thus rests wholly in the mind of the investigator, and
18
not necessarily in the mind of the property buyer.
26
-------
Obviously, the important question arises: "What effects of air pollution
are discounted and what effects are capitalized in the property market?"
In other words, to what extent does the property value estimator reflect
the true or complete damage cost of air pollution?
It is possible that this estimator may be biased if a good deal of air pol-
lution injury is so insidious as to escape consumer notice. Yet public
19
opinion surveys seem to indicate that such has not happened. Ayres
believes that the real estate market primarily reflects the tan-
gible, experiential aspects of pollution: more rapid deterioration and
extra cleaning and maintenance costs; the milder medical symptoms, such
20
as shortness of breath and smarting eyes; plus, smells and dirt.
Most investigators agree that costs associated with organoleptic effects
(including psychic) as well as soiling-caused cleaning and maintenance
21
expenditures are capitalized in this estimator. This assumption appears
consistent with the conclusion recorded by the Surgeon General's Ad Hoc
Task Group on Air Pollution Research Goals which states: "The aspects of
air pollution which are most apparent and of greatest personal concern to
the individual probably are irritation to the eyes, nose, and throat, mal-
odors, and the reduction of visibility. The pollutants responsible for
these effects are undesirable whether or not they cause long-range health
effects or economic losses, because they constitute an annoyance to people.
The nuisance aspects of these effects together with those related to soiling
give rise to the greatest number of complaints received by air pollution
authorities. There is no doubt that a person's well-being is eventually
affected by exposure to these sensory annoyances and that this may result
22
in economic loss."
27
-------
Another difficulty in using this approach was voiced by Freeman: no
general equilibrium model has yet been developed that is capable of pre-
00
dieting land or site values following some given change in air quality.
He argues further that "...empirical studies of land values and air pollution
should await the formulation of general models from which empirically testable
hypotheses can be deduced. Until such models are formulated and tested,
empirical land-value studies will make little or no contribution to our
24
knowledge of the benefits of air pollution abatement." Anderson and
Crocker argue that a general equilibrium model does exist in the form of
imer
26
Or
an assignment model. Furthermore, the model has been subjected to a test
of sorts.'
07
In fact, there seems to be embodied in some work by Strotz an equilibrium
model acceptable to Freeman and at the same time, quite consistent with
Anderson and Crocker's hypotheses. The model appears to be operational
though it has not yet been fully tested. It can be termed a general equil-
28
ibrium model in a pure exchange setting.
Even so, Anderson and Crocker conclude that a partial equilibrium model,
designed to explain the differentials in site values that exist at a given
point in time, can be used to predict the change in the value of a repre-
sentative site that would accompany a change in air quality, other things
being equal. The implied damage cost estimate, properly interpreted as the
marginal capitalized loss due to air pollution, does provide valuable infor-
mation concerning the nature of air pollution damages being suffered.
Then comes the inevitable question: "What portion, then, of pollution damages
are measured by the property value estimator?" Theory would state that if
all consumers do not regard the two sites as perfect substitutes for each
other when each site has equal air pollution dosages, then some air pollution
damages will be registered in other durable assets and losses in consumers'
surplus. Property value differentials, then, can be employed to obtain a
lower bound on air pollution damages, and, if the sites in question have
rather homogenous characteristics, their differential values represent all
29
or nearly all damages. Spore states that at a minimum, since people
28
-------
exposed to pollution dosages will relocate only if the pollution costs they
can avoid plus the costs of moving are greater than the costs of using
some alternative less-polluted site, then the costs of this adjustment
(moving) will be reflected in site-value differentials.
Another type of market study has been identified as the compensating income
approach. That is, people who live in relatively dirty environments are,
on the average, compensated by relatively higher incomes. Only little
31
empirical work has been undertaken in this area.
Opinion Surveys of Air Pollution Sufferers
In an era when Gallup and Harris polls are as commonplace as interest in
the constellations, it seems quite appropriate (and popular) for those who
have responsibility in making decisions about environmental management, to
be concerned about public opinion. And indeed, this approach is closest to
the classical economic approach in that it focuses on estimating utility and
demand functions. For example, a recent opinion survey by Opatow Associates
32
attempted to measure the public's awareness and reaction to pollution.
Of particular interest was the extent to which concern by the public about
environmental pollution affected consumption patterns. And, the effects of
pollution on consumption are what economists want to measure.
In a November 1970 popularity survey, a nationwide poll conducted by Harris
showed that "pollution" ranked as the most serious problem facing many com-
33
munities. In a December 1971 Gallup poll, 52% of the people questioned
expressed a "deep concern" about environmental pollution. Such techniques
have also delved into the economic aspects of the problem. In the same
Gallup poll, 8% of the respondents said that they would be willing to pay
$100 or more per year in added taxes "to improve our natural surroundings,"
and 46% said they would be willing to pay only an extra $10 or less per
year.
29
-------
Investigators employing this method, have attempted to ascertain what
people do and do not perceive as air pollution effects, distinct from
whether or not they know the cause of the effects.36 This distinction
is important, for as with the property value method--contrary to the con-
fusion on the subject—in order to determine air pollution effects, it
37
doesn't matter whether people recognize the cause of these effects.
The notion of cause and effect need rest only in the mind of the investi-
gator.
If it can be assumed that people know explicitly the effects of air pol-
lution, then the objective is to elicit complete information from them in
a way that would dissuade untruthful responses. It is well known that
sample polling questions like "what would you be willing to pay to avoid
(or gain) so and so," often yield misleading answers. Since every person
questioned actually pays nothing to have his opinion recorded, he can
respond by making extreme statements in the hope of indirectly influencing
policy.
Also important in interviewing is the "free rider" aspect. Air pollution
control can involve a "free rider" problem because air pollution is indi-
visible and pervasive in nature and moves about freely. If the respondent
feels that the sum to be collected is large, he will name an arbitrarily
low figure. This is the conventional problem of public goods: the
interviewee reasons that even if he doesn't pay, he will be able to enjoy
something others are paying for. He doesn't want to pay for abating pollution
when it will benefit everyone.
Finally, there is the possibility that -a respondent might not understand
fully the consequences of air pollution on his health, for example. Again,
in the case of health, one might be unable or unwilling to think about such
consequences in purely economic terms.
30
-------
In spite of many of these problems, the opinion survey approach does have
its usefulness. Information on a sufferer's preception of air pollution and
his attitudes toward it can be obtained by the use of questionnaire inter-
view studies. Interviews can also provide the investigator with the suf-
ferer's understanding of the type, nature, and extent of air pollution
effects. To the extent that this knowledge is used as a basis to improve
sufferer information so that he will make more complete adjustments, the air
pollution damage function will be changed. Findings from studies employing
the opinion survey approach are quite sketchy. These kinds of studies will
be reviewed primarily in Sections IV and VIII.
Opinion surveys have shown particular usefulness in understanding: (1) how
attitudes about air pollution are formed and then affected by changes in air
quality; and (2) what people do and do not perceive as air pollution effects.
This method can also provide some insight into what people might be willing
to pay for improvement in the air environment, or perhaps, what their demand
might be for the reduced risk of experiencing certain adverse effects.
Obviously, concern over the environment and individuals' ability to pay are
important factors in determining willingness to pay for the abatement of
air pollution. It can be shown with conventional economic theory that
given one's knowledge, he purchases that much clean air to where the benefit
of the last increment purchased equals the cost of abating by that last incre-
ment. By acquiring more knowledge (at a cost, of course) of the effects of
air pollution, an individual's willingness to pay for different levels of
fi«
air quality (i.e. pollution control) would probably change. Given the diffi-
.'is
culty of measuring people's knowledge, it is likely that other measures will
*j
have to be used in conjunction with the interview method to determine the
demand schedule for a cleaner environment.
The willingness-to-pay, if correctly determined, indicates the demand for
a cleaner environment. However, it does not include the damages (say to
health) that accrue to persons who cannot, or, because of low income, are
unwilling to pay anything" to avoid the damages. Here, the damage to an
affluent person would be valued more than the same damage to a poor person.
31
-------
It seems that one particular area where this method might-prove very useful
is in understanding the aesthetic or psychological effects of air pollution.
Ridker's (1967) attempt at reaching this understanding indicates some pro-
mise. And as the general public becomes more knowledgeable about the effects
of air pollution (and this can be determined through the use of the inter-
view method), then this approach will become even more useful in under-
38
standing what people are willing to pay for varying levels of air quality.
Litigation Surveys
By 1969, after many attempts at estimating pollution costs, it was sensed:
that personal opinion polls often did not yield truthful responses; that
surveys of the technical coefficients of household production functions
failed to pick up the myraid adjustments to pollution loadings; and, that
property value studies were only as good as the data used in them—and the
data were often weak. It was hoped that some new technique could be developed
to circumvent the difficulties of the traditional estimators. Perhaps legal
cases would suggest some way of deriving information on air pollution damages
from the decisions of the judicial system in adjudicating conflicts of
interest over air resources.
A litigation survey project of Philadelphia cases, undertaken by Havighurst
(1969) and his staff, originally had two major objectives: First, they were
to locate and report in sufficient detail, all litigation—at the original
or appellate levels—that might bear on the problems of finding out how
much air pollution costs. Further, they were to determine the extent to
which the people of Philadelphia have turned to the courts for redress.
Second, using the information gathered from the study, they were to eval-
uate judicial data as estimators of damage functions.
The investigators spent many hours talking to lawyers, court clerksi state
and local control officials and anyone who might have knowledge of past or
•*' •
pending litigation relevant to the search. In all, three useful cases in
Philadelphia were found. Havighurst concluded that citizens of urban areas
are much less inclined to attempt to control pollution through private legal
39
action than are citizens of less polluted areas; city dwellers become
32
-------
conditioned to air pollution. And, in a dense industrial city, there is
some difficulty in knowing just what sources are primarily responsible
for the pollution. Due to the paucity of interesting cases in the
Philadelphia area, the project was broadened to include the Berks County-
Bethlehem region. Except for a few cases which turned up, this effort,
too, proved unavailing.
It was obvious that no damage functions, or even much useful data, would
emerge from the records of the few cases that had been located, and therefore
the most important thing to be done was to evaluate carefully the feasi-
bility of using litigation information as a means for measuring pollution
damage.
To proceed on such an appraisal required a careful comparison of the type
of damage information desired by economic analysts and the type yielded by
the courts. Noting that most courts, in practice, make nuisance awards on
the basis of th'e estimated decline in the market value of the injured
property and on the basis of the court's allowance for special "discomfort
and annoyance," Havighurst concluded that the economic usefulness of such
awards depended on the similarity between the preferences of the market and
the preferences of those actually injured. In deriving estimates of economic
damage, the more these preferences coincided, the stronger the case of
40
disregarding the court's special annoyance allowance.
A final product of the project was a recommendation that litigation surveys
of this type be continued. Despite the lack of success in the Philadelphia
area, it was felt that a national survey, perhaps of cases involving odors,
would turn up enough damage awards that some tentative functions might be
drawn. Haviqhurst suggested, however, that legal records as they now
stand are frequently unsatisfactory for this purpose due to a failure to
itemize pollution Injuries and to specify the ambient air quality involved
in the nuisance conditions.
33
-------
Political Expressions of Social Choice
In utilizing this approach, one tries to gauge political expressions,
representations, and exhortations in the hope that their intensity somehow
corresponds to intensity of preference for one outcome over another. Yet,
assessments of the outcomes of political expressions are not likely to be
accurate indicators of what receptor preferences for one state of air quality
relative to another would be in any real market situation. The intensities
of social choice registered by political means represent only the relative
valuations that occur under the constraints imposed by the political process.
As such, they reflect the ability of the voter to alter the relative prices
he faces.41
While no formal efforts have been made to specify the magnitude of the damages
which usually emerge in these processes, the numerous environmental news-
letters provide some appreciation for the intensities of social choices by
focusing on reports where voters or taxpayers have supported (or failed
to support) the passage of bond referendums, or where legislators have
raised taxes to finance the construction of some pollution control activity
42
or facility.
Delphi Method
This approach, as stated by Pill, is "...a method of combining the knowledge
and abilities of a diverse group of experts to the task of quantifying
43
variables which are either intangible or shrouded in uncertainty." Essen-
tially the method is one of subjective decision-making. It is an efficient
way to produce best judgments where the knowledge and opinion of experts ,
are extracted. Desiring a particular output, those who are considered
experts in the relevant area are asked to give their best solution to any
given problem. This method is one that has been used by the U. S. Depart-
ment of Agriculture in forecasting crop production levels. The estimate,
as generated by the U. S. Department of Agriculture, of $325 million in
crop losses due to air pollution, appears to have been developed in such
44
a manner. Another estimate apparently generated in such a manner is the
$40 to $80 million annual cost of the adverse effects of air pollution on
45
air travel.
34
-------
The Delphi method appears to be an approach that can provide quick answers
in a short time frame. Yet, due to its subjective nature, many of the air
pollution damages generated in this manner, have been, and probably should be
largely ignored. Yet as Dal key said, "We can either wait indefinitely until we
have an adequate theory enabling us to deal with socio-metric and political
problems as confidently as we do with problems in physics or chemistry, or
we can make the most of an admittedly unsatisfactory situation and try to
obtain the relevant intuitive insights of experts and then use their judg-
ments as systematically as possible. The use of the Delphi approach repre-
46
sents an effort to proceed along the second of these alternatives."
WHAT METHOD IS BEST?
Of the six methods surveyed, the technical coefficients of production and
consumption, opinion surveys of air pollution sufferers, and a particular
market study application, the property value method, have yielded the most
promising insights into the true nature of air pollution damages. Yet
even the application of these methods has been fraught with many problems.
Air pollution is but one environmental stress, and there are no satisfactory
methods of allocating the observed damages among a number of synergistically
interacting multiple stresses, nor can the damages themselves be easily
measured and reduced to economic terms.
Because of its general ease of measurement and inclusiveness, the property
value, or site value differential technique, is one of the more promising
approaches to the estimation of the economic losses due to air pollution.
The advantage of this method is that the investigator does not have to
discover and evaluate the pollution sufferers' adjustment possibilities,
nor does he have to worry about how to make individual properties commen-
surate so that he can aggregate them. The housing market does it for him
through directly observable market prices. It is simply the investigator's
job to correctly specify the separate influence of each characteristic,
including air pollution, so that each's influence on air pollution can be
discerned by well-known statistical tecnniques.
35
-------
Two significant limitations of the property value technique which result
in the underestimation of true damage costs, are: (1) the extent to which
certain minimum levels of pollution are pervasive over all properties in a
market, nothing could be gained by the receptor in relocating; and, (2)
since there are many long-run, chronic effects that are not easily measured,
it is doubtful that this technique would discern these effects. As stated
earlier, concern over the limited ability of this approach to reflect even
major effects, has been expressed by Ayres and Lave.
Applications of the technical coefficients approach also, can provide
information on the damages of air pollution. Given that all damages will
not be registered in the property value approach, the technical coefficients
approach can provide insight into the fundamental processes of receptor
response where air pollution has its impact. The technical coefficients
and property value approaches, then, can provide complementary information.
The property value approach has the advantage of ease, whereas the technical
coefficients approach has the advantage of providing insight into funda-
mental processes.
A deeper understanding of the fundamantal adjustment processes of the
receptor can also be gained by employing the interview approach. This
approach can be used to determine what effects receptors perceive or fail
to perceive. The information obtained can also be quite valuable for analyzing
the subtle effects of air pollution. Thus, the interview approach can provide
the investigator with information about receptors who suffer from air pol-
48
lution effects. This knowledge can be used as a basis to improve the
information that sufferers have so that they will make more complete
adjustments. This information, in effect, will result in some shift in
the air pollution damage function.
The litigation, political expressions of social choice, and del phi Approaches
have been somewhat less successful in measuring the damages of air'pollution
than technical coefficients studies, market studies, and opinion surveys.
An evaluation of the litigation approach shows that there is a theoretical pro-
blem of distinguishing between economic costs and legal costs. Also, there is
the general problem that court decisions usually lack adequate dose-response
36
-------
information. The severe constraints and complexities of the political
process afford little opportunity in determining the value of marginal
changes in voters' preferences.
Given the dearth of air pollution dose-response information, it is possible
that the delphi method which relies on subjective opinion rather than objec-
tive data, will be used in a more significant way. It seems obvious that
where substantial information is missing, the pooled judgments of experts
could provide useful information to the decision-maker on the general mag-
nitude of damages over a range of pollution levels.
37
-------
SECTION IV
ASSESSMENT OF NON-SPECIFIC AIR POLLUTION DAMAGES
OVERVIEW OF THE PROBLEM
Attempts have been made to assess the effects of air pollution on human
health, materials, vegetation, aesthetics, soiling, animals and the natural
environment. Studies focusing on specific effects have typically used the
technical coefficients of production and consumption approach. These
studies will be reviewed in later sections. It is somewhat more difficult
to identify what effects are measured when the investigator undertakes
market studies or opinion surveys to investigate the behaviour or awareness
of air pollution sufferers.
In those cases where it is not clear what effects are being evaluated,
the term "non-specific" will be used. Studies that have evaluated these
non-specific damages of air pollution, will be reviewed in this Section.
Such a review will afford the opportunity to assess critically the merits
of each study which , in turn, will provide a basis for the extrapolation
of findings to develop a national estimate of the cost of air pollution
damage. This chapter will review then: (1) opinion surveys with air
pollution sufferers; and (2) market studies that have employed the property
value method to measure air pollution losses.
OPINION SURVEYS
Some of the earliest research to assess public awareness and concern about
air pollution was done by deGroot and others (1966) in Buffalo. Their
findings supported the hypothesis that there was direct relationship between
the perception of the seriousness of air pollution and actual' ambient air
quality In the area. With respect to the perception of effects of air
pollution, the study showed that: (1) the majority of people believed that
air pollution is detrimental to human health; (2) three-fourths of the
people queried in 1959 thought that air pollution had a bad effect on real'
estate; and, (3) residents perceived air pollution as bad elsewhere, but not
in their own neighborhood.
38
-------
Another study was performed in Clarkston, Washington, by Medalia and
Finkner (1965) on the impact of odors and smoke emanating from a local pulp mill,
Of approximately 100 interviews, 91% perceived air pollution in the corn-
muni tv as an odor problem, 74% as a visibility problem, and 62% claimed it
to be a problem in nose-throat irritation. Concern with air pollution was
found to be unrelated to location of respondents' residence with respect
to the pollution source. This absence of a relationship was probably due to the
pervasive nature of the pollution. Concern with air pollution was found to
vary directly with social status and with attitudes about civic pride, desire
to eliminate the problem, length of residence, and occupation prestige of the
household head.
A broader study by Williams and Edmisten (1965) in Nashville, Tennessee,
included an examination of individual perceptions of pollution. People
in 3,032 dwelling units were interviewed to test the hypothesis that
perception and concern for air pollution would be directly related to
neighborhood pollution levels. Their findings supported the hypothesis.
Furthermore, the higher the socioeconomic status, the greater was the
correlation between degree of concern and air pollution levels. 'They
also found that the citizens' perception of air pollution was influenced
more by the frequency of high daily levels of pollution than by high
monthly, seasonal, or average levels.
A recent perception and attitudnal survey was reported on by Mason (1972).
His research proceeded in two steps. First, a test was made of the hypo-
thesis which stated that perceived changes in air quality in a community
are defined or redefined, in part, by the local mass media and that such
a definition is sufficient to form attitudes towards the topic. An empiri-
cal test of the hypothesis involved a content analysis of daily newspapers
in two communities where visible changes in air quality levels were known to
occur. The study took place in two cities in the Willamette Valley—Salem and
Eugene, Oregon—and the pollution studied was the visibility-restricting
smoke generated by the annual burning of grass stubble. Measurement of
attitudes of a random sample of adults in these communities was also com-
pleted twice—once when visibility was low and once when 'it was normally
high.
39
-------
Results showed that: (1) attitudes were likely to have been formed by, and
are subject to change as a result of, mass media definition; and (2) attitudes
did not change as a result of known changes in air quality.
In the second step of Mason's study, a theoretical model was tested concern-
49
inq the communication of air quality information to the public. Interview
data, gathered in the same manner as before, were analyzed by two-stage least
squares. The accuracy of a respondent's perception of visibility was also
determined. The results suggested that different communications models
represented different stages of the same underlying communication process.
Mason concluded that an understanding by public relations personnel in
environmental programs "of the complex communication process should enhance
their efforts to speed up the formation of positive attitudes and the
level of knowledge of people on the topic of environmental pollution.
Once information has been gathered on people's perception and attitude
concerning environmental pollution or the effects of that pollution, the
next step is to value the individual's willingness to pay to reduce
either the pollution or the risk of his experiencing certain effects.
i
51
Ridker surveyed the residents near a steam plant to determine the cost
of cleaning up after a malfunction in the boilers caused a fumigation
of the neighborhood of 3,500 residential units with unusually high amounts
of soot. The 1965 survey not only measured costs of cleaning, but also
attempted to measure psychic costs by asking "willingness-to-pay" questions.
The responses of 10 individuals who provided answers for measured and v/i 11 ing-
ness-to-pay costs are shown in Table 1. . Psychic costs are considered to be
aesthetic costs or losses greater than the direct, measured costs that people
suffer, and are valued at what these sufferers are willing to pay to avoid them.
40
-------
Table 1. COST OF CLEANING UP AFTER BOILER MALFUNCTION
Respondent
1
2
3
4
5
6
7
8
9
10
Measured
costs,
$
9.45
43.75
45.61
19.78
15.67
25.32 ,
13.59
12.95
4.25
8.79
Willingness-
to-pay costs,
$
10.00
50.00
45.61
25.00
25.00
30.00
25.00
20.00
4.25
8.79
Psychic
costs,
$
0.55
6.52
0.00
5.22
9.33
4.68
11.41
7.05
-
-
The results indicate that these heads of households generally were willing
to pay at least the measured cost of pollution clean-up to avoid the neces-
sity for clean-up. Usually they were willing to pay more than the measured
cost. The results indicate that the residents were willing to pay, on the
average, 27% above the measured cost of clean-up in order to avoid the clean-
up.
These results must be qualified for several reasons: First, since the
survey was constructed in only a few days, one must allow for the probability
of an incomplete questionnaire and an unreliable sample. Second, the ten
respondents on willingness-to-pay were too few to be representative of the
psychic costs of the sample of 122 households.
Another survey on willingness-to-pay for an improvement in air quality
was conducted by Lawyer (1966) among 362 of the 6,424 families in Morgan-
town, West Virginia. Table 2 shows the percentage of respondents by the
amount each would pay each year "...if all air pollution were reduced
below the point where it was noticeable (or harmful).
41
,,52
-------
Table 2. AMOUNT OF MONEY RESPONDENTS WOULD
PAY ANNUALLY TO REDUCE AIR POLLUTION
IN MORGANTOWN, WEST VIRGINIA*
Amount, $
Zero or no response
1 to 5
6 to 10
11 to 15
16 to 20
21 to 25
26 to 30
31 to 35
36 to 40
40
Total
Respondents, %
38.4
23.9
9.4
2.5
4.7
6.3
0.3
0.6
0.3
13.5
99. 9a
Error resulted from rounding of figures
Source: Robert E. Lawyer. An Air Pollution
Public Opinion Survey for the City of Morgan-
town, West Virginia. West Virginia University,
Morgantown. Unpublished Master's Thesis. 1966.
The average amount respondents would be willing to pay can be calculated
to be $16.46. The "Zero or no response" category is not considered in the
average because it is probable that many of the observations in this class
are non-respondents. The average of the $1 to $5 class is taken to be
$3.50 since it is assumed the class really extends to $5.00. The averages of
each subsequent class are $5.00 higher than the previous one. The highest
class is assumed to have an average mark of $40.
An average payment of $16.46 per year per respondent is much higher than
the willingness-to-pay results of a study by Williams and Bunyard (1966).
42
-------
They reported that 66% of those interviewed in a 1963 survey of the St. Louis
area were willing to pay $5.00 per year in higher living costs for clean
air and that 85% of those interviewed would pay $1.00 per year in higher
taxes.
This brief literture survey has been a review of a rapidly growing body of
knowledge concerning the perception or awareness of and the formation of
attitudes about air pollution, and the willingness to pay for reductions of
the perceived effects of air pollution.
V
PROPERTY VALUE STUDIES
Another method that has been used to measure willingness-to-pay is a particu-
lar type of market study—the property value approach (For detailed discus-
sion of the methodology, see Section III). As with the opinion surveys
reviewed earlier in this section, the property value method measures "non-
specific" effects. By applying classical least-squares regression procedures,
the existence of a statistically significant relationhip between air pollution
and property values can be tested.
Ridker and Henning - St. Louis
The first serious use of the property value or housing market estimator was
made by Ridker and Henning (1967). They used the statistical technique of
multiple regression analysis to isolate the significance of air pollution—
sulfation, in this case—in explaining changes in property values. Using
1960 census data and pollution readings from the 1963-64 Interstate Air
Pollution Study, their regression equation explained over 90% of the varia-
tion in the median property values of the St. Louis Standard Metropolitan
Statistical Area (SMSA) census tracts. The variables in their regressions
were as fol1ows:
MPV = Median property value
SUL = Annual geometric mean sulfation levels
MNR = Median number of rooms
43
-------
PBR = Percentage homes recently built
HPM = Houses per mile
TIZ = Bus travel time to St. Louis central business district
HWA = Accessibility to highways
SCHI = School quality
OCR = Occupation ratio (ratio of craftsmen, foremen,
operatives and laborers to total work force)
PPU = Population density
PNW = Percentage non-white residents
RILL = Dummy variable indicating whether census tract in Illinois
or Missouri, orthogonal to sulfation
RMFI = Median family income, orthogonal to MNR, HPM, OCR
They concluded that the partial regression coefficients for sulfation
can be interpreted as meaning that if the average sulfation levels to
which any single family dwelling unit is exposed were to drop by 0.25 mg of
S03/100 cm /day, the value of that property can be expected to rise by at
least $83 and more likely closer to $245. The coefficient of SUL, then, esti-
mates the minimum sum needed to induce receptors to endure various levels of
whatever pollution a sulfation index measures if all other factors have their
mean values. At any inducement less than this, receptors on the average
would find it to their benefit to demand either a higher payment (which
in essence becomes a cost of production to the emitter to use the atmosphere
for waste disposal) or a cleaner environment.
Zerbe - Toronto
Zerbe (1969) used a theoretic rationale similar to Ridker and
Henning in estimating the air pollution damage to residential property in
Toronto and Hamilton, Ontario, Canada. He related property values for
each census tract to: four variables representing neighborhood characteristics;
five variables representing property characteristics; and, two variables
representing average levels of sulfur dioxide and dustfall pollution. On
the basis of his best fitting equation, which explained almost 96% of the
variation in median property values, Zerbe concluded that for Torpnto,
other things being equal, for each increase in 1 mg S03/day, property
values will fall by an amount between $800 and $1800 for each single family
CO
dwelling. His best estimate was that values will fall by about $966.
44
-------
Jaksch and Stoevener - Toledo, Oregon
In like manner, Jaksch and Stoevener (1970) conducted a study in Toledo,
Oregon, using dustfall measurements as the pollutant variable. Their
hypothesis was that air pollution represents an economic cost to the
affected community and that such cost is reflected in the property values.
They found that reductions in property values due to increasing air pollution
levels are likely to be greater for higher-valued, more-developed properties
than for less-developed, lower-valued properties. Two models were developed
to analyze the relationship between dustfall pollution and residential property
values, the difference being the measure of the price of residential property.
2
In one, it was found that an increase of 1 ton/mi -month in dustfall caused
property values to decline by $277 per acre. In the other, it was found that
a similar increase in dustfall caused a decline in property values of $29 per
market transaction.
Anderson and Crocker - St. Louis, Kansas City, Washington, D.C.
Anderson and Crocker (1970) gave more attention to the specification of
their housing market model and to the formulation of the theoretic rationale
underlying their study, and thus, significantly refined the method
first employed by Ridker and Henning. Anderson and Crocker studied the
covariation of sulfation, suspended particulates, and census tract median
property values in St. Louis, Washington, D.C., and Kansas City. The equations
most successful in explaining the variation in property values in each of
the three cities are listed in Table 3.
A multiplicative (linear in logs) equation form gave the best statistical
2
fit—highest R —in the regression for each city. Without exception, the
signs of the coefficients for all explanatory variables, including the air
pollution variables, were in agreement with a priori expectations. The
hypothesis that air pollution (as Anderson and Crocker defined it) and property
values are inversely related, was confirmed.
45
-------
Table 3. ANDERSON-CROCKER PROPERTY VALUE REGRESSION EQUATIONS
Equation
Degrees^of
Freedom
St. louis
Kansas City
Vlashington.D.C.
In MPV = 3.5407 - .1019 In (PSN) - .1192 In (PPT) + .7660 In (MFI) .7550 228
(.6332)c(.0340) (.0475) (.0772)
- .0802-In (DLP) - .0257 In (OLD) + .0373 In (Nl-IT) - .1387 In (DIS)
(.0087) (.0162) (.0060) (.0382)
In MPV = 3.5775 -,_Q782 In (PSN) - .0876 In (PPT) + .6720 In (MFI) .8231 179
(.7261) (.0396) (.0362) (.0898)
- .0405 In (DLP) - .0721 In (OLD) - .0058 In (NWT) - .0623 In (DIS)
(.0094) (.0124) (.0067) (.0245)
In MPV = 3.3901 - .0712 In (PSN) - .0610 In (PPT) + .7677 In (MFI) .6966 267
(.4012) (.022) (.0318) (.0447)
+ .0044 In (DLP) - .0106 In (OLD) + .0251 In (NWT) - .0582 In (DIS)
(.0059) (:0103) (.0064) (.0158)
Where: MPV = Median property value
PSN = Annual arithmetic mean sulfation
PPT = Annual arithmetic mean suspended particulates
MFI = Median family income
DLP = Percentage homes dilapidated
OLD = Percentaqe homes more than 20 years old
in 1959 "
NWT = Percentage of homes occupied by nonwhites
DIS = Distance to central city
Notes: Shows what part-of-the, variation in property values across census tracts within the SMSA is explained
by the equation.
The degrees of freedom relate to the level of confidence one can have in the stability of the R .
cThe figures in parentheses are the standard errors of the coefficients.
-------
From their equations, Anderson and Crocker concluded that the pollution (sul-
fation and suspended particulates combined) elasticity of MPV falls between
p
-0.1 and -0.2. Thus for every change of 0.1 mg SO^/lOO cm -day in sulfation plus
3
10 ug/m -day of suspended particulates over a given census tract, the best
estimate of the change in that tract's MPV lies in the interval $300 to $700.
Crocker - Chicago
Crocker (1971) in an attempt to extend further and to test the methodology
formulated by himself and Anderson, performed a study in Chicago. The
purposes of this study were: (1) to test new economic hypotheses about the
relation between property values and air pollution; and (2) to remove possible
sources of statistical bias present in previous studies. In studying the
covariation between air pollution dosages (sulfur dioxide and suspended par-
ticulates) and property values (using FHA and census tract data) his results
were consistent with those from other property value studies.
On the average, the sum of the damage elasticities for sulfur dioxide and
suspended particulates in the City of Chicago were found to be between
-.30 and -.40. This would mean that the average marginal capitalized damages
o
are about $450 for: (1) an additional 10 ug/m /day of suspended particulates;
plus, (2) an additional part per billion by volume per day of sulfur dioxide.
Peckham - Philadelphia
Peckham (1970) studied the covariation of urban property values and air pol-
lution to determine if the relationship determined in other studies, would hold
in Philadelphia. Peckham concluded, "It does appear clear that negative and,
in most cases, unambiguously significant coefficients have been found for
both pollution variables, and that in confirmation of Crocker and Anderson, the
elasticity of MPV with respect to both AMS (annual arithmetic mean sulfation)
and SPT (annual arithmetic mean suspended particulates) is about -.2."
Peckham's estimates of the marginal capitalized sulfation damage of .1 mg
SO-/100 cm -day were (for two equations) $600 and $750.
47
-------
Spore - Pittsburgh
Spore (1972) performed a cross-section analysis of the relationship between
air pollution and property values in Pittsburgh. The pollutant measurements
of dustfall and sulfur dioxide (as determined from sulfation data) were re-
gressed against 1970 U.S. Bureau of the Census data. By applying classical
least-squares regression procedures (multiplicative equations), the presence
of a statistically significant inverse relationship between air pollution and
property values was shown, confirming the conclusions of earlier studies. His
analysis showed that for an average property, the effects associated with an
2
additional 5 tons/mi -month of dustfall plus an additional .005 ppm/day S02
(or, a 10-15% increase in air pollution dosages) result in an increase in
pollution damage costs of approximately $150-$200.
Weiand - St. Louis
Weiand (1970) attempted to test the hypothesis that air pollution is negatively
related to property values in St. Louis. His work differed from that of
Anderson and Crocker and Ridker and Henning in that he started from rather
different premises. Weiand used a measure of residential acreage to derive
a measure of land use intensity as the dependent variable representing the
price of housing. Thus, he defined, a commodity of "housing services" as
being representative of two variables, the price and the quantity of housing.
He then demonstrated that housing services and property values do not vary with
55
one another. With his definition of the commodity of housing services, he
/studied the effect of air pollution on that variable, but found no statistically
significant relationship. Anderson and Crocker argue that there is nothing
logically wrong with Weiand's method, but it is fraught with more statistical
problems and thus, it is much harder to identify incremental air pollution
effects.56
Crocker - Polk County, Florida
Crocker's (1968) study in Polk County, Florida shows that the property value
method is as applicable to agricultural areas as to urban areas. In fact, Crocker
believes that property value estimators used in rural areas may capture a greater
proportion of the damages since the health-related costs are likely to be some-
what less important.
In the Florida study"Crocker investigated the economic impact of fluoride emis-
sions that emanated from the production of phosphate fertilizers, on surrounding
48
-------
agricultural land. Such land was used either for the production of citrus
or for the grazing of cattle. To distinguish between submarkets, separate
analyses of the cattle and citrus industries were performed. The most satis-
factory estimating equations for citrus was multiplicative while that for
grazing lands was linear.
The explanatory variables in the citrus equation were statistically significant
and possessed a sign in accordance with that expected from economic theory.
Over time, the magnitude of the negative coefficient for air pollution con-
sistently followed emission patterns, and in the years 1961 - 1962, the average
reduction in sale price for citrus sites in the area was about $150 per acre.
By 1964 when fluoride emissions had been significantly reduced, the differen-
tial sale prices attributable to the presence of air pollution had disappeared.
Flesh and Weddell - Southern California
Flesh and Weddell (1972) studied the relationship betwen odors and property
58
values in Southern California. Using the same method as that posed by
Ridker in his time-series study in St. Louis, Flesh and Weddell attempted
to estimate the economic costs of odors through property value differentials.
The primary contention was this: if the presence of odors in an area does have
negative impact on home values, the impact should manifest itself in slower
growth in home prices, at least at the outset of the problem. They examined
growth rates of property values in two "identical-except-for-pollution" residen-
tial neighborhoods. The study failed to show significant changes in the dif-
ferential property values. Either there was no impact of odors on property
values or the method of analysis and limits of the data masked the effects.
Summary
In summary, the majority of these property value studies use regression
analysis to estimate a partial equilibrium, single-equation model whose
explanatory variables include measures of not only air quality, but also location,
neighborhood, occupant, and physical property characterstics thought to be
determinants pf residential property values or rents. With two exceptions,
49
-------
the data measurements were obtained from cross-sectional samples over either
individual properties or aggregates of properties, such as census tracts. And
with three exceptions,61 these studies established a statistically significant
inverse relationship between air pollution and property values or rents.
Also, most studies have included attempts to uncover the presence of harmful
multicollinearity and correct for its effects. Anderson and Crocker found
significant collinearity for only the two pollution variables. Thus, they
interpret their coefficients as a joint measure or composite of pollution. Both
they and Spore agree that, based on their samples, it seems that the ability to
uncover the true significance of air pollution as a determinant of property
values or rents is not hindered by the presence of multicollinearlity between
air pollution and other independent variable included in the regression equation.
NATIONAL ESTIMATE OF AESTHETIC AND SOILING COSTS
To estimate total damage costs using the property value technique, one would
have to perform separate property value studies for residential, commercial,
industrial, and agricultural land. Given the paucity of information in areas
other than residential, total damage estimates in this report are only for
those damages capitalized in the residential property market and measured
through site and improvement differential values.
The findings of the property value studies reviewed here are summarized in
Table 4. All studies agree, that air pollution is inversely related to MPV.
The magnitude of the marginal capitalized sulfation damage for residential
2
structures for a marginal decrease of 0.1 mg S03/100 cm -day, lies roughly in
the range $100 to $600. This uniformity of results, for six major metropolitan
areas, warrants some confidence in the worth of the housing market estimator
for estimating national pollution damages.
It is well to recognize the difficulties in the way of making straightforward
comparisons of the findings of different regression experiments in different
cities. Ridker and Henning, as well as Zerbe use only one on the pollutant
50
-------
Table 4. SUMMARY OF PROPERTY VALUE STUDIES
Study
Ridker-Henning
Zerbe
Anderson-Crocker
Crocker
Peckham
Spore
City
St. Louis
Toronto
Hamilton
St. Louis
Kansas City
Washington, D.C.
Chicago
Philadelphia
Pittsburgh
Pollution Measure
Sulfationa
Sulfation3
Sulfation9
Sulfation
Suspended particulate
Sulfation
Suspended particulate
Sulfation
Suspended particulate
Sulfur dioxide
Suspended particulate
Sulfation
Suspended particulate
Sulfation
Dustfall
Pollution
Coefficient
-.12
-.08
-.10
-.12*
-!09*
-.07
-.06**
.06**
, -.40
-.10
-.12
+ .03*
-.12
R2
.94
.92
.76
.82
.70
.77
.76
.81
Marginal
Capitalized
Damage
100b
97b
300-700°
470d
600-750°
1 50-2006
*Not significantly different from zero at the .01 level
**Not significantly different from zero at the .05 level
Notes
: ? Single pollution variable probably measures effect of both sulfaction and suspended particulates
Mean change in MPV per change of .1 ma SO-/100 cm?-day -
c Mean change in MPV per change of .1 mq SO-/100 cm'-day plus 10 yq/m' -day change in suspended
. particulates * ~
Mean change in MPV per change of .001 parts per million/72 hrs. S0? plus 10 yq/m -day change in
suspended particulates ?
e Mean change in MPV per change of .005 parts per million/day of SO, plus a 5 tons/mi -month change in
dustfall L
-------
regressors in their equations, while the others always use both. Anderson and
Crocker argue that the identification of the separate influences of the two,
highly correlated pollution variables, can best be achieved when both appear
as regressors in the same equation. If this argument is true, it is probable
that the marginal damages reported by both Ridker and Henning and Zerbe reflect
the influence of both sulfation and particulate variations.62
Despite the pitfall that the pollutant measures in the different studies are
not homogenous, it seems safe to say that in comparing the different studies,
the slopes of the estimated damage functions differ from city to city. Peckham
says that, "...this conclusion is no necessary discredit to the property value
technique for measuring pollution damage, for cities do vary in their meteoro-
logy and emission density, and one would expect some corresponding variation
63
in their damage functions.1 Anderson and Crocker concluded that over the
ranges of both sulfation and suspended particulates that they observed, marginal
capitalized damages and the responsiveness of damages to pollution, appears to
decline with increasing arithemetic mean pollutant concentrations. Thus,
total damages seem to increase at a decreasing rate, and the proportionate
change in damages, relative to the proportionate change in pollution concentra-
tions appears to decline with increasing arithmetic mean pollution concentrations.
This pattern was observed when the results for areas within an individual city
65
were compared, as well as comparison among cities.
Given a marginal capitalized damage coefficient and assuming that sulfation
changes are always evenly distributed among census tracts (i.e., a 10% drop in
the annual average sulfation rate for a city Implies a 10% drop in correspond-
ing rates for each tract), crude estimates of sulfation damage (as captured in
property value differentials) can be calculated by the following equation:
i
DAMAGE = (Marginal capitalized sulfation damage) x (no. of
marginal changes needed to reduce arithmetic annual
mean sulfation rate for the metropolitan area to
desired background) x (no. of housing units).
The damage given by this relation is total capitalized pollution damage,
or the decrease in real property wealth caused by whatever pollution is measured
by sulfation.
52
-------
It should be noted that this damage equation uses linear extrapolation and
consequently it can give rise to both under and overestimation biases. These
biases are demonstrated in Figures 3 and 4. Figure 3 is an example of under-
estimation, and Figure 4 is an example of overestimation. In both figures it
is assumed that the slope (b) at the indicated point is at the relevant pollution
coefficient taken from a property value regression and that air quality level
Q* is "clean air". The solid lines are the "true" property value-pollution
functions. The dashed lines are linear extrapolations of property values to
coincide with "clean air" levels.
Total annual damage, or the decrease in real property income from pollution,
is obtained by multiplying total capitalized damage by a discount rate reflecting
T 1
the average return on capital:
*i
TOTAL ANNUAL DAMAGE = (Discount Rate) x (DAMAGE)
Thus, the total capitalized pollution damage depends on: the choice of a mar-
ginal capitalized sulfation damage coefficient; the desired background sulfa-
tion rate; the annual arithmetic mean sulfation rate for the metropolitan
area; the number of housing units over which the aggregation occurs; and, the
discount rate.
In this report the following choices were made:
1. The studies reviewed here, taken together, show that the magni-
tude of the marginal capitalized sulfation damage for residential
p
structures, for 0.1 mg/SO./lOO cm -day, probably lies in the
3
range of $100 to $600. For purposes of this report a range of
damages will be estimated using extreme marginal capitalized sul-
fation damage coefficients of $200 to $500.
2. Selection of the desired background level at 0.1 mg/S03/100 cm -
day was guided by the annual sulfation averages in selected
fi7
suburban and rural regions.
3. Annual arithemetic mean sulfation rates were established for all
metropolitan areas either by: (a) using sulfation data on annual
arithmetic means from the Interstate Surveillance Project;68
53
-------
Property
Va-lue, $
"True"
Damage
Function
Underestimate
"True" increase in property
value from improved air quality
Air QualityQ*~tciean Air)
Figure 3. Underestimation of Property Value Losses
Property
Value, $
"True"
Damage
Function
Overestimate
"True" increase in property
value from improved air (quality
Air Quality Q* (Clean Air)
Figure 4. Overestimation of Property Value Losses
54
-------
(b) using available sulfation data from EPA's SAROAD data bank
and averaging monthly data; or (c) estimation for those
metropolitan areas for which sulfation data were lacking, based
on a regression of annual sulfur oxides (SO ) emission data on
A
sulfation annual averages.
4. All estimates were calculated for the number of year-round
housing units in each metropolitan area as given for 1970.
5. As an approximation to an average return on all real property
wealth in the economy, a 10% discount rate was uniformly used
in all calculations.
As a result of the calculations described above, and assuming a range of mar-
ginal capitalized damage coefficients of $200 to $500 for each reduction in
2
0.1 mg/SOg/lOO cm-day, the national annual estimate for 1970 of air pol-
lution damages measured via the property value differential method comes to $3.4
to $8.4 billion. A best approximation would probably be a middle estimate for
a MPV of $350, or a total damage of $5.9 billion.
In conclusion, this estimate: (1) spans all housing units within all metro-
politan areas; (2) assumes that pollution changes are spread evenly over all
census tracts; (3) assumes that there is a negative relationship between sul-
fation and MPV; and (4) indicates the approximate amount that residents of
American cities would demand, under emitter liability, to forego asserting
their rights to have pollution abated so that arithmetic mean sulfation rates
in all SMSA's would be 0.1 mg/S03/100 cm -day or lower.
As argued earlier (see Section III), it is believed that the costs associated
with aesthetic effects as well as soiling-caused cleaning and maintenance
expenditures are capitalized in this estimator. These costs reflect the tan-
gible, experiential "disamenities" associated with air pollution. Here, the
$5.9 billion is used as an estimate of the damages to aesthetic properties and
soiling, although it is recognized that other effects may be included in this
estimate.
55
-------
SECTION V
ASSESSMENT OF AIR POLLUTION DAMAGE TO HUMAN HEALTH
OVERVIEW OF THE PROBLEM
It is common to see references that cite health effects separate from the
economic effects of air pollution. One inference can be drawn from this
division: the effects of air pollution on health transcend economic values.
While in fact there is a value, it is not easily measured. The separation
may have been influenced largely by the absence of cost estimates for health
relative to other types of damages.
What we are interested in here is how air pollution affects: illness and
death rates, including partial disability; absence from work and school; and
general expenditures on health protection and care. For example, we would like
to know how many days of good health or work would be>gained by a specific
reduction in pollution. The obvious difficulty is: isolating the subtle,
marginal effect of pollution on the complex human organism.
If health is defined as the general state of well-being, it might be convenient
to think of a five-stage human response spectrum of increasing severity at a
given level of air pollution (see Figure 5). While a significant; percent of
a population might be exposed to air pollution, only a smaller portion will be
adversely affected, and then still a smaller fraction affected to the degree of
death. The dashed line through the middle of the response spectrum in Figure 5
indicates that below this line, the health response is probably not economically
significant. Above this dashed line, the response is expected to be of
economic significance.
As indicated by the dotted line in Figure 5, the distinction between "psychic"
and "normal" morbidity costs is unclear. While it is believed that such
psychic costs do exist, they are probably only partially measured by "normal"
morbidity indicators such as work-loss days, doctor visits, etc. Thus, any
quantification of health costs that considers only morbidity and mortality
costs will underestimate total health costs. While recognizing that these
56
-------
Morbidity
Physiological Sentinel
of Disease
(Possible Psychic Costs)
Physiologic Changes of
Uncertain Significance
Pollutant Burdens
Adverse
Health
Effects
Proportion of Population Affected
Figure 5. Human Response to Pollutant Exposure0
Adapted from: C.M. Shy and J.F. Finklea, Air Pollution Affects
Community Health. Environ. Sci. and Techno!, 17. (3):205, March 1973.
57
-------
subtle psychic costs are important, because of data limitations in this area,
the analysis here will be restricted to the areas of morbidity and mortality.
Investigations focusing on the impact of air pollution on mortality rates,
and on morbidity rates, represent two distinct areas of research.
Death is often the result of a large number of unrelated causes. For example,
a general "urban factor" has been identified as a causal agent in
explaining mortality rates. People in cities tend to lead more tense and
less healthy lives; they smoke more, tend to be more overweight, and get less
exercise. Somehow, the effects of air pollution must be separated from other
causes explaining this "urban factor".
Lave lists and discusses the following factors that might affect mortality:
(a) physiological characteristics which include age, sex, race, and genetic
factors; (b) socioeconomic characteristics which include income, occupational
mix, and fuels used for home heating; (c) environmental factors that include
air pollution, radiation, and climatological factors; and (d) personal factors
that include smoking habits, habits of exercise, medical care received, and
general nutrition. As Lave says almost in despair, "It is virtually impossible
to account for all possible factors that might be the 'real' causes of ill
health."72
INDIVIDUAL STUDIES
Even given the empirical problems, several notable attempts have been made to
quantify the health costs attributable to air pollution. It appears that until
1970, Ridker (1967) had made the only significant published attempt to estimate
the health costs due to air pollution. Since-then, Riggan (1970) has estimated
the health costs associated with automotive emissions. And, a study underway
in the Environmental Protection Agency is attempting to estimate morbidity costs
associated with a pollution composite of sulphur dioxide, total suspended parti -
culates, and suspended sulfates. Finally, estimates ,of mortality costs asso-
ciated with air pollution have been developed by Lave and Seskin (1970).
58
-------
The studies by Ridker, Riggan, and the EPA follow the damage factor method
where: (a) an estimate is made of the value of total health losses; (b) a
proportionality factor is determined for the share of this value attributable
to air pollution; and (c) the product of total health losses times this fac-
tor yields the value of health losses associated with air pollution.
In all three studies, the proportion of health losses associated with air pol-
lution is determined to be a constant. Thus, the cost-of-pollution function
for health may be taken to be linear with a negative slope. This assumption
should be interpreted as an approximation over a limited range of a much more
complex function. By assuming that a marginal improvement in health has a
constant price over the estimated range of the damage function, the linear
damage function gives rise to a linear benefit function.
Ridker - Morbidity and Mortality, Respiratory Diseases
The earliest known published attempt at estimating health costs resulting from
air pollution is that of Ridker (1967). For the year 1958, he estimated the
total national cost of morbidity and mortality for diseases associated with
the respiratory system. The diseases considered were cancer of the respiratory
system, chronic and acute bronchitis, pneumonia, emphysema, asthma, and the
common cold. These diseases were chosen on the basis of crude empirical
determinations and on a priori knowledge.
The cost estimates for each disease included the costs of premature death and
burial, treatment costs, and costs associated with absenteeism. The costs of
dying earlier than expected were estimated by the 1958 present value of lost
future earnings. To put earnings on an annual basis, discount rates of 5 and
10 percent were applied to expected lifetime earnings, assuming full labor force
participation. Costs of premature burial represented the difference between
the present costs of burial and the present value of future expected burial costs
discounted at rates of 5 and 10 percent. Treatment costs were estimated for each
disease using per capita expenditures of drug manufacturers' shipments. Absen-
teeism costs were the product of days lost for each disease and the average
annual earnings of those suffering from that disease. Summing the costs for 1958
(using a discount rate of 5 percent) yielded a total economic cost of these
diseases of $1.99 billion. Table 5 presents Ridker's results.
59
-------
Table 5.
COSTS OF DISEASES ASSOCIATED WITH AIR POLLUTION
Type of
cost
Premature
death
Premature
burial
Treatment
Absenteeism
Total
Costs3 associated with selected diseases, $ million
Cancer
of the
"espiratory
system
518
15
35
112
680
Chronic
bronchitis
18.0
0.7
89.0
52.0
159.7
Acute
bronchitis
6.0
0.2
na
na
6.2
Common
cold
nab
na
200
131
331
Pneumonia
329
13
73
75
490
Emphysema
62
2
na
na
64
Asthma
59
2
138
60
259
Total
992.0
32.9
535.0
430.0
1,989.9
en
o
Using a discount rate of 5 percent.
bNot available.
Source: Ronald G. Ridker, Economic Costs of Air Pollution, New York: Frederick A.Praeqer, 1967,
-------
Ridker estimates that 18 to 20% of the approximate cost of $2 billion in
national health costs for respiratory diseases are due to air pollution. His
proportionality factors of 0.18 and 0.20 represent the proportion of health
losses attributable to air pollution. These coefficients are taken from two
studies relating, in one case respiratory mortality rates and, in the other,
lung cancer mortality rates, to urban and rural areas. The coefficients are
corrected for age, sex, race, smoking habits, and the proportion of U.S. popu-
lation which is urban. Thus, the damage to health from air pollution in 1958
was estimated to be about $.4 billion.
Ridker considers the estimate to be quite conservative. For one thing, he
recognizes the absence of avoidance costs representing the moving costs and
possibly reduced earnings of those who migrate to areas of lower pollution
because of their health. He did not impute the full value of housewife ser-
vices lost due to death or illness related to air pollution. Nor was any attempt
made to include any psychic costs associated with death or illness. Data limi-
tations prevented estimates of full costs for some of the diseases considered
and also prevented the consideration of certain diseases.
Full recognition is given to the inability to relate damages to specific
types of pollution, thus preventing the construction of individual damage
functions.
Riggan - Human Health and Motor Vehicle Pollution
Where the Ridker study will allow for estimation of the economic benefits
to be derived from the abatement of air pollution, Riggan (1970) takes a
more limited view. Riggan1s study was not designed to generate national esti-
mates of total health costs, but rather to investigate the economic costs of
motor vehicle pollution on human health and the resulting losses to the Federal
Government. In using the same general method as Ridker, Riggan estimates
the value of income taxes that would have,been paid to the Federal Government
by people who died prematurely from automotive and related pollution. His
total was $1.9 billion for 1970. He estimates the value of social security
payments to workers disabled by automotive pollution to be $189.4 million.
Riggan also estimates a value of $97.1 million in lost productivity by
61
-------
Federal workers 1n Washington, D.C., due to oxidant-causing eye irritation.
His total dollar health cost attributed to motor vehicle pollution is
estimated at $2.2 billion.
The study by Riggan is fraught with empirical problems, and it lacks a sound
theoretic rationale. The methods used to derive the pollution coefficients
that are applied to the total disease categories, are not clear. With
respect to the theoretical problems, neither is it clear what a "cost to the
Federal Government" really means. It can be argued that the author should
have included estimates of the social security disability payments that the
Government "saved" because of premature death. This inclusion might have
forced the net cost down to the break-even point where the Federal Government
neither lost nor gained because of pollution from automobiles and other related
pollution. Therefore, the results from this study are not used in this report
to generate a national damage estimate for health.
Jaksch - Morbidity and Air Pollution in Portland, Oregon
Jaksch's (1972) study in Portland, Oregon, is a good example of a thorough
morbidity-air pollution study that presents an economic-theoretic rationale
for estimating health costs of air pollution. Using multivariate regression
techniques, Jaksch attempted to isolate the effect of air pollution on the
consumption of medical services by enrollees in the Kaiser pre-paid health
plan. His supposition was that air pollution (suspended particulates in this
case) can aggravate a state of health resulting in increased consumption of
outpatient medical services, and in an increase of the number of contacts with
the medical system, for certain respiratory, cardiovascular, and other diseases
aggravated by air pollution. By considering a host of explanatory variables for
pollution exposure, personal attributes, socioeconomic-demographic charac-
teristics, and meteorological parameters, Jaksch was able to isolate
suspended particulates as having some effect on the consumption of outpatient
medical services used to treat diseases.
62
-------
EPA - Morbidity, Selected Respiratory Diseases
73
In a study being performed within the Environmental Protection Agency,
research findings from the Community Health and Environmental Surveillance
System (CHESS) Program are being evaluated to determine the cost of selected
adverse health effects associated with the pollution composite, sulfur dioxide-
total suspended particulates-suspended sulfates. The following adverse health
effects, which are associated with exposures to these air pollutants, are
included in this study: acute lower respiratory disease and chronic respiratory
disease.
The methodology employed in this EPA health study involved four steps. First,
the impact of these air pollutants on disease rates was determined by statisti-
cal analyses of data collected from a number of CHESS communities offering dif-
ferent pollutant gradients. These communities were specifically selected to
control for major co-determinats that might affect disease rates. Second,
results of the association between these air pollutants and ill health were
extrapolated to other Standard Metropolitan Statistical Areas on the basis of
relative pollution gradients. Third, the population affected was estimated by
using disease rates provided by the National Center for Health Statistics,
DHEW. Fourth, average per-case-costs were then applied to rough estimates
of the affected population to determine the gross disease-specific costs.
The following costs are being evaluated: (1) the direct medical expenditures
for doctor visits, medicine, etc.; (2) work loss days; (3) housewife bed
disability days; and (4) school loss days.
The objective of the EPA health study 1s to estimate the benefits of reducing
1n one case sulfur dioxide and partlculates to levels of the primary air
quality standards, and in the case of sulfates, to an assumed annual average
standard of 6-8 ug/m . There are obvious pitfalls associated with translating
risk factors, sever1ty-of-disease rates and average per-case-costs of selected
diseases Into economic estimates. Yet, such information on health costs should
Improve our understanding of the potential health benefits to be realized by
abating such air pollution.
63
-------
Lave and Seskin - Mortality
Lave and Seskin (1970) expanded the scope of diseases covered by Ridker
(1967). In particular they added air pollution damages to health, in the forms
of heart disease and cancers of the stomach, esophagus, and bladder. In the
case of bronchitis and lung cancer, they consider the evidence relating air
pollution and health to be very reliable. Although evidence relating air
pollution to heart disease and non-respiratory cancers is not so reliable, they
believe that a consideration of all factors suggests causality.
Approximately half the lost income and current medical expenses associated
with morbidity and mortality from bronchitis are ascribed to air pollution.
The proportionality factor for lung cancer is estimated to be 25%. In the other
categories, air pollution is responsible for an estimated 15% of the damages
associated with non-respiratory lung cancers, 25% of all respiratory diseases,
and 10% of cardiovascular diseases. These coefficients were estimated by
regressions that were run on data from published epidemiological studies.
Based on these coefficients, the total annual cost of air pollution in increased
human morbidity and mortality can be estimated to be $4.3 billion for 1963;
or inversely stated, a 50 percent reduction in air pollution would result in
an annual savings of about $4.3 billion. These results are summarized in
Table 6.
Table 6. HEALTH COST OF AIR POLLUTION, 1963
Disease Category
Respiratory
Cancers
Cardiovascular
Total
$ billion
1.9
1.5
0.9
4.3
64
-------
Table 7. FACTORS AFFECTING THE MORTALITY RATE IN U.S. CITIES
O»
01
MR. = 19.607 + .041 mean P, + .071 min S. + .001 P/M2 + .041% NW. + .687 % 65. + e.
1 (2.53) n (3.18) n (1.67) (5.81) n (18.94) 1
Sensitivity
coefficients:
where:
MRi
mean P.
min S.
y i
P/M*
% NW
65i
ei
.53%
.37%
.07%
.57%
6.32%
= the total mortality rate (per 10,000 people) in city i
= the arithmetic mean of suspended participate read|sgs in city i
= the smallest biweekly sulfate reading ^n city_i
= the population density in city i
= the proportion of the population which is nonwhite in city i
= the proportion of the population 65 and older in city i
= an error term for variation in the mortality rate not explained
by the equation
Notes: The figures in parentheses are the t statistics for a test that the estimated coefficient
is not significantly different from zero (no effect).
The sensitivity coefficients show the expected increase in the mortality rate estimated to
come from a 10% increase in each variable in turn.
The equation explains 82.7% of the variation in the mortality rate across 17 cities in 1960.
-------
In their cross-section study of 117 cities, Lave and Seskin (1973) used
multivariate regression techniques to explain variation in mortality rates
as a function of air pollution, socioeconomic variables, and age. Their
regression (Table 7) explains 83% of the total variation in mortality
rates for those cities. The regression is a linear equation which predicts the
mortality rate within a city on the basis of air pollution, population density,
the proportion of non-whites in the population, and the proportion of the
population age 65 and older. All coefficients in the regression are shown to
be significant.
The "sensitivity coefficient" shows how much the mortality rate would be estimated
to increase if one of the variables were to increase by 10%, i.e., if the mean
level suspended particulate increased 10%, results show that the mortality rate
would increase .53% and a 10% increase in both particulates and sulfur oxides
would increase the mortality rate by .90%. Assuming the linear relationship
between air pollution (as defined here) and mortality, a 50% decrease in"air
pollution is associated with a reduction in the mortality rate by 4.5%. Lave
concludes that while the individual sensitivity coefficients cannot be argued
as "true", their general magnitude cannot be doubted.
For several reasons, Lave and Seskin consider their estimate to be conservative.
They argue that for conceptual meaning, the willingness of an individual to pay
for improved health or longevity, given a level of income and wealth, is the
true measure of health costs due to air pollution. To Lave and Seskin, the
sum of income lost and current expenditures resulting from morbidity and mor-
tality caused by air pollution is a gross underestimate of willingness-to-pay.
Also, some health costs may have been overlooked, resulting in a more conser-
vative estimate. Finally, the exclusion of some treatment costs results in the
underestimation of true damages.
The regression survived a number of tests and manuevers designed to identify
any spurious relationships. Air pollution was significant where expected, such
as in explaining the mortality rates of the very young, the very old, and
people dying of cardiovascular and respiratory diseases. Also, air pollution
was not significant where expected, neither in explaining the mortality rates
for young adults, nor for people dying from suicide.
66
-------
NATIONAL ESTIMATE OF HEALTH LOSSES
It is believed that the findings of EPA's CHESS program on morbidity and
Lave and Seskin's work on mortality provide an acceptable basis on which a
gross damage estimate of the health costs of air pollution can be made.
By applying the general methodology utilized in the EPA study to an analysis
of the CHESS findings on the health effects of sulfur dioxide, suspended
particulates, and suspended sulfates, rough measures of some economic health
benefits of controlling these pollutants can be estimated. This crude method-
ology is outlined in Table 8.
In Table 8 the particular health effects are identified in the different reports
from CHESS. Estimates of affected populations are based on: data on popu-
lations-at-risk taken from an internal EPA report; data on disease rates pub-
lished by the National Center for Health Statistics; and, population data re-
corded in the Statistical Abstract of the U.S.-1972. The estimated change for
each health effect is based on an interpretation of the data reported in the
individual CHESS studies. Estimates of cost-per-health effect are based upon
information taken from the Statistical Abstract of the U.S.-1972 and NCHS reports,
tempered with best judgment. Results of this process yield rough estimates of
the benefits to human health of controlling sulfur dioxide, suspended partic-
ulates, and suspended sulfates. The human morbidity costs for 1970 determined
in this manner are estimated to range from roughly $.9 to $3.2 billion.
In extrapolating the Lave and Seskin results, if 1970 air pollution levels
(total suspended particulates) were reduced by 26% (in order to reach the
primary ambient air quality standard), the savings in morbidity and mortality
costs would be $3.73 billion. This was determined here in the following
manner: First, Lave and Seskin's regression of air pollution and mortality
shows that a 26% reduction in air pollution in major urban areas would lower .
•JO
the mortality rate by 2.33 percent. Second, usinq cost figures developec
by Rice, the value for this percent reduction in mortality and morbidity
nil
81
78
the mortality rate by 2.33 percent. Second, usinq cost figures developed
thi
on
in 1963 was $2.24 billion. Third, extrapolating this value to 1970, the
estimate becomes $3.73 billion.
67
-------
Table 8. ROUGH ESTIMATES OF SOME HEALTH BENEFITS THAT CAN BE REALIZED BY THE COflTP^L
SULFUR DIOXIDE, SUSPENDED SULFATES, AMD SUSPENDED PARTICULATES
Health effect
Rough estimate of
population affected,'
million people
Estimated possible change
Rough estimate of
annual health benefits,
$ million
-------
NOTES ON TABLE 8
a. Populations-at-risk data are taken from: W.C. Nelson, et. al., Estimates
of Populations-at-Risk, Environmental Protection Agency, Division of Health
Effects Research, Research Triangle Park, N.C., In-House Technical Report,
January 1972. Population data are taken from: Statistical Abstract of the
United States - 1972 (93rd Edition), U.S. Bureau of the Census, Washington,
D.C. Data on disease rates are taken from: Current Estimates from the Health
Interview Survey, United States - 1970, U.S. DHEW, PHS, MCHS, Rockville, Md.,
Vital and Health Statistics, Series 10, No. 72, May 1972.
b. Source: C.J.< Nelson, et. al. Family Surveys of Irritation Symptoms During
Acute Air Pollution Exposures: 1970 Summer and 1971 Spring Studies. JAPCA
23.(2):81-86, February 1973.
c. Source: C.M. Shy, et. al., Ventilatory Function in School Children:
1970-]971 Testing'in New York Communities; and C.M. Shy, et. al., Ventilatory
Function in School Children: 1967-1968 Testing in Cincinnati Neighborhoods.
Both studies are in: Health Consequences of Sulfur Oxides: A Report from
CHESS, Environmental Protection Agency, Human Studies Laboratory, Research
Triangle Park, N".C. (In press).
d. Source: H.E. Goldberg, et. al. Frequency and Severity of Cardiopulmonary
Symptoms in Adult Panels: 1970-1971 New York Studies. In: Health Con-
sequences of Sulfur.Oxides: A Report from CHESS.
e. Determined by inflating the estimated health expense for a person with a
chronic condition with a limitation in activity (Source: Personal Health
Expenses Per Capita Annual Expenses, United States: July-December 1962.
Vital and Health Statistics, Series 10, Number 27, February 1966, Table 9, p.
30) to 1970 on the basis of the medical care price index (Source: Statistical
Abstract of the United States - 1972, Table 90, p. 65).
f. Source: J.F. Finklea, et. al., Aggravation of Asthma by Air Pollutants:
1971 Salt Lake Basin Studies; and J.F. Finklea, et. al., Aggravation of Asthma
by Air Pollutants: 1970-1970 New York Studies. Both studies are in: Health
Consequences of Sulfur Oxides: A Report from CHESS.
g. Based on a best judgment estimate of $15 per asthmatic attack. This
estimate might include the direct cost of medicine and other health services,
as well as the cost associated with resultant restricted activity-days.
h. Source: W.C. Nelson, et. al., Frequency of Acute Respiratory Disease in
Children: Retrospective Survey of Salt Lake Basin Communities, 1967-1970; J.F.
Finklea, et. al., Frequency of Acute Lower Respiratory Disease in Children:
et. al., Prospective Surveys of Acute Respiratory Disease in Volunteer Families:
69
-------
1969-1970 Chicago Nursery School Study; and G.J. Love, et. al., Prospective
Studies of Acute Respiratory Disease in Volunteer Families 1970-1971 New York
Studies. All of these studies are in: Health Consequences of Sulfur Oxides:
A Report from CHESS.
i. Based on the value of $11 for a restricted activity-day. This value is
determined by dividing the pelr capita income in the U.S. for 1970, by the
number of days in the year. 'Data on restricted activity days for acute
lower respiratory disease are taken from: Acute Conditions: Incidence and
Associated Disability, United States - July 1969-June 1970. Vital and Health
Statistics, Series 10, Number) 77, Table 2, p. 12; the percentage of these
restricted activity days that might be associated with populations within
SMSA's are determined from Table 11, p. 21; and the number of physician
visits is determined from TaBle 4 (p. 14) and Table 11 in much the same
manner. These physician visits are then valued by dividing the per capita
consumer expenditures for physician services (Source: Statistical Abstract
of the U.S. - 1972, Table 92, p. 66) by the physician visits per capita
(Statistical Abstract of the U.S. - 1972, Table 100, p. 69).
j. Source: D.E. House, et. al., Prevalence of Chronic Respiratory Disease
Symptoms in Adults: 1970 Survey of Salt Lake Basin Communities; C.G. Hayes,
et. al., Prevalence of Chronic Respiratory Disease Symptoms in Adults: 1970
Survey of Five Rocky Mountain Communities; J.F. Finklea, et. al., Prevalence
of Chronic Respiratory Disease Symptoms in Military Recruits, 1969-1970; and,
H.E. Goldberg, et. al., Prevalence of Chronic Respiratory Disease Symptoms in
Adults: 1970 Survey of New York Communities. All of these studies are in:
Health Consequences of Sulfur Oxides: A Report from CHESS.
k. Based on the inflation to 1970 of the annual expense to a person with a
chronic condition with no limitation 1n activity on the basis of the medical
care price index (Source: See Note e).
70
-------
This amount of $3.73 billion includes direct health expenditures and the
discounted present value of lost future earnings due to morbidity and mor-
tality. This estimate probably understates the true cost of death caused
by air pollution because: (1) since people are valued according to their
earnings, a person not earning wages is usually valued at zero; and (2)
the willingness-to-pay for improved health normally exceeds the costs
82
incurred for protection and care.
It is possible that Lave and Seskin have stronger faith in the magnitude,
sign, and statistical significance of their regression coefficients than what
their analysis would seem to support. Their many statements about the causes
of these "effects" may not be as justified as they seem to conclude. Some
of the difficulties with their work include the use of aggregate correlations,
inadequate exposure data, no personal covariate information such as smoking
83
or occupational exposure, and mobility. They are also faced with the obvious
problem of regressing aggregated data collected for different reasons. Yet
in fairness, despite the author's questionable data and extended discussion
of their results, their estimate of health costs is believed to be reasonable.
While Lave and Seskin struggle with the multiple causation dilemma and generally
ignore the multicollinearity problem, the CHESS design attempts to separate
major co-determinants of disease, primarily through analysis of variance
techniques. Multiple, repetitive health, personal covariate, meteorologic,
and environmental pollutant exposure measurements were taken in CHESS on
tens of thousands of individuals to estimate the relative effects of multiple
pollutant exposures. Even though the results seem to be reasonable, several
criticisms concerning the CHESS design have been levied. These criticisms
relate to the bias in the samples and the argument that significant socio-
economic covariates are not adequately accounted for.
From the extrapolated CHESS data and the Lave-Seskin study, it seems that a
defensible estimate of health costs for 1970 can be made. The CHESS data
extrapolated in this report estimates morbidity costs associated with selected
respiratory diseases. Lave and Seskin's work estimates morbidity and mortality
costs associated with the same respiratory diseases that the CHESS study con-
siders, as well as many others. The Lave-Seskin estimate of morbidity costs
71
-------
is inferred from their mortality estimate. In contrast, the CHESS estimate
for selected respiratory diseases is based upon direct analysis of morbidity
rates and air pollution. Since the CHESS estimate would seem to be a more
reliable estimate of those respiratory morbidity costs, it is desirable that
this estimate rather than the Lave-Seskin estimate for those same diseases
be included in the aggregate health cost estimate.
Thus, in order to make the two estimates additive, the component identified
in the Lave-Seskin estimate as morbidity and direct expenditures for respiratory,
diseases, must be subtracted from the $3.73 billion. Such an operation results
in a health estimate of $3.51 billion. Given that 73.5% of the total popu-
lation in 1970 lived in urban areas, this further reduces the estimate of
84
health costs to $2.58 billion. If then, the variance about the mean is
considered, a range of $0.7-4.4 billion can be generated for 1970. Adding
this range of estimates to the range of $.9-3.2 billion generated by extra-
polating the CHESS data, the range bf gross estimates of health costs associated
with air pollution for 1970 becomes $1.6-7.6 billion.
This gross health estimate represents the benefits that would be realized
by reducing air pollution in major1 urban areas to the particulate primary
3 3
standard of 75 yg/m , the sulfur dioxide primary standard of 80 yg/m , and
3
the assumed sulfate standard of 6-8 yg/m . Given the lack of knowledge to
the contrary, it is assumed that the estimates generated by the two studies
are additive in the sense that they have been handled here. It can1be con-
cluded that the middle of the range, $4.6 billion, is the "best" estimate
of the true costs of the adverse effects of air pollutants on human health
and longevity.
72
-------
SECTION VI
ASSESSMENT OF AIR POLLUTION DAMAGE TO MAN-MADE MATERIALS
OVERVIEW OF THE PROBLEM
A1r pollution has a variety of effects on materials—the corrosion of
metals, the deterioration of materials and paints, and the fading of dyes.
There have been a number of attempts at estimating the resultant economic
losses due to those detrimental effects of air pollution.
INDIVIDUAL STUDIES
Robbins - Electrical Contacts
It appears that the first materials effects study of major significance
was performed by Robbins (1970) of Stanford Research Institute. The study
intended to determine whether or not air pollution creates problems of economic
significance in the operation of electrical contacts, and if so, to study the
cause-and-effect relationships. Information was compiled by searching the
literature as well as by consulting manufacturers through letters, telephone
calls, and on-site visits. The major costs investigated were: (a) the direct
cost associated,with the plating of contacts with precious metals; and (b)
the indirect cost associated with the preventive measures of air conditioning
and air purification.
The types of contacts that are normally plated are used in switches, relays,
connectors, potentiometers, and commutators, which are used mostly in the
electronics and communications industries. More money is spent combating the
effects of sulfur dioxide (S02) and hydrogen sulfide (H2S) air pollution on
low-voltage electrical contacts than all the other air pollution effects on
electrical devices combined. The effects of organic gases and particulates
are of less importance. Organic gases form "frictional" polymers on sliding
contacts whereas particulate pollutants are a problem due to the fact that
the materials are excellent absorbers of moisture and corrosive agents.
73
-------
It was estimated that $20 million is spent annually on plating contacts with
precious metals because of air pollution, and $25 million is spent annually
on air conditioning and purification which act to prevent corrosion by main-
taining clean air. Another $4 million is expended annually for washing
insulators, $5 million for research and development expenditures by firms
that might be affected by air pollution, and as estimated $10 million for
losses due to failures. This totals approximately $65 million annually.
The most important conclusion drawn from this study was the fact that the
air pollution problem, with respect to electrical contacts, was not as serious
as originally estimated. It was also concluded that the cost of $65 million
is unnecessarily high because two or more individually preventive measures
are being applied simultaneously to minimize losses. It is believed that
losses will decrease as cheaper substitutes, more resistant to air pollution,
are used in electrical contacts.
ITT - Electrical Components
In a similar survey, ITT Electro-Physics Laboratories (1971) assessed the
economic impact of air pollution on electronic components. Information i
concerning pollutant-material damage mechanisms was acquired by surveying
the literature. Interviews with manufacturers and users of electrical and
electronic components provided information on actual experience. The
electronic component categories studied were: semiconductor devices, inte-
grated circuits, television picture tubes, connectors, transformers, relays,
receiving tubes, and crystals. Total damage costs for these receptors summed
to $15.5 million. This included $2.36 million for preventive costs which
included filtering air, etc., and another .$13.2 million for maintenance
costs which were the costs of cleaning, repairing, replacing and in any way
restoring a piece of otherwise defective equipment. Costs that were estimated
for various electronic devices in the Robbins (1970) study were not assessed
in the ITT study.
74
-------
As in the Robbins study, the cost of $15.5 million estimates the cost of
two or more individually effective counter-measures that are being applied
simultaneously to minimize losses. As expected, the data were very sketchy,
and extrapolations from individual cases to the nation are questionable at
best. Attempts to correlate statistically equipment failures and pollutant
levels were less than successful. While the literature reviews indicated
that sulfur dioxide should be expected to account for most of the damage
to electrical components, interviews with manufacturers revealed that par-
ticulate matter was perhaps responsible for most of the electronic compo-
nent and equipment malfunctions currently experienced.
Salmon - General Materials
^
The most comprehensive survey of the economic effects of air pollution on
materials was undertaken by Richard L. Salmon (1970) of Midwest Research
Institute. The objectives of the study were: (1) to identify the materials,
air pollutants, and environmental factors that should be studied in order to
assess the economic damage to materials caused by air pollution; (2) to
analyze systematically the physical and chemical interactions among the
variables identified in (1) for the purpose of determining cause-effect
relationships; (3) to determine, where possible, the pollutant dose-response
relationship for materials that are significant because of their relative
economic value, and indicate how this may be done where such relationships
are presently defined; and (4) to translate the pollutant and dose-response
relationship into a pollutant and dose-cost-damage function.
Information was gathered through literature searches, and by personal, mail,
and telephone interviews. Economic losses of materials were attributed
either to damaged properties or to reduced serviceability. The basic pro-
blem was to determine the extent of economic damage associated with a given
level of physical or functional damage. This problem was approached in
different ways, depending on the material and its application. In some
cases, a "percent condition" approach was adequate, with economic damage
considered to be incurred at the same rate as the physical damage. The
replacement of damaged materials was also included in this category, with
replacement presumably occurring at the 100% damage (or zero percent con-
dition level). In other cases, a cost-of-prevention or cost-of-restoration
basis proved more suitable.
75
-------
The economic value of material exposed to air pollution was calculated
as the product of the annual dollar production volume times a weighted
average economic life of the material (based on usage), times a weighted
average factor for the percent of the material that is exposed to air
pollution. The in*place or as-used value of the material was determined
by including a labor factor. The rate of economic loss was calculated
as the product of the economic value of material exposed to air pollution
times a value of interaction. The value of interaction was calculated by
estimating the difference between the rate of material deterioration in a
polluted environment compared to that in an unpolluted environment. The
interaction value is expressed as dollars lost per year. The results of
the operations described are presented in Table 9. The total value of
materials exposed to air pollution and values of interaction between the
various materials and pollutants have been combined to produce a single
figure representing the extent of economic damage attributable to air
pollutants.
To interpret the results in Table 9, it must be realized that the individual
material loss estimates were made to determine relative importance rather
than actual value. However, the sum of the economic losses, $3.8 billion
in 1968, appears tenable. Such a calculation procedure is valid only within
narrowly circumscribed limits. The difficulties lie in the problems inherent
in the technical coefficients approach—the substitution problem in particular.
The study concluded that if it is assumed that this list of materials repre-
sents only 40% of the total value of materials exposed to air pollution, and
that damage functions for the other 60% are similar, the total loss due to
chemical attack on materials by air pollution is estimated at $9.5 billion.
Salmon found that the pollutants, in decreasing order of economic importance,
and the materials they damage, are as follows:
1. Sulfur oxides: metals, cotton, finishes, coatings, building stone,
paints, paper, and leather.
2. Ozone: rubber, dyes and paints.
3. Nitrogen oxides: dyes and paints.
4. Carbon dioxide: building stone.
5. Particulates: stone, clay, and glass
76
-------
Table 9. RANKING OF MATERIAL ECONOMIC LOSSES CAUSED BY DETERIORATION
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
t
Material
Paint
Zinc
Fibers
Cement and concrete
Nickel
Rubber
Tin
Plastics
Aluminum
Copper
Carbon steel
Building brick
Paper
Leather
Wood
Building stone
Brass and bronze
Magnesium
Alloy steel
Bituminous materials
Gray iron
Stainless steel
Clay pipe
Malleable iron
Chromium
Silver
Gold
Glass
Lead
Molybdenum
Refractory ceramincs
Carbon and graphite
Economic loss,
$ million
1,195.0
778.0
358.0
316 J3T&3M
260.0
194.0
144.0
126.0
114.0
110.0
53.8
24.2
22.1
20.6
17.6
17.6
13.4
13.0
8.7
2.2
1.9
1.6
1.4
0.9
0.8
0.7
0.6
0.3
0.2
0.1
0.02
0.003
Total (Approximate)
3,800.0
77
-------
It was concluded that organic pollutants (hydrocarbons and aldehydes) are
not damaging to materials except, to some extent, to elastomers. Much
information was available on the effects of air pollutants on metals and
rubber. There was some information on fibers such as cotton and nylon and
little on paints, paper and l.eather. Virtually no information existed on
plastics, wood, wool, and concrete. There exist direct quantitative corre-
lations of damage with specific pollutants (dose-response) only for zinc
and S02, several varieties of rubber and 0-, and for cotton and S02. In
summary, the major informational shortages concern the effects of air pol-
lution on concrete, paints, fibers, and plastics.
The Salmon (1970) study is useful in summarizing the effects of air pollution
on materials and lays the groundwork for more intensive studies. Some of the
data used, such as that for formulating interaction and corrosion rates, is
of questionable validity, and particularly in a quantitative analysis such
as was used here, the results are only as good as the data used. Another
problem exists because no clear distinction was made between the stock and
flow of materials. There is very likely a mixture of the two, resulting
1n potential discounting problems. Also, only the direct effects were
investigated; the "value" of service was not assessed. As Salmon cautioned
in his report, the economic loss from material deterioration indicated sus-
ceptibility to economic loss or potential loss. The results could not be
interpreted as actually incurred economic loss. A primary purpose of the
study was the ranking of materials indicating relative measures of air pol-
lution-induced damage. This study should be very useful for setting research
priorities.
Mueller and Stickney - Rubber Products
The Battelle Memorial Institute study by Mueller and Stickney (1970) entitled,
"Survey and Economic Assessment of the Effects of Air Pollution on Elastomers,"
correlates technical information relating to the effects of air pollution on
rubber products and makes an estimate of the yearly cost of this pollution.
Costs are measured as: (1) the increased costs at the manufacturers' level
78
-------
to provide products that are resistant to atmospheric pollutants (these
are normally passed on to the consumer); and (2) the direct costs to the
consumer in the form of shortened useful life of the product. The com-
bined costs at the consumer level are used in estimating the total cost of
pollution.
One phase of the study was a review of the literature for technical information,
In the second phase questionnaires were sent to 60 rubber products firms to
determine the cost of atmospheric pollution at the manufacturing level. Of
the thirty which were returned, only about a third provided complete infor-
mation.
To estimate the cost of atmospheric pollution at the rubber product manu-
facturer's level, two independent calculations were made. One was based
on the information derived by questionnaires sent to the industry, whereas
the second calculation was based on the total of individual compounding
costs.
Extrapolation of the results recorded by individual firms to the industry
as a whole show a yearly cost of $54.0 million due to the detrioration of
rubber products caused by air pollution. To calculate the added cost at the
manufacturer's level, it was necessary to add individually estimated costs
for resistant polymers, antiozonants, waxes, protective finishes, and
wrapping. These estimated costs are listed in Table 10.
Table 10. ESTIMATED COSTS OF AIR POLLUTION RESISTANT MATERIALS
Preventive measure
Resistant polymers
Antiozonants
Waxes (50% of total)
Protective finishes
Wrappings
Research for compounding
Total
$ million
20.6
34.1
5.0
?
?
?
59.7
79
-------
It can be seen that the two figures of $54.0 million and $59.7 million
compare favorably. These are costs at the manufacturer's level. Since
industry estimates suggest that an average retail price is three times the
manufacturing cost, the compounding cost would be approximately $160-180
million at the retail level. Information on the cost of early replacement
of rubber products is summarized in Table 11.
Table 11. COSTS OF SHORTENED LIFE OF RUBBER PRODUCTS
Rubber product
Tires
Innertubes
Footwear
Mechanical goods
Medical goods
Belting
Hoses
Total
$ million
37.0
-
-
29.7
100.5
22.5
36.0
'225.7
The total annual cost of pollution, as ft affects the rubber industry, is
approximated as the sum of $170 million (the middle of the range $160-180 million)
and $225.7, or about $395 million. The first cost is that added at the
manufacturer's level, which is passed on to the consumer in the form of an
increased price for the product. The second costs is for the early replace-
ment of rubber products because of a shortened service life, a cost which must
be borne directly by the consumer. Another cost considered is the labor cost
connected with the early replacement of damaged rubber products. Mueller and
Stickney conservatively estimate this labor cost to be $75 million annually.
In conclusion, almost all damage to rubber is caused by ozone. Very little
is known about effects of other pollutants on elastomers. Virtually no
information is available on the damage threshold for rubber, so that few or
no data are available for the construction of a meaningful damage function.
80
-------
The Mueller-Stickney study lacks detail in that it deals mostly with gross
figures. Early replacement and associated labor costs of rubber products
were "ballpark estimates" at best. Also, the logic of the assumption that
retail costs are three times the manufacturers' cost can be questioned.
Theoretically, the pollution cost at the retail level should only reflect
the incremental pollution-related costs at the manufacturing level. It
is not clear that the additional cost imposed by air pollution necessarily
result in an increase in overhead and marketing costs or any other costs
comprising the factor of "three times." Nonetheless, given the lack of
better information, the estimate of $475 million annual cost determined by
Battelle is considered acceptable in this report.
Spence and Haynie - Paints
Recent work by Spence and Haynie (1972) investigated the deterioration of
exterior paints by: (1) particulate matter primarily; and (2) the interaction
of particulate matter and sulfur oxides. The associated potential economic
loss to manufacturers and consumers because of this deterioration was then
estimated. Their initial review of the literature made them acutely aware
of the almost total lack of information on dose-response relationships with
respect to air pollution damage to paints. This informational gap confirms the
findings of the Salmon (1970) study mentioned earlier. That study concluded
that sulfur dioxide and particulate matter play an important role in the
chemical deterioration of modern-day exterior paints. These pollutants serve
to promote the chemical deterioration of exterior paints.
Spence and Haynie referred to the soiling studies by Michel son and Tourin
(1967 and 1968) and Booz, Allen and Hamilton (1970) in order to derive
information on the frequency of house repainting as a function of suspended
particulate concentrations. The results of Michelson and Tourin are sum-
marized in Table 12, and Booz-Allen in Table 13. In Table 12, it is shown
that the maintenance intervals of the years between repaintings were observed
to decrease as the particulate concentrations increased. While there appears
to be a positive relationship between frequency of repainting and particulate
concentrations, the questionable data generated in the Michelson and Tourin
study (see Section IX) make use of it questionable.
81
-------
Table 12. INTERVAL FOR EXTERIOR REPAINTING AS A FUNCTION
OF PARTICULATE CONCENTRATIONS*
City
Steubenville
Union town
Suit! and
Rockville
Fairfax
Parti cul ate
concentration,
yg/m
235
115
85
75
60
Maintenance
interval ,
year
0.88
1.89
2.93
3.62
3.90
Maintenance
frequency,
number/year
1.14
0.53
0.34
0.28
0.26
Reciprocal of maintenance .interval in years
* Source: J. W. Spence and F. H. Haynie. Paint Technology and Air
Pollution: A Survey and Economic Assessment. Environmental
Protection Agency Publication No. AP-103, Research Triangle Park, M. C.
(February 1972).
Table 13. FREQUENCY FOR EXTERIOR WALL PAINTING IN PHILADELPHIA
AS A FUNCTION OF PARTICULATE CONCENTRATION*
Parti cul ate concentration
ranges, yg/m
<75
75-100
100-125
>*25
Mean annual
frequency
0.28
0.35-
0.35
0.29
Standard error
of mean
0.016
0.053
0.041
0.055
* Source: J. W. Spence and F. H. Haynie. Paint Technology and Air Pollution:
A Survey and Economic Assessment. Environmental Protection Agency
Publication No. AP-103, Research Traingle Park, N. C. (February 1972).
82
-------
Many of the analytical and statistical problems encountered in the Michelson
studies were minimized in the Booz, Allen, and Hamilton study in Philadelphia.
Table 13 shows their maintenance frequencies for exterior wall painting by
mean annual particulate concentrations. Spence and Haynie suggest that the
fact that the proportion of households with incomes less than $6000 increased
with pollution level is a factor that tends to counteract the effect of sus-
pended particulates on paint life. And it is this confounding factor that
partially explains why no statistically significant difference in painting
86
frequency as a function of particulate level was detected.
Table 14 outlines the method used by Spence and Haynie to establish their
estimate of the potential annual consumer cost for repainting that can be
attributed to pollutant damage in urban areas. Estimates of losses are
developed for four paint classes: coil coating, automotive refinishing,
maintenance, and household. Expected service lives were best "guesstimates"
except for household, where expected service lives for rural and urban areas
were estimated from the Michelson and Booz-Allen studies. Estimates of the
distribution of paint in rural versus urban areas were also best judgments
of the authors. Labor factors were estimated from best available information
concerning professional painting costs. The total value lost at the retail
level is estimated as $704 million. Household paints, with a value loss of
$450 million, represent over 75% of the total estimate.
The authors readily admit that these calculations result in only a rough
approximation, and that more information on dose-response, expected service
life, maintenance frequency, and labor factors, must be obtained if more
reliable economic assessment is to be made. Given the lack of better infor-
mation, the figure of $0.7 billion will be taken as more defensible than the
estimate for paint damage of $1.195 billion generated in the Salmon study.
Fink - Corrosion
A recently completed study by Fink, Buttner, and Boyd (1971) attempted to
develop a more realistic assessment of the added cost of corrosion damage
to the nation resulting from the exposure of metallic systems and structures
to polluted atmospheres. In the approach used, applicable national shipment/
83
-------
Table 14. ECONOMIC ASSESSMENT OF AIR POLLUTION EETERIORATION OF EXTERIOR PAINTS*
Column
1
Exterior
paint
classes
Coil
coating
Automotive
refinishing
Maintenance
Household
Total
Column
2
Value at
manufacturer's
level ,
$ million
40
150
100
485
775
Column
3
Area
Rural
Urban
Rural
Urban
Rural
Urban
• Rural
• Urban
Column
4
Expected
service
life,
years
20
15
5
4
5
4
6
3
Column
5
Maintenance
frequency,8
per year
0.05
0.07
0.2
0.25 ,
0.2
0.25
0.17
0.33
Column
6
Estimated
distribution
of paint,
1 population
30
70
30
70
30
70
40
60
Column
7
Paint
consumed
in urban
areas,
77
74
74
74
Column
..8
Value of paint
exposed in
urban areas,
$ million
31
111
74
359
575
Column
9
Service
life,c
% loss
25
20
20
50
Column
10
Loss at
manufacturer's
level,
$ million
8
22
15
180
225
Column
11
Labor
factor
2
4
4
3
Colunn
12
Loss at
consumer's
level ,
$ million
16
88
60
540
704
00
Notes:
Maintenance frequency is the reciprocal of expected service (Column 4).
Calculation of I coil coatings consumed in urban areas (Column 7):
Rural Areas:
Urban Areas:
Col. S Col. 6
0.05 x730- 0.015 0.049 _ inn _
0.07 x .70 -0.049 O6Txl°°
0.064
c Calculation of i service life loss for coil coating (Column 9):
Rural Areas: 20 years I Service Life loss:
Urban Areas: 15 years ,n ,Q
M2Q x 100 - 25
* Source: J. W. Spence and F. H. Haynie. Paint Technology and
Air Pollution: A Survey and Economic Assessment. Environmental
Protection Agency Publication No. AP-103, Research Triangle
Park, N.C. (February 1972).
-------
value data from the U.S. Department of Commerce were employed to compute
average pollution costs on a national basis. Of these data, individual cor-
rosion costs were calculated for nine major categories that survived the
screening process (see Table 15). These categories were regarded as the
most sensitive to and most damaged by air pollution corrosion. The total
of these individual estimates was $1.45 billion.
Two calculations were considered by Fink et. al. for assuming up the
economic marginal cost of corrosion caused by air pollution over the typical
cost in clean air service. The first calculation estimated the extra amount
of protection and maintenance expense required in polluted atmospheres to
prevent serious corrosion attack. The second calculation estimated the
cost due to the shortened life of the structural system caused by corrosion
from polluted air. Thus, costs associated with corrosion include both main-
tenance costs (i.e., painting) and early replacement costs.
The general approach compared maintenance painting costs in clean and pol-
luted air. The steps followed were: (1) the total amount of each item in
use was established; (2) this amount was converted into exposed surface
area; (3) the portion of this area exposed to polluted air was estimated;
(4) the annual extra cost of protection by paint, per unit area, was cal-
culated for each system; and (5) the area and annual cost figures were com-
bined to obtain the total national loss for each item.
As shown in Table 15, the top four structural systems are primarily constructed
of galvanized steel. Based on this corrosion study, the accelerated corrosion
of zinc by sulfur dioxide accounts for more than 90% of the national economic
burden imposed by air pollution corrosion.
The Fink, et. al. corrosion study makes a systematic and reasonable estimate
of the costs of corrosion to materials and the costs of related losses. Yet,
several crittcisms can be levied against this study. The first criticism re-
lates to the problem that typifies most materials studies: no attempt is made
to relate material damage to actual levels of air quality. Material losses are
identified with "clean" or "polluted" areas which are not defined according to the
severity or kind of air pollution. As a general rule, any study of air pol-
lution damage should attempt in its analysis to relate damages to actual
85
-------
Table 15. ANNUAL COST OF CORROSION BY AIR POLLUTION
DAMAGE OF EXTERNAL METAL STRUCTURES, 1970*
Structural System
Outdoor metal work
Chain link fencing
Pole line hardware
Galvanized wire
and rope
Steel storage tanks
Bridges
Street light
fixtures
Power transformers
Transmission towers
Total
Useful life,
years
45
30/20
30
20
50/11
30
20
30
30
Cost basis
Maintenance
Maintenance
Repl acement
Replacement
Maintenance
Maintenance
Maintenance
Maintenance
Maintenance
Cost, $
million
914.0
166.0
161.0
112.0
46.3
30.4
11.9
7.5
1.5
1,450.0
Percent of
total
63.0
11.5
11.1
7.7
3.2
2.1
0.8
0.5
0.1
100.0
*Source: F.W. Fink, F.H. Buttner and W.K. Boyd. Technical Economic
Evaluation of Air Pollution Corrosion Costs in Metals in the United
States. Final Report from Battelle Memorial Institute, Columbus, Ohio,
to the EPA, Research Triangle Park, N.C.. February 1971.
86
-------
levels of air quality, or at least to sound proxy for air quality. The
use of "clean" and "polluted" as measures of the severity of air pollution,
without further definition, is not acceptable. Another problem in the Fink
corrosion study is the fact that the role of relative humidity in the
corrosion and deterioration of materials is not considered in the analysis.
Such an important parameter cannot be ignored.
Gillette - SO Effects on General Materials
A
Gillette (1973), employing basically the same general approach as that used
by Battelle, attempted to rectify in his assessment of materials damages, some
of the problems in the Fink study. Gillette's study, being somewhat broader
than that of Battelle, was designed to assess the economic damages from sulfur
oxides (SO ) to man-made materials. Where Fink assumes that about 80% of materials
y\ .<
are located in "polluted areas," Gillette assumes that materials are distributed
according to human population. Information on population distributions, coupled
with sulfur dioxide data for about 150 SMSA's for the years 1968-1972, provided
a basis for estimating materials populations-at-risk.
Gillette then integrated measures of the average annual relative humidity
by SMSA into his analysis. This consideration of relative humidity is impor-
tant because the corrosion damage function shows relative humidity to be more
important than sulfur dioxide in causing corrosion. Using the best available
damage function data for corrosion and paint deterioration, Gillette estimated
economic losses for the inclusive years. Gillette estimated for 1970 that SO
/\
damage (where sulfur dioxide acts as a surrogate for all sulfur damaging com-
pounds in the atmosphere) to metals and paints was approximately $.4 billion.
Gillette concluded from his analysis of available dose-response data on
the effects of SO on susceptible materials that SO effects on textiles,
X X
building materials, leather and paper products, and dye fading are probably
negligible from an economic standpoint. Gillette based this conclusion on
the following considerations: (1) many materials are exposed primarily to
indoor environments where the exposure is to much lower levels of air pol-
lutants than exposure outdoors; (2) current S02 levels are generally lower than
several years ago, presumably because of the substitution of cleaner fuels; and
(3) the use-life of many materials 1s quite short.
87
-------
Gillette's study, while a substantial improvement upon earlier efforts to
assess materials losses due to air pollution, is not necessarily a definitive
statement of the effects of sulfur oxides and derivatives on man-made materials,
For example, it is difficult to say to what extent SCL is a good surrogate
for all atmospheric sulfur-damaging compounds. It is possible that atmospheric
sulfates are significant in the deterioration of materials, and thus, should
have been accounted for in a more explicit manner in Gillette's analysis.
Also, it appears that in some cases, tenuous conclusions were drawn where
informational gaps existed. For example, some of the assumptions made con-
cerning the paint damage function were based on very "soft" information.
Yet even with these caveats, the estimate of SO -caused material losses
J\
developed by Gillette i's believed to be more realistic and more defensible
than estimates of losses developed in other materials studies.
Salvin - Dye Fading
Victor S. Salvin (1970) of the Unitersity of North Carolina conducted a study
of the economic effects of air pollution on textile fibers and dyes. The
objectives of his study were: (1) to conduct a comprehensive survey to
identify and document known and suspected air pollution-induced effects
on various textiles and dyes; and (2) to assess the economic effects of air
pollution on textile fibers and dyes.
The status of the problem was discussed with manufacturing and industrial
representatives as to the prevalence, mechanisms, prevenative measures,
and research costs. The suppliers of dyes were contacted for costs of
dyes and dyeing processes (preventative measures). A technology committee
of each industry served as a clearinghouse for information. The production
figures for each industry, the awareness of the industry, and actions by
major manufacturers in offering goods of increased performance were documented.
The costs in this case are those of research, quality control, more expensive
dyes and textiles, and the associated, more costly production techniques. The
additional annual replacement costs of air pollution-damaged textile and fiber
products were also estimated. Preliminary economic costs of the fading of
dyes on textiles due to air pollutants are shown in Table 16.
-------
Table 16. ESTIMATED COSTS OF DYE FADING IN TEXTILES
Pollutant
N0x
'
°3
Effect
Fading on acetate and triacetate
Fading on viscose rayon
Fading on cotton
Yellowing of white acetate-Nylon-Spandex
Subtotal
Fading on acetate and triacetate
Fading on Nylon carpets
Fading on permanent-press garments
Subtotal
Total
$ million
$ 72.800
21.600
22.050
5.650
$122.100
24.985
41.500
17.050
83.535
$206.000
89
-------
The costs of fading of dyed fabrics by oxides of nitrogen (NO ) and ozone
(03) generally are based on: the increased cost of dyes more resistant to
fading; the cost of inhibitors for cheaper dyes; the cost of research; the
cost of quality control related to the use of more expensive dyes; and, the
costs to consumers and sellers with respect to any reduction in product-life.
As weak as this information is, it appears as if this is the only information
of its type available. There are little or no supporting economic data and none
that would contribute to the construction of any damage function. Given the
serious concern over the validity of many of the assumptions made in his study,
it is believed that the estimate of $206 million is a very rough approximation
of the actual damage of air pollution on textile and fiber products. The
conservative nature of this estimate of corroborated by the absence of data
in Salvin's study on the effects of sulfur and nitrogen oxides and acids and
87
other particulate matter on the deterioration of textile fibers and dyes.
These effects are assumed to be measured in the "fibers" category in Table 9.
NATIONAL ESTIMATES OF MATERIALS LOSSES
The eight studies reviewed in this report offer substantial evidence on which
a reasonable national gross damage estimate can be based. A total of $1.8
billion is derived if the following are summed: $.5 billion (rounded off)
from the Mueller and Stickney elastomer study; $.4 billion from the Gillette
SO study; $.2 billion from the dye-fading study; and $.7 billion from the
/\
Spence-Haynie paint study. In summing these, it is assumed: (1) that the
Gillette study is more defensible than the corrosion study by Fink, et. al.;
(2) that the Gillette study does not significantly overlap with the Spence-
Haynie study that focuses primarily on the effects of particulates on painting
rates; and (3) that the Gillette study adequately accounts for damages to-
electrical contacts and components. Then if the materials analyzed.in these
studies—zinc, paints, synthetic and natural rubber, carbon and alloy steel,
fibers, cement and concrete, plastics, building brick, paper, leather, wood,
and building stone—are subtracted from the Salmon study, in addition to the
metal categories of aluminum, copper, stainless steel and lead which were
regarded by Fink, et. al. as not significantly affected by air pollution, the
90
-------
total remainder in the Salmon study becomes about $.4 billion. Adding $.4
billion to the $1.8 billion derived earlier, the total gross damage estimate
for material losses in 1970 due to air pollution is approximated at $2.2
billion.
The Salmon study develops estimates of losses for the various material
receptor categories, but where more indepth research was done, it is believed
that those estimates should be given priority. It should be reasonable
to assume that this national estimate of materials losses is representative
only insofar as it falls within an estimated range. Given the lack of more
objective evidence, the percent variation expressed with the property value
estimate will also be applied to the materials receptor- By applying that
same variation (about 43%), a range of $1.3-3.1 billion is generated, with
a "best" estimate of $2.2 billion. Given the nature of the studies reviewed,
this estimate should be viewed not as the "true" cost of material damage
but indicative of the general magnitude of damage in 1970.
91
-------
SECTION VII
ASSESSMENT OF AIR POLLUTION DAMAGE TO VEGETATION
OVERVIEW OF PLANT SURVEYS
Damage to vegetation as a result of air contamination has been recorded
in the United States since the turn of the century. What was once a problem
associated only with point sources has evolved into an air pollution problem
more commonly associated with urban expansion. The continued commercial and
noncommercial production of crops and forests in many areas has been jeopar-
*.
dized and in some locations discontinued.
In general, air pollution adversely affects plants in one of two ways. Either
the quantity of output or yield"is reduced or the quality of the product is
lowered. The biological response of a plant to a fumigation by air pollution
is a function of a complex mix of biological, environmental, and climatic
factors. Such factors include, among others: level and duration of pol-
lution exposure, age of plant, genetic sensitivity of the plant, light, rela-
tive humidity, soil moisture and fertility, and general health o%the plant.
Given this kind of information, one could construct a reasonable, physical dose-
response relationship—the physical damage function. The translation of this
function into an^economic damage function is fraught with another complex set
of variables. Important aspects that one must consider here include: time
and growing season, market value of the plant affected, the aesthetic value
that might be attached to the plant, the nature of the harvesting and culturing
costs for the particular affected crop, the adaptability of the site for
growing a different crop, and the value of the site for alternative uses.
Two general approaches have been used to assess the amount of economic loss
resulting from plant damage by air pollution. One general approach has been
to survey air pollution losses on a statewide basis by using the existing man-
power of county agricultural agents and commissioners at the local level. From
these local estimates of damage, extrapolations can be made to the national
level to arrive at a crude estimate of gross damages. Another general approach
is one of incorporating data on pollutant emissions, crop statistics, and
92
-------
meteorological parameters into a predictive model of plant losses. Such
models that incorporate damage factors are then subject to continual refine-
ment as: (1) greater knowledge about dose-response relationships is gained;
and (2) the true situation at the local level is better defined.
The major strengths in using statewide surveys are: (1) an existing manpower
can be utilized for achieving continual coverage over an area; (2) agents
at the local level have an established rapport with growers in that area,
are familiar with any crop peculiarities, and are probably knowledgeable of
any sources of pollution in the locale; and (3) a field coordinator
supplies expertise to the reporters and provides some degree of standardization
in reporting losses. One problem arises all too often: unjustifiable con-
clusions can be made on one-year estimates when several years are needed to
make accurate assessments of damages.
INDIVIDUAL STUDIES
Middleton and Paul us - California, 1955
The use of manpower at the local level was first used in a California survey
performed in 1949. A second survey in 1955, as reported by Middleton and
Paulus (1956), was designed to determine the location of injury, the crops
injured, and the toxicant responsible for the damage. Specialists in agri-
culture throughout the state were trained as crop survey reporters. The
survey covered four categories of crops: field, flower, fruit, and vegetable.
Lacasse - Pennsylvania, 1969 and 1970
A similar program was established in 1969 in Pennsylvania and reported by
Lacasse-Weidensaul-Carrol (1970). As in California, a training course was
held to teach trained observers how to identify and evaluate air pollution
damage to plants. The objectives of the survey were: (1) to estimate objec-
tively the total cost of agricultural losses due to air pollution in Pennsylvania;
93
-------
(2) to determine the relative importance of the various pollutants in
Pennsylvania; (3) to survey the extent of the air pollution problem in
Pennsylvania; and (4) to provide a base for estimating the nationwide impact
of air pollution on vegetation and for guiding research efforts.
A professional plant pathologist was enlisted to coordinate the field survey.
He assisted reporters in the detection and evaluation of air pollution damage
to crops and performed independent field surveys in areas where sources of
pollution were located. Commercial and noncommercial plants were studied.
Past episodes also were investigated for purposes of detecting possible trends.
Estimates of losses were based on the amount of plant damage, crop value, and
production costs incurred by harvest time. Direct losses to producers or
growers included only production costs. Indirect losses included profit
losses, costs of reforestation, grower relocation costs, and substitution of
lower-value crops for higher-value crops. Costs associated with the destruction
of aesthetic values, erosion, and resultant stream silting, damage to watershed
retention capacity, and farm abandonment, were not considered.
Ninety-two field investigations were made as part of the Pennsylvania study.
The amount of direct losses uncovered 1n the survey were estimated at more
than $3.5 million. The air pollutants, by decreasing Importance, were: ox1-
*
dants, sulfur oxides, lead, hydrogen chloride, particulates, herbicides, and
ethylene. The crops most affected, 1n decreasing order of Importance, were:
vegetables, fruits, agronomic crops, lawns, shrubs, woody ornamentals, timber,
and commercial flowers. Indirect losses were estimated at $8 million, of
which $7 million reflects profit losses, $0.5 million for reforestation
of land, and $0.5 million for grower relocation costs. In total, air
pollution losses in Pennsylvania for 1969 were estimated at approximately
$11 million.
The major criticisms of the effort 1n Pennsylvania reflect the state-of-the-
art. Little 1s known of the extent to which home garden plantings and flowers
are being affected by air pollution; and, 1f they are affected, at what price
they should be valued. The economics of assessing losses 1s somewhat question-
able 1n that grower profit losses were not Included as direct costs, and 1t
94
-------
was not clear as to what constituted an annual cost and what did not. Also,
the translation of physical injury into economic loss is somewhat subjective
and has not been standardized.
Lacasse (1971) made an attempt in 1970 to again survey Pennsylvania. In
using the same concepts of what constituted a "cost" and what did not as in
the previous year's survey, Lacasse estimated direct losses to be $218,630 and
indirect losses of $4,000. In explaining the low estimate of losses, Lacasse
explains that "...the reason for the lack of widespread air pollution injury
to vegatation during the 1970 growing season may have been due to fewer inver-
88
sions and to no unfavorable growing conditions when air stagnation did occur."
Feliciano - New Jersey, 1971
Similar surveys have also been made in New Jersey and in the New England states.
In general, these surveys suffer from many of the same deficiencies as the
Pennsylvania survey. Feliciano (1972) reported that losses to agriculture
in New Jersey due to air pollution were estimated at $1.19 million in 1971.
As in the Pennsylvania surveys, profit losses were not included. A total of 315
reported air pollution incidences were investigated and documented during the
period of the New Jersey survey. A "rule of thumb" evaluation method developed
by Millecan (1971) was used for estimating losses. As Feliciano describes
it, "Where visual inspection of the overall leaf surface of the plants indi-
cated 1 to 5 percent injury, a 1 percent loss was applied for that crop. A
leaf surface injury ranging from 6 to 10 percent was given a 2 percent loss;
11 to 15 percent injury, a 4 percent loss; and 16 to 20 percent injury, an 8
on
percent loss." Estimates of losses were based on the crop value of the
acreage affected.
The damaging pollutants listed by decreasing importance were: peroxyacyl
nitrates (PAN); hydrochloric acid mist and chlorine gas; ethylene; sulfur
dioxide; ammonia; fluoride; and particulates. The first two accounted for
80% of the total incidences reported. Vegetables and field crops
experienced about 85% of the total economic loss reported. Damage
was reported in 16 of the 21 counties in New Jersey.
95
-------
Pell - New Jersey, 1972
To obtain a better understanding of the year-to-year variation in plant losses
caused by air pollution, Pell (1973) continued the work initiated by Feliciano
in 1971. Pell estimated that direct losses of agronomic crops and ornamental
plantings for the 1972-73 growing season were approximately $130,000. As in
the study by Feliciano, costs associated with crop substitution and yield
reductions, were not accounted for. The damaging pollutants, listed by
decreasing importance, were: oxidants, 47% of crop losses; hydrogen fluoride,
18%; ethylene, 16%; sulfur dioxide, 4%; and anhydrous ammonia, 1%. Surprisingly,
the damage reported in Pell's survey was only 11% of that reported in Feliciano1s
1971-72 survey in New Jersey. The significant year-to-year variation is
attributed to altered environmental conditions rather than to decreased air
pollution concentrations. For example, it is believed that the unusual rain-
fall patterns in 1972 placed the plants under water stress and thereby protected
them from air pollution injury.
V
Naegele - New England, 1971-72
Naegele, et. al. (1972) reported on a field survey of agricultural losses in
the New England region caused by air pollution. Some 83 investigations were
made in 40 counties of the six New England states. Direct economic losses
for the 1971-72 season were estimated at approximately $1.1 million. Esti-
mates of economic losses were based on grower costs, crop value at the time
of harvest, and the possibility of crop recovery following the pollution inci-
dent. Here, direct losses also include grower profit losses. It was determined
that fruits, vegetables, and agronomic crops suffered the greatest losses, and
that over 90% of the damage could be attributed to oxidant air pollution.
Millecan - California, 1970
An approach similar to that used in Pennsylvania, New Jersey, and New England
was employed by Millecan (1971) to survey and assess the air pollution damage
to California vegetation in 1970. Because of foreknowledge of the distribution
of air pollution problems, efforts were concentrated in the Los Angeles Basin,
96
-------
San Joaquln Valley, and the San Francisco Bay Area. Estimates of losses were
confined to 15 of the 58 counties in the state. Plant injury from air pol-
lution was observed in 22 counties. Ventura County, with a loss of almost $11
million, experienced the greatest economic crop loss for any one county. Losses
of citrus production in the Los Angeles Air Basin accounted for over $19 million
of the total monetary loss of almost $26 million. The monetary loss estimate
does not include losses attributed to reduction in crop yield or growth except
for losses of citrus and grapes. Nor were monetary losses to native vegetation
including forests or to landscape plantings estimated.
As expected, photochemical smog accounted for most of the economic losses.
Photochemical smog is composed of oxidant-type pollutants like ozone and PAN
that are derived from the interaction of nitrogen oxides and hydrocarbons in
the presence of sunlight. Analysis of field reports showed that six
pollutants accounted for the following percentages of plant injury: ozone,
50%; PAN, 18%; fluorides, 15%; ethylene, 14%; sulfur dioxide, 2%; and
particulates, 1%.
Benedict - Nationwide Survey, 1969 and 1971
A major study to estimate plant losses caused by air pollution was undertaken
by H. M. Benedict (1971) of Stanford Research Institute (SRI) in 1969. SRI
has developed an estimate of the annual economic losses to agriculture in all
regions of the United States resulting from damage to vegetation by air pol-
lutants. Special emphasis was placed on those losses ascribable to automotive
emissions.
Their work progressed in the following manner: First, counties were selected
in the United States were the major air pollutants—oxidants (ozone, PAN, and
oxides of nitrogen), sulfur dioxide, and fluorides—were likely to reach plant-
damaging concentrations. This selection was based on fuel consumption and the
existence of large single-source emitters. Second, relative potential severity
classes of the pollution in each county were then estimated, based on emissions
area, and potential pollution episode days. Third, crop value estimates were
completed for these counties. This necessitated calculating the dollar value
of grass hay produced and of pastures. Fourth, estimates of the potential
97
-------
annual value of forests and the annual maintenance costs of ornamental plantings
were completed and apportioned by area and population. Fifth, a continuing
literature review provided information on the relative sensitivity of different
plant species to the selected pollutants, so the percentage loss that might
be expected to crops and ornamental plantings in the most severely polluted
counties could be determined. Sixth, tables were then prepared showing the
percentage loss that might be expected to crops and ornamentals in counties and in
the different pollution classes described in the second step above. And seventh,
these factors were then applied to the value of the crops, forests, and orna-
mentals grown in the polluted counties, and the dollar loss value for each crop
in each county was recorded. From this, state, regional, and national esti-
mates were obtained.
Thus, dollar loss estimates for agricultural crops and ornamentals were deter-
mined using the following equation:
DOLLAR LOSS = (Plant Value) x (Plant Sensitivity) x (Pollution Potential)
When the loss factors for the various pollution intensities in the 551 selected
counties were applied to the determined crop and ornamental values, the total
annual dollar loss to crops in the United States for 1964, as shown in Table
17, was calculated to be about $85.5 million, and the loss to ornamentals,
about $46 million.
The significant weaknesses in the study seem to be: (1) given the paucity of
knowledge in the literature on pollutant-yield relationships, many of the damage
factors were probably "best guesses" and thus are subject to refinement. The
same can be said of the determination of relative sensitivity; (2) the system-
atic application of factors to determine crop and ornamental plant values,
especially the latter, does not allow for individual variation, thus one would
expect a great deal of variation in error in any particular county estimate;
and (3) ornamentals were under-valued in that only replacement costs were used
as a proxy for aesthetic values. Also, the values of recreational areas were
not assessed. A major benefit from this study is the accumulation of good
background data for the development of more sophisticated predictive models
for estimating losses when better data on dose-response becomes available.
98
-------
Table 17. PLANT LOSSES DUE TO AIR POLLUTION
($ million)
Crops
Ornamental s
Total
Oxidants
78
43
121
so2
3.3
3
6.3
Fluorides
4.3
0.2
4.5
Total
85.6
46.2
131.8
NATIONAL ESTIMATE OF PLANT DAMAGES
The loss estimate of $132 million generated in the SRI study for 1964 is the
most defensible of those reviewed. Also, the SRI estimates are consistent with
the individual estimates for California, Pennsylvania and elsewhere. Because
the SRI study attempted to grapple with losses of ornamentals, the state estimates
generated by the predictive model are consistenly higher than those developed
through statewide surveys. But even then, the SRI estimate still reflects a
lower bound of the true plant-associated losses due to air pollution. This is
so because the losses resulting from reduction in yield are largely ignored.
Also ignored are costs associated with grower relocation, crop substitution,
losses in productivity, and denudation of land and resultant erosion. Due to
the lack of adequate knowledge on many aspects of air pollution effects on
plants, there are many inadequacies inherent in all reported efforts to esti-
mate plant losses due to air pollution. Although all estimates have their
shortcomings, the studies discussed above represent the current state of the art.
By updating the 1964 estimate of $132 million to 1970, the estimated cost of
air pollution damage to vegetation is estimated to he approximately $150 million.
This is based on the assumption that the same percentage value of crops are
90
lost in 1970 as was lost in 1964. Implicit in this assumption is that the
value of ornamental plantings has increased the same as cash crops. By
rounding this estimate off to $0.2 billion and then assuming that $0.2 billion is
representative of a range, a range of 50 percent or $0.1-0.3 billion will be
assumed, with a "best" estimate of $0.2 billion for 1970.91
99
-------
SECTION VIII
ASSESSMENT OF THE EFFECTS ON AIR POLLUTION ON AESTHETIC PROPERTIES
OVERVIEW OF THE PROBLEM
This very nondescript receptor category is concerned mostly with the organ-
oleptic aspects of air pollution—those pertaining to sight and smell, fiiven
the level of public esposure, it can be considered public knowledge that air
pollution restrains progress toward an environment congenial to aesthetic and
no
other socially conditioned needs. Peckham's review of the literature has
provided a good starting place in understanding the seriousness of this aspect
of the air pollution problem.
One possible effect of air pollution is the deterioration of materials with
historic or artistic significance, such as paintings, statuary, and rare books.
Air pollution that reduces visibility and obscures vistas can also have a
depressing psychological effect on individuals. Noxious odors represent another
series of effects that are considered here to be aesthetic effects of air pol-
lution.
Aesthetic effects also belong to the calculus of pollution damages because of
values that could be attached to prevention or avoidance. For example, the
New York City Public Library spent $900,000 between 1952 and 1967 to microfilm
books that were in an advanced state of deterioration due significantly to air
93
pollution. Part of this expenditure represen
to pay to avoid book losses from air pollution.
93
pollution. Part of this expenditure represents what the library was willing
In general, man wants an environment congenial to his aesthetic and phycho-
logical needs. Yet air pollution restrains progress toward such an environ-
ment. Odors from various industrial sources deprive many of the full enjoyment
of their property. Particulates dangerously diminish visibility. Oxides of
sul,fur accelerate the decay of honored works of art and statuary. Emissions
from automotive combustion and their resultant atmospheric interactions injure
the trees that adorn our urban arteries and often cause watering of eyes, thereby
diminishing our quality of life.
100
-------
ODORS
Odors have historically generated a high level of public concern. This is
clear from opinion surveys. For example, in the St. Louis survey,94 926 out
of 1361 complaints received during the 1958-1962 period pertained to odors. In
an opinion survey taken in Clarkston, Washington,95 91% of the respondents made
a similar identification of air pollution with odors.
Regardless of the area they cover, odors can deprive people of the full use
and enjoyment of their property. A survey of the court records might be a
good way to determine the seriousness of this deprivations For example, 31
homeowners brought suit against the Weyerhauser Company to recover damages
caused by odors emitted from the company's kraft pulp mill in Elkton, Maryland.
The testimony was convincing enough in that the court awarded the plaintiffs
the amount of about $18,000. In another case, plaintiffs were awarded over
$35,000 as a result of a suit filed against an industrial concern for emanating
odorous pollutants that resulted in hospitalization costs, loss in earning, and
97
the loss of enjoyment of their properties. Other similar judgments are
extensively recorded in the legal records. Yet some decisions from recent
cases seem to indicate that the rights of individuals to odor-free air have
not been well established. For example, a recent ruling by the United States
no
District Court in Cincinnati ruled that odors from a dump site are not illegal.
A suit had been brought by 22 property owners who complained that odors from
the disposal site fouled their air and deprived them of their rights to clean
air and the free use of their property. The judge ruled that the United States
Constitution does not guarantee citizens the right of protection of their
environment. A similar ruling was made in a Michigan court case where the
plaintiff claimed that hog odors from a neighbor's farm had an "adverse
99
effect" on her sick husband.
Not surprisingly, the major obstacle to initiating action against odor problems
begins with measurement. As reported by Dravnieks, "Odor dimensions are inten-
sity, detectability, acceptability, and quality (character). Human response
to odors is not linearly related to the concentrations of odorants in air
and relates to the chemistry of odorants in a complex way. The response is
101
-------
influenced by criteria which the individuals use to interpret odor sen-
sations and by the form of response expected or accessible, especially in
community setting." Methods for evaluating odor problems have been developed
by Copley International Corporation. These are to be field tested in
selected metropolitan areas throughout the country. Copley found that people
were generally unaware of any adverse economic effects of odors pollution,
in?
or they did not believe such effects existed. it is likely that people
are generally unwilling to admit these kinds of effects do exist.
VISIBILITY
Air pollution not only affects the olfactory senses but also the sense of
sight. Serious hazards to transportation are created by visibility-reducing
air pollutants; they cloud the landscape with haze and smog and discolor the
sky. The visibility-restricting air pollutants are particulates and nitrogen
dioxide.
Particulates in the atmosphere can affect visibility in two ways—either by
absorption or by scattering of light. The nature and magnitude of the effect
are functions of the chemical composition of the particle, parti oilate size,
shape, and concentration. Particulates also exert an indirect effect upon
visibility by facilitating the formation of fogs and by slowing their dissi-
pation, making travel difficult and hazardous.
Poor visibility can also result 1n accidents and disruptions in transportation.
It has been estimated that adverse effects of air pollution on air travel cost
103
from $40 to $80 million annually. The Civil Aeronautics Board, in reviewing
1962 aircraft accidents in the United States, found at least six to be directly
due to what they called "obstruction to vision" caused by smoke, haze, dust,
and sand. Delays in the movement of air traffic are quite common at
major commercial airports during times of poor visibility.
As Peckham has said, "People want safe and dependable transport. They want to
be able to travel without extraordinary risk or delay. Furthermore, most
people also want access to pleasing scenery and bright, clear weather with
102
-------
sunshine. Air pollution can often defeat these wants by depressing visibility,
blocking sunshine, and intensifying fog. This seems clear from the evidence.
What is not so clear, however, is the monetary magnitude of the injuries
suffered."105
Vars and Sorenson (1972) attempted to measure externalities associated with
the visibility-reducing air pollution generated by open field burning in
Oregon's Willamette Valley. Utilizing a market study approach, they con-
structed a theoretical model to explain the relationship between air quality
and consumer behavior, or more specifically, the consumption of recreation-
related activities. The consumer was viewed in the model as literally pro-
ducing consumption activities by combining market commodities and time in
different combinations so as to create an activity he then consumes. While
the theory was straightforward, albeit complex, it was not amenable to
empirical analysis. There was not only the problem of the lack of a priori
information about the comsumption activity production functions, but also a
second difficulty arising from the lack of a priori information on what
activities would be affected negatively by a deterioration in air quality and
which activities would be affected positively, as they perform the role of
substitutes.
These empirical problems led to a redesign In research plan. The subsequent
investigation employed multiple regression analysis in examining the deter-
minants of the following recreation activities: (1) swimming pool use; (2)
golf course attendance; (3) number of visitors at the state Capitol; and
(4) the number of overnight campers at a selected state park. By hypothesis,
participation in these activities should be affected by, among other things,
air quality. It was predicted that all but the number of visitors at the
state Capitol would be negatively affected by a worsening in air quality.
Attendance or use was treated as the dependent variable, while the independent
variables Included minimum daylight visibility, high temperature, and day of
the week. Data recorded between July 15 and September 30, 1970 on the par-
ticipation in these activities were analyzed.
103
-------
The results showed no significant relationships between rounds of golf,
indoor swimming pool attendance in Eugene, Oregon, and minimum visibility.
On the other hand, the hypothesized relationships between minimum visibility
and attendance at outdoor swimming pools in both cities were positive.
Statistically, the relationships were significantly different from zero
at the 95 and 99% levels of significance. Overall, the hypothesis
that air quality affects the extent to which individuals undertake (consume)
outdoor recreational activities, was supported. In addition, the evidence
supports the hypothesis that consumers, in fact, do engage in activity
substitution under variations in air quality.
Sorenson's data on swimming pool attendance as calculated over the range of
minimum visibilities in Eugene and Salen, imply that a 20% improvement in
visibility associated with a ban on open field burning would increase the
number of swim-days for residents by 60,000 to 90,000. Valued at an admission
fee of $0.50, the value of this increase in swim-days amounts to a sum of
$30,000 to $45,000. By assuming an elasticity of aggregate resident recreational
activity with respect to daylight visibility of 0.1, and estrapolating to the
entire Willamette Valley, estimates of increased recreational experiences are
generated. Economic values were based on the range of values suggested by the
Water Resources Council for the evaluation of outdoor recreation'days.
Vars and Sorenson determined best estimates of the value of increases in
Willamette Valley resident outdoor recreation activity, consequent to the
three policies under investigation, to be: (1) $249,000 for a ban on open
burning; (2) $111,000 for a policy allowing alternate year open burning, and
(3) $160,000 for a policy allowing open burning once in three years.
The second portion of the empirical research design consisted of a statistical
analysis of data derived from interviewing 401 tourists travelling in Oregon
during periods of reduced air quality. It was hoped that information on i
tourist perceptions and response to reduced visual range, could be secured.
Yet, again, the lack of a priori information about the consumer activities
production functions complicated the empirical problem. This sample popu-
lation was divided into polluted and non-polluted areas, based on the hypo-
thesis that recreation activities in the "clean" area would be substitutes
104
-------
for recreation activities in the "dirty" area because of the adverse aesthetic
effects of poorer air quality. One particularly interesting aspect of this
was the investigation of whether tourists were sufficiently "tracked" in
their behavior or whether they were flexible to the extent that their plans
were causually subject to change.
Data showed that of the 59 respondents who indicated that their plans had
changed upon entering the state, only six associated that change with air
pollution. Further analysis showed that tourists have fairly fixed plans
about the path of their travels and the amount they will spend. Quite
obviously these six responses cannot be viewed as statistically significant
in relation to any of the hypotheses about behavior or perceptions of interest
to this study.
f
Vars and Sorenson concluded that in terms of the hypothesis about the effects
of air pollution that this research set out to examine, the findings must
be viewed as negative. It is possible that the experience of most Americans
has suggested that air pollution is more or less pervasive, and that they
108
select among available recreational activities as if air quality were constant.
In an indirect criticism of the market approach, Vars and Sorenson observe that
it will be very difficult for social scientists to observe a clearly defined
relationship between air pollution and the behavior of people, even though
the behavior of people would be related implicitly to air pollution.
WORKS OF ART
Another kind of injury to aesthetic sensitivities occurs when air pollution
accelerates the decay of stonework. Calcarous materials such as limestone,
marble, lime plaster walls, and frescoes are subject to chemical assault by
acids which are formed by the interaction of sulphur and nitrogen oxides and
moisture. One such example of international significance is the damage that
has occurred to the 14th century Giotto frescoes in the Scovegni Chapel at
Padua, Italy. These frescoes have been the object of special studies. By
late 1960, these works had experienced severe deterioration and scaling of
paint.109
105
-------
The assualt of atmospheric pollutants on antiquities, metal, and stone art-
work occurs all over the industrial world. In New York City, air pollution
has contributed to the spoiling of the facade of City Hall, resulting in a
$4 million expenditure for restoration.110 And, officials of the Metropolitan
Museum of Art have been forced to coat statuary with beeswax and to air con-
dition exhibition areas. The Operating Administrator has explicitly stated:
"The presence of various forms of sulphur in the air is particularly injurious
to limestone and marble. There is an appreciable, visible etching on marble...
I would say that all of the exposed stonework of ancient elements at the
Cloisters has deteriorated since its erection in New York City...It is pointless
to collect outstanding works of art, many over a thousand years of age, if one
thousand years from now they are going to be so badly deteriorated as to be
virtually worthless."
In Spain, several Titians, Rubens, and other priceless works of the Italian,
Flemish, and Dutch schools are reportedly in danger of serious damage due
to the polluted air in the Prado Museum. Although experts have warned for
12 years that air pollution in Madrid was damaging the valuable canvasses,
the Spanish government only recently ordered emergency measures to protect
112
the 3,000 major paintings that are housed in the former palace. As a
result of these kinds of dangers, it is hoped that preservatives can be
113
applied to help retard decay in marble and limestone.
ORNAMENTAL PLANTINGS
Another category of effects that should rightfully be classified as aesthetic
include the destruction by air pollution of ornamental flowers, shrubs, and
trees that normally provide some sense of aesthetic enjoyment. These could be
ornamental flowers and shrubs that surround our homes, trees that line our
traffic arteries, or trees and other vegetation that normally grow in parks
and other areas used for purposes of recreation. Heggestad has reported phy-
totoxic effects of common urban pollutants on lilacs, petunia, orchid, and
gladioli.114 Also, it has been reported that some 160,000 acres of ponderosa
and Jeffrey pine in Southern California are experiencing severe decline, and
this has been attributed to the oxidant air pollution generated in the Los
Angeles Basin.115 This is significant in that a large part of the natural eco-
system is being affected, and this area is one of very high recreational value.
106
-------
CONCLUSIONS
As Ayres has summarized, "The disutility arising from minor discomfort and
essentially aesthetic objections to air pollution is probably the most under-
estimated and certainly the fastest-growing component of the total problem.
This arises from two interrelated factors: (1) the rising level of education
on the part of the population and even more rapid rate of increase in the
means and possibilities of communications, all of which results in an explosive
increase in the level of awareness and general perception of the pollution problem
as compared with a few decades ago, and (2) the fact that comfort and aesthetic
satisfaction are 'superior goods', as many economists have pointed out, and
the demand for them grows nonlinearly with general prosperity and affluence,
which are themselves rapidly increasing."
Given that the aesthetic qualities of our environment do have some economic
dimension, we are faced with the problem of measurement in quantifiable
economic terms. S. V. Ciriacy-Wantrup makes a strong case for the use of the
terminology, "extra-market" rather than "intangible" to describe those benefits
that are not routinely valued in the market place. Further, he argues that
attempts to quantify such values should be encouraged, and indeed, if the
measurement of air pollution damages is to be realistic, these extra-market
values must be assessed. The SRI vegetation study estimated the value of
ornamental plantings. Copley International Corporation (1971) made an attempt
to estimate odor costs by analyzing property, value differentials. Vars and
Sorenson (1972) made an attempt to estimate the impact and value of air
pollution as it affected recreation-related activities in Oregon. While none
of these studies has been particularly successful in identifying the pollution
exposure-receptor response relationship, progress is being made in understanding
the value of social choices that man explicitly and implicitly makes every
day. The major difficulty lies in isolating the incremental effect of changing
air quality on man's aesthetic and psychological needs. Even though measurement
is difficult, it is quite obvious that society is willing to expend significant
resources to reduce aesthetic damages from pollution.
107
-------
SECTION IX
ASSESSMENT OF THE EFFECTS OF SOILING
OVERVIEW OF THE PROBLEM
Individuals, households, and commercial establishments are affected by air
pollution in many ways, only a few of which are obvious. When dust particles
fall, the need to dust window sills and furniture is distressingly obvious.
But the effects of air pollution in most cases are so much more gradual as
to be unnoticed. Yet the costs of dealing with these effects may involve
considerable extra expense of which the household is usually unaware. In
urban areas some families spend very little as a result of air pollution,
but many spend hundreds of dollarsjnore each year than they would need to
if the air were clean.
INDIVIDUAL STUDIES
Mellon Institute - Pittsburgh Nuisance
The best known of the early studies of economic losses due to air pollution
is the Mellon Institute Study of the Pittsburgh smoke nuisance reported by
O'Connor (1913). The purpose of the study was to assess the economic cost
of the smoke nuisance to the populace in the city of Pittsburgh. The cost
estimates were based upon literature searches, observations, and informal surveys.
The damage estimates obviously included some direct costs as well as some
adjustment costs. The costs reflect losses due to soiling, corrosion, and
the obstruction of sunlight by particulates. Questionable statistical
techniques were used in averaging damage costs, in estimating the number
of units (i.e., stores) affected, and in arriving at the percentage
damage due to air pollution.
Beaver Report - London Smog Episode
The next major attempt to estimate soiling costs was an outgrowth of the
Mellon study. As a result of the London "smog" episode in 1952, a committee
was appointed in 1953 to examine the nature, causes, and effects of air
pollution and the efficacy of preventive measures. The report to Parliament
was released by Hugh Beaver (1954). i°8
-------
Much of the data were secured through literature searches and informal
surveys. The actual method used to make the estimates was similar to
that used in the Mellon study. In the Beaver Report, however, "black"
areas were compared with "clean" areas, whereas in the Mellon study,
Pittsburgh was compared with different cities. Costs were assessed in
the Beaver study by estimating the proportion of the total expenditure
on a specific item that is attributable to air pollution. The necessary
proportional estimates were obtained from additional estimates of the
amount and frequency of expenditures in polluted versus non-polluted areas,
as determined by interviews with local authorities. The polluted areas in
the study were assumed to contain one-half of the total population as thus
one-half of all other items, i.e., painting of buildings, etc.
It is evident that this method resulted in extremely crude lump sum figures
with only simple correlation with pollutant level. Where little or no
information was available, the investigators did not hesitate to use pure
guesswork. It should be noted that the investigator recognized that his
results did little more than suggest a broad order of magnitude.
Michel son and Tourin - Household Costs
In recent years, several attempts have been made to identify the costs of
soiling due to air pollution. For the most part, these studies have worked
with the household as the primary unit of investigation in an attempt to
measure pollution-related cleaning and maintenance costs.
In the area of evaluating household costs of soiling, the work of Irving
118
Michel son has received the most attention. Michel son's method of study
is based upon the hypothesis that if air pollution causes meaningful soiling,
the intensity of soiling should be reflected in a shorter time interval
between successive cleaning and maintenance operations in areas with higher
levels of pollution. If the relationship between particulate level and the
frequency of cleaning and/or maintenance operations could be established,
soiling costs could be calculated by applying a cost factor for each operation
studied.
109
-------
To test this theory, Michelson and Tourin (1966) conducted a survey by
mailed questionnaire in the towns of Steubenville and Uniontown in the
Upper Ohio River Valley. These towns have annual average particulate
levels of 235 yg/m and 115 yg/m , respectively. A high response rate was
achieved through a large publicity campaign, and a positive relationship
was found to exist between the cleaning frequency of the home and of personal
care items and the particulate level.
Cost comparisons were made within two income groups (less than $8000 and more
than $8,000), and the total costs were calculated on the basis of the number
of families and persons in each income group in each city. The differences
in frequencies were calculated and then converted into dollar differences
by applying local market prices for the various household services used in
the survey. The resulting figures showed that the economic cost of air pol-
1
lution for Steubenville was $3.1 million, or $84 per capita higher than in
Uniontown.
In an attempt to validate this study, a subsequent survey by Michelson and
Tourin (1967) was conducted in three suburban cities of the Washington, D.C.
area. The Washington area was chosen for the validating study because it
was thought to offer a severe test to the method. First, the absolute
levels of suspended particulate in the D.C. area were very much lower.
Second, the difference in the levels of suspended participates of the
paired cities was so much smaller in the Washington area as compared to the
paired cities of the Upper Ohio River Valley. Finally, the character of
the two areas was very different as far as industrial mix and population
characteristics.
Again, Michelson found a positive relationship between the frequency of ,
cleaning and maintenance operations and the level of suspended particulates.
Although the findings of the Washington, D.C. study would seem to support
the findings of the Ohio study, there appear to be not only major differences
between the two studies but also inherent problems within each that throw
some doubt upon his conclusion. For example, Income level was the only
110
-------
controlling factor in the analysis. Furthermore, only the responses of the
above-average income group were analyzed in the Washington study. Once the
relationship was found to exist in that income group, it was assumed to
exist for the below-average group in his estimate of total and per capita
soiling costs. In summary, the principal weakness was the lack of statis-
tically reliable techniques.
Since these two major studies, Michel son and Tourin (1968) have applied
their method of estimating the total extra household costs resulting from
air pollution in Connecticut. In this study, no household survey was per-
formed to measure the frequency of cleaning and maintenance operations.
Instead, the frequencies were taken from the Upper Ohio River Valley and
Washington, D.C. area studies. Because these frequencies were not alike,
some kind of "averaging" was done. The local costs of the operations were
investigated, and the demographic figures from census materials were used to
arrive at a total damage estimate for the state of Connecticut. The usefulness
of this method without adequate verification is questionable.
Ridker - Urban Soiling
Ronald Ridker also did research in identifying the soiling costs of air pol-
1 in
lution. In 1965, Ridker conducted a study in high, medium, and low pol-
lution zones of Philadelphia to determine whether family behavior and expen-
ditures were affected by air pollution. Despite the apparently adequate
collection of data, the results of the analysis were inconclusive. Although
there appeared to be some detailed problems and errors in the analysis, the
principal problem involved the use of time spent in routine household cleaning,
which may very well be an inappropriate estimate of these costs. The relative
frequency with which these tasks are performed may be a more appropriate measure,
Ridker also conducted a time-series analysis of a pollution episode in
Syracuse.^20 A questionnaire was developed and administered by personal
interview. Although the results of this household survey were much better
than the cross-sectional analysis in Philadelphia, the approach was obviously
limited to the episode-type situation and could not be put to widespread use.
Ill
-------
The Michel son and Ridker soiling studies have indicated several major
problem areas with regard to evaluating household soiling costs due to
air pollution:
1. Isolation of costs due to air pollution from those due to
other variables.
2. Sample selection and bias.
3. Development of a survey technique that will provide reliable
answers.
4. Inclusion of all household tasks whose costs are influenced by
soiling damage from air pollution.
Booz, Allen and Hamilton - Phil del phia Survey
The Booz, Allen and Hamilton, Inc. (1970) residential soiling study in
Philadelphia was expected to improve upon and extend the methodologies already
developed. A questionnaire consisting of two sections was developed to
determine the frequency of cleaning. The first section included questions
regarding cleaning operations and the second consisted of a set of self-
referent statements designed to determine cleaning attitudes. A total of
1800 personal interviews were conducted in the Philadelphia region.
Booz-Allen employed rigorous statistical survey techniques from the outset
of the"project. These techniques were employed because of the belief that
other, perhaps more dominant, non-pollution variables explain differences
in the frequencies of many residential cleaning arid maintenance operations
to a far greater degree than the variations in the*annual air particulate
levels. Therefore, the survey techniques included: (1) a probability sample
within several zones of the Philadelphia area; (2) group-depth interviews
leading to pre-estimates of attitudes toward cleaning' and the best ways'bf
phrasing survey questions; (3) personally administered questionnaires,jrather
than mail-or telephone surveys; (4) a factor analysis of the questionnaire
respondents to separate the population into attitude groups in order to
better explain why people clean; (5) collection of demographic data on each
respondent and his residence; and (6) the use of qualified interviews
coding, and keypunching operations.
112
-------
The study of residential household soiling costs made an attempt to discern
between cleaning necessitated by pollution and cleaning by habit or other
factors. Before any relationship could be established between the frequency
of these cleaning operations and the level of particulate pollution, other
socioeconomic variables that may contribute to the frequency were identified
and the degree of their interaction established through a factor analysis.
From the study of 27 cleaning and maintenance operations, results indicated
that the range of annual air particulate levels experienced in the Philadelphia
area (approximately 50 to 150 yg/m3) had no statistically significant dif-
ferential effect on the residential cleaning and maintenance costs for over
1,500,000 households in the area. These operations included painting, cleaning,
and washing. Of the 27 operations shown in Table 18, 11 were determined to
be somewhat sensitive to air particulate levels. Each of the sensitive
operations is a low-cost, do-it-yourself item, and many are associated with
being able to see out of the house—washing windows, cleaning screens, and
cleaning Venetian blinds. It must be pointed out that these do-it-yourself
operations were considered to be free of labor cost. The material costs of
these do-it-yourself operations were considered only when such costs were
considered to be non-trivial, such as the cost of painting.
They concluded that some low-cost cleaning and maintenance operations appear
to be sensitive to air particulate levels, but more importantly, the high
cost operations are unaffected by variations in air particulate levels in
the Philadelphia area. Another finding of interest indicated that a higher
proportion of residents of high-pollution areas believed their neighborhoods
were dirtier than did residents of low-pollution areas.
On a smaller scale, an attempt was made to determine the costs of soiling
borne by commercial establishments because of particulate pollution. A
sample survey of 138 stores was conducted and various cleaning operations
were investigated. Because of the poor return in the sample, the results have
proved inconclusive. Also, because of contractual arrangements, cleaning
functions are performed at the stores at regular intervals, whether "needed"
or not.
113
-------
Table 18. RELATIONSHIP OF CLEANING AND MAINTENANCE OPERATIONS
TO AIR PARTICULATE LEVELS
Inside
Clean and oil air conditioners
Clean furnace
Clean Venetian blinds and shades
Dry-clean carpeting
Dry-clean draperies
Paint walls and ceilings
Replace air conditioner filter
Replace furnace filter
Shampoo carpeting
Shampoo furniture
Wallpaper walls
Wash floor surfaces
Wash walls
Wash windows (inside)
Wax floor surfaces
Outside
Clean and repair awnings
Clean and repair screens
Clean and repair storm windows
Clean gutters
Clean outdoor furniture
Maintain driveways and walks
Maintain shrubs, flowers, etc.
Paint outside trim
Paint outside walls
Wash automobiles
Wash windows
Wax automobiles
Relationship
Sensitive
X
X
X
X
X
X
X
X
X
X
X •:->,
Insensitive
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
':
X
114
-------
The data collected in the Booz-Allen study are extensive, but the analysis
performed was of a very limited nature. Some of the conclusions are
believed to be unwarranted. For example, great care was taken to collect
demographic and social motivation data because these variables were assumed
to be perhaps more important than the air pollution variable. Yet, the -analysis
of these data shows that no more than one variable was considered in the
attempts to identify the soiling damage functions. In all likelihood, the
within-zone variability of these factors would so increase the maintenance
frequency scatter that it would be impossible to see any statistically sig-
nificant effects of pollution levels. It is this author's opinion that such
is the case with the activity of painting ouside walls. Booz-Allen con-
cluded that this activity was not sensitive to differences in particulate
levels because the frequencies between zones were not statistically sig-
nificant. These data warrant further analysis to account for some of the
confounding factors considered.
The Booz-Allen report has been criticized on several grounds: (1) statements
concerning the statistical significance of operations have not been adequately
justified in many instances; (2) the sensitivity or insensitivity of the
cleaning and maintenance operations is not fully explained; and (3) accepted
economic principles justify including with the cost of materials some imputed
values for homemakers1 time spent in cleaning and maintenance operations.
CONCLUSIONS
In conclusion, the Michelson, Ridker, and Booz-Allen studies dealt mainly with
the estimation of household cleaning and maintenance costs. Except for
Michelson, the evidence to date indicates air pollution does not have sig-
nificant economic effects in terms of household maintenance and cleaning
operations. A cost estimate will not be derived for this category in this
report because: (1) the Michelson estimates do not appear to be acceptable
for the purpose of extrapolation; and (2) those soiling costs associated with
painting have already been estimated by Spence and Haynie (1972).
115
-------
Yet, intuitively, other than what is implicitly measured in property value
differentials, it is difficult to conclude that there are not significant soiling-
related costs. Some of the significant costs that perhaps deserve attention
include: commercial cleaning and maintenance costs; individual adjustments such
as laundering, dry cleaning, and hair and facial care; car washing; and costs to
quasi pub-lie properties, which might include cleaning and maintenance costs of
buildings and monuments and washing of street luminaries. The magnitude of
soiling costs associated with specific effects undoubtedly runs into the millions
of dollars annually, but because of the lack of data, these soiling costs
will not be estimated in this report.
116
-------
SECTION X
EFFECTS OF AIR POLLUTION ON ANIMALS
OVERVIEW OF THE PROBLEM
Generaly speaking, air pollutants enter the bodies of domestic animals and
1 pi
wildlife via inhalation or ingestion of contaminated vegetation. -Li Hie
122
and Stokinger and Coffin provide good reviews of the effects of air pol-
lution on animal organisms. Fluorine by-products, lead, and arsenic are
the major offending constituents in industrial pollution. Dusts, ammonia,
hydrogen sulfide, and sulfur and nitrogen oxides cause less of a problem.
Pollution of agricultural origin is oftentimes linked to the misuse of
pesticides. Urban air pollution has been implicated as a causal factor in
the poor health of zoo animals. While no empirical studies to estimate air
pollution damages to animals have been attempted, a brief survey of the
literature will, hopefully, place this problem in perspective.
DOMESTIC ANIMALS
Some of the odlest documented cases of the deleterious effects of air pol-
lution have been associated with the Meuse Valley disasters, notably in 1897,
1902, and 1911. Vegetation and cattle were known to suffer from adverse
123
atmospheric conditions, locally called "fog disease." In actuality, the
124
cattle had been stricken with asthma and emphysema. An analysis of data
collected in Donora, Pennsylvania, has shown that a positive correlation
exists between the smog and the health of small domestic animals— dogs, cats,
125
poultry, and rabbits.
Fluoride poisoning of cattle grazing in the vicinities of aluminum reduc-
tion and phosphate fertilizer plants has demanded much attention in the
literature. Fluorosis, a disease common to cattle, occurs when fluorine
compounds are ingested for long periods of time. The animals eat contami-
nated fodder, grass, and hay, and also inhale quantities of fluorine.
Chronic fluorosis is typified by severe dental malformations and bone lesions
Acute fluorosis often results in stiffness, anorexia, weakness, convulsions,
117
-------
126
and cardiac failure. ° Cattle with fluorosis often show a reduction in milk
production and conceive poorly.127 Middleton128 reported that registered
steers in Polk County, Florida, once valued at $3,000 a head, were sold for
as little as $50 or were slaughtered because they were crippled and made
helpless by eating fluoride-poisoned grass.
Losses to livestock have been known to occur in the vicinity of lead smelters
129
and refineries. Molybdenum dust, scattered from the chimney of the neigh-
boring molybdenum-smelting factory, also has allegedly caused damage to live-
stock. The cattle developed diarrhea and malnutrition. Also, decreases in
130
production and in the rate of conception were found. Animals also have
been damaged by the effluents of a copper smelter. The high copper and
arsenic output that deposited on plants and grass caused numerous cases of
poisoning and even the death of domestic animals such as cattle, horses,
131
sheep and poultry.
WILDLIFE
A number of general conclusions have been drawn regarding air pollution
effects on wildlife. From field investigations, the economic poisons-
Insecticides, herbicides, chlorinated hydrocarbons, organic phosphates, etc.--
appear to outweigh by far all other types of air pollutants as hazards to
wildlife. Wildlife 1s chiefly affected by Ingestion of the "fallout" of the
air pollutant.
The relative susceptibility of various species to specific air pollutants
1s far from clear, but 1t would appear that the mammals are considerably
132
more susceptible than birds. Yet air pollution has been implicated as
the causal agent of primary lung cancer 1n birds in the Philadelphia zoo.
i 33
Synder has focused attention on the possibility that the amount of
carcinogens 1n the atmosphere is increasing, because water fowl that were
kept outdoors the year round were those animals most affected. It has also
been reported that lead poisoning of zoo animals has become a significant
problem at the Staten Island Zoo. The major source of the lead appears
I jA
to come from atmospheric contamination.
118
-------
CONCLUSIONS
The damage to animals caused by air pollution has generally been localized
and its economic consequence has probably been relatively unimportant; but
the social consequences of this pollution are potentially more severe. Though
indirect, the risk to the food cycle, especially when pesticides are impli-
cated, could be serious; and it may be true that the economic importance of
heavy metals and other toxic substances may lie in their impact on animals.
In general, little is known about the effects of urban air pollutants on
domestic animals and zoo animals. The pollutant burden in these animals
might offer an area of fruitful research.
Tolerance limits, much less damage functions, have not been developed for
domestic animals exposed to air pollutants except for fluoride with cattle,
swine, and poultry, and ammonia and carbon monoxide with poultry. In general,
air pollution does not appear to constitute a major potential health hazard
to domestic animals. However, the cadmium content of milk throughout the
United States has revealed levels higher than safe limits. This finding,
coupled with evidence that other edible tissues of animals show increasing
concentrations of air toxicants, indicates the importance of the potential
impact that air pollution could have on the food chain.
119
-------
SECTION XI
EFFECTS OF AIR POLLUTION ON THE NATURAL ENVIRONMENT
OVERVIEW OF THE PROBLEM
The general effects of air pollution on the natural environment or biosphere,
have yet to be clearly delineated, let alone evaluated in economic terms.
Nevertheless, it is useful to mention some of the more pronounced effects
that air pollution may have, thereby providing a perspective of economic
dimension. The impacts of air pollution on the natural order of things
are important because again, we are dealing with the problem of scarcity--
the scarcity of natural resources.
Damage studies typically examine specific types of pollution effects and
attempt to isolate the damages of air pollution on a very limited basis.
However, the effects and damages associated with air pollution are likely to
have repurcussions beyond the simple effect investigated. Thus, there is a
need to trace out the interdependent effects of air pollution, and it is this
broader approach that recognizes the effects of air pollution as inherently
related to other aspects of man's activities and his natural environment.
Such a study is relevant, because it forces an examination of the global
and other less quantifiable aspects of air pollution effects. Increasingly,
air pollution is being considered a global problem not because of its indi-
vidually minor effects, but because of its collectively major effects.
When we talk about the natural environment or ecology, we are concerned with
the relation of living things to their environment and to each other. Over
time, the environment is altered, naturally and by man. The so-called eco-
logical balance, then, is a transitory, everchanging state of relationships
of living things to each other and to their environment. We can conclude,
then, that it is not conceivable that there is, ever has been, or ever will
136
be, an ideal, all-inclusive ecological balance. The economist, then, is
interested in what way major perturbations of this changing ecological
balance impact upon man and his welfare.
120
-------
Environmental problems normally arise because the natural assimilative
capacity of the environment 1s exceeded. Bower and Spofford state, "The
natural environment has a capacity to assimilate, in some degree, all forms
and types of residuals through the mechanisms of transport, transformation,
and storage. In effect, the environment acts as a buffer between the dis-
charger and the receptor, that is, it dissipates, absorbs, dilutes, and
degrades or modifies residuals. However, the capacity of the environment
to assimilate residuals varies from place to place and from time to time,
depending both upon local conditions and upon the stochastic nature of some
137
component of the environment, such as stream flow, temperature, and sunlight."
Actually, little is known about the ultimate fate of pollutants once they are
emitted Into the atmosphere. Some of the pollution undoubtedly moves into
the upper atmosphere where 1t can remain for long periods of time, but most
is probably washed out. The continual deposition of pollution on the earth's
surface may be creating irreversible imbalances by affecting the nutritional
content of the soil as well as the delicate balance of soil microbes and
I OO
other organisms important in food chains. When we examine the details of
food cycles, we see that the living and nonliving elements in nature are
found together in the ecosystem through which energy cascades and matter
cycles.139
Intuitively, scientists feel that air pollution should have some bio-climatic
effects. Will the discharge of C02 and heat into the atmosphere create the
infamous "greenhouse effect" and cause the polar ice caps to melt, or does
the slight increase in the earth's temperatures indicate that the discharge
of particulates into the atmosphere causes a reflection of solar rays
resulting in cooler temperatures? Indeed, little is known about the impact
of man's activities on the geophysical and biological world.
Aspects of global ecology enter also into the accumulation of toxic sub-
stances. Woodwell140 argues that there are global, long-term ecological
processes that concentrate toxic substances, sometimes hundreds of thousands
of times, above levels 1n the environment. These processes include not only
121
-------
patterns of air and water circulation, but also a complex series of bio-
logical mechanisms. Over the past decade, detailed studies of the distri-
bution of both radioactive debris and pesticides have revealed patterns that
have surprised even biologists long familiar with the unpredictability of
nature. Yet as Moriarity points out, there is little evidence to suggest
that pollutants do concentrate along the food chain.
Climatic effects have also been associated with air pollution. Scientists
have shown measurable and distinct differences between the climate in the
142
city and that in its environs. Air pollution is recognized as a signi-
ficant causal factor. In the city, temperature and humidity are generally
higher, precipitation and cloud cover are more frequent, and fog is more
common. In extrapolating such findings, we should obviously be concerned
with any major changes in the global climate. Weather has a tremendous
effect on many animal populations. If an entire area warms or cools sig-
nificantly, the reproduction, growth, and survival of organisms in that area
could be affected. It has been reported that in the Northeastern United
144
States, rainfall shows higher acid content than heretofore. This has been
blamed by scientists on air pollution. Oxides of sulfur and nitrogen are
believed to be converting to strong acids, thus, increasing rainfall acidity
10 to TOO times. Some fear has been expressed that this could prove to be
a water supply contaminant. Sweden is experiencing a similar problem, and
Brohult has concluded that the acid rain is leaching nutrients that are
145
essential for forest growth from the forest soils.
CONCLUSIONS
There is much to learn about the effects of pollution as it intervenes into
the life processes of the food web, productivity, populations, distributions,
and the mechanisms of reinoculation. As Porter argues, "As we lengthen and
elaborate the chain of technology that intervenes between us and the natural
world, we forget that we become steadily more vulnerable to even the slightest
failure to that chain."146 Many of these consequences of air pollution are
122
-------
not without some economic value. While macro-economic analysis might be
premature in many areas where the human and natural ecological relationships
are not clearly defined, micro-economic analysis can aid in the identification
of those possibilities that are economically feasible, resulting in a more
efficient allocation of research efforts.
Even though the effects of pollution on ecological systems are not known,
nor the probability of catastrophic events, it is obvious that people are
concerned and are willing to spend money to reduce these effects and prob-
abilities. The amount people are willing to pay to avoid ecological risk
is probably very large in the aggregate and should be included in any esti-
mate of the benefits to society from reducing pollution. However, given the
paucity of such information, no numbers are currently available.
123
-------
SECTION XII
ESTIMATION AND ALLOCATION OF NATIONAL GROSS DAMAGES
GROSS DAMAGE ESTIMATION
Methods and studies have been examined in this report to determine the
economic value of the damages of air pollution. It is concluded from this
review of the six measurement methods that can be used to estimate costs of
pollution, only two have been used successfully in developing defensible
damage estimates—the market study approach employing the property value
method and the technical coefficients approach.
The property value method provides a national estimate of air pollution
damages ranging from $3.4 to $8.4 billion with a "best" estimate of $5.9
bill1on for 1970. Anderson and Crocker argue that the property value esti-
mate can, with great confidence, be considered a lower bound of the eco-
nomic value of the negative effects of air pollution. As argued earlier
1n this report, it 1s assumed here that property value, or site value
differentials measure primarily aesthetic and soiling costs.
Studies of the costs of air pollution associated with human health, materials
and vegetation were also reviewed. These studies have used the technical
coefficients approach. The damages from air pollution determined in this
manner sum to $3.0 to $11.0 billion, with a "best" estimate of $7.0 billion.
These national damage estimates for 1970 are summarized in Table 19. Esti-
mates were not generated in this report for the other effects—animals and
environmental risk—because of data limitations.
The problem now 1s to try to understand how the estimates of $5.9 billion
and $7.0 billion relate to each other. The best that can be done is to make
Intelligent, Intuitive interpretations. The components of the latter cost
estimate have been fairly well defined; the former, much less so. In theory,
the housing market estimator should capitalize all of the economic costs
associated with polluted air. In the real world, however, this is unlikely
because the property market 1s less than perfect. This 1s because some losses
are probably capitalized 1n durable resources that are immobile, and some of the
effects are perhaps so Insidious as to go unnoticed by consumers.
124
-------
Table 19. NATIONAL ESTIMATES OF AIR POLLUTION DAMAGES (UNADJUSTED) 1970.
($ billion)
Effect
Aesthetics and soiling3
Human Health
Materials
Vegetation
Range of Damages
Low High "Best" Estimate
3.4
1.6
1.3
0.1
8.4
7.6
3.1
0.3
5.9
4.6
2.2
0.2
Property value estimator
Does not include estimates of losses attributable to oxidant-related air
pollutants because of data limitations.
Thus, what does the property value differential estimate? As discussed
earlier, many authors agree that what are probably implicitly contained in
this estimator are the aesthetic aspects of air pollution—costs associated
with soiling, odors, visibility-restriction, "psychic" effects, and losses
of plant ornamentals. If such is indeed true, then it would seem justifiable
to add the property value differential estimate to the $7.0 billion estimate
which is the sum of losses that, in general, do not significantly overlap
with those losses capitalized in the residential property, market. This
would sum to a total of $12.9 billion.
At a minimum, there would be two areas of overlap: (1) the value of plant
ornamental losses as estimated in the study by Benedict (1971). Yet even
here this overlap is believed to be very small, since Benedict's estimate
for ornamental losses included only replacement costs, not any "aesthetic"
value—the true value that would normally exceed the replacement value;
and (2) soiling costs associated with household painting thay may have been
partially estimated by Spence and Haynie (1972). Even so, the total value of
these two areas of overlap would be quite small in proportion to the whole.
125
-------
The fact is: we have little idea as to the extent the various effects are
capitalized into property values rather than being capitalized into other
assets or registered as losses in consumer surplus. Because of this lack of
knowledge, it seems reasonable to consider the estimates of $5.9 billion and
$7.0 additive, with minor adjustments.
If one considers the areas of overlap mentioned above, two adjustements must
be made. First, $50 million for ornamental losses as determined by Benedict
(1971) must be subtracted from the property value estimate. Second, $540
million for residential painting as determined by Spence and Haynie (1972)
must be subtracted from the estimate of materials losses. By making these
adjustments, the possibility of double-counting losses for plant ornamentals
and soiling that are implicit in the property value estimator, is minimized.
The adjusted gross damage estimate for 1970 then becomes $12.3 billion.
This estimate can be allocated as follows: Aesthetics and Soiling, $5.8
billion; Health, $4.6 billion; Materials, $1.7 billion; and Vegetation,
$0.2 billion.
SOURCE EMISSIONS
Using the general approach of Barrett and Waddell (1973), it may be instructive
to relate the cost of pollution for each effect to the specific pollutants
considered most responsible for that effect. EPA has estimated national
emissions of principal pollutants by major source category for 1970. The
principal pollutants are carbon monoxide (CO), particulates (part.), sulfur
oxides (SOV), hydrocarbons (HC), and nitrogen oxides (NO ). National emissions
X /v
of these pollutants were estimated to be 266 million tons in 1970 (see
Table 20).
Approximately 54% of all national emissions come from transportation sources,
including automobiles, trucks, buses, trains, aircraft, and other vessels.
Fuel combustion in stationary sources such as public utility and industrial
power plants, commercial and institutional boilers, and residential furnaces
accounts for M% of national emissions. Pollutants from industrial processes
other than fuel combusion make up 14% of national emissions. Dumps and incin-
erators and related solid waste disposal practices generate some 4% of the
126
-------
Table 20. ESTIMATES OF NATIONWIDE EMISSIONS, 1970*
(thousand tons/year)
Source Category
Transportation
Fuel combusion in
stationary sources
Industrial process
losses
Solid waste disposal
Agricultural burning
Miscellaneous
Total
CO
11.0
0.8
11.4
7.2
13.8
4.5
148.7
Part.
0.7
6.8
13.3
1.4
2.4
1.5
26.1
S0x
1.0
26.5
6.0
0.1
Neg.
0.3
33.9
HC
19.5
0.6
5.5
2.0
2.8
4.5
34.9
N0x
11.7
10.0
0.2
0.4
0.3
0.2
22.8
Total
143.9
44.7
36.4
11.1
19.3
11.0
266.4
* Source: J.H. Cavender, D.S. Kircher, and A.J. Hoffman, Nationwide
Air Pollutant Emission Trends 1940-1970, Pub!I. No. AP-115,
Environmental Protection Agency, Research Triangle Park,
January 1973.
127
-------
national emissions. The remaining 15% derives from a variety of sources
including prescriptive burning of agricultural and forest fuels, wild forest
fires, structural fires, coal refuse burning, organic solvent evaporation,
and gasoline marketing.
ASSIGNMENT OF DAMAGE COSTS BY POLLUTANT AND SOURCE
The national air pollution-related health costs for mortality and morbidity
in 1970 were estimated to be $4.6 billion. Most health studies reviewed in
this paper, related health effects with particulates, sulfur dioxide, and
sulfur oxide pollutants. These pollutants have been studied most commonly
because: (a) there is generally more information on dose-response for these
pollutants than for any others; (b) more and better air quality data is
generally available for these pollutants than for any others; and (c) often-
times, particulates measurements seem to be a fairly good index of overall
air quality. Thus, until better information is forthcoming, it is assumed
that the health costs of air pollution stem from particulates and sulfur
oxides and from the sources of these two pollutants shown in Table 20. Costs
will be allocated in this report according to the relative sensitivity
coefficients for these pollutants as determined by Lave and Seskin. In
other words, the relative importance of particulates can be determined as
147
being accountable for 59% of the total costs and SO for the balance, or 41%.
^\
Therefore, 59% or $2.7 billion of the $4.6 billion in health losses, is attri-
buted to particulates and $1.9 billion, or 41% of the $4.6 billion, is estimated
for the sulfur oxides-related health costs. Data deficiencies prohibit the esti-
mation of the value of health effects associated with carbon monoxide, hydro-
carbons, arid oxides of nitrogen.
In the case of materials, pollution damages of $.7 billion to elastomers
and dyes are attributed to oxidants and nitrogen oxides. The $.4 billion
of damages to materials by sulfur oxides, estimated by Gillette, is identi-
fied as such. Because of the difficulty of separating the pollutant inter-
actions, the remaining $.2 billion in the Spence-Haynie study (after adjusting
for double-counting) will be equally divided in this attribution process
between particulates and sulfur oxides. The remainder of the total materials
cost estimate, $.4 billion from the Salmon study, is allocated in proportion
to the emissions of pollutants, except for carbon monoxide, which, according
128
-------
to present knowledge, is not damaging to materials.148 Give that hydro-
carbons and nitrogen oxides react in the presence of sunlight to form photo-
chemical pollutants (oxidants), emissions of these two pollutants will be
combined to represent damage from oxidant pollution and from nitrogen
oxides.
Results of the model developed by Benedict (1971) which predicts air
pollution damage to vegetation, indicate that over 90% of the observable
damage can be attributed to oxidants, with a smaller part for sulfur oxides,
and with a still smaller fraction attributable to fluorides. This assignment
will allocate the total estimate of air pollution damage to vegetation of
$.2 billion to oxidants. There should be a small portion allocated to SO ,
/\
but, because of its magnitude, it will not be displayed.
In considering the nature of the property value estimate, in that by assumption
it measures aesthetic and soiling costs, it seems reasonable to assume that
the total cost of $5.8 billion (adjusted for double-counting) can be allocated
by evenly dividing the damage between particulates and sulfur oxides. Therefore,
$2.9 billion in damage is associated with particulates and $2.9 billion with
149
sulfur oxides.
Results of assignment by effect and by pollutant are given in Table 21. In
like manner, assignment of air pollution damages is made by effect and by
source according to the relative contribution of damaging pollutants. This
relationship is shown in Table 22.
129
-------
Table 21. NATIONAL COSTS OF AIR POLLUTION DAMAGE, BY POLLUTANT AND EFFECT, 1970
($ billion)
Effect
b c.
Aesthetics & soiling '
Human health
Materials0
Vegetation
Animals
Natural environment
Total
S0x
Low
1.7
0.7
0.4
*
?
?
2.8
High
4.1
3.1
0.8
*
?
7
8.0
Best
2.9
1.9
0.6
*
?
?
5.4
Parti cul ate
Low
1.7
0.9
0.1
*
?
?
2.7
High
4.1
4.5
0.3
*
?
?
8.9
Best
2.9
2.7
0.2
*
•
7
5.8
Ox'
Low i High
? j ?
?
0.5
0.1
?
0.6
?
1.3
0.3
?
?
1.6
Best
•
0.9
0.2
•
7
1.1
CO
Best
*
•
*
*
*
7
Total
Low
3.4
1.6
1.0
0.1
?
?
6.1
High
8.2
7.6
2.4
0.3
•
?
18.5
Best
5.8
4.6
1.7
0.2
?
?
12.3
CO
o
Notes:
aAlso measures losses attributable to NO .
/\
Property value estimator
cAdjusted to minimize double-counting
?Unknown
*Negligible
-------
Table 22. NATIONAL COSTS OF POLLUTION DAMAGE, BY SOURCE AND EFFECT, 1970
($ billion)
Effects
Aesthetics & soiling
Human health
Materials
Vegetation
Total
Transportation
0.2
0.1
0.6
0.2
1.1
Stationary source
fuel combustion
3.1
2.2
0.8
*
6.1
Industrial
processes
2.0
1.7
0.3
*
4.0
Solid
Waste
0.1
0.2
*
*
0.3
Agricultural
burning
0.2
0.2
*
*
0.4
Misc.
0.2
0.2
*
*
0.4
Total
5.8
4.6
1.7
0.2
12.3
*Negligible
-------
SECTION XIII
DISCUSSION
SOME LIMITATIONS OF GROSS DAMAGE ESTIMATES
There will be a temptation to use the $12.3 billion estimate of the total
cost of pollution as the measure of total benefits from pollution control.
Yet, in fact, some of the pollution costs associated with the miscellaneous
source category are not likely to become benefits resulting from general
pollution reduction. This is because emissions from structural and wild
forest fires are not normally controlled under traditional air quality
management programs.
Also, there has been no comparative analysis of pollutants in terms of
their relative severity. We do not know, for example, if a ton of SO
/x
causes a greater or lesser effect on vegetation than a ton of NO emissions.
/\
This aspect should temper any use of the damage estimates as allocated
according to pollutant.
There will also be the temptation to use the pollutant cost estimates as
indicative of relative seriousness. While they may be indicative of a
general magnitude of seriousness, it is necessary to point out that few
studies have attempted to assess oxidant-type pollution effects on human
health and aesthetics. While no cost is shown for oxidant effects on
human health, it would be naieve to assume that there are no such effects.
The problem is this: research has not yet progressed to that point where
specific effects can be isolated and quantified. Because of this deficiency,
the author has opted to conclude that instead of placing a zero cost in
that particular cell, it would be more appropriate to indicate a lack of
knowledge.
And there js another possibility: the results of some of these studies are
spurious because sulfation or particulate measurements, for example, are
acting as proxies for the presence of other environmental pollutants.
This is a common problem in all non-laboratory studies. Research has shown
132
-------
that SOX, NOX, and HC all break down to the particulate state; thus, any
individual particulate air quality measurement might also be representative
of those pollutants that were originally emitted as gases. This possibility,
then, complicates and raises serious questions of the validity of allocating
costs by pollutant in the nice, neat way shown in Table 21. Also, these
pollutants act synergistically to cause damage that perhaps would not occur
when acting independently. So again, we have the problem of attaching weights
to the different pollutants, which, by themselves are perhaps harmless, but
which, in the presence of other pollutants, become harmful.
A problem of perhaps a different magnitude is whether or not damages will
become benefits through the abatement of air pollution. In theory, the two
should be the same. But, given the measurement problems that we either assume
away or are somehow rationalized into nonsignificance, it is quite likely that
damages estimated by some of the techniques discussed in Section III (especially
the technical coefficients approach) are not "true" damages. This is so
partly because of the obvious fact that the world is not optimal except for
air pollution, consumers do not have sufficient knowledge about how they
are being affected by air pollution, and because no allowance is made for
substitution possibilities and adjustments that would be expected under a
different set of environmental conditions. Thus, it is possible that the
control of air pollution will result in benefits not heretofore yet measured.
Another inconsistency may occur in estimating gross damages because of some
double-counting. Property value estimates, along with estimates of pollution
effects on health, materials, and vegetation, are included in the total
damage estimate of $12.3 billion. There may be some significant overlap
of property value effects with the other categories. Information is not
sufficient to determine the extent of double-counting.
In summary, the major limitations of gross damage estimates are: (a) esti-
mates are often based on questionable air quality monitoring techniques
or incomplete air quality data; (b) synergistic actions between pollutants
complicates the categorization of effects and pollutant cost; (c) weak
133
-------
assumptions are often made in extrapolating experimental data to the effects
on the true population; (d) since some of the extra-market effects are not
amenable to quantitative assessment, they are lost in these estimates; (e)
the confounding of effects prevents assignment of residual damages to spe-
cific pollutants or sources; and (f) there may be some double-counting
between property value effects and other effects.
COMPARISONS AMONG GROSS DAMAGE ESTIMATES
The $12.3 billion estimate can be compared with those developed by others.
Perhaps the earliest cited figure for the costs of pollution is $11 billion
in 1959 or $60 per capita, which was extrapolated from results of the 1913
Mellon Institute Study on the basis of the commidity price index and popu-
lation.151 Ridker (1966) has suggested a total cost of pollution in 1970
as falling between $7.3 billion and $8.9 billion. Gerhardt (1969) estimated
the cost of pollution to be $8.1 billion for 1968 within a range of $6.8 to
$15.2 billion. The basic procedure of the latter two efforts involved five
steps: (1) the identification of categories of air pollution damage; (2)
an estimation of the total value of category regardless of the air pollution
effects; (3) the assumption of an air pollution damage factor; (4) the
application of this damage factor to the total value of the category; and
(5) the summation of the estimates across all damage categories.
Recently, a $16.1 billion estimate for 1968 was generated by Barrett
and Waddell (1973). It might be of value to mention how the Barrett-WaddelT
estimate of $16.1 billion for 1968 compares with the $12.3 billion esti-
mate for 1970 developed in this paper. From a casual glance, one might'
assume that damages have been reduced by approximately $4 billion between
1968 and 1970. This is not necessarily true. It is hoped that a brief
discussion will put the differences between the two estimates in better
perspective.
There are several significant aspects that account for the differences
between the two estimates. First, in the case of human health, the benefits
of reducing pollution to the primary air quality standards for particulates
and sulfur dioxide were estimated in this study, while Barrett
134
-------
and Waddell estimated the benefits of reducing pollution to zero. This would
tend to result in a lower estimate for 1970.
Second, in the case of the property value estimator of aesthetic and soiling-
related damages, there are two important things: (a) additional research showed
that a marginal capitalized property value of $350 would be more accurate than
the $200 value used in the earlier Barrett-Waddel1 study; and (b) levels of the
air quality data for 1970 that were used were, in general, somewhat lower than
those used for 1968. These tended to balance each other, thus resulting in no
significant difference between the two property value estimates.
And third, in the case of materials damages, additional completed studies
forced a lowering in this study of the economic losses associated with the cor-
rosion of metals and those associated with painting. Also, a revaluation
of the available information suggested that there was no sound basis for
estimating air pollution damages to certain materials such as cement and con-
crete, plastics, and wood; thus, estimates included for these damages in 1968
were dropped in this study. In addition to this fact, the lower'SOp levels in
1970 resulted in a lower materials damage estimate for 1970. Implicit in all
of these dollar values (as with that for vegetation losses), is the fact that
inflation is another factor pushing air pollution damages higher in one year
relative to the preceding one. The same can be said with respect to the
increase in many instances of populations-at-risk. This would particularly
be true in the area of health.
Thus, given that the bases for comparison of the two gross damage estimates are
varied, it would be very difficult and probably not very meaningful to try to
isolate what portion of the $3.8 billion difference could be attributed to the
different assumptions made or different kinds of data used. Compared to the
$16.1 billion estimate for 1968, the $12.3 billion estimated for 1970 in
this paper is considered to be more refined and better specified—more
refined in the sense that more logical and realistic assumptions are made
and better specified in the sense that it is acknowledged that this is only
the best estimate that falls within a specified range of $6.1 to $18.5 billion
with some high degree of probability.
135
-------
Most recently, Justice, et.al. (1973) have estimated that air pollution
damages in 1970 ranged from $2.0 billion to $8.7 billion. While the range
of damage estimates developed by Justice, et.al. overlap with the range
developed in this report, there are significant differences between the two
studies. The most significant difference pertains to health costs associated
with air pollution, which Justice, et.al. estimate to range from $62 million
to $311 million for 1970. This range is significantly lower than tjhat reported
in this report primarily because Justice, et.al. considered neither the addi-
tional work reported by Lave and Seskin after their Science article, nor the
recent findings from EPA's CHESS program. Differences in other costs for
specific effects rest largely on differences in the assumptions made, many of
152
which that are suspect.
The principal difference among all of the national damage estimates, including
those reported in this study, is the determination of damage factors. The
factors applied for national cost-of-pollution estimates for this study are
believed to be determined by more reliable and objective procedures than
in the previous studies.
SOME CONCLUDING REMARKS
It is the author's opinion that the estimate of $12.3 billion is a reasoned,
defensible one. Many pollution effects were not costed simply because of data
limitations. The estimate generated through the use of property value
method is believed to be, at a minimum, the lower bound on the "true"
economic damages resulting from air pollution. To minimize double-counting,
potential areas of overlap in ornamental losses and household painting were
accounted for. By accounting for this overlap, the separate estimates deter-
mined by the technical coefficients approach—health, materials, and vegetation-
were made additive to the estimate for aesthetics and soiling determined via
the property value approach. While acknowledging that there is room for
argument, it is believed that the available evidence suggests that the
two estimates should be added together. While some may exercise the option
of using $7.0 billion as the gross damage estimate for 1970, it is argued
here that $12.3 billion is a sounder, more realistic estimate.
136
-------
With respect to health, the estimates generated from Lave and Seskin (1973)
and EPA can be considered a conservative measure of the real cost. It is
argued that, in general, people are willing to pay more than the expenses
of medical expenditures and lost productivity which they suffer, for air
pollution abatement. While it is doubtful that the assumption of a straight-
line functional relationship of mortality and morbidity and pollution is
accurate, it is perhaps the most reasonable stance that can be taken at this
time. In summary, these two studies provide a basis for taking a significant
step in attempting to assess the economic effects of air pollution on human
health. Again, little is known about the effects of automobile-and-related
pollution on human morbidity and longevity.
The estimate of economic costs associated with materials degradation also
appears to be a reasonable approximation. It is quite obvious from the
numerous studies that only little dose-response information is available
and in particular, little is known about air pollution effects on concrete and
other building materials, paints, and some fibers. Also, little is known
about adjustment costs that can be related to the use of more resistent
materials because of air pollution.
Although vegetation losses due to air pollution are believed to be somewhat
greater in magnitude than the suggested $.2 billion, little empirical
evidence could support such an assumption. The figure is conservative because
the yield and growth effects on plants are not generally considered in this
estimate. There is much to learn about subtle, chronic, low-level-pollution
yield effects. Also, no attempt has been made to quantify the effects of
air pollution on the nutritional content of edible crops. Until some of
these areas are investigated further, the vegetation loss estimate can only
be used with an understanding of its many deficiencies.
There is still a lack of conclusive evidence on the soiling costs attri-
butable to air pollution. Although Booz-Allen concluded that no significant
economic impact of particulate pollution differentials existed with respect
to residential cleaning and maintenance costs, the analysis in their study
appears to be incomplete and warrants further work.
137
-------
While the impact of air pollution on man's aesthetic values is believed
to be considerable, because of data limitations, no direct estimates were
generated. Quite obviously, man is bothered by poor visibility and noxious
odors, but few attempts have been made to quantify these impacts. The lack
of information suggests that only little is understood about the "psychic"
costs people suffer as a result of a deteriorating air environment as well
as the deleterious effects of air pollution on precious works of art.
As mentioned earlier, no known attempts have been made to investigate the
economic effects of air pollution on animals, domestic or wild, even though
pollutants such as chlorinated hydrocarbons, pose a threat to the
balance of animal and related populations. It was also concluded that eco-
nomic analysis of any long-run implications of perturbations to our ecosystem
might be premature-.
Obviously, of the different methods that might be used to estimate pollution
costs, the technical coefficients approach has been the most popular. Why?
Because of its simplicity in handling and translating from physical
or biological damage to economic loss. Market studies, or more specifically,
the property value approach, with its sophisticated econometric handling of
data, has provided the soundest basis for estimating pollution costs. Even
though the assumption is often made that most aesthetic-related costs are
implicitly measured in this approach, some uncertainty exists as to what
effects are actually measured.
It is likely that some combination of the different methods surveyed will
ensure the most accurate assessment of the economic damages resulting from
air pollution insults. The technical coefficients approach should prove
valuable in understanding the basic cause-effect relationships affecting
adjustments in the market place. The property value approach should be applied
to rural areas and should be tested with other pollutant measurements. The
public polling technique will be used in understanding the social, aesthetic,
and psychological or "psychic" effects of adjustments that people experience.
Different problems will require different handling.
138
-------
In any attempt to determine a dose-response relationship, the large number
of variables that must be considered presents a serious problem in the
isolation of those parameters that are significant. Yet of course, excluded
variables introduce a bias only to the extent they are not orthogonal to
the included variables. Also, the application of different discount
rates in the determination of total costs of pollution could change the
relative cost estimates. The ten percent rate of interest used in the
residential property value to estimate results is an understatement of costs
relative to health costs which applied an eight percent interest rate.
RECOMMENDATIONS
Gross damage estimates are only the first step in providing information on
the benefits of pollution abatement to policy-makers. Such estimates do
point to the seriousness of air pollution problems. However, the U.S.
Government and most individuals in the U.S. are already convinced that air
pollution is indeed a serious problem.
Expansion and refinements in pollution effects studies should be undertaken.
Such information on dose-response--damage functions—would provide a sounder
basis for estimating benefits of abatement. However, the information which
is generated should be over a range of realistic ambient air quality or
control levels. Damage functions should be constructed on a pollutant-by-
pollutant basis or group basis when pollutants act, or can be acted upon,
together, and, most importantly, should be analyzed in a regional cost-
benefit, policy-making framework (see the example described in Section III).
Research should also be expanded in the area of the different methods that
can be utilized in the assessment of the social cost of air pollution. It
is likely that some combination of the different methods surveyed in this
paper will ensure that most accurate assessment of the economic damages
resulting from air pollution insults. Also, attempts should be made to
understand and identify the economic and social significance of adjustments
people make because of deteriorating environmental quality.
139
-------
SECTION XIV
REFERENCES
1. For a deeper discussion on how the effects of income distribution
might affect the pareto-optimum, see: J. Krutilla, Welfare Aspects of
Benefit-Cost Analysis, J. Polit. Econ., 6£, June 1961. Other summary
papers on theoretical efficiency in pollution control are: D.C. Ogden,
Economic Analysis of Air Pollution. Land Econ. 62_ (2) May 1966;-and R.
Zerbe, Theoretical Efficiency in Pollution Control. West. Econ. J. 8_ (4),
December 1970.
2. For example, see: Economics of Clean Air, Annual Report of the
Administrator of the Environmental Protection Agency to the Congress.
Senate Document No, 92-67, March 1972.
3. Under most circumstances, proportionate reduction requires a dif-
ferent mix of source-specific pollution control than does marginally allo-
cated reduction._ Therefore, regional pollution control benefits, both in
magnitude and distribution, are likely to differ under the two enforcement
schemes. Here, this complication is ignored in order to keep the analysis
relatively simple; it should, of course, be taken into account in any real-
world analysis.
4. In some instances, marginal costs may also be highly uncertain in
which case upper and lower confidence limits on marginal costs should be
provided to the decision-maker. This would tend to spread even wider the
range of efficient control levels.
5. Economics of Clean Air, op. cit.
6. Cumulative Regulatory Effects on the Cost of Automotive Transportation
(RECAT). -'Final Report of the Ad Hoc Committee, prepared for the Office of
Science and Technology, Washington, D.C. February 28, 1972.
7. For a more thorough discussion of this aspect, see: C. Wright, Some
Aspects of the Use of Corrective Taxes for Controlling Air Pollution Emissions.
Natur. Resour. J. £ (1), January 1969.
8. Kneese, A.V. How Much is Air Pollution Costing Us in the United States?
In: Proceedings of the Third National Conference on Air Pollution, Washington,
D.C. December 12-14, 1966. p. 529-538.
9. Ridker (1967), p. 12-29.
10. Crocker, T.D. The Measurement of Economic Losses from Uncompensated
Externalities. In: Proceedings of a Seminar on the Economics of Air and
Water Pollution. Water Resources Institute, Virginia Polytechnic Institute,
Blacksburg, Virginia, October 1969. p. 180-194.
140
-------
11. Lave, L.B.. Air Pollution Damage: Some Difficulties in Estimating
the Value of Abatement. In: Environmental Quality Analysis: Theory and
Method in the Social Sciences, A.V. Kneese and B.T. Bower, (Eds.). Baltimore,
Johns Hopkins, 1972.p. 213-242.
12. Anderson, R.J., and T.D. Crocker. The Economics of Air Pollution:
A Literature Assessment. In: Air Pollution and the Social Sciences: For-
mulating and Implementing Control Programs. Downing (Ed.), New York,
Praeger, 1971. p. 133-165.
i
13. Crocker, op. cit., p. 183.
14. Anderson and Crocker, op_. cit., p. 147.
15. Ridker (1967), p. 25.
16. The reader interested in the theoretical considerations should see
R.G. Lind's discussion of the classical theory on rents: Land Market Equili-
brium and the Measurement of Benefits from Urban Programs. Presented at the
Committee on Urban Economics Conference, University of Chicago, September
11-12, 1970.
17. For a more in-depth discussion of this limitation, see Lave (1972),
op. cit.
18. Anderson and Crocker (1970), p.3.
19. See: Public Administration and Metrooolitan Affairs Program, S.
Illinois University, Public Awareness and Concern with Air Pollution in the
St. Louis Metropolitan Area. Appendix B. Final Report on Contract No. PH
86-63-131, U.S. Public Health Service, Washington, D.C., Hay 1965-, and W.
Smith, et.al. Public Reaction to Air Pollution in Nashville, Tennessee.
JAPCA 14. (10), October 1964.
20. Ayres, R.V. Air Pollution in Cities, Natur. Resour. J. 9_ (1):17,
January 1969.
21. In particular, see: H.O. Nourse, The Effect of Air Pollution on House
Values. Land Econ. 43_, May 1967: and Ridker (1967), Chapter 6.
22. U.S. Public Health Service. National Goals in Air Pollution Research.
Washington, D.C., August 196t). p. 20-21.
23. Freeman, A. Myrick, III. Air Pollution and Property Values: A Method-
ological Comment. Rev. Econ. Statist. 53_ (4):415-416, November 1971.
24. Ibid, p. 416.
25. In response to Freeman's criticism see: R.J. Anderson, and T.D.
Crocker. Air Pollution and Property Values: A Reply. Rev. Econ. Statis.
54 (4):470-473, November 1972.
141
-------
26. See: R.J. Anderson and T.D. Crocker. Property Market Equilibria and
the Environment: A General Equilibrium Model of Exchange- and Some Empirical
Results: Program in Environmental Economics, University of California,
Riverside. Working Paper No. 20, February 1973.
27. Strotz, R. The Use of Land Rent Changes to Measure the Welfare Benefits
of Land Improvement. Washington, D.C., Resources for the Future, Inc., July
1966. (mimeo).
28. For a recent discussion on this problem, see: E.P. Seskin. Residential
Choice and Air Pollution: A General Equilibrium Model. Amer. Econ. Rev.
63 (5):960-967, December 1973.
29. Crocker (1971), p. 148.
30. Spore (1972), p. 32. Actually, the presence » relocation costs reduces
receptor losses below what they would otherwise be.
31. For some notions on the theoretical basis of this aporoach and some
rough empirical work, see: I. Hoch. Urban Scale and Environmental Quality.
In: Population, Resources and the Environment, Vol. Ill, Ridker, R.G. (Ed.).
Washington, D.C., Commission on Population Growth and the American Future,
1972, p. 235-283. See also: I. Hoch. Income and the City Size. Urban
Studies 9i299-328, October 1972.
32. Survey Expo!ores Consumer Environmental Awareness, Soecial Renort:
Ecology. April 24, 1972.
33. Pollution Rated Top Problem by Communities in Survey. Environ. Health
Letter 10 (1), January 1, 1971. These results are consistent with those
reportecTby: A. Murch. Public Concern for Environmental Pollution. The Public
Op1n. Quart. 35, Spring 1971; and, R. Rankin. Air Pollution Control and Public
Apathy. JAPCJH9 (9), August 1969.
34. Public Opinion on Environment Sampled. Environ. Health Letter 11 (9),
May 1, 1972.
35. Such figures tend to be very misleading for they Ignore the locational
seriousness of, or concern with, the problem. For example, 1n a survey
taken 1n the state of Oregon by Louis Harris (The Public's View of Environ-
mental Problems 1n the State of Oregon. Prepared for the Pacific Northwest
Bell Telephone Company, Study No. 1990, March 1970, 28% of the population
sampled said that they would be willing to accept a $200 increase in family
expenditures per year to improve the environment. Some 4758 said that they
would be willing to spend $100.
36. A number of these studies were reviewed by Ido DeGroot (1967).
142
-------
37. While the economist does not necessarily need to know this kind of
information, work by R. Creer, R. Gray, and M. Treshow (Differential Responses
to Air Pollution as an Environmental Health Problem, JAPCA 20 (12), December
1970) will be helpful in understanding the psychology of differential
responses to environmental pollution.
t
38. For an application in a parallel area, see: M. Clawson and J. Knetsch.
Economics of Outdoor Recreation. Baltimore, Johns Hopkins Press, 1966.
39. For example, see: Nevada Residents Awarded More than $1.8 Million in
Pollution Damages, Special Report: Ecology, December 6, 1971; and ASARCO
Settlement of El Paso Smelter Suit Estimated at Nearly $1 Million, Air/Water
Pollution Report 1_0 (21), May 22, 1972.
40. Hayighurst (1969) provides a helpful review of the subject of legal
vs. economic damages. See also: N. Leonard. The Measurement of Damages:
An Economist's View. Ohio State Law J. 3J_ (4), Fall 1970.
41. Anderson and Crocker (1971), op_. cit., p. 144-145.
42. For specific examples of this kind of information, see: From the
State Capitals. Published by Bethune Jones at 321 Sunset Avenue, Asbury
Park, New Jersey. This bimonthly reviews state, local and municipal poli-
tical decisions relating to environmental pollution.
43. PHI, J. The Delphi Method: Substance, Context, A Critique and an
Annotated Bibliography. Socio-Econ. Plan. Sci. 5_(1):58, February 1971.
This article provides a good description of the methodology, the historical
development, and a critique and annotated bibliography of the Delphi method.
44. Losses 1n Agriculture. Agricultural Research Service, Washington,
D.C. U.S. Department of Agriculture Handbook No. 291. August 1965.
45. Environmental Quality. First Annual Report of the Council on Environ-
mental Quality, Washington, D.C. August 1970, o. 72.
46. Dal key, N.C. The Delphi Method: An Experimental Study of Grouo
Opinion. The Rand Corporation, Santa Monica, California. Paper No.
RM-5888-PR. 1969.
47. Crocker (1969), p. 186.
48. For example, Salvln (1970), using this approach 1n Philadelphia, found
that knowledge of the effects of air pollution on textiles was generally
lacking. This kind of Information can be used by local abatement agencies
1n their public relations-educational programs.
49. For a similar but more simplistic communication model see: M. Crowe.
Toward a Definitional Model of Public Perceptions of Air Pollution. JAPCA
18 (2), March 1968.
143
-------
50. This knowledge should be important for purposes of securing public
support for an abatement program. A number of studies have shown where
people do not consider themselves polluters. For example, see: R.J.
Simon. Public Attitudes Toward Population and Pollution. Public Opinion
Quart. 35_, Spring 1971; and, Public Opinion on Environment Sampled, cm. cit.
51. Ridker (1967), D. 90-114.
52. Lawyer (1966), p. 41.
53. Zerbe (1969), p. 54.
54. - Peckham (1970), p. 15.
55. Such a conclusion is also found in: R.F. Muth. Cities and Housing.
Chicago, The University of Chicago Press, 1969.
56. See: R.J. Anderson and T.D. Crocker. A Comment on: Property Values
and Air Pollution: A Cross Section Analysis of the St. Louis Area, by
Kenneth Wieand. Appendix B. In: T.D. Crocker (1971).
57. For a deeper discussion see: T.D. Crocker. Externalities, Property
Rights, and Transactions Costs: An Empirical Study. J. Law Econ. 1£ (2),
October 1971.
58. See also Copley (1971). Their approach at isolating the nrooerty
value and air pollution relationship is not significantly different from:
Ridker (1967), Chapter 7; and W. 11. Auberle and B. Linsky. A Case Study
of Air Pollution on Property Values. Presented at the 61st Annual fleeting,
Air Pollution Control Association. St. Paul, Minnesota. June 23-27, 1968.
59. Ridker (1967), p. 141-151.
60. Ibid., and Flesh and Weddell (1972).
61. Ibid., and Wieand (1970).
62. An anonymous reviewer suggested that damages will reflect all the influence
of both pollutants if, and only if, there exists a linear relation between
the included and excluded pollutants.
63. Peckham (1970), p. 16.
64. This is consistent with R.F. Muth's (op_. cit.) treatment of consumer
behavior and derived conditions for household equilibrium.
65. Crocker (1969), p. 189-190. This proposition is also supported by
Crocker (1971). Spore's (1972.) attempt to discover the shape of the marginal
damage function was less fruitful (p. 102).
144
-------
66. A study by Crocker (1971) of site value differentials in Chicago pro-
vides some empirical data in support of Figure 3. For other cities with
conditions different than Chicago, the case outlined in Figure 4 is also
plausible. Obviously, in order to make better extrapolations, additional
points on property value-pollution curves are needed. Unfortunately, such
estimates are not now available. In the meantime, we choose to use linear
extrapolation and hope that this is a reasonable first approximation. Thanks
to William Watson for this suggestion. For further discussion on the validity
of these assumptions, see: A. Myrick Freeman, III. On Estimating Air Pollution
Control Benefits from Land Value Studies (In Press).
67. This assumption is based on sulfation data for the following cities:
Columbus, Georgia; Anchorage, Alaska; Lewistown, Idaho; Sioux City, Iowa; Cannon
Mountain, New Hampshire; and Nogales, Arizona. Source: James H. Cavender,
et.al. Interstate Surveillance Project: Measurement of Air Pollution Using
Static Monitors. Environmental Protection Agency, Research Triangle Park, N.C.,
May 1971.
68. Ibid.
69. Data on housing units were taken from: General Housing Characteristics:
United States Summary. Bureau of the Census, U.S. Department of Commerce.
1970 Census of Housing, December 1971. Where estimates of housing units
for metropolitan areas were not available, 49% of the total number of units
in the SMSA were taken. This percentage factor is based on statistics that
indicate 49% of the housing units are in the central cities.
70. Lave (1972), op_. cit., p. 216-217.
71. These are discussed in Lave (1971).
72. Lave (1972), 0£. cit., p. 217.
73. Private communication with Richard J. Johnson, Biometry Branch, Human
Studies Laboratory, National Environmental Research Center, Research Triangle
Park, North Carolina. May 21, 1973.
74. C.M. Shy, et. al. An Overview of CHESS. In: Health Consequences of
Sulfur Oxides: A Report from CHESS, National Environmental Research Center,
Research Triangle Park, North Carolina. (In Press).
75. Health costs by disease were estimated in: Dorothy Rice. Estimating
the Cost of Illness. Health Economics Series Number 6, PHS Publication No.
947-6. U.S. Department of Health, Education, and Welfare, Washington, D.C.
May 1966. Rice included the cost of premature death, of treatment and of
absenteeism. Costs were broken down by major disease category, except for
some types of treatments, and by costs of personal or nonpersonal nature,
such as drugs, eyeglasses, and school health services. The cost of premature
death is the loss of earnings discounted at 6%. All costs are for 1963. This
estimate is developed by taking 4.5% of the sum of total national health
145
-------
expenditures identified in Table 1 in Rice (p. 3) plus the total mortality
and morbidity costs identified in Table 32 (p. 110), also from Rice.
76. Lave (1972), OD_. cit., p. 231.
77. Or, alternatively, the "true measure" would as likely be what a person
would be willing to accept for reduced longevity.
78. Air Quality data for suspended particulates for 1970 taken from: Air
Quality Data for Suspended Particulates: 1969, 1970 and 1971. Environmental
Protection Agency, Research Triangle Park, N.C. Publication No. APTD-1353.
This report showed that the annual arithmetic mean for about 90 SMSA's was
102 yg/m3. Thus, a 26% reduction would be necessary to reduce this to the
primary standard of 75 ug/m3. Since there was no obvious way to relate
Lave and Seskin's minimum sulfation measure to the S02 standard, it was
simply assumed that the mortality rate would respond to a reduction in both
pollutants in like manner. In using the authors sensitivity coefficients,
a 26% reduction in these pollutants would result in 2.34% reduction to the
mortality rate.
79. Rice, op. cit.
80. This estimate is determined by taking 2.34% of the total value of
direct expenditures, of morbidity, and of mortality as given in Rice. This
total value is determined by summing the costs of morbidity and total
mortality (Table 32 in Rice) plus the value of direct expenditures (Table 1
in Rice).
81. Statistics on Private Health Expenditures and Personal Income-Wage
and Salary taken from the 1966 and 1972 Statistical Abstract of the United
States (87th and 93rd Editions), Bureau of Census, Washington, D.C., show
that there was an annual growth rate in private health expenditures of
6.3% from 1963 to 1970 and an annual growth rate in personal income-wage
and salary of 8.2%. By extrapolating the Rice direct expenditures at a
rate of 6.3% and the morbidity and mortality (foregone earnings) costs at
8.2%, an estimate of $3.73 is determined.
82. For a discussion of the conceptual basis for estimating health benefits,
see: T.E. Waddell. Environmental Pollution Control and Health Benefits.
Presented at a Symposium on the Economics of a Clean Environment, sponsored
by the MITRE Corporation and the American Geophysical Union. McLean,
Virginia. January 14-16, 1974.
83. Private memorandum from William C. Nelson to the Chief, Ecological
Research. Branch, Division of Health Effects Research, National Air Pollution
Control Administration. November 3, 1970.
84. Data taken from the 1971 Statistical Abstract of the United States (92nd
Edition), Bureau of the Census, Washington, D.C., Table No. 16, p. 16, shows
that 73.5% of the total population in 1970 lived in urban areas.
146
-------
85. This variance was determined in the following manner: (1) standard
errors of the pollution coefficients were determined by dividing the
coefficients by their t-statistics; (2) then by subtracting and adding
two standard errors to the coefficients, changes in the mortality rate as
a result of a 10% reduction in air pollution levels were determined; and
(3) these percent changes divided by the mean mortality rate, and this
multiplied by 2.6 (the number of 10% reductions) enabled the determination
of the total variance.
86. Spence and Haynie (1972), p. 29.
87. Salvin estimates elsewhere (Textile Pollution Loss is in Billions.
Raleigh News and Observer, March 29, 1970, Section IV, p. 10) that the
total economic damage of air pollution to textiles and fibers is $2 billion
annually. There seems to be little basis for such an estimate.
88. Lacasse (1971), p. 23.
89. Feliciano (1972), p. 13.
90. Agricultural data for 1964 and 1970 on the cash value of commercial
crops was taken from: Agricultural Statistics. U.S. Department of Agri-
culture, Washington, D.C.
91. See also: H.M. Benedict, et. al. (1973).
92. Peckham, B.W. Odors, Visibility, and Art: Some Aspects of Air Pol-
lution Damage. In: Proceedings of a Seminar on the Economics of Air and
Water Pollution. Water Resources Institute, Virginia Polytechnic Institute,
Blacksburg, Virginia, October 1969.
93. Library Battles Decay of Rare Books. New York Times, February 20,
1967, p. 37.
94. Edmisten, N., et. al. Interstate Air Pollution Study, Phase II Project
Report: Introduction. U.S. DHEW, PHS, Taft Center, Cincinnati, 1966, p, 29.
95. Medalia and Finkner, op_. cit., p. 13.
96. Phillips, et. al., v. Elk Paper Manufacturing Co., Equity No. 1577B,
Circuit Court, Cecil County, Maryland. April 17, 1969.
97- Capurro v. Galaxy Chemical Co., Inc., Nos. 3313 and 3357, Circuit Court,
Caroline County, Maryland. June 3, 1972.
98. Odors Ruled Not Illegal. Solid Waste Management, April 1972, p. 73.
99. Conservation News 37_ (5):15, March 1, 1972.
100. Dravnieks, A. Odor Perception and Odorous Air Pollution. Tappi 55_ (5):
737, May 1972.
147
-------
101. See: Cooley International Corp. Procedures for the Identification and
Assessment of Community Odor Problems. Environmental Protection Agency, Research
Triangle Park, N.C., Final Report, Contract CPA 70-116. October 1971.
102. Copley International Corp. (1971), p. 40.
103. Environmental Quality, op. cit., p. 72.
104. A Study of Pollution - Air: Staff Report. Committee on Public Works,
U.S. Senate, 88th Congress, 1st Session, September, 1962, p. 21.
105. Peckham (1969), op_. cit., p. 165.
106. See: Water Resources Council. Evaluation Standards for Primary Outdoor
Recreation Benefits. Supplement No. 1 to U.S. Senate Document No. 97 (87th
Congress, 1964), p. 4.
107. Vars and Sorenson (1972), p. 70.
108. Ibid., p.^H-36.
109. Tintori, L. The State of Conservation of the Frescoes and the Principal
Technical Restoration Problems. Studies in Conservation 8_ (37), May 1963.
110. Special Committee to Investigate Air Pollution. Council of City of
New York. Air Pollution in New York City: Interim Technical Report M-970.
New York, June 22, 1965. p. 29.
111. Special Committee to Investigate Air Pollution. Council of the City
of New York. Blueprint for Clean Air: Final Report M-10. New York, December
1965. p. 24.
112. Conservation News 37 (8):14, May 1, 1972.
113. See: S. Kleinfield. New Protective Coating for Buildings said to
End Graffiti, Pollution Scars. The Wall Street Journal, August 4, 1971.
p. 10.
114. Heggested, H.E. Disease of Crops and Ornamental Plants Incited by
Air Pollution. Phytopathology 58_ (1089), August 1968.
115. Angeles National Forest Hit by Smog. Mews Release, Forestry Research
News, Pacific Southwest Forest and Range Experiment Station, Berkeley,
California. December 2, 1970.
116. Ayres, op_. cit., p. 9.
117. Ciracy-Wantrup, S.V. Economics and Public Water Policy in Water
Resource Development. In: Benefit-Cost Analysis and Public Resource
Development, Smith and E. Castle (eds.). 1964.
148
-------
118. For an excellent review of Michelson's work see: A.C. Jones. Studies
to Determine the Cost of Soiling Due to Air Pollution: An Evaluation.
In: Proceedings of a Seminar on the Economics of Air and Water Pollution,
W. Walker (ed.). Water Resources Institute, Virginia Polytechnic Institute,
Blacksburg, October, 1963. p. 146-156.
119. Ridker (1967), p. 73-89.
120. Ibid., p. 90-110.
121. Lillie, R.J. Air Pollutants Affecting the Performance of Domestic
Animals: A Literature Review. USDA, ARS, Washington, D.C. Agricultural
Handbook No. 380, August 1970.
122. Stokinger, H.E., and D.L. Coffin. Biological Effects of Air Pol-
lution. In: Air Pollution, Vol. 1, A. Stern (Ed.). New York, Academic
Press, 1969.
123. Rubay, M. About the Fog Observed in the Meuse Valley in December
1930 and Its Noxious Effects on Animals. Ann. Med. Veterinaire, 77; 97-110,
March 1932. (Abstract).
124. Firket, J. Comparative Pathology and Air Pollution. Seventh Lausanne
Congr. Intern. Pathol. Com. 7_(2): 57-80, 1955. (Abstract).
125. Schrenk, H.H., H. Heimann, G.D. Clayton, et.al. Air Pollution in
Donora, Pennsylvania. U.S. Public Health Service, Public Health Bull. 306. 1949.
126. Shupe, J.L. Levels of Tocivity to Animals Provide Sound Basis for
Fluoride Standards. Environmental Sci. Techno!. 3^ (8):721-726, August 1969.
127. Bohne, H. Industrial Smoke Damage from Fluoride. M.H. Deut. Landwirtsch.
Ges. 77 (17):575-578, 1962. (Abstract).
128. Middleton, J.T. We Can Have Clean Air. Country Beautiful, 1969.
129. See, for example: MACC Hears Representative Dinger. Missouri Air
News, 2^ (6), July-August 1970.
130. Ogura, Y. Molybdenum Poisoning in Cattle Due to Air and Soil Contamina-
tion as an Industrial Hazard. Tokyo National Institute Animal Health Bulletin
50: 24-29, 1965. (Abstract).
131. Bischoff, 0. Poisoning of Domestic Animals Through Copper and Arsenic
Containing Fly Dust. Deut. Tieravztl. Wochschr. 47^:442-447, 1939. (Abstract).
132. Stokinger, H.E. Effect of Air Pollutants on Wildlife. Conn. Med. 27^
(8):487-492, August 1963.
149
-------
133. Snyder, R.L., and H.S. Ratsliffe. Primary Lung Cancers in Birds and
Mammals of the Philadelphia Zoo. Cancer Research." 26_:514-518, March 1966.
134. Bazell, R.J. Lead Poisoning: Zoo Animals May Be the First Victims.
Science 173 (3992): 130-131, July 9, 1971.
135. Lillie, op_cit.., p. 108-109.
136. For a discussion of this, see: W. Schultz. The Ecosystem Doom.
Bulletin of the Atomic Scientists 28 (4), April 1972.
137. Bower, B.T. and W. 0. Spofford, Jr. Environmental Quality Management.
Natur. Resour. J. 1_0 (4):659, October 1970.
138. For a detailed discussion of this aspect, see: A. March. Smoke, The
Problem of Coal and the Atmosphere, London, Faber and Faber, 1947. For exam-
ple, Marsh has observed that the presence of pollution deposits in the smoky
atmosphere inhibited growth of many plants with a consequent decrease in the
population of insects that fed on the plants, and a corresponding decrease
in the population of birds that fed on the insects.
139. For a very good discussion of ecological systems see: K.E. Watt.
Ecology and Resource Management. New York, McGraw-Hill, 1968.
140. Woodwell, D.M. Toxic Substances and Ecological Cycles. Scientific
American 216 (3):24-31, March 1967.
t '
141. Moriarty, F. Pollutants and Food Chains. New Scientist 53. (787):594,
March 16, 1972.
142. For a good discussion of these factors, see: P.M. Lowry. The Climate of
Cities. Scientific American 217(2):15-23, August 1967.
143. See Watt, op_. c1t., p. 29.
144. A1r and Water News 6. (12):7, March 27, 1972.
145. Brohult, Sven. The Sulfur Problem and A1r Pollution. Annual Report
to the National Academy of Engineers, Sweden. (1967).
146. Porter. The Place That No One Knew, Glen Canyon on the Colorado.
Sierra Club Bulletin No. 50. 1963.
147. These relative weights are determined by dividing each sensitivity
coefficient (.53 and .37 for particulates and sulfur dioxide, respectively,
as taken from Table 7) by the sum of the two sensitivity coefficients (.90).
148. A1r Quality Criteria for Carbon Monoxide. U.S. DHEW, PHS. National Air
Pollution Control Administration, Washington, D.C. Publication No. AP-62.
March, 1970.
150
-------
149. Intuitively, one would predict that particulates would be more closely
associated than sulfur oxides with the disturbances of aesthetic properties,
but for the lack of more definitive information, equal weights will be placed
on these two pollutants.
150. For example, see: Air Quality Criteria Document for Photochemical Oxi-
dants. U.S. DHEW, PHS, National Air Pollution Control Administration,
Washington, D.C, Publication No. AP-63. March 1970; Air Quality Criteria
Document for Hydrocarbons. Publication No. AP-64. March 1970; and, Air Quality
Criteria Document for NitrQgen Oxides. Publication No. AP-84. January 1971.
151. Michelson and Tourin (1966).
152. For a more detailed critique of this study, see: T. Waddell.
Memorandum to files on the subject, "Critique of Economic Costs of Air Pollution
Damage, by C. G. Justice, et. al. of Science, Technology and Research, Inc.
(STAR) May 1973." Human Studies Laboratory, NERC-RTP, EPA, Research Triangle
Park, N.C. March 14, 1974.
151
-------
SECTION XV
BIBLIOGRAPHY OF LITERATURE ON THE ASSESSMENT OF AIR
POLLUTION DAMAGES
Anderson, R. J., Jr., and T. D. Crocker. Air Pollution and Housing:
Some Findings. Herman C. Krannert Graduate School of Industrial
Administration. Purdue University. Lafayette, Indiana. Paper
No. 264. January 1970.
Barrett, L. B. and T. E. Waddell. The Cost of Air Pollution Damages:
A Status Report. Environmental Protection Agency. Research
Triangle Park, North Carolina. Publication Number AP-85.
February 1973. 73p.
Beaver, H. Committee on Air Pollution Report. Her Majesty's Stationery
Officer. London, England. 1954.
Benedict, H. M., C. J. Miller, and R. E. Olson. Stanford Research
Institute. Economic Impact of Air Pollutants on Plants in the
United States. Coordinating Research Council. New York, New York.
Final Report, Contract CRC-APRAC CAPA-2-680-70). November 1971.
77 p.
Benedict, H. M., C. J. Miller, and J. S. Smith. Stanford Research
Institute. Assessment of Economic Impact of Air Pollutants on
Vegetation in the United States: 1969 and 1971. Coordinating
Research Council. New York, New York. Final Report, CRC Contract
CAPA 2-680-71] CPA 70-16. Environmental Protection Agency.
Research Triangle Park, North Carolina. Final Report, EPA
Contract 68-02-0312. July 1973. 96p.
Booz, Allen & Hamilton, Inc. Study to Determine Residential Soiling
Costs of Particulate Air Pollution. U. S. DHEW, PHS, National
Air Pollution Control Administration. Raleigh, North Carolina.
Final Report, Contract Number CPA 22-69-103. October 1970.
Copley International Corp. A Study of the Social and Economic Impact of
Odors. Environmental Protection Agency. Research Triangle Park,
North Carolina. Phase II Final Report-, Contract Number CPA
70-116. November 1971.
Crocker, T. D. Some Economic Aspects of Air Pollution Control with
Special Reference to Polk County, Florida. U. S. DHEW, PHS,
National Center for Air Pollution Control. V'ashinqton, D. C.
Final Report, Research Grant AP-00399-02. January 1968.
Crocker, T. D. Urban Air Pollution Damage Functions: Theory and
Measurement. Environmental Protection Agency. Washington, D. C.
Final Report, Contract Number CPA 22-69-52. January 1971.
152
-------
Degroot, I. Trends in Public Attitudes Toward Air Pollution. Journal
of the Air Pollution Control Association. 17: 679-681, October
1967. —
Degroot, I., W. Lording, A. Rihm, Jr., S. W. Samuels, and H. V'inkelstein.
People and Air Pollution: A Study of Attitudes in Buffalo, N. Y.
Journal of the Air Pollution Control Association. 16: 245-247,
May 1966. —
Feliciano, A. Rutgers-The State University. 1971 Survey and Assessment
of Air Pollution Damage to Vegetation in New Jersey. Environmental
Protection Agency. Research Triangle Park, North Carolina.
Final Report, Contract Number 68-02-0078. October 1972. 43 p.
Fink, F. W., F. H. Buttner, and tt. K. Boyd. Battelle-Columbus Labora-
tories. Technical Economic Evaluation of Air Pollution Corrosion
Costs in Metals in the United States. Environmental Protection
Agency. Research Triangle Park, North Carolina. Final Report.
February 1971. 104 p.
Flesh, R. C., and R. P- Weddell. Social and Economic Effects of Odors:
Methods of Measurement. Presented at the 65th Annual Meeting of
the Air Polluction Control Association. Miami Beach, Florida.
June 18-22, 1972.
Gerhardt, P- H. An Approach to the Estimation of Economic Losses Due
to Air Pollution. U.S. DHEW, PHS, National Air Pollution Control
Administration. Washington, D. C. Unpublished Report. June 1969.
Havighurst, C. C. Duke University. A Survey of Air Pollution Litigation
tn the Philadelphia Area. U. S. DHEW, PHS, National Air Pollution
Control Administration. Raleigh, North Carolina. Final Report,
Contract Number CPA 22-68-112. December 1969.
ITT Electro-Physics Laboratories Inc. A Survey and Economic Assessment
of the Effects of Air Pollutants on Electrical Components. En-
vironmental Protection Agency. Research Triangle Park, North Carolina.
Final Report, Contract Number CPA 70-72. August 1971.
Jaksch, J. A. Outpatient Medical Costs Related to Air Pollution in the
Portland-Vancouver Area. Oregon State University. Corvallis,
Oregon. Unpublished Ph.D Dissertation. June 1973.
Jaksch, J. A., and H. H. Stoevener. Effects of Air Pollution on
Residential Property Values in Toledo, Oregon. Agricultural
Experiment Station, Oregon State University. Corvallis, Oregon.
Special Report 304. September 1970.
Justice, C. G., J. R. Williams, and J. D. Clement. Science, Technology
and Research, Inc. (STAR). Economic Costs of Air Pollution
Damage. Prepared for Southern Services, Inc., Birmingham, Alabama.
Report Mo. STAR-CR-103. May 1973. 159 p.
153
-------
Lacasse, N. L. Assessment of Air Pollution Damage to Vegetation in
Pennsylvania, June 1970-June 1971. Center for Air Environment
Studies, Pennsylvania State University. University Park, Penn-
sylvania. CAES Publication Number 209-71. 61 p.
Lacasse, N. L., and T. C. Meidensaul. Statewide Survey of Air Pollution
Damage to Vegetation - 1969. Center for Air Environment Studies,
Pennsylvania State University. University Park, Pennsylvania.
CAES Publication Number 148-70. January 1970. 51 p.
Lave, L. B. Does Air Pollution Cause 111 Health? Presented at the
50th Annual Meeting of the Mew England Hospital Assembly. Boston,
Massachusetts. March 29, 1971.
Lave, L. B., and E. P. Seskin. Air Pollution and Human Health. Science.
169C3947): 723-733, August 21, 1970.
Lave, L. B., and E. P. Seskin. An Analysis of the Association Between
U. .S. Mortality and Air Pollution. Journal of the American
Statistical Association. 68(342):284-290, June 1973.
Lawyer, R. E.. An Air Pollution Public Opinion Survey for the City of
Morgantown, West Virginia. West Virginia University. Morgantown,
West Virginia. Unpublished Master's Thesis. 1966.
Mason, R. G. Oregon State University. Effects of Air Pollution on
Public Attitudes and Knowledge. Environmental Protection Agency.
Research Triangle Park, North Carolina. Final Report, Contract
Number CPA 70-117. June 1972.
Medalia, N. A., and A. L. Finkner. Community Perception of Air Quality:
An Opinion Survey in Clarkston, Washington. U.S. DHEW, PHS,
Division of Air Pollution. Cincinnati, Ohio. Publication Number
AP-10. June 1965.
Michelson, I., and B. Tourin. Environmental Health and Safety Research
Associates. Report on the Validity of Extension of Economic
Effects of Air Pollution Data from Upper Ohio River Valley to
Washington, D. C. Area. U.S. DHEW, Public Health Service, Washington,
D. C. Final Report, Contract Number PH-27-68-22. August 1967.
Michelson, I., and B. Tourin. The Household Costs of Air Pollution in
Connecticut. Report to the Connecticut State Department of Health,
Hartford, Connecticut. October 1968.
Middleton, J. T., and A. 0. Paulus. The Identification and Distribution
of Air Pollution Through Plant Response. AMA Archives of Industrial
Health. 14: 526-532, December 1956.
- .j,
Millecan, A. A. California Department of Agriculture. A Survey and
Assessment of Air Pollution Damage to California Vegetation in
1970. Environmental Protection Agency. Research Triangle Park,
North Carolina. Final Report, Contract Number CPA 70-91. June 1971.
154
-------
Mueller, W. J., and P. B. Stickney. Battelle-Columbus Laboratories. A
Survey and Economic Assessment of the Effects of Air Pollution on
Elastomers. U.S. DHEW, PHS, National Air Pollution Control
Administration. Raleigh, North Carolina. Final Report, Contract
Number CPA 22-69-146. June 1970.
Naegele, J. A., W. A. Feder, and C. J. Brandt. University of Massa-
chusetts. Assessment of Air Pollution Damage to Vegetation in
New England: June 1971-July 1972. Environmental Protection
Agency. Research Triangle Park. North Carolina. Final Report,
Contract Number 68-02-0084. August 1972.
O'Connor, J. J. The Economic Cost of the Smoke Nuisance to Pittsburgh.
Mellon Institute, Pittsburgh, Pennsylvania. Smoke Investigation
Bulletin Number 4. 1913.
Pell, E. J. Putgers - The State University. 1972 Survey and Assessment
of Air Pollution Damage to Vegetation in Mew Jersey. Environmental
Protection Agency. Research Triangle Park, North Carolina. Final
Report, Contract Number 68-02-0078. June 1973.
Peckham, B. l-f. Air Pollution and Residential Property Values in
Philadelphia. U. S. DHEW, PHS, NAPCA, Division of Economic Effects
Research. Raleigh, North Carolina. Unpublished Report. September
1970.
Ridker, R. G.t The Problem of Estimating Total Costs of Air Pollution: A
Discussion and an Illustration. Report to the U. S. Public Health
Service, Washington, D. C. July 1966.
Ridker, R. G., Economic Costs of Air Pollution. New York, Frederick A.
Praeger, 1967. 215 p.
Ridker, R. G., and J. Henning. The Determinants of Residential Property
Values with Special Reference to Air Pollution. Review of Economics
and Statistics. 49:246-257, May 1967.
Riggan, W. Cost to the Federal Government of'Health Effects Attributed
to Pollution from Motor Vehicles. U. S. DHFW, PHS, NAPCA,
Division of Health Effects Research. Durham, North Carolina.
Unpublished Report. 1970.
Robbins, R. C. Stanford Research Institute. Inquiry into the Economic
Effects of Air Pollution on Electrical Contacts. U.S. DHEW, PHS,
National Air Pollution Control Administration. Raleigh, North
Carolina. Final Report, Contract Number PH-22-68-35. April 1970
(Revised).
Salmon, R. L. Midwest Research Institute. Systems Analysis of the Effects
of Air Pollution on Materials. U.S. DHEW, PHS, National Air
Pollution Control Administration. Raleigh-,-North Carolina. Final
Report, Contract Number CPA 22-69-113. January 1970.
155
-------
Salvin, V. S. University of North Carolina, Greensboro. Survey and
Economic Assessment of the Effects of Air Pollution on Textile
Fibers and Dyes. U.S. DHEW, PHS, National Air Pollution Control
Administration. Raleigh, North Carolina. Final Report, Contract
Number PH-22-68-2. June 1970.
Spence, J. W., and F. H. Haynie. Paint Technology and Air Pollution:
A Survey and Economic Assessment. Environmental Protection- Agency,
Research Triangle Park, North Carolina. Publication Number AP-103,
February 1972.
Spore, R. L. Property Value Differentials as a Measure of the Economic
Costs of Air Pollution. (Ph.D. Dissertation). Center for Air
Environment Studies, Pennsylvania State University. University
Park, Pennsylvania. CAES Publication Number 254-72. June 1972.
Vars, C. R., Jr., and 6. W. Sorenson. Oregon State University. Study
of the Economic Effects of Changes in Air Quality. Environmental
Protection Agency. Research Triangle Park, North Carolina. Final
Report, Contract Number CPA 70-117. June 1972.
Wieand, K. F. Property Values and the Demand for Clean Air: Cross
Section Study for St. Louis. Presented at the Committee on Urban
Economics Conference. Chicago, Illinois. September 11-12, 1970.
Williams, J. D.x and F. L. Bunyard. Interstate Air Pollution Study:
Opinion Surveys and Air Quality Statistical Relationships. U. S.
DHEW, Public Health Service. Washington, D. C. 1966.
Williams, J. D., and N. 6. Edmisten. An Air Resource Management Plan
for the Nashville Metropolitan Area. U. S. DHEW, Public Health
Service. Washington, D. C. 1965.
Zerbe, R. 0., The Economics of Air Pollution: A Cost-Benefit Approach.
Report to the Ontario Department of Public Health. Toronto,
Ontario, Canada. 1969.
156
«UA GOVERNMENT PWIHTINO OFFICE:»74 546-3X9/432 1-3
-------
BIBLIOGRAPHIC DATA !• Report No. 2.
SHEET EPA-600/5-7ll-012
4. Title and Subtitle
The Economic Damages of Air Pollution
7. Author(s)
Thomas E. Waddell
9. Performing Organization Name and Address
Washington Environmental Research Center
Office of Research and Development
Environmental Protection Agency
Washington, D.C. 20460
12. Sponsoring Organization Name and Address
Same as Above
3. Recipient's Accession No.
5' Report Date
May 1974
6.
8. Performing Organization Kept.
No.
10. Project/Taslc/Woik linn \"
1AA004-26BAF-02 !
1 1. Contract/Grant No.
i
i
13. Type of Report & Period
Covered
1
14.
IS. Supplementary Notes
16. Abstracts ^j. pOnutiPn is a problem because, it endangers man's health and the environ-
ment in which he lives. The information researched in this report indicates that the
cost of air pollution damage in 1970 Cfor measured effects only) falls within a range of
$6.1 to $18.5 billion, with a "best" estimate of $12.3 billion. A benefit-cost analytica
framework for environmental decision-making is outlined. The methods that have been or
can be used to estimate the damages of air pollution are identified. The strengths and
weaknesses of each method are discussed. The technical coefficients method is utilized
in estimating the value of air pollution damage to human health, to man-made materials,
and to vegetation. A particular market study method, the property value approach, was
used to estimate aesthetic and soilings-related costs. Economic losses associated with
air pollution effects on domestic animals and wildlife and the natural environment are
not estimated because of data limitations» Damages are allocated by major pollutant and
source category-. The utility and limitations of gross damage estimates are discussed,
and comparison*with, other such estimates is made. Report contains bibliography.
17. Key Words and Document Analysis. 17a. Descriptors
Economic Analysis
Benefit/Cost Analysis
Economic Effects
Economic Losses
17b. Identifiers/Open-Ended Terms
17c. COSATI Field/Group
18. Availability Statement
UNLIMITED
19.. Security Class (This
Report)
UNCLASSIFIED
20. Security Class (This
UNCLASSIFIED
21. No. of Pages
167
22. Price
FORM NTIS-35 (REV. 3-72)
USCOMM-OC I4B82-P72
------- |