EPA-600/5-76-011
September 1976
Socioeconomic Environmental Studies Series
    PHYSICAL AND ECONOMIC DAMAGE FUNCTIONS
               FOR AIR  POLLUTANTS  BY RECEPTORS
                                       Environmental Research Laboratory
                                       Office of Research and Development
                                      U.S. Environmental Protection Agency
                                            Corvallis. Oregon 97330


-------
                   RESEARCH  REPORTING SERIES

  Research reports of the Office of Research and Development, U.S. Environmental
  Protection  Agency, have been grouped into  five  series.  These five broad
  categories were established to facilitate further development and application of
  environmental technology. Elimination of traditional grouping was consciously
  planned to foster technology transfer and a maximum interface 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 on environmental management,
  economic analysis, ecological impacts, comprehensive planning and forecast-
  ing, and analysis methodologies.  Included  are tools for determining varying
  impacts of alternative policies; analyses of environmental planning techniques at
  the regional, state, and local levels; and approaches to measuring environmental
  quality perceptions, as well as analysis of ecological and economic impacts of
  environmental protection  measures. Such topics as urban form, industrial mix,
  growth policies, control, and  organizational structure are  discussed in terms of
  optimal environmental performance. These interdisciplinary studies and systems
  analyses are presented in forms varying from quantitative relational analyses to
  management and policy-oriented  reports.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                       EPA-600/5-76-011
                                       September 1976
PHYSICAL AND ECONOMIC DAMAGE FUNCTIONS FOR AIR
            POLLUTANTS BY RECEPTOR
                      by

           Ben-chieh Liu, Ph.D.
          Eden Siu-hung Yu, Ph. D.
        Midwest Research Institute
            425 Volker Boulevard
       Kansas City, Missouri 64110
          MRI Project No. 4004-D
         EPA Contract No. 68-01-2968
                Project Officer
                  John Jaksch
        Criteria and Assessment Branch
    Corvallis Environmental Research Laboratory
              Corvallis, Oregon 97330
   CORVALLIS ENVIRONMENTAL RESEARCH LABORATORY
          OFFICE OF DEVELOPMENT AND RESEARCH
       U.S. ENVIRONMENTAL PROTECTION AGENCY
            CORVALLIS, OREGON 97330

-------
                           DISCLAIMER
     This report has been reviewed by the Corvallis Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publi-
cation.  Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorsement
or recommendation for use.
                                ii

-------
                            FOREWORD
     The Clean Air Act of 1970 requires substantial reduction in air
pollution.  Under the authority of this and subsequent Acts, the
Environmental Protection Agency has promulgated national ambient air
quality standards for several pollutants.  In geographic regions
where ambient standards are exceeded, the states have been required
to undertake action to comply with the standards.

     The current energy crisis has resulted in a closer look by society
and the Agency at the tradeoffs between energy conservation and improved
environmental quality.  Specifically, the crisis has resulted in the air
quality standards coming under closer scrutiny.  The standards in many
instances are viewed by industry as impediments to the use of alterna-
tive fuels which could alleviate the current energy situation.

     In order to effectively evaluate the environmental tradeoffs, the
decision maker must have information on the costs and benefits of alterna-
tive environmental control strategies.  Providing such information involves
difficult issues of measuring and evaluating the diverse effects of pollu-
tion abatement.  One of the results of the energy crisis has been a
renewed call for a reevaluation of and increased emphasis on the delinea-
tion and quantification of the benefits and costs attributable to air
pollution reduction.

     As most economists who are familiar with the methodology know, benefit/
cost analysis has its limitations in practical application to decision
making problems.  The primary limitations are the difficulties encountered
in placing an economic value on some effect responses, and/or the deriva-
tion of adequate effect responses.  While dependable, systematic estimates
of damages resulting from the effects of air pollution are still quite
rare, progress is being made.  Within the past decade, several studies
have been completed estimating property and material costs of air pollution
and the effects of air pollution on property values and human health.
However, many of these studies are too specific, and, as a result, do not
lend themselves well for use in formulating decisions having national
implications.    The purpose of this study was to see, using existing
studies, whether this limitation could be overcome.

     More specifically, the purpose of this study was to examine past
economic, and other related environmental studies, to determine whether the
results could be utilized in estimating composite parametric damage func-
tions.  The functions, while providing ballpark estimates, could be used
in evaluating the outcomes of implementing alternative environmental
policies.  In the meantime it was hoped that additional economic-environ-
mental studies would be undertaken which would mitigate the shortcomings
and permit a reestimation of more precise damage functions.
                              m

-------
     This report estimates economic, parametric damage function  by  receptor
(human health, household soiling, materials, and vegetation)  for the  sta-
tionary source pollutants - sulfur dioxide and suspended particulates.   The
damage functions are based on existing research results.  The socio-economic
data used in formulating the damage functions for the different metropolitan
areas are derived from the 1970 census.

     The research results have been extensively reviewed by environmental
economists, whose suggestions and comments have been  incorporated into the
study.  The results should be used with appropriate caution.  Some of the
assumptions employed in the study, by necessity, are  uncertain.  Some of
the methodological-statistical techniques employed are  in their infancy
and have not been tested elsewhere.  Despite the existence of these diffi-
culties, it is the general consensus of the reviewers that the  study re-
presents an important step forward in evaluating alternative pollution
control options.  Peer review of the study results by other environmental
economists are welcome, and should be sent to the project officer at the
Corvallis Laboratory.

     This study was initiated by the Washington Environmental Research
Center, Office of Research and Development, Washington, D.C., and completed
at the Con/all is Environmental Research Laboratory (CERL), Office of Research
and Development, Corvallis, Oregon.
                               A. F. Bartsch
                               Director, CERL
                                  IV

-------
                                   PREFACE
     This is the Final Report for the project entitled "Physical and Economic
Damage Functions for Air Pollutants by Receptor," for U.S.  Environmental Pro-
tection Agency, EPA Contract 68-01-2968 and MRI Project No. 4004-D.

     The primary objective of this project is to generate some physical  and
economic damage functions by receptor for sulfur dioxide and suspended par-
ticulates for the U.S. urban areas so that marginal benefit and marginal cost
principal can be applied to air pollution control decisionmaking.  Based  on
existing literature and available data on U.S. metropolitan areas,  1970, aver-
age functions are developed for air pollution damages on human health, house-
hold soiling, materials and vegetation. Various types of air pollution damages
are also estimated on a cross section basis for the metropolitan areas included.
It should be noted that the geographic damage estimates are tentative not only
because the assumptions employed in the study are uncertain but also because
the methodology used is in its infant stage of development.

     This project was completed under the general supervision of Mr. Bruce
Macy, Assistant Director of Economics and Management Science Division and the
project director was Dr. Ben-chieh Liu, Principal Economist.  Research assis-
tance and data process were provided, respectively, by Miss Mary Kies, and
Mr- Jim Miller. Valuable assistance and comments from Dr. Chatten Cowherd,
Messrs. Paul Gorman and Richard Salmon of MRI, Drs. Donald Gillette, Michael
Hay and John Jaksch of EPA, Dr. Fred Able of Energy Research and Development
Administration, Dr. William Watson of Resource for the Future and Dr.  Eugene
Seskin at Urban Institute are gratefully acknowledged. Editorial service was
provided by Mrs. Doris Nagel, Mrs. Sharon Wolverton efficiently performed the
report typing and computer work was carried out at MRI's Computation Center.
Nevertheless, the views expressed in this study are those of the authors.  They
do not necessarily reflect the opinions of the sponsoring agency.
                                   v

-------
                                CONTENTS
 Foreword	   iii

Preface	t	    v

List of Figures	    ix

List of Tables	    x

Executive Summary 	     1

     Section I - Introduction 	     1
     Section II  - Mortality and Air Pollution	     1
     Section III - Morbidity and Air Pollution	     2
     Section IV  - Household Soiling and Air Pollution 	     2
     Section V - Material and Air Pollution	     2
     Section VI  - Vegetation and Air Pollution	     2
     Section VII - Aggregate Damage Losses and Damage Functions:   An
       Overall View	     3

Section I - Introduction	     8

     Damaging Effects of Air Pollution	     9

Section II - Mortality and Air Pollution	    14

     Introduction:  The Problems and the Objectives 	    14
     Estimation  of Physical Damage Functions	    18
     A Linear General Physical Damage Function	    26
     Values of Air Pollution Damages and Economic Damage Functions.  .    28
     Premature Mortality Damages and Suspended Particulates 	    33
     Implications and Concluding Remarks	    36

Section III - Morbidity and Air Pollution	    41

     Problems and Objectives	    41
     Environmental Damage Functions:  Some Theoretical Underpinnings.    45
     Adult Morbidity and Air Pollution	    47
     Adult Morbidity Damages and Sulfur Dioxide 	    49
     Adult Morbidity Damages and Total Suspended Particulates ....    60

Section IV - Household Soiling and Air Pollution	    66

     The Problems and the Objectives	    55
     Soiling Physical  Damage Functions	    57
     Economic Damages  and Economic Damage Functions 	    59

                                    vii

-------
                          CONTENTS (concluded)


                                                                      Page

Section V - Material and Air Pollution	   92

     Problems and Objectives	   92
     A Theoretical Framework	   94
     Exposition of Methodology	   97
     Regional Material Damage Costs 	   99
     Economic Damage Functions	102
     A Summary of Material Physical Damage Functions	108

Section VI - Vegetation and Air Pollution	120

     Problems and Objectives	120
     Dose-Response Relationships	122
     Economic Damage Functions	124
     Concluding Remarks	133

Section VII - Aggregate Economic Damage Costs and Functions:  An
                Overall View	136

     Aggregate Economic Damage Functions	138

Section VIII - References	143

Appendix A - Optimal Policies in the Presence of Environmental
               Pollution:  A Theoretical Framework	153

Appendix B	158
                                   viii

-------
                             LIST OF FIGURES
Number                                                               Page

 II-1     Hypothetical relationship between mortality rate and
            S02 concentration	^. . .  20
              2
 II-2     Heteroscedastic distribution of the residuals 	  27

 III-l    Sample observation from four morbidity studies with respect
            to S02	52

 III-2    Sample observations  from four morbidity studies with respect
            to TSP	61
                                   ix

-------
                              LIST  OF TABLES


 Number

 S-l       Economic  Damages Due  to  Air  Pollution, by Receptors  for
             Selected  SMSA'S  	    4

 II-1      Correlation Coefficients	   25

 II-2      Mortality Costs With  S02 by  SMSA's,  1970.	   31

 II-3      Mortality Costs With  TSP by  SMSA's,  1970	   35

 II-4      Mortality Costs by SMSA's, 1970	   39

 III-l     Morbidity Dose  - Response Observations	   48

 III-2     Adult Morbidity Linear Damage Functions 	   50

 III-3     Mean Values and Standard Deviations  of the Variables.  ...   51

 III-4     Morbidity Costs With  S02 by  SMSA's,  1970	   58

 III-5     Morbidity Costs With  TSP by  SMSA's,  1970	   63

 IV-1      Pollution-Related  Tasks  and  Their Unit Cleaning Costs  ...   68

 IV-2      Mean Frequency, Standard Error and Upper and Lower Limits
             of Frequency and Suspended Particulates 	   70

 IV-3      Soiling Physical Damage  Functions  	   71

 IV-4      Net  Soiling Damage Costs by  Large SMSA's	   73

 IV-5       Net  Soiling Damage Costs by  Medium SMSA' s	   75

 IV-6       Gross Soiling Damage  Costs by Large  SMSA	   78

 IV-7       Gross Soiling Damage  Costs by Medium SMSA's 	   80

IV-8       Per Capita Net and  Gross Soiling Damage Costs  ($) by Large
            SMSA's,  1970	   83

IV-9       Per Capita Net and  Gross Soiling Damage Costs  ($)  by Medium
            SMSA's,  1970	   85

                                    x

-------
                       LIST OF TABLES  (concluded)


Number

IV-10     Net and Gross Soiling Damage Costs in 148 SMSA's by
            Cleaning Operations, 1970  	   88

IV-11     Soiling Economic Damage Functions 	   89

V-l       Soiling and Deteriorating Costs of Paint and Zinc 	  100

V-2       Material Damage by Large SMSA' s, 1970	103

V-3       Material Damage by Medium SMSA1 s, 1970	105

V-4       Economic Damage Functions on Materials	107

V-5       Major Pollutant - Material Interactions 	  109

V-6       Results of Regression Analysis for Soiling of Building
            Materials as a Function of Suspended Particulate Dose .  .  116

V-7       Physical Damage Functions for Materials 	  118

VI-1      Variables Used in Economic Damage Functions 	  126

VI-2      Economic Damage Functions on Vegetation With Pollution
            Relative Severity Indices  	  127

VI-3      Economic Damage Functions of Vegetation, With Sulfur
            Dioxide Annual Mean Level  	  129

VI-4      Economic Damage Functions on Total Crops,  Total Ornamentals
            and All Plants	130

VI-5      Estimated Economic Damages of Total Crops,  Total Ornamentals
            and All Plants	132

VI-6      Mean and Standard Deviations of Variables  in Vegetation
            Damage Functions	134

VII-1     Economic Damages Due to Air Pollution, by  Receptors for
            Selected SMSA1s 	  137

VII-2     Economic Damage Functions 	  140

VII-3     Gross Economic Damages Changes Resulting From a 10 Percent
            Reduction in the Pollution Level	141

                                    xi

-------
                              EXECUTIVE SUMMARY
     The research delineated in this report is primarily concerned with evaluat-
ing regional economic damages to human health, material, and vegetation and of
property soiling resulting from air pollution. This research also attempts to
develop a more plausible exponential physical dose-response function for pre-
mature mortality and morbidity. The comparable and consistent damage loss esti-
mates for a variety of receptors developed in this research are expected to pro-
vide a data base useful for designing national and regional pollution control
strategies.

     The report comprises seven sections. A brief summary of the highlights
from each section follows:
SECTION I - INTRODUCTION

     The project involving the determination of regional air pollution damage
losses for mortality, morbidity, household soiling, material and vegetation can
be divided into four distinct phases:   (1) problem discussion and refinement;
(2) information and data gathering;  (3) damage loss assessment; and (4) physical
and/or economic damage function estimation. Static analyses are performed on
the basis of 1970 data for many metropolitan areas and regions in the United
States.
SECTION II - MORTALITY AND AIR POLLUTION

     A two-step econometric model was developed for estimating a nonlinear mor-
tality physical damage function and net damage costs of premature deaths result-
ing from excess air pollution for the 40 Standard Metropolitan Statistical Areas
(SMSA's) which had a sulfur dioxide level above 25 (ig/m^ between 1968 and 1970.
The model circumvents partial'ly the often recognized but largely ignored econo-
metric problems such as heteroscedasticity and multicolinearity and, hence, gives
credence to our damage loss estimate. In addition, an "average" economic damage
function was developed which relates premature mortality damage losses in dollar
terms to socioeconomic, demographic, climatological and air pollution variables—
sulfur dioxide (SO ) and total suspended particulate (TSP). The estimated mor-
tality damage due to SO  for 1970 varies from less than $0.1 million in Charleston,
West Virginia to $329 million in New York City, whereas mortality damage attribut-
able to TSP ranges from $1.4 million in Lawrence, Massachusetts to $155 million
in New York City. On a per capita basis, the highest damage due to SO  and TSP
is $28.4 in New York City and $27.6 in Detroit, respectively.

-------
SECTION III - MORBIDITY AND AIR POLLUTION

     The damage costs and physical and economic damage functions were developed
and estimated. Regional physical damage functions on adult morbidity were de-
rived by resorting to the classical least-squares linear regression. A Monte
Carlo technique was then used to derive an "average" nonlinear morbidity physi-
cal damage  function for adults. Low estimates for total annual morbidity costs
due to  SO   range from less than $1,000 in Cincinnati to a maximum  of $22 mil-
lion in New York City. Low estimates on morbidity damages attributable to TSP,
however, range from $152,000 in Bridgeport to more than $21 million in Chicago.
On a per capita basis, the highest damage due to S02 and TSP  is respectively
$1.9 in Chicago and $3.7 in Cleveland.
 SECTION IV -  HOUSEHOLD  SOILING  AND AIR POLLUTION

      A system of  soiling physical damage  functions relating various  types  of
 cleaning frequencies  to air  pollution was developed. Net and  gross  soiling dam-
 age costs for the 148 SMSA's were estimated. Finally, national "average" eco-
 nomic damage  functions  for household soiling were developed by relating  soiling
 damages to air pollution, dempgraphic, socioeconomic, and climatological vari-
 ables. Total  net  soiling costs  for 1970 attributable to air pollution  over the
 148 SMSA's were estimated to be more than $5 billion, while total gross  soiling
 costs were about  $17  billion over the 148 SMSA's.
 SECTION V - MATERIAL  AND  AIR POLLUTION

      This section  develops  economic  damage  estimates on  the  two most  economically
 important materials,  i.e.,  zinc  and  paint,  for  the  148 SMSA's  in  the  United
 States. Economic damage functions  relating material damages  to air  pollution
 and other socioeconomic and climatological  variables were  derived.  The  state
 of the art regarding  the  physical  damage  functions  on materials was also  re-
 viewed and summarized. The  soiling damage costs of  zinc  for  1970  range  from
 less than $0.5 million in Dayton,  Ohio  to $1.7  billion in  Chicago,  whereas the
 deteriorating damage  costs  of  zinc range  from less  than  $0.5 million  in Dayton
 to $57 million in  Chicago.  The soiling  damage costs of paint for  1970 range from
 $19 million in Fayetteville, North Carolina, to $2.3 billion in New York  City,
while  the deteriorating damage cost  of  paint is $0.7 million in Fayetteville
and $79 million in New York City.
SECTION VI - VEGETATION AND AIR POLLUTION

     Dose-response relationships for vegetation were reviewed. A set of national
"average" economic damage functions for 10 economically important crops in  the
United States and regional economic damages to vegetation were derived. The eco-
nomic damage functions will be useful to policymakers for forecasting possible
gains as a result of pollution control programs.

-------
SECTION VII - AGGREGATE DAMAGE LOSSES AND DAMAGE FUNCTIONS:   AN OVERALL VIEW

     Range estimates of economic damage losses over some broader categories
of receptors were derived. A number of aggregate economic damage functions were
also developed and summarized for the major pollutants. The aggregate as well
as the disaggregate damage functions developed in the previous sections can be
useful to national and regional policymakers in their quest for obtaining esti-
mates of possible benefits brought about by various pollution abatement strate-
gies.

     The numerically large values of aggregate damage estimates provided by
the experts in this area point to the need for effective control of pollutant
emissions. The question naturally arises as to what constitutes economically
optimal and politically feasible pollution control programs. As an effort in
providing some useful clues for understanding the above question, this study
attempts to estimate net as well as gross economic damages to human health,
material, vegetation and household soiling attributable to and in the presence
of air pollution for the urban areas in the United States. Economic and physi-
cal damage functions relating economic (physical) damages to air pollution,
demographic,  socioeconomic, and climatological variables were also developed
for the United States urban areas. It is hoped that the generalized economic
damage functions in this report are informative and useful for predicting pos-
sible marginal (average) benefits resulting from various air pollution abate-
ment programs.

     Any study of this nature is bound to have a few inherent limitations. The
notable limitations are the uncertainty associated with estimating the physical
damage function and in translating it into economic terms, and the uncertainty
of selecting  the most relevant measure of air pollution and the "correct" form
of relating damages to pollution.

     To provide the reader an overall view of the economic damages of various
receptors due to air pollution, a summary of the damage estimates for the effect
categories of human health, material deterioration, and household soiling is
presented in Table S-l. The selected 40 SMSA1 s which had an SO  level equal to
or greater than the threshold 25 |a,g/m  are listed in Column 1. The low and high
damage estimates of human health are presented, respectively, in Column 2  (HNC1)
and Column 3  (HNC2). Column 4 (MDC) presents the material deteriorating damage
estimates of both zinc and paint; Column 5 (TNSCO) contains the aggregate net
household soiling damages. On the basis of the low and high damage estimates of
human health presented, respectively, in Columns 2 and 3, two sets of low and high
aggregate damage estimates for the three effect categories, i.e., human health,
material deterioration and household  soiling,  were derived and summarized in
Column 6 (TNCl) and Column 7 (TNC2), respectively. The further details on the
estimations of the economic damages of each of the effect categories are  con-
tained in the subsequent Sections II, III, IV, V and VI. The formulas used for
deriving the estimates presented in Table S-l will be discussed in Section VII.

-------
                TABLE   S-l  . ECONOMIC  DAMAGES  DUE  TO AIR  POLLUTION,  BY
                             RECEPTORS  FOR  SELECTED SMSA's
                             (in  $ million,  1970)

(1)
SMSA' s
1. Akron, OH
2. Allen town, PA
3. Baltimore, MD
4. Boston, MA
5. Bridgeport, CT
6. Canton, OH
7. Charleston, WV
8. Chicago, IL
9. Cincinnati, OH
10. Cleveland, OH
11. Dayton, OH
12. Detroit, MI
13. Evansville, IN
14. Gary, IN
15. Hartford, CT
16. Jersey City, NJ
17. Johnstown, PA
18. Lawrence, MA
19. Los Angeles, CA
20. Minneapolis, MN
21. New Haven, CT
22. New York, NY
23. Newark, NJ
24. Norfolk, VA
25. Paterson, NJ
26. Peoria, IL
27. Philadelphia, PA
28. Pittsburgh, PA
29. Portland, OR
30. Providence, RI
31. Reading, PA
32. Rochester, NY
33. St. Louis, MO
34. Scranton, PA
35. Springfield, MA
36. Trenton, NJ
37. Washington, DC
38. Worcester, MA
39. York, PA
40. Youngstown, OH
Total
(2)
HNC1
10
8
48
49
3
6
3
191
22
55
18
129
2
12
12
11
4
3
123
21
3
352
39
13
7
4
107
45
13
16
5
13
44
5
12
3
48
3
4
9
1,475
(3)
HNC2
18
15
80
52
5
6
3
360
22
93
18
161
2
24
19
17
4
5
147
32
5
527
48
13
7
4
158
79
13
25
5
15
61
5
15
3
88
4
4
10
2,166
(4)
MDC
7
3
17
26
6
11
4
105
12
49
9
55
2
8
5
8
1
7
76
12
4
111
14
3
13
9
33
30
8
9
4
7
24
2
3
2
21
8
2
8
736
(5)
TNSCO
16
16
137
117
3
14
10
516
57
216
39
294
5
24
16
17
10
3
388
37
4
418
112
29
9
8
104
147
30
20
15
27
119
23
7
5
86
6
9
23
3,134
(6)
TNC1
33
27
202
192
12
31
17
812
91
320
66
478
9
44
33
36
15
13
587
70
11
881
165
45
29
21
244
222
51
45
24
47
187
30
22
10
155
17
15
40
5,349
(7)
TNC2
41
34
234
195
14
31
17
981
91
358
66
510
9
56
40
42
15
15
611
81
13
1,056
174
45
29
21
295
256
51
54
24
49
204
30
25
10
195
18
15
41
6,045
Note—individual figure  may not add to  totals due  to rounding.
                                           4

-------
     Table S-l reveals that the largest aggregate air pollution damage, in the
order of $1 billion, occurred in New York and Chicago SMSA's in 1970. The small-
est air pollution damage occurred  in Evansville and Trenton, both SMSA's damages
were in the magnitude  of $10 million in  1970.  Human
health damage estimates  (mortality and morbidity) ranged from $1.5 to $2.2 bil-
lion for the 40 SMSA's. Total material deterioration damages were about 0.7
billion, and total household soiling costs were about 3 billion for the 40
SMSA's under study.

     The implication of our  study for pollution abatement strategies is obvious.
Any effort to reduce the current pollution level appears to have a varyingly
significant impact on the economic damages resulting from the harmful effects
of air pollution. Admittedly, the implication of this study must be qualified
by several theoretical and empirical factors. The major difficulties often en-
countered in estimating air  pollution damages involve the lack of knowledge
regarding the shapes of functions describing the relationship between air pol-
lution and various receptors, and the lack of a satisfactory theoretical model
specifying the way air pollution affects various receptors. The impossibility
of accounting for all major  factors which might affect various receptors, the
lack of reliable formulations used for translating physical damages into mone-
tary terms, and the presence of numerous econometric problems have also caused
concern to investigators.

     Despite the existence of these difficulties, this study represents a step
forward in our knowledge of  pollution damages. It seems to be the first attempt
to construct essential frameworks of the physical and economic damage functions
which can be used for calculating comparable regional damage estimates for the
several important receptors--human health, material, and household soiling--
however tentative the damage estimates may appear to be. More importantly, ag-
gregate economic damage functions instrumental for transforming the multifarious
aspects of the pollution problem into a  single, homogeneous monetary unit are
tentatively derived and illustrated. It  is hoped that these results will be of
some use to guide policymakers as they make decisions on the implementation of
programs to achieve "optimal" pollution  levels for this country. Given the ex-
perimental nature of the methodological  and statistical procedures and the de-
gree of uncertainty associated with the  study results, a great deal of caution
should be exercised in using the products of this research.

     Finally, it should be noted that although the availability of information
on average or marginal damages is instrumental in determining the optimal na-
tional or regional pollution control strategies, the current problem is far
more complex than the question of balancing the benefits to polluters with dam-
ages inflicted on the receptors. The issues are pressing and not yet well speci-
fied. The basic difficulty in applying the recent research findings to accurately
estimate the air pollution damage cost stems from our ignorance about the recep-
tors at risk to air pollution. So far, few attempts have been made to identify
who suffers, to what extent, from which  sources, and in what regions. At this
moment, updating and expansion of the available crude estimates, which are

-------
generally restricted to certain regions,  are urgently needed.  To identify the
population at risk to air pollution,  and to measure the damage specifically
for polluted regions are apparently the most logical steps in  the area of future
research.

-------
                           MAJOR NOTATIONS AND VARIABLES
A
C
CC
CMR
CRMR
CROPL
CROPV
DTS
DDCZ
DDCP
EXP or e
GSCO
MR
MB
MBC
MDC
NSCO
OXID
PAGE
PYAP
POOL

PWOP
POP
PDS
RHM
RMR
SDCP
SDCZ
SMSA
SO
TSP
TMBCSO
TMBCTSP
TEMB
TEMA
u
Air pollutants
Conventional mortality rate
Computed conventional mortality rate
Computed mortality rate
Computed residual mortality rate
Economic loss of a particular type of crop
The output value of a particular type of crop
Number of days with thunderstorms
Deteriorating damage cost of zinc
Deteriorating damage cost of paint
Exponential
Elasticity of variable i with respect to variable j
Gross household soiling damage cost
Mortality rate
Morbidity rate
Morbidity cost
Material deteriorating cost
Net household soiling damage cost
Oxidant relative severity index
Percentage of population 65 or older
Percentage of population with income above poverty level
Percent of persons 25 or older who have completed 4 years
  of college
Percentage of white to total population
Population in the area
Population density
Relative humidity
Residual mortality rate
Soiling damage cost of paint
Soiling damage cost of zinc
Standard Metropolitan Statistical Areas
Sulfur dioxide
Possible annual sunshine days (percent)
Total suspended particulates
Total morbidity cost due to SO
Total morbidity cost due to TSP
Number of days in a year with temperature below 33° F
Number of days in a year with temperature above 89° F
The disturbance term

-------
                                 SECTION I

                                INTRODUCTION
     Deterioration in urban air quality constitutes one of the major problems
confronting most American cities today. Air pollution has inflicted a multitude
of damaging effects on human health, material, vegetation, animals, household
and industrial property. In the past decades, numerous research studies have
been conducted to ascertain and to quantify, if possible, the physical and
monetary damage losses to the various receptors due to the presence of exces-
sive concentration levels of the major air pollutants, e.g., sulfur dioxide,
total suspended particulate matter, oxidants, carbon monoxide and other sub-
stances in the urban areas.i/

     The numerical values of aggregate damage estimates provided by the experts
in this area point to the need for effective control of pollutant emissions ..£'
The question naturally arises as to what constitutes economically optimal and
politically feasible pollution control programs. The issues surrounding the
control strategies have been hotly argued and debated. Implementation of some
of the proposed control programs has been postponed for either political or
economic reasons.

     According to estimates prepared by the Bureau of Economic Analysis,
(Cremeans and Segel, 1975) a total of $18.7 billion was spent on domestic air,
water, solid waste and other pollution abatement and control programs in 1972.
The expenditure was -about 1.6 percent of our GNP in that year. Of the total
figure, 35 percent was accounted for by control and abatement activities of
air pollution. This expenditure figure is indicative of the magnitude of sac-
rifice the society has made for the purpose of reducing the problem of air
degradation.

     Is this amount of expenditure sufficient, from an economic point of view,
to attain optimal air quality for this country? The inquiry into this question
is handicapped without information about the corresponding benefit accruable
to the society because of the existing pollution control programs.

     From economic theory, it is well-known that the control policy is optimal
if the marginal benefit due to pollution abatement is matched by the marginal
expenditure incurred to implement the control. In the absence of national mar-
ginal or "average" damage functions of air pollution by receptors and the
marginal (average) damage estimate for each effect category, it is difficult,
if not totally impossible, to estimate the marginal  (average) benefits stemming
from the abatement of the last unit of air pollution in each metropolitan area
and the nation as a whole.
\J For a background information on the cost of air pollution damage, see
     Barrett and Waddell (1973) and Waddell (1974).
2/ For details on the damage estimates and the references, see the beginning
     paragraphs of each of the later sections.
                                      8

-------
     For purposes of analysis the effects of air pollution are customarily
classified into six broad categories:  (1) detrimental effects on human health;
(2) damage to vegetation; (3) deterioration of materials;  (4) soiling of house-
holds and business establishments; (5) injury to animals;  and (6) reduction
of visibility and other atmospheric effects of an aesthetic nature.  Since each
of these categories has direct and indirect economic value, whenever one's
ability and opportunity to enjoy these benefits is reduced, economic damages
result. It is unfortunate that the magnitude and measurement of the  resulting
economic damages is probably the most controversial point  in the entire pollu-
tion control issue.

     The basic objectives of this study were to estimate net as well as gross
economic damages to human health, material, vegetation and household soiling
attributable to and in the presence of air pollution for the urban areas in
the United States. Economic and physical damage functions  relating economic
(physical) damages to air pollution, demographic, socioeconomic, and climate-
logical variables were also developed for the United States urban areas. It
is hoped that the generalized economic damage functions in this report are
informative and useful for predicting possible marginal (average) benefits
resulting from various air pollution abatement programs.

     Any study of this nature is bound to have a few inherent limitations.
The notable limitations are the uncertainty associated with estimating the
physical damage function and in translating it into economic terms,  and the
uncertainty of selecting the most relevant measure of air  pollution  and the
"correct" form of relating damages to pollution.

     Since this study is primarily concerned with the estimation of  the economic
damages of air pollution in the United States urban areas, a brief,  but criti-
cal, review of the economic effects of air pollution is in order. Accumulating
evidence suggests that air pollution results in a number of noticeable and
substantial economic effects. Some of the more obvious of  these effects include
the soiling of materials by dustfall, necessitating additional expenditures
for cleaning; corrosion of materials, requiring replacement and application
of protective coatings; atmospheric haze, reducing visibility and causing aes-
thetic blight; and various respiratory and other health problems associated
with the inhalation of noxious fumes and particles from the atmosphere.
DAMAGING EFFECTS OF AIR POLLUTION

Effects on Human Health

     According to the 1974 National Academy of Sciences reports, two major
pollutants, i.e., total suspended particulates and sulfur dioxide,  are responsi-
ble for the bulk of the deleterious effects on human health. Other pollutants,
like carbon monoxide, nitrogen oxides and photochemical oxidants and ozone
also exert damaging effects. Exposure to high concentrations of carbon monoxide
damages the function of oxygen-dependent tissues and exposure to low concen-
trations of carbon monoxide results in adverse effects both in normal people

-------
and in patients with heart disease. Acute exposure to Low concentrations of
nitrogen oxide can cause visual and olfactory abnormalities. Tentative evidences
indicate that long term exposure to photochemical oxidants can result in eye
irritation and a decrease in lung tissue elasticity. At any one time, several
pollutants are present in the air- Thus, it is difficult to determine the inter-
action of pollutants and the specific health damages caused by a single pollut-
ant. Nevertheless, it has been established that air pollutants can accelerate
disease and death, even at levels generally considered safe and used as the
basis for setting standards. Each of the major air pollutants presents a health
hazard in itself, and harmful effects may be greatly amplified when they occur
in combination. Unfortunately, the degree of the synergistic effects among
the pollutants is not clearly known.

     Particulate emissions include a wide variety of pollutants, each of which
may exert different effects on human health. Carbon or soot particles are the
most commonly emitted kinds of particles. However, even when these are the
only particulates emitted—such as in coal combustion--there are indications
that the toxic effects of sulfur dioxide (also released in the coal combustion
process) are enhanced by their association with the particulate matter. Other
contaminants can absorb on the surface of the particles, thereby coming into
contact with the inner surfaces of the lungs and mucous membranes in far greater
concentrations than would otherwise be possible. The site and extent of parti-
cle deposition in the respiratory tract, and therefore its ultimate effect
on human health, depend upon both physical and physiological factors.

     Sulfur dioxide is highly soluble in body fluids. The principal effect
of this gas is irritation of the tissues lining the upper respiratory tract.
This results in bronchial constriction which, in turn, produces an increase
in respiratory flow resistance. Persons suffering from respiratory or cardiac
diseases may be unable to withstand the increased body burden caused by this
respiratory flow resistance. Adverse effects on ciliary activity and mucous
flow may also result from prolonged exposure to sulfur dioxide. Sulfur dioxide
and other oxides of sulfur can, under certain conditions, combine with water,
soot particles and other aerosols in the atmosphere to produce toxic acid aero-
sols and other contaminants far more dangerous than any of the individual ingre-
dients.

     The damage to human health depends not only on the concentration level
of pollutants,  but also on the physical conditions of each individual. There
is virtually no single threshold of pollutant concentration below which health
damages will not occur. At every level of a pollutant concentration, someone
could be adversely affected. In view of a wide range of physical conditions
of human beings the threshold of pollutants may be viewed as a symmetrical
distribution.  The "mean" level of this distribution is used in the present
study to calculate the economic damages resulting from air pollution.
                                       10

-------
     While the exact role of air pollution in causing illnesses is not known,
there is substantial evidence that air pollution does aggravate existing ill-
nesses, even to the point of causing premature death..I/While high  rates of
asthma attacks have been reported on days with high air pollution surface con-
centrations, greatly increased mortality rates from influenza,  bronchitis,
and pneumonia have been noted during periods of high sulfur dioxide and partic-
ulate levels.

     In estimating the damage cost of morbidity, it should be noted that the
direct, out-of-pocket cost of treating an illness or disease is probably far
less than the value of avoiding the necessity for treatment. When someone
suffers from a pollution-related chronic illness, the cost of pollution to
him is almost infinite; the value of avoiding the pollution-induced discomfort
is, for this person, immeasurably high. For this reason,  it should be cautioned
that the health damage of air pollution estimated in this study, like other
major studies on the basis of the health costs of treating pollution-related
illnesses, may understate the true economic costs or benefits of reducing the
responsible pollutants. Sections II and III present a thorough analysis of
the air pollution effects on human health, i.e., mortality and morbidity,
respectively.

Effects on Materials

     Many external factors influence the reaction rate between pollutants
and materials, with moisture the most important in accelerating corrosion.—
Inorganic gases are likely to cause tarnishing and corrosion of metals; can
attack various building materials such as stone, marble,  slate, and mortar;
and may deteriorate a variety of natural and synthetic fibers.

     The most noticeable effect of particulate pollutants is soiling of the
surfaces on which they are deposited. They may also act as catalysts increasing
the corrosive reactions between metals and acid gases. Additional damages to
surfaces and textiles are incurred by the wear and tear imposed by the extra
cleanings made necessary because of particulate soiling.

     The true economic damage to materials caused by air pollution is difficult
to ascertain. First, it is difficult to scientifically distinguish between
natural deterioration and deterioration caused by air pollution. Secondly,
it is uncertain regarding indirect costs of early replacement of materials
worn out by pollutant soiling.

     The most comprehensive analysis of the economic effects of air pollution
on materials was conducted by Midwest Research Institute (Salmon 1969). In that
study, the damages caused by interactions between specific pollutants and spe-
cific materials were identified. The estimated economic loss resulting from the
various pollutant-material interactions totaled $3.8 billion in 1968.
 I/ For details,  see Section III.
~2f See Section V for further details.
                                     11

-------
     Detailed analyses  of soiling costs and material damages by region are con-
tained in Sections IV and V,  respectively.

Effects on Vegetation

     The air pollutants having the greatest deleterious effects on vegetation
are sulfur dioxide, hydrogen fluoride, photochemical smog and oxidants, ethyl-
ene, and herbicides and fungicides. Sulfur dioxide enters a leaf through the
stoma, causing injury to the blade of the leaf in the form of intervenal col-
lapsed areas. Fluorides may be absorbed from the surface of the leaf and can
be toxic to some plants at extremely low concentrations. Other pollutants may
damage only certain susceptible types of plants.

     Based upon a Stanford Research Institute study (Benedict  et a 1.,  1973),
the national damage cost of air pollution on vegetation is estimated to be $150
million. This damage cost amounts to approximately more than one-half of 1 per-
cent of the total value of crops produced in the United States in 1970. This
figure represents mainly the visible damage to agricultural crops,  and does not
fully recognize the real economic losses due to growth suppression, delayed ma-
turity, reduced yields, and increased costs of crop production.

     Section VI describes and estimates the air pollution damages on vegetation
on a regional basis for different types of crops.

Other Damaging Effects

     Aesthetic damage caused by air pollution is the most difficult to quantify;
yet, intuitively at least, it represents one of the important categories of eco-
nomic loss suffered as a result of degraded air quality. The aesthetic category
encompasses a number of different effects ranging from impaired atmospheric
visibility to decreased property values resulting from the presence of air pol-
lutants .

     Reduction in visibility creates a heavy economic burden on most communities.
Some of the community operations which are most affected by pollution-related
visibility problems include airports, highways, and homes. When an airport's
traffic pattern is slowed due to delays in take-offs and landings caused by re-
duced atmospheric visibility, operational costs are increased, additional safety
hazards are imposed, passengers are inconvenienced, and businesses may be indi-
rectly affected. Similar effects occur on highways where reduced visibility slows
traffic,  causes congestion, and increases the likelihood of injuri/ous and expen-
sive accidents. Additional lighting—both on the streets and in the home—is
required when the sunlight is unable to penetrate a polluted atmosphere.

     Aesthetic damage can sometimes be partially measured indirectly, such as
by comparing property values in comparable residential neighborhoods having dif-
ferent air pollution levels. In other cases, aesthetic damages may be reflected
in the costs that are incurred in connection with their prevention or avoidance,
such as special precautions taken to protect certain values from aesthetic damage

                                     12

-------
by air pollutants. Still other cases may require a willingness-to-pay approach,
estimating the amount that individuals would be willing to pay in order to pre-
vent or avoid the threatened aesthetic damages due to the soiling effect or,
conversely, how much additional they would have to be paid to willingly endure
the aesthetic blight.

     Due to data deficiency, air pollution effects on aesthetics are not studied.

     Considerable damage to animals caused by air pollution has been noted. How-
ever, most cases are localized, the sources are easily identified, and the eco-
nomic consequences are relatively minor- Poisoning of livestock from heavy mef
als--arsenic, lead, and molybdenum--has been reported on numerous occasions,
and cattle and sheep are particularly susceptible to fluorine poisoning. In
addition to the direct economic losses resulting from animal mortality, signifi-
cant losses may come from  such effects as decreased reproductivity, decreased
growth, and lower output of milk, eggs and wool.

     No studies of the economic impact of air pollution on animals have been
reported in the literature. The value of all livestock and livestock products
produced during 1968 was $21 billion; out of this total, perhaps $10 million
could reasonably be attributed to losses of all kinds from air pollution damage
(Park, 1974).

     Due to data deficiency, air pollution effects on animals are not studied.

     In summary, this air  pollution damage function project involves four dis-
tinct phases common to each of the five studies regarding the damaging effects
of air pollution on mortality, morbidity, household soiling, materials and vege-
tation. The four phases are as follows: (1) problem refinement; (2) data and
information gathering; (3) estimation of regional economic damages; and (4)
development of physical and economic damage functions.

     Data on air pollution, demographic, socioeconomic and climatological vari-
ables were collected by a  thorough literature search. Most of the data utilized
for developing the economic damage functions were attained from a comprehensive
quality of life study for  the United States Standard Metropolitan Statistical
Areas (SMSA's) recently completed by Liu (1975).

     Following the selection of the needed data, regression models were devel-
oped to determine the physical and economic damage functions for all these major
air pollutants as well as  the various categories of the damaging effects. Econo-
metric problems and technical difficulties are discussed and dealt with as much
as possible during the process of damage estimation. Furthermore,  several method-
ologies were developed to  evaluate the economic damages by air pollutants and
effect categories for the  SMSA's in the United States.
                                     13

-------
                                  SECTION II

                         MORTALITY AND AIR POLLUTION


INTRODUCTION:  THE PROBLEMS AND THE OBJECTIVES

     Two issues in the area of pollution control have attracted much attention
recently. The first problem is to evaluate from the control efficiency viewpoint
the appropriate governmental policies for handling pollution abatement. While
Kneese  (1972), Peltzman and Tideman (1972), and Lerner  (1974) opted for regional
regulation of pollution, Stein (1974) stressed the role of the federal govern-
ment for controlling various pollution. Another problem involves the determina-
tion of the  optimal level of pollution abatement at which the marginal benefits
are matched  by the marginal expenditures incurred to implement the control. Esti-
mation  of the marginal benefits of pollution control at regional levels, however,
requires information on damage functions and damage estimates for the various
regions in the United States.

     Empirical works in this area for the United States have been advanced  sub-
stantially by Ridker (1967), Lave and Seskin (1970, 1973), Jaksch and Stoevener
 (1974), R. K. and M. Koshal (1974), among others. They confirmed the existence
of a close association between health and air pollution. 1>.2/ The conventional
ordinary least squares, linear or log-linear regression method has been employed
to quantify  the damaging effect of air pollution on mortality. However, often
the major difficulties encountered in estimating such a physical damage function
involve the  problems of errors in variables, nonnormality, heteroscedasticity,
and multicolinearity among air pollution and other explanatory socioeconomic,
demographic  and climatological variables, and the lack of knowledge regarding
the shape of the function which depicts the relationship between air pollution
and health.

     Two major approaches have been suggested in the literature for estimating
a pollution  damage function..3' The first approach involves the assumption that
consumers are explicitly or implicitly knowledgeable about the potential bene-
fit of  pollution control. Therefore, the estimation problem boils down to one
of inducing  the consumers to reveal their "true" preferences about abatement.
Often, unsatisfactory results were obtained in this approach because consumers
generally are not willing to pay their share of cost for abatement, and, hence,
tend to provide misleading information about the benefit accruable to them  if
air quality  is improved.
_!_/ These and earlier studies are subject  to a number  of  limitations.  For a de-
     tailed discussion see, for example,  J. R. Goldsmith (1969).
_2/ Contrary results have also been obtained, for  example,  by  Toyama  (1964) and   *
     Petrilli, Agnese and Kanitz (1966).  There were no controls for  socioeconomic
     factors in their studies. Hence, their results are  subject to bias.
.37 See, for example, Lave (1972), p. 213, for a detailed exposition of  the two
     approaches.
                                     14

-------
     The second method, on the other hand, involves explicit quantification of
the physical damage function and translation of the physical damage into mone-
tary terms. The advantage of this explicit approach is that it requires no inter-
personal utility comparison and cooperation from the consumers. However, the con-
siderable extent of uncertainty present in estimating the physical damage func-
tion and in converting it into an economic damage function casts doubt on the
reliability of the damage estimates.

     The damaging effects on human health by air pollution in New York City have
been well documented by Glasser _et al. (1967), Greenburg et al. (1962a, 1962b),
Hodgson (1970), and McCarroll and Bradley (1966). Recently, Schimmel and
Greenburg (1972) performed a time-series study based on mortality rate and pol-
lution for New York City covering the period between January 1, 1963,  to
December 31, 1968. The excess mortality rate was regressed on two daily mean
pollution variates, SO and smoke shade, for both the same and previous day. They
showed that approximately 80 percent of the excess deaths were attributed to
the effects of smoke shade while only 20 percent were attributed to SO . Again,
methodological problems encountered in national estimates are also prevailing
in these regional estimates.

     Damage costs of premature death and morbidity due to air pollution have
been estimated for the whole nation previously. Ridker (1965) estimated the to-
tal costs of a specific disease and then attributed 20 percent of these costs
to air pollution. Lave and Seskin (1970, 1973) related the amount of mortality
for specific diseases to air pollution and some socioeconomic variables. They
found that the association between air pollution and mortality is significant
and of substantial magnitude; e.g., a 10 percent decrease in the biweekly mini-
mum level of sulfates is associated with a 0.3 percent decrease in mortality
rate per 10,000 live births. Koshal (1974) established a quantitative relation-
ship between respiratory mortality rates and the level of air pollution and two
climatic variables. They estimated a reduction of about 50 percent in the air
pollution would imply a social saving on the order of about $1.9 to $2.2 billion
per year in terms of respiratory disease alone.

     It is noteworthy that although most of these air pollution damage studies
draw tentative conclusions, they suffer from a certain inherent difficulty in
evaluating their results. Difficulty arises because either the statistical pro-
cedures employed are less than perfect or the results obtained are inadequate
for generating statistical inferences needed. With the exception, perhaps, of
those of Lave and Seskin and the Koshals, most of the studies are time-series
analyses with sample observations restricted to a specific area or a small num-
ber of areas. As a result, little information can be deduced from the existing
studies for designing a general air pollution control policy which requires
the knowledge of an "average" damage function expressed in both physical and
economic terms and applicable to all metropolitan areas in the nation. From eco-
nomic theory, it is well known that the control policy is optimal if the marginal
benefits resulting from pollution abatement are matched by the marginal expendi-
tures incurred to implement the control. In the absence of national average dam-
age functions by pollutants and the marginal damages for each pollutant, it is

                                     15

-------
difficult to estimate the marginal benefits stemming from the abatement of the
last unit of air pollution in each metropolitan area and the nation as a whole.

     Lave and Seskin (1973 p. 290) in a well-known article, noted possible spec-
ification errors in the empirical estimates of mortality and air pollution rela-
tion. They cautioned the reader that "[their] analysis is beset by a vast number
of problems including little a. priori knowledge of the true specification of
the relations, omitted variables, and errors of measurement in the variables."
This observation has been recently verified by Smith (1975) by reestimating a
set of mortality air pollution relationships with a new data base. The Ramsey
tests were utilized with the data on mortality rates and suspended particulates
for 50 SMSA's.l/ The research findings indicate that the errors in specification
and heteroscedasticity could constitute technical problems in estimation.

     While the multicolinearity problem between air pollution and other indepen-
dent variables in the damage function makes it difficult, if not totally impos-
sible, to disentangle their influences so as to obtain reasonably precise esti-
mates of their separate independent effects on mortality, the presence of the
heteroscedasticity problem violating one of the assumptions used in the normal
linear regression model (i.e., the disturbances were independently distributed
with constant variances) renders the ordinary least-squares estimates ineffi-
cient ..£/ Despite the fact that these specification errors were observed by Lave
and Seskin, the econometric problems remain largely unexplored in the prior stu-
dies.

     This section attempts to achieve two basic objectives. First, a stepwise
econometric model will be developed to estimate a dose-response relation for
mortality and pollution. Second, "average" economic and physical damage func-
tions for the United States Metropolitan Areas will be constructed by relating
mortality economic damages and mortality rates, respectively, to air pollution,
demographic,  socioeconomic, and climatological variables. Although the method-
ological and  statistical procedures used are experimental, and the statistical
results  are subject to a great deal of uncertainty, it is hoped that the gener-
alized economic damage function and the cost estimates presented in this sec-
tion are informative. They can be useful for predicting possible benefits in
the urban areas resulting from various air pollution abatement programs and to
shed light on the major issues in current and future air pollution research.

     Technically, the heteroscedasticity and multicolinearity problems that  emerged
in estimating the relationships between mortality and pollution damage are par-
tially circumvented via the two-step econometric model. In the first step, ob-
served mortality rates are regressed on several relevant socioeconomic, demo-
graphic and climatological variables. In the second step, the residual mortality
rates obtained by subtracting the computed mortality rates from the observed
 I/ The logic underlying the Ramsey tests was succinctly outlined in Smith  (1975),
     pp. 341-342. For a detailed discussion on the tests, see Ramsey  (1969, 1970,
     1974).
 21 See, for example, Johnston (1963), pp. 207-211, and Goldberger (1964), pp.
     192-194.
                                       16

-------
mortality rates are again regressed nonlinearly on air pollution variable only
to derive the physical damage function. The estimated dose-response relation
is then utilized to derive net damage costs of premature deaths due to excessive
air pollution for  40  Standard Metropolitan Statistical Areas (SMSA's) in the
United States.

     In order to estimate physical and economic damages associated with air pol-
lution the effects of air pollution on human health are classified as:  (1) mor-
tality effect;  (2) morbidity effect; and  (3) combination effect.I/ The mortality
effect refers to the  increase in the excess deaths resulting from increased con-
tamination in the  air, or the decrease in the survival probability of all ages.
The premature mortality affects an individual's probability of being accessible
to future earning  opportunities and nonmarket leisure activities, but it will
not alter the nature  of the existing economic and leisure activities. The mor-
bidity effect, which  will be dealt with in the next section, however, directly
changes the nature of economic and leisure activities. The combination effect
can be viewed as earlier mortality because of increased severity in morbidity.
In this case, both the survival probability and the nature of activities of the
victim are affected.  Schrimper  (1975) has shown that this interaction effect
can be conveniently ignored because of its small  magnitude.

     It may be worth  pointing out, at the outset, that the physical dose-response
relation derived in the present study is probably the first of its kind ever
estimated in the pollution effect studies. Four distinguishing features in the
dose-response relation differentiate our  study from the earlier  studies, say,
Lave and Seskin  (1970, 1972, 1973) and Koshals  (1974). First, the technique of
residualizing the  dependent variable  (mortality rates) is used in estimating
the dose-response  function. Second, the pollution variable is the sole explana-
tory variable included in the dose-response relation. Third, the dose-response
function is specified as a nonlinear relation in  accord with both a^ priori judg-
ment and empirical results regarding human responses to increased pollution
doses. Fourth, a threshold level is adopted before damages are estimated.

     This section, which represents a preliminary effort to estimate  empirically
a nonlinear dose-response function and a  linear "average" pollution  damage func-
tion, is presented in the following subsections:  Estimation of Physical Damage
Functions, A Linear General Physical Damage Function, Values of  Air  Pollution
Damages and Economic  Damage Functions, Premature  Mortality Damages and  Suspended
Particulates, and  Implications and Concluding Remarks.
I/ For a  detailed discussion  on  the  effect  of  air  pollution  on  human  health,
~~     see  Schrimper (1975).
                                      17

-------
ESTIMATION OF PHYSICAL DAMAGE FUNCTIONS

     For analytical purposes, two types of physical damage functions can be
posited: (1) dose-response or stimulus-effect relations; (2) general physical
damage functions which relate mortality not only to pollution, but also to other
relevant socioeconomic, demographic and climatological variables.

     A long-term, generalized physical damage function has been specified, for
example, by Lave and Seskin  (1970), Goldsmith (1965), and Ferris and Whittenberger
(1966) as

                     MR = F(D, S, E, W, A; e)                        (II-l)

where  MR  is mortality rate per 10,000 population and is related to  D   (demo-
graphic factors  such as age, sex, racial and genetic),  S  (the social factor
such as individual's exercise and other habits, nutrition, occupational struc-
ture, population density, and housing conditions),  E  (the economic variables
such as income and the level and quality of medical care received),  W  (weather),
 A  (the air pollutants), and  e  (the disturbance term). To measure the damage
effect of air pollution and other independent variables on mortality, the con-
ventional least-squares linear regression has been the common technique.

     If the objective is to estimate a short-term, day-to-day, physical damage
function for a given study region, demographic, social and economic factors can
then be reasonably assumed to be stable. Hence, the short-term physical damage
function can be  specified as

                            MR = f(W, A; e)                          (II-l1)

Lave and Seskin  (1972) utilized (II-l') to derive acute, day-to-day mortality-
pollution relationships. Lags up to 5 days in the pollution variables were in-
corporated into  the regression equations. Results obtained in their study were
generally negative because no discernible, consistent pattern of statistically
significant coefficients was observed. One of the several questions examined
by  Lave and Seskin which has bearing on policymaking is whether deaths are
merely  shifted by a few days by pollution episodes. Their finding indicates
that the reallocation of mortality extends over a period longer than 10 days.

Nonlinear Dose-Response Function

     A dose-response relation which includes the pollution variables as the sole
explanatory variable can be written as

                       MR = g(A; e)                                (II-2)

The dose-response functions may be estimated via controlled laboratory experi-
ments on human bodies. However, ethical and legal considerations prohibit the
use of human bodies for experimental purposes. Because of this, epidemiological
studies so far have not been fruitful in identifying the true cause-effect re-
lationship underlying mortality and air pollution.
                                       18

-------
     The clinical, laboratory, and experimental studies at relatively high con-
centration levels of sulfur dioxide or other pollutants suggest that tentative
dose-response relationships are available. However, information on such rela-
tionships is largely lacking at concentration levels in the range of current
standards. The best available tentative dose-response function ever produced
by epidemiological studies is presented in Buechley (1971). The function can
be approximated by a nonlinear, flat "S" shape curve, as shown in Figure II-l.
The relation indicates that while the air pollutant SO  is a contributing fac-
tor of premature mortality, the damaging effect is nonproportional. As the SO
concentration level increases, the excess mortality rate increases initially
at an increasing rate and continues to increase, but at a decreasing rate af-
ter a certain inflection level.

     It is generally the opinion of medical experts that the true a_ priori dose-
response is nonlinear. This hypothesis is also recently used by Leung (1974)
who studied the exposure-effect relation between human health and mobile source
air pollution best described by a nonlinear curve as shown in Figure II-l.

     On the basis of _a priori judgment of medical experts and the two empirical
results produced respectively by Buechley and Leung concerning the human health
damage responses to pollutant doses, the exposure-effect function relating mor-
tality rate (MR) to sulfur dioxide (SO ) in the present study is hypothesized
as an exponential function alternatively specified as follows:


                    MR = C + e(a-b/S°2)                  (II_3)
                    MR- c-e                           (H-3-)
                    In (MR - C) = a - b/S02              (II-4)
where  C  is the "conventional" mortality rate in that the mortality rate is
independent of the pollutants and  a and  b  are parameters determining the
shape of the nonlinear function. Since both coefficients  a  and b  can take
any real values, the  semilog, reciprocal equation (II-3) covers a wide range
of nonlinear functions with positive first derivatives.

     The conventional mortality rate  (C) is determined by a host of socioeconomic,
demographic, climatological and personal factors. It is recognized that many of
the factors known to  affect mortality are not amenable to quantification. Factors
such as nutrition, exercise, personal habits, etc., are difficult to measure
conceptually> while data on smoking habits have not been collected and on medi-
cal care are poorly measured. The  exclusion of these relevant factors from the
regression equation because of insufficient data may result in specification
errors and, hence, biased  estimates. Thus, careful interpretation of the regres-
sion results is warranted.
                                      19

-------
o
o
o
 0)
 o_
 D
 

O
 C
<


>-

Zj
<


o
     0
      0
    50


SO2(fig/m3)
75
            Figure  II-l. Hypothetical relationship between mortality rate

                        and SO  concentration.



                                         20

-------
     A number of regressions with data on more than 25 potential explanatory
variables collected from the 40 SMSA's which had a sulfur dioxide level equal
to or greater than 25 iig/m^ between  1968 and 1970 were run during the course
of this study. The selection of 25 ^.g/m^ as the threshold is based on two
considerations:  First, this concentration level is the average level prevail-
ing in rural areas. Second, this level is considered to be the "mean" of the
tolerable threshold distribution of  all individuals in the SMSA. Available evi-
dence suggests that no matter how small the concentration is, adverse health
effects may still occur (National Academy of Science, 1974). Thus, threshold
in a strict sense should be zero concentration. However, the threshold levels
with respect to all individuals in a given region could be reasonably viewed
as a symmetrical distribution with a mean level possibly at 25 (J-g/m^. The use
of this mean level of thresholds will probably result in more accurate damage
estimates than using zero  or other threshold levels.

     It should be noted that damage  estimates cited previously in other pollu-
tion studies were derived  on the basis of a zero threshold. The use of a zero
threshold level tends to overstate the damages.

     It is noteworthy that many of the determinants of mortality are difficult
to quantify, and data are  not readily available for some of the variables.JL'
The data for the above mentioned variables for 40 SMSA's which had a sulfur
dioxide level equal to or  greater than 25 [ig/m^ between 1968 and 1970 were
taken from a comprehensive quality of life study about U.S. SMSA's recently
completed by Liu (1975).   Variables  of no statistical significance or with
wrong signs were accordingly eliminated, and the best regression results with
the remaining seven independent variables were obtained as follows:
    CC = 229.6 + 741.8 PAGE - 119.7 PYAP - 0.12 PCOL - 76.58  PWPO
          (50.5)* (96.4)*      (62.8)**    (0.04)*     (21.6)*
                                                                    (II-5)
                      - 0.54 SUN + 0.23 RUM + 0.04 DTS
                       (0.24)*    (0.22)     (0.07)

                                           R  = 0.82
If In the studies of Lave and Seskin  (1970,  1973) and the Koshals (1974), a
     portion of the explanatory variables was used to estimate a general physi-
     cal damage. Lave and Seskin regressed mortality rates against air pollu-
     tions—particulates and sulfates—population density, proportions of non-
     white, proportions of people over age 64, and proportion of poor families,
     The Koshals selected the population density, the percentage of relative
     humidity and the pollutants-suspended particulate matter and benzene
     soluble organic matter as the explanatory variables in their mortality
     equation.

                                     21

-------
where  CC  denotes the computed conventional mortality rates, PAGE the percent-
age of population 65 or older,  PYAP  percentage of population with income above
poverty level,  PCOL  percent of persons 25 or older who have completed  4 years
of college, PWPO  percentage of white to total population,  SUN  possible annual
sunshine days,  RHM  relative humidity, and  DTS  number of days with thunder-
storms. The figures in the parentheses are standard errors of the estimates.
The estimated coefficients shown in the equation have the correct signs, and
with * and ** to indicate that they are statistically significant at the 1 and
5 percent level.

     The dose-response function embodying the effect of the threshold level of
25 (ig/m-^ is expressed as:

              (MR - CC) = EXP (a - b/(SO  - 25))

or

              RMR = EXP (a - b/(S02 - 25))                 (II-6)

where  CC  is the computed value of conventional mortality rate from equation
(II-5), and RMR = MR - CC is the residual mortality rate.

     The residuals, i.e.,  MR - CC = RMR,  take both positive and negative val-
ues. Since the logarithm of a negative number is undefined,  RMR was squared
prior to its logarithmic transformation. The resultant regression equation was
then adjusted by dividing the coefficients by 2. This adjustment is demonstrated
as follows:

     The regression equation takes the form
              2
     In  (RMR)  = 2 a - 2b/SC>2

     By virtue of a property of logarithm,  we also obtain

     2 In (RMR) = 2a - 2b/SO                                   (II-7)

or

     In (RMR)  = a - b/SO

Note that the coefficients in equation (II-7) are twice as large as those in
equation (II-4) which is the initially specified nonlinear dose-response function.
                                      22

-------
     The  regression result  for  equation  (II-4)  is  shown as  follows:


                     RMR2 =  EXP  (2.50  - 51.04/SO )
                                (1.34)   (4.22)*
or                                                              (II-8)
                    RMR = EXP  (1.25  -  25.52/SO  )
                                                R  = 0.03
The figures below  the  coefficients  are standard errors  with *  indicating  that
the coefficient of  SO  is  significant at  the  1 percent level.  Though  SO   ex-
plains only  3 percent of  the  residual mortality rate,  the nonlinear fit  showed
an explanatory power  150  times larger than the linear  fit. Generally  comparison
of R  when the dependent  variables are different may not be meaningful.  However,
the purpose  of comparing  R2 associated with RMR and In (RMR)  equations here is
to determine which  of the  two specifications  is more suitable for the  estima-
tion  of  the  physical  damage function. For comparison purposes,  such a  linear
regression equation is presented as follows:


             RMR = 29.65 - 0.034 SO                            (II-9)
                  (20.28)  (0.35)

                                          R2 = 0.0002
The  linear  fit  showed  not  only  very  low  explanatory power, but also an incor-
rect sign for S02-  Thus, the nonlinear specification of the dose-response rela-
tion seems  to be  superior  and tends  to support  the _a priori judgment regarding
human responses to  pollution dose  variations.

     To recapitulate,  the  methodological  procedures for estimating the function
between mortality rate and SO   are summarized as follows:

          1. A  linear  multiple  regression model represented by equation (II-5)
was developed for estimating the effects  of the socioeconomic, demographic, and
climatological  factors with the exclusion of air pollution on the conventional
mortality rate, C,  expressed in deaths per 10,000 population.

          2. The  computed values of C, i.e.,  CC, were subtracted from the ob-
served gross mortality rate. The residual, RMR = MR - CC, was then regressed
on S02 alone according to  the specification in equation (II-4). The regression
result was  shown  in equation (II-8). The  nonlinear, exponential dose-response
function was transformed into a linear function with logarithm on  RMR  and
reciprocal  on SO  for  empirical estimation.
                                     23

-------
          The nonlinear physical dose-response function between residual mor-
tality and SO  derived from this stepwise econometric technique is characterized
by the following features:

          1. The nonlinear dose-response function is consistent with the £
priori judgment about dose-response relationship between air pollution and mor-
tality rate. It can also be easily adjusted with whatever is the threshold
level of the SO  concentration when estimating the economic damages.

          2. For the purpose of predicting and computing the marginal mortality
damages due to SO , this nonlinear equation has the right sign and higher ex-
planatory power than its counterpart linear equation in view of its goodness
of fit.

          3. The nonlinear specification circumvents at least partially some
of the econometric problems such as multicolinearity and heteroscedasticity
                          •
which are to be discussed next.

Technical Problems in Estimation--
     Although detecting and treating econometric problems which are often en-
countered in the pollution effect studies are not the main purpose of this
study, the problem of multicolinearity and heteroscedasticity are examined dur-
ing the course of research.

     Multicolinearity--j:/It is well known that multicolinearity problems occur
when some or all of the explanatory variables are highly correlated and that
it becomes  difficult!, if not totally impossible, to disentangle their separate
influences. Of the nine explanatory variables used in this study, PWPO is cor-
related with the pollution variables, SO  and TSP. RHM is correlated with PAGE,
PYAP, PWPO, and SUN. The correlation coefficients are presented in Table II-1.
On the basis of this correlation coefficient table, one may be led to conclude
that not too "strong" multicolinearity appear to be present in this study. How-
ever, it should be noted that the usefulness of partial correlation coefficients
as a diagnosis of multicolinearity is questionable. Wichers (1975) has recently
shown that  a given value of partial correlation coefficient may be compatible
with two very different multicolinearity patterns. Less obtusely stated, a simple
correlation coefficient may not be the appropriate measure of multicolinearity.
 \l For a  detailed  discussion on multicolinearity see Johnston  (1963), p.  207,
     Goldberger  (1964), pp. 192-193, Farrar and Glauber  (1967), and Haitovsky
      (1969). The three-stage test for the detection of multicolinearity patterns
     in the classical regression model was criticized by Kumar  (1975), Wichers
      (1975), and O'Hagen and McCabe  (1975). Kumar cast doubt on the x^ test  sug-
     gested by Farrar and Glauber for the existence of multicolinearity and  on
     the  F and t tests to localize the problem. Wichers  showed that the third
     stage of the  Farrar-Glauber test is ineffective. O'Hagan and McCabe  pointed
     out  a fundamental error which renders meaningless the contribution of
     Farrar-Glauber  to multicolinearity as a  sample problem.
                                     24

-------
                      TABLE  II-l.  CORRELATION  COEFFICIENTS!/
PAGE
PYAP
PCOL
PWPO
SUN
RHM
DTS
S02
TSP
0.
-0.
0.
0.
-0.
0.
0,
0.
0.
74
26
61
36
25
23
05
13
24
-0.12
-0.41
0.72
-0.09
0.35
-0.20
0.05
-0.09
0.25
0.33
-0.01
0.45
-0.19
-0.10
-0.23
-0.38
0.18
-0.04
-0.19
0.08
-0.15

-0.26
0.42
-0.13
-0.27
-0.33


-0
-0
0
-0


.36
.17
.08
.23



0.
-0.
-0.



00
08 -0.04
01 0.06





0.04
        MR
PAGE   PYAP *  PCOL   PWPO   SUN    RHM    DTS
SO,
_a/ Correlation  coefficients  are  statistically  significant at  5 percent  level
     if r^0.32  for  40  observations.
     Thus,  diagnosis  of multicolinearity  could  be  guided by _a priori judgment
with respect  to  the interactions  among  the  explanatory variables. Furthermore,
the existence of multicolinearity poses little  problem if the model is correctly
specified,  because in such  a  case least-squares estimates will be unbiased re-
gardless of the  extent of multicolinearity. The estimates will be biased if a
relevant variable is  omitted  and  inefficient if a  nonrelevant variable is in-
cluded in the regression analysis. The  extent of the biases is dependent on the
degree of correlation between the misspecified  variable and the variables with
significant coefficient.

     In the presence  of multicolinearity, no cut-and-dried technique has been
discovered  to treat the problem.  The  residualization technique was first used
by Ridker (1965, p. 127-135)  in a study of  property value and pollution to al-
leviate the multicolinearity  problem  by attributing to all the nonpollution
variables the covariance between  them and the pollution variable. The two-stage
estimation procedure  is known to  bias the pollution coefficients toward zero
and reduce  their significance in  the  presence of multicolinearity.

     Residualization  technique was later  employed  by Lave and Seskin (1973) to
examine the multicolinearity  problem. However,  the estimated results obtained
by Lave and Seskin indicate that  the  estimated  coefficients of the air pollut-
ants retain their significance and the  parameter estimates are similar to those
in the one-stage regression equation.

     Following Ridker and Lave and Seskin,  the  residuals rather than the gross
mortality rates were  regressed on the air pollution variables. In doing so,
not only the  nonlinear dose-response  function can  be estimated, but also the
                                     25

-------
possible multicolinearity problem existing among the explanatory variables can  be
alleviated. The low R  for the dose-response function is expected from using
this two-stage residualization technique. However, the important result is that
the nonlinearity of dose-response function represents a better fit than the lin-
ear specification as pointed out previously.

     Heteroscedasticity- The violation of the condition of a constant variance
in the disturbance term in any regression analysis is called heteroscedasticity.
The effect of heteroscedasticity is not on the biasness of the estimated regres-
sion coefficient itself, but rather on efficiency of the variance of the coefficient
estimated. It is recognized that the existence of heteroscedasticity often occurs
in the cross-section data. In the present study, heteroscedasticity is detected
by using the eyeballing -method. In terms of Figure II-2, the residuals are
plotted against the dependent variables. The shape of the residual distribution
pattern suggests that the variance of the error term is variable, i.e., there
is likely  a problem of heteroscedasticity. Glejser (1969) and Park (1966) dis-
cussed alternative methods for detecting heteroscedasticity. These methods have
been applied by Smith and Deyak (1975) for testing heteroscedasticity in estimat-
ing air pollution and property value relation.

     The common treatment for heteroscedasticity is to use the weighted regres-
sion method designed to reduce the nonhomogeneity of the variance. The use of
semilog on the dependent variable in this study is a sort of the weighted re-
gression method. The semilog transformation reduces the nonhomogeneous spread
of the variance in the error term (e.g., along the mortality rate axis in Figure
II-2, on page 27), and, hence, partially alleviates the heteroscedasticity
problem.
A LINEAR GENERAL PHYSICAL DAMAGE FUNCTION

     As noted earlier, reliable and useful average damage functions on mortality
rate and air pollution for the United States metropolitan areas are still lack-
ing. To close this gap in the air pollution damage investigation, a generalized
average damage function is developed by regressing jointly, in a linear form,
the sum of the estimated mortality rates from both equations (II-5) and (II-8)
on the four socioeconomic and demographic variables, the three climatological
variables and the SO  . It should be stressed that the results of this gener-
alized average damage function should only be used for prediction purposes,
and any statistical interpretations would be meaningless. Otherwise stated,
this damage function  so derived serves to yield a more accurate prediction with
respect to the changes in the mortality rates in response to a ceretis paribus
change in any of its  determinants. Based on the data of the 40 SMSA's with SO
exceeding 25 |_lg/m3 between 1968 and 1970, the linear regression analysis was
conducted to ascertain the generalized average damage function, estimated ass
                                     26

-------
   10
                                                                 /
•H
CO
                                  /
                         *
-1


-2


-3


-4


-5
      \ 70
    L  \
          \
            \
 -211
              \
                \
                  \
                    \
                      \
                        \
                          \
                            \
                              \
                                         -*-*	1	1	
                 80      •  90        W100       110        120        130
                                \
                                 \




                                             \
                                               \
                                                                                    Mortality

                                                                                      Rate
                                                             \
                 Figure II-2. Heteroscedastic distribution  of  the  residuals.

                                          27

-------
      CMR = CC + CRMR

       = 226.2 + 735.4 PAGE - 113.8 PYAP - 0.12 PCOL - 77.5 PWPO
          (3.4)*  (8.7)*       (5.3)*      (0.003)*    (2.0)*
                                                                      (11-10)
               - 0.55 SUN + 0.23 RHM + 0.03 DTS + 0.023 S02
                (0.02)*    (0.02)*    (0.006)*   (0.003)*
where  CMR  is the computed mortality rate, which is the sum of the computed
conventional mortality rate  (CC)  and the computed residual mortality rate
(CRMR)  from equations (II-5) and (II-8), respectively. All independent vari-
ables on the right-hand side of equation (11-10) were defined previously.

     Admittedly, a usual statistical interpretation for the generalized damage
function summarized by equation (11-10) is not meaningful. However, the purpose
of deriving this equation is to demonstrate that the stepwise econometric model
ameliorates some technical problems of estimation. The advantage of this approach
is clear if equation  (11-10) is compared with the similar physical damage func-
tion using the actual rather than Computed mortality rates as the dependent
variable. Such a physical damage function is summarized as follows:


      MR = 230.1 + 746.4 PAGE - 119.3 PYAP - 0.12 PCOL - 77.7 PWPO
           (51.5)  (10.59)      (63.9)      (0.035)     (24.3)
                                                                     (II-ll)
         - 0.54 SUN + 0.23 RHM + 0.04 DTS - 0.004 SO
          (0.25)      (0.22)     (0.07)     (0.033)
It is noteworthy that the coefficient of SO  in equation (II-ll) is negative
despite the fact that the simple correlation coefficient between MR and SO  is
positive and equal to 0.13. The negativity of the SO  coefficient is probably
due to multicolinearity and other econometric problems discussed earlier. The two-
step  econometric method  seems  to have partially  overcome  these  technical  problems
and yields, if not coincidentally, the expected positive coefficient of SO  in
equation (11-10).
VALUES OF AIR POLLUTION DAMAGES AND ECONOMIC DAMAGE FUNCTIONS

     Air pollution damage to human health in this country has been roughly es-
timated by Ridker (1965), Lave and Seskin (1970, 1973),  Jaksch and Stoevener
(1974), Koshal and Koshal (1974), Park  (1974), and others. However, their es-
timates vary considerably; from $443 million by Ridker to $2.4 billion by Lave
and Seskin, and $6.8 billion by Park, partially because their study scopes and
period are not commensurate with each other. In order to estimate an average
economic damage function for the United States urban areas, it is not meaning-
ful to borrow the national damages estimated by the above authors not only
                                     28

-------
because of this great disparity but also the different methods of estimation.
A method will be developed to quantify regional damage separately for each met-
ropolitan area so that regional control costs and benefits can be evaluated.
Since we considered 25 (ig/m^ as the threshold of SO , only those SMSA's with
average annual SO  levels eual to or greater than 25 i/m^ between 1968 and
                   levels equal to or greater than
1970 and with data on other relevant factors were selected.

     Air pollution has caused high morbidity rates in addition to premature
mortality in this country. This section, however, is mainly concerned with the
mortality damages. The morbidity damages due to air pollution will be discussed
in Section III. To estimate the mortality damages of SO  and the percentage of
pollution-caused damage to total mortality losses, an expected average permanent
income method was developed. Specifically, we computed via equations (II-5) and
(II-8) the conventional and the residual mortality rate for the selected SMSA's.
Assume that each individual in any of the SMSA's is equally affected by air pol-
lution and that the growth in median earnings from 1960 to 1970 represents an
expected normal income rate. The expected future income streams are computed
by a simple formula computed for the conventional and air pollution victims in
the labor force- -between 18 and 64 years of age. The present value of the eco-
nomic damages was  derived by discounting the future incomes at a rate of 4 per-
cent which is the long-term bond rate. Finally, we regressed the computed economic
losses of both conventional and pollution victims on the demographic, socioeco-
nomic, and weather variables, and SO  for the selected SMSA's to derive the so-
called "average" economic damage function.

     In functional form, this part of the work for each SMSA can be succinctly
expressed as follows:.!;/
                 V = Y
                V = Y
                         n
                                  fc/(i + i)fc  •  L  •
                         n
(1 + r)
(1 + r)  /(I  + i!
[cc + CEMR]   (11-12)
                        t=l
                                              L  • CG
              (11-12')
I/ A somewhat different formula was developed and employed by Ridker for esti-
~~   mating damage costs due to premature death. A drawback of his method, as
    noted by Ridker himself, is the lack of adjustment for increase in labor
    productivity over  time. A similar framework was also used by Schrimper
    (1975) to calculate mortality costs for Chicago. The expected income formula
    developed here considers the improvement in labor productivity, though all
    workers are assumed to live through and be employed until the age of 65.
    The bias in the resulting estimates is believed to be negligible.
                                     29

-------
where  V  and  V  are, respectively, the computed value of regional economic
damages with or without air pollution;

     Y  is the weighted median income of 1970 between males and females with
        the weights being their respective share in the labor force;

     r  is the expected family income growth rate which partially reflects the
        growth in labor productivity assumed to be equal to the average from
        1960 to 1970;

     i  is the discount rate, set at 4 percent per year, a rather conservative
        rate;

     L  is the labor force or population between 18 and 64 years of age;

     CRMR  and  CC  are the computed excess mortality rates and the computed
        conventional mortality rate, respectively;

     n  is the difference between regional median age and 64; this assumes that
        the number of deaths due to air pollution with age younger than the me-
        dian age is offset by those who fall short of reaching the age of 64.

     The damage costs without and with air pollution and the per capita damage
costs for 1970 by SMSA are estimated using equations (II-8), (11-12), and (II-
12') and are contained in Table II-2. All dollars reported in the table are in
1970 value. Under the heading of mortality damage due to SO , total and per
capita mortality cost for each SMSA can be found in Columns 1 and 2. Mortality
damages in the absence of air pollution is presented in Column 3, and Column
4 presents the ratio or the relative magnitude of total mortality cost attribut-
able to SO  and the mortality damage with and without SO . The higher the ratio,
the more serious is the pollution damage.

     It should be noted that the damage estimates presented in the table depend
vitally on the assumptions made in this study. The most critical assumptions
are the threshold levels of SO  the natural mortality rate, the growth in in-
come and the discount rates. Change in any of these assumptions would result
in modification in the damage estimate.

     As readily revealed in the table, total mortality costs in the presence
of SO  amounted to $887 million for the 40 SMSA's which had average annual SO
concentration beyond 25 (0,g/m^ between 1968 and 1970. Given that total mortality
cost in the absence of air pollution in the 40 SMSA's is $60.2 billion, as mea-
sured, the air pollution damage accounted for 1.4 percent of the total. Among
the 40 SMSA's, New York City had the highest total and per capita mortality air
pollution damage, about $329 million and $28.4 respectively, partially because
it had the highest SO  concentration level between 1968 and 1970, i.e., 210
      . The highest percentage of air pollution damage was found in Chicago and
New York City; 2.7 percent of total gross mortality values in these areas could

                                    30

-------
                          TABLE 11-2. MORTALITY COSTS WITH S02 BY SMSA1 s, 1970

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.
33.
34.
35.
36.
37.
38.
39.
40.

SMSA
Akron, OH
Allentown, PA
Baltimore, MD
Boston, MA
Bridgeport, CT
Canton, OH
Charleston, WV
Chicago, IL
Cincinnati, OH
Cleveland, OH
Dayton, OH
Detroit, MI
Evansville, IN
Gary, IN
Hartford, CT
Jersey City, NJ
Johnstown, PA
Lawrence , MA
Los Angeles, CA
Minneapolis, MN
New Haven, CT
New York, NY
Newark, NJ
Norfolk, VA
Paterson, NJ
Peoria, IL
Philadelphia, PA
Pittsburgh, PA
Portland, OR
Providence, RI
Reading, PA
Roche ster, NY
St. Louis, MO
Scranton, PA
Springfield, MA
Trenton, NJ
Washington, D.C.
Worchester, MA
York, PA
Youngs town, OH
Total
( g5/™3)
51
57
54
31
40
30
27
120
25
64
25
38
25
58
57
75
25
52
35
38
40
210
37
26
28
26
84
57
26
67
30
32
40
30
87
32
47
31
31
30

Mortality
Due to
Total
(in 106)
(1)
8.4
7.5
28.4
1.3
2.6
0.1
—
178.0
--
34.3
__
26.0
--
10.6
10.5
9.6
—
2.7
15.9
9.1
2.2
329.0
7.0
--
--
--
97.9
30.0
--
14.6
0.1
0.8
13.3
—
10.6
0.2
35.5
0.2
0.1
0.1
886.6
Damage
S02
Per
Capita
(2)
12.4
13.8
13.( 7
0.5
6.7
0.3
..
25.5
.-
16.6
_„
6.2
--
16.7
15.8
15.8
--
11.6
2.3
5.0
6.2
28.4
3.8
--
--
--
20.3
12.5
--
16.0
0.3
0.9
5.6
0.01
20.0
0.7
12.4
0.6
0.3
0.2

Mortality Damage
Without Air
Pollution Ratio
(in 106) (lH((l)+(3)
(3) (4)
570.6
462.5
1891.6
2398.7
353.4
330.9
159.0
6292.0
1160.0
1875.7
398.0
4884.0
214.0
555.4
552.5
529.4
263.0
204.3
4964.1
1380.9
341.8
11671.0
1633.0
511.0
1150.0
295.0
4322.1
2000.0
922.0
777.4
257.9
784.2
2156.7
185.0
458.4
254.8
1194.5
319.8
263.9
482.9
60,221.4
0.0145
0.0160
0.0148
0.0005
0.0073
0.0003
__
0.0275
--
0.0180
__
0.0053
..
0.0187
0.0187
0.0178
--
0.0130
0.0032
0.0065
0.0064
0.0274
0.0043
--
--
--
0.0221
0.0148
--
0.0184
0.0004
0.0010
0.0061
--
0.0226
0.0008
0.0175
0.0006
0.0004
0.0002


Note:  	 individual figure may not add to  totals due  to  rounding.
                                          31

-------
be attributed to SO .  In order of magnitude, New York City, Chicago, and
Philadelphia all had air pollution damages of more than $50 million. In terms
of the ratio of net mortality damage to gross mortality damage, i.e., Column
4, again New York and Chicago which had ratio values of 2.7 percent lead all
the other SMSA's. As noted earlier, the degree of the pollution damage is par-
tially reflected by the magnitude of this ratio.

     Although the economic damage costs derived in this section are more detailed
than prior estimates, they are still crude information and should be used with
caution under the stated conditions. In order to develop a marginal economic
damage function useful for prediction and control purposes, the "total of eco-
nomic costs of mortality" is related not only to SO , but also to various socio-
economic, demographic, and climatological characteristics of different regions.
The stepwise  regression technique was used with inputs from the 40 sample ob-
servations to estimate the economic damage function. The regression results are
shown as follows:
         V = 10,295 + 47.02 SO  - 8,128.4 PWPO + 98.5 RHM + 72.3 SUN
             (11,023)  (6.97)*    (5,195.9)       (46.9)*    (53.4)
                                                                      (11-13)
           - 15.98 DTS -  16,191.8 PYAP +7.7 PCOL + 3,772 PAGE
             (15.99)     (13,659.9)       (7.6)      (22,650)

                             R2 = 0.74
where   V  is  total mortality cost obtained from equation  (11-12) and all the
explanatory variables are defined earlier.

     The coefficients and standard errors in  (11-13) are  reduced by a factor
of  10  .  The values of standard error are presented below  the coefficients, with
* to indicate that the coefficient is  significant at the  1 percent level.

     The economic damage function derived can be useful to policymakers in esti-
mating  the marginal and average damages  (benefits) resulting from a pollution
control  program. To serve as an illustration, an example  involving the computa-
tion of  the partial elasticity of an explanatory variable and  the associated
marginal benefit due to the changes in that variable is presented. Suppose the
federal  government is considering the  implementation of a pollution abatement
program which is expected to reduce the average SO  level in the urban areas
by, say,  10 percent. What will then be the dollar worth benefit of the reduced
premature mortality rate as a result of the pollution abatement program? Since
the average total damage cost due to premature mortality  is $1,530.8 million
and the average  SO   level is 47.95 (j/g/m  among the  40 SMSA's,  the partial  elas-
ticity of the  damage cost with respect to SO  is derived by using the formula
that

                                     32

-------
      E. QH  =  @c/o-(S00)) x  (SO./c) =  47.02 x  (47.95/1,530.8)  = 1.45.
       Cjdu            /        z


Note (9c/d(SO )) in the formula denotes  the coefficient  of  SO  in  the economic
damage function; SO  and  c   are, respectively,  the mean values of  SO   and  the
total damage cost for the 40  SMSA's  included in  the  sample.

     The distinguishing property  of  the  concept  of elasticity is that it  is  a
unit-free measure of the percentage  change  in the dependent  variable with re-
spect to the percentage change in any of the explanatory variables  while  hold-
ing other things equal. Given the computed  elasticity  of damage cost with re-
spect to SO  , EC SQ2  = 1.45, it  is  in  general  expected  that a  10 percent de-
crease in the SO  concentration level will  result in a 14.5  percent reduction
in the premature mortality damage cost.  Since the mean value of the regional
damage cost  for the 40 SMSA's is  $1,531  million, when  the SO level decreases
from 47.95 (J,g/m3 to 43.15 |j,g/m3,  it  is  expected  that on  the  average the damage
cost will be reduced by the amount of $1,530.8 x 14.5  percent = $221.9 million.
Likewise, the elasticities for the other explanatory variables  can be analogously
computed and interpreted.
PREMATURE MORTALITY  DAMAGES  AND  SUSPENDED PARTICULATES

     Earlier  studies have  established  a positive qualitative relationship be-
tween mortality and  suspended particulates. Recently Lave and Seskin (1970,
1973), and the Koshals  (1974) further  confirmed the existence of a quantitative
association between  mortality and  the  particulates. As discussed earlier, the
threshold effects  of the air pollutant and the heteroscedasticity problems in
the empirical estimation of  the  relation were, by and large, ignored in the prior
studies. A two-step  residualization  technique was, however, developed earlier
to cope with  these problems  in estimating a nonlinear dose-response function.
The same methodology is used in  this section to establish a dose-response func-
tion relating mortality to suspended particulates.

     The nonlinear dose-response relation (II-3)  was used for regressing the
residual mortality rate  (MR-C)  on  total suspended particulates (TSP). To be
consistent with earlier SO  estimates, the particulate level is also adjusted
by a threshold of  25 |i.g/m3 in computing the physical damage. As noted earlier,
although 25 (ig/m^  is a reasonable  level for capturing the threshold effect,
alternative thresholds may also  be considered. Changes in the threshold will
cause modifications  in the damage  estimates. It is conceivable that, other
things being  equal,  a lower  threshold  level implies higher damage cost.
                                       33

-------
     The least-squares regression yields the following nonlinear physical damage
functions for suspended particulates:
         RMR = EXP [1.30 - 65.75/TSP]
                   (0.83) (70.84)                                   (11-14)

                                          R2 = 0.02
The values below the coefficients are standard errors. The explanatory variable
TSP is not statistically significant. With the availability of the physical dam-
age function, the methodological procedures used earlier for estimating the pos-
tulated function relating mortality rate to SO   can be employed to estimate
the economic damages and the associated economic damage function for suspended
particulates.

Economic Damage Functions

     For policymakers, economic damage functions may be more relevant than physi-
cal damage functions. An economic damage function, or a monetary damage function,
relates levels of pollution to the amount of compensation which would be needed
in order that the society is not worse off than before the deterioration of the
air quality. The economic damage function is useful to decision makers since
the multiple dimensions of the decision problem are reduced into one dimension
only, i.e., money. It should be noted, however, that transformation of a physi-
cal damage function into an economic damage function often involves value judg-
ment on the part of the policymaker- A related question as to the degree of
conformity of the values of the policymaker with those of the consumer sover-
eignty is largely unresolved.

     The expected permanent income method delineated earlier was employed to
estimate premature mortality damages due to total suspended particulates. The
damage costs associated with total suspended particulates are presented in
Table II-3. Columns (1) and (2) present total and per capita mortality damage
attributable to TSP. Mortality damage without air pollution is presented in Col-
umn (3). Column (4) presents ratio of total mortality damage due to TSP to total
mortality damage with TSP. This ratio reflects the relative magnitude of the
damage attribute TSP to total mortality damage. An examination of the table re-
veals that the mortality damages range from $1.4 million in Lawrence,
Masssachusetts, to $155 million in New York City. The air pollution damage in
Lawrence is 0.7 percent of the total gross mortality damage, while in New York
City suspended particulate causes about 1.3 percent of total mortality damage.
The highest ratio of pollution damage to total mortality damage of the magnitude
of 4.0 percent is observed for Dayton, Ohio.

     Generalized economic damage functions were derived by regressing the pre-
mature mortality damage costs associated with TSP (TMRCT) which is the sum of
Columns (1) and (3) in Table II-2 on the demographic, socioeconomic, and

                                      34

-------
                        TABLE II-3. MORTALITY COSTS WITH TSP BY SMSA's,
                                  (in dollars)
                                                                1970
Mortality Damage
Due to TSP

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.
33.
34.
35.
36.
37.
38.
39.
40.

SMSA
Akron, OH
Allentown, PA
Baltimore, MD
Boston, MA
Bridgeport, CT
Canton, OH
Charleston, WV
Chicago, IL
Cincinnati, OH
Cleveland, OH
Dayton, OH
Detroit, Ml
Evansville, IN
Gary, IN
Hartford, CT
Jersey City, NJ
Johnston, PA
Lawrence, MA
Los Angeles, CA
Minneapolis, MN
New Haven, CT
New York, NY
Newark, NJ
Norfolk, VA
Paterson, NJ
Peoria, IL
Philadelphia, PA
Pittsburgh, PA
Portland, OR
Providence, RI
Reading, PA
Rochester, NY
St. Louis, MO
Scranton, PA
Springfield, MA
Trenton, NJ
Washington, B.C.
Worchester, MA
York, PA
Youngs town, OH
Total
TSP
(ug/tn3)
80
87
147
108
57
103
105
155
106
201
114
153
75
105
74
83
103
65
118
76
60
95
134
113
56
78
78
135
86
77
117
90
120
189
64
71
90
72
85
110

Total
(in 106)
(1)
7.2
6.0
42.0
42.9
2.0
5.3
2.5
147.0
18.9
47.8
16.4
116.0
2.0
10.6
6.4
5.4
3.7
1.4
106.0
18.8
1.9
155.0
33.6
11.2
6.0
3.3
45.8
38.5
11.1
8.0
4.3
•11.7
38.5
3.7
3.1
2.4
43.3
2.8
3.3
8.2
1044.0
Per
Capita
(2)
10.6
11.0
20.3
15.6
5.1
14.2
10.9
21.1
13.6
23.2
19.3
27.6
8.6
16.7
9.6
8.9
14.1
6.0
15.1
10.4
5.3
13.4
18.1
16.5
4.4
9.6
9.5
16.0
11.0
8.8
14.5
13.3
16.3
15.8
5.9
7.9
15.1
8.1
10.0
15.3

Mortality Damage
Without Air
Pollution Ratio
(in 106) (l)-K(D+(3)
(3) (4)
570.6
462.5
1891.6
2398.7
353.4
330.9
159.0
6292.0
1160.0
1875.7
398.0
4884.0
214.0
555.4
552.5
529.4
263.0
204.3
4964.1
1380.9
341.8
11671.0
1633.0
511.0
1150.0
295.0
4322.1
2000.0
922.0
777.4
257.9
784.2
2156.7
185.0
458.4
254.8
1994.5
319.8
263.9
482.9
60,221.4
0.0125
0.0128
0.0217
0.0176
0.0056
0.0158
0.0155
0.0228
0.0160
0.0249
0.0396
0.0232
0.0093
0.0187
0.0115
0.0101
0.0139
0.0068
0.0209
0.0134
0.0055
0.0131
0.0202
0.0214
0.0052
0.0111
0.0105
0.0189
0.0119
0.0102
0.0164
0.0147
0.0175
0.0196
0.0067
0.0093
0.0212
0.0087
0.0124
0.0167

Note:  — individual  figure may not  add to totals due  to rounding.
                                           35

-------
climatological variables and the suspended participate level for the 40 SMSA's.
The stepwise regression result for the generalized economic damage function is
summarized below:


     TMRCT = 13,109 + 7.63 TSP - 20,225 PWPO + 85.05 RHM + 6.85 SUN
            (18,034) (11.68)     (8,339)*     (72.85)     (87.49)

           - 14.88 DTS - 10,554 PYAP + 11.39 POOL + 54,606 PAGE
            (24.89)     (21,076)      (12.25)      (33,019)

                                         R2 = 0.38             (11-15)
     The values below the coefficients are standard errors. The symbol * indi-
cates that the coefficient is significant at the 1 percent level. The coeffi-
cients and standard errors are reduced by a factor of 106. The explanatory
variables are the same as those appearing in the sulfur dioxide economic damage
function and explain about 38 percent of the variations in the dependent vari-
ables .

     Given the mean values of total damage cost and total suspended particulate
level are $1,543.5 million and 100.9 (j,g/m^, respectively, the partial elasticity
of damage cost with respect to total suspended particulate is:
                  Ec TSP = ?*63 X (100-9/1»543) = 0.49
     Thus, a 10 percent decrease in the average TSP level as a result of pollu-
tion control programs will cause a reduction of 4.9 percent in the premature
mortality damage cost. That is, when the TSP is reduced from 100.9 [^g/m3 to
89.81 |j,g/m3 the damage cost, on the average, will be reduced by the amount of
$1,543 x 4.9 percent = $75.6 million.
IMPLICATIONS AND CONCLUDING REMARKS

     This  study is the first attempt to estimate a physical-nonlinear damage
function between excess mortality rates and the SO  concentration with consid-
erations of circumventing certain econometric problems such as multicolinearity
and heteroscedasticity, and accounting for the effects of the threshold  levels.
Through a  two-step adjustment procedure, the average physical mortality.function
was generalized with a rather complete specification. That is, the  generalized
average mortality model includes not only the major demographic, socioeconomic,
and climatological determinants but also air pollution variables. The two-step
econometric model developed here represents a constructive response to the  call
recently made by Lave and Seskin (1973) and Ferris (1970) in connection  with
the urgent need to improve on the existing studies in the area of air pollution
and human  health.
                                     36

-------
     This study is also the first attempt to present comparable estimates for
premature mortality damages due to "excessive" air pollution—sulfur dioxide
and total suspended particulates—for individual urban areas in the United
States. To assist the policymakers in estimating possible marginal damage
(benefit) resulting from a given pollution control strategy, "average" economic
damage functions which transform the multiaspect of the problems into a single,
homogeneous monetary unit were also developed separately for the two major pol-
lutants.

     It should be noted that the present federal standards were derived on the
assumption that threshold levels for various pollutants exist. These threshold
levels are considered to be the safe levels below which essentially no person
is hurt. This threshold level concept has been attacked by many medical experts
on the grounds that evidence has failed to support a genuine clear-cut lower
limit. It is our contention that the threshold model of health effects, however,
should not be taken literally, as some experts suggested.

     The threshold of 25 |j,g/m3 was used in this study in deriving the damage
estimates because it is viewed as the mean level of the underlying distribution
of tolerable threshold levels of all the individuals in a given SMSA. Further-
more, it is also the average concentration level in the rural areas where lit-
tle air pollution damage on human health is observed. Thus, while the annual
average concentration level is below the threshold level, the majority of the
population in the rural areas is assumed not hurt from the presence of air pol-
lution. In order to derive more accurate "average" damages of pollution in a
given region, it is imperative to establish threshold level which is the mean
level of the actual threshold distribution. Our model is easily adaptable for
any threshold levels that one would like to consider as tolerable.

     Another issue which merits discussion is the possible chemical interactions
among the pollutants. It is generally recognized that the total effect of sev-
eral pollutants present at the same time in the air may be greater or less than
the sum of their individual effects. In other words, the interaction effect may
be additive, synergistic or even antagonistic. Two types of interactions should
be noted:  (1) physiological, and (2) chemical. Both types of interactions are
expected to occur. The crucial question is how and to what extent air pollutants
interact with each other. Stated differently, the question is whether the inter-
action effects are of sufficient magnitude to negate the present method of estab-
lishing the air quality standards. All panel experts, according to a recent Na-
tional Academy of Sciences study (1974), found that synergistic effects are not
important enough to invalidate the current methods which set air quality stan-
dards for each major pollutant.

     Since the synergism occurs when SO  and TSP are present at the  same time,
the independency assumptions employed in this study may result in underestimating
the damages. However, it has been well recognized that both SO  and  TSP may be
merely convenient indexes of all major damaging pollutants. This measurement
problem of the pollutants contributes to overestimating the damages. In view
of these two opposing factors, we are unable to judge whether our procedure
tends to result in upward or downward biased estimates.
                                     37

-------
     Other conceptual and empirical problems often encountered in estimating
air pollution damage also should be noted. The major difficulties include the
lack of knowledge regarding the shape of the function which describes the re-
lationship between air pollution and health, the lack of a theoretical model
specifying the way air pollution affects health, the virtual impossibility of
accounting for all factors that might affect human health, and errors of ob-
servations in the data. Some of these problems, however, have been tackled in
the present paper. For example, the nonlinear dose-response relation was spec-
ified for the excess mortality rate and the pollutant concentration level. The
specification of this more plausible physical dose-response function would par-
tially account for the credibility in our air pollution damage estimates. The
semilog transformation reduces the heteroscedasticity while the use of the re-
siduals ameliorates  the multicolinearity problem.

     Although the income  foregone or productivity models have been employed  by
economists,  their application  to mortality  or deaths in individual metropolitan
areas  due to  SO  , TSP, and other causes opens up another avenue for air pollu-
tion damage  quantification which seems to be much more desirable than earlier
studies unveiling only some aggregate figures for the nation as a whole. Fur-
thermore,  the average air pollution damage  functions derived in this study with
observations  from a  selected set of SMSA's  with pollution  level above the thresh-
old are conceivably  more  meaningful than prior  studies which included all SMSA's
as sample observations regardless  of the concentration level of air pollution.

     This section presents a set of more recent estimates  of air pollution  dam-
age for each of  the  40 SMSA's  with concentration  levels higher than 25 p-g/m3.
Based  on  the conservative assumption employed,  it is found that while SO    alone
 in 1970 costs approximately  $887 million, or  about  1.4 percent of total mortality
 costs  in  these  areas, TSP imposes  about $1,047 million damages,  or about  1.7 per-
 cent on  the total mortality  costs  in the  same areas.

      The  mortality  damages  due to  SO  and TSP and the mortality  damages without
 air pollution from  Tables II-2 and II-3 are reproduced in  Table  II-4. Column
 4 of Table II-4 presents  the total mortality  damage  costs  which  are the  sum of
 the three component  damages  listed in Columns 1,  2,  and 3.  The SO  damage  esti-
mates  derived in this section  should replace  the  earlier  estimates reported by
 Liu (1975).

      The  results presented  in  this  section  are  only suggestive and tentative.
 Given  the tentativeness  and  experimental nature of  the methodological  and
 statistical procedures,  and  the degree  of uncertainty associated with the
 study results,  a great deal  of caution  should be  exercised in using  the  prod-
ucts of  this research. However,  the  availability  of average or marginal  damages
 is instrumental in  determining the optimal  national or  regional  pollution con-
 trol strategies.

      The  current problem seems far more complex than the  question of  balancing
 the benefits to polluters against  damages  inflicted on  the receptors. The  issues
are pressing and not yet  well  specified.  The  basic  difficulty in applying  the

                                     38

-------
                          TABLE II-4. MORTALITY COSTS BY SMSA's, 1970

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.
33.
34.
35.
36.
37.
38.
39.
40.

SMSA
Akron, OH
Allentown, PA
Baltimore, MD
Boston, MA
Bridgeport, CT
Canton, OH
Charleston, WV
Chicago, IL
Cincinnati, OH
Cleveland, OH
Dayton, OH
Detroit, MI
Evansville, IN
Gary, IN
Hartford, CT
Jersey City, NJ
Johnstown, PA
Lawrence, MA
Los Angeles , CA
Minneapolis, MN
New Haven, CT
New York, NY
Newark, NJ
Norfolk, VA
Paterson, NJ
Peoria, IL
Philadelphia, PA
Pittsburgh, PA
Portland, OR
Providence, RI
Read ing , PA
Rochester, NY
St. Louis, MO
Scranton, PA
Springfield, MA
Trenton, NJ
Washington, D.C.
Worchester, MA
York, PA
Youngs town, OH
Total
Total Mortality
Damage Due to
S02 (106) (1)
8.4
7.5
28.4
1.3
2.6
0.1
--
178.0
--
34.3
--
26.0
--
10.6
10.5
9.6
--
2.7
15.9
9.1
2.2
329.0
7.0
--
--
--
97.9
30.0
--
14.6
0.1
0.8
13.3
--
10.6
0.2
35.5
0.2
0.1
0.1
886.6
Total Mortality
Damage Due to
TSP (in 106) (2)
7.2
6.0
42.0
42.9
2.0
5.3
2.5
147.0
18.9
47.8
16.4
116.0
2.0
10.6
6.4
5.4
3.7
1.4
106.0
18.8
1.9
155.0
33.6
11.2
6.0
3.3
45.8
38.5
11.1
8.0
4.3
11.7
38.5
3.7
3.1
2.4
43.3
2.8
3.3
8.2
1044.0
Mortality Damage
Without Air
Pollution
(in 106) (3)
570.6
462.5
1891.6
2398.7
353.4
330.9
159.0
6292.0
1160.0
1875.7
398.0
4884.0
214.0
555.4
552.5
529.4
263.0
204.3
4964.1
1380.9
341.8
11671.0
1633.0
511.0
1150.0
295.0
4322.1
2000.0
922.0
777.4
257.9
784.2
2156.7
185.0
458.4
254.8
1994.5
319.8
263.9
482.9
60,221.4
Total Mortality
Damage
(in 106)
586.2
476.0
1962.0
2442.9
358.0
336.3
161.5
6617.0
1178.9
1957.8
414.4
5026.0
216.0
576.6
569.4
544.4
266.7
208.4
5086.0
1408.8
345.9
12155.0
1673.6
522.2
1156.0
298.3
4465.8
2068.5
933.1
800.0
262.3
796.7
2208.5
188.7
472.1
257.4
2073.3
322.8
267.3
491.2
62,152.0
 Note: -- denotes  less than $0.1 million.
Note:  — individual figure may not add  to totals  due  to rounding.
                                        39

-------
recent research findings to accurately estimate the air pollution damage cost
stems from our ignorance about the populations at risk to air pollution. So
far, few attempts have been made to identify who suffers, to what extent, from
which sources, and in what regions. At this moment, updating and expanding the
available crude estimates which are generally restricted to certain regions are
urgently needed.
                                   40

-------
                                 SECTION III

                         MORBIDITY AND AIR POLLUTION
PROBLEMS AND OBJECTIVES

     A great number of epidemiological studies have suggested that there is a
significant relationship between various morbidity rates and air pollution.
Even in the early 17th century it was quite generally suspected that sulfur
dioxide in coal smoke was responsible for the high morbidity and mortality
associated with the notorious smoke disasters such as those that later occurred
in Belgium's Meuse Valley in 1930, in Donora, Pennsylvania in 1948 and in London
in 1952.

     The relationship between air pollution and health can be acute response--
dramatic increases in air pollution concentration exert an immediate adverse
effect on human health. However, it is well known that air pollutants contin-
uously react dynamically in the environment. The effect of pollutants on health
should also be examined over an extended period. Lave and Seskin (1973b, p. 17)
remarked that "a long, or chronic exposure to low concentrations might be just
as harmful to health as a short, or episodic exposure to high concentrations."I/

     The diseases which are known to be related to air pollution include the
following: bronchitis and emphysema; pneumonia, tuberculosis and asthma; total
respiratory diseases; lung cancer; nonrespiratory-tract cancers; and cardiovas-
cular diseases. A review of the existing literature on the diseases attributable
to air pollution is given in the following paragraphs for better understanding
of the problems under study.

Bronchitis and Emphysema

     Six specific bronchitis rates have been found by Stocks  (1959) to be cor-
related with a deposit index and smoke. This result was corroborated by
Ashley (1969) who found a positive correlation between deaths due to bronchitis
and sulfur dioxide and smoke. However, a contrary result was obtained by Burgess
and Shaddick (1959) who failed to reveal a significant relationship between bron-
chitis death and air pollution.

     Holland and Reid (1965) and Reid (1968) found that the health status of
postmen was inversely affected by fog and air pollution. Cornwall and Raffle
(1961) found a positive correlation between sickness absence and fog.
I/ A comprehensive literature review on the effect of air pollution on human
     health was provided, for example, by Lave and Seskin (1975).
                                     41

-------
     Higgins (1966) found lower peak expiratory flow rate in urban areas than
in rural areas. Hammond (1967) confirmed that heavy smokers in cities suffered
a much higher morbidity rate than those in the rural areas. Ishikawa et al. (1969)
found that the incidence and severity of emphysema was higher in St. Louis than
in Winnipeg, which had a lower pollution level than St. Louis.

     Petrilli et al. (1966) also discovered that the incidence of bronchitis
was  signficantly correlated with pollution. Toyama (1964) and Yoshida et al.
 (1966) confirmed the positive relationship between bronchitis and pollution.

Pneumonia, Tuberculosis, and Asthma

     Stocks  (1960) discovered a high correlation between smoke index and pneu-
monia mortality. Mills  (1943) found substantial correlation between pneumonia
mortality and pollution levels. Significant sample correlations for pneumonia
mortality and fuel consumption, and for tuberculosis mortality and fuel con-
 sumption were reported by Daly (1969).

     Sultz et al.  (1969) found a significant relation between air pollution
 levels and the incidence of asthma and eczema among boys under 5 years of age.
Yoshida et al. (1969) found that bronchial asthma among Japanese residents was
proportional  to the  sulfur dioxide levels.

Total Respiratory Disease

     Skalpe  (1964) found that pulp mill workers under 50 years of age exposed
to sulfur dioxide suffered from a significantly lower maximal expiratory flow
rate. Speizer and Ferris (1963) reported more prevalent chronic respiratory
disease in those working in the tunnel for more than 10 years than for those
with shorter employment periods.

     Winkelstein and Kantor (1969) discussed a positive reaction between cough
with phlegm and suspended particulates. However, the association was not found
between cough and sulfur dioxide. Rosenbaum (1961) found that British servicemen
from an industrial region exhibited a greater liability to respiratory diseases.

     Feidbert et al. (1967) discovered that total respiratory disease mortality
in Nashville was directly related to the degree of sulfation and soiling. Lepper
et al.  (1969) found that total respiratory deaths were related to the levels of
sulfur dioxide across areas of Chicago with various socioeconomic variables being
controlled.

Lung Cancer

     Dean (1966)  discovered that lung cancer death rates are higher in urban
areas than in rural areas.  Gardner et al. (1969) found the lung cancer death
rate in males is  positively related to air pollution when other social and en-
vironmental  factors are controlled. Somewhat inconsistent results regarding the
relationship between sulfur dioxide and lung cancer were obtained by Buck and
Brown (1964).
                                     42

-------
     Stocks (1966) discovered a significant correlation between lung cancer and
air pollution. Clemmesen and Nielsen (1951) reported the lung cancer morbidity
for males in Copenhagen was about four times greater than in rural areas in
Denmark.

     Manos and Fisher (1959) and Griswold et al. (1955) found that urban lung
cancer rates are significantly higher than rates in rural or nonmetropolitan
areas. Greenburg et al. (1967) reported correlation between lung cancer and air
pollution. However, negative results were obtained by Zeidberg et al. (1967)
and Winkelstein et al.  (1967).

Nonrespiratory-Tract  Cancers and Cardiovascular Disease

     Winkelstein and Kantor (1969) found that stomach cancer mortality was
twice as high in high pollution areas as in low pollution areas.

     Levin et al.  (1960) discovered that the incidence rate for both sexes for
each of 16 categories of cancer was higher in urban than in rural areas. Contrary
results have also been reported by Greenburg et al. (1967a), among others.

     Higher incidence rates of cardiovascular diseases in urban than in rural
areas were reported by Enterline et al. (1960). Zeidberg et al. found heart dis-
ease rates were correlated with air pollutants in Nashville. Manos and Fisher
(1959) also found positive relationships between heart disease and air pollution.

     The results of many of the epidemiological studies discussed above indicate
that incidence rates  of various kinds of diseases are generally much higher in
the urban areas than  in the rural areas. Many of these disparities in morbidity
rates between urban and rural areas can be attributed to air pollution. The ratio
of urban incidence to rural incidence of morbidity has been termed the urban
factor. This urban factor has been used for estimating health damage due to air
pollution. The rationale for the urban factor technique is that if air pollution
levels in the urban areas could be reduced to the rural levels, then the dif-
ferences between the urban and rural morbidity rates adjusted for smoking, age,
sex, and race should be eliminated.

     The crucial question is what portion of this urban factor is attributable
to air pollution. In a pioneering study of air pollution damage, Ridker (1965)
assumed that 100 percent of the urban factor is attributable to air pollution
and derived a damage value of $2 billion for 1958. Williams and Justus  (1974)
assumed that a minimum of 10 percent and a maximum of 50 percent of the urban
factor is due to air pollution and estimated that the total 1970 nationwide
health cost due to air pollution was between $62 million and $311 million. The
                                     43

-------
figures are much lower than the estimate of $6.22 billion for respiratory dis-
ease in the United States.JL'  The damage estimates derived by using the urban
factor of health deterioration due to air pollution are apparently subject to
a large margin of error because of the difficult assignment problem of the ur-
ban factor. The urban factor method is also replete with several other concep-
tual and practical difficulties. For example, the distinction between urban
and rural pollution levels is hard to define because there exists a continuous
scale of pollution intensity instead of a simple dichotomy between urban and
rural pollution levels. Thus, after all, the question as to what percentage of
this urban factor is actually accounted for by air pollution remains largely
unresolved.

     A recent  study performed by Shy et al. (1974) on the Community Health and
Environmental  Surveillance System (CHESS) examined the adverse effects of air
pollution on acute and chronic respiratory disease. The methodological proce-
dures employed in the CHESS  study involve statistical analysis with varying pol-
lutant gradients and concentration levels. Each CHESS set which consists of a
group of communities selected to represent an exposure gradient for designated
pollutants  generally includes High, Intermediate, and Low exposure communities.
The community  selection is subject to the following criteria: The communities
have  similar climates and are made up of a predominantly white, middle-class
population with as much homogeneity in socioeconomic and other demographic fac-
tors  as  possible. The research findings point to a clear trend toward excess
illness  in  the High exposure community.

      Since  the national and  regional annual damage cost figures greatly assist
policymakers in determining  optimal pollution control strategies, the effort
to derive  a  set of internally consistent and relatively accurate damage estimates
is warranted.  The primary purpose of this study is to derive such damage esti-
mates. Specifically, physical and economic damage functions will be derived re-
lating morbidity rate and morbidity costs to air pollution, socioeconomic, demo-
graphic, and climatological variables. The morbidity damage costs will be esti-
mated for the  40 SMSA's included in the preceding section on mortality and air
pollution.

     The balance of this section, which represents an exploratory effort to
estimate morbidity dose-response functions for adult morbidity damage costs
for the  40 SMSA's selected in our study, discusses the following subjects:
_!/ For a detailed discussion on some of the problems in using the urban factor
     for calculating health costs, see J. R. William and C. F. Justus, "Evalu-
     ation of Nationwide Health Costs of Air Pollution and Cigarette Smoking,"
     Journal of the Air Pollution Control Association (November 1974), pp. 1063-
     1066. The figure $6.22 billion was derived by William and Justus by ad-
     justing Ridker's value of $2 billion for 1958.
                                     44

-------
Environmental Damage Functions: Some Theoretical Underpinnings,  Adult Morbidity
and Air Pollution, Adult Morbidity Damage and Sulfur Dioxide, Economic Damages
and Economic Damage Functions, and Adult Morbidity Damages and Total Suspended
Particulates.
ENVIRONMENTAL DAMAGE FUNCTIONS: SOME THEORETICAL UNDERPINNINGS

     An economic damage function, which is usually derived on the basis of a
physical damage function, is defined, for example, by Maler (1974) as the com-
pensating variation or the amount the individual (or society) should be com-
pensated so as to maintain his initial preference level in the presence of a
deterioration in the environment. This definition is clearly applicable to any
situations in which the effect of environmental degradation enters directly into
the individual's utility function.

     We assume that the consumer's preferences can be represented by a twice
differentiable, concave utility function, defined on Rm + n
                       U = U(C,H(A))                                  (III-l)
where  C  is an m-vector representing  m  private commodities and services,  with
positive components indicating consumption, and negative ones, supply of labor
services. H  denotes the health status, which is influenced by air pollution;
A  is an n-vector characterizing environmental quality, which is exogeneously
given to the community. H  can be viewed as the dose-response function.

     Each individual wants to maximize (III-l) subject to the following budget
constraint:
                               PC
where  P  is the price vector associated with  C, and  Y  is the individual's
income.

     The economic damage function as registered in the compensation variations
due to changes in the individual's health condition because of changes in  A
can be derived by minimizing the total expenditures subject to a given utility
level, say  U.
                                     45

-------
     The familiar first order necessary conditions are


                       a U.  = P., i = l,...,m                          (III-3)


where  &  is the Langrangean multiplier.

     Solving (III-3) yields the following compensated demand functions


                       C = C(P,H(A);U)                                 (III-4)
     The minimum income required to maintain the same utility level when one
 or  several components in  A  changes is denoted by±'
                    I = I(P,A; U)                                      (IH-5)
      Assuming  the individual always exhausts his budget, the economic damage
 function  is  simply the difference between (III-5) and the individual's initial
 income,   Y,
                    D = I - Y = (P,H(A); U)                            (III-6)
      Regional  economic damages and the economic damage function can be opera-
 tionally  expressed as:
          MBC j = MB (A) x PC j + HSj(A) x HC j + DUj(A) x DC j x POP j      (IH-7)
                   MBC = f(H(E,D,S,W,A;e),P)                           (III-8)
where  MBC.   denotes total morbidity cost in the jth urban area,  MB is the
morbidity  rate,  HS  hospitalization rate,  DU  drug use rate, PC  physician
cost,  HC  hospitalization cost,  DC  drug cost, and POP  is the population in
the  area.  The notations in equation (III-8) were defined in Section II. That  is,
_!/ Equation 5 was labeled by Maler as the expenditure function. The analytical
     properties of such expenditure functions are delineated in K. G. Maler,
     Studies in Environmental Economics, in press.

                                    46

-------
E  for the economic factors,  D  the demographic factors,  S  the social fac-
tors,  W  climatological factors,  A  air pollution,  e  error term, and  P
the commodity prices.
ADULT MORBIDITY AND AIR POLLUTION

     Physical damage functions on adult morbidity are derived by the classical
least-squares linear regression technique  and the random sampling, simulation
technique. The few aggregated dose-response observations obtained from the CHESS
study  (1974) form the data base for the regression analysis in this study. The
dose-response observation reported in the CHESS study related morbidity prevalence
rate to particulates and sulfur dioxide in 1971 for four regions:  Salt Lake
Basin, Chicago, Rocky Mountain, and New York.JV

     The CHESS communities in the Salt Lake Basin are located near the major
copper smelter, and the local meteorological pattern provides an area gradient
of  exposure to sulfur oxides. The selected communities include Magna, Kearns,
Salt Lake City, and Ogden. Magna was designated the high exposure area because
it  had a high sulfur dioxide level due to its proximity to the smelter. Kearns,
Salt Lake City, and Ogden were designated as Intermediate II, Intermediate I
and Low exposure areas. These three cities had a descending exposure gradient
to  sulfur oxides.

     The CHESS communities in the Chicago area include urban core, suburban
areas and the relatively clean area, designated as High I, High II and Low pol-
lution exposure areas for 1969-1970. The five communities selected in the Rocky
Mountain area for the CHESS study are Anacenda, Kellogg, East Helena, Bozeman
and.Helena, designated, respectively, as High I, High II, Low III, Low I and
Low II exposure areas. For the New York City area, Riverhead, Long Island was
chosen as a Low exposure community, the Howard Beach section of Queens as the
Intermediate exposure community, and the Westchester section of the Bronx as
a High exposure community.

     The dose-response observations collected from the 15 CHESS communities
in  the four selected regions are summarized in Table III-I. The adjusted
bronchitis prevalence rates expressed in percentages for the selected exposure
areas are presented in Column 3 of the table. The annual average sulfur dioxide
and total suspended particulates levels for the same set of communities are
presented respectively in Column 4 and Column 5. It should be noted that the
bronchitis prevalence rates presented in the CHESS report for Utah, Rocky Moun-
tain and New York were adjusted for smoking status (e.g., nonsmoker, ex-smoker
and smoker) and sex (e.g., mother and father), while the rates for Chicago were
adjusted for education level, race and smoking status.


I/  For a general description about the EPA's CHESS Program, see Shy and Finkles
     (1973).


                                     47

-------
                              TABLE III-l. MORBIDITY DOSE  - RESPONSE OBSERVATIONS
oo

Area
Salt Lake Basin



Chicago


Rocky Mountain




New York


Adjusted Bronchitis
Community Prevalence Rate (%)
Low
Intermediate I
Intermediate II
High
Low
High I
High II
Low I
Low II
Low III
High I
High II
Low
Intermediate I
Intermediate II
6.71
6.92
8.54
10.77
25.97
25.30
21.22
1.78
5.10
4.88
4.23
3.98
9.17
16.49
13.93
Pollution Levels (ug/m3)
S02 (1971)
8
15
22
62
19
96
217
10
26
67
177
374
23
51
51
TSP (1971)
78
81
45
66
71
155
103
50
45
115
65
102
34
63
86

-------
     The adjusted bronchitis prevalence rates were regressed on the two pollut-
ants to derive the dose-response functions for Salt Lake Basin, Chicago, Rocky
Mountain and New York separately by the least-squares technique. The regression
results are summarized in Table III-2. The regression fit between morbidity and
SO  for New York, Chicago and Salt Lake Basin is fairly good, with R  having
the values of 0.50, 0.88 and 0.94, respectively. Furthermore, SO  is significant
at the 1 percent level for the New York and Salt Lake Basin regression equations.
For total suspended particulates, good regression fit was obtained for Chicago
and New York. However, TSP is consistently insignificant in expressing the vari-
ations in morbidity- These regression equations, coupled with the mean values
and standard deviations of the pollutants and the morbidity prevalence rates
presented in Table III-3, were used for a random sampling and simulation study
to generate a "national" dose-response function which can be used for estimating
morbidity damage costs in the various SMSA's.
ADULT MORBIDITY DAMAGES  AND  SULFUR  DIOXIDE

     Epidemiological  studies have demonstrated that deterioration in air quality
results  in  increased  consumption of medical  services and, hence, in economic
loss to  the pollution victims.  To estimate such damage  loss for the 40 SMSA's
and to estimate an average economic damage function on  adult morbidity, a ran-
dom sampling technique for deriving a  "representative"  dose-response function
was employed.

Random Sampling Simulation Study and the Physical Damage Function

     "Simulation"  is  the technique  of  setting up a stochastic model of a real
situ.ation  so that  sampling experiments can be performed upon the model (Harling,
1958). Simulation  study  differs from the classical sampling experiment in that
the former  involves the  construction of an abstract model, while the latter in-
volves direct  experiment with the new  data.  The term "simulation" is often used
interchangeably with  the term "Monte Carlo"  technique.

     The Monte Carlo  technique, which  was employed to generate the "average"
nonlinear  dose-response  damage function vis-a-vis existing time series and cross-
section  studies, involves the study of probability models. As described by
Dienemann  (1966) the  Monte Carlo technique can be defined as follows:

           Assume a system planner can  describe each parameter with
           a probability  distribution.  This distribution is then treated
           as a theoretical population  from which random samples are
           obtained. The  method of taking such samples,  as well as
           problems which rely on these sampling techniques, are often
           referred to as Monte Carlo methods.
                                     49

-------
              TABLE III-2. ADULT MORBIDITY LINEAR  DAMAGE  FUNCTIONS
 I.   S02
      (1)  Rocky Mountain
                            MB    (7»)  = 3.84 + 0.001  S02           R2  =  0.016
                                       (0.94)*(0.005)
      (2)  Chicago

                            MB    (7o)  =
                                        (2.49)* (0.023)
MB    (7o)  =  22.14 + 0.018 S02          R2 =  0.50
      (3)  New York
           Salt Lake Basin
                           MB    (7o) = 4.2 + 0.21 S02             R2 =  0.88
                                       (3.46) (0.083*
                             MB   (7=) =  6.22 + 0.075 S02           R2 = 0.94
                                      (0.46)*  (0.013)*
II.  TSP

      (1)  Rocky Mountain
      (2)   Chicago
      (3)  New York
     (4)  Salt Lake Basin
                            MB    (7o) = 2.94 + 0.014 TSP           R2 = 0.109
                                      (1.84)  (0.023)
                            MB   (70) = 18.42 +  0.05  TSP           R2 = 0.74
                                       (3.52)*  (0.03)
                            MB   (7o) = 7.19 + 0.098 TSP            R2 = 0.47
                                      (6.66) (0.10)
                           MB    (7=) = 11.97 - 0.05 TSP            R2  = 0.23
                                       (4.90)* (0.07)
                                     50

-------
TABLE III-3. MEAN VALUES AND STANDARD DEVIATIONS OF THE VARIABLES
	Mean  Value  (X)    Standard Deviation (S)

Utah
  Prevalence Rate                   8.2                  1.9
  S02                               26.8                 24.2
  TSP                               67.5                 16.3

Chicago
  Prevalence Rate                   24.2                  2.6
  S02                              110.6                 99.8
  TSP                              109.6                 42.4

Rocky Mountain
  Prevalence Rate                   4.0                  1.3
  S02                              132.8                150.8
  TSP                               75.4                 31.4

New York
  Prevalence Rate                   13.2                  3.7
  S02                               41.2                 16.2
  TSP                               61.0                 26.1
     A random sampling experiment was performed on the four sample regions in
this study for deriving an "average" morbidity dose-response function. These
four sample regions were constructed in the two dimensional space with the aid
of the four regional dose-response functions shown in Part I of Table III-2,
coupled with the data on the mean values and the standard deviations of the
dependent and independent variable (see Table III-3). The four regional blocks
are shown in Figure III-2, the vertical axis represents the morbidity rate ex-
pressed in number of incidences per 100 residents, and the horizontal axis de-
notes SO  pollutant concentrations level expressed in fig/m . For each sample
block, the height of the block is the difference between the morbidity rate com-
puted from the dose-response function with the coefficient of SO  in the function
taking the value of (b + s) and (b - s), where  b  is the coefficient of SO  and
s  the associated standard error. The width of the block is, however, measured
by the mean value of_ SO  plus and_minus one standard deviation of the mean,
i.e., (X + S) and (X - S) where  X  denotes the mean value of SO  and  S  the
associated standard deviation.

     Thus, the four sample blocks shown in Figure III-2 were defined on the
basis of the four prior studies regarding the morbidity effect of SO  in the
four different regions. The construction of these four blocks permits us to

                                     51

-------
Oi
to
>

5
en
0£.

i
                                 Figure III-1.  Sample observation from four morbidity studies

                                                with respect to SO .

-------
perform random sampling experiments.  A random sample of 800 observations
with 200 chosen  from each block was obtained.  To eliminate possible bias
in the probability  of being randomly  selected resulting from the overlapping
of the blocks, another random  sampling was performed on the basis that two
sorting schemes  yield better results  than one sorting procedure.  A smaller
sample of 81  observations, i.e.,  10 percent  of 800, was chosen.  These 81
observations  were used to develop a nonlinear "average" dose-response function
specified alternatively as follows:


                    MB = C + EXP  (a-b/SO  )                 (III-9)
or
                    MB  - C = EXP  (a-b/S02)

                     (MB-C) = a-b/S02                        (111-10)
      As  in  the mortality  study  reported  in Section II, the physical dose-
 response function  in  this morbidity  study is again expressed as an exponential
 function which is  consistent with _a  priori judgment and empirical results of
 medical  experts  regarding plausible  human dose-responses to changes in pollution
 levels.  The geometrical counterpart  of this exponential relation is a long flat
 "S"  curve,  implying that  while  the air pollutant contributes to the morbidity
 incidence rate,  the damaging effect  is not proportional. In the presence of in-
 creased  SO   level, the morbidity rate initially increases at an increasing rate
 and  continues to increase,  but  at a  decreasing rate after a certain inflection
 level.

      Unlike the  mortality study in which the intercept term  C, conventional
 mortality,  is expressed as  a function of a number of  socioeconomic, demographic
 and  climatological variables, no such conventional morbidity function was esti-
 mated due to the lack of  a  systematic collection of morbidity data by the var-
 ious  SMSA's. Of  necessity,  the  C  term  in equation (III-9) above is assumed
 to take  the value  of  11 since 11 is  the  arithmetic mean of the morbidity rates
 calculated  from  the four  regional dose-response functions with the explanatory
 variable, SO being at the  threshold of  25 )J,g/m  for  the sake of consistency
 with  the earlier mortality  study.

      In  estimating equation (III-9), the classical least-squares technique was
 applied.  Since (MB -  .11)  may be negative, and the logarithm of a negative number
 is undefinable,  (MB - 11) was therefore  squared prior to its logarithm transfor-
 mation.  The resultant regression equation was then adjusted by dividing the co-
 efficients  by 2. A detailed discussion on the rationale of this procedure was
 presented in Section  II.
                                    53

-------
     The regression  results  for  equation  (III-9)  look as follows:
                   MB = 11 + EXP(0.65  - 4.96/SO )
                                (0.11)* (1.99)*           (111-11)

                                       R2  = 0.072
     The figures below the coefficients are standard errors,  with * indicating
that the coefficient  of SO   is significant at the 1 percent level. However,
the pollution variable SO  explains only about 7 percent of the variations in
the residual morbidity rate, i.e., (MB - 11).

     A linear morbidity equation was also fitted, with the regression result
shown as follows:
                        MB = 12.06 - 0.01 SO-
                            (1.28)* (0.01)                  (111-12)

                                        R2 = 0.011
     Comparing the result of equation (III-ll) to that of (111-12) the exponen-
 tial dose-response function is apparently a better fit than the linear one be-
 cause the former showed an explanatory power seven times larger than the latter
 equation. Furthermore, the coefficient of SO   in the exponential equation is
 statistically significant, whereas it is insignificant and has a wrong sign in
 the linear equation. Thus, the empirical results suggest that the nonlinearity
 in the dose-response relation is more consistent with 
-------
          c. These 81 randomly selected observations were fitted to an exponen-
tial reciprocal equation to derive an "average" dose-response function for the
four regions.

     Like the mortality dose-response function, the nonlinear morbidity dose-
response function has a number of distinguishing features:  (1) the nonlinear
dose-response function is not only more in accord with _a priori judgment re-
garding human morbidity response to pollution doses, but also it is more amen-
able to being adjusted with whatever the assumed threshold level of SO  is in
estimating the economic damages than the linear functions;  and (2) for the pur-
pose of predicting and estimating the marginal morbidity damages due to SO ,
the nonlinear equation has  shown better fit and hence, will yield more accurate
prediction over the linear one.

Economic Damages and Economic Damage Functions

     Given the preceding nonlinear physical damage function, the economic costs
of diseases related to air pollution can be estimated by transforming the addi-
tional morbidity rate into monetary units. Economic damages of morbidity, as
discussed earlier, represent the amount that an individual or a society is will-
ing to spend so as to maintain the previous preference level in the presence
of the deterioration of air quality,.

     Morbidity damages generally are comprised of two parts: direct and indirect
costs of illness. Included in the direct costs of illnesses are the expenditures
for prevention, detection, treatment, rehabilitation, research, training, and
capital investment in medical facilities. Indirect costs of illness include the
loss of output to the economy because of disability and the imputed costs such
as opportunities foregone. A comprehensive framework for calculating the direct
and indirect economic costs of illness and disability has been developed by Rice
(1966) and others.

     Both direct and indirect morbidity costs were estimated in the present study.
Direct morbidity costs were computed by summing up the costs of physician visits,
hospitalization costs, and drug costs. According to a recent study by Jaksch
(1975), the average cost per physician visit for all ages combined in 1970 was
$14, and the average cost of a hospital day for all ages combined was $82. To
estimate total morbidity costs, further information is needed on the average
number of physician visits and the average length of hospital stay per pollution-
related disease incidence. A number of assumptions were made to obtain conserva-
tive morbidity damage estimates, as follows: (1) each pollution-related morbidity
incidence results in one visit to consult a physician; (2) 1 of 8.3 physician
visits, i.e., 12 percent, results in hospitalization; (3) drug costs run about
50 percent of the physician costs; (4) if hospitalization is required, each
patient stays 1 day in the hospital for treatment.I/

I/ Various information on national data about the number of visits to doctors
     and the hospital days stayed per treatment can be obtained from Public
     Health Service (1973).
                                     55

-------
     The conservative nature of both assumptions (1) and (4) leads to under-
estimations of the morbidity costs. The bias could be partially removed by as-
suming a greater number of physician visits and a longer hospital stay, however.
The  estimates presented in this study can be regarded as low estimates for mor-
bidity costs. Assumption  (2) is based on the calculated proportion of physician
visits resulting in hospital discharge for four categories of diseases related
to pollution  (Jaksch, 1975). The figure 12 percent is the average of such pro-
portions of physician visits in the four disease categories. Assumption  (3)
is,  however,  based on a ratio of total drug costs to total physician costs at-
tributable to the use of  oxidation catalyst as estimated by (Jaksch, 1975),
i.e., 11.4/23.2 = 0.5.

     The direct morbidity costs attributable to SO  were estimated with the aid
of  the following formulas:


        PCS02 = $14 x EXP [0.65 - 4.96/(S02 -25)] x POP x NPV        (111-13)


        HCS02 = $82 x EXP [0.65 - 4.96/(SO  -25)] x 0.12 x POP x HSD (111-14)


        DCS02 = 0.5 x PCS02                                          (111-15)


  where            PCSO  = physician cost atttibutable to S0_.

                   HCSO  = hospitalization cost attributable to SO .

                   DCS02 = drug cost attributable to SO .

                   POP   = SMSA population.

                   NPV   = number of physician visits per incidence
                           = 1 (by assumption (1))

                   HSD   = number of hospital stay days = 1 (by assumption (4))


     Recall the physical  dose-response function for SO  as expressed in equa-
tion (III-ll) which has an intercept value of 11. If the exponential term in
equations (111-13) and (111-14) is replaced by the value of the intercept of
the dose-response function, then we can derive another set of cost estimates
for morbidity in the absence of SO .
                                  2
     Another dimension of morbidity health costs is the indirect component re-
garding the changes in earnings and leisure opportunities because of disability
and debility.  A shortcut to estimate the indirect morbidity cost attributable
to pollution was found by applying to the direct morbidity cost a multiplier

                                    56

-------
of 2.4, which is the ratio of the best estimates of total indirect net costs and
the total direct costs of morbidity (Jaksch, 1975). Hence, the following formula
was used for estimating the indirect morbidity costs attributable to SO :


          IMBCSO  = 2.4 x  (PCSO  + HCSO  + DGSO )         (111-16)


     The estimated morbidity costs for the 40 SMSA's with an SO  level equal
to or greater than 25 p,g/m , i.e., the threshold level, are presented in Table
III-4. Columns 1, 2, and 3 present, respectively, the physician costs, hospital
costs and drug costs attributable to SO  . Indirect morbidity costs due to SO
are presented in Column 4. It should be noted that the figures in Column 4 are
2.4times the sum of Columns 1, 2, and 3. Total morbidity costs due to SO  cal-
culated by summing Columns 1, 2, 3 and 4 are presented in Column 5, and per
capita total morbidity costs are in Column 6. Total morbidity costs in the
absence of SO , direct and indirect, are presented in Column 7. The cost figures
in this column were estimated with the aid of equations (111-13) to (111-16)
with the modification of replacing the exponential term by the intercept term
of the dose-response function. Finally, Column 8 presents the ratio of total
morbidity cost attributable to SO  to total morbidity cost with and without SO ,
that is, Column 8 = Column 5/(Column 5 + Column 7). The extent of pollution
damage to human health is partially reflected by the magnitude of this ratio.

     Upon examination of the  low estimates of morbidity costs in Table III-4,
it is readily revealed that the annual morbidity costs due to SO  range from
a minimum value of less than  $1,000 in Cincinnati, Dayton, Evansville and
Johnstown to a maximum of  $22 million in New York City. Per capita morbidity
costs attributable to SO  in  1970 vary between cost of negligible magnitude to
$1.96 in New York City- Total morbidity damages attributable to SO  over the
40 SMSA's were at least $99 million in 1970.

     It should be stressed that the cost figures presented in the table repre-
sent low estimates for the morbidity damages due to the two conservative as-
sumptions made for the calculation of the costs. If five instead of one is the
average number of doctor visits, and the average number of days in the hospital
is 5 days rather than 1 day per pollution-related disease incident, then by
assuming the same costs incurred per visit to consult doctors and per hospital
day for treatment, the cost figures in Columns 1 to 7 should be revised accord-
ingly. In other words, the direct and indirect morbidity costs and the per cap-
ita total morbidity cost attributable to SO  should be five times as  large as
the low cost estimates calculated for the SMSA's.

     An "average" economic damage function was derived for the purpose of predict-
ing marginal and average changes in the morbidity costs in response to changes
in the pollution or in other variables. The morbidity cost in the presence of
SO , which is the sum of morbidity costs due to SO  and morbidity cost in the
absence of pollution, was regressed on a host of socioeconomic, demographic and
climatological variables. The stepwise regression results are shown as follows:
                                     57

-------
                          TABLE III-4. MORBIDITY COSTS WITH S02 BY SMSA's, 1970
Indirect
Morbidity Costs Total Morbidity
Direct Morbidity Costs Due Due to S02 Morbidity Cost Due Cost Ratio
to SO? (in $103) (in S103) to SO? Without S02 (in S1Q3) (8) = (5 U( (5K(71 )
SMSA
1 AKR
2 ALL
3 BAL
4 BOS
5 BRI
6 CAN
7 CHA
8 CHI
9 CIN
10 CLE
11 DAY
12 DET
13 EVA
14 GAR
15 HAR
16 JER
17 JOH
18 LAW
19 LOS
20 MIN
21 NHA
22 NYO
23 NEW
24 NOR
25 PAT
26 PEO
27 PHI
28 PTB
29 FOR
30 PRO
31 REA
32 ROC
33 STL
34 SCR
35 SPR
36 TRE
37 WAS
38 WOR
39 YOR
40 YOU
Total
PCS02
(1)
151
125
468
323
75
37
5
1775
--
487
--
769
--
146
152
148
	
52
1149
332
69
3021
329
1
70
1
1188
551
2
218
29
117
455
23
131
40
612
40
39
53
13,183
HCS02
(2)
106
88
329
227
53
26
4
1248
--
343
--
541
--
103
107
104
--
36
808
233
48
2123
231
1
49
--
835
388
1
153
21
82
320
16
92
28
430
28
27
37
9,266
DCS02
(3)
75
62
234
162
38
19
3
888
--
244
--
385
--
73
76
74
--
26
575
166
34
1511
165
1
35
—
594
276
1
109
15
58
228
12
66
20
306
20
19
27
6,597
1MB C SO 2
(4)
796
660
2474
1708
397
196
27
9386
--
2577
--
4066
--
773
806
782
--
274
6075
1756
362
15900
1741
7
369
3
6280
2916
10
1150
156
616
2407
123
694
212
3238
214
204
282
69,637
Total
(51 (in $103)
1127
935
3505
2420
563
277
39
13200
--
3651
--
5760
--
1095
1142
1108
--
389
8607
2487
513
22600
2467
10
522
5
8897
4131
14
1629
221
873
3410
174
982
301
4587
303
290
399
98,633
Per Capita
(6) ($)
1.66
1.72
1.69
0.88
1.44
0.74
0.17
1.91
--
1.77
--
1.37
--
1.73
1.72
1.82
--
1.67
1.22
1.37
1.44
1.96
1.33
0.01
0.38
0.01
1.85
1.72
0.01
1.78
0.74
0.99
1.44
0.74
1.85
0.99
1.60
0.88
0.88
0.74

(7)
7834
6269
23883
31763
4500
4293
2647
80240
15973
23809
9807
48280
2685
7305
7656
7027
3031
2681
80920
20919
4120
133280
21414
7850
15673
3944
55420
27696
11639
10530
3419
10181
27255
2700
6112
3506
33001
3975
3801
6182
783,202
(8)
0.13
0.13
0.13
0.07
0.11
0.06
0.01
0.14
..
0.13
—
0.11
--
0.13
0.13
0.14
--
0.13
0.10
0.11
0.11
0.14
0.10

0.03

0.14
0.13

0.13
0.06
0.08
0.11
0.06
0. 14
0.08
0.12
0.07
0.07
0. 06

Note:  -- denotes less than $1,000.
Note: —  individual figure may not add  to totals  due  to rounding.
                                          58

-------
    TMBCSO  = 52.4+ 0.60 SO   -  135.0 PWPO + 1.4 SUN + 1.3 RHM -
              (80.3) (0.09)*    (67.9)*      (0.7)**   (0.6)*

          0.3 DTS + 0.09 PCOL +  34.4 AGE                      (111-17)
          (0.2)     (0.10)       (310.4)

                                      R2 = 0.73
where  TMBCSO   denotes the morbidity cost in the presence of SO ,  and all
seven explanatory variables are the same as those defined previously in Sec-
tion II. The values below the coefficients are standard errors, with * and **
to indicate that the coefficients are significant at the 1 and 5 percent level,
respectively. All coefficients and the corresponding standard errors are re-
duced by a factor of 10 . It should be pointed out that the primary use of equa-
tion (111-17) is only for prediction. "Wrong" signs as well as other statistical
questions do not constitute a great problem if they are understood and accounted
for.

     In predicting and estimating the responsiveness of morbidity damages to
changes in any one of the explanatory variables, the partial elasticity of the
morbidity cost with respect to the variable of interest merits some discussion.
Suppose a policymaker would like to estimate what the marginal changes will be
in the morbidity cost if the pollution level of SO  in the SMSA' s is lowered,
on the average, by, say, 1 percent. In order to aid this policymaker to make
the prediction, the partial elasticity of the morbidity cost with response to
SO  (K,__ ,,„ ) is calculated as follows:
  2.   MCB,bU


                 n  = °'6 x 10  x  (47.95/22.7 x 106) = 1.27    (111-18)
where  (0.6 x 10  ) is the coefficient of SO  in the economic damage function,
and 47.95 and  (22.7 x 10°) are, respectively, the mean level of SO  and total
morbidity cost.

     In view of  the SO  partial elasticity value of 1.27, the estimated morbidity
cost would decrease by 1.27 percent, for every 1 percent reduction in SO  level,
other things being equal. Stated differently, if the air pollution control program
lowers the SO  level by 4.7 |ig/m  from 47.9 to 43.2 [ig/m^ (10 percent reduction),
adult morbidity  costs on the average would decrease by $2.72 million, from $22.7
million to $19.98 million. In a like manner, the coefficients of other variables
in equation (111-17) can be used to compute the partial elasticities associated
with the variables and can be analogously interpreted as conditional marginal
inpact when others are held constant.
                                     59

-------
ADULT MORBIDITY DAMAGES AND TOTAL SUSPENDED PARTICULATES

     Total suspended particulates are directly harmful to human health. The
poisonous substances or hydrocarbons contained in the particulates may cause
cancer. Other particulates multiply the potential harm of irritant gases. For
example, the interaction of sulfur dioxide gas with particulate matter will
penetrate deep into the lungs and cause much greater harm. Some particulates
expedite chemical reactions in the atmosphere to form harmful substances.

     Arsenic, a well-known poison, may also cause cancer. Asbestos fiber is re-
sponsible for chronic  lung disease. Beryllium has produced malignant tumors in
monkeys. Cadmium, a respiratory poison, induces high blood pressure and heart
disease. Lead, a cumulative poison, impairs the functioning of the nervous  sys-
tem in  adults.

     Adult morbidity costs attributable to TSP were estimated by invoking the
 same methodology delineated above for deriving morbidity costs due to  SO  .  The
aggregate dose-response observations relating morbidity rate to TSP are presented
 in Table III-l, page 48. The  observations, obtained from the report on the  CHESS
 study,  were  used to estimate  four separate regional, dose-response functions
 for the four study regions, i.e., Salt Lake Basin, Chicago, Rocky Mountain  and
New York. The  regression results for the regional dose-response relations are
 shown  in the lower half of Table III-2, page 50. The mean values and standard devi-
 ations  of suspended particulates and the morbidity prevalence rates are pre-
 sented  in Table III-3, page 51.

     The random sampling and  simulation techniques delineated above were again
applied to  derive an "average" nonlinear dose-response function relating mor-
bidity  rates to suspended particulate levels. A total of 82 observations was
randomly selected in the two-round sampling experiments from the four  "blocks"
defined in  the two-dimensional morbidity and suspended particulate space as
shown in Figure III-3. Given  these 82 observations, least-squares regressions
were run and the results are  shown as follows:
                      MB = 11 + EXP  (1.75  -  (87.7/TSP))
                                    (0.22)*  (15.7)*                 (111-19)

                                                     2
                                                   R  = 0.28
     Again, the values below the coefficients are standard errors with *  to  in-
dicate that the coefficients are significant at the  1 percent  level.  It should
be noted that the intercept term 11 in equation (111-19) is the arithmetic mean
of the morbidity rates calculated from the four regional dose-response functions
with the dependent variable TSP being at the threshold level of 25
     As in the case of SO , (MB - 11) was squared prior to its logarithmic trans-
formation when the regression was run. The coefficients in equation (111-19)
                                     60

-------
  30 -
  20
£
(JU
i
 .
O
  10
                   25
                                   50
                                                  75
                                                                  100
                                                                                 125
                                                                                                 150
                                                                                                                  175
                                                                                                            TSP
                          Figure III-2. Sample observations  from four morbidity studies
                                        with respect to TSP.

-------
were obtained by dividing also the regression coefficients Log 2. The coefficient
of TSP in this nonlinear dose-response function is also statistically significant
at the 1 percent level and has a correct sign.

     The direct morbidity costs attributable to TSP were estimated with the aid
of the following formulas:


     PCTSP =  $14 x EXP  [1.75 - 87.7/(TSP - 25)] x POP x NPV        (111-20)

     HCTSP =  $82 x EXP  [1.75 - 87.7/(TSP - 25] x POP x HSD         (111-21)

     DCTSP =  0.5 x PCTSP                                           (III-22)


where  PCTSP  = physician  cost attributable to TSP.

       HCTSP  = hospitalization cost attributable to TSP.

       DCTSP  = drug  costs attributable to TSP.

       POP, NPV  and HSD  are the same as those defined in (111-13)
 and (111-14).

      Applying the  same  multiplier of  2.4 used in the case of SO  ,  the indirect
morbidity  costs  due  to  TSP  (IMBCTSP)  were computed by


      IMBCTSP  = 2.4 x (PCTSP + HCTSP + DCTSP)              (III-23)
      Morbidity  costs  for  the  40 SMSA's with a TSP  level  equal  to  or  greater
 than 25  |Jig/m3 are  tabulated in Table III-5. Physician costs, hospital  costs,
 and drug costs  attributable to TSP are presented in Columns  1  to  3,  and  indirect
 morbidity costs due to TSP in Column 4. Total and  per capita morbidity costs  at-
 tributable to TSP  are presented in Columns 5 and 6. The  ratio  of  total morbidity
 cost attributable  to  TSP  to total morbidity cost associated  with  or  without TSP
 is  given in Column 8.

      It  should  be  again noted that the cost figures presented  in  this  table,
 as  those in the  case  of SO , are low estimates  for the morbidity  damage  associ-
 ated with TSP.  If  each pollution-related incidence results in,  on the  average,
 five rather than one  visit to doctors, and the  patients,  if  admitted to  a  hos-
 pital, will stay in the hospital for 5 days instead of 1  day,  then,  by assuming
 a constant cost  for consuming medical services, the morbidity  cost estimates
 in  Columns 1 to  7  in  Table III-5 will be magnified five  times.  Consequently,
 the  total morbidity costs over the 40 SMSA's for each category (column)  will
 also increase five times.
                                     62

-------
                       TABLE III-5. MORBIDITY COSTS WITH TSP BY SMSA's,  1970
Indirect
Morbidity Costs Total Morbidity
Direct Morbidity Cost Due Due to TSP Morbidity Cost Due Cost
to TSP (in 1103) (in S103) to TSP Without TSP (in $103
SMS A
1 ARK
2 ALL
3 BAL
4 BOS
5 BRI
6 CAN
7 CHA
8 CHI
9 CIN
10 CLE
11 DAY
12 DET
13 EVA.
14 GAR
15 HAR
16 JER
17 JOH
18 LAW
19 LOS
20 MIN
21 NHA
22 NYO
23 NEW
24 NOR
25 PAT
26 PEO
27 PHI
28 PTB
29 POR
30 PRO
31 REA
32 ROC
33 STL
34 SCR
35 SPR
36 TRE
37 WAS
38 WOR
39 YOR
40 YOU
PC TSP
(1)
111
106
813
771
20
97
62
2862
378
1010
256
1705
32
170
89
108
69
21
2208
262
23
2663
669
202
65
53
742
872
193
136
92
134
756
110
45
36
598
43
62
154
HC TSP
(2)
78
75
571
542
14
68
43
2012
266
710
180
1199
23
120
63
76
48
15
1552
184
16
1872
470
142
45
37
521
613
136
96
65
130
532
78
32
26
420
30
43
108
DC TSP
(3)
56
53
406
386
10
49
31
1431
189
505
128
853
16
85
45
54
34
10
1104
131
12
1332
334
101
32
26
371
436
97
68
46
92
378
55
23
18
299
21
31
77
IMBCTSP
(4)
587
563
4298
4077
107
515
327
15100
1998
5342
1352
9016
172
901
472
572
364
111
11600
1384
124
14000
3537
1070
342
278
3923
4608
1021
720
487
975
3998
584
238
192
3162
227
325
814
Total
(5) (in 103)
832
797
6089
5776
152
730
463
21400
2830
7567
1915
12700
243
1277
669
810
515
157
16500
1960
175
19900
5011
1516
484
394
5557
6528
1446
1020
689
1382
5664
828
337
253
4479
322
461
1153
Per Capita
(6)' ($)
1.22
1.47
2.94
2.10
0.39
1.96
2.02
3.07
2.04
3.67
2.25
3.04
1.04
2.02
1.01
1.33
1.96
0.67
2.35
1.08
0.49
1.72
2.70
2.23
0.36
1.15
1.15
2.72
1.43
1.12
2.33
1.57
2.40
3.53
0.64
0.90
1.57
0.93
1.40
2.15
(7)
7834
6269
23883
31763
4500
4293
2647
8240
15973
23809
9807
48280
2685
7305
7656
7027
3031
2681
80920
20919
4102
133280
21414
7850
15673
3944
55420
27696
11639
10530
3419
10181
27255
2700
6112
3506
33001
3975
3801
6182
Ratio
) (S) = (5M(5U(7))
(8)
0.10
0.11
0.20
0.16
0.15
0.15
0.15
0.21
0.15
0.24
0.16
0.21
0.08
0.15
0.08
0.10
0.15
0.06
0.17
0.09
0.04
0.13
0.19
0.16
0.03
0.09
0.09
0.19
0.11
0.09
0.17
0.12
0.17
0.24
0.05
0.07
0.12
0.08
0.11
0.16
 Total
          18,848
            13,251
                             9,425
                             99,483
                                              140,981
711,202
Note:
—  individual figure may not add to  totals due to  rounding.
                                            63

-------
     The table reveals that the low estimate for morbidity damages attributable
to TSP range from $0.15 million in Bridgeport to more than $21 million in
Chicago. On a per capita basis, the low damage estimates for morbidity range
from $360 in Paterson, New Jersey to $3,000 in Chicago. Total morbidity dam-
ages due to TSP over the 40 SMSA's were estimated to be at least $140 million
in 1970.

     Comparison of Tables III-4 and III-5 reveals that the morbidity costs as-
sociated with TSP are larger than the costs associated with SO  . The total mor-
bidity cost due to TSP is $141.2 million, while the total morbidity cost attrib-
utable to SO  is $98.4 million. The ratio between these two costs is 1.43. The
larger morbidity cost  due to TSP is attributable to the fact that the average
TSP level (100.87 |j,g/m  is larger than the average S02 level  (47.95 p,g/m ) and
that TSP has a more responsive dose-response function than SO .

     Note that an important assumption on the independency between SO  and TSP
is made  so  that we can estimate the damage cost separately. In  reality, the
costs of SO and TSP may be larger than the sum of the two component damages
because  of  the possible interaction effects between the two pollutants.

     However, another note of  caution is warranted in interpreting the cost es-
timates  presented in  this study. The effect of SO  as indicated in the regres-
sion equation may represent the effect of not only the single pollutant SO
but also the effect of other pollutants, say TSP, as well. The  prior pollution
studies  suggested that the variable SO  may serve as a proxy variable for air
pollution.  If this is the case, then the pollution damage estimates yielded by
summing  the two computed damages attributable to SO  and TSP may not necessarily
be  smaller  than the actual pollution damages, even if the effect of interaction
is accounted for. Whether the  sum of the two component damages  estimates is
larger or smaller than the actual damages attributable to the concomitant pres-
ence of  the two major pollutants depends on the balance of the  magnitudes of
the two  opposing factors, i.e., the interaction effect versus the double count-
ing effect.

     An "average" economic damage function for TSP with respect to the 40 SMSA's
was developed by the  least-squares technique. Morbidity costs in the presence
of TSP, i.e., the sum of the morbidity costs due to TSP, and  the morbidity costs
in the absence of pollution, were regressed against the same  set of socioeconomic,
demographic and climatological variables appearing earlier in the SO  economic
damage function. The regression results are shown as follows:
     TMBCTSP = -43 + 0.55 TSP - 131.7 PWO + 1.3 SUN + 1.2  RHM
               (74) (0.09)*     (63.3)*       (0.7)**    (0.6)**

                       - 0.2 DTS + 0.07 PCOL +  35.0 AGE
                        (0.2)     (0.09)      (289.7)               (111-24)
                                          R2 = 0.72
                                     64

-------
where  TMBCTSP  denotes the total morbidity cost in the presence of TSP, and
all seven explanatory variables are identical to those defined previously in
Section II. The values below the coefficients are standard errors, with * and
** to denote that the coefficients are significant at the 1 and 5 percent levels.
All coefficients and the corresponding standard errors are reduced by a factor
of 106.

     Since equation (111-24) is developed mainly for prediction purposes, the
"unexpected" signs and possible colinearity among the independent variables
should not present a problem to the use of this equation for estimating TMBCTSP
provided that the signs and the multicolinearity will persist in the future.
However, the use of partial elasticity between the dependent and the independent
variable with wrong signs does cause difficulty in interpreting the results.

     This average economic damage function again is useful for forecasting and
estimating the changes in adult morbidity costs in response to changes in any
of the climatological, demographic, and socioeconomic characteristics, and the
suspended particulate variable. The partial elasticity of the morbidity damages
with respect to suspended particulates is computed as follows: El     qp = 0.55
x  (100.87/708) = 0.08, as measured from the respective mean levels 6f total mor-
bidity costs and suspended particulates. Thus, if the suspended particulate level
in the air is lowered by 10.1 (Jig/m  from 100.87 to 90.76 ^g/m  (i.e., 10 percent
reduction), gross adult morbidity costs on the average would reduce by $5.66
million from $708 to $702.3 million nationwide.
                                     65

-------
                                  SECTION IV

                     HOUSEHOLD SOILING AND AIR POLLUTION
THE PROBLEMS AND THE OBJECTIVES

     In addition to human health,  air pollution has also a multitude of damag-
ing effects on material, vegetation, animals, and residential and commercial
establishments, etc. Ronald Ridker (1967) designed a framework for identifying
and quantifying these damage costs. He suggested that the effects of air pollu-
tion and their costs can be categorized into:  (1) cost of direct effects, (2)
adjustment costs, and (3) market effect costs. The damage costs of human health
derived in the previous two chapters are costs of direct effects of air pollu-
tion. The present section is concerned with the second category; i.e., adjustment
costs or the cost of individual adjustments to the effects of air pollution.

     The best known and the pioneering contribution to the estimation of soiling
loss due to air pollution is the Mellon Institute Study of the Pittsburgh smoke
nuisance (1913). The $20.00 per capita soiling cost figure of the Mellon Insti-
tute Study has been used as a basis for extrapolating to the $11 billion na-
tional damage estimate. The validity of this damage estimate, often quoted by
public officials, has been questioned by Jones (1969) and others. A serious
problem with the national damage estimate arises because of the strong assump-
tion that the air pollution level in Pittsburgh is representative of the entire
nation.

     The two studies of quantifying the soiling costs in the Upper Ohio River
Valley and Washington, D. C. carried out by Michelson and Tourin (1966) have
also attracted public attention. Their methodology is based on the hypothesis
that significant soiling due to air pollution may be reflected in shortened
time intervals between successive cleaning and maintenance operations. Michelson
and Tourin established a positive relationship between frequency of cleaning
operations and the levels of air pollution in both studies. However, the prob-
lems with the sample survey design and the lack of a statistically reliable
technique cast doubt on the reliability of their findings. Michelson and Tourin
(1968) employed the same methodology and estimated the extra household soiling
costs due to air pollution in Connecticut. They found that an average household
spent about $600 each year for coping with the effect of suspended particulates,
with the range from $230 per year in Fairfield to $725 per year in Bridgeport.
These cost estimates are conservative since the cleaning operations studied did
not cover the full gamut of operations affected by air pollution.

     Ridker (1967) conducted interurban studies to determine the relation between
per capita soiling costs and air pollution level for 144 cities in the United
States. Soiling damage costs were approximated by per capita expenditures on
laundry and dry cleaning services. Ridker found that no discernible patterns
between soiling costs and the suspended particulate levels were detected,
whether the effects of climate, per capita income, and price differentials were
                                      66

-------
controlled for or not. The problem often encountered in identifying the soiling
damages, as noted by Ridker, is that cleaning and maintenance operations are
often undertaken on a rigid schedule which is independent of the location of
the operation. This is especially true for commercial and industrial buildings.
Furthermore, nonpollution factors which could not be controlled for may be im-
portant in explaining the cleaning and maintenance procedures.

     The primary objectives of this study are threefold:   a system of soiling
physical damges functions which relate various types of cleaning frequencies
to air pollution level are derived. The physical damage functions are then uti-
lized to estimate net and gross soiling damage costs for the 148 SMSA's. Finally,
"average" economic damage functions over the United States metropolitan areas
are developed by relating soiling damages to air pollution, demographic, socio-
economic, and climatological variables. It is hoped that the generalized eco-
nomic damage functions presented in this section are informative and useful for
predicting possible benefits as a result of the reduction in air pollution when
air pollution abatement programs are implemented.

     This section, which represents a first exploratory effort to estimate aver-
age air pollution soiling damage functions and soiling damage costs for the 148
SMSA's individually, contains subsections: Soiling Physical Damage Function, and
Economic Damages and Economic Damage Functions.
SOILING PHYSICAL DAMAGE FUNCTIONS

     Soiling as a result of falling total suspended particulates compels  house-
holds as well as business and industrial establishments to increase cleaning
activities. Thus, soiling has resulted in extra economic losses not only  to house-
holds but to business and industrial firms as well. As noted above, a number
of attempts have been undertaken to identify and quantify the soiling damages
due to air pollution. However, a recent study by Booz, Allen and Hamilton,  Inc.
(1970), offers the needed data base for our purpose of developing the soiling
physical damage functions.

     Sophisticated and rigorous statistical survey techniques were employed by
Booz-Allen researchers. The Renjerdel area around Philadelphia, Pennsylvania,
was used as the data gathering area. Frequency of cleaning by the residents was
determined by a carefully developed questionnaire containing queries regarding
cleaning operations and a set of self-referent statements with respect to clean-
ing attitudes. Among the 27 cleaning and maintenance operations, the study shows
that 11 were somewhat sensitive to air-suspended particulate levels. Because
of the lack of certain needed information for evaluating the costs, only  9 of
these 11 cleaning tasks were considered in this study. A list of these nine
pollution-related cleaning tasks together with the information on unit cleaning
costs is contained in Table IV-1.
                                     67

-------
       TABLE IV-1.  POLLUTION-RELATED TASKS AND THEIR UNIT CLEANING COSTS
     Tasks                                      Unit Market Value ($)
1
2
3
4
5
6
7
8
9
Replace air conditioner filter
Wash floor surface
Wash inside window
Clean Venetian blinds/shades
Clean/repair screens
Wash outside windows
Clean/repair storm windows
Clean outdoor furniture
Clean gutters
1.00
6.00
0.50
3.50
0.20
1.50
2.00
10.00
15.00

     A set of physical damage functions was derived via the technique delineated
in Section III above, which combines the simulation and regression analysis.
The areas under study were divided into four zones according to their air pollu-
tion levels. This breakdown in the study areas allows one to construct four pop-
ulation "blocks" for each pollution-related cleaning task in the two-dimensional
pollution level and cleaning frequency spaces. For ease of description, let X
and Y  denote respectively the suspended particulate level and cleaning fre-
quency. The vertices of each "block" then consist of the following four combina-
tions:  [Max X, Max Y]; [Max X, Min Y];  [Min X, Max YJ; and [Min X, Min Y], where
Max and Min denote the upper and^ower limits of the two variables. The annual
average particulate levels ((j,g/m ) in the four sampling zones were given in the
Booz-Allen report as follows:
              Zone 1            x < 75

              Zone 2            75 < X < 100

              Zone 3            100 < X < 125

              Zone 4            125 < X

                                       68

-------
     Thus, the suspended particulate levels,  X, vary from 75 (J,g/m  to 100
    3 in Zone 2 and from 100 |ig/m3 to 125 |0,g/m3 in Zone 3. The upper limit of
X in Zone 1 is 75 [ig/m3 and the lower limit of X in Zone 4 is 125 u.g/m3. Assuming
that 25 [ig/m3 of suspended particulate is the background concentration level
and 175 u.g/m3 is the upper limit in the study areas then the values of Min X
and Max X (in (j,g/m3) for the four study zones are tabulated as follows:
               Zone  1

               Zone  2

               Zone  3

               Zone  4
     The minimum and the maximum values for the dependent variable Y (Min Y
and Max Y) for each zone were calculated by subtracting and adding one standard
error of the mean from  the mean value of the cleaning frequency. The computed
values for Min Y and Max Y, the mean frequency of cleaning and the standard er-
ror of the means are presented in Table IV-2.

     The Monte Carlo sampling technique, delineated in Section III, was applied
to the four blocks for  generating a random sample for the regression analysis.
A total of 800 such random observations for each cleaning task were selected.
For the sake of computational simplicity, a smaller random sample, about 20 per-
cent of the 800 random  observations, was further obtained. The 160 observations
included in this sample were fitted via both linear and nonlinear least-squares
techniques. The linear  fit is more superior than the nonlinear fit in all cases
except for Task 8. The  linear regression results for Task 1 through 7 and Task
9 and the nonlinear regression result for Task 8 are summarized in Table IV-3.
ECONOMIC DAMAGES AND ECONOMIC DAMAGE FUNCTIONS

     Given the preceding nine physical damage functions for the nine pollution-
related cleaning tasks and the associated unit cleaning costs which were ob-
tained through telephone conversations with various cleaning firms in Kansas
City, the economic costs of  soiling or of individual household adjustment to
air pollution can be derived by transforming the increased cleaning frequency
into monetary units, via the following two formulas:!.'
_!/  For Task 8, NSC08 = EXP(0.85  -  0.015/(TSP  -  45))  . UC  . U  . HU  and
      GSCOg =  2 + EXP(0.85  - 0.015/(TSP  -  45))  . UC  . U  .  HU.
                                      69

-------
TABLE IV-2. MEAN FREQUENCY,  STANDARD ERROR AND UPPER  AND  LOWER LIMITS OF
            FREQUENCY AND SUSPENDED PARTICULATES
Mean Frequency Standard Error
of Cleaning of Means Mtn Y
Task 1




Task 2




Task 3




Task 4




Task 5




Task 6




Task 7




Task 8




Task 9





Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

Zone 1
Zone 2
Zone 3
Zone 4

0.36
0.50
0.30
0.98

40.55
42.06
42.74
45.17

10.06
11.78
12.74
18.45

4.04
6.17
9.13
9.21

0.80
0.93
0.79
1.50

4.25
4.59
6.17
10.09

2.07
1.60
2.12
3.69

2.50
4.29
3.52
1.19

1.12
1.54
1.35
2.80

0.06
0.08
0.07
0.34

0.84
0.84
0.98
0.93

0.61
0.70
0.82
1.10

0.53
0.66
0.91
0.49

0.07
0.16
0.10
0.32

0.35
0.38
0.60
0.88

0.28
0.23
0.39
0.63

0.45
0.65
0.71
0.47

0.22
0.33
0.44
0.69

0.30
0.42
0.23
0.64

39.71
41.22
41.77
44.24

9.45
11.08
11.93
17.85

3.51
5.51
8.22
8.22

0.75
0.77
0.70
1.18

3.90
4.21
5.57
9.21

1.79
1.37
1.73
3.60

2.05
3.64
2.81
0.72

0.91
1.21
0.91
2.11
Max Y

0.42
0.58
0.37
1.32

41.39
42.90
43.72
46.10

10.17
12.48
13.55
20.05

4.57
6.87
10.04
10.20

0.87
1.09
0.86
1.82

4.60
4.97
6.77
10.97

2.35
1.83
2.51
4.32

2.95
4.94
4.23
1.66

1.34
1.87
1.79
3.49
Min X

25
75
100
125

25
75
100
125

25
75
100
125

25
75
100
125

25
75
100
125

25
75
100
125

25
75
100
125

25
75
100
125

25
75
100
125
Max X

75
100
125
175

75
100
125
175

75
100
125
175

75
100
125
175

75
100
125
175

75
100
125
175

75
100
125
175

75
100
125
175

75
100
125
175
                                   70

-------
TABLE IV-3. SOILING PHYSICAL DAMAGE FUNCTIONS!/
»——•—     -   .  __ _       .

            A.  Frequency =  a + b TSP

   Task       a            b           R2
1
2
3
4
5
6
7
9
0.03
(0.05)
38.6
(0.18)
5.6
(0.4)
2.3
(0.2)
0.42
(0.06)
1.00
(0.28)
0.85
(0.15)
0.27
(0.12)
0.00510
(0.00048)*
0.0400
(0.0017)*
0.078
(0.036)*
0.048
(0.002)*
0.0059
(0.0049)*
0.0530
(0.0025)*
0.015
(0.001)*
0.0140
(0.0011)*
0.43
0.80
0.76
0.79
0.48
0.74
0.48
0.55
         B.  Frequency = c +
8
(c = 2)
0.67
(0.10)
53.2
(7.4)*
0.26

  a/  The values below the coefficients are standard
        errors, with * to indicate that the coefficient
        is significant at the 1 percent level.
                       71

-------
     NSCO. = b.(TSP-45) • UC • U • HU                 (IV-1)


     GSCO. = a. + b.(TSP-45) • UC • U *  HU            (IV-2)

where  NSCO.  and  GSCO.  are, respectively, the net (extra) and gross soiling
damage cost1for the it^type of cleaning task. Coefficients  a±  and  b±  are the
estimated coefficients in the physical damage functions in Table IV-3. i = 1
through 7, and 9. Variables UC, U and HU stand for the unit market value, num-
ber of cleaning objects per household and number of households in a metropolitan
area, respectively.

     To capture the "real" effect of suspended particulates on soiling damages,
the suspended particulate level was adjusted by a threshold level because a low
level of  suspended particulate might have a negligible effect on the household
cleaning  activities. A threshold level of 45 |j,g/m3 for suspended particulate
was assumed as the background concentration level in this study because the low-
est 1970  annual mean level for total suspended particulates was 46.7 (j,g/nP for
Charleston, South Carolina. Alternative reasonable threshold levels can also
be considered. Other things being equal, a higher threshold level is generally
associated with a lower damage cost, and the marginal changes in the damage cost
in response to a unit change in the threshold level is the value of  b.  for
the ith type of cleaning task.

     Given  the data collected for the variables in the formula (IV-1) and (IV-2)
the net and gross household  soiling costs for each of the nine cleaning operations
by the 65 large SMSA's (with population greater than 500,000) in the United States
were derived and presented in Tables IV-4 and IV-6. Similar damage costs for
each of the nine cleaning operations by the 83 medium SMSA's (200,000 to 500,000
people) were presented in Tables IV-6 and IV-7- An examination of the table re-
veals that  Chicago, New York, and Los Angeles, in order of magnitude, suffered
the most  in terms of total net soiling damages. The net soiling damages in these
three SMSA's in 1970 are, respectively, $516 million, $418 million, and $388 mil-
lion. It  is noteworthy that  the cleaning activities of Tasks 4 and 6 in response
to air pollution had resulted in an economic damage of about $1,956 million and
$925.7 million, respectively, in the 40 metropolitan areas. These two tasks con-
stitute the largest damage categories among the nine pollution-related clean-
ing tasks.

     Per  capita net and gross soiling damage costs in the presence of air pollu-
tion for  large SMSA's and medium SMSA's for 1970 are presented, respectively,
in Tables IV-8 and IV-9. Per capita net soiling costs (PCNSCO) and per capita
gross soiling costs (PCGSCO) are summarized in the second and the third columns
of the tables. These cost figures indicate that the soiling damages attributable
to air pollution in large SMSA's range from $5 per person in San Antonio, Texas,
to $104 per person in Cleveland, Ohio, whereas the net soiling damages in medium
SMSA's vary from less than a dollar per person in Charleston, South Carolina,
to $67.35 per person in Wichita, Kansas. These estimates for individual SMSA's
appear to be compatible with the overall per capita soiling damage estimates
of $20.00 by Mellon Institute and of $200 by Michelson and Tourin.
                                      72

-------
                                                         TABLE IV-lt. NET SOILING DAMAGE COSTS  BY LARGE  SMSA's£/
                                                                     (million $)

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.
33.
34.
35.
Large SMSA' s
AKR, OH
ALB, NY
ALL, NJ
ANA, CA
ATL, GA
BAL, MD
BIR, AL
BOS, MA
BUF, NY
CHI, IL
CIN, OH-KY-IN
CLE, OH
COL, OH
DAL, TX
DAY, OH
DEN, CO
DET, MI
FOR, FL
FOR, TX
GAR, IN
GRA, MI
GRE, NC
HAR, CT
HON, HI
HOU, TX
IND, IN
JAC, FL
JER, NJ
KAN, MO-KS
LOS, CA
LOU, KY-IN
MEM, TN-AR
MIA, FL
MIL, WI
MINN.MN
NSC01
..
0.1
--
0.1
0.1
0.3
0.2
0.3
0.2
1.2
0.1
0.5
0.1
0.1
0.1
0.2
0.7
--
0.1
0.1
__
--
--
--
0.1
0.1
--
--
0.1
0.9
0.1
0.1
--
0.1
0.1
NSC02
1.7
4.0
1.8
6.1
3.8
15.3
7.7
12.9
8.1
57.5
6. .3
24.3
2.3
6.8
4.4
10.1
32.7
0.9
2.6
2.7
1.1
1.9
1.8
1.2
6.4
2.5
1.2
1.9
4.0
42.8
6.3
2.6
1.7
4.8
4.1
NSC03
1.4
3.2
1.4
5.0
3.1
12.4
6.2
10.5
6.6
46.7
5.1
IV. 7
1.9
5.5
3.5
8.2
26.6
0.7
2.1
2.2
0.9
1.5
1.4
1.0
5.2
2.0
0.9
1.5
3.3
34.8
5.1
2.1
1.4
3.9
3.3
NSC04
6.3
14.0
6.2
21.3
13.1
53.6
26.7
45.3
28.3
201.0
21.9
85.0
8.2
23.7
15.2
35.3
114.0
3.2
9.2
9.3
4.0
6.5
6.2
4.1
22.4
8.9
3.8
6.7
14.1
150.0
21.9
9.2
6.0
16.9
14.4
NSC05
..
0.1
--
0.2
0.1
0.4
0.2
0.3
0.2
1.4
0.2
0.6
0.1
0.2
0.1
0.2
0.8
--
0.1
0.1
__
--
--
--
0.2
0.1
--
--
0.1
1.1
0.2
0.1
--
0.1
0.1
NSC 06
2.9
6.7
2.9
10.1
6.2
25.3
12.6
21.4
12.4
95.2
10.4
40.2
3.9
11.2
7.2
16.7
54.3
1.5
4.3
4.4
1.9
3.1
2.9
2.0
10.6
4.2
1.8
3.2
6.7
71.0
0.2
4.4
2.9
8.0
6.8
. NSC07
1.1
2.5
1.1
3.8
2.4
9.6
4.7
8.1
5.1
35.9
3.9
15.1
1.5
4.2
2.7
6.3
20.4
0.6
1.6
1.7
0.7
1.2
1.1
0.7
4.0
1.6
0.7
1.2
2.5
2.7
10.3
1.6
1.1
3.0
2.6
NSC08
0.9
2.2
1.0
3.4
2.0
7.3
3.1
7.2
4.2
26.2
3.5
9.0
1.2
3.8
2.4
4.7
15.1
0.2
1.5
1.5
0.5
1.0
0.8
0.5
3.5
1.2
4.7
1.0
2.2
2.3
3.9
1.5
0.3
2.7
1.9
NSC09
1.5
3.5
1.5
5.3
2.2
13.4
6.7
11.3
7.1
50.3
5.5
21.2
2.1
5.9
3.8
8.8
28.6
0.8
2.3
2.3
1.0
1.6
1.6
1.0
5.6
2.2
0.9
1.7
3.5
3.8
5.5
2.3
1.5
4.2
3.6
TNSCO
15.5
36.4
15.9
55.4
34.0
137.0
68.2
117.0
73.1
516.0
57.0
216.0
21.1
61.4
39.4
90.7
294.0
7.8
23.7
24.2
10.1
16.9
15.8
10.5
58.1
22.7
9.7
17.2
36.6
383.0
56.3
23.8
15.0
43.9
36.9
a/  NSC01   stands for the net soiling cost for  the  1th  type of operation,  i  =  1, 2,.  .  .,9.
      TOTNETSL Is the sum of NESOC01 over 1 and  "--•'  Indicates that  the  figure  Is less  than 0.05.

-------
                                                          TABLE IV-4 (Concluded)
•vl
-P-

36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.

Large SMSA's
NAS, TN
NEW, LA
NEW, NY
NEW, NJ
NOR, VA
OKL, OK
OMA, NE-LA
PAT, NJ
Pill, PA-NJ
PI10, AZ
PIT, PA
FOR, OR-WA
PRO, RI-MA
RIG, VA
ROC, NY
SAC, CA
SAI, MO-IL
SAL, UT
SAN, TX
SAN, CA
SAN, CA
SAN, CA
SAN, CA
SEA, WA
SPR, MC-CT
SYR, NY
TAM, FL
TOL, OH-MI
WAS, DC-MD-VA
YOU, OH
Total
NSC01
0.1
0.1
1.0
0.3
0.1
__
0.1
--
0.2
0.2
0.3
0.1
—
0.1
0.1
--
0.3
--
--
0.1
—
0.1
--
--
--
0.1
0.1
0.1
0.2
0.1
9.9
NSC02
3.2
2.7
46.0
12.4
3.2
1.2
3.8
1.1
11.5
10.4
16.3
3.3
2.2
2.7
3.1
1.0
13.1
1.9
0.5
7.9
1.4
4.0
1.2
1.4
0.8
3.0
2.7
4.0
9.7
2.5
474.5
NSC03
2.6
2.2
37.4
10.1
0.6
0.9
3.1
0.9
9.4
8.5
13.2
2.7
1.8
2.2
2.4
0.8
10.6
1.5
0.4
6.3
1.1
3.2
0.9
1.2
0.6
2.5
2.2
3.3
7.8
2.0
382.7
NSC04
11.1
9.3
161.0
43.5
11.0
3.8
13.2
3.8
40.4
36.4
57.1
11.6
7.7
9.4
10.2
3.5
46.0
6.6
1.8
27.5
4.8
14.0
4.0
5.0
2.7
10.6
9.4
14.1
37.7
8.9
1.662.0
NSC05
0.1
0.1
1.1
0.3
0.1
..
0.1
—
0.3
0.3
0.4
0.1
0.1
0.1
0.1
__
0.2
--
--
0.2
—
0.1
--
--
--
0.1
0.1
0.1
0.2
0.1
10.6
NSC06
5.3
4.4
76.3
20.6
5.2
1.8
6.3
1.8
19.1
17.2
27.0
5.5
3.6
4.4
4.9
1.7
21.7
3.1
0.8
13.0
2.3
6.7
1.9
2.3
1.3
5.0
4.5
6.7
15.9
4.2
774.0
NSCO7
2.0
1.7
28.7
7.8
2.0
6.8
2.4
0.7
7.2
6.5
10.2
2.1
1.4
1.7
1.8
0.6
8.2
1.2
0.3
4.9
0.9
2.5
0.7
0.9
0.5
1.9
1.7
2.5
6.0
1.6
284.7
NSC08
1.7
1.3
25.8
6.3
1.7
0.3
1.8
0.1
5.6
4.1
8.2
1.8
1.1
1.5
1.6
0.2
7.0
1.2
--
3.9
0.2
0.6
0.2
0.1
0.2
1.7
1.3
2.1
5.42
1.4
216.8
NSC09
2.8
2.3
40.3
10.8
2.8
1.0
3.3
1.0
10.1
9.1
14.2
3.0
1.9
2.3
2.5
0.9
11.5
1.7
0.4
6.9
1.2
3.5
1.0
1.2
0.7
2.7
2.4
3.5
8.4
2.2
379.7
TNSCO
28.9
23.9
418.0
112.0
28.5
9.6
34.0
9.3
104.0
92.8
147.0
30.1
19.7
24.2
26.5
8.8
119.0
17.2
4.3
70.8
11.9
34.8
10.0
12.1
6.7
27.4
24.2
36.4
85.5
23.0
2,465.9
        Note: —  individual  figure may not add  to   totals  due to rounding.

-------
TABLE IV-5. NET SOILING DAMAGE COSTS BY MEDIUM SMSA'
            (million $)
Medium SMSA' 8
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
ALB,
ANN,
AFP,
AUG,
AUS,
BAK,
BAT,
BEA,
BIN,
BRI,
CAN,
CI1A,
CHA,
CHA,
CIIA,
COL,
COL,
COL,
COR,
DAV,
DBS,
DUL,
ELF,
ERI,
EUG,
EVA,
FAY,
FLI,
FOR,
FRE,
GRE,
HAM,
HAR,
HUN,
IRJN,
MM
MI
WI
GA-SC
TX
CA
LA
TX
NY^PA
CN
OH
SC
WV
NC
TN-CA
CO
SC
GA-AL
TX
IA-IL
IA
MN-WI
TX
PA
OR
IN-KY
NC
MI
IN
CA
SC
OH
PA
WV-KY.OH
AL
NSC01 NSC02
I.I
0.5
0.9
0.3
0.5
2.2
0.3
0.3
0.3
0.4
1.6
..
1.1
1.6
1.4
0.8
0.4
0.1
1.1
2.3
0.9
0.5
2.2
1.1
0.7
0.5
0.3
0.1 3.0
0.6
2.1
0.7
0.6
1.0
1.0
0.3
NSC03
0.9
0.4
0.7
0.3
0.4
1.8
0.3
0.3
0.2
0.3
1.3
--
0.9
1.3
1.2
0.7
0.4
0.1
0.9
1.8
0.7
0.4
1.9
0.9
0.5
0.4
0.2
2.4
0.5
1.7
0.6
0.5
0.8
0.9
0.2
NSC04 NSC05
3.
1.
3.
1.
1.
7
1
1
1
1
5
0.
3.
5,
5.
1.
1.
0.
4,
8.
3.
1.
7.
3,
2.
2.
0.
10.
2,
7.
2.
2,
3,
3.
1,
.7
.7
1
,1
9
.7 0.1
.1
.2
.0
.2
.6
.1
.7
.7
.0
,9
3
.3
.0
.0 0.1
,1
,9
,9
.9
,3
.0
.9
.2 0.1
.2 00
.5 0.1
.4
.0 0.1
.6
.5
.0
NSC06
1.8
0.8
1.4
0.5
0.9
3.7
0.5
0.6
0.4
0.6
2.6
0.1
1.8
2.7
2.4
1.4
0.6
0.1
1.9
3.8
1.5
0.9
3.8
1.9
1.1
0.9
0.4
4.9
1.0
3.5
1.1
1.0
1.7
1.7
0.5
NSC07
0.7
0.3
0.5
0.2
0.3
1.4
0.2
0.2
0.2
0.2
1.0
--
0.7
1.0
0.9
0.5
0.2
--
0.7
1.4
0.6
0.3
1.4
0.7
0.4
0.3
0.2
1.8
0.4
1.3
0.4
0.4
0.6
0.6
0.2
NSC08
0.6
0.2
0.5
0.1
0.2
1.1
0.1
0.1
--
--
0.9
--
0.6
0.9
0.8
0.5
0.1
--
0.6
1.1
0.5
0.2
1.1
0.6
0.4
0.3
0.1
1.5
0.3
1.2
0.3
0.3
0.5
0.6
0.1
NSC09
0.9
0.4
0.8
0.3
0.5
1.9
0.3
0.3
0.2
0.3
1.4
--
0.9
1.4
1.2
0.7
0.3
0.1
1.0
2.0
0.8
0.5
2.0
1.0
0.6
0.5
0.2
2.6
0.5
1.9
0.6
0.5
0.9
0.9
0.9
TNSCO
9.7
4.3
7.9
2.7
4.8
19.9
2.7
3.0
2.5
3.0
14.4
0.3
9.6
14.6
12.9
7.6
3.2
0.7
10.2
20.4
8.1
4.8
20.1
10.0
6.0
5.0
2.2
26.5
5.6
19.3
6.2
5.3
9.2
9.1
9.7

-------
TABLE IV-5 (Continued)

101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118,
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.

JAC,
JOH,
KAL,
KNO,
LAN,
LAN,
LAS,
LAW,
LITT
LOR,
LOW,
MAC,
MAD,
MOB,
MON,
NEW,
NEW,
NEW,
ORL,
OXN,
PEN,
PEO,
RAL,
RE A,
ROC,
SAG,
SAL,
SAN,
SAN,
SCR,

MS
PA
MI
TN
PA
MI
NV
MA-NH
, AK
OH
MA
GA
WI
AL
AL
CN
CN
VA
FL
CA
FL
IL
NC
PA
IL
MI
CA
CA
CA
PA
NSC01 NSC02
1.1
1.1
0.2
1.7
1.5
0.9
1.2
0.4
0.7
0.1 2.7
0.1
0.5
0.6
1.6
0.7
0.4
0.2
0.2
1.0
1.9
1.0
0.8
0.2
1.7
1.2
1.3
1.2
1.5
1.2
0.1 2.6
NSC03
0.9
0.9
0.2
1.3
1.2
0.7
1.0
0.3
0.6
2.2
0.1
0.4
0.5
1.3
0.6
0.4
0.2
0.1
0.8
1.5
0.8
0.7
0.1
1.4
1.0
1.0
1.0
1.2
1.0
2.1
NSC04 NSC05
3
4
0
5
5
3
4
1
2
9
0
1
2
5
2
1
0
0
3
6
3
3
0
6
4
4
4
5
4
9
.8
.0
.7
.8
.2
.1
.1
.2
.5
.5 0.1
.3
.9
.2
.6
.6
.4
.9
.6
.4
.5
.7
.0
.6
.0
.2
.5
.2
.2
.2
.2 0.1
NSC06
1.8
1.9
0.3
2.7
2.4
1.5
1.9
0.6
1.2
4.5
0.1
0.9
1.0
2.7
1.2
0.7
0.4
0.3
1.6
3.1
1.7
1.4
0.3
2.8
2.0
2.1
2.0
2.5
2.0
4.3
NSC07
0.7
0.7
0.1
10.
0.9
0.5
0.7
0.2
0.4
1.7
..
0.3
0.4
1.0
0.5
0.2
0.2
0.1
0.6
1.2
0.7
0.5
0.1
1.0
0.8
0.8
0.7
0.9
0.7
1.6
NSC08
0.6
0.6
--
0.9
0.8
0.4
0.7
0.1
0.3
1.0
__
0.3
0.3
0.9
0.4
0.1
0.1
--
0.5
1.0
0.6
0.4
--
0.9
0.7
0.7
0.6
0.8
0.6
1.0
NSC 09
0.9
1.0
0.2
1.5
1.3
0.8
1.0
0.3
0.6
2.4
0.1
0.5
0.5
1.4
0.7
0.3
0.2
0.2
0.9
1.6
0.9
0.-7
0.1
1.5
1.0
1.2
1.0
1.3
1.0
2.3
TNSCO
9.7
10.1
1.7
15.0
13.3
7.9
10.6
3.1
6.4
24.1
0.7
4.9
5.5
14.5
6.8
3.5
2.1
1.5
8.8
16.9
9.5
7.6
1.4
15.3
10.9
11.6
10.8
13.4
10.8
23.3

-------
                                                      TABLE IV-5 (Concluded)

131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.

SHR,
SOU,
SPO,
STA,
STO,
TAC,
TRE,
TUC,
TUL,
UTI,
VAL,
WAT,
WES,
WIC,
WIL,
WIL,
WOR,
YOR,

LA
IN
WA
CN
CA
WA
NJ
AZ
OK
NY
CA
CN
FL
KS
PA
DE.NJ.MD
MA
PA
Total
NSC01 NSC02
1.3
0.6
1.2
0.2
0.3
1.4
0.6
1.4
1.5
0.7
0.3
0.5
0.5
0.1 2.9
2.2
0.1 2.9
0.7
1.0
1.9 84.3
NSC03
1.1
0.5
1.0
0.2
0.3
1.2
0.5
1.1
1.2
0.5
0.2
0.4
0.4
2.4
1.8
2.3
0.6
0.8
71.4
NSC04
4.6
2.2
4.2
0.7
1.1
5.1
2.0
4.9
5.1
2.3
0.9
1.9
1.7
10.2
7.8
10.0
2.4
3.5
294.3
NSC05 NSC06
2
1
2
0
.2
.1
.0
.3
0.5
2
1
2
2
1
0
0
0
0.1 4
0.1 3
0.1 4
1
1
.4
.0
.3
.4
.1
.4
.9
.8
.8
.7
.7
.1
.7
1.3 151.7
NSC07
0.8
0.4
0.7
0.1
0.2
0.9
0.4
0.9
0.9
0.4
0.2
0.3
0.3
1.8
1.4
1.8
0.4
0.6
64.5
NSC08
0.7
0.3
0.7
—
--
0.8
0.2
0.8
0.8
0.3
--
0.3
0.1
1.4
1.1
1.5
0.3
0.5
58.3
NSC09
1.2
0.6
1.0
0.2
0.3
1.3
0.5
1.2
1.3
0.6
0.2
0.5
0.4
2.6
1.9
2.5
0.6
0.9
109.2
TNSCO
11.9
5.7
10.8
1.6
2.8
13.0
5.2
12.8
13.1
8.9
2.3
4.9
4.3
26.2
20.0
25.9
6.2
9.0
767.2

Note:  — individual  figure may not  add to totals due  to  rounding.

-------
                                                            TABLE  IV-6. GROSS SOILING DAMAGE COSTS BY LARGE SMSA*/
                                                                        (million $)
oo

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.
33.
34.
35.
Large SMSA's
AKR, OH
ALB, NY
ALL, NJ
ANA, CA
ATL, GA
BAL, MD
BIR, AL
BOS, MA
BUF, NY
CHI, IL
CIN, OH-KY-IN
CLE, OH
COL, OH
DAL, TX
DAY, OH
DEN, CO
DET, MI
FOR, FL
FOR, TX
GAR, IN
GRA, MI
GRE, NC
HAR, CT
HON, HI
HOU, TX
IND, IN
JAC,, FL
JER, NJ
KAN, MO-KS
LOS, CA
LOU, KY-IN
MEM, TN-AR
MIA, FL
MIL, WI
MINN, MN
GSC01

0.1
—
--
0.1
0.3
0.2
0.3
0.2
1.3
__
0.5
0.1
0.2
0.1
0.2
0.7
--
0.1
0.1
„.
--
--
--
0.2
0.1
--
--
0.1
1.0
„_
0.1
--
0.1
0.1
GSC02
49.4
57.2
42.2
107.0
103.0
159.0
61.3
212.0
104.0
563.0
106.0
174.0
67.8
120.0
65.2
100.0
326.0
52.5
58.4
45.0
38.1
45.1
60.8
39.3
147.0
82.8
38.5
49.8
98.7
606.0
67.8
55.2
100.0
105.0
133.0
GSC03
0.9
1.3
0.8
2.2
1.8
4.2
1.2
4.5
2.5
15.4
2.2
5.8
1.2
2.5
1.4
2.7
8.9
0.8
1.1
1.0
0.6
0.8
4.1
0.7
4.8
1.4
0.6
0.9
1.8
1.4
1.7
1.1
1.5
2.0
2.2
GSC04
14.3
23.3
13.2
38.9
20.5
78.8
36.1
80.1
45.2
289.0
39.4
111.0
19.6
43.5
25.8
51.2
166.0
12.1
18.9
16.8
10.4
14.0
16.5
10.8
47.1
22.8
10.3
15.0
30.7
248.0
32.3
18.3
23.3
34.4
36.9
GSC05
0.1
0.2
0.1
0.3
0.3
0.6
0.3
0.7
0.4
2.3
0.3
0.9
0.2
0.4
0.2
0.4
1.3
0.1
0.2
0.1
0.1
0.1
0.2
0.1
0.4
0.2
0.1
0.1
0.3
2.1
0.3
0.2
0.2
0.3
0.3
GSCO6
4.4
8.4
4.2
13.3
9.5
30.0
14.4
27.9
16.5
111.0
13.6
45.1
6.0
14.9
9.2
19.6
63.8
3.2
6.2
5.8
3.1
4.5
4.8
3.2
15.2
6.8
3.0
4.7
9.8
89.2
12.3
6.1
6.1
11.2
10.9
GSC07
2.8
4.5
2.6
7.5
6.0
14.8
6.8
15.4
8.6
54.5
7.6
20.7
3.9
8.4
5.0
9.7
31.2
2.5
3.7
3.2
2.1
2.8
3.3
2.1
9.2
4.5
2.1
3.0
6.0
4.7
6.1
3.6
4.7
6.7
7.3
GSC08
5.0
6.8
4.5
12.1
10.5
19.7
7.7
24.4
12.5
69.9
12.1
22.0
6.8
13.6
7.6
12.5
40.4
4.6
9.6
5.2
3.7
4.8
5.9
3.8
15.7
8.1
3.7
5.2
10.3
71.6
8.1
6.0
8.9
11.3
13.0
GSC09
2.3
4.5
2.3
7.1
5.0
15.9
7.6
14.8
8.8
59.1
7.2
23.9
3.2
7.9
4.9
10.1
33.8
1.7
3.3
3.1
1.6
2.4
2.6
1.7
8.1
3.6
1.6
2.5
5.2
47.3
6.5
3.2
3.2
6.0
5.9
TGSCO
79.3
106.0
69.9
188.0
166.0
324.0
136.0
380.0
199.0
1160.0
188.0
405.0
108.0
212.0
119.0
207.0
672.0
77.5
98.0
80.8
59.8
74.6
95.1
61.8
246.0
130.0
60.0
81.2
163.0
1,120.0
133.0
93.7
148.0
117.0
209.0
        a/  GSCOi  denotes  the  gross  soiling damage  cost  for  the ith type of cleaning  operation,  i  = 1,  2,.
              TGSCO  is  the sum of  GSCOj  over i.
9,

-------
                                                   TABLE IV-6 (Concluded)

36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.

Large SMSA's
NAS, TN
NEW, LA
NEW, NY
NEW, NJ
NOR, VA
OKL, OK
DMA, NE-IA
PAT, NJ
PHI, PA-NJ
PHO, AZ
PIT, PA
FOR, OR-WA
PRO, RI-MA
RIC, VA
ROC, NY
SAC, CA
SAI, MO-IL
SAL", UT
SAN, TX
SAN, CA
SAN, CA
SAN, CA
SAN, CA
SEA, WA
SPR, MC-CT
SYR, NY
TAM, FL
TOL, OH-MI
WAS, DC-MD-VA
YOU, OH
Total
GSC01
0.1
0.1
0.1
0.3
0.1
	
0.1
--
0.3
0.2
0.4
0.1
0.1
0.1
0.1
	
0.3
--
—
0.2
—
0.1
--
--
--
0.1
0.1
0.1
0.2
0.1
9.5
GSC02
42.3
76.3
939.0
147.0
47.6
49.7
42.0
99.9
354.0
80.5
192.0
82.5
67.9
40.4
65.6
60.3
183.0
38.7
57.0
91.9
99.5
265.0
75.9
110.0
38.9
47.4
88.3
53.3
247.0
40.0
8,063.0
GSC03
1.0
1.3
18.2
3.7
1.1
0.8
1.1
1.4
6.0
2.5
4.8
1.5
1.2
0.9
1.2
0.9
4.2
0.6
0.8
2.3
1.4
3.8
1.1
1.6
0.6
1.0
1.5
1.3
4.1
0.9
169.8
GSC04
18.0
22.1
317.0
67.1
18.8
12.3
19.9
21.0
100.0
48.7
87.8
25.4
19.1
15.9
21.1
13.8
75.8
13.0
11.6
42.1
21.9
59.7
17.0
24.0
9.3
18.3
24.4
22.7
70.1
15.4
2,967.8
GSC05
0.2
0.2
2.8
0.6
0.2
0.1
0.2
0.2
0.9
0.4
0.7
0.2
0.2
0.1
0.2
0.1
0,6
0.1
0.1
0.3
0.2
0.6
0.2
0.2
0.1
0.2
0.2
0.2
0.6
0.1
25.4
GSC06
6.6
6.8
105.0
25.0
6.7
3.4
7.5
5.0
30.2
19.5
32.7
8.1
5.7
5.7
6.9
3.6
27.3
4.3
2.7
15.7
5.5
15.1
4.3
5.9
2.5
6.5
7.2
8.3
22.7
5.4
1,029.7
GSC07
3.4
4.3
61.5
12.7
3.6
2.5
3.8
4.3
19.8
9.1
16.6
5.0
3.8
3.1
4.1
2.8
14.4
2.5
2.4
8.0
4.5
12.1
3.5
4.9
1.9
3.4
4.8
4.3
13.6
3.0
531.3
GSC08
5.1
7.7
103.0
17.9
5.6
4.6
5.1
8.6
32.2
10.1
23.3
8.6
6.7
4.7
7.0
5.3
21.7
4.2
4.9
11.1
8.6
23.1
6.6
9.6
3.5
4.4
8.7
6.4
23.3
4.6
883.8
GSC09
3.5
3.6
55.9
13.2
3.5
1.8
4.0
2.7
16.1
10.3
17.3
4.3
3.1
3.0
3.7
1.9
14.4
2.3
1.4
8.3
2.9
8.1
2.3
3.2
1.3
3.4
3.9
4.4
12.0
2.9
546.6
IE SCO
80.1
122.0
1,600.0
288.0
87.0
75.1
83.6
143.0
563.0
181.0
375.0
135.0
107.0
73.9
110.0
88.8
342.0
66.1
80.9
180.0
144.0
388.0
111.0
160.0
58.2
86.0
129.0
101.0
364.0
72.4
14,162.8
Note:  — individual figure may not add to  totals due to  rounding.

-------
                                                                            TABLE IV-7. GROSS SOILING DAI1AGE  COSTS  BY MEDIUM SMSA's
                                                                                        (million $)
00
O
Medium SMSA's
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
ALB,
ANN,
APP,
AUG,
AUS,
BAR,
BAT,
BF,A,
BIN,
BRI,
CAN,
CHA,
CHA,
CHA,
CHA,
COL.
COL,
COL,
COR,
DAY,
DES,
DUL,
ELP,
ERI,
BUG,
EVA,
FAY,
FLI,
FOR,
FRF,,
GRE,
HAM,
HAR,
Hl'N,
HUN,
NH
MI
WI
GA-SC
TX
CA
LA
TX
NY-PA
CN
OH
SC
WV
NC
TN-GA
CO
SC
GA-AL
TX
IA-IL
IA
MN-Wr
TX
PA
OR
IN-KY
NC
MI
IN
CA
SC
OH
PA
WV-KY-OH
AL
GSC01 GSC02
22.8
16.4
18.9
16.7
21.6
25.8
19.0
22.8
21.8
28.1
28.2
19.2
18.1
30.5
24.1
16.5
20.2
15.5
19.6
28.9
22.4
19.9
24.4
19.4
16.4
17.9
12.2
0.1 36.2
20.5
31.5
21.7
16.1
31.5
19.7
15.3
GSC03
0.4
0.3
0.4
.0.2
0.3-
0.6
0.3
0.3
0.3
0.4
0.6
0.2
0.4
0.6
0.5
0.3
0.3
0.2
0.4
0.7
0.4
0.3
0.6
0.4
0.3
0.3
0.2
0.9
0.3
0.7
0.4
0.3
0.5
0.4
U.2
GSC04
7.5
4.5
6.2
0.3
5.6
11.8
4.4
5.1
4.8
6.1
10.2
3.5
6.7
10.7
8.9
5.7
4.8
3.0
7.1
12.5
7.0
5.3
11.7
7.1
5.1
5.0
3.0
16.1
5.7
12.5
0.1
4.8
9.0
6.8
3.6
GSC05
0.1
--
0.1
--
0.1
0.1
—
--
--
0.1
0.1
--
0.1
0.1
0.1
__
--
--
--
0.1
--
--
0.1
0.1
--
--
--
0.1
0.1
0.1
1.8
--
0.1
0.1
--
GSC06
2.5
1.3
2.0
1.0
1.6
4.4
1.1
1.3
1.2
1.5
3.5
0.7
2.3
3.6
3.1
1.9
1.3
0.6
2.5
4.6
2.1
1.6
4.4
2.5
1.6
1.5
0.8
6.0
1.7
4.5
1.2
1.5
2.7
2.3
1.0
GSC07
1.5
0.9
1.2
0.8
1.1
2.2
0.9
1.0
1.0
1.2
2.0
0.7
1.3
201
1.7
1.1
1.0
0.6
1.4
2.4
1.3
1.0
2.2
1.4
1.0
1.0
0.6
3.1
1.1
2.4
1.2
0.9
1.8
1.3
0.7
GSC08
2.5
1.6
2.0
1.5
2.0
3.1
1.7
2.0
1.9
2.4
3.2
1.7
2.1
3.4
2.8
1.8
1.8
1.3
2.2
3.5
2.3
1.9
3.0
2.2
4.7
1.8
1.1
4.4
2.0
3.7
2.2
1.6
3.1
2.2
1.4
GSC09
1.3
0.7
1.0
0.6
0.8
2.3
0.6
0.7
0.6
0.8
1.9
0.4
1.2
1.9
1.6
1.0
0.6
0.3
1.3
2.4
1.1
0.9
2.4
1.3
0.9
0.8
0.4
3.2
0.9
2.4
1.0
0.8
1.4
1.2
0.5
TGSCO
38.6
25.7
31.9
24.8
33.1
50.6
28.0
33.3
31.5
40.6
49.6
26.4
32.2
53.0
42.8
28.5
30.1
21.7
34.7
55.2
36.8
30.9
49.0
34.4
27.0
28.1
18.4
70.0
32.3
57.9
34.4
26.0
50.2
33.9
22.7

-------
                                                                                    TABLE IV-7 (Continued)
00

101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126
127.
128.
129.
130.

JAC,
JOH,
KAL,
KNO,
LAN,
LAN,
LAS,
LAW,
LITT
LOR,
LOW,
MAC,
MAD,
MOB,
MON,
NEW
NEW,
NEW,
ORL,
OXN,
PEN,
PEO,
RAL,
REA,
ROC,
SAG,
SAL,
SAN,
SAN,
SCR,

MS
PA
MI
TK
PA
MI
NV
MA, NH
, AK
OH
MA
GA
WI
AL
AL
CN
CN
VA
FL
CA
FL
IL
NC
PA
IL
MI
CA
CA
CA.
PA
GSC01 GSC02
18.2
19.8
13.8
21.0
24.1
26.5
21.5
17.4
24.5
0.1 19.6
13.9
14.9
21.2
26.8
14.6
26.5
14.3
19.3
32.0
26.4
17.4
25.8
15.9
24.3
20.6
15.8
17.6
20.9
16.9
0.1 20.2
GSC03
0.4
0.4
0.2
0.6
0.5
0.5
0.4
0.3
0.4
0.6
0.3
0.3
0.3
0.6
0.3
0.4
0.2
0.3
0.5
0.6
0.4
0.4
0.2
0.5
0.4
0.4
0.4
0.5
0.4
0.6
GSC04
5.7
7.2
5.3
10.9
9.1
7.6
7.7
4.2
5.7
12.4
3.7
5.4
5.8
10.0
5.0
6.0
3.3
4.0
8.8
10.8
6.5
7.3
3.3
9.9
7.6
7.1
7.0
8.6
6.9
12.2
GSC05
0.1
0.1
--
0.1
0.1
0.1
0.1
--
0.1
0.1
..
0.3
0.1
0.1
--
0.1
--
==
0.1
0.1
0.1
0.1
--
0.1
0.1
0.1
0.1
0.1
0.1
0.1
GSC06
2.3
2.5
0.8
3.7
3.2
2.3
2.6
1.1
2.0
5.1
0.6
1.4
1.7
3.5
1.7
1.5
0.9
0.9
2.6
3.9
2.3
2.2
0.8
3.6
2.5
2.6
2.5
3.1
2.5
4.9
GSC07
1.3
1.4
0.6
2.1
1.8
1.5
1.5
0.9
1.3
2.3
0.6
0.9
1 .1
1.9
1.0
1.2
0.7
0.8
1.8
2.1
1.3
1.4
0.7
1.9
1.5
1.3
1.3
1.6
1.3
2.3
GSC08
2.1
2.3
1.2
3.5
2.8
2.7
1.4
1.6
2.4
2.5
1.2
1.5
2.1
3.1
1.6
2.3
1.3
1.7
3.1
3.1
2.0
2.6
1.4
2.9
2.4
1.9
2.1
2.5
2.0
2.5
GSCP?
1.2
1.3
0.4
2.0
1.7
1.2
1.4
0.6
1.1
2.7
0.3
0.7
0.9
1.8
0.9
0.8
0.5
0.5
1.4
2.1
1.2
1.2
0.4
1.9
1.4
1.4
1.3
1.6
1.3
2.6
TGSCO
32.3
35.0
20.1
54.0
43.2
42.3
37.6
26.2
38.4
45.4
19.5
24.1
33.1
47.8
25.2
38.8
21.2
27.5
50.3
49.1
31.1
41.1
22.7
45.1
36.6
30.6
32.4
38.9
31.5
45.6

-------
                                                              TABLE IV-7. (Concluded)
00
CO
GSC01 GSC02
131.
132.
133..
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146
147.
148
SHR,
SOU,
SPO,
STA,
STO,
TAG,
TRE,
TUG,
TUL,
UTI,
VAL,
WAT,
WES,
WIG,
WIL,
WIL,
WOR,
YOR,
LA
IN
WA
CN
CA
WA
NJ
AZ
OK
NY
CA
CN
FL
KS
PA
DE, NJ, MD
MA
PA
Total
22
20
22
15
21
29
22
27
38
24
17
15
28
0.1 31
28
0.1 37
25
25
3.3 1,821
.3
.7
.9
.0
.6
.9
.1
.1
.2
.5
.8
.3
.9
.8
.1
.3
.0
.3
.9
GSC03
0.5
0.3
0.5
0.2
0.3
0.6
0.4
0.5
0.7
0.4
0.3
0.3
0.4
0.8
0.7
0.9
0.4
0.5
34.2
GSC04
8.3
5.7
8.0
3.2
4.9
10.0
5.8
9.4
11.5
6.5
4.0
4.5
6.7
15.2
12.2
16.0
6.7
7.7
616.3
GSC05
0.1
0.1
0.1
—
--
0.1
0.1
0.1
0.1
0.1
_.
--
0.1
0.1
0.1
0.1
0.1
0.1
5.2
GSC06
2.9
1.7
2.7
0.8
1.21
3.3
1.7
3.2
3.6
1.9
1.0
1.4
1.7
5.8
4.5
5.9
1.9
2.4
198.2
GSC07
1.6
1.1
1.5
0.7
1.0
1.9
1.2
1.8
2.3
1.3
0.8
0.9
1.4
2.9
2.3
3.1
1.3
1.5
160.3
GSC08
2.6
2.0
2.6
1.3
1.9
3.3
2.1
3.1
3.9
2.3
1.6
1.6
2.6
3.9
3.4
4.5
2.4
2.6
195.4
GSC09
1.5
0.9
1.4
0.4
0.7
1.8
0.9
1.7
1.9
1.0
0.5
0.7
0.9
3.1
2.4
3.1
1.0
1.3
105.4
TGSCO
39.7
32.7
39.6
21.6
31.6
50.9
34.1
46.9
62.3
37.9
26.1
24.7
42.7
63.7
53.8
71.0
38.8
41.4
3,204.3
    Note:  — individual figure may not  add to totals due to  rounding,

-------
TABLE IV-8. PER CAPITA NET AND GROSS SOILING DAMAGE COSTS ($) BY LARGE SMSA1 s,
            1970


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.
33.
34.
35.
36.
37.
38.
39.
40.
SMSA
AKR, OH
ALB, NY
ALL, NJ
ANA, CA
ATL, GA
BAL, MD
BIR, AL
BOS, MA
BUF, NY
CHI, IL
CIN, OH-KY-IN
CLE, OH
COL, OH
DAL, TX
DAY, OH
DEN, CO
DET, MI
FOR, FL
FOR, TX
GAR, IN
GRA, MI
GRE, NC
HAR, CT
HON, HI
HOU, TX
IND, IN
JAC, FL
JER, NJ
KAN, MO-KS
LOS, CA
LOU, KY-IN
MEM, TN-AR
MIA, FL
MIL, WI
MINN, MN
NAS, TN
NEW, LA
NEW, NY
NEW, NJ
NOR, VA
PCNSCO
22.83
50.49
29.23
39.01
24.46
66.15
92.29
42.48
54.19
73.94
41.16
104.65
23.03
39.46
46.35
73.86
70.00
12.58
31.10
38.23
18.74
27.98
23.80
16.69
29.27
20.45
18.34
28.24
29.19
55.18
68.08
20.91
11.83
31.27
20.34
53.42
22.85
36.26
60.31
41.85
PCGSCO
116.79
147.02
128.49
132.39
119.42
156.45
184.03
137.98
147.52
166.21
135.74
196.22
117.90
136.25
140.00
168.57
160.00
125.00
128.61
127.65
110.95
123.51
143.22
98.25
123.98
117.12
113.42
133.33
129.98
159.27
160.82
121.69
116.72
83.33
115.21
148.06
116.63
138.78
155.09
127.75
                                       83

-------
TABLE IV-8 (Concluded)

41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
SMSA ' s
OKL, OK
OMA, NE-LA
PAT, NJ
PHI, PA-NJ
PHO, AZ
PIT, PA
POR, OR-WA
PRO, RI-MA
RIG, VA
ROC, NJ
SAC, CA
SAI, MO-IL
SAL, UT
SAN, TX
SAN, CA
SAN, CA
SAN, CA
SAN, CA
SEA, WA
SPR, MC-CT
SYR, NY
TAM, FL
TOL, OH- MI
WAS, DC-MD-VA
YOU, OH
PCNSCO
14.98
62.96
6.84
21.59
95.87
61.22
29.83
21.62
46.72
39.01
10.99
50.36
30.82
4.98
61.94
8.76
11.19
9.39
8.51
12.64
43.08
23.89
52.53
29.88
42.91
PCGSCO
117.16
154.81
105.22
116.85
186.98
156.18
133.80
117.45
142.66
124.58
110.86
144.73
118.46
93.63
157.48
106.04
124.76
104.23
112.52
109.81
135.22
127.34
145 . 74
127.23
135.07
          84

-------
TABLE IV-9. PER CAPITA NET AND GROSS SOILING DAMAGE COSTS ($) BY MEDIUM
            SMSA's, 1970


66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.

ALB
ANN
APP
AUG
AUS
BAK
BAT
BEA
BIN
BRI
CAN
CHA
CHA
CHA
CHA
COL
COL
COL
COR
DAV
DES
DUL
ELP
ERI
EUG
EVA
FAY
FLI
FOR
FRE
GRE
HAM
HAR
HUN
HUN
SMSA'S
, NM
, MI
, WI
, GA-SC
, TX
, CA
, LA
, TX
, NY- PA
, CN
, OH
, SC
, wv
, NC
, TN-GA
, CO
, SC
,GA-AL
, TX
, IA-IL
, IA
, MN-WI
, TX
, PA
, OR
, IN-KY
, NC
, MI
, IN
, CA
, sc
, OH
, PA
, WV-KY.OH
, AL
PCNSCO
30.
18.
28.
10.
16.
60.
9.
9.
8.
7.
38.
0.
41.
35.
42.
32.
9.
2.
35.
56.
28.
18.
55.
37.
28.
21.
10.
53.
20.
46.
20.
23.
22.
35.
42.
70
38
52
67
22
49
47
49
25
71
71
99
74
70
30
20
91
93
79
20
32
11
99
88
17
46
38
32
00
73
67
45
38
83
54
PCGSCO
122.15
109.83
115.16
98.02
111.82
153.80
98.25
105.38
103.96
104.37
133.33
86.84
140.00
129.58
140.33
120.76
93.19
90.79
121.75
152.07
128.67
116.60
136.49
130.30
126.76
120.60
86.79
140.85
115.36
140.19
114.67
115.04
122.14
133.46
99.56
                                    85

-------
TABLE IV-9 (Continued)

101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
SMSA
JAC, MS
JOH, PA
KAL, MI
KNO, TN
LAN, PA
LAN, MI
LAS, NV
ALW, MA-NH
LITT, AK
LOR, OH
LOW, MA
MAC, GA
MAD, WI
MOB, AL
MON, AL
NEW, CN
NEW, CN
NEW, VA
ORL, FL
OXN, CA
PEN, FL
PEO, IL
RAL, NC
RE A, PA
ROC, IL
SAG, MI
SAL, CA
SAN, CA
SAN, CA
SCR, PA
SHR, LA
SOU, IN
SPO, WA
STA, CN
STO, CA
TAG, WA
TRE, NJ
TUG, AZ
TUL, OK
UTI, NY
PCNSCO
37.45
38.401
8.42
37.50
41.56
20.90
38.83
13.36
19.81
93.77
3.29
23.79
18.97
38.46
33.83
9.83
10.10
5.14
20.56
62.13
39.09
22.22
6.14
51.69
40.07
52.73
43.20
50.76
52.68
99.57
40.34
20.36
37.63
7.77
9.66
31.63
17.11
36.36
27.46
26.18
PCGSCO
124.71
133.08
99.50
135.00
135.00
111.90
137.73
112.93
118.89
176.65
91.55
116.99
114.14
126.79
125.37
108.99
101.92
94.18
117.52
180.51
127.98
120.18
99.56
152.36
134.56
139.09
129.60
147.35
153.66
194.87
134.58
116.79
137.98
104.85
108.97
123.84
112.17
133.24
130.61
111.47
        86

-------
                          TABLE IV-9 (Concluded)
.	SMSA	PCNSCO	PCGSCO

141.  VAL, CA                                    9.24       104.82
142.  WAT, CN                                   23.44       118.18
143.  WES, FL                                   12.32       122.35
144.  WIC, KS                                   67.35       163.75
145.  WIL, PA                                   58.48       157.31
146.  WIL, DE, NJ, MD                           51.90       142.28
147.  WOR, MA                                   18.02       112.79
148.  YOR, PA                                   27.27       125.45
                                    87

-------
     A summary of the net and gross soiling damages by cleaning operations is
contained in Table IV-10. The total net soiling damage as a result of falling
suspended particulates for the 148 SMSA's in 1970 amounts to $5 billion. This
damage figure is far smaller than the $11 billion and $30 billion national esti-
mate extrapolated from the per capita damage figures reported respectively in
the Mellon Institute study and the study by Michelson and Tourin. As noted
earlier, the validity of the $11 billion and $30 billion estimates is  seriously
undermined by the assumptions used in the extrapolation technique. Regarding
the gross soiling damage costs, New York, Chicago and Los Angeles had the high-
est damages among the 148 large SMSA's, about $1.6 billion, $12 billion, and
$1.1 billion, respectively, partially because of the relatively high suspended
particulate levels and a large number of household units in these three cities.
Total gross soiling damage which is the sum of soiling damages attributable to
air pollution and other factors amounts to $17.4 billion per year for the 148
SMSA's.
           TABLE IV-10. NET AND GROSS SOILING DAMAGE COSTS IN 148
                        SMSA's BY CLEANING OPERATIONS,  1970


                        Net                Gross               Net/Gross
                  Soiling  Damages      Soiling Damages      Soiling Damage
Tasks
1
2
3
4
5
6
7
8
9
Total
(million $)
11.8
558.8
454.1
1,956.3
13.6
925.7
349.2
275.1
488.9
5,033.0
(million $)
12.8
9,884.9
204.0
3,584.1
30.6
1,227.9
691.6
1,079.2
652.0
17,367.1
Cost
0.91
0.05
0.45
0.55
0.44
0.75
0.50
0.25
0.75
0.28

       Note: — individual figure may not add to totals due to rounding.
                                    88

-------
00
                                                               TABLE IV-11


                                              TABLE IV-11.  SOILING  ECONOMIC  DAMAGE FUNCTIONS£iW
Dependent
Variable
GSC01
GSC02
GSC03
GSC04
GSC05
GSC06
GSC07
GSC08
GSC09
TGSCO y
MANFV
66.18
(2.02)*
42.9
(1.45)*
957.6
(25.9)*
17.2
(0.42)*
144.3
(3.90)*
6.12
(0.17)*
3.29
(0.08)*
4.92
(0.15)*
3.23
(0.08)*
78.9
(2.3)*
TSP
1,128.83
(147.36)*
-108.5
(105.5)
6,802.6
(1,887.5)*
160.0
(33.1)*
1,031.6
(284.2)*
84.0
(12.5)*
27.7
(6.4)*
9.07
(10.9)
44.6
(6.5)*
226.4
(166.9)
PCOL
2,377.55
(1,288.8)
2,252.2
(915.7)*
42,426.0
(16,507.0)*
727.4
(29?.. 3)*
6,384.9
(2,485.6)*
236.9
(109.9)*
142.0
(56.3)*
223.0
(94.9)*
126.9
(579.4)*
3.766.0
(1,460.0)*
RUM
-995.17
(538.89)
670.6
(383.4)
-14,677.0
(6,902.0)*
-262.6
(122.4)*
-2,210.7
(1,039.3)*
-92.6
(45.9)*
-50.3
(23.5)*
-74.8
(39.7)
-48.9
(24.2)*
-1,219.8
(610.7)*
DTS
271.69
(193.52)

1,946.0
(2,478.0)
42.577
(43.8)
294.6
(373.2)
20.9
(16.5)
7.6
(8.4)

11.2
(8.6)
90.9
(219.3)
PDS
-1.78
(3.81)
2.2
(2.7)
12.4
(48.9)

1.85
(7.36)
-0.083
(0.325)
0.025
(0.166)
0.17
(0.28)

2.3
(4.3)
PAGE
764.7
(1,940.9)
2,401.1
(1,392.6)
32,626.0
(24,859.0)
501.9
(439.4)
4,898.9
(3,743.2)
116.6
(165.6)
102.2
(84.8)
212.9
(144.4)
60.4
(87.0)
3,432.3
(2,199.5)
a
-100,181.3
(46,192.6)
5,400.0
(32,763.0)
652,046.0
(591,650.0)
14,825.0
(10,445.0)
98,799.0
(89,087.0)
-7,565.5
(3,941.2)
-2,607.0
(2,018.7)
877.7
(3,396.3)
4,047.7
(2,070.1)
-25,621.0
(52,347.0)
R2
0.92
0.89
0.93
0.93
0,93
0.93
0.93
0.91
0.93
0.92
         a/  All  coefficients and standard errors  are reduced by  a  factor of 103, except equation GSC01, GSC03 and GSC05. The
         ~    standard errors are the values  below the  coefficients, and * indicates that the coefficient is significant at
              the  I percent level.
         b/  TGSCO  denotes the total gross soiling cost  for  the  ith cleaning task, and TGSCO is the sum of GSCOi over i, i = 1, 2,

-------
     The regional soiling damage costs and the national damage cost deduced in
this study should be used, however,  only as crude estimates. There are uncertain-
ties embodied in the two major assumptions:  (1) the physical damage functions
for the variety of cleaning tasks estimated on the basis of the Philadelphia
study are "representative" of the physical damage functions of the 148 SMSA's;
(2) the unit  market value figures obtained in the Kansas City area are applicable
to other SMSA's.

     In order to develop "average" soiling economic damage functions for each
of the nine cleaning tasks which can be used for prediction and control purposes,
the individual metropolitan damage costs were regressed against not only the
SO , but also to several socioeconomic, demographic, and climatological charac-
teristics of different regions. The independent variables include MANFV (value
of manufacturing), PCOL (percentage of persons 25 or older who have completed
4 years of college), RHM  (relative humidity), DTS (number of days with thunder-
storm), PDS  (population density), PAGE (percentage of population 65 or older)
and TSP- The inclusion of these variables is to account for the variations in
educational  level, economic and age structure and density differentials among
the study regions. The stepwise regression technique was used with inputs from
the 148 sample observations for the purpose of estimating the economic damage
functions. The regression results are summarized in Table IV-11. It is note-
worthy that  all the coefficients of TSP are of correct signs except the one in
the second regression equation. Since the partial correlation coefficient be-
tween GSC02  and TSP is positive and equal to 0.18, the negative coefficient ob-
tained for TSP in the regression equation may be attributable to multicolinearity
between TSP  and other independent variables or other econometric problems or
data deficiency.

     It is interesting to note that aside from total suspended particulates PCOL,
RHM, and MANFV are significant factors in determining the household soiling costs.
While the effect of educational level on soiling adjustment cost is ambiguous
£ priori , relative humidity is likely to have a cleansing effect which reduces
the soiling  costs.

     The soiling economic damage functions derived in this study are useful to
policymakers at either the local or national level in estimating the marginal
and average  benefits of implementing a particular pollution abatement program.
The responsiveness of gross soiling damages for a particular cleaning task to
changes in climatological, demographic, and socioeconomic variables and the con-
centration level of suspended particulates can be easily estimated. The partial
elasticity of the gross soiling costs of,  say, cleaning Task 4, i.e., cleaning
Venetian blinds and shades, with respect to suspended particulate  level, can
be estimated by
             Esc4jTSP = |1SC41 .  TSP
                        o-(TSP)    SC4
                                      90

-------
where  8(SC4)/B(TSP)  is the coefficient of TSP in the soiling economic damage
function with   GSC04   as the dependent variable, and is equal to 160,000. TSP
and  SC4  are respectively the mean values of total suspended particulates and
of soiling damage cost associated with cleaning Venetian blinds and shades. Given
TSP  = 94.5 M-g/™3> and SC4  = $24.2 million,
               ESC4,TSP = °'16 X <94-5/24-2> = °-62


Thus, for every 1 percent reduction in suspended particulate level, the soiling
damage cost of cleaning Venetian blinds would decrease by 0.62 percent, holding
other characteristics unchanged. Thus, if the suspended particulate level in
the air is lowered by 9.45 p,g/m3 from 94.5 to 85.5 l-lg/m^ (i.e., 10 percent re-
duction), gross soiling damage cost associated with cleaning Venetian blinds
alone, would reduce, on the average, by $1.5 million from $24.2 to $22.7 million
nationwide. Of particular policy interest is the estimation of possible benefit
in terms of the reduction in the overall soiling damage cost as a result of a
pollution control program. Note that the coefficient of TSP in the overall soil-
ing economic damage function is 226,400 and the mean value of overall soiling
damage cost is $117.3 million.
             E       = 226400 x (94.5/117.300000) = 0.18
              SC,TSP


Thus, if the suspended particulate level is lowered, on the average, by 10 per-
cent, from 94.5 to 85.5 [j,g/m , overall gross soiling damage cost would reduce
by 1.8 percent or by $2.1 million, from $117.3 million to $115.2 million.
                                     91

-------
                                  SECTION V

                          MATERIAL AND AIR POLLUTION
PROBLEMS AND OBJECTIVES

     The damaging effects of air pollution to materials have been well recognized
by the Air Pollution Control Office of the Environmental Protection Agency for
some time. Effects of air pollution on materials range from soiling to chemical
alteration. Corrosion of metals has attracted the most attention, while many
other important areas were found to have been largely neglected. Most materials
exhibit a high degree of chemical resistance to oxides of nitrogen, while sulfur
dioxides were found to seriously attack about a third of the materials. And
while some materials (such as glass) are highly resistant to chemical attack
by most air pollutants, certain plastics and metals are highly susceptible to
damage by a number of different commonly encountered air pollutants.

     Among the adverse effects of air pollution on material included are the
corrosion of metals, the deterioration of rubber, the fading of paint and soil-
ing of materials. Many external factors influence the reaction rate between pol-
lutants and materials, with moisture the most important in accelerating corro-
sion. Inorganic gases are likely to cause tarnishing and corrosion of metals;
they can attack various building materials such as stone, marble, slate, and
mortar and may deteriorate a variety of natural and synthetic fibers.

     The most noticeable effect of particulate pollutants is soiling of the
surfaces on which they are deposited. They may also act as catalysts increasing
the corrosive reactions between metals and acid gases. Additional damages to
surfaces and textiles are incurred by the wear and tear imposed by the extra
cleaning made necessary because of particulate soiling. The true economic dam-
age to materials caused by air pollution is difficult to ascertain because of
the difficulty of distinguishing between natural deterioration and deterioration
caused by air pollution and the uncertainty regarding indirect costs of early
replacement of materials worn out by excessive cleaning.

     Some of the material damage estimates attributable to air pollution in this
country are as follows: In a pilot study by Uhlig (1950), the total corrosion
bill was estimated at $5.4 billion, though air pollution was merely implicated
as a causal agent of corrosion. Uhlig's estimate was updated and estimated by
the Rust-Oleum Corporation (1974) to be $7.5 billion in 1958. Stickney, Mueller
and Spence (1971) estimated that the pollution damage cost of rubber to U.S.
consumers amounts to at least $398 million. Haynie (1973) estimated a value of
$1.4 billion for the cost of corrosion of galvanized steel. The total damage
cost due to air pollution inflicted on textiles and fibers was estimated to be
$2 billion annually by Salvin (1970).
                                    92

-------
      Robbins (1970)  of  Stanford Research Institute conducted  one  of  the  first
major material  damage  studies  and found that the air pollution  damage with  re-
 spect to  electric  contacts was not as  serious as originally estimated. Two
 types of  major  costs were investigated. They are the direct cost  associated
with  the  plating of  contacts with precious metals and  the  indirect cost  incurred
 because of  the  preventive measures of  air conditioning and air  purification.
 It was estimated that about $65 million is spent annually  on  electric contacts
 because of  air  pollution.

      Economic impact of air pollution  on electric components  was  estimated  by
 International Telephone and Telegraph  (ITT) Electro-Physics Laboratories  (1971).
 The damage costs to the following electric components were  estimated:  semi-
 conductor devices, integrated  circuits, television picture tubes, connectors,
 transformers, relays, receiving tubes, and crystals. Total damages to these
 categories  amounted  to $15.5 million.

      A most comprehensive study on pollution damage on materials was conducted
 by Midwest  Research  Institute  (Salmon, 1970)- The MRI  study presented a systematic
 analysis  of all of the physical and chemical interactions  between materials, pol-
 lutants and environmental parameters.  Fifty-three economically  important mate-
 rials which represent about 40 percent of the economic value  of all materials
 exposed to  air  pollution were  identified and selected  for  the study. An esti-
mated $100  billion in added cleaning costs would be necessary to keep these mate-
rials in  polluted areas as clean as they would be in a nonpolluted environment,
while deterioration  of these materials causes yearly direct damage losses of
approximately $4 billion. Paint and zinc are the two materials most affected
by both soiling and  deterioration caused by air pollution, accounting for more
than  half of the total losses  in each  category. Based  on the MRI estimate,
Barrett and Waddell  (1974) estimated an annual material deterioration damage
in the United States to be $4.75 billion.

      Some of the methodological procedures for estimating material effect was
critically  reviewed  by Gillette and Upham (1973). It is generally understood
that  in order to develop reasonable estimates of pollution damages, the follow-
ing information is very relevant: (1)  geographical and temporal distribution
of air quality  levels and receptors' exposure to various pollution levels;  (2)
physical  damage functions on important receptors; and  (3)  data on other socio-
economic, demographic and environmental factors on a regional basis.

      Pollution  level varies within any given SMSA. In  the  absence of population-
at-risk information, it is usually assumed that the entire SMSA population was
exposed to  the  same  pollution  level as recorded by the  station(s) which in all
probability is  (are) located in the central city of the SMSA. Cost estimates
derived under this assumption  tend to overestimate the actual damage due to air
pollution,  since the pollution concentration level is  likely to be higher in
the central city than in the suburban areas.

     Physical damage functions relating material damage to air pollution have
recently  been derived in a series of in-house experiments. Haynie and his
                                     93

-------
associates (1974, 1975) obtained such physical dose-response relations sep-
arately for different kinds of steels, zinc, oil-base house paint and selected
fabrics. Economic damage functions for materials which translate material phys-
ical loss into monetary terms, however, are still lacking. Although some damage
estimates for certain types of materials at the national level are now avail-
able, detailed regional cost estimates on material damages due to air pollu-
tion are virtually nonexistent. Since the information on such regional damage
costs of air pollution is indispensable for providing guidance for establishing
pollution controls, it is imperative to develop a set of comparable damage cost
estimates for each as well as a system of economic damage functions on differ-
ent types of materials.

     This section represents a first exploratory effort to estimate not only
urban material damages attributable to air pollution for all 148 SMSA's with
population greater than 250,000, but also the "average" air pollution damage
functions on materials for this country. This section contains the following
subsections: A Theoretical Framework, Exposition of Methodology, Regional Mate-
rial Damage Costs, Economic Damage Functions, and A Summary of Material Phys-
ical Damage Functions.
 A THEORETICAL FRAMEWORK

      A theoretical framework is developed in this section for defining and de-
 veloping urban  economic costs and economic damage functions of materials. The
 economic costs  of a material are defined as the decrease in the values of this
 particular material as a result of increased contamination in the environment.
 An economic  damage function of material which relates the economic costs to a
 host  of relevant variables including pollution, socioeconomic, demographic and
 climatological  variables is also estimated.

      It is noteworthy that materials generally do not directly affect an individ-
 ual's  utility or preferences. Thus, materials can be regarded as "pure" inter-
 mediate commodities which are differentiable from the traditional interindustry
 flows.  The pure intermediate commodities utilized the primary inputs, i.e., la-
 bor and capital, in their production and are themselves solely utilized as in-
 puts  in the  production of "final" commodities which enter into one's utility
 function. Interindustry flows, however, refer to those commodities which are
 intermediate inputs to be used in other industries as well as final outputs
 for consumers.

     The degree of the effect of air pollution on materials depends on a number
 of factors:  (1) the extent of the reduction in the normal service or use life
 of material; (2) the frequency of maintenance and preventive measures on the
 part of users;  and (3) the changes in the quality and quantity of the services
 rendered by  the product which contains the materials being affected by air pol-
 lution.


                                     94

-------
      Let  D.  represent  the  level of physical deterioration of the ith type of
material.  &L. will  be  the  service life, SP. the service performance, ME  the
maintenance  effort  of  the  same ith type of material, and  A  the air pollution
concentration level  to which the material is being exposed.

      Thus, we can write
                 D± = D   [SLXA), SPi(A), MEi(A); e]
                                      (V-l)
     Equation  (V-l) states that the deterioration of the ith type of material
is functionally related to SL., SP.  and ME., directly. Each of the explanatory
variables is, however, being influenced by the prevailing pollution levels.

     The plausible signs of the partial derivations are as follows:
     Thus,
           o-(SL) d(A)
.
'
 BO   . d(SP)
a(SP)   d(A)
                                                .
                                                ' *
                           a(ME)    d(A)
                                                                     > 0.
     Assuming noninteractions among the independent variables, total physical
damage of material as a result of an increased pollution is expressed asi—'
        dD =
5D
a(sL)
d(SL)
d(A)
+ 5D
a(sp)
d(SP)
d(A)
+ 5D
9 (ME)
, d(ME)
d(A)
                             d(A)
                                                                     (V-2)
     Note that changes in  SL  and  SP  of the product represent direct damages
attributable to pollution, whereas pollution-induced changes in  ME  are indirect
damages due to pollution.

I/ The assumption of noninteraction is made for the sake of simplicity. In the
     real world, it is observed that service life, performance and maintenance
     effort are interrelated. Increased maintenance should also increase service
     life and performance.

-------
     A simplified and yet commonly adopted formulation of material  physical  dam-
age function is written as follows:
                   D.  = D.(A, RHM;  u)                     (V-3)


where  RHM  is relative humidity, D  and  A  are the same as in (V-l) and  u
is the error term.

     In order to develop a theoretical framework for estimating economic damage
costs of materials, let us assume for the sake of simplicity, but without loss
of generality, that there is only one type of material. The initial endowment
of this material in a given urban area is M . Further, it is assumed that the
material stock grows at an exogenously determined rate of  r  over the planning
time horizon.JV Thus, total  stock of this material in the absence of air pollution
in the area at time  t  is given by:
                          t

                     g
M  =  / M  er dt                         (V-4)
     J    o
                         t=0
     If the area in question is subject to an air pollution level which is above
the threshold level, and if the material depreciates at a rate  i  because of
the air pollution, then the net existing material stock at the time  t  is given
by:
                    M  =  /* M  e      dt
                     n   */    o
     Note that i = dD/D and D = D(A; RHM) with 9D/d(A) > 0. dD/d(RHM)  > 0.

     The economic damage of this material (ED) is defined to be that portion
of the material loss attributable to air pollution evaluated at the prevailing
market prices of the material. Let  P   be the market price of the material,
which is determined by the supply ana demand conditions for this material.
JL/ A more realistic approach is to consider that the growth of stock of a mate-
     rial within an area is endogeneously determined. It is determined by the
     need for that material within an area and the ability to acquire it. One
     would expect that the stock growth rate to be a function of population
     growth rate, per capita income growth rate and the change in demand for
     that material with respect to replacement material.

-------
Thus,
                      f      X.        *~V*    ^*-i-r I i> \fjLmj.\.iU.J./         .    .
              ED = P  (M  -M  ) = P   /Me              dt     (V-6)
                    m g  n     m S   o
                       &          t=o
     It may be remarked that  (V-6) reflects the economic damage associated with
the air pollution level through a change in demand for the material. It  does
not reflect the economic damage associated with increased flow of the material
through the area caused by pollution-induced decreased service life.

     It follows from (V-6) that a general economic damage function for material
can be expressed as

                    ED = ED(P  , A, RHM; u)                   (V-7)
                             m
     Those socioeconomic and other variables which influence  P   could be in-
cluded in the general economic damage function in addition to  A  and  RHM.
EXPOSITION OF METHODOLOGY

     Ideally, the information on the distribution of materials, of pollutants,
the value of the products made from the materials, the service life of the prod-
ucts in the absence of pollution and the physical dose-response function should
be gathered in order to accurately assess economic costs of material deteriora-
tion due to excessive air pollution. Empirical data on the distributions of both
materials and pollution are unavailable for most urban areas. However, sketchy
estimates for product values and the service life have been derived by Fink et
al. (1971) by resorting to the annual production figures in the Standard Indus-
trial Classification statistics and product useful life statistics issued by
Internal Revenue Service. Dose-response functions for a variety of materials
have recently been estimated via the technique of in-house controlled experi-
ment s.

     In the absence of the relevant material and pollution distribution data,
an alternative "top-down" method is developed in this section to derive a con-
sistent set of urban economic costs of material damage as a result of air pol-
lution. The existing valuable information on national material damage and the
dose-response functions were obtained through literature survey. The national
damage estimates were then allocated down to various SMSA's by utilizing the
dose-response functions and other relevant regional data.

     For the sake of illustration, but without loss of generality, let the phys-
ical dose-response function for the ith type of material be written as:

                                     97

-------
                       i    i  '


where  D, A  and  C  denote, respectively, the physical damage, the air pollu-
tion level and climatological conditions.

     Given the physical damage function and the national damage estimates, the
regional damage costs for the ith type of material can be estimated by using
the following formula:

                               D. .            SE.
                                 ij              j
                RED   = NED.  • 	  • 	     (V-9)
 I
j/
                                      n        SE.n
                                                 J
where  RED. .  and  NED.  are, respectively, the regional and national economic
damage costs for the ith type of material. SEj  stands for the relevant socioeco-
nomic characteristics which are thought to affect material damages, e.g., manu-
facturing  establishments and  D-M is the dose-response relation for the ith type
of material. The subscript  j  denotes the jth SMSA.

     Substituting  (V-8) into (V-9) yields:
            RED   =NED.  .  W  V     •  Jfj	           CV-10)
               1J      X  SD..(A.,  C.)
                            ij   j    j        SSE^ n
      Equation  (V-10) can be used to  derive the regional  economic  cost  for  the
 ith  type  of material. The  data on  NED   is available from an earlier  material
 damage  study conducted by  Salmon (1970) at Midwest Research  Institute,  and  AJ ,
 Cj   and  SEj   can be respectively  secured from the Air Quality Data, published
 by the  Environmental Protection Agency; Local Climatological Data, published
 by the  National Oceanic and Atmospheric Administration;  and  1972  County and  City
 Data Book. Substituting the values for  NED-^, A j, Cj, and  SEj  into equation
 (V-10), a series of consistent estimates for material damages for various  SMSA's
 due  to  air pollution is obtained.

      According to the earlier MRI  study, material damage as  a result of air  pol-
 lution  can be  categorized  into two major effects:   (1) soiling effects attribut-
 able to particulate pollutants; and  (2) chemical effects attributable  to gaseous
 pollutants. The national soiling cost  (SC) and deterioration cost (DC) of  various
 materials were estimated with the  aid of the following formulas:

                                     98

-------
                        SC = SIF . Q                         (V-ll)

                        DC = DIP . Q                         (V-12)

                         Q = P . N . F . R                   (V-13)
where  SC  represents material soiling costs, SIF  the soiling interaction fac-
tor, Q  in-place unprotected material value, DC  material deterioration cost,
DIF  deterioration interaction factor, P  annual production value of the mate-
rial, N  economic life of the material based on usage, F  weighted average fac-
tor for the percentage of the material exposed to air pollution, and  R  labor
factor reflecting the in-place or as-used value of the material.

     The rate of soiling interaction factor,  SIF, is computed by complex for-
mulas which are different for fibers and nonf ibers .JL/ The rate of deterioration,
DIF, is computed by estimating the difference between the deterioration rate
in polluted and unpolluted environments divided by the average thickness of the
material.

     Finally, it is noted that two methods are generally feasible for estimating
regional damage costs. The first method is the "top-down" technique which is to
allocate via weighting and adjusting schemes a national damage estimate down
to various regions. The second method is the "bottom-up" technqiue which involves
direct estimation of the regional damages. The bottom-up method, which incorpo-
rates uncertainties of the assumptions into regional estimates generated,  was
used to derive the damage costs for human health and household soiling adjustment
in the preceding three sections. The top-down technique was employed, however,
in this study to estimate material damage costs because of the lack of distribu-
tion information about materials and pollutants and the technical difficulties
encountered in direct estimation of the regional damages.
REGIONAL MATERIAL DAMAGE COSTS

     This section is concerned with estimating regional material damages by us-
ing the top-down method delineated above. In view of the fact that there are
virtually infinite categories of materials, and the fact that zinc and paint
are most important from an economic point of view, only these two materials were
selected for this study. The damage costs of zinc and paint account for over 50
percent of the total economic damage losses of the 53 economically important
materials selected in the earlier MRI study (Salmon, 1970).
I/ The formulas are:  SIFfibers = °-10 Af/Rw and SIFnonfibers = O.lOAf/Rwpt
     where  Af  is the increased frequency of cleaning due to pollution,  R
     the labor factor,  w  the material price per pound, p the density and  t
     the average thickness.
                                     99

-------
     The soiling and deterioration costs of paint and zinc, and the percentage
of these two costs in terms of total costs, are summarized in Table V-l-i/
These figures are admittedly artificial because they were calculated on the
assumption that the material would be maintained completely clean at all times.
In practice, each individual will have an acceptable level of soiling for each
material.
        TABLE V-l. SOILING AND DETERIORATING COSTS OF PAINT AND ZING




Paint
Zinc
53 Materials
Soiling
Cost (SC)
(billion $)
35.0
24.0
100.0


SC/Total SC
0.35
0.24
0.59
Deterioration
Cost (DC)
(billion $)
1.2
0.8
3.8


DC/Total DC
0.31
0.20
0.51

     The physical dose-response functions  for zinc  and oil-base  house paint  are
 obtained from two recent studies  by Haynie  and Upham (1970),  Spence,  Haynie and
 Upham (1975).
      Zinc - Corrosion = 0.001028  (RH -  48.8)  SO                           (V-14)

      Paint - Erosion = 14.32 + 0.01506  SO +  0.3884 RH                    (V-15)


     Recalling equation (V-10),  and  substituting  the above  physical  dose-response
 function for zinc,  and the  relevant  socioeconomic  data  into  equation  (V-10)
 yields the estimates of soiling and  deterioration  cost  of  zinc  for  the  jth SMSA.
 Thus,
                                O.OOr028(RH.-48.8)50              ME
         zinc.      zinc  * —•	          1        *
             j              148
                            S   (0.001028(RH.-48.8)SO   /148)    ZME./148
                            j=l             J         j         j  j
                                    /148
                                   J  S  POP ,
                                   V    J
   *
NED .    = NED .    Jj  E  POP /POP   .
   zinc      zinc 1 .       j'   us I                       (V-17)
y The estimates are taken from Table XII and Table XIII of the MRI research
     report, "System Analysis of the Effects of Air Pollution on Materials,"
     Kansas City (January 1970).
                                   100

-------
where  NED*  .    is the total damage cost of zinc over the 148 SMSA's included
           7 1 T) f
for the present study, RHj  and  MEj  the relative humidity and manufacturing
establishment in the jth SMSA,  POPj  and POPus  represent, respectively, the
population in the jth SMSA and in the whole country.

    The earlier MRI study estimated the national soiling and deterioration dam-
age costs of zinc to be $24 billion and $778 million, respectively. Given the
ratio of 0.63 which represents the ratio of the 148 SMSA population to the na-
tionwide population in 1974, the soiling and deterioration damage costs of zinc
for the 148  SMSA's are calculated as follows:
                  $24 x 0.63 = $15.12 billion (soiling)

                  $778 x 0.63 = $496 million (deterioration)
     A threshold level of zero  (J,g/m3 for SC>2 is implicit in the physical dose-
 response functions for zinc and paint. Thus, a zero threshold level is used in
 estimating the regional damage costs of materials. It should be noted that the
 value of RED*z£nc as calculated from the weighting scheme expressed by (V-16)
 can be greater than NED*z:j_nc. Thus, the damage costs for each SMSA are further
                                 J 148
 adjusted by the ratio (REDzj;nCj j  .J5. REDz^nCj j) to preserve the equality be-

 tween the sum of the regional damage costs obtained via equation (V-16) over
 the 148 SMSA's and the total damage costs evaluated for the same 148 SMSA's,
 i.e., NED*.

     Similarly, the soiling or deterioration damage costs of paint for the jth
 SMSA are computed by using the following formula:
                                  (14.3 + 0.01506 S02  + 0.3884 RH.)
      RED         = NED*     .	j	J
         paint, j      paint   ^
                                S (14.3 + 0.01506 SO   + 0.3884 RH.)/148
                               J=l                   J            J

                     HU.         YP.
                       J           J
SHU /148  * EYP./148
j  j        j
                                    .
                                   J                               (V-18)
              NED*     = NED   .
                 paint      paint
                     POP /POP                 (V-19)
                        j    us
                                     101

-------
where  NED*paint  is the total cost of zinc over the 148 SMSA's,  HUj  and  YPj
are housing units and per capita income in the jth SMSA's. POP, S02>  RH,  NED
and  RED  are the same as those in equation (V-15).

     The nationwide soiling and deterioration damage costs of paint were esti-
 mated in 1970 to be $35 billion and $1.2 billion,  respectively. Thus,  applying
 the population ratio of 0.63 to these two estimates, NED* is calculated to be
 $22 billion and $753 million for soiling and deterioration,  respectively.  The

 damage estimates are further adjusted by RED        ./  (S  RED        )  such
                                            pirint jj    j
 that  the sum of the  damage  costs  over  the  148  SMSA's  equals  the  total  damage
 cost  evaluated for these  SMSA's by  using (V-19).

     The soiling and deterioration damage costs of zinc and paint for the 65 large
 SMSA's and the 83 medium SMSA's are computed by using equations (V-16) through
 (V-19). The results of the regional damage costs are summarized in Tables V-2
 and V-3. Since the national damage figures employed are extremely large because
 they were estimated on the stringent assumption that the materials would be main-
 tained completely clean at all times,  it is not surprising that the top-down
 method yields relatively large damage figures for the study regions. An examina-
 tion of the tables reveals that Chicago scores the highest damage costs on zinc
 among all the 148 SMSA's in both the soiling and deterioration damages (an an-
 nual soiling damage of $1.7 billion and deterioration damage of $57 million in
 1970). However, regarding the paint damages, New York City,  which had an annual
 soiling and deterioration damage cost  of paint of $2.3 billion and $79 million,
 respectively, surpassed all the SMSA's included in this study.

     It  should be noted again that the assumptions made in deriving the national
 damage figures and the physical damage functions in the earlier studies are in-
 herited by this study, especially the theoretically maximum national damage esti-
 mates in both categories of soiling and deterioration. Should there be changes
 in these assumptions, the estimates developed and presented in the table would
 be modified accordingly. Thus, the results presented in this section are only
 suggestive and tentative. Given the tentativeness and experimental nature of
 the methodological and statistical procedures, and the degree of uncertainty
 associated with the estimates, a great deal of caution should be exercised in
 using the product of this research.
 ECONOMIC DAMAGE FUNCTIONS

     In order to develop marginal equivalent economic damage functions which can
 be used for damage (benefit) prediction and for designing pollution control
 strategies, the economic costs of material soiling and deterioration are re-
 gressed not only against S02 and relative humidity, but also against other rele-
 vant socioeconomic and climatological variables. The stepwise regression tech-
 nique was used with inputs from the 148 sample observations for estimating the
 economic damage functions. The regression results for soiling and deterioration
 damages of zinc and paint are presented in Table V-4. Consistent with a priori
 expectation, the coefficients of  ME, S02, TSP, RH, HU  and  YP  are all posi-
 tive, and the coefficient of  SUN  has ambiguous signs depending on the type
                                      102

-------
                     TABLE V-2. MATERIAL DAMAGE BY LARGE SMSA's,
                                (in million $)
1970
I
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.
33.
34.
35.
36.
37.
38.
39.
40.
•arge
AKR,
ALB,
ALL,
ANA,
ATL,
BAL,
BIR,
BOS,
BUF,
CHI,
CIN,
CLE,
COL,
DAL,
DAY,
DEN,
DET,
FOR,
FOR,
GAR,
GRA,
GRE,
HAR,
HON,
HOU,
IND,
JAC,
JER,
KAN,
LOS,
LOU,
MEM,
MIA,
MIL,
MINN
NAS,
NEW,
NEW,
NEW,
NOR,
SMSA's
OH
NY
NJ
CA
GA
MD
AL
MA
NY
IL
OH-KY-IN
OH
OH
TX
OH
CO
MI
FL
TX
IN
MI
NC
CT
HI
TX
IN
FL
NJ
MO-KS
CA
KY-IN
TN-AR
FL
WI
, MN
TN
LA
NY
NJ
VA
Soiling Damage
Cost of Zinc
(SDCZ)'
108.0
54.0
16.5
62.6
25.2
190.0
12.9
289.0
32.8
1,770.0
134.0
1,110.0
91.5
15.3
125.0
--
910.0
8.8
11.9
144.0
29.4
18.8
19.1
2.3
28.6
79.3
2.1
128.0
122.0
730.0
171.0
17.3
24.6
178.0
32.5
12.6
25.6
1,040.0
94.9
10.3
Deteriorating Dama
Cost of Zinc
(DDCZ)
3.533
1.751
0.537
2.029
0.818
6.166
0.421
9.395
1.064
57.600
4.372
36.000
2.969
0.496
4.080
--
29.500
0.284
0.387
4.680
0.955
0.611
0.621
0.078
0.929
2.572
0.068
4.152
3.956
23.600
5.545
0.563
0.800
5.799
1.054
0.411
0.832
34.000
3.077
0.335
ige Soiling Damage
Cost of Paint
(SDCP)
107.0
121.0
85.6
279.0
224.0
320.0
98.8
480.0
219.0
1,350.0
225.0
392.0
153.0
273.0
145.0
167.0
741.0
158.0
127.0
94.0
81.9
86.0
140.0
83.7
338.0
192.0
70.8
101.0
229.0
1,530.0
127.0
94.5
239.0
247.0
307.0
80.8
147.0
2,310.0
327.0
81.4
Deteriorating Damage
Cost of Paint
(DDCP)
3.684
4.141
2.925
9.542
7.667
10.900
3.374
16.300
7.503
46.300
7.716
13.300
5.239
9.343
4.952
5.728
25.300
5.414
4.352
3.217
2.798
2.937
4.792
2.860
11.500
6.563
2.418
3.475
7.824
52.300
4.345
3.227
8.164
8.435
10.500
2.759
5.035
79.100
11.100
2.781
Note:
       -- denotes damage of value less than $0.5  million.
                                           103

-------
                                TABLE V-2 (Concluded)
Large SMSA's
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
OKL,
DMA,
PAT,
PHI,
PHO,
PIT,
FOR,
PRO,
RIC,
ROC,
SAC,
SAI,
SAL,
SAN,
SAN,
SAN,
SAN,
SAN,
SEA,
SPR,
SYR,
TAM,
TOL,
WAS,
YOU,
OK
NE-IA
NJ
PA-NJ
AZ
PA
OR-WA
RI-MA.
VA
NY
CA
MO-IL
UT
TX
CA
CA
CA
CA
WA
MC-CT
NY
FL
OH -MI
DC-MD-VA
OH
Total
SDCZ
2.
13.
101.
520.
--
504.
36.
137.
18.
57.
__
308.
--
1.
12.
9.
57.
26.
143.
15.
40.
13.
28.
38.
152.
10,114
6
3
0
0

0
3
0
7
1

0

8
1
8
3
7
0
3
4
3
4
4
0
.4
DDCZ
0.
0.
3.
16.
-
16.
1.
4.
0.
1.
_
10.
-
0.
0.
0.
1.
0.
4.
0.
1.
0.
0.
1.
4.
328.
083
434
292
800
-
300
780
51&
061
852
_
000
-
060
393
316
858
868
641
498
311
432
922
246
946
651
SDCP
106
85
271
767
90
386
179
140
78
151
114
411
58
97
198
224
725
199
293
75
104
175
108
578
84
18,272
.0
.1
.0
.0
.5
.0
.0
.0
.7
.0
.0
.0
.8
.2
.0
.0
.0
.0
.0
.6
.0
.0
.0
.0
.9
.3
DDCP
3.
2.
9.
26.
3.
13.
6.
4.
2.
5.
3.
14.
2.
3.
6.
7.
24.
6.
10.
2.
3.
5.
3.
19.
2.
624.
641
907
271
200
091
200
116
798
688
516
900
000
Oil
321
760
670
700
826
000
584
557
975
712
700
900
854
!-• • -
Note: —  individual figure may not  add to totals due  to  rounding,
                                104

-------
TABLE V-3. MATERIAL DAMAGE BY MEDIUM SMSA's,  1970
           (In million $)
Medium SMSA's
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
ALB, NM
ANN, MI
APP, WI
AUG, GA-SC
AUS, TX
BAR, CA
BAT, LA
BEA, TX
BIN, NY-PA
BRI, CN
CAN, OH
CHA, SC
CHA, WV
CHA, NC
CHA, TN-GA
COL, CO
COL, SC
COL, GA-AL
COR, TX
DAY, IA-IL
DES, IA
DDL, MN-WI
ELP, TX
ERI, PA
BUG, OR
EVA, IN-KY
FAY, NC
FLI, MI
FOR, IN
FRE, CA
GRE, SC
HAM, OH
EAR, PA
HUN, WV-KY.OH
HUN, AL
JAC, MS
JOH, PA
KAL, MI
KNO, TN
LAN, PA
LAN, MI
LAS, NV
LAW, MA-NH
LITT, AK
LOR, OH
SDCZ
..
572.0
0.0
2.3
3.3
-2.9
75.8
164.0
93.4
155.0
287.0
3.8
76.6
119.0
96.6
-.
7.7
20.5
5.9
34.6
28.1
31.4
--
131.0
41.0
104.0
4.6
0.0
67.2
--
30.2
22.2
6.7
91.4
38.8
2.7
2.4
45.2
46.9
62.6
118.0
--
162.0
12.3
17.8
DDCZ

18.500
0.000
0.076
0.107
-0.095
2.459
5.342
3.030
5.041
9.332
0.122
2.485
3.859
3.134
--
0.251
0.665
0.192
1.123
0.912
1.020
--
3.924
1.332
3.389
0.150
0.0
2.180
--
0.979
0.722
0.218
2.965
1.260
0.086
0.081
1.466
1.523
2.032
3.839
--
5.270
0.399
5.771
SDCP
29.6
46.2
39.7
26.5
44.8
35.9
38.5
49.1
47.9
70.3
59.0
32.1
33.6
65.1
43.9
24.8
34.8
26.5
112.0
60.3
52.0
43.9
24.3
37.0
33.2
34.7
19.1
76.4
46.5
46.7
37.6
32.3
62.7
34.8
31.1
28.7
32.3
32.3
55.3
45.0
61.0
29.7
42.3
46.2
37.9
DDCP
1.014
1.580
1.357
0.905
1.532
1.228
1.317
1.678
1.638
2.404
2.016
1.096
1.148
2.224
1.501
0.849
1.189
0.905
3.831
2.060
1.778
1.501
0.830
1.264
1.135
1.186
0.653
2.609
1.591
1.596
1.287
1.106
2.143
1.190
1.065
0.980
1.104
1.106
1.889
1.539
2.085
1.015
1.444
1.579
1.297
                    105

-------
                                    TABLE V-3 (Concluded)
Medium SMSA's
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
Total
LOW -MA
MAC, GA
MAD, WI
MOB, AL
MON, AL
NEW, CN
NEW, CN
NEW, VA
ORL, FL
OXN, CA
PEN, FL
PEO, IL
RAL, NC
REA, PA
ROC , IL
SAG, MI
SAL, CA
SAN, CA
SAN, CA
SCR, PA
SHR, LA
SOU, IN
SPO, WA
STA, CN
STO, CA
TAC, -WA
TRE, NJ
TUC, AZ
TUL, OK
UTI, NY
VAL, CA
WAT, CN
WES, FL
WIG, KS
WIL, PA
WIL, DE-NJ-MD
WOR, MA
YOR, PA

SDCZ
151.0
2.8
18.5
34.7
2.9
66.3
42.8
5.3
18.5
22.5
34.6
240.0
10.9
90.6
56.8
56.3
5.0
11.6
14.9
25.8
38.5
154.0
--
84.6
--
35.0
18.6
--
473.0
97.7
11.0
8.7
3.8
25.4
60.3
96.1
201.0
9.7
5 , 005 . 5
DDCZ
4.924
0.089
0.602
1.126
0.092
2.152
1.388
0.174
0.601
0.732
1.122
7.782
0.355
2.939
1.843
1.828
0.164
0.377
0.485
0.838
1.249
4.993
--
2.744
--
1.137
0.604
--
15.300
3.169
0.357
0.282
0.122
0.824
1.956
3.117
6.537
0.313
168.309
SDCP
31.2
25.5
49.1
42.8
24.7
64.4
32.3
39.8
62.4
57.2
30.3
61.4
30.6
48.3
48.3
32.1
36.9
44.8
38.9
31.1
38.3
46.4
38.6
63.0
37.8
63.7
48.3
32.9
97.4
50.7
39.3
31.8
79.1
62.5
46.0
78.0
56.1
49.6
3,777.8
DDCP
1.067
0.871
1.678
1.464
0.845
2.200
1.106
1.359
2.132
1.954
1.037
2.099
1.046
1.650
1.651
1.099
1.261
1.531
1.329
1.065
1.310
1.587
1.318
2.153
1.292
2.176
1.651
1.126
3.328
1.733
1.345
1.087
2.702
2.136
1.572
2.663
1.918
1.695
127.996

Note: —  individual figure may not  add to totals due  to  rounding.
                                    106

-------
           TABLE V-4. ECONOMIC DAMAGE FUNCTIONS ON MATERIALS^/


SDCZ      -23,328.4 + 43.1 ME + 943.3 S02 + 148.1 TSP - 235.0 SUN
          (19,929)    (3.4)*   (171.6)*     (356.0)*    (1820.4)

                                                                       (V-20)
               + 2,679.3 RHM + 21.9 YP       R2 = 0.64
                (1,750.2)      (18.9)
DDCZ    = 7,562.2 + 1.4 ME + 30.5 S02 + 47.9 TSP - 76.2 SUN
          (6,460.4)   (0.1)*  (5.5)*     (11.5)*    (59.0)
               + 86.8 RHM  + 712.6  YP        R  = 0.63
               (56.7)       (615.5)
 SDCP     - -141,199.7 + 577.2 HU +  15.2 YP + 911.3 RHM + 69.1  S02
           (259.861.3)    (3.4)*     (2.6)*   (235.3)*     (23.2)*

                                                                       (V-22)
               + 305.3 SUN                   R  = 0.995
                 (245.9)
DDCP     = -4,820.1 + 19.7 HU + 0.5 YP + 31.1 RHM + 2.3  S02 + 10.4  SUN
           (887.2)*    (0.1)*   (0.08)*  (8.0)*     (0.8)*     (8.4)

                                              9                        (V-23)
                                             RZ = 0.995
a/  The values below the coefficients are standard errors with * to indicate
      that the coefficients are significant  at  the  1  percent  level. All coefficients
      and standard errors are reduced by a factor of  10  .
                                    107

-------
of material. The coefficients of ME, SC>2, and TSP are significant at the 1 per-
cent Level in equations (V-20) and (V-21), whereas the coefficients of  HU, YP,
RH and S02 are significant at the 1 percent level in equations (V-22) and  (V-23).

    The economic damage functions on zinc and paint summarized in Table V-4 were
estimated simply for national decisionmaking. They offer some short-cut tech-
niques for  rough  computations and can be used in determining the marginal as
well as average damages (benefits) resulting from a pollution control strategy.
To serve as an illustration, an example involving the computation of the par-
tial elasticity of SDCZ with respect to SC>2, and the associated marginal
benefit due to a reduction in SC>2, is presented. Suppose the federal government
is contemplating the implementation of a pollution abatement program which is
expected to reduce the average S02 level in the urban areas by, say, 10 percent.
A question arises as to what the dollar benefit will be for the reduction  in
soiling damage of zinc as a result of the pollution control program. Since the
average, gross soiling damage due to S02 is $102 million and the average SC>2
level is 55.73 [ig/m3 among the 148 SMSA's, the partial elasticity of the damage
cost with respect to S02 is obtained:
                E  __  = 0.9433 x  (55.73/102) = 0.52.
                 c,SO


     Thus,  it  is in general expected that a 10 percent decrease in the SC>2 con-
 centration level will result in a  5.2 percent reduction in the soiling damage cost
 cost of  zinc.  Since  the mean value of the regional damage cost for the 148  SMSA's
 included in this study is $102 million, when the 862 level decreases from 55.73
 (J,g/m3 to 50.16 |ig/m3, it is also expected that on the average the damage cost
 will be.  reduced by the amount of $102 million x 5.27o = $5.73 million. Likewise,
 the  elasticities for the other dependent variables with respect to S02 and  other
 explanatory variables can be analogously computed and interpreted.
 A SUMMARY OF MATERIAL PHYSICAL DAMAGE FUNCTIONS

    After a careful review of the  literature,  some recent publications  as well
 as many unpublished manuscripts were identified as major sources providing use-
 ful information for our future on material damage*. The damage resulting from
 air pollution  includes corrosion of metals,  deterioration of paints and materi-
 als, fading of fabric dyes, etc. It is also  worth noting that the physical damage
 functions expressed in equations (V-30) and  (V-32) below were utilized  in deriv-
 ing the economic damage estimates  in the preceding section.

 Major Pollution and Material Interactions

     Fred H. Haynie (1974), based  on an MRI  study, summarized in Table  V-5 the
 relative extent to which effects of pollutants on materials are known.  Haynie
 considered an  effect (1) "well established"  when evidence was corroborated by
 several high quality references; (2) "some evidence"  in the presence  of one  or
 two references; and (3) "suspected" when interactions are based on behavior  of
 material at pollutant level much higher than the ambient level.

                                     108

-------
TABLE V-5. MAJOR POLLUTANT - MATERIAL INTERACTIONS
                                       Material
  Pollutants
Metals   Paints   Textile   Elastomers   Plastics
Sulfur dioxide
Particulate
Ozone
Nitrogen dioxide
Hydrocarbons
1
2
2
2
--
2
1
2
3
3
1
2
1
1
__

_ _
1
3
--
3
—
3
3
—

Metals--
     Steel, S02_ and Photochemical Oxidant--Haynie and Upham (1971) conducted
an in-house field study of the effect of atmospheric pollutants on steel cor-
rosion, Good dose-response relationships were estimated as follows:
     Carbon steel  y = A.013
                                0.00161 SO,
                                    0.7512 - 0.00582 OX
                            (4.768t)
(V-24)
     Copper bearing steel
                   y = 8.341
                                0.00171 SO,
                                    0.8151 - 0.00642 OX
                            (4.35U)
(V-25)
     Weathering steel
                   y = 8.876
                                0.0045 SO,
                                    0.6695 - 0.00544 OX
                            (3.389t)
(V-26)
   is the depth of corrosion in microns (m);
   up   t  is time.
                           S02  and  OX  are expressed in
     Enameling Steel, Sulfate in Suspended Particulate--Havnie and Upham (1974)
in a companion paper examined correlation between erosion behavior of steel and
gaseous S02, total suspended particulate, sulfate in suspended particulate, and
nitrate in suspended particulate. Multiple linear regression and nonlinear curve
fitting techniques were used to analyze the relationship between corrosion of
enameling steel and the atmospheric data. The resulting best empirical function
has the form:
                                     109

-------
        Corrosion = 183.5   x/TEXP  [0.06421 Sul - 163.21/RH ]         (V-27)

where

        corrosion = depth of corrosion, p-m

                t = time, years

              Sul = average level of sulfate in suspended particulate
                      (p.g/m ) or average level of sulfur dioxide  (|J,g/m )

               RH = average relative humidity

The  statistical analysis shows that differences in average temperature, average
total  suspended particulate, and average nitrate in suspended particulate exert
insignificant effects on steel's corrosion behavior. Covariance between sulfate
and  S02 and the relative accuracies of the two sets of data make  it impossible
to statistically identify the causative agent. Laboratory experiments  suggest
that S02 is the major cause.

     Galvanized Steel and S0?--In an unpublished paper, Spence, Upham, and
Haynie (1975) derived a  dose-response relation for galvanized steel, S02, N02
and  Ozone  in controlled  environmental chambers. Of the three pollutants, S02
was  shown  to be a major  factor in determining the corrosion rate  of galvanized
steel. The corrosion of  the galvanized panels fits the relationship:

                                       •u   p /RT
                     Corr = (d. SOo +  e        ) ft~                  (V-28)
                              0   ^             v w
 where          Corr = corrosion in micrometer ((im)

                  dQ = 0.0187,  b = 41.85,  E = 23,240

                  t  = time of wetness
                   w

     Weathering Steel, S02, Relative Humidity and Temperature—A  similar cham-
ber  study  for weathering steel was conducted by Spence, Haynie, and Upham
(1975b) who developed a  corrosion function that accounts for  99 percent of  the
variability for the  clean air and pollutant experimental data.

                        r               (55  44 _ 31.150 "I  _.
               Corr =    5.64  \/ SO. +  e    *       RT   J V  w           (V-28)
 where
      S02  is

        R  is  1.9872  cal/g -  mole °K
                                      110

-------
        T is the goemetric mean temperature of the specimen when wet
            in °K

        t is time of wetness in years
         w
     This chamber has shown that S02 is a major factor determining the corrosion
of weathering steel.

     Zinc and S02—Haynie and Upham (1970) have shown that the amount of S02
in the air is the major factor in determining the rate of corrosion of zinc.
They found that little zinc corrosion would occur in an environment in which
SC>2 was not present.

     The dose-response relationship between zinc corrosion rate and SC>2 was
estimated as follows:
          Y= 0.00104 (RH - 49.4) S02  - 0.00664 (RH - 76.5)           (V-30)

or alternatively, Y = 0.001028 (RH - 48.8) S02

where

          Y = zinc corrosion rate, |j,m/year

         RH = average relative humidity, percentage

        SO  = average sulfur dioxide concentration, \},g/m.


     Pitting of Galvanized Steel--J. W. Spence and  F. H. Haynie  (1974) exposed
specimens of galvanized steel to polluted and clean air in controlled environ-
mental chambers. They found that corrosion of the zinc films was  essentially
a linear function of time for polluted and clean air condition. Uniform corro-
sion of the zinc occurred in the polluted exposures, whereas pitting corrosion
of the zinc was observed ^.n the clean air exposures.

     The pitting corrosion, expressed as a uniform thickness loss, fits the re-
lationship:


           Corr = t  exp   |30.53 - (16,020/RT )1                       (V-31)
                   w       [_                  m J
 where
           Corr = amount of pitting corrosion  (|im)

                                    111

-------
            t  = time of wetness,  years
             w
            T  = geometric mean specimen temperature when wet, °K
             m
     Catastrophic Failure of Metals--Air pollution has contributed to the cata-
strophic failure of metal structure. John Gerhard and Fred H. Haynie (1974) es-
timated the loss of metal failure to be between $50 million and $100 million
annually for the United States. The dose-response relationship between air pol-
lution and the occurrence of catastrophic failure of metals has not been estab-
lished in the literature.

     Three types of catastrophic failure of metals that are associated with en-
vironmental corrosion were identified: (1) stress-corrosion cracking, (2) cor-
rosion fatigue, and (3) hydrogen embrittlement. Notable examples of problem
areas involve failure of essential structures, aircraft and aerospace components,
and communication equipment.

     In the case of a 40 percent reduction in pollution, the per capita cost
would drop from the present level of $7.10 to $4.36 by 1980. Assuming a 60 per-
cent reduction in pollution, the per capita cost will drop from $7.10 to $2.20
in 1980.

     Air Pollution Corrosion Costs on Metals--Fink, Buttner and Boyd (1971)
examined air pollution corrosion costs on metals in the United States from both
technical and economic viewpoints. They calculated corrosion costs for the nine
major categories which were most sensitive to and most damaged by air pollution
corrosion. The grand total cost was estimated at $1.45 billion, or approximately
$7.10 per person per year.

     Fink, Buttner and Boyd considered SC>2 as the most important pollutant from
a corrosion point of view. They projected the damage costs of metals due to SO
under a variety of SO  concentration levels for 1980. The annual loss would in-
crease from the present $1.45 billion to $2.1 billion by 1980 if there is a 55
percent increase in pollution. A 10 percent increase in SO  level would result
in an increase in annual loss by $0.3 billion to $1.73 billion in 1980.

Paint

Paint technology and Air Pollution--
     J. W. Spence and F. H. Haynie (1972a) in their recent survey on paint tech-
nology and air pollution, identified the characteristics of pollutant attacks
on exterior paints, and estimated the annual cost of air pollutant damage to
such paints. They assessed the chemical damage of air pollutants for four classes
of exterior paints: (1) household, (2) automotive refinishing, (3) coil coating,
and (4) maintenance. The cost at the consumer level is more than $0.7 billion
annually. Household paints sustain damage representing 75 percent of the total
annual dollar loss.

                                     112

-------
Exterior Paint, SO?, and Particulates--
     Spence and Haynie (1972b) investigated the deterioration of exterior paints
due to 862 and particulate matter, and the associated potential economic loss
to manufacturers and consumers. A breakdown of the damage loss of exterior
paints is summarized as follows:
                          Loss at Consumer
                         Level (million $)

                    Coil coating               16
                    Automotive refinishing     88
                    Maintenance                60
                    Household                 540
                      Total                    704
Oil Base House Paint, Acrylic Latex House Paint, Vinyl Coil Coating and Acrylic
Coil Coating--
     A chamber study of the effects of gaseous pollutants on paints was carried
out by J. Spence, F. Haynie, and J. Upham (1975c). Regression analysis showed
that S02 concentration and relative humidity accounted for 61 percent of the
variability in the case of oil base house paint which experienced the highest
erosion rates. Vinyl and acrylic coil coatings experienced very low erosion
rates.

     The multiple linear regression of oil base house paint on SO  concentration
and relative humidity gave the relationship:
           erosion rate = 14.323 + 0.01506 SO  + 0.3884 RH             (V-32)
where erosion rate is (j,m/year
                          3
               S02 is

               RH is percent relative humidity

     In the case of vinyl and acrylic coil coating, the regression equations
are respectively given by:
                                     113

-------
           erosion rate = 2.511 + 1.597 x 10"5 RH x SC>2                (V-33)
and
          erosion rate = 0.159 + 0.000714 03                           (V-34)

where  0   is in (J,g/m .

Fabric Fading

Selected Drapery Fabrics and N02--
     Upham, Haynie and Spence (1975) have studied and assessed the fading charac-
teristics of three drapery fabrics after exposure to air pollutants and other
environmental factors in the chamber. The experimental results indicated that
N02 is a major factor in determining the fading rate for one of the fabrics,
a plum-colored cotton duck material. The other two fabrics did not fade signifi-
cantly in the presence of the air pollutants.

     The dose-response relationship for the plum fabric was estimated as follows:
             A E = 30
1 - EXP(-(257 + 3.38 x  IQ-^M x N02>t)
(V-35)
    where
              A E = amount of fading,  fading units

                M = amount of moisture,  (J,g/m ,  at 25 C and one atmosphere

              N02 =

                t = exposure time, year

               R2 = 0.70
Dyed Fabrics, N02, Ozone, S02 and Nitric Oxide--
     Beloin  (1973) assessed 20 dye-fabric combinations which were exposed to
two levels each of N02, ozone, S02 and nitric oxide, at four combinations of
temperature and humidity for a period of 12 weeks. The study showed that NC^,
ozone and, to a lesser extent, S02, can cause appreciable dye fading, and that
nitric oxide has little or no effect. In an earlier article, Beloin (1972)
evaluated the color fastness of 67 dye-fabric combinations exposed to atmos-
pheric gases. Multiple regression analysis of pollutant concentrations indicated
that SC>2, N02 and ozone are major factors determining fabric fading.

Rubber, Ozone, Nitrogen

     Stickney, Mueller and Spence (1971) estimated the yearly cost of air pol-
lution damage to rubber industry. The total estimated cost at the consumer level
                                    114

-------
is at least $500 million yearly. Of this,  $398 million can be accounted for in
detail. Mueller and Stickney (1970) identified ozone as the only major pollutant
to shorten the life of rubber products. S02 is not known to have harmful effects
on rubber products.

Soiling and Suspended Particulates

     Beloin and Haynie (1975) studied the soiling of building materials. Six
building materials were exposed at five sites in Birmingham, Alabama,  to deter-
mine the rate of soiling by different levels of suspended particulate. Excellent
dose-response relationships were obtained for the white-surfaced painted cedar
siding and asphalt shingles. Similar regressions for brick can account for 34
to 50 percent of variability. Poor correlations were obtained for concrete,
limestone, and window glass. The regression results are summarized in  Table V-6.

Material Damage

Material Damage From SO 2--
     An overall assessment of material damage from S02 was made by Gillette and
Upham  (1973) and is summarized as follows:

     Me ta Is-- Cor ro si on of metals by acids derived from airborne S02 is most im-
portant. Zinc and  steel are particularly vulnerable to attack by atmospheric
S02.

     Cotton Fabrics--Not significant since most cotton fabrics are not exposed
continuously to the external environment.

     Synthetic Fabrics and Blends--Not significant with the exception  of nylon
hosiery.

     Dye Fading- -Unimportant at present SO  concentrations.

     Paper and Leather Products- -Strongly influenced by S02» tend to disintegrate
or discolor after prolonged exposure to relatively high levels of
     Plastics- -Little is known about the effects of S02 on plastics.

     Concrete, Marble, Roofing Slate, Mortar and Other Limestone—Subject to
attack from acids derived from S02- Most of the concrete and limestone used in
the construction of highways and buildings in the United States is not seriously
affected by the present level of atmospheric S02«

     In assessing the damage loss resulting from S02> the following observations
are noteworthy:

     - Most materials are not substantilly damaged when pollution levels are
less than 250 |j,g/m3.
                                     115

-------
                  TABLE V-6. RESULTS OF REGRESSION ANALYSIS FOR SOILING OF BUILDING MATERIALS
                             AS A FUNCTION OF  SUSPENDED PARTICULATE DOSE
_. Material Independent variable
Oil Base Paint JsP(pg/m3) x t (months)
V
Tint Base Paint Ditto
Sheltered Acrylic "
Emulsion Paint
Acrylic Emulsion "
Faint
Shingles SP(ug/m ) x t (years)
10
Shingles -JsP(ug/m ) x t (months)
Concrete Y Ditto
Coated Limestone "
Uncoated Limestone "
Coated Red Brick "
Uncoated Red Brick "
Coated Yellow Brick "
Uncoated Yellow Brick "
Glass "



N Number of data sets (dependent upon the
A Intercept of linear regression
B Slope of linear regression
N
400

400
400

720

48

48
160
80
80
80
80
80
80
45 0



number


A
89.43

86.13
91.54

90.79

41.69

43.50
41.45
44.57
46.99
12.95
14.88
45.05
43.21
.2806




B
-0.2768


-0.2618
-0.

-0.

-0.

-0.
-0.
+0.
-0.
-0.
-0.
-0.
-0.
+0.



593

4131

331

199
0458
0779
0503
0296
0374
1133
1133
0314



of controlled






s2,
0.0641

0.0571
0.1156

0.0497

0.1895

0.5771
0.1338
0.2464
0.1500
0.0223
0.0331
0.5337
0.2740
0.008077



variables


S2B
0.000069

0.000061
0.000123

0.000026

0.000312

0.000258
0.000080
0.000164
0.000089
0.000013
0.000020
0.000317
0.000168
0.000007



S2
E
7.6510

6.8265
13.8143

8.3791

3.8685

7.6992
7.5011
6.9046
4.2035
0.6255
0.9274
14.9533
7.6773
0.6851



in the factorial





0.

0.
0-.

0.

0.

0.
0.
0.
0.
0.
0.
0.
0.
0.



P2
745

738
880

902

884

769
143
347
266
459
477
342
503
340



Remarks
Excludes all data
beyond 12 months
Ditto
11











Excludes 2 24 month
Tarrant readings
Ditto
















Haze readings include
ing 3 periods for
right panes prior
12 months

to

experiment)








Estimated variance of intercept
Estimated variance of slope
Residual variance (error)
Correlation index (fraction of variability accounted for by regression)
Suspended particulata

-------
     - Given the present 862 levels, damage to most materials other than certain
ferrous and nonferrous metals is probably not significant.

     - The only important materials adversely affected by 862 are iron,  steel
and zinc products.

     - There are two important factors determining the corrosion rate for gal-
vanized products: (1) relative humidity, and (2) S02 level.

     - The relationship between SO  and damage to paint is not as clear  as it
is between SO  and corrosion to galvanized products.

     - The threshold or minimum level of SO  required to produce an economic
loss ^ (10 |j,g/m3). For lack of better data, it is reasonable to assume a thresh-
old level of 20 p,g/m  before any loss is achieved.

     Finally, to recapitulate, the various physical dose-response relationships
for metals, paints, fabrics, and building materials are summarized in Table V-
7.
                                     117

-------
                               TABLE V-7. PHYSICAL DAMAGE FUNCTIONS  FOR MATERIALS
                 Material
                                       -  Response Relationships
1.  Metals
          A.  Steel - Carbon steel
                      Weathering steel B


                      Enameling steel A
                      Galvanized steel
          B.  Zinc
2.   Paints
          A.  Oil base house paint
                                                Y = 9 013
Copper-bearing steel      Y = 8 341
                      Weathering steel A        Y = 8 876
                                      0.00161 S02
                                                             . 00171
                                      0.0045 S02
                                                                                  0.7512 - 0.00582 OX
                                                    (4.768t)
                                                            0.8151 - 0.00642 OX
                                                                          (4.351t)
                                                                                 0.6695 - 0.00544 OX
                           corr =|5.64 \|S02 + e
                                                   (3.389t)

                                               (55.44 - 31,150/RT)
                      'Enameling steel B         corr = 325
                                          [0.06421 Sul-163.21/RH]
                          corr = 183.5 >jte

                                        [0.00275 S02 - (163.2/RH)]
                          corr =
                                                                       41.85 -  23,240/RT
0.0187 S02 + e
                            Y* = 0.001028 (RH - 48.8) S02
                          erosion rate = 14.323 + 0.01506 S02 + 0.3884 RH
                                                                                                             0.91
                                                    0.91
                                                    0.91
                                                    0.91
                                                                                                             0.9
                                                                                                             0.9:
                                                                                                              0.6

-------
                                        TABLE V-7  (Concluded
                                                                                 -5
          B.  Vinyl coil coating

          C.  Acrylic coil coating


3.  Fabrics - Plain fabric


4.  Soiling of Building Material

          A.  Oil base paint

          B.  Tint base paint

          C.  Sheltered acrylic emulsion
                 paint

          D.  Acrylic emulsion paint

          E.  Shingles

          F.  Coated yellow brick
  erosion rate = 2.511 + 1.597  X10   RH x S02

  erosion rate = 0.159 + 0.000714 03

              -(2.57  + 3.38 x 10"  MX N02)t
AE = 30
           L-e
  Reflectance = 89.43 -  0.2768  \JSP x t*
  Reflectance = 86.13 -  0.2618 ^SP x t*
  Reflectance = 91.54- -  0.593 \JSP x t*

  Reflectance = 90.79 -  0.4131 NJSP x t*
  Reflectance = 43.50 -  0.199 XJSP x t*
  Reflectance = 43.21 -  0.1133 \(SP x t*
                                                                                                             0.34
   0.70
   0.74

R2 = 0.738


R2 = 0.88

R2 = 0.902
R
     0.769
R2 = 0.503
where      Y = depth of corrosion in microns  (u)
                   13
         S02 =ug/nr
          OX = ug/m3
           t = time, years
        corr = depth of corrosion in micrometer (um)
         sul = average  level of sulfate in suspended particulate (ug/m )  t* = time, months
          RH = average  relative humidity, percent
          tw = time of  wetness
           R = 1.9872 cal/g - mole  °K
           T = geometric mean temperature of the specimen when wet in °K
          Y* = Zinc corrosion rate,/urn/year
Erosion rate = urn/year
          03 = Ozone, ug/m3
         /\E = amount of fading, fading units
                                                        A
                            M = amount of moisture, ug/m at  25°c and one
                                  atmosphere
                           N02 = ug/m3
                  Reflectance = a measure of  soiling, percent
                            SP = suspended particulate (ug/m3)

-------
                                  SECTION VL

                         VEGETATION AND AIR POLLUTION
PROBLEMS AND OBJECTIVES

     Air pollution is a fact of contemporary life. It is not only deleterious
to human health, material, and household and commercial establishments as dis-
cussed in the preceding sections, but it is also recognized as a causal agent
of damage to vegetation. Urban expansion and industrialization have resulted
in deteriorated air quality in many major cities in the United States. Though
social concern with the problem of contaminated air can be dated back to as
early as the 13th century, the biological effects of degraded air are not thor-
oughly understood even now. Some progress has been made, however, in recent
years. According to Naegele (1973), laboratory and chamber studies of individual
plants under somewhat controlled environments have contributed to the awareness
of the complexity of plant response to toxicants. Acute and even chronic responses
of plants to deteriorated air are being studied and documented.

     There  are three principal air pollutants of major interest to agricultural
plants; namely, sulfur dioxide, fluorine compounds and smog. Regarding smog,
 there are two distinct types, with numerous intermediate grades:  the London
type, which is a mixture of coal smoke, fog and sulfur dioxide, and the Los
Angeles type which is a mixture of ozone and peroxidized organic compounds.!.'

     Studies on the effect of sulfur dioxide (S02) on vegetation are voluminous.
Stoeckhard  (1871) reported SO  injury to plants as early as 1871. Since then
more than 700 articles have been published regarding the effects of SO  upon
vegetation. The documents point to a great variation in plant responses to the
pollutant.  This variation in plant responses can be accounted for by such fac-
tors as genetic composition, stage of development, climatic factors, interactions
between pollutants, the time of day of exposure, and soil moisture.

     The effects of air pollution are customarily classified into two categories:
(1)  visible effects, which are identifiable pigmented foliar patterns as a re-
sult of major physiological disturbances to plant cells, and (2) subtle effects,
which are not visibly identifiable, and may be identified when physiological
change occurs in the plant. The disturbance of biochemical processes at the
molecular level is the cause of both the visible and subtle effects. Within the
category of visible effects, acute and chronic injury can be identified. Acute
injury is a severe injury as a result of a short-term, but high concentration
of the pollutant. Chronic injury is light to severe injury; it develops from
exposure to long-term, low pollutant concentration.
_!/ For a detailed discussion on the types of air pollutants causing  damage  to
     vegetation, see Thomas (1961).
                                     120

-------
     The effects of oxidant on vegetation have been studied since the early part
of this century. Oxidant or smog type symptoms were identified with the reac-
tion product of ozone and reactive hydrocarbons. The symptoms were also associ-
ated with a new toxicant, peroxiyacetyl nitrate (PAN),  which was generated ex-
perimentally by photochemical reaction of a mixture of  nitrogen dioxide and re-
active hydrocarbon (Stephen et al., 1960). Nitrogen dioxide is also a phototoxi-
cant at high concentration levels. Benedict and Breen (1955) found tissue col-
lapse with nitrogen dioxide concentration above 20 ppm.

     Generally speaking, agricultural plants are adversely affected by air pol-
lution vis-a-vis reductions in the quantity of output and/or degradation of the
quality of the product. With the information on the determinants of the biologi-
cal response of a plant to contaminated air, a reasonable, physical dose-response
relationship could be constructed. In translating the physical damage function
into a monetary damage function, the following factors  should be considered:
time and growing season, market value and price of the  plant, the possibility
of growing a different crop and the opportunity cost of the site for growing
the plant.

     Waddell (1974) identified two general approaches to assess the economic
loss of plants due to air pollution-}J One approach is  to survey the damage loss
on a statewide basis. Included in this category are the studies by Middleton
and Paulus (1956), Weidensaul and Lacasse (1970), Feliciano (1972), Pell (1973),
Naegele et al.  (1972), and Millecan (1971).

     Another approach is to construct predictive models by relating data on crop
losses to crop values, pollution emission and meteorological parameters. The
landmark study by Benedict and his associates (1971, 1973) at Stanford Research
Institute (SRI) is probably the only study undertaken so far which provided some
essential background material for further investigation. The SRI study estimated
plant losses caused by air pollution in those U.S. counties where major pollutants
(oxidants, SO , and fluorides) are expected to produce  adverse effects on plants.

     The major contribution of the SRI study is the provision of a wealth of
data for the development of economic damage functions or of more sophisticated
predictive models when better dose-response data are available. However, the
study also contains the following weaknesses: (1) the damage factors were at
best educated guesses and are subject to criticism; (2) yearly variations in
climate and meteorology were not allowed for; (3) ornamentals were undervalued
since only replacement costs were used as a proxy for aesthetic values; and (4)
the subtle effects of air pollution which causes no visible injury were ignored.
However, some subtle injuries were indeed included, contrary to most critics.
The amount was a rough guess and, with the exception of citrus and grapes, could
II See Waddell (1974) for a detailed discussion.
                                     121

-------
have been much larger or much smaller  depending on  the plant  species.  Latest
information  shows that  such  losses to  forests  and perhaps  cotton  in  California
are much greater than previously realized.i'

     A review of some previous  damage  estimates at  both  the national and state
 levels would give us a  rough idea as to  how serious the  damage loss  is because
 of air pollution. Benedict and  his associates  estimated  the national total damage
 of visible  injury to vegetation to be  $132  million  each  year. Lacasse-Weidensaul
 estimated the amount of direct  losses  uncovered in  the survey to  be  more than
 $3.5 million in Pennsylvania in 1969.  Indirect losses were estimated to be $8
 million. Feliciano  reported  the losses to agriculture in New  Jersey  due to air
 pollution were about  $1.19 million in  1971. Naegele estimated direct economic
 losses for  the 1971-72  season at $1.1  million. Finally,  Millecan  estimated a
 monetary loss of  $26 million in crops  in California in 1970.

      In summary,  the problems in the field  of  vegetation and  air  pollution are
 similar to  those  delineated  previously in other categories, i.e., the lack of
 reliable scientific damage functions and the presence of a wide range of damage
 estimates.  The primary  objective of this section  is to review the state of the
 art and derive,  through existing documentation and  data, an integrated economic
 damage function  of  air  pollution on vegetation for  purpose of prediction.  The
 remaining part of this  section  contains  the following subsections:   Dose-Response
 Relationships, Economic Damage  Functions, and  Concluding Remarks.
 DOSE-RESPONSE RELATIONSHIPS

      Some crude dose-response relationships  for  various  types  of  crops have been
 derived. O'Gara (1972) estimated the first  such  function for alfalfa under condi-
 tions of maximum sensitivity, as follows:
                  (C-0.33O  = 0.92                                      (VI-1)


 where  C  is  the  concentration  level  to  be  estimated with respect to time  t
 in  hours.  The constant  0.33 ppm represents  a  concentration that presumably can
 be  endured indefinitely,  i.e.,  the  threshold  level,  without prolonged fumigation.
 That  is  to say that  C = 1.25 ppm for  t = 1.0.

      The O'Gara equation  was generalized by Thomas and Hill (1935) for any degree
 of  leaf  destruction  and any degree  of susceptibility.  The generalized equation
 can be specified  as:
                               t(c-a) = b                               (VI-2)
JY Personal correspondence with Dr. H. M.  Benedict.

                                      122

-------
where t = time, hours, c = pollutant concentrations above a, a = threshold concen-
tration below which no injury occurs, and b = constant.

     W^th maximum susceptibility, the generalized equations were shown as follows:


                    t(c-0.24) = 0.94 traces of leaf destruction

                    t(c-1.4)  =2.1  50 percent leaf destruction

                    t(c-2.6)  = 3.2  100 percent leaf destruction
     Zahn  (1963) developed an equation which modified the O'Gara equation and
provides better fit over a longer period of time. The equation is shown as fol-
lows:
                                  t =  1 + 0.5C                         (VI-3)
                                       C(C-a)

The threshold  level  a  was given  as 0.1  for alfalfa;  b  is the dimensional
resistance factor which incorporates the  influence of environmental conditions,

     An alternative experimental formula  was suggested by Guderian, Van Haut
(1960) and Stratmann (1963). The formula  gives best fit to their observations
for either short- or long-term exposures.


                                 t  =  Ke-b(C  - a)                        (VI-4)
where  K = vegetation life time, in hours, t;  a,  b,  and  C  are the same as
in (VI-3). These parameters may vary with species, environmental conditions,
and degree of injury.!.'

     Although several physical dose-response relationships have been determined,
economic damage functions for vegetations are largely nonexistent. The economic
damage functions described in the following section employed input data on vege-
tation losses obtained from the Benedict study (1971,1973).
I/ The dose-response equations developed by Zahn, and Guderian, Van Haut and
     Stratmann were summarized in Environmental Protection Agency, Effect of
     Sulfur Oxides in the Atmosphere on Vegetation, op cit. The references were
     contained therein.
                                    123

-------
     Benedict et al.  derives crop Loss estimates by using the product of three
     i Y- C _  T . (^ . -
factors, i.e
          Crop Loss = crop value .  crop sensitivity to the pollutant .
                        regional pollution potential                   (VI-5)


     The regional pollution potential is a relative severity index of pollution,
estimated for each county selected in the Benedict study on the basis of emission
rates which are, in turn, derived from fuel consumption data. The relative sensi-
tivity of various plant species to the pollutants was determined from a litera-
ture review. Each crop or ornamental was classified as to whether the part of
the plant directly affected by the pollutants had high, medium or no economic
value.

     Despite the fact that the ceteris paribus type of dose-response functions
has been developed and refined for certain types of vegetation, such functions
are still unavailable for a majority of vegetations even now. Furthermore, the
multivariate physical damage functions relating plant damage to several relevant
explanatory factors are yet to be developed. In the absence of reliable plant
dose-response functions, only rough estimates of economic damages for various
plants can be derived.
 ECONOMIC  DAMAGE FUNCTIONS

      Of more  relevance to policymakers at both the national and local levels,
 however,  are  the monetary or economic damage functions which transform all as-
 pects of  dose-response relationships into one common unit of measurement, i.e.,
 money. An attempt was made in this study to estimate such economic damage func-
 tions which relate economic losses of a variety of crops to air pollution con-
 centration levels and climatological variables.

      The  crops and agricultural products for which the economic damage functions
 were  estimated include corn grain, soybean, cotton, vegetable, other vegetable,
 nursery,  floral, forestry, field crop, fruit and nuts, total crops, total orna-
 mentals,  and  all plants. The selection of the crops is based mainly on the eco-
 nomic  importances of these crops to the United States. However, it is understood
 that  different cultivating procedures and methods as well as relocation  of crop
 growing patterns in the United States will result in reduction in air pollution
 damage to crops.

     A stepwise linear multivariate regression model was developed for determin-
 ing the economic damage functions for the selected crops and plants, as  follows:


     CROPLi= a + b CROPVi + c TEMB + d TEMA + e SUN + f RHM + g DTS
               + h S0  + j OXID                                        (VT-6)
                                      124

-------
where  CROP   denotes the economic loss (in $1,000) of the  ith  type of crops
by county from the Benedict study;  CROPVi  the output value (in $1,000) of the
ith   type of crops by county;  TEMB  and  TEMA stand for, respectively, the
number of days in a year with temperature below 33°F and above 89°F;  SUN  rep-
resents possible annual sunshine days;  RHM, relative humidity;  DTS number of
days with thunderstorm;  S02  sulfur dioxide concentration or relative severity
index; and  OXID  the oxidant relative severity index.

     Data used for the regression analysis were obtained from prior studies on
vegetation losses and the official publication on climatological data. As noted
earlier, the disaggregated data on the vegetation losses and the values of the
crops by county were obtained from the Benedict study. It should be pointed out
that only the aggregate data on vegetation by regions are presented in Benedict
et al. (1973). The crop data in the published form were integrated so as to pre-
serve some anonymity about certain single sources of pollution. The data for
CS02 and OXID were taken from Table 7 of Benedict et al. (1973), and the data
for TEMB, TEMA, SUN, RHM, DTS were secured from the U.S. Department of Commerce,
Local Climatological Data. Since the climatological data were not available for
all counties or cities, data for a nearby city were, hence, substituted for the
missing information for a number of counties. Finally, the annual mean level
for S02 was taken from the U.S. Environmental Protection Agency, Air Quality
Data - 1972 Annual Statistics.

     Although estimates on crop values and crop losses are available for a total
of 679 counties in the United States, a thorough examination of the data reveals
that some counties have zero crop damage estimates and, hence, are not suitable
for inclusion in the study sample. In addition, both climatological and pollution
data are unavailable for a number of counties, but for which positive crop loss
estimates were available. Only 74 counties have both positive crop loss esti-
mates and data on climate and pollution levels. Thus they were selected for
this study for deriving the vegetation economic damage functions.

     The dependent and explanatory variables used in the regression analysis
are described in Table VI-1. It should be noted that for sulfur dioxide two al-
ternative measures were available: the first measure is the relative severity
index constructed on the basis of pollutant emissions, concentration rate factor
and episode days by Benedict et a1.. (1971), i.e., CS02«.The second alternative
measure, S02, is the annual mean level for sulfur dioxide (fig/m^). Both measures
were used in the regression analysis, and the regression results are separately
reported in Tables VI-2 and VI-3. With regard to oxidants, the relative severity
index for oxidants was also provided by Benedict et al. (1971). However, data
on the annual mean level of oxidants are insufficient for this study. Thus, only
the former measure was used in the regression analysis. The regression results
containing oxidants are presented in Tables VI-2 and VI-4.

     Some remarks on the regression results are in order. The values below the
regression coefficients are standard errors with * indicating that they are
significant at the 1 percent level. The signs of the regression coefficients
                                    125

-------
           TABLE VI-1. VARIABLES USED IN ECONOMIC DAMAGE FUNCTIONS
A.  Dependent variables - vegetation loss (in $1,000)
       CORNL
       SOYBL
       COTNL
       OVGTL
       NUSRL
       FLORL
       FRSTL
       FCROL
       FRNTL
       VEGTL
       TOCRL
       TOORL
       ALPLL
Corn grain loss.
Soybean loss.
Cotton loss.
Other vegetable loss.
Nursery loss.
Floral loss.
Forestry loss.
Field crops loss.
Fruit and nuts loss.
Vegetable loss.
Total crop loss.
Total ornamentals loss.
All plant loss.
B.  Explanatory Variables

       CROPV
       TEMB
       TEMA
       SUN
       RHM
       DTS
       so2
       OXID

       CS00
The value of the vegetation in question (in $1,000)
Number of days with temperature 32°F or below.
Number of days with temperature 90°F or above.
Possible annual sunshine days.
Relative humidity.
Number of days with thunderstorm.
Annual mean level for sulfur dioxide (ug/m ).
The relative piant-damaging oxidant pollution
  potential index.
The relative plant-damaging sulfur dioxide
  pollution potential index.
                                     126

-------
                               TABLE VI-2.  ECONOMIC DAMAGE FUNCTIONS ON VEGETATION WITH POLLUTION
                                           RELATIVE SEVERITY INDICES (in $1,000)
Ni
-J


(1) CORNL
(2) SOYBL
(3) COTNL
(4) OVGTL
(5) NUSRL
(6) FLORL
(7) FRSTL
(8) FCROL
a
4.4
(32.1)
-2.2
(0.3)
-5.8
(6.9)
133.6
(58.5)*
-113.1
(300.2)
-616.4
(485.2)
-616.4
(485.2)
520.5
(222.3)*

CROPV
0.001
(0.001)
0
(0
0
(0
0
(0
0
(0
9
(0
0.
(0.
.003
.001)*
.0063
.0002)*
.006
.001)*
.11
.02)*
.10
.01)*
071
003)*
0.003
(0.002)
TEMB
0.02
(0.04)
0.01
(0.03)
0.0006
(0.0094)
-0.03
(0.08)
1.12
(0.42)*
0.93
(0.57)
1.93
(0.70)*
0.28
(0.32)
TEMA
0.09
(0.10)
0.04
(0.07)
-0.054
(0.028^
-0.44
(0.22)
-0.19
(1.03)
-0.30
(1.41)
-2.33
(1.63)
1.17
(0.82)
SUM
-0.13
(0.35)
-0.04
(0.28)
0.067
(0.077)
2.02
(0.63)*
0.35
(3.27)
-0.79
(4.37)
5.20
(5.34)
-5.61
(2.44)*
RHM
0.16
(0.34)

0.03
(0.07)
0.10
(0.65)
-2.95
(3.26)
-6.7
(4.4)
-1.88
(5.23)
-3.26
(2.44)
DTS CS09
-0
(0
0
(0
0
(0
0
(0
2
(1
3
(1
4,
(1.
-1
(0
.041 6.73
.10) (1.84)*
.05 3.58
.74) (1.49)*
.03 0.05
.02) (0.40)
.06
.21)
.34
.02)*
.03
.37)*
.77
.71)*
.20
• 77)
OXID
-0.85
(2.18)
0.24
(1.65)
0.57
(0.48)
97.73
(3.71)*
191.51
(33.09)*
356.3
(30.8)*
370.52
(30.71)*
54.07
(14.20)*
R?
0.28
0.26
0.98
0.96
0.90
0.93
0.96
0.35

-------
                                               TABLE VI-2 (Concluded)
(9) FRNTL

(10) VEGTL

-90.9
(281.2)
-308.8
(168.4)
0.061
(0.006)*
0.011
(0.002)*
0.83
(0.43)*
-0.33
(0.23)
0.43
(1.00)
-1.66
(0.64)*
-2.28
(3.18)
4.92
(1.80)*
0.28
(3.09)
1.05
(1.85)
1
(0
0.
(0.
.74
.98)
08
60)
121
(18
136.
(10.
.3
.02)*
02
69)*
0.82

0.89

N5
oo

-------
VO
                           TABLE VI-3. ECONOMIC DAMAGE FUNCTIONS OF VEGETATION,  WITH SULFUR DIOXIDE
                                       ANNUAL MEAN LEVEL (In $1,000)£/

a
(1)
(2)
(3)
(4)
(5)
CORNL 10.
(31.
COTNL -9
(6
OVGTL -803
(181
NUSRL -780.
(350.
FRSTL-3,315
(785
.6
,0)
.4
•2)
.0
.1)*
,2
,6)*
.6
.0)*
CROPV
0.0013
(0.0007)
0.0063
(0.0002)*
0.009
(0.003)*
0.20
(0.01)*
0.065
(0.005)*
TEMB
0.
(0.
-0
(0
-0
(0
0.
(0.
015
045)
.0004
.0089)
.72
.27)*
77
51)
-3.52
(2.90)
TEMA
0.11
(0.09)
-0.05
(0.03)
-1.05
(0.77)
-0.98
(1.25)
-1.29
(1.17)
SUN
0.38
(0.32)
0.11
(0.66)
9.44
(1.95)*
7.79
(3.81)*
36.53
(8.57)*
RHM
0.21
(0.30)
0.07
(0.07)
6.82
(2.00)*
1.19
(3.86)
24.3
(8.63)*
DTS

0.02
(0.02)
-1.58
(0.67)*
1.87
(1-25)
-3.05
(2.82)
S02 (ug/m3)
0.0008
(0.0960)
0.0005
(0.0195)
0.27
(0.57)
0.40
(1.06)
2.30
(2.44)
R2
0.10
0.98
0.60
0.85
0.87

          a/  For the 10 types of vegetations, the economic damage functions for CORNL COTNL OVGTL NUSRL and
                FRSTC yields a positive S02, while the remaining regression equations contain a negative S02.
                Only those five damage functions with a positive S02 are reported here.

-------
                                TABLE VI-4. ECONOMIC DAMAGE FUNCTIONS ON TOTAL CROPS, TOTAL
                                            ORNAMENTALS AND ALL PLANTS (in $1,000)£/
CO
o

(1) TOCRL
(2) TOORL
(3) ALPLL
(4) TOCRL
(5) TOORL
(6) ALPLL
a
-375.7
(762.8)
-519.7
(965.5)
-2,251.3
(1,908.9)
-8,247.2
(2,302.4)
-5,927.4
(1,630.8)*
-14,350.5
(3,835.7)*
CROPV
0.011
(0.003)*
0.074
(0.004)
0.039
(0.006)*
0.032
(0.011)*
0.069
(0.008)*
0.05
(0.01)*
TEMB
-0.66
(1.07)
3.18
(1.38)
0.50
(2.73)
-9.07
(3.38)
-3.039
(2.41)
012.53
(5.69)*
TEMA
-6.81
(2.80)*
-1.61
(3.28)
-16.02
(6.76)*
-16.40
(9.03)
-4.03
(6.03)
-24.98
(14.52)
SUN
12.93
(8.45)
-0.91
(10.59)
18.02
(21.49)
93.30
(26.04)*
RHM
-8.12
(8.40)
-6.36
(10.63)
7.84
(20.83)
74.72
(25.07)*
61.38 46.27
(17.82)* (17.97)*
150.45
(43.62)*
128.37
(41.68)*
DTS OXID S09(ug/m3)
1.25 1,262.5
(2.77) (50.26)*
9.09 769.31
(3.42) (60.87)
9.34 1,892.46
(7.21) (121.58)*
-15.44 3.15
(8.76) (7.19)
-6.23 4.29
(5.89) (5.07)
-20.13 6.87
(15.06 (11.84)
R2
0.96
0.92
0.92
0.59
0.74
0.6

         al  Equations (1) through (3) are economic damage functions of total crops, total  ornamentals  and
               all plants with OXID as the sole pollution variable, while equations  (4)  to  (6)  are  similar
               economic damage functions with S09 rather than OXID as the sole pollution variable.

-------
are mostly compatible with a. priori expectations. Specifically, the signs of
the pollution variables are mostly correct except in equation  (1) of Table VI-
2 in which a negative sign for  OXID  appears. The negativity of  OXID  may be
substantially attributable to the multicolinearity between the two pollution
variables, CSC>2  and  OXID (r = 0.31) because  OXID  changes sign from positive
to negative immediately when  CSO   was picked up by the regression equation,±.'

     Utilizing pollution severity indexes in the regression, a wide range of
R^  is obtained, ranging from 0.25 for soybeans to 0.98 for cotton. However,
when the annual mean level of S02 was included as the sole pollution variable,
the independent variables explain a minimum of about 10 percent of the variations
in corn losses and a maximum of  98 percent of the variations in cotton losses.
The coefficients for the pollution severity indexes, i.e.,  CS02 which were
constructed on the basis of pollutant emissions, concentration rate factor and
episode days and  OXID, are mostly significant at the 1 percent level whereas
no coefficients for S02 are significant even at the 10 percent level. This re-
sult lends support to the hypothesis that it may not be appropriate to use pollu-
tion measures mostly recorded in the central city to represent countywide pollu-
tion level. Furthermore, it should be noted that the variable  DTS  was intention-
ally excluded from equation (1) of Table VT-3 to preserve the positive sign of
S02-.2/

     Using Equation (4), (5) and (6) in Table VI-4, economic damages of total
crops, total ornamentals and all plants were estimated for the 74 counties. The
results are presented in Table VI-5. The table reveals that while total crop
damages reached about $4 million in Los Angeles, Orange and San Diego counties
all in California, San Bernadino suffered the largest ornamental damages and
all plant damages in the order of $8.5 million and $10.6 million, respectively.

     Intercorrelation among explanatory variables may not constitute a serious
problem if prediction is the primary objective, provided, of course, the inter-
correlation is expected to persist in the future. However, if multicolinearity
results in an incorrect sign of the key variable, S02, a statistical interpreta-
tion of the S02 coefficient would be meaningless, and the exclusion of DTS is,
hence, warranted.
 I/ It should be noted that  RHM  was  intentionally excluded  from equation
     2 in Table VI-2 because  the inclusion of  RHM  resulted in a negative
     OXID.

 2/ When  DTS  is included, the regression equation, however, changes
     to read as follows:

     CORNL = 9.6 + 0.0013 CROW + 0.02 TEMB + 0.14 TEMA  - 0.41 SUN
            (31.3)  (0.0007)      (0.05)        (0.12)     (0.33)

           + 0.27 RHM + 0.05  DTS - 0.004 SO
              (0.35)    (0.10)     (0.097)
                                            R2 =0.10
                                       131

-------
    TABLE VI-5. ESTIMATED ECONOMIC DAMAGES OF TOTAL CROPS,  TOTAL
                ORNAMENTAL AND ALL PLANTS2/
                (In $1,000)
(1)
Counties
Jefferson, Alabama
Maricopa, Arizona
Alameda, California
Los Angelesi California
Orange, California
San Bernadino, California
San Diego, California
Fairfield, Connecticut
New Haven, Connecticut
New Castle, Delaware
Santa Rosa, Florida
Chatham, Georgia
Fulton Georgia
Honolulu, Hawaii
Cook, Illinois
Lake, Indiana
Marion, Indiana
St. Joseph, Indiana
Vanderburgh, Indiana
Polk, Iowa
Se dgwl ck , Kan sa s
Shawnee, Kansas
Wyandotte, Kansas
Boone , Kentucky
McCracken, Kentucky
Cumberland, Maine
Anne Arundel, Maryland
Bal timore, Maryland
Harford, Maryland
Howard, Maryland
Montgomery, Maryland
Prince Georges, Maryland
Berkshire, Massachusetts
Bristol, Massachusetts
Middlesex, Massachusetts
Worcester, Massachusetts
St. Louis, Missouri
Douglas, Nebraska
Lancaster, Nebraska
Rockinghara, New Hampshire
Mercer, New Jersey
•Bernalillo, New Mexico
Albany, New York
Erie, New York
Monroe, New York
Niagara, New York
Oneida, New York
Forsyth, North Carolina
Clark, Ohio
Cuyahoga, Ohio
Franklin, Ohio
Hamilton, Ohio
Jefferson, Ohio
Mahoning, Ohio
Mon tgomery , Ohio
Stark, Ohio
Sunmit, Ohio
Mul tnomah, Oregon
Indiana, Pennsylvania
Washington, Rhode Island
Greenville, South Carolina
Hamilton, Tennessee
Shelby, Tennessee
Tom Green, Texas
Nanseraond, Virginia
York, Virginia
King, Washington
Pierce, Washington
Spokane, Washington
Dane , Wisconsin
Milwaukee, Wisconsin
Natrona, Wyoming
(2)
Estimated
Total
Crop b/
Damages"
..
1,947
3,708
4,591
4,330
3,780
4,008
703
847
136
—
—
139
3,178
149
161
356
442
407
522
159
122
375
—
281
381
144
281
148
84
—
—
—
280
894
185
196
327
563
117
--
—
—
294
..
257
—
114
187
151
—
..
..
281
72
..
-.
102
..
175
—
--
267
303
	
995
803
834
711
—
1,070
340

(3)
Estimated
Total
Ornamental
Damaees^
„..
605
1,527
3,658
1,249
8,481
2,434
530
423
2
—
28
185
885
766
82
144
106
277
58
103
38
231
—
217
209
139
1,257
—
--
—
--
—
87
444
423
230
227
265
333
--
—
—
389
—
—
—
51
—
322
—
„
—
124
42
..
„
„
—
--
~
--
77
387
	
224
327
1,249
837
286
132
147

d)
Estimated
All
Plant
Damages^'
	
3,214
5,677
8,350
6,675
10,606
6,908
1,140
1,290
89
—
—
250
4,596
888 .
335
602
692
633
841
301
222
565
—
407
549
177
1,192
83
--
—
--
—
311
1,359
533
405
527
990
286
--
—
-.
835
-.
405
	
142
266
458
--
,_
„
494
180
..
	
__
-.
21
—
—
393
595
	
1,364
1,079
1,876
1,407
..
1,759
517

_§_/  "--" denotes that the estimates are ei ther Insignificant  or  unreliable.
t>/  Estimates based on equation (4) in Table VI-4.
_£/  Estimates based on equation (5) in Table VI-4.
Al  Estimates based on equation (6) in Table VI-4»
                        132

-------
      Utilizing the "average" economic damage functions  presented  in  this  section,
the changes in crop losses brought about by changes  in the pollution  or  climato-
logical variables can be easily estimated.   For the  sake of illustration;  but
without loss of generality, consider equation (6)  of Table VI-4.   The partial
elasticity of ALPLL with respect to S00 evaluated  at their mean values  (see
Table VI-6) is                        l
                    EPLSO  = 6'87 x (20-5/79°)  = °-18-
      Thus, if the SC>2 level in the air is lowered on the  average by 2 yg/
m
from 20.5 yg/m3 to 18.5 yg/m  (i.e., 10 percent reduction),  then  economic damage
to all plants, on the average, could reduce by $14,220,  $790,00 x 1.8 percent
from $790,000 to $775,780.  The partial elasticities  for other variables of in-
terest in the economic damage functions can be similarly computed, and the results
are amenable to analogous interpretation.   It should  be  noted, however, that the
estimates are based on the assumption that the presence  of any S02 is harmful to
vegetation regardless of its level of concentration.   Although  California has
been reported to have very low S02, Equations (4)  to  (6)  do  indicate the positive,
though not statistically significant, damaging effect of SC>2 on crop losses.

CONCLUDING REMARKS

     Economic  damage functions  estimated  in this  section are replete with con-
ceptual  difficulties.  The task  of translating physical  damage functions into
monetary damage  functions involves a rather anthropocentric-egocentric evalua-
tion procedure.  This is  generally the case because the  evaluation, and subse-
quently  adoption,  of the  physical damage  functions by Benedict et al. is mainly
based on our  own value judgments  rather than on any  scientific substance. Fur-
thermore, the  damages  suffered  or anticipated by  the receptors may well lead
to changes  in  the  market  behavior, and hence, the market prices  may not correctly
reflect  the welfare  loss  associated with  the physical damages.!/

     In  spite  of the various conceptual difficulties associated  with translating
physical damages into  dollar worth equivalents, economic damage  functions were
estimated for  a  variety  of vegetation in  this study. In view of  the numerous
inherent weaknesses  in the prior  study and other  conceptual and  empirical  diffi-
culties  associated with  the estimation of economic damage functions, the  damage
functions presented  in this section, though useful for  estimating possible  damage
reductions  brought about  by pollution abatement programs, should be  interpreted
and employed with  proper  caution.

     Finally,  it is  widely recognized that the best  way to  determine the  occur-
rence and severity of  an  air pollution episode is to install a network of  re-
corders  to  measure the daily and  hourly concentration of various pollutants  and
the physical  effects simultaneously. Although such nationwide networks have  been
JL/ For a  detailed  discussion  on  some  conceptual difficulties with economic dam-
     age  functions,  see Hans  Opschoor,  "Damage Functions,  Some Theoretical and
     Practical  Problems,"  in  Environmental  Damage Costs, Paris,  OECD  (1974).

                                     133

-------
TABLE VI-6. MEAN AND STANDARD DEVIATIONS OF VARIABLES
            IN VEGETATION DAMAGE FUNCTIONS^/
Variable
CORNL
SOYBL
COTNL
OVGTL
NUSRL
FLORL
FRSTL
FCROL
FRNTL
VEGTL
TOCRL
TOORL
ALPLL
CORNV
SOYBV
COTNV
OVGTV
NUSRV
FLORV
FRSTV
FCROV
•FRNTV
VEGTV
TOCRV
TOORV
ALPLV
S02
SUN
DTS
TEMA
TEMB
OXID
SC02
RHM
Mean
6.3000
3.8838
2.8176
32.3865
72.2946
150.8000
208.3392
48.7608
56.2436
58.2176
436.6257
353.8257
790.4486
1199.6392
562.8405
439.0622
992.6432
728.7108
1441.2243
2734.4284
6435.3541
1154.7284
1660.6703
11905.0000
5576.9757
17473.4527
20.4595
59.5135
34.5811
26.2027
82.4595
0.4586
0.7927
58.8108
Standard
Deviation
13.5973
11.0622
18.7760
120.5110
370.8508
622.7670
894.3120
106.6920
257.2594
197.5723
1502.4422
1334.1816
2651.6519
2347.7278
1330.7283
3088.3767
4300.7229
1755.0623
3055.6803
10586.1798
8530.7033
3037.2926
5341.7627
16015.2235
12421.3958
22764.5429
17.7148
6.6647
18.8999
24.2735
40.6061
1.0726
0.9200
6.9394

zil  The values of crop losses and crop values are
      expressed in $1,000.
                         134

-------
set up, the individual stations are unfortunately mostly located in the center
of large metropolitan areas or industrialized areas. Few stations have been lo-
cated in agricultural areas or in suburban areas where most of the vegetation
is grown. Furthermore, a substantial amount of S02 is produced by power plants
and various smelter operations which are generally located outside of SMSA's.
This difficulty of a lack of meaningful information on pollution levels in subur-
ban or rural areas has motivated earlier investigators to resort to fuel consump-
tion, number of pollution episodes, and the tendency of atmospheric conditions
to derive the air pollution damaging potential estimates. After all, it is imper-
ative to conduct research directed at obtaining information on vegetation-at-
risk isopleths for various counties in the United States, so that more reliable
economic damage estimates for vegetation can be derived for policy decisions.
                                     135

-------
                                 SECTION VII

       AGGREGATE ECONOMIC DAMAGE COSTS AND FUNCTIONS:  AN OVERALL VIEW
     Air pollution constitutes a modern problem which goes beyond the technology
of simply controlling the pollutants. The need for effective control is  generally
recognized, but arguments against control proposals also prevail. These  arguments
are mainly based on economic grounds—whether or not the cost of attaining  a
specified level of ambient air quality exceeds the economic benefit that would
be realized from a control program. The regional damage estimates developed in
the preceding six sections provide some of this much needed information, how-
ever crude it may be, for evaluating the economic feasibility of a specific air
pollution control program.

     This final section presents an overall view of the economic damages and
damage  functions of various receptors that were derived in the preceding six
sections. Further, "aggregate" economic damage functions defined with respect
to  several effect categories are developed by regressing the aggregate damages
to  the  same set of explanatory variables used earlier in the development of
the "individual" effect economic damage functions. Aggregate damage estimates
for  selected categories of damaging effects are also computed and presented.

     The  economic damage estimates for the effect categories of human health,
material, and household soiling are summarized in Table VII-1, for the 40 SMSA's
having  an S02 level equal or greater than 25 (j,g/m3- These 40 SMSA's are  listed
in  Column 1. Column 2 (HNC1) and Column 3 (HNC2) present, respectively,  the low
and  the high damage estimates of human health; the material deterioration damage
estimates of both paint and zinc as derived in Section V are summarized  in  Column
4 (MDC).  Column 5 (TNSCO) contains the total net household soiling damages  as
described in Section IV. Based upon the low and high damage estimates of human
health  presented in Columns 2 and 3, respectively, two sets of low and high ag-
gregate damage estimates for the three effect categories were estimated  and pre-
sented  in Column 6 (TNC1) and Column 7 (TNC2).

     Specifically, the following two equations were used for computing HNC1
and HNC2  for the 40 SMSA's.
              HNC1 = Maximum of (HNCSO  ,  HNCTSP)                         (VII-1)


                   HNC2  = HNCS02 + HNCTSP                               (VII-2)


where  HNCS02  and  HNCTSP  are, respectively, the net health damages attributable
to S02 and TSP. These two aggregate damage estimates were computed by summing the
mortality and morbidity costs due to S02 and TSP derived in Sections II and  III;
namely,

                                     136

-------
TABLE VII-1. ECONOMIC DAMAGES DUE TO AIR POLLUTION, BY
             RECEPTORS FOR SELECTED SMSA's
             (in $ million, 1970)
(1)
SMSA1 s
1. Akron, OH
2. Allentown, PA
3. Baltimore, MD
4. Boston, MA
5. Bridgeport, CT
6. Canton, OH
7. Charleston, WV
8. Chicago, IL
9. Cincinnati, OH
10. Cleveland, OH
11. Dayton, OH
12. Detroit, MI
13. Evansville, IN
14. Gary, IN
15. Hartford, CT
16. Jersey City, NJ
17. Johnstown, PA
18. Lawrence, MA
19. Los Angeles, CA
20. Minneapolis, MN
21. New Haven, CT
22. New York, NY
23. Newark, NJ
24. Norfolk, VA
25. Paterson, NJ
26. Peoria, IL
27. Philadelphia, PA
28. Pittsburgh, PA
29. Portland, OR
30. Providence, RI
31. Reading, PA
32. Rochester, NY
33. St. Louis, MO
34. Scranton, PA
35. Springfield, MA
36. Trenton, NJ
37- Washington, DC
38. Worcester, MA
39. York, PA
40. Youngstown, OH
Total
(2)
HNC1
10
8
48
49
3
6
3
191
22
55
18
129
2
12
12
11
4
3
123
21
3
352
39
13
7
4
107
45
13
16
5
13
44
5
12
3
48
3
4
9
1,475
(3)
HNC2
18
15
80
52
5
6
3
360
22
93
18
161
2
24
19
17
4
5
147
32
5
527
48
13
7
4
158
79
13
25
5
15
61
5
15
3
88
4
4
10
2,166
(4)
MDC
7
3
17
26
6
11
4
105
12
49
9
55
2
8
5
8
1
7
76
12
4
111
14
3
13
9
33
30
8
9
4
7
24
2
3
2
21
8
2
8
736
(5)
TNSCO
16
16
137
117
3
14
10
516
57
216
39
294
5
24
16
17
10
3
388
37
4
418
112
29
9
8
104
147
30
20
15
27
119
23
7
5
86
6
9
23
3,134
(6)
TNC1
33
27
202
192
12
31
17
812
91
320
66
478
9
44
33
36
15
13
587
70
11
881
165
45
29
21
244
222
51
45
24
47
187
30
22
10
155
17
15
40
5,349
(7)
TNG 2
41
34
234
195
14
31
17
981
91
358
66
510
9
56
40
42
15
15
611
81
13
1,056
174
45
29
21
295
256
51
54
24
49
204
30
25
10
195
18
15
41
6,045
                          137

-------
     HNCSO~ = Mortality cost due to SO 2 + morbidity cost due to SO ^

     HNCTSP = Mortality cost due to TSP + morbidity cost due to TSP


Total material damages  (MDC)  in Column 4 is the sum of deterioration dam-
ages on both materials, zinc and paint.  Specifically, it was calculated as
follows:


                         MDC = DDCZ + DDCP                            (VII-3)


with DDCZ,  and  DDCP  defined and computed previously in Section V.

     Finally,  Column  6  (TNCl) and  Column  7  (TNC2), which represent the  low and
 high human  health  damages,  respectively,  plus other  damages, were calculated
 as  follows:


                   TNCl = HNC1 + MDC + TNSCO                           (VII-4)


                   TNC2 = HNC2 + MDC  + TNSCO                           (VII-5)
      An inspection of  Table  VII-1  reveals  that while New York  and  Chicago  SMSA's
 had the largest aggregate  air  pollution  damages,  in the order  of  $1  billion,
 the smallest air pollution damages occurred  in Johnstown and York, Pennsylvania,
 in the magnitude of $15 million  in 1970.

      Total material deterioration  damage,  including deterioration  for zinc and
 paint, amounted to $0.7 billion  for the  selected  48 SMSA's  under  study.  The
 corresponding figures  for  net  household  soiling was estimated  at  $3  billion,
 respectively. The damage on  vegetation for this nation was  estimated, according
 to Benedict, to be $132 million. These damage figures  employed in  this study
 were taken from earlier studies  which were completed under  various  stringent
 assumptions.
 AGGREGATE ECONOMIC DAMAGE  FUNCTIONS

      In order to develop marginal equivalent economic damage functions for the
 purpose of predicting damage or benefit,  and for designing pollution control
 strategies,  the overall  economic costs  of human health in the presence of S02
 (HCS02) and that in the  presence of  TSP (HCTSP) were respectively regressed
 not only against pollution and relative humidity, but also against other
 relevant socioeconomic and climatological variables, e.g., PWPO, PAGE, PCOL,
 PDS,  DTS, SUN,  etc.  The least-squares  regression technique was used with
 input from the 40 sample observations for estimating the economic damage
                                     138

-------
functions.  The regression results pertaining to overall human health damage
are presented in Column 1 to Column 4 in Table VII-2.  The overall economic
damage functions for zinc and paint, for household soiling and for plants
derived in the previous sections are also presented  in the table in Columns
5, 6, 7,  8, 9, and  10.

     The  existence  of an economic  damage function does not in itself provide
us with sufficient  information  to make any policy recommendations. Quantitative
estimates of the magnitudes of  the relationship are  required. As discussed ear-
lier, this information can be obtained directly from the estimated regression
coefficients. The coefficients  in  the regression equation indicate the changes
in the dependent variable in response to a one unit  change in the associated
explanatory variable ceteris paribus. The coefficients can be used for computing
the elasticities under given conditions. A distinguishing feature of the concept
of elasticity is that it is a unit free measure of the percentage change in the
dependent variable  with respect to the percentage change in the independent var-
iable. Given the elasticity estimates, we are able to answer the question, "What
would the effect of a reduction in the pollution level be, ceteris paribus, on
the level of economic damages of various receptors?"

     Table VII-3 contains estimates of a hypothetical reduction in the air pollu-
tion concentration  level for the several pollution receptors analyzed and presented
in Table  VII-2. The first column in this table presents the dependent variables.
Column 2  shows the  estimated values of the coefficients of the S0£ or TSP var-
iables. The next two columns list  the mean values of SC^j TSP, and the economic
'damages of the various receptors.  The estimated elasticity of economic damages
of a particular receptor with respect to SC>2 or TSP; evaluated at the means of
both variables, is  found in Column 5. These elasticities indicate the percentage
change in the economic damages  that would result, on an average, from a  1 per-
cent change in S02  or TSP.

     Of particular  interest to  the policymaker is the effect of a given  discrete
change in the pollution level on the economic damages of a particular receptor.
Assuming  that the federal government is considering  the implementation of a pol-
lution control program which is expected to lower the pollution level, on  the
average,  by 10 percent, the average benefit of a receptor can be calculated
by multiplying the  coefficient  of  S02 or TSP by 0.10 times the mean value of
S02 and TSP. These  estimates can be found in Column  6.

     The  study of Table VII-3 reveals that the partial elasticities of gross
economic  damages of the receptors  included in our study vary from 0.004  to 1.28.
Furthermore, a 10 percent reduction of the air pollution level would result in
a decrease in the annual economic  damages in the range of $0.01 million  for plants
(ALPLL) to $5.26 million for the soiling effect of zinc (SDCZ).

     The  implication of our study  for pollution abatement strategies is  obvious.
Any effort to reduce the current pollution level appears to have a varyingly
significant impact  on the economic damages resulting from the harmful effects
of air pollution. Admittedly, the  implication of this study must be qualified

                                     139

-------
                                                     TABLE VII-2.  ECONOMIC DAMAGE FUNCTIONSa»b,c/
Dependent Variables HCS02

Intercept

FWPO

PAGE

COL

PD

DTS

RHM

SON

S02

TSP

MANFV

ME

YP

HU

CROPV

TEMA

TEMB
2
R
(1)
37,775
(35,512)
0.02
(0.14)
189,112
(207,266)
70
(89)


94
(196)
222
(489)


593
(78)*
















0.66
HCTSP
(2)
-9,939
(10,525)


74,828
(36,937)*
5
(16)


39
(35)
156
(94)
179
(114)


0.0003
(0.002)














0.25
HCAP1
(3)
-54,687
(42,107)


100,580
(179,860)
70
(83)


54
(156)
120

242
(563)
611
(74)*
0.00006
(0.00009)














0.69
HCAP2
(4)
-46,751
(57,923)


146,324
(204,915)
70
(89)


75
(195)
120
(518)
139
(632)
601
(77)*
0.00004
(0.00012)














0.68
SDCZ
(5)
-23,328.


.4
(19,929)










2,679.
(1,750.
-235.
(1,820.
943.
(171.
148.
(356.


43.
(3.
21.
(18.








0.










3
2)
0
4)
3
6)*
1
0)*


1
4)*
9
9)








64
DDCZ
(6)
7,562.2
(6,640.4)










86.8
(56.7)
-76.0
(59.0)
30.5
(5.5)*
47.9
(11.5)*


1.4
(0.1)*
712.6
(615.5)








0.63
SDCP
(7)
-141,199.7
(259.8)*










911.3
(235.3)*
305.3
(245.9)
69.1
(23.2)*






15.2
(2.6)*
577.2
(3.4)*






0.99
DDCP GRSOC
(8) (9)
-4,820.1 -25,621
*887.2 *(52,347


3,432
(2,199
3,766
(1,460
2
(4
90
(219
31.1 -1,219
(8.0)* (610
10.4
(8.4)
2.3
(0.8)*
226
(166
78
(2


0.50
(0.08)*
19.7
(0.1)*






0.099 0


.0
.0


.3
• 5)
.0
.0)*
.3
• 3)
.9
.3)
.8
.7)*




.4
.9)
.9
.3)*












.92
ALPLL
(10)
-14,350.5
(3,835.7)*








-20.13
(15.06)
128.37
(41.68)*
150.45
(43.62)*
6.87
(11.84)










0.05
(0.01)*
-24.98
(14.52)
12.53
(5.69)*
0.64
at  The values in the brackets are standard errors of the coefficients, with * to indicate that the coefficient Is significant at
      the 1 percent level.  The coefficients and standard errors in equations (5), (6), (7), (8) (9) and (10) are reduced by a
      factor of 10
b/  HCSO2 = Overall health cost In the presence of S02, HCTSP = overall health cost in the presence of TSP.
    HCAP1 = HCS02 + HCTSP = high health damage estimates.
    HCAP2 = Maximum (HCS02, HCTSP) = low health damage estimates, SDCZ, DDCZ, SDCP, DDCP, GRSOC and ALPLL are defined previously in
      Chapters IV, V, and VI.
cj  The sample observations for HCS02, HCTSP, HCAP1 and HCAP2 are the 40 SMSA's with SO2 level equal or greater than 25  g/m , whereas
      the sample observations for SDCZ, DDCZ, SDCP, DDCP and GRSOC are the 148 SMSA's with population greater than 250,000.   In the
      case of ALPLL, 74 counties were selected in the sample observation.

-------
                   TABLE  VII-3.  GROSS ECONOMIC DAMAGES CHANGES RESULTING FROM A
                                 10 PERCENT REDUCTION IN THE POLLUTION T,F.VF.T,a,b/

(2) (3) (4)
(1) Coefficients Mean Values of Mean Value of
Dependent S02, TSP S02, TSP Economic Damages
Variables (103) (ug/m3) ($ million)
HCS02 593 47.25
HCTSP 0.0003 100.87
HCAPl(a) 611 47.25
(b) 0.00006 100.87
HCAP2(a) 601 47.25
(b) 0.00004 100.87
SDCZ (a) 943.3 55.73
(b) 148.1 93.81
DDCZ (a) 30.5 55.73
(b) 47.9 93.81
SDCP 69.1 55.73
DDCP 2.3 55.73
GRSOC 226.4 93.81
ALPLL 6.87 20.45
5,575.7
2,431.7
8,007.4
8,007.4
6,789.2
6,789.2
107.3
107.3
3.5
3.5
150.0
3.3
434.2
0.8
(5)
Partial Elasticity
E = (2)-(3)/(4)
0.050
— —
0.004
—
0.004
— —
0.480
0.130
0.480
1.280
0.026
0.039
0.049
0.180
(6)
Economic Damage
Reduction
= 0.1-(2)-(3)
($ million)
2.80
— —
2.89
—
2.83
~—
5.26
1.39
0.17
0.45
0.39
0.01
2.12
0.01

a/  This table  is  calculated  on the  basis  of the 10 economic damage equations presented in Table VII-2.
"b/  "__" denotes value  smaller  than  $10,000.

-------
by several theoretical and empirical factors. As discussed in the previous sec-
tions, the major difficulties often encountered in estimating air pollution dam-
ages include the lack of knowledge regarding the shapes of functions describing
the relationship between air pollution and various receptors, and the lack of
a satisfactory theoretical model specifying the way air pollution affects various
receptors. The impossibility of accounting for all major factors which might
affect various receptors, the lack of reliable formulations used for translating
physical damages into monetary terms, and the presence of numerous econometric
problems have also caused concern to investigators.

     Despite the existence of these difficulties, this study represents a major
step forward in our knowledge of pollution damages in that it seems to be the
first attempt to construct essential frameworks of the physical and economic
damage functions to calculate comparable regional damage estimates for the sev-
eral important receptors—human health, material, and household soiling, however
tentative  they may be. More importantly, various aggregate economic damage func-
tions instrumental for transforming the multifarious aspects of the pollution
problem into a single, homogeneous monetary unit are tentatively derived and
illustrated. It is hoped  that these will be useful to policymakers as they make
decisions  on the implementation of programs to achieve "optimal" (where social
MR =  social MC) pollution  levels for this country, although proper caution must
be  exercised in interpreting and employing the various economic damage functions
presented  in this study.

      Finally, it should be noted that although the availability of information
on  average or marginal damages is  instrumental in determining the optimal na-
tional or  regional pollution control strategies, the current problem is far more
complex than the question  of balancing the benefits to polluters against damages
inflicted  on the receptors. The issues are pressing and not yet well specified.
The basic  difficulty  in applying the recent research findings to accurately
estimate  the air pollution damage  cost stems from our ignorance about the  recep-
tors  at risk to air pollution. So  far, few attempts have been made to identify
who suffers, to what  extent, from which sources, and in what regions.JL^ At this
moment, updating and  expansion of  the available crude estimates, which are gener-
ally  restricted to certain regions, are urgently needed. To  identify the popula-
tion  at risk to air pollution, and to measure the damage specifically for  pol-
luted regions are apparently the most logical steps in the area of  future  re-
search.
_!/  We  are  aware  of  only one  study in the area of  estimating population at risk.
     Namely,  1stvan Jakaces  and G. Bradford  Shea,  Estimation  of  Human Population^
     at-Risk  to  Existing Levels of Air Quality, Enviro  Control,  Inc., Rockville,
     Maryland (February 1975). This study  reports  the number  of  people within
     each  major  social and economic classification who  were  exposed to 1973 levels
     of  various  air pollutants within each standard metropolitan statistical
     area  and EPA regions. Estimates of the  population  at  risk for other major
     receptors,  e.g., material and vegetation, have not been  derived to date.

                                     142

-------
                               SECTION VIII

                                REFERENCES
Ashley, D. J. B., "Environmental Factors in the Aetiology of Gastric Cancer,"
  Brit. J. Prev. Soc. Med. 23, 187 (1959).

Barrett, L. B. and Thomas Waddell, Cost of Air Pollution Damage:   A State^
  Report,  (Research Triangle Park, N. C.:  National Environmental Research
  Center, 1973); and Thomas E. Waddell, The Economic Damage of Air Pollution,
  (Washington B.C.:  Washington Environmental Research Center, U.W. Environ-
  mental Protection Agency, 1974).

Beloin, Norman J., "Fading of Dyed Fabric Exposed to Air Pollutants,"
  Journal of Textile and Colorist, Volume 5, No. 7, July 1973.

Beloin, Norman J., "Fading of Dyed Fabrics by Air Pollution," Journal of
  Textile Chemist and Colorist, Volume 4, Nos. 3, March 1972.

Beloin, Norman J. and Fred H. Haynie, "Soiling of Building Material,"  Air
  Pollution Control Association Journal,  April 1975.

Benedict, H. M. and W. H. Breen, "Use of Weeds as a Means of Evaluating
  Vegetation Damage Caused by Air Pollution," Proc. Nat Air Poll Symposium,
  3rd Symp., Pasadena, Calif., 1955.

Benedict, H. M., C. J. Miller and R. E. Olson, Stanford Research Institute,
  Economic Impact of Air Pollutants on Plants in the United States, Coor-
  dinating Research Council, New York, New York, Final Report, November
  1971.  And H. M. Benedict, C. J. Miller and J. J. 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, Environmental Protection
  Agency, Research Triangle Park, North Carolina, Final Report, July 1973.

Booz-Allen and Hamilton, Incorporated, Study to Determine Residentail
  Soiling Costs of Particulate Air Pollution, Washington, D. C., October
  1970.

Buck, S. F.  and D.  A. Brown, "Mortality from Lung Cancer and Bronchitis in
  Relation to Smoke and Sulfur Dioxide Concentration, Population Density and
  Social Index," Tob. Res. Counc., Research Paper No. 7 (1964).

Buechley, R. W., Arch. Env. Health. 27/3):137,  (1971).

Burgess, S. G. and C. W. Shaddick, "Bronchitis and Air Pollution,"  Roy.  Soc.
  Health J. 79, 10 (1959).

                                    143

-------
Clemmensen, J. and J. Nielson, "The Social Distribution of Cancer in
  Copenhagen, 1943 to 1947," Brit.  J.  Cancer 5, 159 (1951).

Coordinating Committee on Air Quality  Studies,  National Academy of Sciences
  National Academy of Engineering,  Committee on Public Works, U.S. Senate,
  "Air Quality and Automative Emissions Control," Serial No. 93-24,
  September 1974.

Cornwall, C. J. and P. A. B. Raffle, "Bronchitis-Sickness Absence in London
  Transport,"  Brit. J. Ind. Med. 18,  24 (1961).

Cremeans, John E. and Frank W. Segel,  "National Expenditures for Pollution
  Abatement and Control, 1972," Survey of Current Business, March 1975.

Dean, G., "Lung Cancer and Bronchitis  in Northern Ireland,"  Brit. Med. J.
  1,  1506  (1966).

Dienemann, Paul F.,  "Estimating Cost Uncertainty Using Monte Carlo Technique,"
  Memorandum  RM-4854-PR, Rand Corporation  (January 1966).

Dole, Malcolm Jr.,  "The Economics of the Deterioration Dilemma," Proceeding
  of the 68th Annual Meeting  of  the Air Pollution Control Association,  1975.

Douglas, J. W. B. and  R. E. Walker, "Air Pollution and Respiratory Infection
  in Children," British Journal  of Pre. Soc. Med., (1967),  21,  7-16.

Enterline, P. E., A. E. Rikli, H. I. Sauer, and M.  Hyman, "Death Rates for
  Coronary Heart Diseas in Metropolitan and Other Areas,"  Pub. Health Rep.
  75, 759 (1960).

Environmental Damage Costs, Record of  a Seminar Held  at the OECD in August
  1972.  Paris:  Organization for Economic Cooperation and Development,
  1974.

Farrar, D. E. and R. R. Glauber,  "Multicolinearity in Regression Analysis:
  The Problem Revisited," Review of Economics and Statistics, 49 (February
  1967) pp. 92-107.

Feliciano, A., 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-
  00078.  October 1972.

Ferris, B. J., Jr.,  "Tests to Assess Effects of Low Levels of Air Pollutants
  on Human Health,"  Archives of Environmental Health, 21 (1970).
                                    144

-------
Ferris, B. G. Jr. and J. S. Whittenberger, M.  Engl.  J.  Med.,  275,  1413
  (1966).

Fink, F. W., F. H. Buttner and  W. K.  Boyd, Technical-Economic  Evalua-
  tion of Air Poolution Corrosion Costs on Metals in the  United States,
  Battelle Memorial Institute, Columbus Laboratories,  Columbus, Ohio
  43211, 1971.

Gardner, M. J., M. D. Crawford, J. N.  Morris,  "Patterns of Mortality  in
  Middle and Early Old Age in the County Boroughs of England  and Wales,"
  Brit. J. Prev. Soc. Med. 23, 133 (1969).

Gerhard, John and Fred H. Haynie, "Air Pollution Effects  on Catastrophic
  Failure of Metals," manuscript, Environmental Protection Agency,  November
  1974.

Gillette, Donald G. and James B. Upham, "Material Damage  from S02  A
  Reassessment," manuscript, National  Environmental Research  Center,
  Research Triangle Park, Environmental Protection Agency, July 1973.

Glasser, M.,  L Greenburg and  F. Field, "Mortality and Morbidity During
  a Period of High Levels of Air Pollution, New York,  November  23-25,  1966,"
  Archieves of Environmental Health. Vol. 15,  pp. 684-694 (1967).
Glejser, H.,  "A New Test for Heteroscedasticity," Journal of  American
  Statistical Association,  64:316-323,  (1969).

Goldberger, Arthur S., Econometric Theory, New York; John Wiley & Sons,  Inc.,
  1964.

Goldsmith, J. R., Arch. Environ. Health, 18, 516 (1969).

Goldsmith, J. R., Medical Thoracalis 22, 1 (1965).

Greenburg, L., et al., "Air Pollution and Morbidity in New York City,"
  Journal of American Medical Association, 182, 161 (1962).

Greenburg, L. M. , F. Field, J. T. Reed and M.  Glasser, "Air Pollution and
  Cancer Mortality, Study on Staten Island, New York,"  Arch. Environ. Health
  15, 356  (1967).

Greenburg, L., M. B. Jacobs, B. M. Drolette, F. Field and M.  M. Braverman,
  "Report of an Air Pollution Incident in New York City, November 1953,"
  Public Health Reports, Vol. 77, pp.   7-16, 1962.

Griswold, M. H., C. S. Wilder,  S. J. Cutler and E. S. Pollack, Cancer in
  Connecticut.  1935-1951,   Hartford, Connecticut State Department of Health
  (1955).

                                    145

-------
Guderian, R. ,  "Method to Determine SC>2 Tolerance Limits for Agricultural
  and Forestry Cultures in the Open Countryside Experiments in Biersdorf,"
  (in German)  Staub.  20 (9):   334-337, 1960.

Haitovsky, Y., "Multicolinearity in Regression Analysis:   Comment," Review
  of Economics and Statistics, pp. 486-489 (November 1969).

Hammond, E. C., "Epidemiological Evidence on the Effects  of Air Pollution,"
  60th Annual Meeting, APCA, No. 67-68, 1967.

Haynie, F. H., "Estimation of Cost of Air Pollution as the Result of
  Corrosion of Galvanized Steel,"  U.S. Environmental Protection Agency,
  Research Triangle Park, N.C., Unpublished Report, 1973.

Haynie, F. H., "The Economics of Clean Air in Perspective,"  Material
  Protection and Performance, Volume 13, No. 4, April 1974, pp. 33-38.

Haynie, F. H. and J. B. Upham, "Correlation Between Corrosion Behavior of
  Steel and Atmospheric Pollution Data," American Society  for Testing and
  Material, ASTM STP 55, 1974.

Haynie, F. H. and James B. Upham, "Effects of Atmospheric  Pollutants on
  Corrosion Behavior of Steel," Material Protection and Performance,
  Volume  9, No. 11, November 1971.

Haynie, F. H. and J. Upham, "Effects of Atmospheric Sulfur Dioxide on the
  Corrosion of Zinc,"  Material Protection and Performance, Volume 9, No. 8,
  August  1970.

Higgins, I. T. T., "Air pollution and Chronic Respiratory  Disease," ASHRAE J.
  37  (August 1966).

Hodgson, Thomas A., Jr., "Short-Term Effects of Air Pollution and Mortality
  in New York City,"  Environmental Science and Technology, Vol. 4, pp.
  589-597  (1970).

Holland, W. W. and D. D. Reid, "The Urban Factor in Chronic Bronchitis,"
  Lancet  1,45 (1965).

Hu, Teh-Wei, Econometrics, Baltimore; University Park Press, 1973.

Ishikawa, S.,  D.  H. Bowden, V. Fisher and J. P. Wyatt, "The Emphysema
  Profile in Two Midwestern Cities in North America," Arch. Environ.
  Health 18, 660 (1969).

Jaksch, John A.,  "Some Economic Damages to Human Health Resulting from the
  Catalytic Converter."  Air Pollution Control Association Proceedings. 1975.

                                  146

-------
Jaksch, John and Herbert Stoevener, Outpatient Medical Costs Related to
  Air Pollution in the Portland, Oregon Area, Washington Environmental
  Research Center, United States Environmental Protection Agency (July
  1974).

Johnston, J., Econometric Methods. New York; McGraw-Hill Book Company,
  1963.

Jones, A. Craig, "Studies to Determine the Costs of Soiling Due to Air
  Pollution:  An Evaluation," in:  Economics of Air and Water Pollution.
  Virginia Polytechnic Institute, Water Resources Center, Blacksburg,
  Virginia, April 1969.

Krishna, Kumar T., "Multicolinearity in Regression Analysis," Review of
  Economics and Statistics, pp. 365-366 (August 1975).

Kneese, Allen V-, "Pollution and Pricing," American Economic Review,
  December 1972.

Koshal, R. K. and  M. Koshal, "Air Pollution and the Respiratory Disease
  Mortality in the U.S.--A Quantitative Study."  Social Indicator Research,
  Vol. 1, No. 3 (December 1974), pp. 263-278.

Lave, Lester B., "Air Pollution Damage:  Some Difficulties in Estimating
  the Value of Abatement," in Allen V- Kneese and Blair T. Bower, eds.,
  Environmental Quality Analysis; Baltimore:  John Hopkins Press for
  Resources for the Future, 1972, pp. 213-242.

Lave, Lester B. and E. P. Seskin, "Air Pollution and Human Health," Science,
  169, 723 (1970).

Lave, Lester B. and E. P- Seskin, "An Analysis of the Association Between
  U.S. Mortality and Air Pollution," Journal of American Statistical Associa-
  tion, Vol. 68, (June 1973).

Lepper, M. H.,  N.  Shioura, B. Carnow, S.  Andelman and  L. Lehrer, "Respiratory
  Disease in an Urban Environment," Industr. Med. 38, 36, 1969.

Lerner, A. P-,  "Priorities and Pollution:  Comment," American Economic
  Review, (September 1974).

Leung, Steve, Elliot Goldstein, and Normal Dalkey, Human Health Damage
  from Mobile Source Air Pollution. California Air Resources Board,
  Sacramento, California (July 1974).
                                  147

-------
Levin, M. L., et. al., "Cancer Incidence in Urban and Rural Areas of New
  York State," J. Nat. Cancer Inst.  24, 1243 (1960).

Liu, Ben-Chieh, "Functions of Air Pollution Damages on Human Health,"  Air
  Pollution Control Association Proceedings, (1975).

Liu, Ben-Chieh, Quality of Life Indicators in U.S. Metropolitan Areas, 1970.
  Washington, D.C., Government Printing Office,  1975; New York,  The Praeger
  Publishers (1976).

Liu, Ben-Chieh and Eden S. H. Yu, "Mortality and Mr Pollution:  Revisited,"
  Journal of the Mr Pollution Control Association (in press).

Lunn, J. E., J. Knowelden, and A. J. Handyside,  "Patterns of Respiratory
  Illness in Sheffield Infant School Children,"  Brit. J. Prev. Soc. (1967),
  21, 7-16.

Manos, N. E. and G. F. Fisher, "An Index of Air  Pollution and its Relation
  to Health," J. Air Pollut. Contr. Ass.  9, 5 (1959).

McCarrol, J. and W. Bradley, "Excess Mortality as an Indicator of Health
  Effects of Air Pollution,"  American Journal of Public Health, Vol. 56,
  pp. 1933-1942  (1966).

Michelson, Irving and Boris Tourin, "Comparative Method for Studying Costs
  of Air Pollution,"  Public Health Report, 81,  June 1966.

Michelson, Irving and Boris Tourin, The Household Costs of Air Pollution in
  Connecticut, the Connecticut State Department  of Health and Environmental
  Health and Safety Research Associate, Hartford, Connecticut, October 1968.

Middleton, J. T. and A. 0. Paulus, "The Identification and Distribution of
  Air Pollutant Through Plant Response."  AMA Archives of Industrial Health,
  14: 526-532, December 1956.

Midwest Research Institute, System Analysis of the Effects of Air Pollution
  on Materials; January 1970.

Millecan, A. A., 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.

Mills, C. A., "Urban Air Pollution and Respiratory Diseases,"  Amer. J. Hyg.
  37, 131 (1943).

Mueller, W. J. and P. B. Stickney, A Survey and  Economic Assessment of the
  Effects of Air Pollution on Elastomers,  Battelle Memorial Institute,
  Columbus Laboratories, Columbus, Ohio   43201, 1970.

Naegele, John A., ed., Air Pollution Damage to Vegetation,  Washington  D.C.
  American Chemical Society, 1973.

                                    148

-------
Naegcle, J. A., W. A. Fcder and C. J. Brandt, Assessment of Air Pollution
  Damage to Vegetation in New England:   June 1971 -  July 1972,  Environ-
  mental Protection Agency, Research Triangle Park,  North Carolina,  Final
  Report, Contract Number 68-02-0084, August 1972.

National Academy of Science, Air Quality and Automobile Emission Control
  Volumes 1, 2, 2, and 4 (September 1974).

O'Connor, John J., Jr., The Economic Cost of the Smoke Nuisance to
  Pittsburgh, Mellon Institute of Industrial Research and School of
  Specific Industries, University of Pittsburgh, Pittsburgh, Pennsyl-
  vania (1913).

O'Gara, P. J., "Sulphur Dioxide and Fume Problems and Their Solutions,"
  J. Ind. Eng. Chem. 14:  744,  1922.

O'Hagan, John and Brendan McCabe, "Tests for the Severity of Multicolinearity
  in Regression Analysis:  A Comment,"  Review of Economics and  Statistics,
  pp.  318-370 (August 1975).

Park, R. E., "Estimation  With  Heteroscedastic Terms,  " Econometrica, 34  (1966).

Park, William R.,  The Economic Impact of SO? Emission in Ohio,  Midwest
  Research Institute Report (January 1974).

Pell, E. J. , 1972  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, June 1973.

Peltzman, Sam and  T. Nicolaus  Tideman,  "Local Versus National Pollution
  Control:   Note," American Economic Review, (December 1972).

Petrilli, F. L.,  G. Agnese  and S. Kanitz, Arch, Environ. Health, 12, 733
  (1966).

Public Health Service,  Current  Estimates  From  the Health Interview Survey United
  States,  1970. Vital  and  Health Statistics,  Series 10, Number 72, Rockville,
  Maryland,  Health Services and Mental  Health  Administration, Department  of Health,
  Education  and Welfare,  August 1973.

Rice, Dorothy P.,  Estimating the Costs of Illness.  U.S. Department of Health,
  Education and Welfare, 1966.

Ridker, Ronald G., Economic Costs of Air Pollution (New York:  Frederick A.
  Praeger. 1967).

Rosenbaum, S., "Home Localities  of National Servicemen with Respiratory
  Disease; Brit. J. Prev.  Soc. Med. 15,  61  (1961).

Rust-Oleum Corporation, The Rust  Index and What it Means, Evanston,
  Illinois,  1974.
                                      149

-------
Salvin, W. S., "Textile Pollution Loss is in Billions," Raleigh News and
  Observers, March 29, 1970, Section 4, p. 10.

Schimmel, Herbert and Leonard Greenburg, " A Study of the Relation of
  Pollution to Mortality, New York City, 1963-1968," Journal of the
  Air Pollution_Control Association, Vol. 22, No. 8 (August 1972).

Shrimper, R. A., "Estimating Benefits of Reduced Mortality and Morbidity
  Associated with Improved Air Quality," Manuscritp, 1975.

Shy, Y. Carl, et al., Health Consequences of Sulfur Dioxide;  A Report
  from CHESS, 1970 to 1971.  Human  Studies Laboratory, U.S. Environmental
  Protection Agency, May 1974.

Shy, Carl M. and John F. Finklea, "Air Pollution Affects Community Health,"
  Environmental Science and Technology, March 1973.

Skalpe, I. 0., "Long-Term Effects of Sulfur Dioxide Exposure in Pulp Mills,"
  Brit. J. Ind. Med. 21, 69 (1964).

Smith, V. Kerry, "Mortality-Air Pollution Relationships: A Comment," Journal
  of American Statistical Association (June 1975).

Smith, V. Kerry and Timothy A. Deyak, "Measuring the Impact of Air Pollution
  on Property Values," Journal of Regional Science, Vol. 15, No. 3 (1975).

Speizer, F. E. and B. G. Ferris, Jr., "The Prevalence  of Chronic Nonspecific
  Respiratory Disease in Road Tunnel Employees," Amer.  Rev.  Resp.  Pis.
  88, 204 (1963).

Spence, J. W. and F. H.  Haynie,  "Chemical Attack and Economic Assessment
  of Air Pollutants on Exterior  Paints," Journal of Paint Technology,
  Volume 44, November 1972.

Spence, J. F. Haynie and J. Upham, "Effects of Pollutants on Weathering
  Steel:  A Chamber Study," Manuscript,  Environmental  Protection Agency,
  January 1975.

Spence, J. W. and F. H.  Haynie,  Paint Technology and Air Pollution:  A
  Survey and Economic Assessment,  Environmental Protection Agency, Office
  of Air Programs,  Research Triangle Park, North Carolina, February 1972.

Spence,  J. W. and F. H.  Haynie,  "Pitting of Galvanized Steel in Controlled
  Clean Air Environments,"  presented at ASTM 1974 Material Engineering
  Congress,  Detroit, Michigan, pp. 21-24, October 1974.
                                    150

-------
Spence, J. j. Upham and F. Haynie, "The Effect of Pollutant on Galvanized
  Steel:  A Chamber Study," manuscript, Environmental Protection Agency,
  January 1975.

Stein, Jerome L., "Priorities and Pollution:   Reply," American Economic
  Review, (December 1974).

Stephen, E.  R., E. R. Darley, 0. C. Trylor and W. E.  Scott, "Photochemical
  Reaction Products in Air Pollution, Proc. API,  4 (III):   325-338,  1960.

Stickney, P.  B. W. J. Mueller and J.  W. Spence, "Pollution Versus Rubber,"
  Rubber Age, 103 (9), September 1971.

Stocks, P-,  "Cancer and Bronchitis Mortality in Relation to Atmospheric
  Deposit and Smoke,"  Brit.  Med. J.   1, 74 (1959).

Stocks, P.,  "On the Relations Between Atmospheric Pollution in Urban and
  Rural Localities and Motality from Cancer,  Bronchitis, Prewmonic,  with
  Particular Reference to 3:4 Benzopyrene, Beryllium, Molybdenum, Vanadium
  and Arsenic," Brit. J. Cancer 14, 397 (1960).

Stocks, P.,  "Recent Epidemiological Studies of Lung  Cancer Mortality,
  Cigarette  Smoking and Air Pollution, with Discussion of a New Hypothesis
  of Causation," Brit. J.  Cancer 20,  595 (1966).

Stoeckhardt,  A., A Untersuchungen ueber die Schaedliche Einwirkung des
  Huetten-und Steinkohlenrauches auf das Wachsthum des Pflanzen,
  insbesondere der Fichte  und Tann.  Tharandt.  Forstl. Jahrb. 21: 218-
  254, 1871.

Stratmann, H., "Fidd Experiments to Determine the Effects of S02 on
  Vegetation, (in German)  Forschungsberichte des  Landes Nordrhein -
  Westfalen.   Essen, W. Germany - No 1984. 1963.

Sultz, H.  A., J. G. Feldman,  E. R. Schlesinger, and  W. E. Mosher, "An
  Effect of  Continued Exposure to Air Pollution on the Incidence of Chronic
  Childhood  Allergic Disease," presented at the 97th Annual Meeting of the
  American Public Health Association, Epidemiology Section, Philadelphia,
  Pa., November 11, 1969.

Thomas, Meyer D., "Effects of Air Pollution on Plants," in Air Pollution,
  World Health Organization, Columbia University Press, New York, 1961.

Thomas, M. D. and G. R. Hill, Jr., "Absorption of Sulfur Dioxide by Alfalfa
  and As Relation to Leaf Injury," Plant Physiology. 10: 291-307, 1935.

Toyama, T.,  "Air Pollution and Health Effects in Japan," Arch. Environ
          ,   153  (1964).                                         ~

                                    151

-------
Toyama, T. and Y. Tomono, Japanese Journal Public Health, (1961), 8, 659.

Uhlig, H. H., "The Cost of Corrorsion in the United States," Corrosion,
  51  (1):  29-33, 1950.

Upham, James B., Fred H. Haynie and John W. Spence, "Fading of Selected
  Drapery Fabrics by Air Pollutants," manuscript, Environmental Protection
  Agency, January 1975.

U.S. Environmental Protection Agency, Effect of Sulfur Dioxide in the
  Atmospher on Vegetation, National Environmental Research Center,
  Research Trinagle Park, North Carolina, 27711, September 1973.

Weidensaul, T. C. and N. L. Lacasse, "Results of the Statewide Survey of
  Air Pollution Damage to Vegetation," Presented at the 63rd Annual
  Meeting of the Air Pollution Control Association, St. Louis, 1970.

Wichers, C. Robert, "The Detection of Multicolinearity:  A Comment,"
  Review of Economics and Statistics, pp. 366-368 (August 1975).

Winkelstein, W., Jr. and S. Kantor, "Respiratory System and Air Pollution
  in  an Urban Population of Northeastern United States," Arch. Environ.
  Health 18, 760 (1969).

Wilkenstein, W., Jr. and S. Kantor, "Stomach Cancer, Positive Association
  With Suspended Particulate Air Pollution," Arch. Environ. Health 18,
  544 (1969).

Winkelstein, W., Jr., S. Kantor, E. W. Davis, C. ~S. Manert and W. E. Mosher,
  "The Relationship of Air Pollution and Economic Status to Total Mortality
  and Selected Respiratory System Mortality in Men. I. Suspended Particulates,"
  Arch. Environ. Health 14, 162 (1967).

Wyzga, R. E., "A Survey of Environmental Damage Functions," in Environmental
  Damage Costs, OECD, Paris (1974) p. 60.

Yoshida, K., H. Oshime and M. Imai, "Air Pollution and Asthma in Yakkaichi,"
  Arch. Environ Health 13, 763  (1966).

Yoshida, K,  Y. Takatsuda, M. Kitabatake, H. Oshima and  M. Imai, "Air
  Pollution and its Health Effects in Yokkaichi Area-Review on Yokkaich:
  Asthma," Mie Med. J. 18, 195  (1966).

Zahn, R., "Investigation on Plant Reaction to Continuous and/or Intermittent
  Sulphur Dioxide Exposure," (in German) Staub, 23  (7): 334-352,  1963.

Zeidberg, A. D., R. J. Miltortan and E. Landau, "The Nashville Air Pollution
  Study. V. Mortality From Diseases of the Respiratory System in Relation to
  Air Pollution," Arch. Environ. Health 15, 214 (1967).

                                     152

-------
                                APPENDIX A
OPTIMAL POLICIES IN THE PRESENCE OF ENVIRONMENTAL POLLUTION:  A THEORETICAL
  FRAMEWORK

     Before we systematically present the economic damage and damage function
of air pollution for a variety of receptors,  a general equilibrium framework
explicitly incorporating the effect of environmental pollution is described in
this section. Optimal intervention policies are also derived in this framework
for policy consideration. More importantly, optimal policy prescriptions are
suggested for meeting the acceptable pollution levels predetermined by the author-
ity.

     For analytical purposes,  the following assumptions are made:—

     1. Air pollution adversely affects social welfare.

     2. There are two types of industries;  pollution emitting and pollution
nonemitting, and air pollution is a joint product of the commodities produced
by the pollution emitting industry.

     3. Air pollution adversely affects the productivity of the labor input used
in other industries.

     4. By holding capital constant, labor is the only variable factor of produc-
tion in all industries in the short run.

     The social utility function for the economy under consideration is written
as
                           U = U(XL,  X2,  A)                            (A-l)


where X^ and X£  denote, respectively, the vectors of commodities produced by
the first and the second industries.  The first industry refers to one in which
the labor productivity is adversely affected by air pollution, and the second
industry consists of those firms which, in the process of producing commodities
X2, emit pollution into the air- A represents a vector of n pollutants existing
in the air, i.e.,{A= [a^,  . . a , .   -j^}.

     The partial derivatives of  U  are subject to the following sign restric-
tions:

I/ The assumptions are made mainly for facilitating the exposition. Relaxation
~~    of any of  the postulates will not affect the conclusions.

                                      153

-------
                                            2     2
              U  = 3U/BX  > 0;       U   = a U/ax   < 0


              U2 = SU/3X2 > 0;       U22 = a U/BX2  < 0


              U  = au/aA  < 0
               A

     In view of assumption (2),  the amount  of  air pollution emitted to the air,
Ae is proportional to X£.


                                  A  = aX0                              (A-2)
                                   e     2

where  a  is a matrix with elements showing the quantity of each type of pollut-
ant being emitted per unit of the commodities  produced by the industry 2.

     Assumption (3) permits the production  function  of the first industry to
be represented by
                              X, = F,  [L,  - bAL,]                      (A-3)
where  L-^  is the amount of labor employed in industry 1, and  b  is the vector
with elements indicating the loss of efficiency in  LI  due to a unit of the
jth pollutant produced by industry 2, j = l,...,n. To ensure that net labor in-
put is positive, it is imposed that bA < 1.

     Since industry 2 is assumed to be unaffected by, or at least compensated
for, air pollution, an externality or by-product, its production function is
represented by
                                 X2 = F2 (V                          (A"4)


where  L   is the amount of labor utilized in industry 2.

     Also assume that there is a pollution control sector with the following
production function
                                          
-------
where  A^,   is  the  quantity  of  air pollution  abated  and   1/3   the amount  of  labor
utilized in the  pollution control activities.

     Thus,  the pollution existing in the air at any point of time  is  simply  the
difference  between the  quantity of pollution emitted and quantity  of  pollution
abated.
                      A = A  - A  = aX0 - A  (X)                        (A-6)
                           e    c     2    c  3
     Finally, the economy is subject to a labor availability constraint
                              Ll + L2 + L3 * L                          (A"7)


     The first order optimality conditions for this economy which is subject
to an environmental externality are.derived by maximizing  (A-l) subject  to  the
constraints (A-2), through (A-7) and
               Lt, L2, L3, Xlf X2, A> 0                                (A-8)


     Form the Lagrangean:
     - Y [aX2 - aF2(L2)] - u [A -aX£ + A^)] - w  (1^ + 1^ + 1^- I)    (A-9)


     Partially differentiating (A-9) with respect to X , X , A, L , L   and  L
yields:                                                              2      3

                 U   - X = 0

            M =  UA -x aF; - bLj_ - u = 0
                                     155

-------
                                    - w
                 50 =  (p+ya) S_F2  - w = 0                               (A-14)


                                                                       (A-15)
                 50 =  -u 5A  - w  = 0
     Note that the shadow prices of X ,  X   and  A  are, respectively, X, (3
and jj,. Both  X  and  (j,  are positive by  assuming nonsatiation in consumption
of both  X   and  X   .   (j, is negative since 5U/5A < 0. The interpretation
of equations (A-10) through (A-15) is straightforward. The optimality in the

presence of the pollution externality requires that  U  /U  =X/[J3+ a(y-u)];


  U /U  = X/(u + bLn + ^dF,  )   and w = X(l-bA) 5_F  =  -u 5_Ac =  0+ ya) 5F0 .
   1  A            l   X
     In view of (A-ll), and remembering a > 0 the optimal policy is to impose
a consumption tax of  aCii+y)  per unit of  X£. From (A-12), it is clear that
a subsidy of  n, + bL^ +X5F  should be given to consumers who suffer from the
                        5L*

air pollution. In view of (A-13), a production subsidy of  XbA  per unit of
Xl  is required for efficient production. Also in view of (A-14), a production
tax of  ya  per unit of  X£  should be imposed. In short, the optimal policies
in the presence of the environmental pollution involve a consumption and prod-
uction tax on  X2, a consumption subsidy on  A  and a production subsidy on X,.
ACCEPTABLE POLLUTION LEVEL

     Suppose the pollution level is constrained by the authority not to exceed
the statutory acceptable level. This problem amounts to introducing an addi-
tional constraint in the model.
                       A
-------
                       **                  5Fo
                       If   = O + Ya + eva)ol? - w=o
                       o-'-'n
_J?  = (-|j, -ec )
                                          - w = o
where  a  is the shadow price associated with the acceptable pollution constraint.
The constraint will be binding because otherwise the objective can be attained
without statutory regulation. This means a > 0. It is clear, in view of (A-14')
and (A-151) that the optimal interventions to constrain the pollution in the
air not to exceed the acceptable level are to apply an additional tax of aa
per unit of  X£  and a subsidy of o;  per unit of  AC  to the pollution control
sector of the economy. Thus, a penalty on the pollution producing industry coupled
with a subsidy on the pollution abatement industry is the second best optimal
combination of policies to achieve the objective of reducing the pollution concen-
tration below the "threshold" level.
                                     157

-------
                                                                      APPENDIX  B
                                                                        LIST  A
                                                   SMSA'S WITH POPULATION  OVER 500,000
01
oo
                    SMSA

 1  Akron, Ohio
 2  Albany-Schenectady-Troy, N.Y.
 3  Allentown-Bethlehem-Easton, Pa.-N.J.
 4  Anaheim-Santa Ana-Garden Grove,  Calif.
 5  Atlanta, Ga.
 6  Baltimore, Md.
 7  Birmingham, Ala.
 8  Boston, Mass.
 9  Buffalo, N.Y.
10  Chicago, 111.

11  Cincinnati, Ohio-Ky.-Ind.
12  Cleveland, Ohio
13  Columbus, Ohio
14  Dallas, Texas
15  Dayton, Ohio
16  Denver, Colo.
17  Detroit, Mich.
18  Fort Lauderdale-Hollywood, Fla.
19  Fort Worth, Texas
20  Gary-Hammond-East Chicago, Ind.

21  Grand Rapids, Mich.
22  Greensboro-Winston-Salem-High Point,
      N.C.
23  Hartford, Conn.
24  Honolulu, Hawaii
25  Houston, Texas
26  Indianapolis, Ind.
27  Jacksonville, Fla.
,28  Jersey City, N.J.
29  Kansas City, Mo.-Kans.
30  Los Angeles-Long Beach, Calif.
                                                Code
Population,  1970
  (in 1.000)
AKR
ALB
ALL
ANA
ATL
BAL
BIR
BOS
BUF
CHI
GIN
CLE
COL
DAL
DAY
DEN
DET
FOR
FOR
GAR
GRA
CRE
HAR
HON
HOU
IND
JAC
JER
KAN
LOS
679
721
544
1,420
1,390
2,071
739
2,754
1,349
6,979
1,385
2,064
916
1,556
850
1,228
4,200
620
762
633
539
604
664
629
1,985
1,110
529
609
1,254
7,032
                    SMSA

31  Louisville, Ky.-Ind.
32  Memphis, Tenn.-Ark.
33  Miami, Fla.
34  Milwaukee, Wis.
35  Minneapolis-St. Paul,  Minn.
36  Nashville-Davidson, Tenn.
37  New Orleans, La.
38  New York, N.Y.
39  Newark, N.J.
40  Norfolk-Portsmouth, Va.

41  Oklahoma City, Okla.
42  Omaha, Nebraska-Iowa
43  Paterson-Clifton-Passaic,  N.J.
44  Philadelphia, Pa.-N.J.
45  Phoenix, Ariz.
46  Pittsburgh, Pa.
47  Portland, Oreg.-Wash.
48  Providence-Pawtucket-Warwick,  R.I.-Mass.
49  Richmond, Va.
50  Rochester, N.Y.

51  Sacramento, Calif.
52  St. Louis, Mo.-111.
53  Salt Lake City, Utah
54  San Antonio, Texas
55  San Bernadino-Riverside-Ontario,  Calif.
56  San Diego, Calif.
57  San Francisco-Oakland, Calif.
58  San Jose, Calif.
59  Seattle-Everett, Wash.
60  Springfleld-Chicopee-Holyoke, Mass.-Conn.

61  Syracuse, N.Y.
62  Tampa-St. Petersburg,  Fla.
63  Toledo, Ohio-Mich.
64  Washington, D.C.-Md.-Va.
65  Youngstown-Warren,  Ohio
                                                                                                                                       Code
                                                                                                                                           Population,  1970
                                                                                                                                             (in  1.000)
LOU
MEM
MIA
MIL
MIN
NAS
NEW
NEW
NEW
NOR
OKL
OMA
PAT
PHI
PHO
PIT
POR
PRO
RIC
ROC
SAC
STL
SAL
SAN
SAN
SAN
SAN
SAN
SEA
SPR
SYR
TAM
TOL
WAS
YOU
827
770
1,268
1,404
1,814
541
1,046
11,529
1,857
681
641
540
1,359
4,818
968
2,401
1,009
911
518
883
801
2,363
558
864
1,143
1,358
3,110
1,065
1,422
530
636
1,013
693
2,861
536

-------
                                                                    LIST B
                                             SMSA'S WITH POPULATION  200,000-500,000  (M)
                     SMSA

  66  Albuquerque, N. Hex.
  67  Ann Arbor, Mich.
  68  Appleton-Oshkosh, His.
  69  Augusta, Ga.-S.C.
  70  Austin, Texas
  71  Bakersfield, Calif.
  72  Baton Rouge, La.
  73  Beaumont-Port  Authur-Orange, Texas
  74  Blnghamton, N.Y.-Pa.
  75  Bridgeport, Conn.

  76  Canton, Ohio
  77  Charleston, S.C.
  78  Charleston, W.  Va.
  79  Charlotte, N.C.
  80  Chattanooga, Tenn.-Ga.
  81  Colorado Springs, Colo.
  82  Columbia, S.C.
  83  Columbus, Ga.-Ala.
  84  Corpus  Christ!, Texas
  85  Davenport-Rock  Island-Moline, Iowa-Ill.

  86  Dee Moines, Iowa
  87  Duluth-Superior, Minn.-Wis.
  88  El Paso, Tex.
  89  Erie, Pa.
  90  Eugene,  Oreg.
  91  Evansville, Ind.-Ky.
  92  Fayetteville, N.C.
  93  Flint, Mich.
  94  Fort Wayne, Ind.
  95  Fresno, Calif.

  96 Greenville,  S.C.
  97  Hamilton-Middleton,  Ohio
  98  Harrisburg,  Pa.
  99  Huntington-Ashland,  W.  Va.-Ky.-Ohio
100  Huntsville, Ala.
101  Jackson, Miss.
102  Johnstown,  Pa.
103  Kalamazoo, Mich.
104  Knoxvllle, Tenn.
105  Lancaster, Pa.
         Population,  1970
Code       (in 1.000)

 ALB            316
 ANN            234
 APP            277
 AUG            253
 AUS            296
 BAK            329
 BAT            285
 BEA            316
 BIN            303
 BRI            389

 CAN            372
 CHA            304
 CHA            230
 CHA            409
 CHA            305
 COL            236
 COL            323
 COL            239
 COR            285
 DAV            363

 DES            286
 DUL            265
 ELP            359
 ERI            264
 BUG            213
 EVA            233
 FAY            212
 FLI            497
 FOR            280
 FRE            413

 GRE            300
 HAM            226
 BAR            411
 HUN            254
 HUN            228
 JAC            259
 JOB            263
 KAL            202
 KNO            400
 LAN            320
                  SMSA

106 Lansing, Mich.
107 Las Vegas, Nev.
108 Lawrence-Haverhill, Mass.-N.H.
109 Little Rock-North Little Rock, Ark.
110 Lorain-Elyria, Ohio
111 Lowell, Mass.
112 Macon, Ga.
113 Madison, Wis.
114 Mobile, Ala.
115 Montgomery, Ala.

116 New Haven, Conn.
117 New London-Groton-Norwich, Conn.
118 Newport News-Hampton, Va.
119 Orlando, Fla.
120 Oxnard-Ventura, Calif.
121 Pensacola, Fla.
122 Peoria, 111.
123 Raleigh, N.C.
124 Reading, Pa.
125 Rockford,  111.

126 Saginaw, Mich.
127 Salinas-Monterey, Calif.
128 Santa  Barbara,  Calif.
129 Santa  Rosa, Calif.
130 Scranton,  Pa.
131 Shreveport, La.
132 South  Bend, Ind.
133 Spokane, Wash.
134 Stamford,  Conn.
135 Stockton,  Calif.

136 Tacoma, Wash.
137 Trenton, N.J.
138 Tucson, Ariz.
139 Tulsa,  Okla.
140 Utica-Rome, N.Y.
141 Vallejo-Napa,  Calif.
142 Waterbury, Conn.
143 West Palm  Beach,  Fla.
144 Wichita, Kans.
145 Wilkes-Barre-Hazleton,  Pa.

146 Wilmington, Del.-N.J.-Md.
147 Worcester, Mass.
148 York,  Pa.
Population, 1970
Code (in 1,000)
LAN
LAS
LAW
LIT
LOR
LOW
MAC
MAD
MOB
MON
NEW
NEW
NEW
ORL
OXN
PEN
FED
RAL
REA
ROC
SAG
SAL
SAN
SAN
SCR
SHR
SOU
SPO
STA
STO
TAC
TRE
TUC
TUL
UTI
VAL
WAT
WES
WIC
WIL
WIL,
WOR
YOR
378
273
232
323
257
213
206
290
377
201
356
208
292
428
376
243
342
228
296
272
220
250
264
205
234
295
280
287
206
290
411
304
352
477
340
249
209
349
389
342
499
344
330

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
1. REPORT NO.
            EPA-600/5-76-011
                             2.
                                                           3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
 Physical and Economic  Damage  Functions for Air
 Pollutants by Receptors
                                                           5. REPORT DATE
             6. PERFORMING ORGANIZATION CODE
                         September  1976
7. AUTHOR(S)
 Ben-chieh Liu and Eden  Siu-hung Yu
                                                           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Midwest Research Institute
 425 Volker Boulevard
 Kansas City, Missouri   64110
                                                           10. PROGRAM ELEMENT NO.
             11. CONTRACT/GRANT NO.

              EPA Contract No. 68-01-2968
 12. SPONSORING AGENCY NAME AND ADDRESS
 Washington Environmental Research Center
 Office of Research and Development
 Environmental Protection Agency
 Washington, D.C.  20460
             13. TYPE OF REPORT AND PERIOD COVERED
              Final
             14. SPONSORING AGENCY CODE
                 EPA/ORD
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
 This study is primarily concerned with evaluating regional economic  damages to human
 health, material, and vegetation  and of property soiling resulting from air pollution.
 This study represents a step  forward in methodological development of  air pollution
 damage estimation. It attempts  to construct essential frameworks of  the physical and
 economic damage functions which can  be used for calculating comparable regional damage
 estimates for the several important  receptors—human health, material,  and household
 soiling—however,tentative the  damage estimates may appear to be. More importantly,
 aggregate economic damage functions  instrumental for transforming the  multifarious as-
 pects of the pollution problem  into  a single,  homogeneous monetary unit are tentatively
 derived and illustrated. It is  hoped that  these results will be of some use to guide
 policymakers as they make decisions  on the implementation of programs  to achieve "op-
 timal" pollution levels for this  country.  Given the experimental nature of the method-
 ological and statistical procedures  and the degree of uncertainty associated with the
 study results,  a great deal of  caution should  be exercised in using  the products of
 this research.
17.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.lDENTIFIERS/OPEN ENDED TERMS
                          c.  COSATI Field/Group
Physical  Damage Functions
Economic  Damage Functions
Air Pollutants  - S02,  Suspended Particulates
Receptors  -  Health,  Materials, Household
  Soiling, Vegetations
Standard Metropolitan  Statistical Areas
 egions	
 8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
 Unclassified
21. NO. OF PAGES
   172
                                              20. SECURITY CLASS {Thispage)
                                              Unclassified
                           22. PRICE
EPA Form 2220-1 (9-73)
                                            150

-------