EPA-600/5-77-012
August 1977
Socioeconomic Environmental Studies Series
IDENTIFICATION AND SPECIFICATION OF
INPUTS FOR BENEFIT-COST MODELING
OF PESTICIDE USE
Office of Air, Land, and Water Use
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports ot the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
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 fore-
casting, 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 environ-
mental quality perceptions, as well as analysis of ecological and economic im-
pacts 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 sys-
tems 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-600/5-77-012
August 1977
IDENTIFICATION AND SPECIFICATION
OF INPUTS FOR BENEFIT-COST
MODELING OF PESTICIDE USE
by
Donald J. Epp
F. Roger Tellefsen
Gary A. Shute
Robert M. Bear
Kenneth P. Wilkinson
The Pennsylvania State University
University Park, Pennsylvania 16802
Grant No. R863247-01-1
Project Officer
Darwin R. Wright
Agriculture and Non-Point Source Management Division
Office of Air, Land, and Water Use
Washington, D.C. 20460
OFFICE OF AIR, LAND, AND WATER USE
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
-------
DISCLAIMER
This report has been reviewed by the Office of Air, Land and Water Use,
U. S. Environmental Protection Agency, and approved for publication. 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.
n
-------
ABSTRACT
Benefit-cost analysis requires inputs which are inclusive, valid, quan-
tifiable, and reliable. Unfortunately, much of the data presently available
on pesticde use does not measure up to these qualifications. The purpose of
this report is to define, delineate, and improve data inputs to benefit-cost
analysis. Specific objectives established to support this purpose are:
• to provide a comprehensive classification scheme incorporating all
impacts resulting from a favorable pesticide-use decision;
• to establish conceptually valid approaches to the quantifications of
major effects of pesticide use;
• to identify and usefully display areas of inputs where quantification
is currently inappropriate and, if attempted, could produce misleading
or seriously biased results;
• to illustrate the application of benefit-cost analysis through
examples based on the use of aldrin on corn.
Proper attention to the six sections of the procedure developed in this
report will give a broad, although not necessarily complete, review of the
effects of a pesticide-use decision and will greatly assist comparison of the
benefits and costs. The six major sections are:
•A taxonomy of pesticide use effects -- This section provides an overview
of pesticide-use decision variables.
•Economic production -- Improvement in economic production is the major
benefit of pesticide use. This section identifies competing means of
evaluating production effects. Although alternative cost assumptions
are commonly used in benefit-cost modeling, the opportunity cost approach
is argued to be the only conceptually valid technique and a generally
superior procedure.
•Human health -- This section identifies statistical methods employed in
the measurement of health costs resulting from pesticide use. Because
chronic health effects are the most difficult to analyze, an emphasis is
placed on evaluation techniques for measuring costs of chronic illness
which are shown to be measurable in both dollar and utility values.
•Environmental impacts -- Relevant factors here and in the next sections
are not generally quantifiable to a benefit-cost framework. They remain,
however, as important considerations to pesticide-use decision making
11
-------
and Impacted Organism/Effects matrices are developed as a tool to insure
comprehensive consideration of environmental impacts.
•Aesthetic impacts -- This section provides a process for aesthetic impact
assessment through the use of Organism/Effects matrices.
•Distribution effects -- Like environmental and aesthetic impacts, distri-
bution effects are treated as exogenous to a benefit-cost scheme. This
final section considers the distributional effects of pesticide use when
stratified by social, income and international trade classifications.
iv
-------
CONTENTS
ABSTRACT
FIGURES
TABLES i*
ACKNOWLEDGEMENTS x
CHAPTER 1: INTRODUCTION AND FINDINGS ................. 1
The Nature and Use of Chemical Pesticides ............ 1
Purpose of Investigation ..................... "I
Regulation of the Pesticide-Use Decision ............. 1
Benefit-Cost Procedures ..................... 2
Use of money as a common unit of measure .......... 3
Objectives and Limitations of Study ............... 4
Outline of Subsequent Sections .................. 5
Major Findings and Conclusions .................. 6
CHAPTER 2: A TAXONOMY OF PESTICIDE USE EFFECTS ............ 7
The Form of Taxonomy ....................... 7
Human and Non-Human Health Effects ................ 7
Agricultural Impact of Pesticide Use ............... 11
The Effects of Pesticide Use on Materials
and Property .......................... 13
Human Health Impact Analysis ................... 13
Environmental and Aesthetic Impact Analysis ........... 13
Regulatory Control Costs ..................... 17
Distributional Aspects of Pesticide-Use Decisions ........ 17
CHAPTER 3: ECONOMIC PRODUCTION .................... 18
The Nature of Agricultural Production in
the United States ....................... 19
General Characteristics of the Agricultural
Production Model ........................ 19
Pesticide Output Information From An
Agriculture Production Model .................. 20
Methods of Analyzing Pesticide-Use Effects ............ 21
The alternative cost approach ................ 21
The opportunity cost approach ................ 23
Comparison of Methods ...................... 25
-------
CONTENTS (Continued)
Choosing Between Alternative Cost and
Opportunity Cost Methodologies 27
Additional Information Needs 29
Illustrations From Aldrin Use on Corn:
A Case Example 30
The model 30
Model results 32
Weakness of the model 39
Summary and Comment on Recent Literature 40
Bibliography for Economic Production Chapter 43
CHAPTER 4: HUMAN HEALTH 47
Review of Literature 47
The Human Health Model 48
Measuring Pesticide Related Illness 53
Acute injury 53
Chronic injury 54
Techniques for Evaluating Risk Levels 55
"No positive result" method 55
Mathematical extrapolation 56
Techniques for Evaluating a Change in
Mortality Rate 62
Foregone earnings approach 63
Other approaches 65
Inadequacy of evaluation techniques 65
Willingness to pay 66
Human Health Effects of Corn on Aldrin:
A Case Study Example 72
Bibliography for Human Health Chapter 81
CHAPTER 5: ENVIRONMENTAL IMPACTS 86
Introduction 86
General Review of Pesticide Movement and Problems 87
Insects 88
Birds 88
Effects of Pesticides on the Terrestrial Ecosystem 90
Soil invertebrates 90
Mammals 91
Fresh Water Aquatic Effects 91
Environmental Impact Assessment 92
Importance of species survival 92
Measurement of environmental effects 93
The food chain model 95
The model 95
DDT example 101
The impact matrix 103
The Delphi technique 105
vr
-------
CONTENTS (Continued)
Environmental Impacts of Aldn'n Applied to Corn:
A Case Study Example 106
Soil complex 107
Higher terrestrial trophic levels 108
Aquatic sector 110
Summary Ill
Aldrin-Dieldrin Impact Matrix 113
Matrix operation 113
Summary 117
Bibliography for Environmental Impacts Chapter 119
CHAPTER 6: AESTHETIC EFFECTS 127
Problems of Quantification 128
Aesthetic Impact Assessment 129
The Aldrin Case 131
Bibliography for Aesthetics Effects Chapter 132
CHAPTER 7: DISTRIBUTION EFFECTS 133
Geographic Distribution of Effects 133
Social Effects 134
A social accounting system 135
Community level assessment 137
Balance of Payment Effects 138
Impact of foreign trade 139
The effect on the U. S. trade balance
of restricting aldrin on corn 142
Bibliography for Distribution Effects Chapter 143
-------
FIGURES
Number Page
1 Health effects and impact matrix 8
2 Analysis of a shift in supply 24
3 Consumption regions 31
4 Change in the number of acres
of corn by region 36
5 Change in corn production by region 38
6 Option demand 49
7 Human health assessment model 50
8 Dose-risk extrapolation using
Mantel-Bryan procedure 59
9 Effect of changing confidence levels
in Mantel-Bryan procedure 61
10 Safety indifference map 68
11 General pesticide movement model
i.n fauna! sub-system 96
12 Pesticide movement in the total system 98
13 General impact matrix 104
14 Concentration of dieldrin in marine organisms 112
15 Aldrin-dieldrin impact matrix 114
16 Pesticide impacts on activities with
aesthetic components 130
vm
-------
TABLES
Number page
1 Summary of Health Effects 9
2 Summary of Impacts 10
3 Agricultural Impacts of Pesticide Decisions 12
4 Material and Property Damage Impacts of
Pesticide Decisions 14
5 Human Health Impacts of Pesticide Decisions 15
6 Natural Environment and Aesthetic Impacts
of Pesticide Decisions 16
7 Change in the Number of Acres in
Specified Crops by Region 33
8 Change in Production of Specified
Crops by Region 35
9 Change in Acres of Corn in Each Soil
Productivity Group by Region 37
10 Extrapolated Values of "Safe" Doses for
Three Different Dose-Response Curves
Describing Observed Responses in 2%
to 50% Response Range Equally Well 58
11 Fraction of Experimental Dose Using
Probit Extrapolation with Different
Slopes for an Estimated Risk of
1/100,000,000 60
12 Dose-Risk Relationships for Dietary
Dieldrin Using the Mantel-Bryan
Extrapolation Technique 74
13 Relative Composition of Diet by Food Class 75
14 Residues of Dieldrin by Food Class 77
15 Major U. S. Export Crops (1974-1975) 140
16 United States Exports of Major Crops
As A Percentage of Total United States
Production
-------
ACKNOWLEDGMENTS
Several people were particularly helpful to the authors at various
stages in the preparation of this work. Dr. Michael J. Hay was the EPA
Project Officer during most of the period of the grant and provided valuable
advice and assistance. Dr. Terry A. Ferrar, Director of the Center for the
Study of Environmental Policy, The Pennsylvania State University, also pro-
vided generous assistance in solving unusual administrative problems which
arose during the grant period.
Technical advice on pesticide effects and related biological science
material was provided by Dr. Herbert Cole, Jr., Professor of plant pathology,
The Pennsylvania State University, Dr. James W. Gillet, Terrestrial Effects
Branch, Environmental Research Lab, EPA, and Dr. Frank M. Fisher, Professor
of Biology, Rice University.
-------
CHAPTER 1
INTRODUCTION AND FINDINGS
THE NATURE AND USE OF CHEMICAL PESTICIDES
The widespread use of chemical pesticides, be they insecticides, herbi-
cides, fungicides, or rodenticides, has resulted in broad scientific and
public concern about the effects of such usage. It is widely appreciated
that pesticides play a critical role in the production of food in modern
agricultural nations, contributing to hitherto unparalleled levels of pro-
ductivity. In the home and in the community, pesticides can result in control
of noisome, unpleasant, and sometimes dangerous pests.
It is no less widely appreciated that pesticides themselves may be
dangerous. Many pesticides are generally believed to persist in the environ-
ment, and to biomagnify through the food chain with adverse or unknown con-
sequences. Some pesticides are believed to be tumorogenic or carcinogenic.
Some pesticides, in relatively small dosages, can cause acute human health
problems. Some pesticides can cause adaptations of target species into more
pesticide-resistent strains for which control is more difficult. Some
pesticides may seriously damage non-target species.
PURPOSE OF INVESTIGATION
In the broadest sense, it is the purpose of this investigation to
identify comprehensively both positive and negative consequences of pesticide
use and to specify all impacts in a manner most conducive to the evaluation
of pesticide effects. It is not the objective of this study to develop a full
benefit-cost decision model, but only to extend knowledge of the relevant
inputs to such a model.
REGULATION OF THE PESTICIDE-USE DECISION
The Federal Insecticide, Fungicide and Rodenticide Act (FIFRA), as
amended and supplemented by the Federal Environmental Pesticide Control Act
(FEPCA) of 1972, mandates the Administrator of the Environmental Protection
Agency to exercise comprehensive regulatory control over the production, dis-
tribution, and use of pesticides. The fundamental regulatory authority
accorded the Administrator is that requiring registration as a prerequisite
to "distribute, sell, offer for sale, hold for sale, ship, deliver for ship-
ment, or receive and (having so received) deliver or offer to deliver to any
-------
person . . ." (FIFRA, Section 3). As a condition of registration, the use
(e.g., home use for flying insects) must be specified in the application for
registration, along with the pesticide name, formula, pertinent test data,
and a request that the pesticide be classified under the law for general use,
restricted use, or both.
Classification is made at the discretion of the Administrator (subject
to review) accordingly as he or she determines whether or not the pesticide,
if used in accordance with instruction, will generally cause unreasonable
adverse effects on the environment (including injury to the applicator).
Where it is determined that such adverse effects would generally result, the
pesticide is classified for restricted use and may be applied only by, or
under the direct supervision of, a certified applicator in accordance with
labeling instructions. If such effects are not expected, the pesticide is
classified for general use and may be generally applied provided only that
applications be in accordance with labeling instructions.
FIFRA and FEPCA, along with regulations promulgated thereunder, provide
further for issuance of permits for experimental use of pesticides, for a
procedure for cancellation of the registration of pesticides in some or all
uses, for redassification for general use and restricted use (or both), and
for suspension of registration for some or all uses pending the outcome of
cancellation proceedings where an imminent hazard exists.
The principal criterion in the law on which the Administrator is to base
the pesticide-use decision is his assessment of the degree to which the pesti-
cide usage in question would result in "unreasonable adverse effects on the
environment." According to FIFRA, the meaning to be given this phrase is
"any unreasonable risk to man or the environment, taking into account the
economic, social, and environmental costs and benefits of the use of any
pesticide" (FIFRA, Sec. 2(bb)). The Act is thus quite clear that the
Administrator's decision is to take explicit account of costs and benefits
(economic, social, and environmental).
BENEFIT-COST PROCEDURES
In developing an analytical procedure to take explicit account of costs
and benefits, it is important to focus on the type of decision for which this
procedure is most critical. Some chemicals presented for registration have
no known harmful side effects. There is very little problem about registering
these for general use. Chemicals of this type do not present a serious
hazard even if the user misapplies them or incorrectly prepares the solution
from the purchased items. While not advocating sloppy constitution or
application of the pesticides, these chemicals do not present a significant
hazard even if carelessly handled. A second group of chemicals are extremely
dangerous. These require specialized equipment for handling and extreme
caution and careful monitoring in order to insure that side effects are kept
at a minimum. These chemicals are so hazardous that they can only be approved
for experimental use. A third group of chemicals have some harmful side
effects but are extremely useful (or potentially so) in controlling
economically important pests. The registration decision for these chemicals
-------
requires a determination of whether the pest control effects (benefits) are
sufficiently important to offset the potential harmful effects (costs) of
using these chemicals. It is for this third group that benefit-cost analysis
is required.
The ideal of the benefit-cost approach is to reduce all of the effects
to a common denominator so that a ratio of benefits to costs for selected
alternatives can be constructed and weighed. The decision rule for pesticide
use would be:
P.V. [i Benefits (Pesticide XXX)]
P.V. [I Costs (Pesticide XXX)]
1.0
where: XXX = pesticide use under regulatory review
P.V. = present value
i = summation of dollar amounts
The expression "greater than 1.0" recognizes that any situation where the
present value of the benefits stream exceeds the present value of the costs
stream is efficient.
In formulating a benefit-cost analysis, the analyst must exercise care
that the appropriate alternatives are compared in computing benefits and costs,
In pesticide-registration decisions this would include consideration of
alternative degrees of restriction (e.g., general or restricted use class-
ifications), alternative geographic restrictions (e.g., registered for use
only in particular states), as well as alternative (substitute) pesticides.
If an alternative pesticide is likely to be used if the one under consider-
ation is not registered for use, the analyst must be certain that the pro-
duction effects, human health effects and environmental effects are all
included in the analysis. Once fully specified and calculated the decision
criterion is clear: if the best ratio is greater than one, the pesticide is
socially useful; if the ratio is less than one, the pesticide should not be
registered for use.
Use of Money as a Common Unit of Measure
In this report the term benefits refers to the desirable effects which
would result from a decision to register the pesticide under review. The
term costs, on the other hand, refers to the undesirable effects that would
result from the decision to register the pesticide. Either benefits or costs
can be measured in terms other than dollars. For example, benefits from
registering a particular pesticide might be measured in bushels of grain not
destroyed by insects while the costs might be measured by the percentage
reduction in the songbird population of the area. It is clear that com-
parison of benefits and costs is very difficult when each is measured in
different units.
-------
However, in practice it is extremely difficult to reduce all pesticide-
use effects to a common unit of measure. Several sectors permit measurement
in terms of money units. These can give a quantified benefit-cost ratio in
dollar terms. Most prominent of these sectors is economic production where
most, if not all, of the effects can be measured in dollar terms. Some human
and non-human health effects can be measured in dollar terms but others cannot.
This poses a question as to whether the effects that can be quantified should
be included with a common benefit-cost ration or if all human health and non-
human health effects should be retained as a separate (exogeneous) analysis.
We find that while many effects can be reduced to dollar terms, some cannot
without risk of bias. Given the current state of the art, a benefit- cost
procedure by itself is not all conclusive: separate consideration of
exogeneous factors must be undertaken. This problem is addressed in some
detail in the next section.
OBJECTIVES AND LIMITATIONS OF STUDY
Benefit-cost analysis requires data which is inclusive, valid, quantifi-
able, and reliable. Unfortunately, much of the data presently available on
pesticide use does not measure up to these qualifications. Jhe_ purpose of
this report i_s_ _tp_ define, delineate, and_ improve data inputs^ to_ benefit-cost
analysis. Specific objectives established to support this purpose are:
•to provide a comprehensive classification scheme incorporating
all impacts resulting from a favorable pesticide use decision;
•to establish conceptually valid approaches to the quantifica-
tion of major effects of pesticide use;
•to identify and usefully display areas of inputs where
quantification is currently inappropriate and, if attempted,
could produce misleading or seriously biased results;
•to illustrate the application of benefit-cost analysis
through examples based on the use of aldrin on corn.
In dealing with econometric problems covering a broad range of modeling
applications, it becomes necessary to limit analysis to obtain meaningful
results. The major limitations accepted by this study are the following:
•Emphasis throughout is placed on agricultural-use
decisions. Commercial, industrial and home use
applications are not explicitly considered in depth.
•Although a complete taxonomy of effects is developed,
only the most critical areas of benefits and costs
are subjected to detailed analysis. For example,
production benefits and health costs are emphasized
but production costs and health benefits of pesticide
use are not fully treated.
-------
• As a result of the limited focus, a comprehensive
benefit-cost example illustrated with a general case
is inappropriate. Instead and where appropriate, the
specific impact categories considered will be illus-
trated with partial case analyses.
• The issue of discounting future benefits and costs
to present value is not considered. All flows are
presumed to be discounted at a zero rate.*
OUTLINE OF SUBSEQUENT SECTIONS
To obtain our stated objectives, we have developed a procedure with the
following six basic sections: (1) a taxonomy of pesticide use effects;
(2) economic production; (3) human health; (4) environmental impacts;
(5) aesthetic impacts; and (6) distribution effects. It is our feeling that
proper attention to these six sections and subpoints developed under them will
give a broad, although not necessarily complete, review of the effects of a
pesticide-use decision and will greatly assist comparison of the benefits and
costs. Most of the sections are relatively self-explanatory and may be
briefly described as follows:
A Taxonomy of Pesticide Use Effects -- This section provides an overview
of pesticide use decision variables.
Economic Production -- Improvement in economic production is the major
benefit of pesticide use. This section identifies competing means of
evaluating production effects. Although alternative cost assumptions are
commonly used in benefit-cost modeling, the opportunity cost approach is
argued to be a generally superior procedure.
Human Health -- This section identifies statistical methodology employed
in the measurement of health costs resulting from pesticide use. Because
chronic health effects are the most difficult to analyze, an emphasis is
placed on these effects and evaluation techniques for measuring costs of
chronic illness are presented. Acute health effects must also be included in
actual analysis, but because the methods and concepts involved are straight-
forward, they are given only passing treatment in this report.
Environmental Impacts -- Relevant factors here and in the next sections
are not generally quantifiable to a benefit-cost framework. They remain,
however, as important considerations to pesticide-use decision making. This
section develops Impacted Organism/Effects matrices as a tool to insure com-
prehensive consideration of environmental and aesthetic impacts.
*
This assumption was made at the suggestion of the Environmental Protection
Agency. While the assumption obviously begs some significant questions, it
was agreed that discussion of these topics should be left to some other report.
-------
Aesthetic Impacts -- This section provides a process for aesthetic impact
assessment through the use of Organism/Effects matrices.
Distribution Effects -- Like environmental and aesthetic impacts, distri-
bution effects are treated as exogenous to a benefit-cost scheme. This final
section considers the distributional effects of pesticide use when stratified
by social, income, and international trade classifications.
For this study we have developed procedures whereby most of the signifi-
cant human health effects can be quantified in dollar terms suitable for
inclusion with the economic production effects in a benefit-cost ratio.
Because the current state-of-the-art does not permit many of the non-human health
effects of significance to be quantified, we have chosen to maintain all of
the effects on non-human species on a descriptive basis rather than incor-
porating portions of it into a benefit-cost ratio. The first two sections
provide the basic core and all of the quantified benefit-cost ratio consider-
ations fpr the study. The distribution sections and the section on costs of
environmental and aesthetic impacts will be presented predominantly as dis-
plays of descriptive materials.
MAJOR FINDINGS AND CONCLUSIONS
Specific findings are found in each major section of the report. In
summary, findings may be classified as related to one of the following three
broad areas of outcomes and conclusions:
• Opportunity cost methodology is the only conceptually
valid technique in benefit-cost modeling. Consumer
surplus is the output of opportunity cost analysis
and is the proper measure of both economic and health
effects associated with pesticide use.
• Chronic health risks can be reduced to both dollar and
utility values. The study reviews the recent develop-
ment of a method for determining such values.
• The present state-of-the-art is not sufficient to adequately
quantify value in the areas of environmental and aesthetic
impacts. However, the treatment developed in this study
takes the first steps toward a conceptually sound quanti-
fied measurement.
-------
CHAPTER 2
A TAXONOMY OF PESTICIDE USE EFFECTS
To be certain of comprehensive coverage of the effects of a pesticide-
use decision, the analytical procedure must include a taxonomy of effects.
This taxonomy should include beneficial as well as harmful effects and must
include all categories of effects which might be significant in a pesticide-
use decision. Obviously, no single decision will require detailed analysis
of every point in the taxonomy. Rather, the analyst will examine a given
use decision in light of the taxonomy to select the effects relevant to the
particular pesticide and use under consideration.
A taxonomy to display benefits and costs for a pesticide-use decision
has been developed in a matrix format. It is believed that this format
makes most clear the effects of a pesticide and foreshadows the organization
of subsequent sections of this report which describe methods appropriate for
use in analysis of the effects. It also assures that all relevant entities
receiving benefits or incurring costs due to pesticide use are considered,
and that all critical economic or environmental impacts are identified.
An important contribution of the taxonomy approach is its assurance that no
relevant impact is inadvertently excluded from consideration.
THE FORM OF TAXONOMY
Figure 1 outlines the major headings of the taxonomy, broadly divided
into health effects and impacts. Table 1 divides the health effects into
human and non-human components, while Table 2 divides the impacts into more
specific subheadings. Subsequent tables give a more detailed breakdown of
each impact identified in Table 2. Thus, to construct the complete taxonomy
with all details included, it is only necessary to place Table 1 and Tables
3 to 6 in the appropriate spaces as noted in Figure 1.
HUMAN AND NON-HUMAN HEALTH EFFECTS
A pesticide is designed to kill an unwanted organism. A perfect
pesticide would stop there and would have no effect on non-target organisms
in the ecosystem. However, such perfect pesticides do not yet exist and
presently available pesticides can and do have certain deleterious effects
on non-target organisms. It is the lack of specificity in these pesticides
which requires a careful examination of effects on all organisms which are
affected by the pesticide in question. If an organism is affected by a
pesticide but its economic or environmental impact is small, it can be
-------
IMPACTS
Acute
13
C/1
LU
Chronic
(Table 1
Part I)
Acute
fO
3
^
I
C
O
Chronic
(Table 1
Part II)
re
o
01
(Table 3)
O)
CT>
03
X3 E
C «
03 Q
tO •(->
••- i.
S- 0)
O) Q.
4-> O
ca i.
o
(Table 4)
(O
C
-------
TABLE 1. SUMMARY OF HEALTH EFFECTS
I. Human
A. Acute
1. Death due to direct exposure
2. Illness
B. Chronic
1. Mutagen
2. Teratogen
3. Oncogen
4. Reproductive Impairment
5. Chronic Physiologic Malfunction
II. Non-Human
A. Acute
1. Death due to direct exposure
2. Contributory to death
B. Chronic
1. Mutagen
2. Teratogen
3. Oncogen
4. Reproductive Impairment
5. Chronic Physiologic Malfunction
-------
TABLE 2. SUMMARY OF IMPACTS
I. Agricultural
A. Yield of crop
B. Quality of crop
C. Cost of production
D. Quality of land
II. Material and Property Damage
A. Right-of-way maintenance
B. Structural integrity of buildings
C. Damage to commodities during storage
D. Personal belongings
III. Human Health
A. Manufacturing worker
B. Formulator worker
C. Distributor - wholesale and retail
D. Applicators
E. Non-occupationally exposed
F. Disease vector control
G. Accident attenuation
IV. Environmental and Aesthetic
A. Non-renewable resources
B. Sporting activities
C. Tourism
D. Home and gardens
10
-------
relegated to a footnote. The important point here is that the effect will
have been considered, lessening the likelihood of a future "surprise." We
have grouped this examination of species susceptibility under the general
rubric of health effects and have broken these effects into human and non-
human components (Table 1).
The human health and the non-human health components are organized in
the same form, that is, division into acute and chronic effects. The major
difference between the two health effects components is that the human health
part includes a section on acute effects which do not lead to death while
the non-human health part does not include such a section. This difference
reflects the value position that sickness and discomfort in humans is an
effect that should be considered in its own right. The same sickness and
"discomfort" in plants and other animals is not of concern unless it contrib-
utes to the death of individuals. The non-human component is meant to be
applied to animals and plants (be they terrestrial, fresh water, or esturine/
marine) exposed to the pesticide under analysis. Since a pesticide can have
varying effects depending on the organism, this component is further broken
down into acute and chronic effects to allow a more thorough evaluation.
It is intended that the human and non-human health effects be iterative.
This means that information about a non-human effect can be fed back into
other non-human effects or into the human effects category where appropriate.
This feature allows layers of effects to be worked through to their final
impact, as in the case of biomagnification, while allowing each layer to be
examined for its economic or environmental impact.
AGRICULTURAL IMPACT OF PESTICIDE USE
All crops produced for sale in the marketplace are considered as part
of agricultural production. Thus, food crops such as wheat, corn, vegeta-
bles, and fruit, fiber crops such as cotton and timber, animal feed crops
such as soybeans, oats, and hay as well as commercially produced flowers and
house plants are included in this category.
The four considerations that are important in assessing the impact of a
pesticide use decision in this area are change in crop yields, change in
crop quality, change in cost of production, and change in land quality due
to effects on soil microorganisms, soil erodability, etc. (Table 3). Col-
lectively, these four considerations make it possible to predict the change
in consumer surplus--the correct measure of benefit or cost.
For commodities which enter into international trade, it is not enough
to measure the change in domestic consumer surplus. It is also important
to obtain a measure of the change in foreign exchange earnings due to a
pesticide-use decision. For commodities such as wheat and corn where a
large percentage of the total output is exported, the balance of payments
impact could be large. The measure of this distributional impact would be
considered along with the change in domestic consumer surplus to evaluate
the agricultural impact.
11
-------
TABLE 3. AGRICULTURAL IMPACTS OF PESTICIDE DECISIONS
I. Yield - compared to best alternative method of pest control
II. Quality - compared to best alternative method of pest control
III,. Cost of Production
A. Machinery
B. Land (quantity required)
C. Fuel
D. Labor
E. Other inputs
F. Constraints on production
1. Spray intervals
2. Re-entry intervals
3. Cultural practices
IV. Land Quality
A. Soil erosion
B. Soil compaction
C. Soil microorganisms
12
-------
THE EFFECTS OF PESTICIDE USE ON MATERIALS AND PROPERTY
Included in this category are the commercial uses of maintaining rights-
of-ways and maintaining the structural integrity of buildings and other man-
made edificies as well as controlling damage to agricultural commodities
during storage and controlling damage to personal belongings. An outline
of these factors is presented in Table 4.
HUMAN HEALTH IMPACT ANALYSIS
Table 5 provides a complete list of groups who are likely to be exposed
to pesticides and possibly suffer the adverse consequences or benefits
listed under the human health effects category. Pesticide participants
typically receive direct exposure to the pesticide through occupation or use
and are most likely to be exposed to high concentrations of the chemical.
As such, they are particularly prone to acute effects and should be the first
group to experience any of the possible chronic health effects. Pesticide
bystanders typically are exposed through residues, which are either ingested,
inhaled or absorbed through the skin. It is in this type of exposure where
the feedback feature of the health effects rows (Figure 1) permits a complete
consideration of the problem. For example, pesticide run-off from crop land
into streams may result in the magnification of pesticide residues as one
moves up the food chain. At the top of the food chain, man may eat contam-
inated fish, become contaminated himself, and suffer health effects. Pesti-
cide participants can also be pesticide bystanders within the definition of
these terms. Obviously, a farmer can spray his fields and later eat contam-
inated fish from his stream.
Also included in the human health impact category are human health
benefits from pesticide use in the form of disease vector control. Health
benefits are also obtained from the reduction of human injury and death in
accidents caused by road hazards visually obstructed by vegetation. Again,
the health effects rows prevent any oversight in considering these benefits
of pesticide use.
ENVIRONMENTAL AND AESTHETIC IMPACT ANALYSIS
In this category, we assess the impact of the pesticide on species
which are not sold in a marketplace (Table 6). In many cases the impact
will be negligible until a fairly high point on the food chain is reached,
at which point the benefits or costs will be assessed. The most important
impact will be in the case where a species is threatened with extinction
due to health effects from a pesticide. Species extinction is an example
of a loss of a non-renewable resource, and the impact will have to be care-
fully assessed.
The aesthetic impact deals with people's perception of their surround-
ings. Pesticide usage can materially affect this perception with the poten-
tial for either increasing or decreasing the perceived quality. Although
13
-------
TABLE 4. MATERIAL AND PROPERTY DAMAGE IMPACTS OF PESTICIDE DECISIONS
I. Right-of-v/ay Maintenance
A. Damage to utility lines by vegetation
B. Damage to property by fire along railroads and
industrial sites
C. Damage to property by motor vehicles due to visual
obstruction of hazards by vegetation
II. Structural Integrity of Buildings
A. Wood decay - railroad ties, etc.
B. Insect damage - termites, carpenter ants, etc.
C. Rodent damage - wiring, timbers, etc.
III. Damage of Agricultural Commodities During Storage
A. Insects
B. Fungus
C. Rodents
IV. Personal Belongings
A. Clothing - clothes moths
B. Other belongings
14
-------
TABLE 5. HUMAN HEALTH IMPACTS OF PESTICIDE DECISIONS
I. Pesticide Participants
A. Manufacturing worker
B. Formulator worker
C. Distributors - wholesale and retail (container breakage)
D. Applicators
1. Farmer and farm employee
2. Commercial applicators
a. Industrial
b. Home
c. Farm
d. Recreational
3. Government workers
a. Health control
b. Right-of-way maintenance
c. Large-scale insect control programs
E. Government inspectors
II. Pesticide Bystanders
A. Non-occupationally exposed
III. Disease Vector Control Benefits
IV. Accident Attenuation Benefits
A. Human injury and death element
15
-------
TABLE 6. NATURAL ENVIRONMENT AND AESTHETIC IMPACTS OF PESTICIDE DECISIONS
I. Environmental
A. Non-renewable resources
1. Terrestrial
a. Plant
b. Animal
2. Fresh water aquatic
a. Plant
b. Animal
3. Esturine/marine aquatic
a. Plant
b. Animal
II. Aesthetics
A. Sporting activities
1. Land based
a. Hunting
b. Camping and hiking
c. Picnicking
d. Bird watching
2. Water based
a. Direct human contact
1. Swimming
2. Water skiing
3. Other
b. Indirect human contact
1. Fishing
2. Boating
B. Tourism
C. Home and garden aesthetics
16
-------
difficult to quantify, it is possible to at least identify groups within the
society who are particularly concerned with aesthetic properties. Through
their reaction to a change in the surroundings caused by pesticide usage,
an estimate of the value placed on an aesthetic component by society as a
whole can often be made.
REGULATORY CONTROL COSTS
Completeness requires that the costs of regulating pesticides be
included in a benefit-cost analysis. This study is not a benefit-cost
analysis of the entire pesticide control program, so it is only necessary
to identify the control costs which are added if the pesticide is registered
for restricted use. This value is added as a subtotal after the large
matrix (Figure 1) is complete.
DISTRIBUTIONAL ASPECTS OF PESTICIDE-USE DECISIONS
Although this topic will not enter directly into a benefit-cost analysis,
its consideration by the policy maker is imperative so that the effects
of pesticide-use decisions on sectors as well as on the national economy
are recognized. Some of the possible distributional effects might be a
shift in agricultural production location causing unemployment in a geo-
graphic sector for a time, a change in the number or size of farms causing
disruption of traditional land holding patterns, or a change in farm
employment and employment trends (possibly a movement back to the farm, for
example).
17
-------
CHAPTER 3
ECONOMIC PRODUCTION*
One of the most important uses of chemical pesticides is to assist in
the production of economically valuable goods and services. Many chemical
compounds are used in the production of agricultural crops and livestock,
forestry products, and for the protection of buildings and stored materials.
By definition, a pesticide is a chemical compound used to protect these
various crops and materials from damage by a particular pest such as insects,
rodents, plants, and fungi.
Within this category of activities, the use of pesticides in the pro-
duction of agricultural crops and livestock and forestry products has been a
major consumer of chemical compounds. Many crops, such as cotton, require
large quantities of chemical pesticides in order to be profitable in many
areas of the country, whereas, for other crops, such as corn, the need for
chemical pesticides depends upon the area. Some regions with many different
pests that seriously affect corn production require large amounts of chemical
pesticides; other areas of the country may require very small amounts of
pesticide or perhaps no chemical compounds at all in order for corn pro-
duction to be profitable. Thus, the amount and type of pesticide used in
agricultural production and for forestry production varies greatly from one
region to another. This variation in the use of pesticide is documented in
the report by Andrilenas (2).**
Additional quantities of pesticide are used to protect materials that
have been stored or to protect structures from damage by a variety of pests.
These uses also vary from one region to another depending on the magnitude
and degree of damage caused by various oests present in the regions and types
of materials that are stored or the structures that need protection. For
example, certain areas of the country are much more seriously infested with
termites and require more chemical protection of structures from termite
damage than do other regions of the country. Generally speaking, the use of
chemical pesticides to protect structures and stored materials is a minor use
of chemical pesticides. The majority of uses, and therefore the major
*Participating researcher in this section is Principal Investigator
Dr. Donald J. Epp, Associate Professor of Agricultural Economics, The
Pennsylvania State University.
**Numbers in parentheses refer to items in the bibliography at the end of the
chapter.
18
-------
concern over pesticide use, involves production of agricultural and
forestry products.
THE NATURE OF AGRICULTURAL PRODUCTION IN THE UNITED STATES
The foregoing indicates that one of the major problems in any benefit-
cost analysis of a pesticide registration decision will be the determination
of the benefits to be obtained from the use of that pesticide. The majority
of cases requiring detailed benefit-cost analysis are likely to involve the
use of chemical pesticides in the production of agricultural and forestry
products. It is the purpose of this section to critically survey procedures
for evaluating the contribution of pesticides to such production and to
discuss the considerations involved in selecting a particular method
appropriate to the decision at hand.
GENERAL CHARACTERISTICS OF THE AGRICULTURAL PRODUCTION MODEL
The production of agricultural crops in the United States takes place,
for the largest part, in a complex and interconnected industry. Many of the
inputs used to produce one commodity could be used to produce different
crops. Most of the firms in this industry produce several to many different
products and alter the proportions as production costs and product prices
change. Due to a large public and private research effort, many aspects of
agricultural production experience technological change. Frequently these
changes are interrelated. Corn production provides a good example, for the
hybrid varieties not only produce greater yields with a given set of other
inputs, they also exhibit greater response to chemical fertilizers, show
greater need for pest protection via chemical pesticides and respond to
altered tillage and irrigation practices.
Because of the complex and interconnected nature of the agricultural
industry it is not possible to accurately predict all of the likely conse-
quences of altering production practices (such as pesticide use) using simple
analytical approaches. More sophisticated analytical methods are needed to
determine all of the important changes, some of which are not obvious. Models
have been developed of sectors of the agricultural industry which permit
refined analysis of changes within that sector.
To evaluate pesticide use in the production of agricultural crops, it is
necessary to utilize a sophisticated model of United States agriculture.
Such a model must be national in scope, recognizing that the markets for
agricultural products and individual production regions are not isolated from
one another. The model used must be regionalized to recognize the differences
in the physical relationships which are a function of the location of
agricultural production such as: soil type, crop yield, climate, pest
infestation, and other factors.
The model developed must permit the treatment of several crops as
endogenous to the model. The use of several crops simultaneously recognizes
that land use may be shifted from one crop to another depending upon the
19
-------
prices and the production cost. The model also needs to handle one or more
livestock activities. It would be desirable that the model have activities
relating to the production of grains and forage, although some portions of
that sector can be handled with exogenous variables.
Another desirable feature of a national agricultural production model
would be the ability to disaggregate by soil types to define various pro-
duction activities on different types of soil. The production activities in
such a model should be disaggregated by variable input so that labor,
fertilizer, pesticides, machinery, and other relevant variables are specified
and can be updated or changed as additional cost data becomes available.
The demand side of the model may be either developed by equation or
specified exogenously. For purposes of a benefit-cost analysis, an endogen-
ously determined demand for each of a set of demand regions would be best,
but in some circumstances an exogenously specified demand may be necessary.
Export demand should be explicitly considered and .may be either added to the
demand of a region from which it will be exported or assigned a regional
location at a specific port.
In most cases the mathematical programming technique will be some form
of linear programming model. This model should guarantee that all regional
demands are satisfied, that no regional resource restraints are exceeded, and
that the objective function is cost minimization, which seems most desirable
for this purpose. Such a model will show the land use and production pattern
emerging from the least cost solution which satisfied all constraints.
It is likely that a specific national model may need to be developed for
evaluating pesticide decisions. It may be possible, however, to use existing
models developed by the USDA or by other agencies, such as the models
developed at Iowa State University as a part of the National Science
Foundation—Research Applied to National Needs Program, or the agricultural
model being developed by the Office of Pesticide Programs of EPA.
PESTICIDE OUTPUT INFORMATION FROM AN AGRICULTURAL PRODUCTION MODEL
The model solution will indicate the effect of a pesticide decision on
the farm cost of producing not only the crop on which the pesticide is used,
but also the effect on costs of producing related agricultural products.
Assuming that markets for agricultural products are reasonably competitive,
these changes in costs will be reflected in changes in the cost of consumer
products. Thus, the production effect on consumer surplus of a pesticide can
be calculated and compared with other consumer surplus effects of the
pesticide decision. The actual conversion of farm price changes to consumer
price changes involves the use of formulas developed by the U.S. Department
of Agriculture. These formulas take into account the processing and trans-
portation costs associated with the conversion of farm products to consumer
products.
These changes in consumer prices, coupled with measures of consumer
price elasticities for each product, permit calculation of the new
20
-------
In concept, the alternative cost approach of the prohibition decision
for a particular pesticide in a particular use requires the following
procedures.
• Identify the production function for the product
using the pesticide under consideration, z=h(x^ . . . xn),
where z is the output and Xj (j=l, 2, . . . n) are
inputs, one of which is the pesticide under
consideration (xn).
• Letting ZQ represent the output of the production
process that would obtain under the status quo
with pesticide use and C be the cost of producing
that output under the status quo find:
The minimum c = Z p,- Xj
such that: h (x) > XQ
given xn = 0.
In this formulation, c gives the minimum cost of
production assuming xn cannot be used in this
production process.
• Then c - C is the alternative cost measure of the
production benefits of using xn in this production
process.
The alternative cost approach has been used in several studies analyzing
the effect of pesticide restrictions in crop production. Andrilenas (1)
summarized three studies conducted by the U.S. Department of Agriculture and
reported the costs to farmers of restrictions on three different pesticides or
pesticide groups commonly used in agricultural production. All of the studies
assumed that production was maintained at the level obtained prior to banning
or restricting the pesticide. The studies then calculated the costs of such
a ban on pesticide use. Typical of the results that can be obtained with this
approach are the conclusions reported by Andrilenas:
Banning the use of 43,000,000 pounds of phenoxy
herbicides, primarily 2, 4-D and 2, 4, 5-T on
62,000,000 acres of crops would increase farmers'
direct production costs about $290,000,000. This is
about 1.5 percent of the farm value of all crops and
6.6 percent of farm value of crops treated . . .
Besides these losses, farmers would use about
20,000,000 more hours of operator and family labor
(1, P. 53).
Another study implying the alternative cost method was reported by
Edwards (18). In his attempt to measure changes in external costs and
benefits related to reductions in the use of chlorinated hydrocarbons, he
assumed that the quality and yield of crops would be maintained by
22
-------
substituting organic phosphates for the restricted chlorinated hydrocarbons.
Langham (33) discussed possible applications of the Edwards study to a
national investigation of changes in external effects.
The Opportunity Cost Approach
A simplified partial analysis of the difference between the alternative
cost method and the opportunity cost method is presented in Figure 2. The
decision to ban a pesticide presently in use is illustrated by the shift in
the supply curve. Curve SQ represents the industry supply curve using the
pesticide while curve Sj represents supply under the higher cost next best
alternative method of production which does not use the pesticide. The
alternative cost method examines the increase in cost of obtaining the
initial equilibrium quantity, Qg. This increase in cost is represented by
the area ABP}PQ. The opportunity cost method, on the other hand, recognizes
that a new equilibrium price and quantity will emerge after the shift in
supply. Quantity will shift to Qi and price to ?2- Comparing the new
equilibrium with the old, the analyst using the opportunity cost method
observes that consumer surplus has been reduced by an amount equal to area
ACP2PQ in Figure 2. This reduction in well-being is clearly less than the
one calculated with the alternative cost method. Because the opportunity
cost method recognizes changes in production and consumption, it is the pre-
ferred method. The alternative cost method overstates the cost of restricting
pesticide use.
The above diagramatic analysis presents the essential features of the
budgeting appraoch to the opportunity cost method. This is discussed in more
detail below. Another approach, modeling, carries the above analysis further
to show changes in the markets for factors of production as well as other
markets. While this method cannot employ graphic analysis due to the
complexity of relationships, the mathematical modeling techniques described
and illustrated below permit a much more comprehensive analysis of likely
effects of a pesticide-use decision.
As was noted above the opportunity cost approach measures the change
resulting from the pesticide restriction as the change in consumer surplus in
the economy. There is abundant literature discussing the problems of
estimating consumer surplus; some authors even question the value of the
concept in many empirical contexts. For the purpose of this exercise, how-
ever, it seems appropriate that consumer surplus be the criterion used. The
ultimate consumer of the products and services produced with the pesticides
should provide the basis for evaluation. This becomes particularly crucial
where ramifications of the production shift may include a variety of products,
not just the product using the pesticide as a direct input. Consumer surplus
becomes the common denominator allowing comparison among a variety of
different production effects. In essence, the opportunity cost approach
calculates a shift in the supply curve for the product under consideration,
calculates a new equilibrium quantity and price, and determines a change in
consumer surplus by comparing the new equilibrium point with the old
equilibrium point. As a practical matter, the opportunity cost approach may
employ either of two methods: budgeting or modeling.
23
-------
Figure 2. Analysis of a shift in supply.
24
-------
Budgeting--
With budgeting, the analyst acknowledges that production shifts are
likely for those products which are produced using the pesticide input which
will now be restricted. Budgeting assumes that other commodities will be
produced in the same amounts and at the same prices as with the status quo.
Likewise, the budgeting approach has great difficulty in adjusting the factor
prices of other inputs into the production of the products being considered.
The budgeting approach is a more limited and restricted approach than the
modeling one, but for some circumstances may prove advantageous.
Model ing—
The second and more complete method of applying the opportunity cost
approach is through the use of production models. With this method,
mathematical models of the economy (more likely the relevant portions of the
total economy, such as the agricultural sector) are employed to trace the
shifts in supply curves and the changes in the amount and price of various
commodities and factors which result from the restriction on the use of a
particular pesticide in the production of a particular crop (e.g., 11, 39).
If these models are properly specified to include geographic areas and the
various alternative production activities which take place in each of these
areas, they are able to project changes in the location of production of
particular crops and changes in the use of various production factors in each
region of the country. This more realistically describes the likely reactions
to the change in pesticide use and permits the calculation of a more accurate
estimate of the change in consumer welfare resulting from the pesticide
decision.
A few studies have attempted to employ an opportunity cost approach to
examining pesticide restriction decisions. Casey and Lacewell (9) examined
some of the effects of pesticide withdrawals on cotton production. They
employed a limited regional model of cotton production to examine the likely
impact of restricting herbicides and insecticides. The model, however,
permitted only a limited number of substitute activities for cotton pro-
duction and was able to only approximate the likely impacts from this change.
Also, the authors did not follow through the changes in production costs to
estimate an impact on consumer surplus.
Additional work has been done in the U.S. Environmental Protection Agency
Office of Pesticide Programs (41). Much of this work has incorporated
national production models on a regional basis using sub-state regions as the
units of production and multi-state regions as the consuming areas. A variety
of crops are permitted as alternatives in the models developed, and
facilitate a more detailed examination of likely production shifts. Again,
the production cost changes have not been translated into consumer surplus
changes in this study.
COMPARISON OF METHODS
The alternative cost and opportunity cost methods are not equally
adapted to handling all problems—each has several advantages and disadvan-
tages. The alternative cost approach has the advantage of using data that is
more easily obtained and not requiring the extensive development of
25
-------
mathematical models prior to the analysis. Since the alternative cost
approach compares the status quo with the next best alternative, subject to
a pesticide restriction, the only additional data needed beyond present
production techniques is an estimate of how the same amount of product could
be produced under the best alternative available with pesticide restrictions.
This data can usually be obtained either from material supplied with the
application for registration or from a few experts in the production of the
crop under consideration. In addition, the alternative cost approach is
usually less expensive since production adjustments are not considered,
thereby eliminating the need for expensive models to incorporate changes in
the use of factors of production and adjustment in production enterprises.
The major disadvantage of the alternative cost approach is that it is
conceptually erroneous. Our knowledge of economic adjustments to changing
production technology recognizes that adjustments will be made in the enter-
prise combinations and in factor combinations for producing a particular
crop. If a significant factor of production is removed from use, the ensuing
production adjustments will not be limited only to factors of production but
may also include changing the crops produced or the location of crop pro-
duction. In other words, restricting the use of a particular factor of pro-
duction changes not only the production function but also the regional com-
parative advantage of various crops. The alternative cost approach ignores
all of this knowledge by assuming that the crop in question, using the next
best alternative, will be produced at the same level as before.
This conceptual error leads to a second error—the overstatement of the
costs associated with pesticide restriction. Because this approach ignores
the adjustments in the amount of agricultural product produced, it leads the
analyst to a cost of production figure that is greater than the one that
would actually result. As production costs increase it is likely that the
quantity produced will decrease. Thus, the alternative cost approach has the
analyst multiply a higher cost per unit by more units than would actually be
produced. This leads to an erroneous calculation of the change in consumer
surplus and the overstatement of the cost of making the pesticide restriction
decision.
The major advantage of the opportunity cost approach is conceptual: it
allows the analyst to see the adjustments that the economy is likely to make
in response to changing production costs from a specific pesticide-use
decision. The use of an econometric model of the agricultural sector,
particularly, facilitates the analysis of changing comparative advantage and
the resulting production pattern. This gives a more accurate indication of
the ultimate cost of making the pesticide-use decision.
The use of models also can facilitate distributional analysis if those
models have spatial or regional variables introduced into them. Through the
use of regional supply and demand models, the analyst is able to estimate not
only the market equilibrium supply and demand adjustments but also the
regional production adjustments in response to the overall market changes.
It is thus possible to note changes in the regional location of production
(an adjustment that is likely to take place with pesticide restrictions) and
to estimate the impact of these changes on various income and social groups
26
-------
that are distributed differently in the various regions of the country. Thus,
the models that are used for the opportunity cost approach can facilitate a
more detailed and sophisticated analysis of the effects of a pesticide-use
decision.
The major disadvantage of the opportunity cost approach is that it
requires a very large amount of data. Production costs, outputs, factor
requirements, factor prices and a variety of other bits of information are
required for each alternative crop activity in each region including alter-
native ways of producing the crop under consideration. Even if the supply
equations and activities are limited to those most likely to enter into the
solution, the requirements are formidable for most agricultural commodities.
The demand side also requires regional consideration with specification of the
demand for each of the commodities included in the supply side of the model.
Careful attention must be given to the inclusion of complements and substi-
tutes so that the model will give a reasonable approximation of actual market
adjustments.
A second disadvantage of the opportunity cost approach is that the models
developed for analyzing production and market shifts are usually short-run and
static. This means that these models must be revised periodically in order
to include new developments in factor prices, product prices and production
technology. Thus, the models are expensive to maintain. They are also
expensive to create in the first place. Most models require a great deal of
prior research on the technical relationships in production and specification
of the factors related to consumption. While much of this work has been done
for agricultural commodities, it is recognized that the material is frequently
inexact and often the models are out of date. It would be necessary, there-
fore, to undertake some rather expensive research in order to incorporate the
opportunity cost approach for a commodity that did not already have sub-
stantial prior work.
Obviously, the detailed knowledge of production relationships and there-
fore the expense is reduced if one uses the budgeting method rather than the
modelling method in the opportunity cost approach. Budgets usually involve
a partial analysis of the adjustments and therefore do not require the
development of production relationships for commodities not closely related
to the commodity under consideration. Even so, the budgeting approach
requires more information than the alternative cost approach because of the
consideration of production changes. It is unlikely that the budgeting
approach can be used for a regionalization analysis that involves more than a
few regions. Thus, the lesser cost is somewhat offset by less information.
CHOOSING BETWEEN ALTERNATIVE COST AND OPPORTUNITY COST METHODOLOGIES
Since the two approaches, alternative cost and opportunity cost, differ
in the amount of data that they require and in the cost of acquiring and pro-
cessing that data, it is necessary for an analyst to choose between these two
methods. Several points should be considered when making a selection of the
most appropriate analytical method:
27
-------
Demand elasticity--If the product for which the pesticide use decision is
being made has a very inelastic demand, it may be appropriate to use the
alternative cost approach. With a very inelastic demand, it is likely that
the assumption of producing the same quantity regardless of cost is a reason-
ably close approximation to reality. Thus, the disadvantages enumerated
above for the alternative cost approach are of much less significance for
many agricultural products, such as vegetables, potatoes, cereals for human
consumption and dry beans and peas.
Information availability—Since the opportunity cost approach requires a
great deal of information about production relationships, not only for the
commodity under consideration but for alternative commodities, it is helpful
to have these relationships previously developed. Much of this information is
so time consuming to develop that it would be virtually impossible to create
a research program that could give results in time to be useful for any
pesticide decision if a great deal of groundwork had not been laid prior to
that research effort. If no information is available concerning production
of t,he product in question or the alternative product that might be produced,
it is likely that the alternative cost approach would be adopted. This
approach requires less data and the material that it requires probably can be
produced in a time frame that is reasonable for a pesticide-use decision.
Complexity of interrelationships—If the product is produced in many
areas with a large variety of production alternatives, both with regard to
factors of production and to alternative crops, one needs to use an
opportunity cost approach. In such a situation the interrelationships are so
complex that it is difficult to judge from casual observation of the data the
likely combination that will result from a pesticide-use restriction. Under
these conditions it is very desirable to use the opportunity cost approach if
at all possible.
Availability of mathematical models—The stage of development of the
mathematical programming models for a particular product or sector of the
economy is an important consideration in choosing a method. If programming
models are fairly well developed for most of the important alternative
commodities as well as the commodity under consideration, it may be possible
to modify the existing work relatively inexpensively to obtain the information
needed on the pesticide-use decision. For example, a great deal of work has
been done with the feed grain-food sectors of American agriculture. Less
work has been done on the livestock sectors, although there are some models
that incorporate feed grains and livestock. Pesticide-use decisions in-
volving feed or food grains should give serious consideration to using an
opportunity cost approach and some of the models that have been developed
with appropriate modification. On the other hand, there has been very little
modelling that includes speciality crops such as fruit and vegetables into a
general agricultural model. It would be rather expensive and of dubious
value to develop a programming model for pesticide use decisions involving
those crops.
It is readily apparent that the key assumption of the alternative cost
approach—no change in the quantity of product produced after restricting
pesticide use—is not realistic in most cases and can lead to substantial
28
-------
error in estimating the social costs of a decision to restrict pesticide use.
It is the opinion of the authors that the alternative cost approach is
appropriate only in those cases where a substitute pesticide is readily avail-
able at virtually the same effective price as the pesticide under consider-
ation or where alternative production techniques are available at no increase
in per unit cost of production and which would involve not more than
negligible shifts in the location of production. These are very restrictive
conditions. For most pesticide-use decisions involving major agricultural
crops these conditions do not hold and the opportunity cost approach is
strongly advised.
ADDITIONAL INFORMATION NEEDS
Aside from the information concerning models and production relationships
described above, the analysis of pesticide-use decisions requires additional
specific information both on the crop production side and on the commodity
demand side. As should be apparent from the preceding discussion the
alternative cost approach needs only a limited amount of information compared
to the opportunity cost approach. This section outlines the data needs for
the opportunity cost approach, which includes all needs of the alternative
cost approach.
Crop production data are needed for each region that produces a par-
ticular crop. Data needed include the inputs used, such as labor, machinery,
gasoline, oil, fertilizer, chemical pesticides and other inputs, and the
productivity of each of those inputs. This can best be summarized by stating
the need for an empirically estimated production function for each crop,
including alternative ways of producing the same crop for each region where
this crop might be produced. In addition to the production function, the
supply estimation side of the problem will require data on factor prices for
each region. All of this information must be provided not only for the crop
under consideration in the pesticide-use decision but for all alternative
crops that might appear in any of the regions where the primary crop is pro-
duced.
The demand side of a model to be used in the opportunity cost approach
is less restrictive concerning the kinds of data that must be provided. It
would be desirable that a demand model be available and data be provided for
each variable entering into the demand such as disposable income, age, ethnic
composition of population in each of the consumption regions. If such a model
or such data are not available, it is possible to use the opportunity cost
approach employing an estimate of the quantity demanded for each commodity
involved in the production side of the analysis. Such an approach assumes
that changes resulting from production shifts will not significantly affect
the demand for various commodities. For most products under consideration,
this assumption is reasonably accurate.
29
-------
ILLUSTRATIONS FROM ALDRIN USE ON CORN: A CASE EXAMPLE
Some of the advantages and difficulties associated with the modeling
method of the opportunity cost approach that were discussed above can be
illustrated using the registration decision of aldrin for use on corn.
This is an example of a registration decision that has actually been con-
sidered by the EPA (41). In the course of its review, the EPA developed a
mathematical model of the agricultural sector to analyze some of the pro-
duction effects surrounding a decision not to permit chlorinated hydrocarbon
pesticides to be used on corn. The model developed by the EPA included many
of the desirable characteristics discussed above and illustrates the type of
information that can be obtained from such an analysis as well as illustrating
some of the problems with data and model specification.
The Model
The basic linear programming model contains activities which include
crop production disaggregated by soil type and region, commodity transpor-
tation, conversion of commodities into feed nutrients, and conversion of feed
grains .into corn equivalents. Resource constraints are defined for land by
land class. Demands include those for specific commodities, livestock
nutrients, feed grains for export, and specific commodity exports.
Seven crops are included in the activities of the model--barley, corn,
cotton, soybeans, oats, sorghum, and wheat. Other agricultural land uses are
projected exogenously and subtracted from the total land base. The livestock
sector is also projected exogenously and fixed at a predetermined level. The
nonfeed grain portion of total livestock nutrient demands as well as the non-
feeding demands for endogenous crops (such as barley malt for brewing or corn
for breakfast cereals) are projected using historical trends and population
estimates. These are added to regional demands which must be met through
local production or import from other regions.
The United States is partitioned into 129 producing areas, each of which
is an aggregation of counties wholly contained within one state except for
New England which is considered one producing region. The production areas
are defined so as to have similar cropping conditions throughout each area.
Because of certain features of the model, it is most meaningful to
present production results by consuming region. Therefore, production changes
in the producing regions are aggregated into the 27 consuming regions composed
of contiguous producing areas (Figure 3). In most cases, the consuming
regions are whole states or aggregates of whole states. State boundaries are
not observed, however, in those cases where commodity markets do not follow
them. Where a state is split, it is because each part of the state lies in a
separate market area.
Even though livestock feeding requirements are determined exogenously in
this model, they are specified in terms of total digestible nutrients and
crude protein. This allows the model to substitute various grains to satisfy
these requirements. Likewise, the export requirement is partly specified
exogenously for specific grains and partly specified in corn equivalents.
30
-------
Figure 3
Consumption regions
-------
This allows the model to substitute feed grains as would be done in practice
by livestock feeders in export markets.
The model is solved through a mathematical algorithm which simultaneously
guarantees that: (a) all regional demands are satisfied, (b) no regional
resource constraint is exceeded, and (c) the objective function is minimized.
In so doing, the model solution depicts the land use and production pattern
which satisfies all constraints at least cost.
The model was applied to the analysis of several alternative pesticide
situations to determine the production effects and the impact on farm income
by region of these alternative decisions. A first run of the model permitted
chlorinated hydrocarbon insecticides to be used on corn in their historic
amounts and areas. This was followed by a run which prohibited the use of any
chlorinated hydrocarbons on corn. A particular situation is very similar to
the one chosen for analysis in this illustration, that being prohibiting the
use of aldrin on corn with no other chlorinated hydrocarbons available as a
substitute. This second run did permit the use of organophosphate and
carbamate pesticides as a substitute for the chlorinated hydrocarbon
pesticides. Yields under the various pesticide assumptions were included in
the model that are based on previous studies in the U.S. Department of
Agriculture and the EPA relating to the effect of pesticide withdrawal and
probable insect infestation.
Recognizing that farmers do not immediately respond to changing price in
production conditions, the model included a flexibility constraint in the
form of a flexibility charge. The flexibility charges impose an additional
cost of production if production deviates beyond a specified percentage of
historic trend for a given productive region.
The basic features of the EPA model are further elaborated by Arnold (3).
One difference from the situation discussed by Arnold is that the present
analysis did not assume yield impacts of restricting pesticide use which
might occur with a moderate to heavy insect infestation. Rather, the analysis
used for this illustration assumed a "most likely" or "typical" insect
infestation with the corresponding adjustment in yield impacts of restricting
chlorinated hydrocarbon pesticide use.
Model Results
On a national basis, the model shows an aggregate increase in planted
corn acreage of about 88,000 acres. This results from the fact that corn is
a more desirable feedgrain than alternative feedgrains in many areas and that
farmers would increase their plantings slightly to offset minor yield
reductions due to the restriction on chlorinated hydrocarbon pesticide
(aldrin) use. Total changes in land used for each crop are reported in Table
7. Idle cropland (reported in the last column of Table 7) is reduced by
almost 33,000 acres. This means that in total, 33,000 acres more cropland
would be in production after a restriction on the use of chlorinated hydro-
carbons than would be used if those pesticides remained available.
32
-------
TABLE 7. CHANGE IN THE NUMBER OF ACRES IN SPECIFIED CROPS BY REGION
Barley Corn Cotton Soybeans Oats Wheat Idle
(Thousands of Acres)
Boston
New York
Baltimore
Atlanta
Jackspnville
0.53 -0.53
140.48 -140.48
Memphis
Houston
New Orleans
Louisville
Cincinnati
Detroit
Indianapolis
Chicago
St. Louis
Des Moines
Milwaukee
Duluth
Minneapolis
Omaha
Kansas City
Amarillo
Billings
Denver
Salt Lake City
Portland
San Francisco
Los Angeles
Total
(Percent)
0.86 -84.84 84.84 -0.86
-38.45 6.62 13.11 5.61
85.91 -8.27 -7.44 -54.48
-15.35 15.35
0.79 0.39
2.07 -2.07
-0.22 -0.59 0.82
2.94 88.29 0.57 -49.03 -8.80 1.74 -32.71
(0.03) (0.16) (0.01) (-0.09) (0.09) (0.00) (-n.21)
Grain sorghum: No change in any region.
33
-------
The regional distribution of land use can also be seen in Table 7 and
Figure 4. For example, the changes in acres planted to corn following the
pesticide restriction vary greatly between regions. The largest reduction in
corn acreage occurs in the Chicago region (84,840 acres) with lesser
reductions in the Duluth and Kansas City regions. Offsetting these
reductions in corn acreage are increases in the Atlanta region (140,480 acres)
and the Minneapolis region (85,910 acres). Similar shifts can be observed for
soybeans, barley, oats and wheat. Several instances can be found where land
is shifted from one crop entirely into another. The Atlanta region is shown
to shift 140,480 acres from soybean to corn production while the Chicago
region shifts 84,840 acres from corn to soybeans. In other cases, the model
shows where land is taken out of crop production. The Kansas City region
shifts 15,350 acres from corn to idle while the Duluth region idles part
(5,810 acres) of its 38,450 acre reduction in corn land. While the magnitude
of these shifts is small 1n this example, it does illustrate the richness of
detail that is possible from modeling approaches.
Changes in production in various regions may be even more important than
changes in acres planted. Regional production changes for each crop are
shown in Table 8 and Figure 5. While again, most of the changes are small
relative to total, production in the region and in total are a negligible
change in national output, some regions do show fairly large production
changes. The 11.9 million bushel reduction in corn production for the Chicago
region might have a noticeable impact if it were concentrated in one part of
the region. If distributed uniformly throughout the region, it is likely to
have little effect since it only amounts to about one percent of total corn
production. Likewise the 8.5 million bushel increase in corn production in
the Minneapolis region and the 10.3 million increase in the Atlanta region
only amount to 2.3 percent and 4.3 percent respectively of the corn pro-
duction in those regions. Users of corn (livestock feeders and mixed feed
manufacturers) likely will adjust to changes of this magnitude without major
shifts in their location.
In some regions the model shows changes in the type of land used for
crop production, even though the total acres devoted to a particular crop in
the region remain constant. The land resource in this model was classified
into three productivity groups, with group one soils the most productive and
group three soils the least productive cropland. The change in acres planted
to corn for each soil group in each region is shown in Table 9. In several
regions (Memphis, Cincinnati, Detroit, Des Moines and Denver) the increases
in corn acreage in one soil group are offset by decreases in corn acreage in
other soil groups. In other regions the change in one soil group reflects
the total regional change (for example, Atlanta or Chicago). The ability
of models to illustrate changes in the types of soil used to produce each
crop permit consideration of the location of production within the region and
consideration of environmental impacts associated with production on different
type soils. For example, the shift of nearly 200,000 acres of corn from
group one soils to group two and three soils in the Des Moines region may
result in greater erosion with accompanying sedimentation of lakes and
streams.
34
-------
TABLE 8. CHANGE IN PRODUCTION OF SPECIFIED CROPS BY REGION
Barley
(bu)
Corn
(bu)
Cotton
(bu)
Soybeans
(bu)
Oats
(bu)
Wheat
(bu)
(Thousands of Units)
Boston
New York
Baltimore
Atlanta
Jacksonville
Memphis
Houston
New Orleans
Louisville
Cincinnati
Detroit
Indianapolis
Chicago
St. Louis
Des Moines
Milwaukee
Duluth
Minneapolis
Omaha
Kansas City
Amarillo
Billings
Denver
Salt Lake City
Portland
San Francisco
Los Los Angeles
40.43
10,289.38
114.70
-66.64
220.96
26.90 -11,889.55
-1,281.46
-1,736.74
-53.76
-2,126.51
-0.10 8,533.24
-38.07
-1,805.99
0.44
40.28
-0.44
-19.63 -0.13
-2,936.19
-0.02
2,833.08 31.99 -80.53
126.48 415.92
-450.51 -221.75
-7.51
-49.36
-56.63
Total
(Percent)
67.08
(0.02)
199.99
(0.00)
0.00
(0.00)
23.35
(0.00)
-438.15
(-0.10)
Grain sorghum: No change in any region.
35
-------
A A A
A A A
A A
(-12.97)
// - 5,000 ACRES ADDITION
A = 5,000 ACRES REDUCTION
() • PERCENT CHANCE BASE ACREAGE
Figure 4
Change in the number of
acres of corn by region
-------
TABLE 9. CHANGE IN ACRES OF CORN IN EACH SOIL PRODUCTIVITY GROUP BY REGION
Boston
New York
Baltimore
Atlanta
Jacksonville
Memphis
Houston
New Orleans
Louisville
Cincinnati
Detroit
Indianapolis
Chicago
St. Louis
Oes Moines
Milwaukee
Duluth
Minneapolis
Omaha
Kansas City
Amarillo
Billings
Denver
Salt Lake City
Portland
San Francisco
Los Angeles
Total
Group 1 Group 2 Group 3
(Thousands of Acres)
0.53
140.48
-9.54 9.54
-131.72 80.80 50.92
-33.50 33.50
-84.84
-199.72 117.69 82.03
-3.28 -28.68 -6.49
-0.76 101.46 -14.78
-15.35
-0.89 0.89
-338.24 304.41 122.11
Total
0,53
140. AS
0.00
0.00
o.no
-84.84
0.00
-38.45
85.91
-15.35
o.oo
88.29
37
-------
CO
CO
A A A A
(-10.77)
// 9 t! ft II II
ti ii o ii a
9 (I II (I II
LESS THAN 250,000 BUSHEL ADDITION
LESS THAN 250,000 BUSHEL RKDUCTION
PERCENT CHANGE BASE PRODUCTION
500,000 BUSHEL ADDITION
500,000 BUSHEL REDUCTION
Figure 5
Change in corn production by region
100 200 300
-------
The regional model also permits the analyst to notice some aggregate
effects that might otherwise go unnoticed. Reductions in corn production
are concentrated in the Western Corn Belt states of Illinois, Iowa, Missouri,
Kansas and Nebraska and the Lake States of Michigan, Wisconsin and the north-
ern part of Minnesota. Most of the increases in corn production take place
in the Southeast and southern Minnesota with lesser amounts in the eastern
Corn Belt and Middle Atlantic states (Figures 4 and 5). If the production
shifts were of larger magnitude, these regional aggregations of changes
could have significant impacts on the location of feeding and processing.
The minor changes in total production of each crop shown by the rise of
this model would have no effect on the national price levels. Any regional
price changes would be so small that minor changes in interregional grain
transportation would offset them. Because there is virtually no change in
the farm level prices of grains, there would be no change in consumer level
prices that could be attributed to the pesticide decision and thus, no
change in consumer surplus in this example. There is, therefore, no economic
production effect, per se, to incorporate into a final benefit-cost calcu-
lation from this example. The procedure for making such a calculation from
farm level price changes is available and may be needed in other examples or
after adjustment of the model to take care of weaknesses discussed in the
next section.
Weakness of the Model
The model used to illustrate points in this section had weaknesses which
were acknowledged by the developers, but which should be eliminated or re-
duced in future applications. The three major concerns involve (a) the lack
of regional demand equations, (b) a need for further refinement in production
regions, and (c) a need to refine the "flexibility charges" to more
accurately reflect the production adjustments of farmers.
The model used for this illustration assumes that demand for agri-
cultural products is primarily a function of population with adjustments for
price and income. The regional demands, therefore, are fixed exogeneously
based on population projections. While the assumption and procedure are
basically correct and the errors in using them are inconsequential in the
example used, it is possible to have serious difficulty or substantial error
in other applications. Since total production of any agricultural crop was
not significantly changed in the "typical case" of a restriction on the use
of chlorinated hydrocarbon pesticides on corn, there were no significant
changes in product prices. Thus, the inability of the model to adjust
quantities demanded to reflect changes in price was not harmful. It is poss-
ible that a future pesticide decision will cause significant price changes.
The analysis in such a case will be in error if the model does not include a
demand function with price as a variable. In any case, the analyst does not
know ahead of time whether price changes will be significant or not.
Continued work on this aspect of the model is needed.
Due to the nature of specifying a single production equation for each
producing region, it was not logical to report production results by the 129
production regions, but rather, they were aggregated into the 27 consuming
39
-------
regions. This technique overcame problems associated with data availability
and gives the analyst a view of statewide changes for most of the important
corn growing region. As was noted above, however, some of the changes in
production that are insignificant when viewed relative to the entire region
(for example, the 11.9 million bushel reduction in corn production in the
Chicago region is only about one percent of total corn production in the
region) might be important if they are concentrated in one area of the region.
The 129 production regions of the model seem appropriate for most analysis.
Efforts should be made to develop a model that gives useful results at that
level of geographic disaggregation.
The third weakness of the model concerns the need to refine the
"flexibility charges" which were inserted to reflect reluctance of farmers to
make major changes in acres devoted to specific crops. This feature is unique
to the EPA model and is an appropriate technique to add realism to the model.
We agree with the developers of the model, however, that the specific set of
flexibility charges used in this example may have been too high. This
resulted in less production response to the pesticide decision than one would
have expected. While this weakness does not negate the usefulness of the
results from this application, it does indicate the need to refine the
technique. Either the charges need to be adjusted or the threshold levels of
production change at which the charges become effective need to be changed.
In spite of the weaknesses discussed, the model shows how an analyst can
examine likely adjustments in production and cropland use in a highly inter-
related agricultural system. When a pesticide decision results in significant
shifts in production, the modeling method of the opportunity cost approach
permits consideration of regional changes which can affect other businesses
in the community. This more comprehensive analysis increases the confidence
of the public in the adequacy of the review leading to the pesticide-use
decision.
SUMMARY AND COMMENT ON RECENT LITERATURE
Two major approaches to evaluating likely economic production shifts
have been discussed. One is the alternative cost approach which assumes that
there will be no change in the amount produced. The other is the opportunity
cost approach which measures changes in consumer surplus resulting from the
restriction using market models or budgeting approaches. The two methods
differ greatly in the conceptual acceptability and the data requirements. The
alternative cost is less satisfactory from a conceptual standpoint since it
assumes that the demand for a product is perfectly inelastic. It has the
advantage of requiring much less knowledge of production relationships and
production data. In general the alternative cost approach is less accurate
but less expensive, whereas the opportunity cost approach is more complete
and more accurate, but more expensive. For most applications to decisions
involving pesticides used on major agricultural crops, the opportunity cost
approach should be followed.
The literature dealing with the effects of pesticide use decisions on
agricultural crop production is quite large. This is not surprising since 59
40
-------
percent of all pesticides used in 1971 went into farm production (2) and a
decision to terminate registration of a pesticide could have a large at-the-
farm impact. In addition, a pesticide's position as an input into the
production function of a product which has a market-derived value makes
relatively easy the calculation of the costs of restricting the use of
pesticide. Various agricultural economists have addressed themselves to the
economic effects of a ban on farm uses of heptachlor (13), lindane and BHC
(14), chlordane (31), phenoxy herbicides (25), 2, 4, 5-T in rice production
(26) and aldrin in corn production (15). All of these reports found that the
direct impact on the farmer would be in the areas of higher production costs,
as the farmer switched to other methods of insect and weed control, and yield
losses where the alternative methods were less effective. Gerlow (26)
pointed out the costs of reduced quality due to weed seeds in rice.
There is some question as to how precise the above measurements of in-
creased production costs might be. Casey and Lacewell (9) used a survey
method called the "Delphi" technique to obtain estimates of the yield and
cost changes in the production of cotton due to pesticide restrictions.
Briefly, this technique consists of polling a number of experts as to their
opinion on the question, tabulation of their responses around the median
response, and return of the questionnaire to each expert for revision and
comment. Through an iterative procedure, the range of the responses is re-
duced and the final concensus is taken as the best measure available. The
obvious problem with the Delphi technique is that the final result is not
testable for accuracy. However, in the absence of better data with which to
estimate the production function and thus the marginal value of the pesticide,
the Delphi technique may be quite cost effective.
In addition to measuring the increase in production costs on the
existing output, a measure of the effect on consumer prices must be made.
This is because the correct method of calculating the benefit of allowing a
pesticide to continue in use is to sum the increased costs of production of
the existing output if the pesticide use is cancelled and the area of the
triangle of consumer surplus (which represents the excess of value over cost
of the decreased output demanded at the higher price). Davis et^ aj_. (12) in
addition to making estimates of the increase in production costs, predicts
that the effect of the consumer will be small if effective alternative methods
of control are available, because for most farm products pesticides are a
small part of the total input. Chapman (11) concurs with this view and adds
that the percentage increase in consumer costs would be less than the per-
centage increase in farm costs due to the large segment (about two-thirds) of
consumer cost which is accounted for by processing and marketing costs. He
also points out that there is a wide variation in the percentage of process-
ing and marketing costs according to the commodity involved; thus the con-
sumer surplus estimations must be made on a per-commodity basis.
Several studies have looked at the effect of increasing the acreage de-
voted to production of a commodity as a substitute for the use of pesticides.
Delvo (15) held the price/bushel of corn to the consumer constant by in-
creasing land use. Pimentel and Shoemaker (39) investigated the effect of a
complete elimination of chemical insecticides on cotton and corn production
and found that when combined with an elimination of government acreage
41
-------
controls the costs of production of these commodities rose 11 percent and 1
percent, respectively and required 2.1 million more acres. This cost increase
was considered a maximum because they did not include the benefit from pest
control to be derived from biological, cultural, or integrated methods. Their
results indicate that acreage controls which restrict the crop from being
grown intensively in regions suited to its production affect crop prices more
than do restrictions on pesticides. This conclusion is echoed in Dixon e_t a]_.
(16) where it is found that a reduction in pesticide usage can be obtained
by movement of crop production to areas where pests are less of a problem; in
this case the movement of cotton production from the pesticide intensive
Southeast to the Southwest.
42
-------
BIBLIOGRAPHY FOR ECONOMIC PRODUCTION CHAPTER
1. Andrilenas, P. A. Evaluating the Economic Consequences of Banning or
Restricting the Use of Pesticides in Crop Production. In (42),
1970. pp. 48-62.
2. Andrilenas, P. A. Farmers Use of Pesticides in 1971—Quantities.
AER-252, Econ. Res. Serv., U.S. Dept. of Agr., 1974.
3. Arnold, Fred T. Mathematical Modelling to Improve the Quality of
Economic Impact Assessment: Studies in Pesticide Policy Analysis.
Office of Pesticide Programs, U.S. Environmental Protection Agency,
Washington, D.C., 1975.
4. Bailey, G. W., R. R. Swank, and H. P. Nicholson. Predicting Pesticide
Run-off from Agricultural Land: A Conceptual Model. Journal of
Environmental Quality, 3(2): 95-102, 1974.
5. Brewer, M. F. Agrisystems and Ecoculture, or: Can Economics Inter-
nalize Agriculture's Environmental Externalities? American
Journal of Agricultural Economics, 53(5): 848-858, Dec. 1971.
6. Carlson, Gerald A. A Decision Theoretic Approach to Crop Disease
Prediction and Control. American Journal of Agricultural
Economics, 52(216), May 1970.
7. Carlson, Gerald A., and E. N. Castle. Economics of Pest Control.
Pest Control: Strategies for the Future. National Academy of
Science, Washington, D.C., 1972, pp. 79-99.
8. Casey, J. E. and R. D. Lacewell. Regional and Aggregate Economic
Impact of Withdrawing Alternative Pesticides from Cotton Pro-
duction. Tech. Article 10961, Agr. Exp. Sta., Texas A & M
University, College Station, Texas.
9. Casey, J. E. and R. D. Lacewell. Estimated Impact of Withdrawing
Specified Pesticides from Cotton Production. So. J. of Agr.
Econ., 5(1): 153-159, 1973.
10. Casey, J. E. and R. D. Lacewell. Pesticide I: A General Model to
Estimate Regional and Aggregate Effects of Pesticide Withdrawals.
Departmental Program and Documentation 73-3, Dept. of Agr. Econ.
and Rural Soc., Texas Agricultural Experiment Station, College
Station, Texas, September 1973.
11. Chapman, Duane. The Economic Consequences of Regulation or Pro-
hibition of Agricultural Chemicals in American Agriculture.
In Environment and Quality of Foods, P. L. White and Dianne
Robbis, eds. Futura Publishing Co., Mt. Kisco, N.Y., 1974.
43
-------
12. Davis, V. W., A. S. Fox, R. P. Jenkins, and P. A. Andrilenas. Economic
Consequences of Restricting the Use of Organochlorine Insecti-
cides on Cotton, Corn, Peanuts, and Tobacco. Agricultural Report
No. 178, USDA Economic Research Service, 1970.
13. Delvo, H. W., et al. Economic Impact of Discontinuing Farm Use
Heptachlor. ERS-509, Econ. Res. Serv., U.S. Dept. of Agr., 1974.
14. Delvo, H. W., et al. Economic Impact of Discontinuing Farm Use of
Lindane and BHC. USDA, ERS #524, 1973.
15. Delvo, H. W. Economic Impact of Discontinuing Aldrin Use in Corn
Production. ERD-557, Econ. Res. Serv., U.S. Dept. of Agr., 1974.
16. Dixon, D., P. Dixon and J. Miranowski. Insecticide Requirements in an
Efficient Agricultural Sector. Review of Economics and
Statistics, 55(4): 423-432, 1973.
17. Edwards, W. F. Economic Externalities in the Agricultural Use of
Pesticides and an Evaluation of Alternative Policies. Ph.D. Thesis,
University of Florida, 1969.
18. Edwards, W. F. Economic Externalities in the Farm Use of Pesticides
and an Evaluation of Alternative Policies. In (42), 1970. pp. 63-
70.
19. Eichers, T. P., et al. Quantities of Pesticides Used by Farmers in
1966. Agricultural Report No. 179, USDA, Economic Research
Service, April 1970.
?0. Farris, D. E., and J. M. Sprott. Economic and Policy Implications of
Pollution from Agricultural Chemicals. American Journal of
Agricultural Economics, 53(4): 66-72, Nov. 1971.
21. Fisher, L. A. The Economics of Pest Control in Canadian Apple
Production. Canadian Journal of Agricultural Economics, 18(3):
89-96, Nov. 1970.
22. Fox,. A. S., et al. Extent of Farm Pesticide Use on Crops in 1966.
Agricultural Economics Report No. 147, USDA, Economic Research
Service, October 1968.
23. Fox, A. S. Economic Impact of Restricting Herbicide Use. Journal of
Environmental Quality, 1(4): 349-352.
24. Fox, Ar S. Economic Consequences of Restricting or Banning the Use of
Pesticides. In (42), 1970. pp. 34-48.
25. Fox, A. S., et al. Restricting the Use of Phenoxy Herbicides—Cost to
Farmers. Agricultural Economics Report No. 194, USDA, Economic
Research Service, 1972.
44
-------
26. Gerlow, A. R. Economic Impact of Cancelling the Use of 2, ", 5-T in
Rice Production. USDA, ERS #510, 1973.
27. Giese, R. L., R. M. Peart and R. T. Huber. Pest Management. Science,
187: 1045-1051, 1975.
28. Hall, Darwin C., and Richard B. Norgaard. On the Timing and
Application of Pesticides. American Journal of Agricultural
Economics, 55(2): 198-201, May 1973.
29. Headley, J. C. Estimating the Productivity of Agricultural Pesticides.
American Journal of Agricultural Economics, 50(1): 13-23, 1968.
30. Headley, J. C., and Lewis, J. N. The Pesticide Problem: An Economic
Approach to Public Policy, Johns Hopkins Press, Baltimore, Md.,
1967.
31. Jenkins, R. P., et a!. Economic Impact of Discontinuing Farm Use of
Chlordane. AER-231, Econ. Res. Serv., U.S. Dept. of Agr., 1972.
32. Jones, Lawrence A. Insuring Crop and Livestock Losses Caused by
Restricted Use: An Appraisal. USDA, E.R.S. #512, January 1°73.
33. Langham, M. R. Potential for Applying the Dade County Pesticide Model
to a Wider Geographic Area. In (42J, 1970. pp. 71-79.
34. Langham, M. R., J. C. Headley and W. F. Edwards. Agricultural
Pesticides: Productivity and Externalities. In Environmental
Quality Analysis, A. V. Kneese and B. T. Bower, eds. The Johns
Hopkins Press, Baltimore, Md., 1972. pp. 181-212.
35. Metcalf, Robert L. Agricultural Chemicals in Relation to Environmental
Quality: Insecticides Today and Tomorrow. Journal of
Environmental Quality, 1(1): 10-14, Jan. - March, P72.
36. Mighell, R. L. Economic Research on Pesticides for Policy Decision-
making. Econ. Res. Serv., U.S. Dept. of Agr., Aoril 1971.
37. Nagadevara, U. S., E. 0. Heady and K. J. Nichol. Implications of
Application of Soil Conservancy and Environmental Regulations
in Iowa Within a National Framework. CARD Report 57, Center for
Agricultural and Rural Development, Iowa State University, Ames,
Iowa, 1975.
38. O'Connell, P. F., and H. E. Brown, Use of Production Functions to
Evaluate Multiple Use Treatments on Forested Matershed. Water
Resources Research, 8(5): 1188-1198, Oct. 1972.
39. Pimentel, David and Christine Shoemaker. An Economic and Land Use
Model for Reducing Insecticides on Cotton and Corn. Environmental
Entomology, 3(1): 10-20, 1974.
45
-------
flf). Smith, V. L. Economics of Production from Natural Resources. American
Economic Review, 58: 409-431, June 1963.
41. U.S. Environmental Protection Agency, Economic Analysis Branch, Criteria
and Evaluation Division, Office of Pesticide Programs, Detailed
Information on Linear Programming Analysis of Organochlorine
Suspension of Corn Use in the United States, September 1975.
42. U.S. Department of Agriculture, E.R.S. Economic Research on
Pesticides for Policy Decisionmaking. Proceedings of a Symposium,
April 27-29, 1970, Washington, D.C., 1971.
43. Van der Plank, J. E. Disease Resistance in Plants, Academic Press,
New York, 1968.
44. von Rumker, R. E. S. Raun and F. Moray. Substitutes for Aldrin,
Dieldn'n, Chlordane and Heptachlor for Insect Control on Corn and
Apples, (in preparation) Office of Pesticide Programs, U.S.
Environmental Protection Agency, 1975.
46
-------
CHAPTER 4
HUMAN HEALTH*
To begin a discussion of the methods of measuring the human health cost
of pesticides it must first be recognized that until recently there were no
studies which specifically linked a pesticide to any chronic debilitating
disease. The only documented human effect had been the accumulation of
certain long-lived pesticides in human tissue and body fluids (49). The re-
cent discovery of chronic health effects connected to the production of the
pesticide Kepone and the increased use of short-lived, though more toxic
organophosphate and carbamate pesticides, however, makes the development of a
technique to estimate the cost of early death or chronic illness important.
Also, in the event that the suspected link between pesticides and cancer is
established, the existence of a human "value of life" technique will make the
benefit-cost procedure more flexible.
REVIEW OF LITERATURE
The various measures that have been used to estimate the value of life
have been criticized by Mishan (38). Lave and Seskin (27) used the present
value of foregone earnings to estimate the cost of dying due to air pollution.
This technique only measures the individual's contribution to Gross National
Product and gives no weight to the preference of retired or unemployed
individuals. Fromm (17) measured an individual's value of life by the amount
of insurance taken out to insure against a particular risk of death. This
measure is incorrect since failure to insure implies placing a zero value on
life. The purchase of insurance does not change the probability of death;
it only alters the distribution of assets if death occurs.
The correct technique to measure the value of life is to estimate the
aggregate consumer surplus involved in the reduction in the mortality rate.
This maintains consistency with other measures of societal welfare. The
correct question to ask then is, "What is the compensation required to have
the person do without the decreased probability of dying?". One method for
answering this question is to observe risk compensation in the market. If
Principal researcher in this section is Mr. F. Roger Tellefsen, Center
for the Study of Environmental Policy, The Pennsylvania State University.
This section was prepared under the supervision of Dr. Donald J. Epp,
Principal Investigator.
47
-------
agricultural workers are paid extra for handling hazardous pesticides, this
could serve as a measure of the compensation necessary to assume the higher
risk. Of course, this measure would not be applicable to the general popul-
ation whose risk preference would be likely to be different.
Feldstein (16) points out the distributional problem with this willing-
ness to pay method. Equal weights are assigned to each person's willingness
to pay, which may not be true. Feldstein suggests a weighting technique be
employed by the decision makers after the benefit-cost study is done.
One human health effect to which it is difficult to impute a cost due
to the long time interval over which it acts is that of pesticide induced
mutations. While it is true that no human mutations can be traced to
pesticides, it is argued (by Epstein and Legator, (11)) that the human genetic
material is not so different from that of test animals in which pesticides
have caused mutations. Also, they argue that due to high living standards
and improved health care, the process of natural selection, which once kept
an approximate equilibrium between spontaneously occurring mutant genes and
non-viable older mutant genes, is no longer in effective operation. This
combined with the near eradication of infectious diseases leads them to the
conclusion that "in the future our medical problems will be increasingly of
genetic origin." Lederberg (30), using the discredited present value of
foregone earnings approach to calculate the cost to society of a one percent
increase from the natural mutation rate, comes up with a rough estimate of
$200 million/year.
The concept of option demand would seem to be relevant to the discussion
of the human health effects of pesticides. In the presence of certainty,
consumer surplus will measure the benefits to the consumer. However if there
is 1) uncertainty about the future demand for the commodity or service, 2)
an element of irreversibility in the decision to supply the commodity or
service in the future, and 3) no way for the resource owner to exclude those
who do not pay for the benefit, an additional element of benefit must be
added to the measure of consumer surplus. Cicchetti and Freeman (2) present
a theoretical explanation of option value and show that at low levels of
certainty with a large number of users, to not consider option value benefits
would "result in significant understatement of benefits." Their model is
shown by Figure 6.
THE HUMAN HEALTH MODEL
The human health model follows a logical progression with well defined
data needs and several decision points (see Figure 7). We begin with the
question of whether or not health problems can be traced to the chemical
pesticide being evaluated. The answer can come from either human epidemic-
logical studies or from laboratory animal testing. Care should be taken to
design the screening procedure to minimize the possibility of passing a
chemical as "safe" because of inappropriate choice of laboratory animal or
sample size.
48
-------
Probability of Survival 1
Source: Cicchetti and Freeman (2, p.538}-
Figure 6. Option demand.
49
-------
What is the
"tolerable"
dose level?
What will the
dose rate be
if registered?
tn
o
Does a
potential human
health problem
exist?
What is the number
of individuals in
each exposure
level?
What is the
willingness to pay
in each exposure
level?
Figure 7. Human health assessment model.
Sum the
willingness to pay
in each exposure
level
Sum the
willingness to pay
across exposure
levels
-------
If the screening procedure shows the chemical does cause chronic health
problems, the assessment of those problems as they affect humans may begin.
A decision rule must first be established which delineates a "tolerable"
level of risk. Some decision rules may be legislatively mandated—the
"Delaney Clause"* for example. In instances where no legislated levels exist,
administrative decisions as to what constitutes "virtual safety" will have to
be made.
With a tolerable risk level decided upon, the dose level which gives
that level of safety can be found. This "tolerable" dose level is compared
with the dose rate expected if the pesticide is registered. In the case of
a pesticide which has been previously registered, this expected dose rate
can be established fairly accurately by examining human monitoring studies
conducted by Office of Pesticide Programs of the Environmental Protection
Agency (EPA). Additional information can be obtained from the Total Diet
Studies conducted regularly by the Food and Drug Administration (FDA). The
FDA study samples "market baskets" of food representative of the dietary
habits at the sampling location. The nationwide scope of the project permits
an accurate estimate of the pesticide residue load through food contamination
to which the average American is exposed. However, since pesticide residue
can reach the human body from sources other than food (for example, pesticide
residues can be transported through the atmosphere as a particle and enter
the body through either respiration or absorption through the skin), the
Total Dietary Study figures can only be taken as a partial estimate of total
body intake. It should be made clear that the desired information is the
total body intake from all sources.
For those registration decisions which must be made on pesticides which
have not been previously used, estimation of the expected dose rate is much
more difficult. Knowing the chemical and biological attributes of the
pesticide, if an analogue pesticide exists which has previously been regis-
tered, it is possible to use that dose rate as a first estimate of the dose
rate which might be expected if the registrant pesticide were allowed into
use. Taking the application rate of the analogue pesticide and comparing
this with the dose rates shown by human survey gives a proportion which can
be used in this procedure. Estimates of the quantity of the registrant
pesticide (taken from the production model) can be used with the proportion
obtained from the analogue pesticide to estimate what the dose rate will be.
If the pesticide is registered under these data conditions, special care
should be taken during the ensuing five-year registration period to monitor
the actual dose levels in the human population. If the levels found to exist
during actual use exceed those levels estimated and used in the human health
evaluation procedure, a strong case may be made for a revaluation of the
pesticide's human health effects—especially if the initial benefit-cost
ratio was marginally in favor of registration.
*The Delaney Clause is part of the Food, Drug and Cosmetic Act of 1958 and
its implications are described in the section headed "'No Positive Result1
Method" (p. 55).
51
-------
Having established what the "tolerable" dose level is and what the
actual dose rate is, it is possible to decide if a potential human health
problem exists. If the expected dose rate is above the "tolerable" rate,
assessment of the degree of health damage must be performed.
It is clear that it is impossible to speak of a "single" dose exposure
level. Keeping the discussion on a conceptual plane, a production worker
who deals with the pure active ingredient is likely to be exposed to higher
levels of the pesticide than the population at large. Whether the pro-
duction worker's exposure level is also higher than the pest control oper-
ator's or the farm laborer's level is not so clear. However, in all
probability, the sub-population should be ranked according to exposure levels.
The ordering will depend on the characteristics of the pesticide, the methods
and conditions under which the pesticide is manufactured and formulated, and
the intensity and geographic breadth with which the pesticide is applied.
We have divided the total population into sub-populations (refer to
Table 5) which can logically be expected to have different opportunities for
pesticide exposure. It is unlikely that each member of a sub-population will
have the same exposure level, however, due to the non-homogeneity of job
risk within a sub-population. This makes an estimate of the number of people
in each exposure level very difficult to make since, theoretically at least,
there can be an infinite number of different exposure levels. As the number
of delineated exposure levels increases, the number of individuals in each
level decreases (the total population remaining the same). Under the present
state of monitoring for human intake of pesticides it is unfortunately imposs-
ible to estimate an individual's exposure level precisely, making moot the
problem of an infinite number of exposure levels.
The sub-populations should be broad enough to allow a meaningful number
of individuals to be included. The best estimates of dose exposure levels
from monitoring studies should be used if available and, when combined with
the number of individuals in each aggregated grouping, gives the number and
characteristics of individuals in each exposure level.
Quantification of the human health effects embodied in the various
exposure levels depends on the theoretically correct concept called "willing-
ness to pay." This method is described in the section headed "Willingness to
Pay" (p. 66), Briefly, it measures the amount an individual would be willing
to pay to achieve a small change in his or her mortality rate. As an ex ante
concept, it can be answered by the individual either directly or through the
amount he is willing to be compensated for undertaking higher risks.
Two individuals who face the same possible change in mortality rate may
not agree on the amount they would pay to attain the improved position. A
major difference between them might be different attitudes toward risk. It is
to control for this factor that the population has been broken down along
occupational lines. It is felt that those who engage in the direct production
and usage of pesticides are aware of the increased danger and therefore have
a different risk function than individuals who avoid occupational pesticide
exposure. It is then obvious that if subgroups and exposure levels corres-
pond, the willingness to pay for a given change in mortality rate may not be
52
-------
transferable among subgroups. Ideally then, a complete willingness to pay
schedule for each subgroup (risk preference level) should be established.
Once the willingness to pay schedule for a subgroup is established, the
dose exposure level is translated into a mortality rate. The average will-
ingness to pay within that subgroup is taken as the amount which would be
paid to move from the higher risk level to the "tolerable level of safety"
embodied in a lower dosage. Summing this amount across the number of
individuals in the subgroup results in the amount that that subgroup would
be willing to pay to reach the "tolerable level of safety." Performing the
above procedure for each subgroup and summing across all subgroups results
in an aggregate willingness to pay figure for the expected lower mortality
rate.
It should be pointed out that for those individuals who are exposed to
pesticides by their own choice (predominantly those involved in pesticide
production, formulation, and application), only that portion of their will-
ingness to pay which is not returned by their compensation (earnings,
employer paid health and life insurance, retirement benefits, etc.) should be
included in the benefit-cost procedure. To do otherwise would be to double
these items since they are all costs to the employer and enter into his cost
of production. At the present, it is assumed that these employees perceive
the riskiness of their occupations and that they receive compensation for
their perceived risk. Whether they are successful of course depends on their
market power. Lacking data as to the extent of their success, the assumption
that their willingness to pay has been completely returned to them can be
made.
This effectively reduces the population to be considered in the benefit-
cost analysis to those whose injury or residue intake is due to non-occu-
pational exposure. Their willingness to pay is not returned to them in their
earnings, and they have little choice as to their exposure level.
MEASURING PESTICIDE RELATED ILLNESS
Establishing a scientific basis for evaluating the hazards of chemical
pesticides on human health is a major problem. If a causal connection be-
tween a pesticide and human health can be demonstrated, the estimation of the
benefits or costs of that effect can proceed according to the method to be
described. Establishing a connection through actual observation of humans
exposed to pesticides is the job of epidemiologists while biostatisticians
attempt to establish a connection via animal testing.
Acute Injury
Acute injury due to pesticide poisoning does occur, causing disruption
in the life of the victim. Pain, nausea, fever, and death are some of the
results of pesticide poisoning. The fact that virtually all poisonings are
the result of accident or human error (the suicidal fringe of the population
which uses pesticides to end a life is disregarded here) does not lessen the
need to consider this aspect of pesticide use. To suppose it does invites an
53
-------
analogy to automobile injuries. Most injuries in automobile accidents are
the result of human error. Yet one does not consider these injuries of little
consequence. Acute pesticide injuries may be several orders of magnitude
fewer in number, but their individual effects can be just as great.
Assessing the risk of pesticide poisoning is possible by referring to
published studies. The EPA is currently conducting a National Hospital
Survey which will greatly improve data in this area. The difficulty, of
course, is in apportioning the risk among various groups. For example, the
data may indicate that the nationwide risk of pesticide poisoning is 7 in
100,000. However, the risk faced by the agricultural worker in vegetable
fields may be hundreds of times greater.
In order that the risks of acute pesticide poisoning may be estimated,
the method of exposure as well as the specific chemical involved must be
determined. Once these parameters have been established, it is possible to
identify groups which are at risk and to rank these groups according to their
probability of acute poisoning.
Chronic Injury
Health problems making their appearance in man only after a long period
pose different problems. First there is the problem of recognition. In test
animal studies, a subject can be sacrificed at regular intervals and an
autopsy performed in order to determine if a disease is developing. For
obvious reasons this course of action is neither feasible or desirable for
human studies. Furthermore, the short life span and rapid reproduction of
most species of test animals allows any effects to be observed over several
generations in a fraction of the time elapsed for a comparable human test.
Thus, any detection system for chronic effects in humans must await the
development of the disease to the acute stage. At this advanced stage it is
often too late to cure the victim of disease. Another problem is that the
information gained from one case is limited to the recognition that one
person has contracted a disease which has developed over some time period and
whose cause is uncertain. This brings up the problem of attribution.
It is often very difficult to attribute a chronic health effect in
humans to the correct cause. This is due to the lack of experimental
controls. In the case of an ill person, the problem could be ascribed to
genetic, occupational, environmental, dietetic or some other factor to which
the individual has been exposed. Until a body of evidence builds up linking
the illness to one or more of these factors, epidemiologists are unable to
say with -any degree of assurance what the causal connection might be.
Animal testing programs have the advantage of allowing controls to be
placed on the genetic, environmental, and dietetic factors, thereby holding
to a minimum the extraneous effects which so often confuse human investi-
gations. Typically they are performed at higher dose levels than are en-
countered by humans in the environment. This allows a chemical to be
screened for harmful effects more quickly by using the experimental direct
correlation between dose level and chronic effects. Thus, experimental
animals exposed to high but sublethal doses of a suspected oncogen have a
54
-------
greater probability of contracting a disease above the spontaneous rate with-
in the time frame of the experiment and establishing the level of danger of
the chemical. The high dose levels are also necessary in order to keep the
screening procedure within financial bounds and still maintain statistical
precision. This follows from the well known proposition that information is
not costless.
In the case of mammalian studies the cost of increasing the number of
experimental subjects climbs rapidly due to the expense of housing, feeding,
supervising, and final analysis of larger numbers of animals. Furthermore,
experiments large enough to detect high incidences of harmful effects at low
dose rates have an increased chance that faulty laboratory technique may be
introduced merely by virtue of the difficulty in managing an experiment of
such size. These factors mitigate against the use of huge numbers of test
animals for screening purposes, and, when combined with the earlier
mentioned association between dose and effect, imply that the most infor-
mation for resources expended is obtained in high dose exposure testing.
TECHNIQUES FOR EVALUATING RISK LEVELS
Given this experimental state of affairs, what should be the technique
for establishing human risk? It is obvious that dose levels which induce
mutations, cancer, tissue growths or any other harmful effect in test animals
cannot be allowed in the human experience. But what of the lower dose levels
to which humans may be exposed which have not been directly tested on
laboratory animals? What methods exist to evaluate the human risk at these
low doses? The next section considers the possible approaches.
"No Positive Result" Method
One approach is to remove from the human environment any chemical which
has been shown to cause major harmful effects in humans or animals. This is
the approach taken against cancer causing food additives by the so-called
"Delaney Anti-Cancer Clause" of the Food, Drug and Cosmetic Act of 1958. It
prohibits the addition of any additive which has been shown to induce cancer
in man or animal. It further prohibits the addition of any cancer inducing
additive to food-producing animals unless there are zero residues of the
additive present when the animal is consumed as human food.
Such a draconian decision rule would be satisfactory if one could
conduct the laboratory screening with great statistical precision. With
limitations on experimental size, however, the possibility is great that an
actual carcinogen could pass this test due to small sample size.
For example, a test result of zero tumors in 100 animals gives a tumor
rate of 0 percent. Such a test result would fulfill the requirements of the
Delaney clause. However, statistically it can be shown that such a result is
consistent with a true risk of 4.5% when an assurance level of 99% is
employed.
55
-------
(1-p)100 = .01
100 log (1-p) = log .01 = -2
log (1-p) = -.02 = 9.98-10
1-p = .955
p = .045
Increasing the test to 0 tumors in 1,000 animals would only lower the upper
limit on tumor risk to .45% (36, p. 460).
In an effort to make up for this deficiency, a fraction of the highest
"no-effect dose" is taken as safe for human consumption. This fraction,
usually 1/100, has no basis in biology and is merely an arbitrary safety
factor. Furthermore, since the "no-effect dose" is an independent variable
in the experiment, it opens the possibility that unduly high levels of
residues may be allowed because of uninformative experimentation, i.e., the
use of low dosages or small sample sizes in the screening experiment. On the
subject of "no-effect dose" Weisburger and Weisburger have written:
"It seems to us a 'no-effect dose1 for a carcinogen is a highly
relative level which applies only for the precise experimental
conditions generated. While similar considerations hold for
drugs, the risk is not nearly so intense. More often than not,
an improper dose rate for rapidly acting drugs is detected al-
most immediately and appropriate remedial action can be taken.
With chemical carcinogens and their long latent period, the
disease condition resulting from inappropriate selection of
dose levels and alteration of environmental conditions leading
to potentiation may become visible only years after the exposure.
At that time remedial action is obsolete and often worthless."*
Our conclusion is that the "no positive result" method with its
arbitrary "safety factor" is inappropriate. Its use leads to a situation
where the pesticide is banned before any consideration can be made of its
true risk. Indeed, risk is never really considered under this method. It
is assumed to be reduced virtually to zero by the use of a safety factor.
Finally, as previously described, this method tends to reward uninformative
experiments by establishing potentially high permitted dose levels.
Mathematical Extrapolation
By mathematical extrapolation we mean any statistical investigation of
dose-response relationships beyond the range of relationships which are
experimentally observable.
There has been and there continues to be a great deal of discussion among
toxicologists and biostatisticians on the subject of mathematical extra-
polation models. It appears clearly ill-advised to join the fray over which
*Weisburger and Weisburger, cited in (5, p. 161).
56
-------
model is the "best." Each model has its weaknesses as well as its strong
points. The intention instead is to briefly describe the various models and
mention the attributes which lend themselves to the problem at hand.
Finally, a determination is made of a recommended model for this project
based on currently available information. Since the field is changing very
quickly with refinements and new techniques appearing annually, different
models may be used in future studies.
The three extrapolation procedures examined differ mainly in their
prediction of the shape of the lower end of the dose-response curve. The
probit model postulates that the individual members of the population have
differing levels of susceptability to a carcinogen and that this variability
can be statistically approximated by the normal distribution (the so-called
"bell-shaped curve"). It further postualtes that since the dose-response
curves at low doses are likely to be concave upward to the observed region,
placing a straight line extrapolation with a low slope from the observed
region sets a conservative upper limit to the risk. The logistic model
postulates that response is tied to the number of available chemical bonding
sites. As the sites become occupied, the response curve levels off, taking
on a shape best described by a log curve. The one-hit model postulates that
a response is induced by a single exposure to a cell by some disruptive agent
such as a chemical. This implies a dose reduction proportional to the risk,
resulting in a linear extrapolation curve from the observed region extending
to zero dose. For example, the dose which resulted in one tumor in 100
animals would be reduced by a factor of 1,000,000 to obtain an extrapolated
upper limit to risk of 1/100,000,000.
The degree of conservatism between these models varies widely. To
illustrate this fact a table from the FDA Advisory Committee on Protocols for
Safety Evaluation (14, p. 433) is cited (in Table 10). It can be seen that
at high levels of estimated risk the models do not differ dramatically in
their permitted dose levels. However, as the estimated risk falls, the
permitted dose levels between models become quite different.
In the absence of experimental results in the low dosage range to
demonstrate the true dose-response curve and thus the true dose-risk relation-
ship, a case can be made for the choice of the more conservative extrapol-
ation models—the one-hit model or the logistic model. If more than one
pesticide contributes to the total risk, in that their actions were similar
to each other, then use of the one-hit rule will minimize the possibility
that their combined "safe" doses will become unsafe. Further, if synergistic
effects between these pesticides possibly exist, the one-hit rule would give
greater protection than a less conservative rule merely by virtue of setting
lower individual safe doses. The true risk of synergistic effects would
have to be investigated via multiple-agent experiments (32, p. 1382).
If biological considerations such as those mentioned above do not exist,
the unnecessary use of very conservative models could lead to regulatory
problems. If the "safe" dosages are almost always so low as to virtually
prohibit any usage at all, the regulatory agency involved, in this case the
EPA, would be under pressure by registrants to devise an alternative pro-
cedure which would be more amenable to the "real world situation." What is
57
-------
desired is an extrapolation method whose conservatism has basis in biology.
This attitude makes defense of any decision based on extrapolation tenable.
TABLE 10. EXTRAPOLATED VALUES OF "SAFE" DOSES FOR THREE DIFFERENT
DOSE-RESPONSE CURVES DESCRIBING OBSERVED RESPONSES IN
2% TO 50% RESPONSE RANGE EQUALLY WELL
Estimated risk
1/100
1/1000
1/1,000,000
1/100,000,000
P rob it
4.0 X 10"2
1.5 X 10"2
1.4 X 10"3
4.1 X 10"4
Logistic
2.2 X 10"2
3.1 X 10"3
9.8 X 10"6
1.6 X 10"7
One-hit
1.4 X 10"2
1.4 X 10~3
1.4 X 10"5
1.4 X 10"8
SOURCE: F.D.A. Advisory Committee on Protocols for Safety Evaluation
(14, p. 433).
The probit model, often called the "Mantel-Bryan model" after the
authors of the article (36) which first discussed its use in extrapolation,
has several features which indicate its suitability for use in this project.
First, it is a linear extrapolation method which weighs dose against
probability making it relatively simple to associate a specific dose with a
risk level. To illustrate: Figure 8 shows the estimation of the "safe"
dose of a potent carcinogen methylcholanthrene. The solid line is fitted to
the data and for the region of observation has a slope of 2.64 normal
deviates per log dose. The extrapolation is performed by extending a line
with conservative slope of one normal deviate per log dose from the lowest
upper confidence limit, in this case that associated with the second dose
level. Although the procedure is intended to set maximum permitted dose
levels given a prespecified risk level (in the illustration, a risk of
1/100,000,000), there is no reason why the procedure cannot be reversed and
used to estimate the risk associated with a prespecified dose level. In this
manner the upper limit to risk of exposure to 'x' parts per million (ppm) of
pesticide could be estimated from laboratory experiments at much higher
dosages.
It should be noted that there are two variables in the Mantel-Bryan
method which can change the dose-risk relationship. The most obvious is the
slope of the extrapolation line. Mantel-Bryan recommend the slope of one
probit per log dose as being conservative from the standpoint that dose-
response curves in the observable region are virtually always of steeper
slope. A conservative extrapolation line is necessary to guard against the
possibility of setting permissable dose levels which are unduly high. Other
researchers have argued for steeper extrapolation slopes. The effect of
using different probit slopes upon the permitted dose with a given estimate
risk of 1/100,000,000 is shown in Table 11. It can be seen that the steeper
58
-------
o>
00
a
o
n>
i
x
c+
"O
O
c
en
10
CU
a>
i
-s
o
o
ft)
Qu
C
ro
o
a>
0)
o
cr
O)
(99.87%)
2
(97.7%)
1
(84%)
0
(50%)
-1
(16%)
-2
(2.3%)
j
-3
- (0.13%)
-4
(0.0032%)
-5
(0.000029%)
I I Observed outcomes
* Maximum p values, 99% confidence
conservative
extrapolation
with slope of
1 normal deviate
per log
maximum likelihood line,
slope = 2.64 normal
deviates per log
l I
-26
2-18 2-14 2-12
-6 -
2° 22
DOSE
-------
slopes allow much higher dose levels—consequently, for a given dose exposure,
the use of higher slopes will result in a lower estimate of risk being
attached to the dose. A conservative bias on the subject of potential gross
health effects should prevail, and in the absence of contrary evidence the
unitary probit slope is used to make risk estimates.
TABLE 11. FRACTION OF EXPERIMENTAL DOSE USING PROBIT
EXTRAPOLATION WITH DIFFERENT SLOPES FOR AN
ESTIMATED RISK OF 1/100,000,000
Observed tumors
Slope = 1
Slope = 1.5
Slope = 2
0/50
0/100
0/500
0/1000
1/18,000
1/8,300
1/1,800
1/1000
1/690
1/410
1/150
1/100
1/135
1/91
1/42
1/32
SOURCE: Cranmer (5, p. 164).
The other variable which can effect the dose-risk relationship is the
choice of the upper confidence limit. Willingness on the part of the re-
searcher to accept as true a false hypothesis is reflected in the choice of
confidence limit. In the case under consideration this translates into a
willingness to accept a no-response result as true when that result could
have been caused by statistical variability. The choice of a low confidence
limit results in a wide confidence interval. Under the Mantel-Bryan model
such a choice leads to lower risks being attached to given dose levels than
would be the case for a high confidence limit. This feature is illustrated
in Figure 9. Mantel-Bryan suggest a 99 percent confidence limit as appro-
priate for large populations.
A further attribute of Mantel-Bryan is that the method "rewards" good
experimentation. Registrants who test their chemicals with large numbers of
animals and/or at high dose rates are "rewarded" with high "safe" levels.
The increased information which results allows the confidence intervals to
become smaller, thereby moving the extrapolation line to the right. Con-
versely, poor experimentation results in wide confidence intervals and the
establishment of low "safe" levels. This is helpful from a registration
standpoint since it makes the registrant responsible for setting his own
permissable residue levels. From the point of view of the benefit-cost
procedure it allows the estimation of risk levels to be made with greater
confidence.
60
-------
to
CI
n>
-*>
o
-h
O
3
oa
o
o
O-
fB
O
a.
(99.87%)
2
(97.7%)
(84%)
0
(50%)
-1
(16%)
-2
(2.3%)
-3
(0.13%)
-4
(0.0032%)
-5
(0.000029%)
o
-s
o
n>
o
cr
cu
cr
* Maximum p values, 99% confidence
Maximum p values, 95% confidence
slope of 1 /
confidence S
level of ,'/ slope of 1
.95 ,''/ confidence
DOSE
-------
Since their 1961 article was published (36), work has been done on the
Mantel-Bryan procedure to correct several statistical deficiencies. A recent
report (33) details the improvements made. One of the most interesting is the
ability to handle several sets of independent data. This allows the risk
level associated with a dose to be updated as new experiments are performed.
Information from registrants, independent laboratories, and government
laboratories can be combined—the result being improved confidence in the
extrapolation.
A caveat must be made at this point. Strictly speaking, what is being
done by any of the mentioned extrapolatory models is the setting of a "safe"
dose for the animal upon which the experiment is performed. For example, if
it is a mouse experiment, then we have set the safe dose to protect mice from
a cancer-causing chemical; the problem of translating animal studies into
human terms is still present. Unfortunately, mice are not men, and studies
which compare the respective metabolic processes of laboratory animals and
humans are needed. About all that can be done at this time is to maintain
comparability between pathways of exposure in the laboratory animal and man.
Various conversion factors have been suggested depending on the way dosages
are stated.* The existence of this problem accentuates the need for a con-
servative extrapolation method in order to protect what must, in the absence
of contrary evidence, be assumed to be the most vulnerable species—Homo
sapiens.
TECHNIQUES FOR EVALUATING A CHANGE IN MORTALITY RATE
The dose-risk extrapolation for chronic effects outlined in the pre-
ceding section can be used by the decision maker in a number of ways. Most
simply, it can be translated into mortality and morbidity figures. Taking
the extrapolated risk and multiplying by the size of the population exposed
will give an upper-limit estimate to the number of acute or chronic illnesses
which will be expected to result from the pesticide under review.
Constructing an example for illustration, suppose one million farmers
were exposed to a pesticide residue of .01 mg and that extrapolation from
animal testing associated a risk of 7.5/100,000 of contracting cancer with
such a dosage. It then follows that as many as 75 farmers can be expected
to contract cancer as a result of their pesticide exposure. It must be
emphatically pointed out that 75 is the upper limit to the number of cases
which are expected in this example; that is, since the extrapolation
procedure aims toward conservative estimates, the true risk (which cannot
be known) may be much less than 7.5/100,000.
*The surface area rule, for example, relates surface area of an organism to
its weight; the surface area is proportionate to 2/3's root of its weight.
"If mouse dose is stated in flat amounts, obtain dose for man by multiplying
the mouse level by 2/3's root of the weight ratio, i.e., 200." Mantel and
Schneiderman, (32, p. 1385f).
62
-------
If desired, the estimation of the human health effects of a pesticide
can be left in these terms. Even if further refinement is desired, this first
step serves as a valuable screening stage. Those who are at risk are
identified and an estimate is made of the number of potential victims. If the
number is small (the determination of "small" to be arrived at by scientific
and political concensus), the examination of human health effects can be
terminated here. If the number is large, the remainder of the assessment
procedure should be completed.
A benefit-cost procedure is most complete when as many of the relevant
entities as possible are reduced to a common denominator. This is a
difficult objective in areas which are not normally exposed to the equili-
brating forces of the market place—one such area being the desired level of
human health. It is felt that techniques exist which allow the human health
effects to be made commensurable with the output of the production effects
section (Chapter 3).
Attempts at systematic evaluation of the value of human life have been
made since at least 1931 when an effort was made to measure the costs of
World War I to the American people (3). In 1946, Louis Dublin and Alfred
Lotka published a work which was designed to find the amount of insurance
which an individual should carry (8). Problems arose when economists began
to use variations of Dublin and Lotka's technique to estimate the costs of
disease and thus the cost of death due to the disease. Most of these en-
deavors can be discussed under the heading of "foregone earnings."
Foregone Earnings Approach
Generally, the foregone earnings approach posits that the discounted
value of the future stream of earnings is the appropriate estimate for the
cost of an untimely death; and, since the analysis is symmetrical, the
appropriate estimate for the value of a life saved from an untimely death.
An advocate of this technique stated that "if an individual had not died or
become ill this year, he would have continued to be productive for a number
of years. It is the present value of these future losses that is required
as an appropriate evaluation measure of the costs of a disease for program
planning purposes" (44, p. 438). Immediately, the question arises over what
is the correct measure of earnings to be used in the calculation—gross or
net.
Those who support the use of gross earnings cite the fact that when an
individual ceases to receive earnings due to death, production is lost to
society. If what is of interest to society is lost production, and the
individual decedent's marginal productivity is measured by his or her earnings
stream, the loss to society can be expressed as
63
-------
where Yft is the expected gross earnings of the person during the t th year,?*
1s the probability in the current (r th) year of the person being alive v
during the t th year (41, p. 688). Among those who have used the gross esti-
mate of earnings are: Rashi Fein, who estimated the loss to society caused by
admissions to mental institutions (12), Herbert Klarman who estimated the loss
to society due to syphilis (24), and Dorothy Rice who estimated the loss to
society due to illness of all kinds (45).
Those who prefer the use of net earnings disagree with the above approach
by arguing that although society loses the production of the decedent, it also
does not have to supply goods for the decedent's consumption. Death releases
resources to the rest of society for their use. Thus, the appropriate loss
to society under this method is
L = 2 P ( Y - C Xl-d)-^ (2)
where Ct is the consumption of the individual during the t th period that is
expected at time r (41, p. 688). It is easy to follow the logic of this
argument. However, closer examination of the equation reveals that it is
possible that the cost to society of a death could be negative (i.e. a
"benefit") if the individual consumes more than he produces. Retired and
unemployed individuals are likely to have such negative life valuations.
Weisbrod points out that the problem here is one of defining the
"society" which is to be included in the evaluation. If the definition of
"society" excludes the individual whose value of life is being estimated,
then the net earnings correctly estimates the contribution he makes to society.
If "society11" includes the individual, then the gross earnings correctly
estimates the contribution he makes to his society. Including the individual
whose value of life is being measured in "society" removes the possibility of
negative valuations but does admit the possibility of a zero valuation (53,
pp. 35-36).
The possibility that the procedure may result in a negative or a zero
valuation of life has always been a major criticism of the foregone earnings
approach. Individuals whose marginal product is not measured in monetary
terms (the largest single example is output from housewives being excluded
from Gross National Product measurement), who are temporarily non-productive
(those who are unemployed), who are retired from the work force are all
victims of a methodological problem which equates their value to society to
zero. In the case of housewives, imputation of value to their services at
the average earnings of a domestic worker has been carried out (45, p. 429).
This imputation results in a lower valuation to causes of illness and death
which predominantly affect women, due of course to the low wage of domestic
workers relative to the average wage of male workers.
It becomes clear from this brief discussion of the foregone earnings
approach that the entity missing from the analysis is the individual at risk.
64
-------
The procedure dwells on only one portion of an individual's contribution
to society (and typically that contribution is measured against the standard
of Net National Product). If the goal of society is the maximization of NNP,
then assessing the costs or benefits of a health impacting project according
to the foregone earnings approach can be rationalized. If society holds other
goals as well, then the use of foregone earnings is merely an expedient
approximation.
Other Approaches
Another method of estimating the value of life may be thought of as the
"insurance premium" approach. In this method the premium paid for insurance
to cover a probability of death from a particular cause is translated into
the value placed on life. Gary Fromm (17) applied this concept to the case
of a traveler who buys flight insurance. If the risk (r) and the correspond-
ing premium (p) are known, then the value of his life is set at P. . A one-
tenth of a percent risk with a premium of $10 implies a value ofrlife of
$10,000. It has been pointed out that the flaw in this procedure is that the
payment of an insurance premium does not alter the probability of death, it
only alters the distribution of assets within society if death occurs. Thus,
the purchase of insurance can be interpreted only as an indication of the
purchaser's concern for his beneficiaries and not as an indication of an
individual's value of his own life (41).
Another concept which is often discussed when the value of life is con-
sidered is that of quantifying intangibles. Pain, suffering, anxiety and
grief are some of the psychic costs which are widely believed to be part of
disease and death and should somehow be included when measuring the value of
life. Rice takes a common position when she says:
Intangible or psychic costs of disease, such as pain and
grief, are omitted [from her study]. These costs do not
directly involve a loss of output and are not readily
measurable. Several economists feel that ignoring the
intangibles may distort the over-all economic and social
costs because the implicit assumption is that the
economic value of intangible losses is zero.
It is of course not true that the economic value of intangible losses is zero.
These negative factors are an important part of the reason why people have
such a high desire to avoid illness and death. The reactions of those indi-
viduals who are completely covered by health and income insurance indicate
that there is more involved to these individuals than loss of income or high
health care costs when death or illness threaten. These personal feelings
of the value of avoiding intangible costs should somehow be included.
Inadequacy of Evaluation Techniques
The most telling argument against the previously discussed evaluation
techniques is that they are conceptually incorrect for use in a benefit-cost
analysis. For whatever reason, they have long been giving answers to the
wrong question. It is for this reason that this study is aimed at correcting
65
-------
conceptual inconsistencies and at asking the correct question.
Dan Usher (50, p. 195) discusses three related but distinct questions
which can be used to point out the erroneous use of past measurement tech-
niques. Question one, the insurance question, asks "What is my life worth to
my wife and children?" Question two, the birth control question^ asks "How
much better or worse off would society be if I ceased to exist or had never
been born?" Question three, the valuation-of-life question, asks "How much
would I pay to avoid a small probability of my death?"
The present value of foregone earnings approach answers the insurance
question. The amount of money which satisfied L-j is the amount which will
keep his family in the same financial condition as it would have been had he
not died. It is the appropriate amount of insurance for him to carry. It
says nothing about what his answer to the valuation-of-life question might be
since it does not address itself to that concept. The fact that a man has
insured himself for the full amount L-] does not imply that he is now in-
different to death. The same criticism applies to the "insurance premium"
approach.
The discounted net earnings approach answers the birth control question.
Since l_2 excludes the individual's preferences, as previously discussed, it
is quite possible that society would be better off if the individual were to
die. Once again, the procedure does not address itself to the valuation-of-
life question.
The question that should be asked is the valuation-of-life question--
How much will someone pay to avoid a small probability of death? This con-
cept is the theoretically correct one to use in a benefit-cost study since it
maintains consistency with other measures of societal welfare; in the case of
measuring the value of life, it estimates the aggregate consumer surplus in-
volved in a reduction in the mortality rates. The next section develops this
concept in detail.
Willingness to Pay
Discussion of what has come to be known as the "willingness to pay"
concept is burgeoning in the recent economic literature. Perhaps the best
known advocate is the British economist E. J. Mishan who has written widely
on welfare economics (42) and benefit-cost analysis (38). His 1971 article
forms much of the theoretical background upon which the breakaway from
"foregone earnings" is founded. One of the earlier statements of the willing-
ness to pay concept was made by Thomas Schelling (46). An attempt to apply
willingness to pay for auto safety features was made by Lester Lave and
Warren Weber (28). They found that auto buyers were, for the most part,
rational in refusing to buy certain safety devices. In a major theoretical
work (with practical illustration of the effects of imputing value to
declining mortality rates) Dan Usher (50) developed a model which estimates
9 Utility subject to the parameters d (rate of discount) and 6 (the
3 Mortality Rate
elasticity of annual utility with respect to consumption).
66
-------
Bryan Conley (4) developed a theoretical model which serves to connect
the discounted earnings technique with the willingness to pay technique (at
least in terms of relative magnitude). He found that the value of saving a
life was greater than lifetime discounted earnings.
The practical difficulty with the willingness to pay concept has been
the measurement of the amount individuals are willing to pay for small reduc-
tions in the mortality rate. This is unlike the type of purchase decision
commonly faced by the average consumer. The concept, however, is not totally
alien to consumers. For example, when one purchases, installs, and uses
safety equipment, one is changing the probability of death and injury that
must be faced. A home fire alarm system lowers the probability of asphyxia-
tion while asleep. The use of automobile seat belts reduces the risk of death
or serious injury in an accident. The purchase of safe laboratory approved
home appliances gives some assurance against electrical shock. Even the
installation of non-slip material in the shower stall affords an improved
probability of survival. Other examples where safety is a component of the
purchase decision can be found.
Another technique to ferret out people's willingness to pay is to look
at the extent to which they are willing to be compensated to incur higher
risks. This behavior is prevalent in those occupations which have been shown
to be actually riskier than the average occupation. In these occupations, a
risk premium may be paid to encourage workers to engage in higher risk
activities. For example, coal mining is inherently riskier than office work.
Within the coal mining industry, underground work is riskier than aboveground
work. Underground miners receive extra compensation for working under
riskier conditions and this amount can be used to estimate their willingness
to pay to obtain small reductions in mortality.
The theoretical equivalency between "willingness to pay" and "willingness
to be compensated" can be demonstrated graphically. In Figure 10 we have
portrayed a small part of an individual's indifference map. Each point along
an indifference curve (U-| and U^) yields the same level of utility to the
individual. That is, the curves trace the combinations which result in.equal
levels of satisfaction to the individual. The vertical axis is denominated
in money terms (Y) and the horizontal axis is denominated in terms of
"quantity of safety" (S). The trade-off is thus income vs. safety and the
individual chooses his position subject to the constraints of total income
and the "price" of safety.
The following two sections examine the effect of a change in the price
of safety. First we examine the case where the price of safety falls. This
would be the situation if, using pesticides regulation as the instigating
event, the government were to unilaterally ban a pesticide which was linked to
increases in human mortality. The individual would then have to pay fewer
dollars to obtain the same level of safety.
67
-------
Quantity of Safety (S)
Figure 10. Safety indifference map.
68
-------
Decreases in Mortal ity--
The individual has an income of OM2 and is initially faced with the price
ratio (dY\ of P He maximizes his utility such that the price of safety (PJ
dS;
equals the marginal rate of substitution (MRS) along utility curve U-, (at
point A) and obtains OSa of safety. With the fall in price to P^ he again
maximizes his utility such that P^ equals the MRS along utility curve U2 (at
point B) and obtains OSfc of safety. In the process he has become better off,
moving to a higher indifference curve and increasing his consumption of safety.
To find out in monetary terms how much better off he is at his new equilibrium
point, we construct a line M-jZ with slope of Pb tangent to the lower in-
difference curve U]. If the individual's income had fallen by the amount
M2Mi at the same time the price fell from Pa to P^, he would have been just as
well off at C as he was at A (since he finds himself on his original indiff-
erence curve LJ-j). Since income remains constant and the price falls to PD»
then M2M] measures how much better off B is than A. Hicks (19) called this
amount the "compensating variation" for the price fall.
From this point it is possible to describe the change in the consumer
surplus which was obtained in the move from quantity OS, to OS^. The new
lower price allows the individual to obtain OS^ of safety. The net benefit
to the individual of the higher quantity is given by the vertical distance
between indifference curve U-[ and indifference curve U2 at the new quantity
obtained (OS^). This net benefit (or change in consumer surplus) is repre-
sented by BD. In Hicksian terminology, BD is the "compensating surplus" for
the price fall.
The above has described the situation of willingness to pay for reduc-
tions in probability of mortality. He now turn to the case of willingness to
be compensated for increases in probability of mortality.
Increases in Mortality—•
We again refer to Figure 10 to discuss the case where the price of
safety rises. This is the situation which exists when an employer asks an
employee to move to a riskier job. The individual would then have to pay
more dollars to obtain the same level of safety.
The individual has an income of OM2 and is initially faced with a price
ratio of P^. He is at equilibrium at point B and obtains OS^ of safety.
When the price of safety rises to Pa the individual maximizes his utility
where MRS = Pa (point A) and obtains OSa of safety. In the process he has
become worse off, moving to a lower indifference curve and lowering his con-
sumption of safety. How much worse off he is can be investigated by con-
structing a line (M]Z) with the slope of the old price ratio (P^) and setting
it equal to the MRS of the new indifference curve (U]). This portrays the
case where the individual's income is cut while at the same time holding the
price constant. At the new point of intersection (C) he is as bad off as he
is at A, yet the fall in utility from B to C is entirely the result of a fall
in income. This amount, M^M-], is called the "equivalent variation" for the
price rise, since it is the amount the individual's income would have to fall
in order for his utility level to be equivalent to that obtained with a price
rise alone.
69
-------
The decline in consumer surplus due to the rise in price of safety can
likewise be shown with the aid of Figure 10. The price rise results in a
decline in the amount of safety purchased. The loss in net benefit to the
individual is given by the vertical distance between indifference curves 1)2
and Ui at the old quantity. If BD is taken from the individual's income, he
is returned to the same indifference curve that he obtains with the price
rise. The amount represented by BD equals the amount which is lost from his
income and which lowers him to the new indifference curve at the old quantity.
This amount is called the "equivalent surplus" for the price rise.
It has been shown that the compensating variation for the price fall
(the case of improved opportunities for safety) is equal to the equivalent
variation for the price rise (the case of reduced opportunities for safety).
The compensating variation for the price fall (M2M-]) is the monetary measure
of how much the individual must be compensated in order to entice him into
the higher mortality rate situation embodied in the change of price from P^
to Pa.
For the purpose of the benefit-cost procedure, the appropriate measure
of gain or loss is the consumer surplus involved in the price change. For a
price fall, the compensating surplus (BD) is the gain in satisfaction that
the individual derives due to the higher quantity of safety obtainable. For
a price rise, the equivalent surplus (again BD) is the individual's loss in
satisfaction due to the lower quantity of safety obtainable. As a practical
matter, for small price changes, MoM] is virtually identical to the consumer
surplus measure needed in the benefit-cost analysis. This concept can be
illustrated with the aid of Figure 10. As the price changes embodied in Pa
and Pb become smaller (the slopes of the price lines anchored at income level
M2 approach each other), the difference (ED) between the appropriate variation
(M2^l) and the appropriate surplus (BD) becomes smaller. As ED approaches
zero, the surplus measurement (BD) approaches the variation measurement
(M2Mi).
How small a change in price is necessary to allow the assumption that
K^MI closely approximates the theoretically correct surplus measurement (BD)?
This cannot be precisely defined since the two values are asymptotically
approaching each other and will only equal each other when the slopes are the
same (i.e., the price is unchanged). However, for purposes of this procedure,
it is felt that for changes in the price of safety associated with changes in
mortality rates below 1 in 10,000 (.0001 change in the probability of death),
M2M] approximates the consumer surplus involved (see Mann (31) for a dis-
cussion of individual's perception of risk).
Once it is established that the amount an individual is willing to pay
for small changes in mortality rate is the correct measure of the change in
consumer surplus» the question now arises as to how the monetary amount of a
person's willingness to pay is found. Two approaches have theoretical
validity: compensation for risk-taking and questionnaires of willingness to
Pay.
70
-------
Compensation for Risk-Taking--
Compensation for risk-taking as a technique for estimating M£M] has the
advantage of being observable in the marketplace—specifically in the risky-
job example cited earlier. If the individual agrees to undertake a hazardous
occupation which increases his probability of death in exchange for a given
sum of monetary compensation, that sum can be used as an estimate of his
willingness to pay for safety. That incremental amount presumably reflects
his indifference map in that it returns him to the higher utility curve he
occupied before he took the riskier job. Thaler and Rosen (48) used data
which measures the relative riskiness of jobs to estimate the trade-off
between wages and risk. Thirty-seven broadly defined job classifications
shown to be actuarially riskier than the average occupation were matched
against a cross-sectional earnings survey. The results of their regression
indicate that individuals in these risky occupations were willing to be com-
pensated approximately $200 a year (in 1967 dollars) to incur an extra one in
a thousand risk of death.
Thaler and Rosen point out that there are several reasons to believe
that this estimate is conservative, especially when applying it to the general
population. The occupations surveyed in their study were approximately five
times riskier than the average U.S. occupation. People who take these jobs
have different reservation prices for risk than the individual who picks a
much less risky job. To use the derived estimate for the case of risk-
averting individuals places a severely conservative bias on the aggregate
willingness to pay. A further rationale for believing this estimate to be
conservative is that the riskiness of these jobs can in some measure be
affected by the individual employee. He thus is to some degree in control of
his personal risk level within the context of the risk level of his occupation.
If the individual is not in control of his personal risk exposure (for
example, the person far removed from the pesticide application site who
ingests pesticide residue in the food he eats), it is quite likely that he
would need to be compensated to a greater extent to undertake the unwanted
risk. This consideration is graphically illustrated by some people's
willingness to accept a relatively high probability of death (.25 in a
thousand) in an automobile accident while refusing to accept relatively low
probabilities of death (on the order of 1 in a million) due to a reactor
break-down at a nuclear power plant.
Questionnaires--
The other technique for assessing willingness to pay is the question-
naire method. It is widely recognized that this method has two basic
problems. The first is the "free-rider" aspect, in that the individual cannot
be excluded from the market. As a member of society he will derive benefits
based on the willingness to pay responses of others if the project results in
improved safety for all. The understatement of consumer surplus is likely to
be large if it is generally believed that the project will be instituted
regardless of the response to the questionnaire. This leads to the second
problem, that of assuring unbiased answers. If the respondent feels he will
actually be charged for the improved conditions according to his response to
the questionnaire, self-interest provides a powerful incentive to understate
his true willingness to pay. The necessity of couching the survey in terms
with an immediate impact on the individual ("How much would you be willing to
71
-------
pay?") makes this the major conceptual problem of the survey technique.
In spite of these difficulties, a methodology for determining willingness to
pay for small changes in human mortality using the questionnaire technique
has been developed by Mann, et.al. (31).
The results of the survey are somewhat disappointing since they diverge
substantially from findings of previous studies. While there is no way of
being sure of true risk preferences, the market derived estimate obtained by
Thaler and Rosen (48) of $200 for a 10'3 increase in risk is much lower than
that found in the Mann survey. On the other hand, the survey indicated that
individuals were able to answer the type of question posed, and that with an
improved survey instrument, the willingness to pay questionnaire may be a
workable concept.
Because of the limited population surveyed and the need to further
develop the questionnaire, it is inappropriate to extrapolate results of the
Mann survey to the general population. Until estimates are made on the
willingness to pay for small changes in mortality by the general population,
the results of the Thaler and Rosen occupational-risk study will be used in
further calculations.
HUMAN HEALTH EFFECTS OF ALDRIN ON CORN: A CASE STUDY EXAMPLE
The insecticide aldrin is primarily used as a treatment to control soil
insects which attack corn. Aldrin is rapidly metabolized into dieldrin which
is itself a commercial pesticide product, although much less widely used.
Since aldrin degrades so quickly into dieldrin, the remainder of this section
will consider the acute toxicity effects of aldrin and the chronic effects of
dieldrin on human health. The assessment follows the form described in
Figure 7.
We begin the health effects section with the question of whether any
health problems can be attributed to aldrin use. Certainly aldrin can be
acutely toxic. High dosages of aldrin affect the central nervous system
resulting in ataxia, tremors, respiratory failure and death. Ingestion of
from 1 to 3 grams (g) of aldrin will cause severe poisoning symptoms (13).
Consumption of such large dosages is restricted to occupationally exposed
workers who come in contact with high concentrations of the active ingredients.
These high concentrations, combined with aldrin's ability to be absorbed into
the skin, give rise to the possibility of acute poisoning. However, instances
of death due to aldrin poisoning in humans are quite rare and are the result
of gross mishandling of the pesticide (23). The probability of death due to
aldrin poisoning for a member of the non-occupationally exposed group is
exceedingly small.
The attribute which makes aldrin an attractive agricultural pesticide--
the ability to persist for a long period of time—is cause for concern when
related to human health. In addition, it is lipophilic; that is dieldrin
has a strong affinity for fat and is fat soluble. These two factors combine
to raise the question of what effect, if any, does the storage of a long-
lived metabolite of aldrin (dieldrin) in adipose tissue have on human health?
72
-------
The answer to this question may be obtained either through epidemiologi-
cal studies or through laboratory experimentation. The problems inherent in
the epidemic logical method have been detailed in the section headed "Chronic
injury" (p. 54). Briefly, the lack of controls, the usually small sample
size being examined, the relatively short time period over which a study
exists, and the potentially irreversible damage to humans if the pesticide
does result in a health effect, combine to make the epidemiological method a
poor approach to use to discover if the chemical does cause health problems.
It was on these grounds that an epidemiological investigation of aldrin by
Jager (22) was found to be inadequate.
The laboratory approach has been criticized on the grounds that animal
testing does not accurately reflect human susceptibility to chemicals. While
this point is arguable, it is in fact true that, except for arsenic, all
chemicals which have been shown to be oncogenic in humans have also demon-
strated their oncogenicity in laboratory animals. On at least one occasion,
laboratory results of oncogenicity presaged the equivalent finding in humans.
The discovery that exposure to polyvinyl chloride gas caused a rare form of
liver cancer in laboratory animals and man (6) lends credance to the use of
laboratory testing for screening purposes.
In a laboratory experiment using CF-1 mice, Walker, Thorpe, and Steven-
son (51) investigated the effect of dietarily ingested dieldrin. The experi-
mental concentrations ranged from 0.1 part per million (ppm) to 20 ppm and
the test periods lasted up to 132 weeks. A statistically significant
increase in tumors was found at all concentration levels. The conclusion
that may be drawn is dietary dieldrin does cause liver tumors in mice at the
dose levels tested. The potential for health problems in humans exists and
the analysis should proceed.
To ascertain the "tolerable" dose level for dietary dieldrin, a
statistical analysis of the Walker, Thorpe, and Stevenson data is performed.
The results of a pair of reports prepared by Mantel (34) (35) are given in
Table 12. The reason for the choice of the probit model as well as the
rationale for the selected levels of the variable parameters are given in the
section headed "Mathematical Extrapolation." Briefly, it is felt that the
choice of a slope greater than one and/or a confidence interval smaller than
99 percent results in an inadequately conservative estimate of the dose-risk
relationship. (The dose-risk relationships for alternative parameters are
also given in Table 12 for illustrative purposes.) The choice of "tolerable"
risk has been set at 10*8 (1 in 100,000,000); this number being not too far
from the intent of the "Delaney Clause."
Statistical extrapolation indicates that the dose associated with a
1/100,000,000 risk of contracting a liver tumor due to ingestion of dieldrin
is 11 parts per trillion (ppt) per day in the diet. This "tolerable" dose
level must be compared to the dose rate to which humans are exposed.
Because the laboratory experiment from which the extrapolation was drawn
delt in dietary exposure, it is correct to examine residues of dieldrin
present in the human diet. The Total Diet Studies series conducted by the
Food and Drug Administration is the best source of data on the average
73
-------
TABLE 12. DOSE-RISK RELATIONSHIPS FOR DIETARY DIELDRIN
USING THE MANTEL-BRYAN EXTRAPOLATION TECHNIQUE
RISK
10"8
io-7
io-6
io-5'
io-4
io-3
SLOPE = 1.0
c.l. = 99%
11
29
81
250
774
2438
ASSOCIATED
SLOPE = 1.0
c.l. = 95%
14
36
99
300
DOSE (ppt)
SLOPE =1.5
c.l. = 99%
580
1090
2160
4560
SLOPE =1.5
c.l. = 95%
650
1220
2430
5140
SOURCE: Mantel (35)
residues of dieldrin in food. The residue levels found in the survey are
associated with the food consumption of a 15-20 year old male (see Table 13).
This composition was chosen in order that residue estimates would be made on
a high consumption diet, lowering the probability that the dietary exposure
to pesticides would be underestimated. In the strict sense, the only indi-
viduals who have residue intakes as high as those found in the survey are
those on similar high consumption diets. However, in the case of dieldrin,
it is only necessary that a higher proportion of a smaller diet be made up of
dairy products and meat in order to receive dosages of the same order of
magnitude as found in the survey.
There is some evidence that, in the case of dieldrin, the diet survey
gives a conservative estimate of total exposure, in spite of the attempt to
make a liberal estimate of residue exposure by using high consumption diets.
Robinson (44) investigated the relationship between dietary intake of
dieldrin and the concentration of dieldrin in human adipose tissue and blood.
The relationship between total intake and residues was found to be:
74
-------
TABLE 13. RELATIVE COMPOSITION OF niET BY FOOD CLASS
1.
2.
3.
4.
5.
6.
7.
8.
9,
10.
n.
12.
FOOD GROUP
Dairy
Meats
Grain and Cereal
Potatoes
Leafy Vegetables
Legume Vegetables
Root Vegetables
Garden Fruits
Fruits
Oil and Fats
Sugars and Adjuncts
Beverages
TOTAL
AVERAGE GRAMS
PEP. DAY
769
273
417
200
63
72
34
89
222
51
82
1130
3402 g
PERCENT OF
TOTAL DIET
22.6
8.0
12.3
5.9
1.9
2.1
1.0
2.6
6.5
1.5
2.4
33.2
100.0%
SOURCE: FDA Total Dietary Study (15).
75
-------
Average total equivalent intake of
dieldrin in micrograms/day (ug/day) =
mean (geometric) concentration of dieldrin in adipose tissue (ug/g)
.0185
or
mean (geometric) concentration of dieldrin in blood (ug/ml)
.000086
The EPA's Human Monitoring Program is the source for data on U.S. human
adipose tissue residues. In fiscal year 1970 the nationwide geometric mean
was found to be 0.18 ug/g (25). Using Robinson's procedure, this translates
into an average total equivalent intake of dieldrin of 9.73 ug/day. This
compares with the 1970 average daily intake found in the Total Diet Study of
5 ug/day (9, p. 339). Since the diet is the dominant source of dieldrin
exposure (when compared to respiration or absorption) for the general public,
the diet study figure is probably low; how low cannot now be answered. To
allow for this we assume that the figures for dieldrin found in the Total
Diet Study are representative of the average dose per day in food for the
entire population of the U.S.. This results in a conservative estimate of
the dosage of dieldrin to which the general population is exposed.
In the fiscal year 1973 survey (15), typical market baskets of food were
collected in 30 U.S. cities and examined for pesticide residues. The 30
composite average for dieldrin is given in Table 14. The ma.ior sources of
dieldrin in the diet (on a ug/kg of body weight basis) are the dairy and meat
groups. As an indication of the ubiquitous nature of the pesticide, twenty-
six of the thirty dairy composites and twenty-nine of the thirty meat
composites contained measurable residues of dieldrin. These two food groups
accounted for 81 percent of the total average dieldrin residues found in the
survey. The grain and cereal, potato, garden fruits and oil and fats groups
each contributed minor amounts to the total.
As this is an examination of aldrin on corn, some assessment of the
proportion of the residues which can be attributed to corn usage must be made.
Aldrin on corn accounts for the vast majority of the pounds of active ingre-
dient applied, and after subtracting the pounds committed to uses not
associated with movement of residues (soil treatment for termites and treated
hot caps, for example), the percentage devoted to the corn rises further.
The assumption made in this analysis is that all dairy and meat residues are
connected to the soil-insect treatments on corn. This assumption follows
from the fact that most of the dairy and meat products in the U.S. either
originate in geographic areas where aldrin is applied to corn (where the
animals may be exposed to the pesticide directly through the air) or the
animals are fed corn and fodder which has been treated with aldrin. The
residues found in the other food groups would be difficult, though not im-
possible, to link to aldrin use on corn.
76
-------
TABLE 14. RESIDUES OF DIELDRIN BY FOOD CLASS
FOOD CLASS ppm ug/kg/day
1. Dairy .0016 .0180
2. Meats .0037 .0147
3. Grain and Cereal .0002 .0012
4. Potatoes .0009 .0025
5. Leafy Vegetables
6. Legume Vegetables
7. Root Vegetables
8. Garden Vegetables .0029 .0038
9. Fruits
10. Oil and Fats .0001 .0001
11. Sugars and Adjuncts
12. Beverages -- --
SOURCE: FDA Total Dietary Study (15).
.0403
77
-------
The result of this assumption is to set the average dietary intake of
dieldrin residues due to corn usage at .0327 microgram/kilogram (ug/kg) of
body weight/day. To convert this figure into parts per trillion of the diet,
the following procedure is followed:
[Residue in food] [Mass of organism]
Dose = [Quantity of Food]
ug
m
To express in parts per trillion: ppt = 10 , -^ = 10
Ppt 10"12 6
Iig7g = 10-6 - 10
conversion factor becomes: 1 = 10 ^P .
ug/g
So, Dose = r Residue] [Mass] ,
[Quantity of Food] L|U J
Using data from fiscal 1973 Total Diet Study,
Residue = .0327 ug/kg
Mass of organism = 69.1 kg (average mass of a teenaged boy)
Quantity of food = 2402 g/day*
Dose (ppt) - [-0327 ug/kg] [69.1 kg] _
[2402 g/day] L'O J -
The average daily dietary consumption of dieldrin is 940 ppt. This is above
the calculated "safe" dose and results in a finding that a potential human
health problem does exist.
All previous discussion of residue levels was in terms of an "average"
of the 30 composite market baskets examined. If the data is available, it
is preferable to establish actual residue levels and the number of people
*The weight of the average diet (3402 grams/day) is reduced by the quantity
of water included in the average diet in order to maintain comparability
with the animal experiment from which the extrapolation is performed.
This results in an average diet of 2402 grams/day, excluding water.
78
-------
associated with each level. As a practical matter, however, such a fine
delineation between residue levels for the general population is not possible.
The results of the Total Diet Studies show wide variations of residues in
food groups obtained in the different sample cities. An attempt to fine tune
the procedure by weighting the residue found times the population surrounding
the sample city is an unwarranted use of the study. It is best to use the
average figure, keeping in mind residue levels at. any given location may be
above or below the average. With this understanding, the estimation of the
average risk is performed.
The equation which may be used to convert dose into the upper limits of
risk depends upon the variable parameters. With the slope equal to one and
the confidence level equal to 99 percent, the equation'(derived from data in
(33) and (34) is closely approximated by:
ppt = (77 x 103) & + 3.3
where: R is the risk level, R ? 0
ppt is the dose level
Under the probit extrapolation technique (with the variable parameters
set as previously described) the 940 ppt dose translates into an upper limit
to risk of 1.5 x 10-4 (1.5 in 10,000). An individual of the U.S. faces a
risk as high as 1.5 in 10,000 of contracting tumors due to the ingestion of
dieldrin residues resulting from the use of aldrin on corn.
o
The value attached to reducing this risk to the "tolerable" risk of 10"
is estimated in the following manner.
Recognizing that an individual accepts higher risks up to the point
where the compensation equals the expected value of the higher risk, Thaler
and Rosen (47), as previously discussed, found a willingness to be compensated
for higher risk jobs of $200 per year per an increased risk of 10~3.
Assuming a linear extrapolation,
compensation = (value of life) (risk level)
$200 = V (10~3)
$2 x 105 = V
The evaluation under consideration in the dieldrin example is a movement from
a higher risk level to a lower risk level—namely, 1.5 x 10"4 to 1 x 10-8-
Accordingly, the equation becomes:
compensation = (V) (risk level #1) - (V) (risk level #2)
= ($2 x 105) (1.5 x 10"4) - ($2 x 105) (1 x 10"8)
$30 - $.002
= $29.998
79
-------
Thus, an individual with the given risk preference and facing the above
change in probability will be willing to pay* $29.998 to attain the lower risk,
Taking 214 million individuals to be the approximate U.S. population, the
aggregate willingness to pay to avoid liver tumors connected with aldrin use
on corn is $6,420,000,000. This estimate is in terms of 1967 dollars. A
conversion to current dollars should be made. If the consumer price index is
used as a measure of the change in the value of a dollar, the aggregate will-
ingness to pay (in 1975 dollars) is $10,349,000,000.
The change in consumer surplus ($10,349,000,000), is the figure which
enters the benefit-cost procedure from the human health section and is
compared with the change in consumer surplus from the production section.
*
The equivalence between willingness to be compensated and willingness to
pay is shown in sections headed "Decreases in Mortality" (p. 66 and
"Increases in Mortality" (p. 66).
80
-------
BIBLIOGRAPHY FOR HUMAN HEALTH CHAPTER
1. Brown, J. D. Effect of a Health Hazard Scare on Consumer Demand.
American Journal of Agricultural Economics, 51(3): 676-679,
August 1969,
2. Cicchetti, Charles J. and A. Myrick Freeman III. Option Demand
and Consumer Surplus: Further Comment. Quarterly Journal of
Economics, 85(3), August 1971.
3. Clark, John Maurice. The Costs of World War to the American People.
Yale University Press, New Haven, 1931.
4. Conley, Brian C. The Value of Human Life in the Demand for Safety.
American Economic Review, 66(1): 45-55, March 1976.
5. Cranmer, Morris F. Reflections in Toxicology. Journal of the
Washington Academy of Sciences, 64(2): 158-579, 1974.
6. Creech, J. L., and M. N. Johnson. Angiosarcoma of Liver in the
Manufacture of Polyvinyl Chloride. Journal of Occupational
Medicine, 16: 150-151, 1974.
7. Darby, William. The Philosophy of Acceptable Risk and the Practicality
of Maximum Safety. In: Environment and Quality of Food,
P. L. White and Dianre Robbis, eds. Futura Publishing Company,
Mt. Kisco, N. Y., 1974.
8. Dubin, Louis I. and Alfred J. Lotka. The Money Value of a Man.
Ronald Press, 1946.
9. Duggan, R. E. and P. E. Corneliussen. Dietary Intake of Pesticide
Chemicals in the U.S. (Ill), June 1968-April 1970. Pesticides
Monitoring Journal, 5(4), March 1972.
10. Edwards, W. F. Economic Externalities in the Agricultural Use of
Pesticides and an Evaluation of Alternative Policies.
Ph.D. Thesis, University of Florida, 1969.
11. Epstein, Samuel S. and Marvin S. Legator. Mutagenicity of Pesticides.
The MIT Press, Cambridge, Mass., 1971.
12. Fein, Rashi. Economics of Mental Illness. Joint Commission on
Mental Illness and Health Monograph Series No. 2. New York:
Basic Books, Inc., 1958.
81
-------
13. Finkelstein, Harold. Preliminary Air Pollution Survey of Pesticides,
A Literature Review. Prepared for Public Health Service, U.S.
Department of Health, Education and Welfare, October 1969.
14. F.D.A. Advisory Committee on Protocols for Safety Evaluation. Panel
on Carcinogenisis Report on Cancer Testing in the Safety Evaluation
of Food Additives and Pesticides. Toxicology and Applied Pharma-
cology, 2: 433, 1971.
15. F.D.A. Total Diet Studies: Fiscal Year 1973, Compliances Program
Evaluation. Bureau of Foods, Unpublished Report, January 9, 1975.
16. Feldstein, Martin S. Distributional Preferences in Public Expenditure
Analysis. In: Redistribution Through Public Choice, Hochman and
Peterson,eds. Columbia University Press, 1974.
17. Fromm, Gary. Civil Aviation Expenditures. In: Measuring the Benefits
of Government Investment, R. Dorfman,ed. Washington, D.C.:
Brookings Institute, 1965.
18. Headley, J. C. and J. N. Lewis. The Pesticide Problem: An Economic
Approach to Public Policy. Johns Hopkins Press, Baltimore, Md.,
1967.
19. Hicks, J. R. A Revision of Demand Theory. Oxford University Press, 1956.
20. Hirshleifer, J., T. Bergstron and E. Rappaport. Applying Cost-Benefit
• ' " ' -" School of
les, California,
iiener, u., i. oeryburun aim c.. r\ap(jajjur L. M(jpiyiny
Concepts to Projects Which Alter Human Mortality. School of
Engineering and Applied Science, U.C.L.A., Los Angeli
1974.
21. Hoover, Robert and Joseph F. Fraumeni, Jr. Cancer Mortality in U.S.
Counties with Chemical Industries. Environmental Research, 9:
196-207, 1975.
22. Jager, K. W. Aldrin, Dieldrin, Endrin and Telodrin: An Epidemiological
and Toxicological Study of Long-Term Occupational Exposure.
Elsevier Publishing Company, Amsterdam, 1970.
23. Kazantzis, G. A., I. G. McLaughlin, and P. R. Prior. Poisoning in
Industrial Workers by the Insecticide Aldrin. British Journal of
Industrial Medicine, 21(46), 1964.
24. Klarman, Herbert E. Syphilis Control Programs. In: Measuring the
Benefits of Government Investments, Robert Dorfman, ed. Washington,
D.C.: Brookings Institute, 1965.
25. Kutz, F. W. et al. Pesticide Residues in Adipose Tissue of the General
Population of the United States, Fiscal Year 1970. Bulletin of
the Society of Pharmacological and Environmental Pathologist, 2(3),
March 1976.
82
-------
26. Lave, L. B. and E. P. Seskin. Health and Air Pollution: The Effect
of Occupation Mix. Swedish Journal of Economics, 73(1): 76-95,
March 1971.
27. Lave, Lester and Eugene P. Seskin. Air Pollution and Human Health.
Science, 16°: 723-732, August 1970.
28. Lave, Lester and Warren Weber. A Benefit-Cost Analysis of Auto Safety
Features. Applied Economics, 2: 265-275, 1970.
29. Lawless, E. W. et al. The Pollution Potential in Pesticide Manufacturing.
Pesticide Study Series No. 5, Office of Water Programs, Environ-
mental Protection Agency, June 1972.
30. Lederberg, J. 'Forward1 to S. Epstein and M. Legator. The Mutagenicity
of Pesticides. MIT Press, Cambridge, Mass., 1971.
31. Mann, Stuart H., Thomas J. Douglas, Jack L. Nasar and F. Roger Tellefsen.
The Determination of the Economic Effects of Pesticide Use on Human
Health: A Gaming Approach to Willingness-to-Pay. The Center for
the Study of Environmental Policy Working Paper Series, No. 32,
The Pennsylvania State University, University Park, Pa., 1976.
32. Mantel, Nathan and Marvin A. Schneiderman. Estimating 'Safe' Levels--
A Hazardous Undertaking. Cancer Research, 35(6): 1379-1386, 1975.
33. Mantel, Nathan, R. Bohidar Neeti, Charles C. Crown, Joseph L. Ciminera,
and John W. Turkey. An Improved 'Mantel-Bryan1 Procedure for
'Safety1 Testing of Carcinogens. Cancer Research, 35(4): 865-872,
1975.
34. Mantel, Nathan. Estimating Limiting Risk Levels From Orally Ingested
DDT and Dieldrin Using an Up-Dated Version of the Mantel-Bryan
Procedure. Prepared for Environmental Protection Agency, April
1974.
35. Mantel, Nathan. Estimating Limiting Risk Levels From Orally Ingested
DDT and Dieldrin Using an Up-Dated Version of the Mantel-Bryan
Procedure. Supplementary Report, Prepared for Environmental
Protection Agency, December 1974.
36. Mantel, Nathan and W. Ray Bryan. 'Safety' Testing of Carcinogenic
Agents. Journal of the National Cancer Institute, 27(2): 455-470,
1961.
37. Mishan, E. J. Economics for Social Decisions. Praeger Publishers, Inc.,
1973.
38. Mishan, E. J. Cost-Benefit Analysis. Unwin University Books, 1972.
39. Mishan, E. J. Pangloss on Pollution. Swedish Journal of Economics,
73(1): 113-120, March 1971.
83
-------
40. Mishan, E. J. The Postwar Literature on Externalities: An Interpreta-
tive Essay. Journal of Economic Literature, March 1971.
41. Mishan, E. J. Evaluation of Life and Limb: A Theoretical Approach.
Journal of Political Economy, 687-705, July-August 1971.
42. Mishan, E. J. Welfare Economics: An Assessment, North Holland,
Amsterdam, 1969.
43. Mishan, E. J. Reflections on Recent Developments in the Concept of
External Effects. Canadian Journal of Economics and Political
Science, 31: 3-34, 1965.
44. Robinson, J. The Relationship Between the Dietary Intake and the Body
Burden of HEOD (Dieldrin). European Journal of Toxicology, 2(4):
210-211, 1969.
45. Rice, Dorothy P. Estimating the Costs of Illness. American Journal of
Public Health, pp. 424-44 , March 1967.
46. Schelling, Thomas. The Life You Save May Be Your Own. In: Problems in
Public Expenditure Analysis, S. Chase ed. Washington, D. C.:
Brookings Institute, 1968, pp. 127-176.
47. Smith, Wayne E. Factors Associated with Age-Specific Death Rates,
California Counties, 1964. American Journal of Public Health,
58: 1937-1949, October 1968.
48. Thaler, Richard and Sherwin Rosen. The Value of Saving a Life:
Evidence from the Labor Market. NBER Conference on Income and
Wealth, 1973.
49. U.S. Department of Health, Education and Welfare. Report of the
Secretary's Commission on Pesticides and their Relationship to
the Environmental Health. Washington, D.C., December 1969.
50. Usher, Dan. An Imputation to the Measure of Economic Growth for
Changes in Life Expectancy. In: The Measurement of Economic
and Social Performance, M. Moss ed. Studies in Income and
Wealth, No. 38, National Bureau of Economic Research, 1973.
51. Walker, A. I. T., E. Thorpe and D. E. Stevenson. The Toxicology of
Dieldrin (HEOD). Food and Cosmetic Toxicology, 11: 415-432,
1972.
52. Warnick, S. L. and J. E. Carter. Some Findings in a Study of Workers
Occupationally Exposed to Pesticides. Archives of Environmental
Health, 25(4), 1972.
84
-------
53. Weisbrod, Burton A. Economics of Public Health, Measuring the Economic
Impact of Diseases. Philadelphia: University of Pennsylvania
Press, 1961.
54. Winch, D. M. Analytical Welfare Economics. Penguin Books, Middlesex,
England, 1971.
55. Zeckhauser, Richard. Procedures for Valuing Lives. Public Policy,
23(4), 1975.
85
-------
CHAPTER 5
ENVIRONMENTAL IMPACTS*
INTRODUCTION
The use of pesticides has saved countless lives by controlling human
disease vectors and by increasing the yields of agricultural endeavors. How-
ever, there is a growing consciousness that in addition to endangering human
health,the environment is becoming polluted by chemicals that can harm plants
and animals as well.
This chapter is concerned with assessment of the effect of pesticides on
plants"and animals not raised for sale in the market place. Included under
this heading are all manner of interrelated plant and animal associations and
aesthetic considerations. It is in this area that a thorough investigation
of the pesticide-caused effects will pay a dividend in the avoidance of
"surprises" after initial registration or reregistration. By following chem-
icals, energy, etc. up the food chains, the basis for an impact identification
and assessment scheme can be developed and, if desired, the impacts avoided.
The inherent problems of pesticide-use are surveyed in this section and
analytical tools are developed which will afford decision-makers a more
realistic and comprehensive view of potential pesticide impacts on the en-
vironment.
First, the general effects of pesticide use will be delineated through
discussion of the chemically induced alterations in the terrestrial sector
(flora, soils and soil organisms, and higher trophic levels) and the aquatic
sector (fresh water and esturine/marine flora and fauna).
Second, we will review some of the existing measurement techniques and
illustrate why these schemes do not readily lend themselves for application
to pesticide-use decisions.
This chapter relies primarily on descriptive material to examine pesti-
cide impacts on the environment and aesthetic considerations. Quantification
techniques commensurate to benefit-cost analysis, such as those set forth in
the chapters on economic production and human health, do not exist at the
present state-of-the-art for this chapter. Descriptive material is used, in
*Principal researcher in this section is Mr. Gary A. Shute, Center for
the Study of Environmental Policy, The Pennsylvania State University.
This section was prepared under the supervision of Dr. Donald J. Epp,
Principal Investigator.
86
-------
lieu of superior approaches, to lay the basis for the development of a work-
able and conceptually correct evaluation technique for environmental and
aesthetic impacts.
The development of an assessment technique consisting of adaptations of
the Leopold matrix and Delphi method will then be presented. This section,
through the use of food chain models and the evaluation matrix, will show the
potential environmental impacts of pesticides based on the information
presented earlier in the text.
Lastly, the evaluation procedure will be used with a specific example
pesticide (aldrin-dieldrin) to further illustrate the envisioned use of the
assessment scheme in actual pesticide-use decisions.
GENERAL REVIEW OF PESTICIDE MOVEMENT AND PROBLEMS
It is commonly acknowledged that the world cannot afford to abandon the
use of pesticides. Crop production must be maintained and increased without
significantly increasing the amount of agricultural land. Decreases in
productivity due to pesticide restrictions cannot be fully offset by bringing
marginal lands into production (11). A shift in demand towards high protein
food as a country develops economically may bring about the discarding of
former dietary habits; seriously impacting land use and pesticide-use
decisions. Developing countries need cheap, effective pesticides to enable
them to expand agricultural output (1) (47) (55) (61) (78) (81). Thus,
pesticide residues will remain an acute environmental problem in the future
(7) unless economically viable, low-residue pesticides or other pest control
methods can be developed.
Even though pesticides have greatly benefited man, chemical and biologi-
cal control agents employed to ameliorate factors that adversely affect man
or his activities have their inherent problems. It is these imperfections
that can cause deleterious effects in the biosphere.
Pesticides are formulated to kill or control a wide array of species
considered harmful to man. The basic problem with these chemical formulations
is that their effects are not limited to the intended or target organism.
Through the use of pesticides, non-target organisms can be killed either
directly or indirectly via the consumption of contaminated foodstuffs, plants,
or animals. Pesticides can be transferred from their original application
sites to other locations by erosion, drift, runoff and biological transfer.
The acute effects of pesticides, derived from a single exposure to the
parent compound or its metabolite, are more readily observable than chronic
effects. There are many natural factors affecting the uptake and accumulation
of residues throughout the biosphere: choice, longevity, and formulation of
the pesticide based on the pest to be controlled, and the target area's
degree of isolation and climate. Formulation and application are greatly
dependent on plants' abilities for translocation from the root or leaf (23)
(80). In addition, particular plant species will differ in penetration,
absorption and translocation between species and compounds. Soil
87
-------
Insects
In reviewing the literature on the work that has been done on the routes
of pesticide movement in insects, most experimentation has been confined to
beetles which are predators or associated with decaying animal or vegetable
matter. Three species of beetles (Harpalus oaliginosus, H. pennsylvanious3
and Poecilus ehalcites) were taken from a corn plot that had been treated
with aldrin for fifteen years and analyzed for residues (52). The results of
the analysis showed that Harpalus had 3.5 times the soil residue and P.
ehaleites had 31 times the soil residue and 9.0 times the residues found in
Rar>palus. The difference in residues was attributed to the fact that P.
ohaloites eat more animal matter that may already have residues whereas
Harpalus eat mainly vegetation. This may be taken as an indication of bio-
magnification along the food chain.
Birds
The effects of pesticides on birds have been studied in great detail and
include altered response to stress, altered liver function, altered vitamin
content, abnormal testicular development, delayed ovulation, increased
metabolism of steroids, failure to lay eggs, failure to deposit calcium in
shells of eggs, decreased number of eggs, increased egg breaking and eating,
inability to renest, decreased viability of the young, and increased tetra-
togenesis. More detailed information on pesticide residues and their effects
on birds can be found in Edwards (26).
Several specific examples are worth noting. Ratcliffe (77) has reported
widespread, rapid and synchronous declines in weight and thickness of shells
of eggs layed by British peregrine falcons (Falao peregrinus) and sparrow-
hawks (Accipiter nisus). A causal chain between the phenomenon of frequent
egg breakage, subsequent status of breeding populations and exposure to
persistent organic pesticides has been made (77). Pesticides stimulated
hepatic microsomal oxidation of drugs and microsomal hydroxylation of steroids
(12). Bengalese finches (Lonahura striata) fed DDT had decreased fertility,
hatachability and fledgling success (48). With pheasants (Phas-ianus oolahicus)
the feeding of near lethal doses of dieldrin had no apparent effect on that
generation; however, the second generation of pheasants had significantly
decreased hatchability and fertility, which seems to imply the possibility of
genetic aberrations (3).
Laboratory experimentation is indispensable in studying the effects of
pesticides on the environment, but it is of paramount importance that the
laboratory findings be correlated with actual field observations. Most of
the laboratory experimentation is concerned with the determination of LDr^'s
and LC 's for various organisms in an attempt to indicate what the
impact of a particular pesticide will be on the species in question.*
0 is lethal dosage that results in 50 percent mortality of the test
species; 1.050 is lethal concentration in air, feed or water resulting in
50 percent mortality.
88
-------
characteristics and weather will markedly influence residue accumulation. The
effects of residues and acute toxicities on plant species are also dependent
uoon plant metabolism.
Soil type also affects the availability of residues to both target and
non-target organisms. If there is organic matter in the soil, there is less
uptake of pesticides due to soil binding. For example, wheat, orchard grass,
alfalfa, and corn had six times as much of a specific pesticide residue
present when grown in sand than when grown in soil (97). Carrots grown in
mineral soil with low applications of pesticides contained higher concentra-
tions than carrots grown in organic soil with higher pesticide applications
(41). Additionally, soils with clay colloids are highly significant sorbers
of chemicals.
According to a study done by the Agricultural Research Service, USDA,
under interagency agreement with the Office of Research and Development, EPA
(93), even if there is heavy rainfall shortly after pesticide application the
concentrations of the chemical in the runoff are very low—approximately five
percent of the total application per year. However, this low amount of pesti-
cide in the runoff is important since some chemicals are highly toxic to fish
and other aquatic fauna and their persistence in the aquatic environment can
be quite lengthy. In eva-luating possible environmental impacts due to pesti-
cide residue in runoff, toxicity and persistence must be considered together.
When a pesticide residue is dissolved in water it can move from the applica-
tion site and may not be tied into the sediment. Therefore, if runoff and
erosion can be controlled, the probability of residue contamination in other
areas will be lessened.
Several methods can minimize pesticide movement from agricultural
application areas. These methods include:
• Adopt product!on techniques which do not use pesticides;
• The using of alternative pesticides that are not water
soluble;
• Placement of pesticide with respect to movement;
• Optimizing pesticide formulation;
• Reducing excessive treatments;
• Optimizing the time of day for pesticide application;
• Optimizing the dates of pesticide application;
• Using lower pesticide application rates;
• Managing aerial application.
89
-------
All of these measures are aimed at reducing the impacts of pesticides on the
environment, but it should be noted that any one or combination of these
control techniques may not be implemented due to physical or economic
constraints.
In general, there are three major impact areas which encompass environ-
mental effects: the terrestrial, the fresh water aquatic, and the esturine/
marine aquatic environmental sectors. Our discussion will be focused on the
first two of these sections since esturine/marine effects have not been as
extensively documented. Environmentally there are basically three types of
ecological consequences which have been identified (55). Pesticides can
affect individual survival, reproductive success, and the genetic constitution
of future generations.
EFFECTS OF PESTICIDES ON THE TERRESTRIAL ECOSYSTEM
Soil Invertebrates
Soil invertebrates can be readily affected by pesticides due to their
proximity and contact with application sites (e.g., direct application to soil
or drift and fallout from application to plants). There has been comparatively
little work done in this area, but there are a few examples of the possible
impacts pesticides could have on non-target soil invertebrates. Earthworms
help to establish and maintain the structure and fertility of the soil, an
activity with obvious agronomic importance (84). Therefore, increased
earthworm mortality may affect soil fertility. In areas where DDT was used
to control Dutch elm disease, residues in earthworms were eight times greater
than in worms from nearby but untreated soil (45). Residues in earthworms
are a function of earthworm species and habitat and residue concentration and
stability. While there are numerous examples of the uptake and storage of
pesticide residues by earthworms in forests, orchards, and arable land, resi-
due concentration tends to be greater from forests and orchards than from
arable lands. Other soil invertebrates that have been studied to date include
molluscs (snails and slugs). Reports indicate that residue retention in slugs
tends to be'greater than in earthworms taken from the same site (91). A
comparatively small amount of work has been done, however, and what has been
done seems to suggest that the uptake mechanism by snails and slugs is
similar to that of earthworms.
The more persistent pesticides may cause more sub-lethal effects (66).
The extent of these sub-lethal effects is quite important. Earthworms, for
example, become easier prey for birds which tend to accumulate residues
themselves thus illustrating food chain transfer of pesticide residues and a
possibility for biomagnification (18) (50). Sub-lethal doses of pesticides
can also affect the predator-prey relationship in that the predator may now
become prey due to altered behavior patterns (52).
90
-------
However, it is extremely difficult for these in vivo studies to indicate food
chain reactions successfully in the environment due to the absence of labora-
tory controls in the field. That is, single route exposures in the laboratory
may not correctly represent polyvalent field exposure through dermal uptake,
inhalation, drinking, and ingestion. Some field studies indicate that the
breeding success of the woodcock (Philohela minor) in New Brunswick varied
inversely with the amount of DDT used and with the total area sprayed (99).
In Wisconsin, when elm trees were sprayed with greater than ten pounds per
acre of DDT, the wild robin (Twdus migratorius) population had a greater than
85 percent spring mortality and, due to the loss of adults, virtually no
reproduction (46). In addition, there was reported an increase in reproduc-
tive failure of Western grebes (Aecfanorphorus occidentalis) due to DDT at
Cedar Lake, California (42). Poor reproduction in the golden eagle (Aquila
chrysaetos) in Scotland was correlated to the amount of organochlorines
ingested (59). The list of comparable examples is quite extensive and has
been condensed to the above mentioned examples to afford some understanding
of the possible consequences of pesticide use or misuse and the long term
effects of pesticidal residues in the biosphere.
Mammals
One of the major sectors of the environment which can be affected by
pesticidal agents is mammals. The impacts upon mammals will be con-
sidered in detail in the section on Aldrin-Dieldrin, and only two specific
effects will be noted here. Frist, shrews (Mcrosorex hoyi) have been shown
to have a particular propensity for residue accumulations (22). Second,
mink (Mustela vision) accumulated greater DDT residues than the hares (Lepus
americanus) on which the mink fed, once again indicating possible biomagnifi-
cation. Examples of wild mammal mortality due to aldrin, dieldrin, and
endrin can be found in Dustman and Stickel (24).
FRESH WATER AQUATIC EFFECTS
When exposed to pesticides, fresh water flora and fauna exhibit many of
the same reactions as those discussed in the terrestrial section. In the
case of fresh water flora, there was shown to be a decrease of 40-90 percent
in phytoplankton (Scenedesmus quadrieauda) production due to organochlorine
exposure (76). Fish represent the sector of the fresh water environment that
has the greatest amount of research effort. Certain species of fish have
demonstrated the ability to develop strains resistant to chlorinated hydro-
carbons. These resistant nontarget organisms include some aquatic inverte-
brates and six species of fish and amphibians (29) (30). Resistant mosquito
fish (Gcmbusia affinis), when bred in a non-contaminated system, retained
their resistance to the pesticide.
Resistance to pesticides can be attributed to a number of different
factors, alteration of permeability at sites of action, changes in respiratory
surfaces, increases in fat content, changes in the excretory system,
detoxifying enzyme systems, novel metabolic pathways and rates, and changes
in antibody reactions (25). The major problem afforded by the development of
resistant strains of fish is that the resistant strain will carry greater
91
-------
body burdens to higher trophic levels. Resistance enables the adapted species
to survive the environmental stress (e.g., pesticides) so that the survivors
are physiologically prepared to cope with a particular stress.
Alteration of normal activity patterns has been noted in the Atlantic
salmon (Salmo solar) which underwent changes in its temperature selection
gradient (69). It has also been noted that pesticide exposure can alter the
salinity preference of fish. One must consider what effect the alterations
in activity patterns will have on the distribution and survival of the
organisms.
ENVIRONMENTAL IMPACT ASSESSMENT
Where pesticides are used to a great extent there exist possibilities for
synergistic or antagonistic interactions between different pesticides. Studies
have shown that there was an interaction between dieldrin and DDT which
affected the accumulation in pyloric caccae of rainbow trout (Salmo gairdneri),
so that DDT residues increased and dieldrin residues decreased over the amounts
taken up from the amount of single insecticides (60).
There is a dual threat of pesticidal accumulation in fish. The first
threat is that the residues in fish may exceed man's toxicity level (tolerance)
but this has been judged unlikely. Secondly, pesticides could bring about the
elimination of certain species that are an important source of food and
recreation by exceeding threshold toxicities (25). One must be cognizant of
the distinction between the elimination of a resource because of its physical
decline (toxicologic impact) and the elimination of resource availability due
to regulation (exceeding tolerance levels). In addition, secondary losses
through increased mortality via reproductive failures and increased predation
resulting from alterations of mobility and activity patterns is of the utmost
importance.
Our understanding of nature's interrelationships is extremely limited.
The tentative nature of human health effects noted in the previous section
shows that man is only beginning to understand the process by which chemicals
affect our own species. By comparison, our level of understanding of the
effects of chemicals on the myriad of lifeforms in nature is in its infancy.
Even within a related family of animals (rodents, for example), the effect of
chemicals is likely to differ among the individual species.
Importance of Species Survival
If a species can be identified as threatened by the application of a
pesticide, the assessment of the effect of its demise can begin. It is at
this point that a thorough understanding of its place in the ecosystem is
important, since its extermination will ripple through the environment as
illustrated in the previous section. In general, there will be an impact on
the organisms upon which it feeds—without the downward pressure of predation,
the food organism may multiply greatly, causing its own impacts to change.
(We see a small scale example of this effect when, in trying to limit the
growth of a harmful insect, a pesticide inadvertently kills the target
92
-------
insect's natural enemies. Without the depressive effect of the predator, the
target insect rapidly reestablishes itself, often at levels higher than before
the spraying.) There will also be an impact upon the organisms which feed on
the threatened species—removal of a food supply could lead to extinction of
a higher predator, if it is a specialized feeder, or merely a temporary dis-
location until other food sources can be utilized by a general feeder.
The concept of option demand, discussed in Chapter 4, is useful in
measuring the importance of species survival. Once the species is gone the
possibility of ever observing it has passed. Certainly the attributes which
describe option demand existence are present; there is uncertainty about the
future demand for the commodity, there is an element of irreversibility in
the decision to supply a commodity in the future, and there is no way for the
resource owner to exclude those who do not pay for the benefit (9). Thus if
consumer surplus can ever be measured for environmental preservation, an
additional amount will have to be added for option demand maintenance.
Measurement of Environmental Effects
This section examines the question of whether environmental effects of
pesticides can be measured. The past decade has seen rapid growth of concern
about the conditions of the natural environment. Just from looking at the
complexity of the problem, one could say measurements would be quite difficult.
Evaluation of man's impacts on the environment is still in its infancy. Un-
fortunately, present econometric techniques do not allow us to make the type
of measurements we would like. We are dealing with a public good—the
environment—all of which is not impacted in each pesticide-use decision.
Incorporation of environmental impacts into a benefit-cost or other
framework is imperative. According to Howe the problems inherent in environ-
mental measurement would be minimized "if there were only two ingredients to
be considered, (a) infallible indices of environmental conditions and, (b)
value weights that, when multiplied by the indices, would permit environmental
values to be added in with the normal benefits and costs of the action under
consideration." (44, p.26) Such a procedure is clearly not available at
present, but this is not to say that environmental impacts should not be
considered at all when attempting to evaluate the desirability of a particular
action.
Much of the work to date has been confined to the area of water resource
planning and evaluating the impacts of aesthetic considerations. Thus far,
studies which have attempted to measure environmental effects quantitatively
have usually used the expenditures approach. In this method, the expenditures
of those who partake in the use of environmental resources are totaled and
the sum is used as a proxy for the value of the wildlife involved. On a
nationally aggregated basis, this technique may measure the value of goods
and services consumed in the enjoyment of the wild. However, as a measure of
the value of the wildlife, it has serious theoretical problems. Furthermore,
the environmental impact of a pesticide is likely to impact on only a small
number of animals or plants. Imputing a value from such a national survey
would be impossible. Use of this technique on a local level would still only
give a theoretically defective measure of aggregated wildlife value with no
capability for species specificity.
93
-------
There are numerous studies which attempt to measure the value of a
specific outdoor activity—usually hunting or fishing. There is some question
of what should be used as the value variable—the license fee paid to the
state, the license fee olus travel costs of reaching the site, or the net of
all costs over that which would have been spent if the activity was not
engaged in. These studies do, however, answer questions of concern to hunting
and fishing agencies.
The problem with relying on this method of valuation in pesticide matters
is that often the threatened species is not hunted or fished. The implication
of the expenditures method is that these species are not valued. If a species
was once hunted but now has been placed on the protected list, is its value
now zero? If it is protected, the species has value to someone, so clearly
the answer to this question is no.
The previously described methodologies do not provide a complete assess-
ment of environmental impacts in a manner which can be successfully integrated
into a benefit-cost framework providing a comprehensive view of environmental
consequences. In addition to these methods, there are numerous other schemes
which attempt to incorporate environmental values into the decision-making
process, but are mainly limited once again to water resource planning,
aesthetic consideration, and highway construction planning. An evaluation of
seventeen such measurement techniques can be found in "A Review of Environ-
mental Impact Methodologies" by Warner and Preston (95). Most of the models
are designed to be site-specific, such as the Dee (19) (20) techniques for
water resource projects and the Little (57), McHarg (62) and Smith (87)
techniques which are primarily concerned with the construction of transporta-
tion facilities.
It is unlikely that the application of pesticides will be confined to a
relatively small area and thus the use of site-specific models does not
possess the ability to account for all the possible environmental consequences
that are attached to pesticide introduction into the biosphere. In fact, no
such technique exists today, but it is quite clear that the environmental
consequences of pesticide use cannot be ignored.
In reviewing the techniques available to date, it appears that a tech-
nique with less site-specificity and general overall flexibility would be
the most appropriate. In an attempt to incorporate pesticidal impacts into
a benefit-cost analysis, we have developed a methodology that is a meld of
the Leopold matrix (56) and the Delphi technique (16). The Leopold matrix is
a grid-structure interaction device in which pesticide effects and impacted
organisms can be identified and then assessed on a relative qualitative and
quantitative basis; with the perspective of the evaluator critical to assess-
ment determination. The information obtained from the matrix analysis is
then further scrutinized via the Delphi technique which consists of a multi-
round iterative examination of the problem culminating in a concensus, but at
the same time, permitting individual differences of opinion to remain at the
final decision point.
94
-------
The Food Chaiji Mode1_
We envision that the evaluator will begin the procedure by first
examining the target organism and its place in the ecosystem. The first step
of the evaluation procedure is a formidable task in itself. General food
chain relationships are available in almost any biology textbook, but this
particular evaluation scheme dictates that the interrelationships among
organisms within the specific impact or pesticide-use area be examined quite
closely. This is not to say that the environment as a whole should not be
examined, but it is of the utmost importance that the evaluation of the
pesticide-use area be as complete as possible from the onset so as not to
omit other possible environmental ramifications.
Chemicals can be introduced into the environment during manufacturing,
distribution (transport to application site, storage, etc.), application and
disposal of unused portions of the formulation and/or the containers. The
introduction mode will affect the impact on the environment and the point at
which the analysis will begin. Regardless of the point of introduction,
however, the thorough consideration of the initial effects of the pesticide
on target and non-target organisms and the food chain relationships of these
organisms is imperative.
The Model
Gillett, ert aJL (35) have collected a series of conceptual models
illustrating the possible movement of pesticides through the environment.
Figure 11 is a representation of the movement possibilities of a pesticide
in the faunal sector of the environment. Upon entry into the system, a
pesticide's possible movement can be traced through the model, thus affording
information on organisms that may be affected. The impacted segments of the
environment will then be examined in the evaluation matrix. Note that this
faunal model is only a sub-section of the environment and the evaluator will
be charged with the responsibility of examining other facets of the biosphere,
such as the floral, atmospheric and fresh water and esturine/marine aquatic
systems, in the same fashion. Each sub-section of the environment can be
integrated with the other to provide a total summary such as illustrated in
Figure 12.
The conceptual models can be modified to be more explicit for a
particular chemical agent in a specific environment. This systematic
scrutiny of the food chain interrelationships and pesticide movement through
the environment will help to minimize the "surprises" after initial registra-
tion or re-registration of a pesticide. Through the use of these models, the
evaluator will possess a more congruent picture of the physical, biological
and chemical interrelationships between pesticides and the environment.
Closer examination of Figure 12 will afford better comprehension of the
envisioned operation of the model which is divided into three major sections:
atmospheric, terrestrial and aquatic. Each of these categories is further
broken down into its integral parts such as, 1) Atmospheric—mesosphere,
stratosphere and troposphere; 2) Terrestrial—soil divisions, fauna and
flora and 3) Aquatic—marine, fresh water and esturine. More specifically
95
-------
HERBIVORES <
> A
FROM
AQUATIC
Figure 11. General pesticide movement
model in faunal sub-system
[Adapted from Gillett, (35, p. 34)].
96
-------
DEFINITION OF TERMS FOR FIGURE
Herbivore - A plant-eating animal, also considered to be
primary consumer.
Carnivore - Any order (Carmivora) of flesh-eating animals.
Animals preying upon primary consumers of
primary production, e.g., predaceous mites
feeding on phytophagous mites.
Decay - Decomposition of proteins chiefly by bacteria.
Conversion of particulate organic carbon (POC)
to dissolved organic carbon (DOC).
Ingestion - To take in for digestion by mouth.
Inhalation - Process by which animals and plant cells utilize
oxygen, produce carbon dioxide and conserve the
energy of food stuff molecules in biologically
useful terms, such as ATP.
Excretion - Removal of metabolic wastes by an organism.
97
-------
CO
Figure 12, Pesticide irgverwnt in total system
fAdaoted from SiJJett
-------
DEFINITION OF TERMS FOR FIGURE 12
Biota
Primary Producers
Primary Consumers
Decomposer Organisms
Cuticle
Root
Fruiting Bodies
Foliage
Flora
Fauna
Mesosphere
Stratosphere
- The fauna and flora of an area, first order
producers and consumers, second order
consumers, higher consumers and omnivores.
- Green plants that manufacture organic
compounds from simple inorganic substances.
- Organisms that prey on the primary producers.
- Bacteria and fungi which break down the
organic compounds of dead cells from pro-
ducer and consumer organisms into inorganic
substances that can be used as raw materials
by green plants.
- A waxy layer that appears on the exposed
surface of the cell.
- Sub-surface portion of a plant which is
structurally adapted for the functions of
absorption, anchorage or storage.
Structure which produces fruits which
seed-bearing structures derived from
flowers.
are
- A cluster of leaves, flowers and branches.
- The plant life characteristic of a region,
period of special environment.
- The animals or animal life of a region, period
or geological stratum adapted for life in a
specified environment.
- A layer of the atmosphere extending from the
top of the stratosphere to an altitude of
approximately 50 miles.
- An upper portion of the atmosphere which is
above approximately 7 miles, depending on
latitude, season and weather and in which
temperature changes little with changing
altitude and clouds are rare.
99
-------
Troposphere - The portion of the atmosphere which is
below the stratosphere which extends outward
about 7-10 miles above the earth's crust,
and in which generally temperature decreases
rapidly with altitude, clouds form and
convection is active.
Particulates - Minute separate particles.
Omnivores - Organisms feeding on both plant and animal
substances.
100
-------
each of these sub-sections can be expanded to be even more explicit depending
upon the needs of the particular evaluation. The atmospheric section of the
environment, which has already been divided into the mesospheric, stratospheric
and tropospheric zones, can be expanded to illustrate the interaction of gases,
aerosols and suspended particulates in the troposphere which also is affected
by fallout, washout, co-distillation and erosion.
The fauna! sub-section of the terrestrial sector illustrates the inter-
action of herbivores, primary carnivores, higher carnivores and man. The
floral sub-sector shows the interrelationships between plant cuticles, foliage,
roots and fruiting bodies. Notice also the relationships of the soil divisions
with one another, i.e., soil surface, soil sub-surface and soil organisms.
Each of these divisions contain important interactions which can affect the
fate of pesticidal movement and have deleterious effects in the environment.
For example, the soil surface compartment contains components such as surface
water, inorganic particulates, organic particulates, etc. which may affect
the retention, mobility, degradation, and toxicity of a chemical agent. The
same components are identifiable within the sub-surface soil. The entire
soil complex is integrated into the environmental model through erosion (wind,
mechanical, abrasion), decay of matter, imbibing or leakage of water and
inhalation just to illustrate a few interaction mechanisms.
Examination of the aquatic sector once again shows a wide array of
interdependences that will affect the fate of chemicals in the environment.
The marine, fresh water and esturine spheres all have common basic components
that are crucial in tracing pesticide movements and effects, i.e., water,
organic and inorganic particulates and the aquatic biota itself consisting of
primary producers and consumers, secondary consumers, higher consumers and
omnivores.
All of the environmental sectors interface in some manner, showing the
essence of the pesticide problem. That is, the chemical formulation affects
not only the target organism in the target area, but through these interfaces
and food chain relationships, can be transported to other segments of the
biosphere posing serious environmental ramifications. Once a pesticide's
possible movements and effects are established through these types of food
chain movement models, the impacts can then be transferred to the evaluation
matrix for further examination.
DDT Example
It may prove advantageous at this time to trace the movement of the
pesticide through the models to illustrate further the envisioned operation.
Assume that DDT is sprayed to control Dutch elm disease. Barring any
accidental introduction into the environment and not considering disposal of
containers and unused formulation,the point of entry will mainly be in the
floral sector of the environment with some introduction into the atmosphere.
The DDT entering the atmosphere (mainly the troposphere) will be in the
suspended particulate form and can be transported via wind movement to other
locations. The atmosphere is a transport medium and the pesticide (DDT)
will not do much until taken up by an organism even though it remains
biologically active as a vapor or particulate and can react chemically or
101
-------
photochemically. Eventually, the DDT will be deposited on the aquatic and/or
terrestrial segments of the environment with the concentration being diluted
by dispersion. Deposition is usually via dry deposition, sedimentation and
washout by rain. This serves to illustrate the possible movements of the DDT
application from the site to other sectors of the environment in accordance
with Figure 12. Note: However, this is not a statement as to whether the
redeposited DDT is toxic, either acutely or chronically. There are many
factors within the atmospheric sector that will affect the pesticide's move-
ment and concentration which the evaluator must consider (96). This proce-
dure traces only one path of a pesticide's possible movement. Now the
evaluator must return to the point of entry (floral sector) and examine alter-
native routes of displacement.
Recalling that DDT was sprayed on the floral sector to control Dutch elm
disease, the pesticide is now in direct contact with the foliage and cuticle.
The DDT that does not adhere the foliage is deposited on the soil surface in
which the pesticide can be introduced into the surface water and the inorganic
and organic particulate complex. The compound can remain bound in this sector
or can be transported to the soil sub-surface through percolation, to the
aquatic sector through runoff, erosion and sedimentation, to soil organisms
through ingestion and physical contact (absorption) and to the atmosphere via
abrasion or volatilization. Soil organisms (e.g., earthworms) are ingested
by primary carnivores (e.g., robins) in the faunal sector (63) (5). The
primary carnivores are the prey of the higher carnivores leading ultimately
to man. Therein exists the possibility of residue accumulation, biotransfer
and biomagnification.* Dead organisms or excreta from the faunal sector
(through residues) may threaten the aquatic, soil, atmosphere and floral
sectors of the environment. Leaf drop from the floral sector can also add to
the soil residue reservoir.
Examining the aquatic sector more specifically, surface runoff and aerial
transport are the major vehicles for translocating pesticides into the aquatic
ecosystem (70). The contaminant is then made available to the aquatic
organisms in the dissolved or suspended particulate form. Primary producers
(phytoplankton) have shown a 40-90 percent decrease in production due to DDT
exposure (76). This will affect the primary consumers by decreasing the food
supply and/or through the ingestion of residues. The aquatic food chain can
be followed through to ascertain what other organisms may be affected. Con-
taminants from the aquatic sector can reenter the other sectors of the
environment via evaporation and precipitation, ingestion of aquatic organisms
by terrestrial organisms (possibly by man directly or indirectly) and by
plants and animals imbibing contaminated water.
*Residue accumulation is a function of (a) input rate and (b) output rate
(excretion and metabolism). When a > b, residues will accumulate. The
critical point is the metabolism/excretion factor. Therefore, bioaccumula-
tion is a consequence of biotransfer when a_ exceeds b_with biomagnification
occuring when the accumulation in the organism is significantly higher than
concentrations in the air, food, or water (34).
102
-------
We have shown that the tracing of a pesticide's movement through the
environment is a formidable task, but it is also both possible and rewarding.
Using the characteristics inherent to a particular pesticide, application
information (i.e., application rate, frequency area, etc.), knowledge of
existing or probable effects and food chain models such as the ones discussed,
the deleterious impacts of pesticides on the environment can be determined.
It has been illustrated that soil invertebrates can accumulate large residues
of organochlorine insecticides. The same is true for beetles, molluscs and
mites. Birds accumulate these residues when they ingest contaminated food
stuffs and in turn are eaten by vertebrates on a higher tropic level. It is
important to indicate however, that ecologically significant flows of carbon
and energy will follow these same pathways whereas pesticide effects will be
seen in the models where they tend to control (stimulate or inhibit) the flux
along the pathways. Furthermore, each effect must be measured and quantified
in dose-response terms in order to facilitate mathematical explicitness in
model ing.
The Impact Matrix
Using the proposed type of model, the movement of a particular pesticide
through the environment can be suggested and the organisms it may affect
possibly detected. Once this procedure is complete an evaluation of the
impacts of the pesticide on the environment can begin using the impact matrix.
It is recognized that the model system is not absolutely all inclusive;
however, a systematic approach such as this will help to insure that fewer
environmental "surprises" will result.
It is possible that such a rigid procedural matrix may not be the most
effective method of achieving the desired level of completeness, and therefore
each of the categories of effects and impacted organisms will be left general
subject to specific expansion needs. When the expansion of these categories
is complete, the matrix will have a detailed breakdown of effects and organisms
affected with slash marks where the two intersect.
Figure 13 is an example matrix of the general impacts on the environment
sector based on the information set forth in the section discussing the
general impacts of pesticides on the environment. To briefly review the
matrix as it stands before the assessment begins, each effect which has been
found (listed vertically) has a potential impact on the environment (listed
top horizontally). A slash will be placed at the point of intersection to
mark this impact. Within each box a number in the upper left hand corner
(from 1-10) which indicates the relative magnitude of the impact will be
inserted by the evaluator. After this, the evaluator will place a number
(1-10) in the lower right hand corner which indicates the relative importance
of the impact (one represents a low importance or magnitude while ten
represents high magnitude or importance). The term "magnitude" is used in
the sense of degree, extensiveness or scale. The term "importance" is a
relative measure of the significance of the impact and is based on the value
judgments of the evaluator. (55)
To illustrate the procedure, suppose that the pesticide under considera-
tion is linked to reproductive problems in California condors. If the pesti-
cide is to be used in areas where condors rarely venture, it might be
103
-------
§
c
Son-Hunan Health Effects
A. Acute
1. Death Duo to Direct Exposure
2. Contributory to Death
a. Alter ResDunsrs to Stress
1. Chanftr Behavior Patterns
2. Chanse ReMstAncc to Disease
at ore Jnd S.-il Initv
B. Chronic
1. Hut. men
a. ChanRC Genetic Code in Offprint-
3. On.c.OBnr
d. Alter Metabolic P;atc
e. Smaller Thvroid ni.-ind«
f. Increase Post-Partuci Mortal Irv
g. Rg8ldiie Accunn.lnrion
a
^
j
S
5
Z
/
/
^
X
H
i
9
?.
c
c
-
•^
"~
&
<
/
&
i
r
i/i
/
/
^
'=•
e
&
/
/
/
/
_/
cr
/
/
/
/
/
/
/
3
/
/
/
/
/
\
EC
t
=
j
/
/
C
K
/
«
(0
c
/
4i
t,
5
i
i
c
s
2
£
/
/
e
^
r
c.
<
/
X
3
fl
e
u
/
/
/
/
•f
<
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
I
'
/
/
~
/
/
i
;t
/
/
/
/
/
s
i
-
s
/
/
/
/
/
/
/
/
/
£
O.
E
V
ci
d
^
/
/
/
/
/
/
/
/
/
«
£
/
/
/
/
/
/
/
/
/
/
/
/
/
/
f
Figure 13. General impact matrix
-------
possible to assess the magnitude of the impact as fairly low (say 1 or 2).
However, if the importance of the condors' survival is deemed high enough,
the importance might be assessed at a high level (say 9 or 10).
Conversely, if a pesticide is applied to a particular area and a sub-
stantial part of the soil invertebrate population (slugs and snails) is
killed, the magnitude of the impact is high, but the relative importance of
the impact may be deemed low or high, depending on point of view.
The impacts illustrated on the general matrix refer to no pesticide in
particular and the evaluation numbers have been omitted at this time since
no specific chemical agent or dosage is indicated. However, it is assumed
that the inferred application rate is that which would be associated with the
pesticide's particular use. Many of the impacts indicated in Figure 13 cannot
be foreseen without the use of the food chain models developed previously.
Only after an impact has occurred will it be entered into the matrix, or it
may even be omitted if the particular event is not readily attributed to a
particular pesticide's use.
An example may assist in clarifying this line of thought. Returning to
the use of DDT to control Dutch elm disease, it was initially speculated that
spraying operations might have some deleterious effects on birds in the spray
area. Therefore, tests were carried out on animals in the proposed spray
area in an attempt to ascertain the possible consequences. In doing this
robins were sprayed with DDT and the effects were thought to be minimal. But,
when the Dutch elms were sprayed, there was a dramatic decrease in the wild
robin population. It was finally deciphered that the earthworms within the
sprayed area, upon which the robins fed, contained DDT residues that were
eight times greater than those found in earthworms from untreated soil. This
experience serves to illustrate that examination of food chain relationships,
such as those discussed in the previous material, is imperative prior to a
pesticide-use decision. Through the use of such food chain models and environ-
mental check lists, the number of environmental hazards posed by a pesticide-
use decision will be kept to a minimum. It should be noted that the general
pesticide impact matrix presented here is based only on those illustrations
of pesticide effects presented in the text and therefore is by no means
comprehensive. The abscissa and ordinate on the matrix are assumed to be
"open-ended" in the sense that the evaluator is encouraged to enter additional
effects or impacted organisms into the basic matrix as appropriate.
The Delphi Technique
Dal key (16) added iteration and controlled feedback to a system designed
for decision-making using group prediction. In general, this "Delphi"
procedure has three main attributes: a) anonymity, b) controlled feedback,
and c) statistical group response (15). Anonymity serves to suppress the
influence of a dominant member of the response group. This is usually
accomplished through the use of a questionnaire, thus eliminating face-to-face
contact. This feature of Delphi lends itself well to the type of analysis we
haveenvisioned in that where there is a lack of data, experts from various
geographic areas can be called upon to provide input (even if subjective) for
a pesticide-use decision. This is a highly desirable feature since each
105
-------
pesticide has inherent to it certain characteristics that will dictate its
environmental and aesthetic impacts. Since it is unlikely that experts will
be assembled at the decision-making unit for each pesticide-use inquiry, the
Delphi technique is useful in that information input can be collected from a
large geographic area by mail or telephone.
Controlled feedback is a device by which the results of previous assess-
ment rounds are summarized and then communicated to the participants for
further comment. This process will also tend to limit the amount of specu-
lative or uninformed opinions entering the analysis. The evaluators could be
given a basic (such as that in Figure 13) matrix at the onset and be asked to
identify any effect and organism affected that they can and rate the impact
for the particular pesticide under consideration. These results will be com-
piled and summarized and then redistributed to the evaluators for iteration.
This process will continue until the chief investigators deem termination to
be appropriate.
Finally, statistical group response is a way of reducing pressure for
group conformity. This results in the possibility of widespread opinion at
the end of the controlled feedback stage, permitting the representation of
individual opinion in the final response.
In evaluating environmental and aesthetic impacts, the evaluators will
be basing their numerical ratings primarily on value judgments, tending to
obscure the validity of the procedure. However, if the respondents will
accept statistical aggregation of weights supplied by the group, the pro-
cedure is still feasible (15, p. 73).
The Delphi technique is a rapid and efficient means to aggregate the
views of knowledgeable personnel. When Delphi is integrated with the food
chain models and environmental matrices, it will provide a relatively re-
liable representation of pesticide-use decision impacts on the environment.
ENVIRONMENTAL IMPACTS OF ALDRIN APPLIED TO CORN: A CASE STUDY EXAMPLE
The purpose of this section is to provide an example for use of the
model presented in Figure 12 by tracing the movement of aldrin-dieldrin
through the environment. Aldrin (l,2,3,4,10,10-hexachloro-l,4,43,5,8,83-
hexayhdro-exo-l,4-endo-5,8-dimethanonapthalene) is readily broken down into
dieldrin, at lesst 85 percent active ingredient HEOD, (1,2,3,4,10,10-hexa-
ch loro-6,7-expoxy 1-1,4,4a, 5,6,7,8,8a-octa-hydro-eo:o-l,4-endo-5,8-
dimethanonapthalene). It is the latter compound that will be examined more
closely so as to determine its movement through the environment and its
possible effects.
106
-------
The majority of aldrin used is in the agricultural sector on crops such
as corn. As aldrin* is applied to soil and vegetation, residues will undoubt-
edly be present in the floral sector, atmosphere and soil complex shortly
after application. Gannon and Biggar (33) have shown that aldrin is readily
converted to dieldrin in plants and Thompson ejt _al_. (92) have studied the
residue accumulation of pesticides in plant tissues such as peanut hay,
shells and in the peanut itself. Dieldrin residues can be stored in plant
tissue without ill effects providing the residues are not massive. The
environmental ramifications would not be extremely complex if the residues
were to remain in the floral sector or be broken down quickly (non-persistent).
But, as with the majority of chlorinated hydrocarbon pesticides, dieldrin is
highly resistant to biological and chemical modification. Therefore, dieldrin
residues persist in the environment for long periods of -time and are available
for interaction with the remainder of the biosphere via various interface
mechanisms. With this type of information one can begin to trace dieldrin's
movement through the environment starting in the floral sector. Figure 12
indicates that the residue can be transferred directly to herbivores and man
through ingestion and drinking, enter the soil complex (soil surface, soil
sub-surface and soil organisms) via decay, and be translocated to the
aquatic sector by runoff and erosion. The atmosphere acquires dieldrin re-
sidues by various means during the application process, primarily through
particulate or vapor drift. Also, residues in soil can enter the atmosphere
via wind erosion (abrasion). The atmosphere is in essence a transport
mechanism in that there is little or no chemical or biological modification
of pesticides while airborne. However, aldrin is epoxidized into dieldrin
and both are photolyzed to photo-dieldrin (a much more highly toxic substance)
in the air or on surfaces exposed to the UV component of sunlight. The
distance over which a pesticide residue can be transported is dependent upon
the compound's properties in accordance with physio-chemical principles.
Removal of atmospheric residues is accomplished through washout by rain and
dry deposition with the material entering the aquatic, soil and floral
sectors. Therefore, the atmosphere may be seen as a cyclic system which
absorbs, transports and deposits pesticide residues.
Soil Complex
When dieldrin is applied to corn, the soil complex is readily contami-
nated, with the soil surface receiving the greatest quantity of residues
when compared to the soil sub-surface. Surface water containing residues
may be leached (mostly in light sandy soils) to the groundwater and the soil
sub-surface. The organic and inorganic particulates in both the surface and
sub-surface soil complexes can accumulate dieldrin residues, with the
organic matter having a greater affinity for the accumulation of chlorinated
hydrocarbons.
*Aldrin is a soil-incorporated insecticide used to control lepidopterous and
coleopterous larvae emerging from the soil to feed on the plants or feeding
directly on the roots.
107
-------
Within the soil complex, specifically the soil organisms, considerable
concentration of dieldrin residues has been found to occur. Korschgen (52)
found that earthworms (Lumbricus terrestrius) accumulated 4.8 times the
amount of dieldrin found in average soil and had a "storage ratio"* of 5.68.
Korschgen (52) also found, as noted in the section on insects p. 88,
indications of biotransfer and biomagnification of dieldrin in three species
of beetles. Gish (37) found that slugs usually have higher residues than
earthworms from the same site. Thus far this evidence appears to suggest
some food chain transfer from the residues in the soil (either from decaying
plant or animal matter or from direct pesticide application to the soil) to
the organisms inhabiting this part of the ecosystem. A note of caution must
be observed, however. Since earthworms process soil, litter, microbiota and
even minute macroinvertebrates, determination of the route of exposure and
mechanism of accumulation is difficult.
The consequences of such residue transfer and magnification are of
paramount importance to species' survival. Coaker (10) noted that sub-lethal
residues of dieldrin caused extreme hyperactivity in carabid beetles
(Bembidion lampros). In conjunction with the alteration of individual
activity patterns, sub-lethal residue accumulation can alter the predator-
prey relationship. For example, after application of insecticides, the
number of cabbage root flies increased, due to the demise of the root fly's
natural enemies and the resultant lack of downward pressure on the cabbage
root fly population. Sub-lethal residue accumulation can also result in the
alteration of behavior patterns. For example, the prey (earthworm) may spend
more time in an exposed position and may become more noticeable to the
predator (birds). This results not only in the demise of large numbers of
prey organisms but also in the consumption of greater numbers of contaminated
prey organisms by the predator. This may allow accumulation of residues to
a lethal dose in the predator or allow the higher concentration of residues
to be passed on to a higher trophic level (49) (89).
Ingestion and consequent biotransfer represents only one mode of residue
translocation from the soil complex. Runoff and erosion transport residues
from the soil complex to the aquatic sector of the biosphere. Vegetation can
also imbibe pesticide residues contained in the groundwater and wind erosion
introduces residues into the atmosphere. This segment of the cycle is com-
pleted when residues return to the soil complex through decaying contaminated
animal and vegetable matter, excreta deposition and rainfall and dry
deposition.
Higher Terrestrial Trophic Levels
In the case of birds, Scott et a]_. (85) found that dieldrin applied to
control insect pests in Illinois caused heavy mortalities in birds and
mammals. Dustman and Stickel (24) noted that dieldrin used to control white-
fringed beetles in Tennessee resulted in numerous bird and mammal deaths.
Therefore, it is safe to surmise that dieldrin is deleterious to some species
'Storage Ratio - "
ppm residues in soil
108
-------
of birds and mammaln. Birds that are carnivores (eating other birds and
fish) usually have higher residues of dieldrin than those birds ingesting
primarily seeds and vegetation, once again suggesting biotransfer of dieldrin
residues from vegetation or soil to higher trophic levels. For more detailed
analysis of residues in birds see Keith and firuchy (51); Vermeer and Reynolds
(94); Prestt, et_ aj_. (75). Studies done by Cummings e_t_ aj_. (13) (14) showed
that dieldrin was readily stored in avian tissues, and Bailey et al. (2) and
Friend (31) showed that dieldrin accumulated in the brain tissue of pigeons
(Columba Hvia) and mallard ducks (Anas platyrhynchos).
First generation pheasants containing dieldrin had no abnormal
physiological symptoms but second generation pheasants had significantly
lower hatchability and fertility than controls (3). The chicks also had
visual impairments. Shag (Phalarocorax aristotelis) eggs contained 0.5-3.0
ppm of dieldrin suggesting the possible relationship between clutch viability
and dieldrin residues since six of the eight clutches had no surviving chicks
at the end of the tenth day (74). Dieldrin in sheep dip and residues in the
environment seem to have caused poor reproduction in golden eagles in
Scotland (58).
On still a higher level, large residues of dieldrin were found in the
tissues of dead mammals in areas of the southeast U.S. treated with dieldrin
for fire ant control. The affected organisms included cotton rats, white-
footed mice (Peromyscus sp.), cottontail rabbits (Sylvilagus floridanus),
raccoons (Prooyon lotor), armadillos (Dasypus novemcinctus) and red foxes
(Vulpes fulva) (21). Similar findings occurred in Illinois when dieldrin
was used to curb a Japanese beetle (Pop-ilia japonica) infestation. Ground
squirrels (Citellus trideoemlineatus and C. franklinii), muskrats (Ondatra
zibethica), cottontail rabbits, short-tailed shrews (Blarnina brevicauda),
fox squirrels (Sciurus niger), woodchucks (Marmota monaxj and prairie voles
(Miarotus ochrogaster) suffered heavy mortalities (85). Blackmore (4)
reported increased mortality in British foxes (Vulpes vulpes) and it was
suggested that the foxes had died from eating dead birds that had eaten grain
treated with dieldrin. This incident serves to illustrate the importance of
examining food chain relationships prior to pesticide-use decision in an
attempt to forego serious environmental reactions. In this case dieldrin
was transferred, in the form of residues, from the original application site
(floral sector) up the food chain to birds and then to carnivores via
ingestion. It is important to point out that movement of residues within
the environment does not stop with the death of the foxes. Scavengers (e.g.,
omnivores) can ingest the contaminated carcasses, the residues can be
returned to the soil reservoir via decay and the chemical residues can be
transported to the acquatic sector through runoff, erosion, sedimentation
and leaching from the terrestrial ecosystem.
The white-fringed beetle, fire ant, Japanese beetle and treated grain
examples do not specifically refer to aldrin application on corn, but these
incidents are germane to the evaluation. That is, since aldrin is employed
quite extensively on corn, it is relatively safe to suspect that organisms in
the application areas may be impacted similarly. It is important to note,
however, that susceptibility to pesticides can vary greatly within organisms
of the same species.
109
-------
Residue accumulation may not only cause death directly, but may also
produce chronic physiologic malfunctions. In a three-year study on the effect
of dieldrin on the physiology and reproduction of white-tailed deer
(Odocoileus virginianus), Murphy and Korschgen (68) found that fawns were
smaller at birth, there were greater post-partum mortality, the fawns had
smaller thyroid and pituitary glands, and there was placental transfer of
dieldrin.
Aquatic Sector
When aldrin is applied to corn, introduction of the pesticide contaminant
into the aquatic ecosystem is primarily achieved through the surface runoff,
aerial transport and direct application of pesticide to aquatic systems.
The aquatic sector possesses massive sorptive capacity and acts as a
pollutant sink yielding a continual source of residues (90). Once the
residues are in the aquatic sector, organisms can acquire contaminants by
direct uptake of water-borne residues or indirectly through the ingestion of
contaminated food stuffs, sediment and detrial material.
The uptake of dieldrin from food and water by fish and fish-food
organisms has been studied by many including, Gakstatter and Weiss (32),
Reinert (79), Hansen (40), and Hoi den (43). Some species can accumulate
residue concentrations that are several thousand times greater than those in
the water.
Petrocelli and Anderson (71) designed a study to test the bio-
magnification of dieldrin residues in the food chain of clams and blue crabs.
Analysis of the data found that animals fed small amounts of dieldrin at
relatively low rates over short periods of time accumulated dieldrin up to
6.8 times the daily exposure. For actual experimental design and results see
Petrocelli and Anderson (71). These results imply that under natural
conditions, animals consuming contaminated food for months or years can
accumulate high residue concentrations.
In another research effort by Petrocelli, Hanks and Anderson (72), the
uptake and accumulation of dieldrin by esturine molluscs, Rangia ouneata,
was examined. It was found that the mollusc accumulated dieldrin residues
from extremely low levels in the water.
It is imperative that the fate of chemical agents in the ecosystem be
established prior to a pesticide's wide-scale use. Metcalf £t ^]_. (65)
developed a model ecosystem in which the movement of a pesticide and its
metabolites could be examined. Using the Metcalf model, Sanborn and Yu (83)
added one crab (Uoa manilensis) and a water plant, Elodea, to the original
model and traced the movement of dieldrin. The fish at the top of the food
chain in the model concentrated dieldrin approximately 6,000 times the
concentration in the water. Physa snails accumulated 115,000 times the
concentration in the water (about nineteen times the value found in fish).
It was found that dieldrin is highly resistant to chemical or biological
modification which, when related to the real environmental case, indicates
that dieldrin residues from any given application will remain in the
environment for some time to come.
110
-------
Hannon (39) observed that fish at high tropic levels have large concen-
trations of residues, one of which is dieldrin. Exposure to these sub-
lethal concentrations and the subsequent accumulation of chemicals by fish
can result in physiologic malfunctions having far-reaching environmental
ramifications. Research done by Burdick e^t aj_. (6) indicates that exposure
to sub-lethal concentrations of dieldrin may result in higher mortality
through increased reproductive failures (gamete residues) and greater
susceptibility to predation due to alteration of fry mobility and activity
patterns. In addition, dieldrin exposure reduced both growth and fecundity
in Poecilia latpinna (54). Cain (8) found that exposure to dieldrin reduced
the resistance to disease and produced below normal feeding rates. Johnson
(50) observed thickening of gill membranes, lack of osmoregulation, lower
blood counts, brain damage, and reduced body weight in fish exposed to
dieldrin.
A suitable example of the movement of dieldrin in the marine food chain
has been done by Robinson and is illustrated in Figure 14. This food chain
model shows the biotransfer and biomagnification potential of dieldrin
within this section of the ecosystem. It is by a composite of models such as
shown in Figure 14 that an evaluator may identify a pesticide's possible
movement through the environment.
Summary
Since dieldrin is resistant to chemical and biological modification, it
can and does persist in all facets of the environment. This persistence
affords the opportunity for biotransfer and biomagnification of residues in
the biosphere as exemplified in the test. Food chain models such as
Figure 12 will afford greater comprehensiveness of analysis and consequently
facilitate a comprehensive view of environmental interdependences and
pesticide impacts.
With the application of aldrin to corn, there exists a myriad of
possibilities for environmental contamination. The aforementioned effects of
aldrin aid in the illustration of the flow of a pesticide through the food
chain model, but the models are operational even if little or nothing is
known about actual in vivo consequences.* Based on laboratory experimentation,
the inherent properties of the pesticide in question, and knowledge of food
chain relationships, the effects of a pesticide can be foreseen, to some
extent facilitating the minimization of ecological "surprises."
Once the potential impact areas are identified, there still remains the
question of how to evaluate these environmental impacts and integrate them
into a benefit-cost framework. In tracing the movement and subsequent
environmental effects of aldrin applied to corn, the impacted organisms are
entered into the environmental impact matrix where an attempt at quantifi-
cation can commence.
*Unfortunately, little is known about environmentally unstable compounds that
do not leave residues but, which, by virtue of repeated applications or some
sustained release mechanism, provide iterative or chronic exposure.
Ill
-------
Seaweed
0.0001 ppm
Microzooplankton and Crustacea
0.002 ppm
Crustacea and Fish
0.03 ppm
Cormorant Eqgs
1.2 ppm
Birds
Shags
Kittiwakes
Terns
Eggs
1.2 ppm
0.1 ppm
0.2 ppm
Figure 14. Concentration of dieldrin in marine organisms [Robinson, (82)],
112
-------
ALDRIN-DIELDRIN IMPACT MATRIX
Matrix Operation
Using the environmental impacts delineated earlier in this section an
impact matrix such as that shown in Figure 15 can be developed. The abscissa
consists of those species (both flora and fauna) which are impacted and the
ordinate is comprised of aldrin-dieldrin induced effects. A slash is placed
at the intersection of the impacted organism and the effect. This process is
continued until the evaluator has entered all the impacts that he is aware of
based on his knowledge of the relevant literature, consultation with
colleagues, etc. Evaluation of the impacts can then commence. A number
(ranging from a high of ten to a low of one) in the upper left-hand corner
indicates the magnitude or extent of the impact on the organisms under
consideration while the number (high of ten to low of one) in the lower
right-hand corner is a relative imoortance rating of the effect. The im-
portance rating is based on the organism's place and function in the food
chain as well as its intrinsic or market value to man.
For example, earthworms readily accumulate dieldrin residues which are
retained in the organism affording greater possibilities for food chain
transfer of residues. 11; is important to note that uotake of residues by
organisms does not imply storage and accumulation. But, in the case of the
earthworm, the untake and accumulation warrants a magnitude rating of five.*
Since the dieldrin resource can alter behavior patterns making the earthworm
more noticeable and therefore easier prey, the magnitude and importance of
behavior alterations is assessed at four and eight, respectively. Earthworms
are a food source for many birds and some mammals which also tend to accumu-
late dieldrin residues quite readily thus reinforcing the residue accumulation
importance rating of nine for earthworms. It should be noted that as higher
trophic levels are reached, the importance rating shifts from strictly food
chain criteria to one based on man's valuation of the impacted organisms.
This is illustrated with the case of the golden eagle. It is at the top of
its food chain and has a low importance rating in that sense but is assumed
to be highly valued by man since it is protected by law. From this per-
spective, the importance of the impact is high, assessed at ten.
The significance of food chain relationships to impact assessment can
readily be seen since the importance rating of residue accumulation in a
species is a function of its level and importance in the food web hierarchy.
In tracing the impact assessment of an organism through the matrix, keeping
in mind the underlying food chain interdependencies, a comprehensive picture
of pesticide impacts can be developed and then evaluated. This concept is
*A11 numerical ratings in this section were developed by the author for
illustrative purposes only and do not necessarilv reflect the judgment of
experts in biological effects of chemical pesticides as would be the case
in an actual assessment for registration decision purposes.
113
-------
Nun-Hun.in Health Effects
A. Acute
.'. C. urlbutorv to Death
... Alter K,.spons,- to Str«ii«
1<- Cjh.inRo Behavior Patterns
h. Chr.,nlr
J- H..I.IXCI
h. ChJiiRi Susceptibility to Post 1 . ( cU-S
c. Alter SDL-clos
1. Irra'ORcn
j. A.bnortftjl TestiruUr I'rvflopni-nt
3. OncoKvn
.». C.ir. inusu of Orpins
i . Nt-priidurl 1 vo Ittipo 1 rncnt
b. hfducc t" J«tch Viability
. . l.owt-r f.-rt Illtv
d. Dccfi-.isc Ability lu ! cfit
c. Canute T. tlurc
.1. Tl.lckvnln*! of GUI Hofthr.inv
1.. tack o( Osmorcgolatlon
. . l.owt-r Blocil Counts
.1. Hr.iin D.IMR*.
.. Hcduft- Body t.YlRlil
f. Louor Fffdlns Habits
K. Altt-r I Ivor Kunrt Inn
h. -\li.-r HetAhoIlr K.itc
k. AHtt EtC PaitvTAS
1. Increase Hw^rt Hntr
n>. UffxprlnR So.i))<-r .it Rtrth
o. ftisiduo Accumulation
i
\
u
/
/
>
•=
-
!
c
...
r
^.
a
-
/.
-
X
o
X
5
•o
^
^
:
Y*
i
/I
•
\
I
c
u.
I
<
I
£
'
4
£
J/J"
^
>
|
-
X
b
c
/
E
^
"X
V
%
/,
^
IT
f.
(
%
"A
i
js
^
••/.
Figure 15. Aldrin-dieldrin impact matrix.
114
-------
-t-
TH
.-J--
32
Zt
{'A
Figure 15 (continued)
115
-------
important since a pesticide may strongly impact a particular segment of the
environment but the importance of this effect on the entire ecological scheme
may be small.
An example will help to clarify this point. Snails and slugs are
adversely impacted by dieldrin and accumulate residues to approximately the
same extent as earthworms. However, the accumulation of residues in snails
and slugs is not as important as earthworms since snails and slugs are not as
readily preyed unon as earthworms. This is most likely due to the slimy
mucus coating of the soil molluscs to which potential nredators display
aversion. Without predation from higher trophic levels, the residue-laden
snails and slugs do not present a serious environmental threat by way of food
chain transfer. Therefore, a residue-accumulation importance rating of five
seems appropriate.
It may prove advantageous to trace the impact assessment through the
matrix to nrovide a holistic view of the integrated food chain model and
assessment matrix in operation. When aldrin is supplied to corn for soil
insect control, residues (dieldrin) become present in the soil and are
available to non-target organisms both directly and indirectly. Soil
organisms (e.g., earthworms) can accumulate the dieldrin residue by ingestion
and absorbtion yielding an impact rating of four and eight (magnitude and
importance, respectively). Since dieldrin residues can alter behavior in the
earthworm, and it is a food source for many carnivores (both birds and
mammalian insectivores), the residue impact is assessed at five and nine.
Small mammals, such as shrews, voles, etc., accumulate residues from ingesting
contaminated food stuffs (earthworm) and in turn can be killed directly due
to dieldrin toxicity or can accumulate residues themselves. Once again, the
importance of residue accumulation in these organisms is embodied in the
animals' food web integration. Since these small mammals are consumed by
predators (both birds and mammals), residue accumulation warrants an
importance assessment of eight. If the golden eagle ingests these contami-
nated organisms, the magnitude of the impact (reproductive failure) may be
high due to a several-fold increase in the proportion of birds that do not re-
produce successfully. The importance rating (10) is also quite high due in
part to species scarcity and man's valuation of this somewhat rare bird. The
significance of the golden eagle as a food source is minimized by virtue of
its standing near the top of the predator-prey hierarchy.
The importance rating on organisms that exhibit acute toxicity to aldrin
and dieldrin will depend on'manly on species abundance and specialized food
chain relationships. Therefore, the importance rating of an organism's
death may substantially differ from the importance evaluation of that
organism's residue accumulation potential. Critical to the importance
rating is the definition of perspective. The majority of pesticide expertise
is centered around protection/production with this particular illustrative
effort taking the latter perspective in that protection is applied to bio-
logical systems. The inherent bias in these stances could possibly be counter
balanced by the selection of experts in the field employed by non-production
groups.
116
-------
reason
Clearly, the food chain relationship presented is only one of many
possibilities which could be presented, but the example suffices to illus-
trate the basic rationale underlying the impact assessment scheme. p/!any of
the impact ratings in Figure 15 have not been discussed, but the assessment
of effects follows the same strategy set forth in the example.
Figure 15 represents the efforts of only one evaluator and it is re-
cognized that there may be numerous other entries possible. For this rei
the Delphi technique and an "open-ended" matrix are proposed to insure that
the evaluation of possible aldrin-dieldrin induced impacts is comprehensive.
The aggregation of many responses toamatrix such as Figure 35 will yield a
composite impact matrix which decision-makers may use in evaluating a par-
ticular pesticide. The output from the environmental assessment, when inte-
grated with the information obtained from production, human health, aesthetic,
and distribution analysis, will afford the decision-maker vital data for
benefit-cost analysis of pesticide-use decisions.
Summary
The majority of aldrin (which degrades to dieldrin) is used by the
agricultural sector on corn. The persistence of aldrin is agronomically
desirable, but this characteristic also has serious environmental ramifi-
cations. The long life of aldrin permits translocation from the application
site to other sectors of the environment. This chemical stability also
facilitates the contamination of non-target organisms, possibly man himself.
The examples presented in Section 15 indicate the types of effects that
aldrin-dieldrin induce in the environment. This ex ante view of environmental
impacts, however, is not the most desirable method by which to evaluate
aldrin-dieldrin effects since impacts can be of an irreversible nature. The
systematic examination and evaluation of impacts facilitated by the food
chain models and measurement techniques developed in this research effort
conclude that the environmental impacts of aldrin-dieldrin are extensive.
And, since this is in essence only a partial analysis, it is reasonable to
suggest that the ecological consequences induced by this particular pesticide
could be even more widespread.
Man only becomes aware of environmental alteration at a relatively high
trophic level. The acute toxicity of aldrin is of great biotic importance,
but is subordinate to the chronic effects of dieldrin. Through residue
accumulation and biomagnification, dieldrin moves from the application point
and low trophic levels to non-target area and higher trophic levels, possibly
to man himself. This environmental evaluation scheme begins with a micro-
system stance in an attempt to predict macrosystem consequences. The material
presented illustrates that the ecological effects of aldrin-dieldrin are
extensive in both a micro and macro system framework while, at the same time,
emphasizing environmental interdependences. Residue accumulation and bio-
magnification of dieldrin, as related to man's perception of ecological
alteration, is by far the most important effect.
As previously discussed, incorporation of environmental data into a
benefit-cost or other analytical frame works is imperative. The most de-
sirable state, when using benefit-cost analysis, would be to reduce
117
-------
environmental impacts to monetary terms making the units commensurate with
other variables in the analysis. Unfortunately, this conversion can only be
accomplished in part and under highly specialized analytical conditions. The
expenditures approaches used by hunting and fishing agencies are not adequate
vehicles for environmental cost estimation in that the schemes possess no
capability for species specificity and yield defective measures of aggregate
wildlife value. If environmental parameters cannot be totally and correctly
reduced to monetary terms, it would be inadvisable to incorporate these
values into a benefit-cost structure. Rather than using conceptually
erroneous measures of environmental impacts, we recommend the evaluation
scheme set forth in this section and comparison of environmental consequences
among alternative pesticide-use decisions.
118
-------
BIBLIOGRAPHY FOR ENVIRONMENTAL IMPACTS CHAPTER
1. Appert, J. Docum. Inst. Rech. Agron. Madagascar, 113, 1967.
2. Bailey, S., J. Bunyan, B. D. Rennison and A. Taylor. The Metabolism
of 1,1-Di (p-chlorophenyl)-2,2-dichloroethylene and 1,1-Di
(p-chlorophenyl)-2-chloroethylene in the Pigeon. Toxicol.
Appl. Pharmacol., 14: 23-32, 1969.
3. Baxter, W. L., R. L. Liner and R. B. Dahlgren. Dieldrin Effects on
Two Generations of Penned Hen Pheasants. J. Wildl. Mgrot.,
33(1): 96-103, 1969.
4. Blackmore, D. K. The Toxicity of Some Chlorinated Hydrocarbon
Insecticides to British Wild Foxes (Vulpes vulpes). J. Comp.
Pathol. and Therap., 73: 391-409, 1963.
5. Boykins, E. A. DDT Residues in the Food Chain's of Birds.
Atlantic Nat., 21: 18-25, 1966.
6. Burdick, G. E., E. J. Harris, H. J. Dean, T. M. Walker, J. Shea and
D. Colby. The Accumulation of DDT in Lake Trout and the Effect
on Reproduction. Trans. Am. Fish. Soc., 93: 127, 1964.
7. Butler, L. I., B. J. Landis and L. M. McDonough. Circular No. 522,
Washington Agr. Exp. State., College of Agr., Washington State
University, 1970.
8. Cain, S. A. Pesticides in the Environment with Special Attention to
Aquatic Biology Resources. Rept. U.S. Japan Meeting on Pesticide
Research, 1965.
9. Cicchetti, Charles J. and Myrick Freeman. Option Demand and Consumer
Surplus: Further Comment. Q.J.E., 85(3): 538, August 1971.
10. Coaker, T. H. The Effects of Soil Insects on the Predators and
Parasites of the Cabbage Root Fly (Erioisohia brassicae (Bouche))
and on the Subsequent Damage Caused by the Pest. Ann. Appl.
Biol., 57: 397-407, 1966.
11. Collingwood, C. A. Persistant Pesticides - A Look At Future Needs.
Ann. Appl. Biol., 55: 335-337, 1965.
12. Conney, A. H., R. M. Welsh, R. Kuntzman and J. J. Burns. Clin.
Pharmacol. Ther., 8: 2, 1967.
119
-------
13. Cummings, J. G., M. Eidelman, V. Turner, D. Reed, K. T. Zee and
R. E. Cock. Residues in Poultry Tissues from Low Level Feeding
of Five Chlorinated Hydrocarbon Insecticides to Hens. J. Assn.
Offic. Anal. Chem., 50: 418, 1967.
14. Cummings, J. G., K. T. Zee, V. Turner and F. Quinn. J. Ass. Offic.
Anal. Chem., 49: 354, 1966.
15. Dalkey, N. C. The Delphi Method: An Experimental Study of Group
Opinion. RAND Corp., RM-5888-OR, 1969.
16. Dalkey, N. C. Delphi, the RAND Corp., P-3704, October, 1967.
17. Dalkey, Norman C. and Daniel L. Rourke. The Delphi Procedure and
Rating Quality of Life Factors. In The Quality of Life Concept:
A Potential New Tool for Decision Makers. EPA, Office of
Research and Monitoring, Environment Studies Division, 1973.
18. Davis, B. N. K. Soil Animals as Vectors of Organochlorine Insecticides
for Ground-Feeding Birds. J. Appl. Ecol., 3: 133-139, 1966.
19. Dee, Norbert, et al. Environmental Evaluation System for Water
Resource Planning. Report to the U.S. Bureau of Reclamation,
Columbus, Ohio. Battelle Memorial Institute, Jan., 1972.
20. Dee, Norbert, et a!_. Planning Methodology for Water Quality Management:
Environment Evaluation System. Columbus, Ohio, July, 1973.
21. Dewitt, J. B., C. M. Menzie, V. A. Adomaitis and W. L. Reichel.
Pesticidal Residues in Animal Tissues. Trans. 25th N. Amer.
Wildl. Conf. 227, 1960.
22. Dimond, J. B. and J. A. Sherburne. Persistence of DDT in Wild
Populations of Small Mammals. Nature (London), 22(5179)
486-487, 1969.
23. Dorough, H. W. and N. M. Randolph. Comparative Residual Nature of
Certain Insecticides Applied as Low Volume Concentrate and Water
Emulsion Sprays. Bull. Environ. Contam. and Toxicol., 2(6):
340-348, 1967.
24. Dustman, E. H. and L. P. Stickel. Pesticides and Their Effect on Soil
and Water. Amer. Soc. Agron. Spec. Publ. No. 8. Soil Sci. Soc.
Amer., 1966, pp. 109-121.
25. Edwards, C. A., ed. Environmental Pollution by Pesticides. Plenum
Press, London and New York, 1973.
26. Edwards, C. A. Persistant Pesticides in the Environment. CRC Press,
Cleveland, Ohio, 1970.
120
-------
27. Edwards, C. A. Critical Reviews in Environmental Control. Vol. 1,
Issue 1, Chem. Rubber Company, 1970, pp. 7-76.
28. Edwards, C. A., J. R. Lofty, A. E. Whiting and K. A. Jeffs. Rep.
Rothanest. Exp. Sta. for 1970, Vol. 193, 1970.
29. Ferguson, D. E. Trans. 32nd N. Amer. Wildl. and Natural Resources
Conf., 1967, pp. 103-107.
30. Ferguson, D. E. In Reservoir Fisheries Resources Symp. Reservoir
Committee on the So. Div. Amer. Fish Soc., University of
Georgia, Athens, 1968, pp. 531-536.
31. Friend, M. Pesticide-Infectious Disease Interaction Studies in the
Mallard. Ph.D. Thesis, University of Wisconsin, 1971.
32. Gakstatter, J. H. and C. M. Weiss. The Elimination of DDT-C14,
Dieldrin-Cl^, and Lindane-C^ from Fish Following A Single
Sublethal Exposure in Aquaria. Trans. Am. Fish. Soc., 96(1):
301-307, 1967.
33. Gannon, N. and J. H. Bigger. The Conversion of Adrin and Heptachler to
Their Epoxides in Soil. J. Econ. Entomol., 51(1): 8-11, 1958.
34. Gillett, James W. Personal communication, 1976.
35. Gillett, James W,, James Hill IV., Alfred W. Jarvinen, and W. Peter
Schoor. A Conceptual Model for the Movement of Pesticides
Through the Environment. National Ecological Research Lab., Gulf
Breeze Environmental Research Lab., National Water Quality Lab.,
Southeast Environmental Research Lab., National Environmental
Research Center, Corvallis, Oregon, Dec. 1974.
36. Gillet, James W., ed. The Biological Impact of Pesticides in the
Environment. Proceedings of the Symposium, August 18-20, 1969,
Oregon State University, Corvallis, Oregon, 1970.
37. Gish, C. D., Pestic. Monit. J., 3(241), 1970.
38. Goldberg, E. D., P. Butler, P. Meier, D. Menzel, C. Paulik, R.
Risebrough and L. F. Stickel. National Academy of Sciences,
Washington, D.C., 1971.
39. Hannon, M. R. M. S. Thesis, South Dakota State Univ., 1966.
40. Hansen, D. J. An Annual Report of the Bureau of Commercial Fisheries
Biological Lab. Gulf Breeze, Fla., for the Fiscal Year Ending
June 30, 1965. Bureau of Commercial Fisheries Circ. No. 247,
10, 1966.
41. Harris, C. R. Proc. Entomol. Soc. Ontario, 100: 14, 1969.
121
-------
42. Herman, S. G., R. L. Garret and R. L. Rudd. Chemical Fallout. C. C.
Thomas, Springfield, 111., 1969, pp. 24-51.
43. Holden, A. V. Organochlorine Insecticide Residues in Salmonid Fish.
J. Appl. Ecol., 3:45-53, 1966.
44. Howe, C. W. Benefit-Cost Analysis for Water System Planning. Water
Resource Monograph 2, American Geophysical Union, 1971.
45. Hunt, L. B. Kinetics of Pesticide Poisoning in Dutch Elm Disease
Control - The Effects of Pesticides on Fish and Wildlife.
U.S.D.I. Fish and Wildl. Serv. Circ. 226. Washington, D.C.
12, 1965.
46. Hunt, L. B. and R. J. Sacho. Response of Robins to DDT and
Methoxychlor. J. Wildl. Mgmt., 33(2): 336-345, 1969.
47. Ingram, W. R. Cott. Grow. Rev., 44: 203, 1967.
48. Jefferies, D. J. Meded. Fakult, Landboux. Gent., 36: 34, 1971.
49. Jefferies, D. J. and C. H. Walker. Uptake of pp'-DDT and Its Post-
mortem Breakdown in the Avian Liver. Nature (London),
212: 533-534, 1966.
50. Johnson, D. W. Pesticides and Fishes - A Review of Selected
Literature. Trans. Amer. Fish. Soc., 97: 398, 1968.
51. Keith, J. A. and I. M. Gruchy. Proc. 15th Int. Ornith, Congress, 1970.
52. Korchgen. L. J. Soil - Food Chain - Pesticide Wildlife Relationships
in Aldrin-Treated Fields. J. Wildl. Mgmt., 34(1): 186-199, 1970.
53. Kulash, W. M. Soil Treatment for Wireworms and Cutworms. J. Econ.
Entomol., 40(6): 851-854, 1947.
54. Lane, C. E. and R. J. Livingston. Some Acute and Chronic Effects of
Dieldrin on Sailfin Molly, Poecilia latipinna. Trans. Am. Fish.
Soc., 99(3): 489-495, 1970.
55. Lazarevic, B. M. Effects of the Number of Sprayings and Different
Chemicals in Controlling Cotton Pests in the Sudan. J. Econ.
Entomol., 63(2): 629-633, 1970.
56. Leopold, L. B., et. al. A Procedure for Evaluating Environmental
Impact. Geological Survey Circular 645, U.S. Geological Survey,
Washington, D.C., 1971.
122
-------
57. Little, A. D., Inc. Transportation and Environment: Synthesis for
Action: Impact of National Environmental Policy Act of 1969 on the
Department of Transportation, Vol. 3, Options for Environmental
Management, prepared for Office of the Secretary, Department of
Transportation, July, 1971.
58. Lockie, J. D., D. A. Ratcliffe and R. Balharry. Breeding Success and
Organochlorine Residues in Golden Eagles in West Scotland. J.
Appl. Ecol., 6: 381-389, 1969.
59. Lockie, L. N. and D. A. Ratcliffe. Insecticide and Scottish Golden
Eagles. Brit. Birds, 57: 89-102, 1965.
60. Macek, 0., Environmental Health Sciences Series No. 1, Oregon State
University, Corvallis, Oregon, 1970.
61. Materu, M. E. A. and D. Hopkins. E. Afr. Agric. For J., 35(78), 1969.
62. McHarg, Ian. A Comprehensive Highway Route-Selection Method. Highway
Research Record, Number 246, 1968, pp. 1015, and McHarg, Design
With Nature, Natural History Press, Garden City, New York: 1969,
pp. 31-41.
63. Mehner, J. F. and G. J. Wallace. Atlant. Nat., 14: 4, 1959.
64. Metcalf, R. L. Proc. 2nd Intemat. Plant. Prot. Conf., Vol. 129, 1957.
65. Metcalf, R. L., I. P. Kapoor and S. K. Gurchavan. Model Ecosystem for
the Evaluation of Pesticide Biodegradability and Ecological
Magnification. Environ. Sci. Tech., 5(8): 709-713, 1971.
66. Moriarty, F. The Toxicity and Sublethal Effects of p, p'-DDT and
Dieldrin to Aglais urtioae (L.) (Lepidoptera: Nymphalidae) and
Chorthippus brwnes (Thunberg) (Saltatoria: Acrididae). Ann.
Appl. Biol., 62: 371-393, 1968.
67. Murphy, P. G. The Effect of Size on the Uptake of DDT from Water by
Fish. Bull. Environ. Contain, and Toxicol., 6(1): 20-23, 1971.
68. Murphy, D. A. and L. J. Korschgen. Reproduction, Growth, and Tissue
Residue of Deer Fed Dieldrin. J. Wild!. Mgmt., 34(4): 887-903,
1970.
69. Ogilvie, D. M. and J. M. Anderson. J. Fish. Res. Bd. Canada, 22: 503,
1965.
70. Peterle, T. J. In The Biological Impact of Pesticides in the
Environment. J. W. Gillett, ed., Environmental Health Sciences
Center, Oregon State University, Corvallis, Oregon, 1970.
123
-------
71. Petrocelli, S. R. and J. Anderson. Biomagnification of Dieldrin
Residues by Food-Chain Transfer from Clams to Blue Crabs Under
Controlled Conditions. Bull. Environ. Contam. and Toxicol.,
13(1): 108-122, 1975.
72. Petrocelli, S. R., A. R. Hanks and J. Anderson. Uptake and Accumulation
of an Organochlorine Insecticide (Dieldrin) by an Esturine
Mollusc, Rangia owieata. Bull. Environ. Contam. and Toxicol.,
10(5): 315-320, 1973.
73. Pill, J. The Delphi Method: Substance, Context, A Critique and
Annotated Bibliography. Socio-Econ. Plan. Sci., 5(1), 1969.
74. Potts, G. R. Success of Eggs of the Shag on the Fame Islands,
Northumberland, in Relation to Their Content of Dieldrin-and
pp'-DDE. Nature (London), 217: 1281-1284, 1968.
75. Prestt, I., D. J. Jefferies and N. W. Moore. Polychlorinated Biphenyls
in British Wild Birds. Env. Poll., 1(1): 3-27, 1970.
76. Pritchard, H. N. and D. M. Dines. DDT Effects on Oxygen Evolution
During Photosynthesis. Proc. Pa. Acad. Sci., Vol. 46, 1972,
pp. 25-26.
77. Ratcliffe, D. A. Changes Attributable to Pesticides in Egg
Breakage Frequency and Eggshell Thickness in Some British Birds.
J. Appl. Ecol., 7: 67-107, 1970.
78. Reed, W. Use of Insecticides on Cotton in Western Tanzania. Agric.
Vet. Chem., 8(1): 8-10, 1967.
79. Reinert, R. E. The Accumulation of Dieldrin in an Alga (Scenedesmus
obliquus), Daphnia (Daphnia magna)t Guppy (Lebistes reticulatus)
Food Chain. Ph.D. Thesis, Univ. of Michigan, Ann Arbor,
Michigan, 1967, pp. 76.
80. Ridgeway, R. L., D. A. Lindquist and D. L. Bull. Effect of Method of
Application on Uptake of Di-Syston by the Cotton Plant. J. Econ.
Entomol., 58(2): 349-352, 1965.
81. Robertson, I. A. D. E. Afr. Agric. For. Ind. J., 35: 181, 1969.
82. Robinson, J. Organochlorine Insecticides and Birds. Chem. Brit.,
4: 4, 1968.
83. Sanborn, J. R. and C. C. Yu. The Fate of Dieldrin in a Model
Ecosystem. Bull. Environ. Contam. and Toxicol., 10(6): 340-346,
1973.
84. Satchell, J. E., A. Burges and F. Raw, ed. Soil Biology, Academic Press,
London and New York, 1967.
124
-------
85. Scott, T. G., Y. L. Willis and J. A. Ellis. Some Effects of A Field
Application of Dieldrin on Wildlife. J. Wildl. Magmt., 23(4):
409-427, 1959.
86. Skutch, M. and J. L. Schofer. Goals - Delphis for Urban Planning:
Concept in Their Design. Socio-Econ. Plan. Sci., 7(3), 1973.
87. Smith, W. I. Quantifying the Environmental Impact of Transportation
Systems. Van Doren-Hazard-Stallings-Schnacks, Topeka, Kansas,
undated.
88. Stickel, N. H., L. F. Stickel and J. W. Spann. N. W. Miller and
G. G. Berg, ed. Chemical Fallout. C. C. Thomas, Springfield,
111., 1969, pp. 174-204.
89. Stickel, W. H., D. W. Hayne and L. F. Stickel. Effects of Heptachlor-
Contaminated Earthworms on Woodcocks. J. Wildl. Mgmt., 29: 132-
146, 1965.
90. Stout, V. F. Pesticide Levels in Fish of the Northeast Pacific.
Bull. Environ. Contam. and Toxicol., 3(4): 240-246, 1968.
91. Stringer, A. and J. A. Pickard. The DDT Content of Soil and Earth-
worms in an Apple Orchard at Long Ashton. Rep. Agric. Hort.
Res. Stn. Univ. Bristol for 1963. p. 127-131, 1965.
92. Thompson, N. P., W. B. Wheeler and A. J. Norden. Residues in Three
Peanut Varieties Grown in Dieldrin Treated Soils. J. Agri.
Food Chem., 18(5): 862-863, 1970.
93. USDA, Agricultural Research Service and Office of Research and
Development, EPA. Control of Water Pollution From Cropland,
Volume I - A Manual for Guildline Development. Revised Draft,
April 1975.
94. Vermeer, K. and L. M. Reynolds. Organochlorine Residues in Aquatic
Birds in the Canadian Prairie Provinces. Can. Field-Naturalist,
84: 117-130, 1970.
95. Warner, M. L. and E. H. Preston. Review of Environmental Impact
Assessment Methodologies. U.S. Environmental Protection Agency,
April 1974.
96. Westlake, W. E. and F. A. Gunther. Organic Pesticides in the
Environment. R. F. Gould, ed. Advances in Chemistry Series 60,
American Chemical Society, Washington, D.C., 1966, p. 110.
97. Wheatley, G. A. The Assessment and Persistence of Residues of
Organochlorine Insecticides in Soils and Their Uptake by Crops.
Ann. Appl. Biol., 55:325-329, 1965.
125
-------
98. Wheeler, W. B., D. E. Frear, R. 0. Mumma, R. H. Hamilton and R. C.
Cotner. Absorption and Translocation of Dieldrin by Forage Crops.
J. Agr. Food Chem., 15(2): 231-234, 1967.
99. Wright, B. S. Some Effects of Heptachlor and Dieldrin on New
Brunswick Woodcocks. J. Wild!. Mgmt., 29(1): 172-184, 1965.
126
-------
CHAPTER 6
AESTHETIC EFFECTS*
Closely related to the environmental effects is society's perception of
its surroundings -- what we have chosen to call aesthetic effects. The close
relationship can be seen when one considers that the natural environment plays
a large part in one's perception of the quality of life. Even the urban
resident can be affected if he realizes that the birds which were once in the
park or the trees which were once on the corner are no longer there. In fact,
the urban resident may be the most avid pursuer of the natural environment
for its aesthetic component since research has shown that "individuals tend
to seek elements which are lacking in their non-leisure environment" (5, p. 132),
Pesticide use can materially affect society's perception of its sur-
roundings. The home use of pesticides is almost entirely devoted to increasing
or at least maintaining an aesthetic level. Perhaps the only home uses of
pesticides which have major non-aesthetic components are the use of pesticides
on home vegetable gardens, and for home rodent control, termite control and
anti-fungal treatments. Even in these cases, the aesthetic element cannot be
dismissed. The use of pesticides by public agencies sometimes have major
aesthetic components. Spraying for control of gypsy moth in the Northeast
can be included here. The denuding of whole forests by insects can have a
major effect on those viewing the scene as well as producing massive
ecological disruptions. The aesthetic value of the view has been decreased
and spraying the area, killing the insects and allowing the forest to
recover may restore the value which had been lost. There obviously are very
real aesthetic benefits which can stem from the use of pesticides. The use
of pesticides for aesthetic improvement is not, however, the major problem.
As previously discussed under environmental effects, pesticide residues
can have a major impact on various species of animals and plants. When this
impact is large enough, there may be a perceived lowering in the aesthetic
component. A well-known example of this is the robin poisonings due to their
consumption of earthworms which had fed on elm leaves sprayed with a pesticide
for control of Dutch elm disease. Judging from the response to the poisoning,
one would say that robins were an important part of the surroundings of that
city. The social cost of the spraying seemed to be high. Can this component
of pesticide use be measured?
Principal researcher in this section is Mr. Gary A. Shute, Center for the
Study of Environmental Policy, The Pennsylvania State University. This
section was prepared under the supervision of Dr. Donald il. Epp, Principal
Investigator.
127
-------
Some progress has been made in the past ten years in measuring social
costs of externalities. The focus has been on air and water pollution, however,
and it was hoped that the measurement techniques developed for these external-
ities could be adapted to pesticide residues and their externalities. Such
an adaptation has proved impossible.
PROBLEMS OF QUANTIFICATION
Negative pesticide-related aesthetic affects are different from the
negative aesthetic affects attributable to water and air pollution. This is
true even though pesticide residues are a part of the total water and air
pollution problem. This can be seen more easily if the discussion breaks
down the attributes which have made some progress possible in quantification
of water and air pollution aesthetic effects.
A major difference between water pollution and pesticide residue caused
aesthetic effects is the possibility of assessing the source. Water pollution
is commonly associated with major point sources. These sources are present
over a long period of time and the fact of their existence becomes common
knowledge. As pollution becomes apparent, these sources are identified and
the amount of effluent discharged is measured. To the extent that the water
district can do so, pressure can be brought to bear on the polluters to
reverse the aesthetic damages. This is not to say, however, that in some
cases non-point sources may be the major cause of pollutant loading (e.g.,
agricultural operations, forests, urban areas, etc.).
The fact that the aesthetic damages of water pollution can impact the
viewers directly is what makes attempts at quantification somewhat successful.
Water pollution is readily identifiable. Fish kills may become common, scum
may float on the surface, vegetation may grow out of control, foul smells may
emanate from the river. All of this makes human activity less enjoyable.
Swimming and picnicking are curtailed. Tihansky (10) has made estimates of
the loss to consumers of swimming at public beaches due to the closing of
the beaches for health reasons. To do this, he drew on the fact that people
were willing to pay for improved swimming conditions plus the unstated fact
that closed beaches could be considered a homogenous entity for the purpose
of the study. Further, it is reasonable to assume that the major component
of the aesthetic loss was due to reduced swimming opportunities. As a first
estimate of the aesthetic cost of water pollution, the study did succeed.
Finally, the impact of an increased dose of water pollution is predic-
table. The increased discharge of untreated waste into the waterway can only
further degrade the aesthetic impression for an observer.
Measurement of the aesthetic effects of airborne pollution is just begin-
ning. The sources of urban air pollution (a typical non-point source) can be
generally specified into broad categories (transportation related, production
related, and residentially related) but identification of individual air
polluters and quantification of their contribution to the air pollution
problem is difficult. Some of the large polluters can be identified, however,
and if their contribution is a large proportion of the exhaust load, clean-up
128
-------
pressure can be brought to bear. Air quality indexes have been constructed
and are in use in major cities. This serves as a warning device in the
present and as an indicator of air quality trends in the long run. The
aesthetic component of air pollution -- dirt, soot, odor and obscured views --
is real enough but difficult to measure in the urban environment due to the
gradual nature of the pollution levels. With a continuous monitoring system
this problem may be overcome and assessment such as took place in New Mexico
may be possible.
In the Four Corners area of New Mexico, Arizona, Colorado and Utah, a
major source of air pollution was introduced. This coal burning power plant
caused an aesthetic impact on an area whose major attribute was visual splen-
dor. Randell et al. (8) undertook a survey of people's willingness to pay for
restoration of the lost aesthetic opportunity and found significant support
for clean air. The study was possible because the degradation was due to a
clearly defined entity which was making an impact directly upon the individual.
AESTHETIC IMPACT ASSESSMENT
The state-of-the-art measurement techniques are by no means adequate
vehicles for the quantification of pesticide impacts on aesthetic consider-
ations. The majority of evaluatory schemes are site-specific since they were
developed for air and water pollution and construction projects and are not
readily adapted for the examination of aesthetic impacts of pesticide use. A
number of assessment procedures have been critiqued by Redding (9).
In an attempt to insure that as few aesthetic consequences as possible
are overlooked, we again propose the implementation of a checklist or matrix
evaluation technique. The rationale for an aesthetic matrix is essentially
the same as for an environmental effects matrix in that existing evaluation
schemes do not readily permit quantification of environmental and aesthetic
impacts. The matrix is designed in such a manner so as to insure that the
evaluator will at least be aware of possible aesthetic ramifications.
Figure 16 illustrates the envisioned aesthetic impact matrix.
The abscissa is comprised of the various components of the environment
which may interact with the ordinate which is comprised of activities with
aesthetic components. If an impact occurs, a slash is placed at the inter-
section of the two categories being examined. When attempting to assess the
potential impacts (both beneficial and harmful) of a pesticide, using a matrix
of this type would be inadequate unless coupled with an examination of food
chain links and other environmental interrelationships to insure the consider-
ation of not only first-round effects but also second- and third-round conse-
quences of chemical compounds in the ecosystem. It may prove advantageous to
implement a scheme consisting of a series of matrices examining first-,
second-, and third-round effects. For example, an algacide may improve boating
and fishing by reducing the amount of algae in the water (a "+" indicating a
beneficial impact) but, may at the same time adversely affect fishing by
having harmful effects on aquatic invertebrates and aquatic vertebrates. Pes-
ticide use can affect the aesthetic component of hunting if the trees, shrubs
and other terrestrial flora are adversely impacted. Conversely, if a pesticide
129
-------
0
o"
0"
X
X
\;
X
X
x,
X
X
\
X
W
O
»
X
x
\
x
x
X
CD
o
3
x
-:
x
x
X
tr
Q.
•1
ct Hunan Contact
o
s
VI
M-
3
X
X
X;
X
x
;/)
3
x
X
X;
X
x
x
Human Contact
-
"
?
a.
c.
c.
Jatching
x
x
X
3
00
X
X
sr
3
•D
7)
fr*
50
x
x
3
w
X
x
X
X
r
Q.
x
•3
rf
>
ttvit ies
Aesthet
Environmental
Non- Renewable
1. Flora
a. Terrestrial
1 . Trees
2. Shrubs
3. Grasses
4. Microf lor.i
5. Agr icul tural Crops
b. Aquae ic
1. Algae
2. Submerged Macrophvtes
3. Emergent Macrophytes
2 . Fauna
I . Arthropods
2. Anncl ids
3. Mollusca
1 . Freshwater
a. Protozoa
b. Hoi lusca
c. Annel ids
d. Arthropods
1. Insects
2. Crustaceans
2. Est urine/Marine Invert .
n. Protozoa
1. Dinofl.igcl latt's
*>• Mollusca
c . Anncl ids
1 . Poly chaete
tl. Arthropods
1 . Crustaceans
c. Terrestrial Vertebrates
I. Aves (Birds)
2. Mammals
a. Domestic
b. Wildlife
d. Aquatic Vertebrates
!• Freshwater
a. Pisces
b. Amphibian
c. Reptile
e. Esturine/Marine Vert.
1 . Pisces
. 2. Aves
Figure 16. Pegtlclde Impact on activities with aesthetic components
130
-------
reduces the number of insect pests in an area, this may improve the aesthetics
of hunting. This insect-aesthetic relationship holds true with the remainder
of the sporting activities and tourism. Pesticide-use decisions will, of
course, severely impact upon hunting if there is an adverse effect on wildlife.
Assume that an insecticide is applied to control insects in the home and
garden. At the onset, this action will have a positive impact on trees,
shrubs, ornamental, etc. in that it kills arthropods (insects) and molluscs
(slugs and snails). On the other hand, the pesticide may directly or
indirectly kill annelids (earthworms) that are important to soil fertility and
bees that are important in pollination. Beneficial birds and mammals may also
succumb to the pesticide. Thus, an action to protect flora (e.g., garden
crops) may result in production losses and pose serious.environmental
consequences.
The comprehensiveness of the analysis will depend upon the amount of
available information on the particular pesticide, organisms and activities
being considered. When considering effects other than first-round impacts,
the analysis can continue ad infinitw, In this case the evaluator must set
the limits of analysis.
THE ALDRIN CASE
In dealing specifically with aldrin application on corn, the information
on direct aesthetic impacts is meager at best. This is most likely a result
of tempering intermediates — plant or animal -- tending to blur the causal
chain between aldrin-dieldrin and environmental and aesthetic impacts.
Environmental and aesthetic impacts must be considered together since an
aesthetic impact is a consequence of an environmental alteration.
Though we are not able to establish a direct aesthetic impact related to
aldrin use on corn, a cause and effect relationship has been drawn between
dieldrin application on rice and the death of snow geese which fed on the
treated rice. Therefore, the aesthetic impact of dieldrin can be alluded to
since the death of snow geese will affect hunting and possibly landscape
aesthetics. It is quite likely that incidents of this type are more common
than reflected in the literature since scavengers tend to remove the carcasses
very quickly and many birds may fall into cover.
The aesthetic impact of pesticides will most likely be less if they are
used in a production sense -- to protect corn from insect damage -- than if
used for aesthetic purposes at the onset (i.e., gypsy moth control in
Northeastern forests). In any event, the causal chain must be established
prior to aesthetic impact assessment since "natural death" is not as obtru-
sive to people as one attributed to preventable causes. That is, people will
tend to assess the death of an organism differently if it is thought to have
died of natural causes rather than the victim of pesticide poisoning. There-
fore, the establishment of the correct cause and effect relationship
surrounding aesthetic impacts is imperative if the assessment of such effects
is to be correct.
131
-------
BIBLIOGRAPHY FOR AESTHETIC EFFECTS CHAPTER
1. Birkhoff, G. D. Aesthetic Measure. Harvard University Press,
Cambridge, Mass., 1933.
2. Coomber, N. H. and A. K. Biswas. Evaluation of Environmental
Intangibles. Genera Press, Bronxville, New York, 1973.
3. Diffey, T. J. Evaluation and Aesthetic Appraisals. British Journal
of Aesthetics, 7(4): 358-373, 1967.
4. Grimes, 0. F. Evaluation of Recreation and Aesthetic Uses of Water
in an Urban Setting. Paper presented at the Urban Economics
Workshop, University of Chicago, February, 1970.
5. Knopp, T. B. Environmental Determinants of Recreation Behavior.
Journal of Leisure Research, 4(2): 129-138, 1972.
6. Leopold, L. B. Method for Measuring Landscape Appeal. Natural History,
8: 36-45, October 1969.
7. Pebbles, J. J. A Methodology Study to Develop Evaluation Criteria
for Wild and Scenic Rivers. Report of Flood Control Subproject,
Water Resources Institute, University of Idaho, February 1970.
8. Randall, A. B., et al. Benefits of Abating Aesthetic Environmental
Damage from the Four Corners Power Plant. New Mexico State
University Agricultural Stat. Bulletin, Fruitland, New Mexico, 1974.
9. Redding, M. J. Aesthetics in Environmental Planning. U. S. Environ-
mental Protection Agency, Washington, D. C., 1973.
10. Tihansky, D. P. Recreational Welfare Losses from Water Pollution
Along U. S. Coasts. Journal of Environmental Quality,
3(4): 335-342, 1974.
132
-------
CHAPTER 7
DISTRIBUTION EFFECTS*
The preceding sections have emphasized the quantifiable benefits and
costs as well as some descriptions of the effects of pesticide decisions on
the health of humans and non-humans. In this section, procedures will be
described for analyzing and presenting the effects of pesticide-use decisions
that relate to the distribution of benefits and costs. Specific attention is
directed to the impact on geographic distribution, social effects, and
effects on the national balance of payments.
GEOGRAPHIC DISTRIBUTION OF EFFECTS
The geographic distribution of effects from the pesticide-use decision
comes primarily from models of economic production effects and from the dis-
persion of human health effects. These models will permit the determination
of the balance of benefits and costs by geographic region. Appropriate
regions will be determined by the particular pesticide decision under inves-
tigation. For some, the geographic areas may be rather small and carefully
delineated. In other decisions, the geographic area may encompass large
portions of the nation such as the Southeast, the Northeast, and the West.
The requirements of the particular pesticide decision determine the appro-
priate geographic areas for each report.
One of the most likely geographic effects from a pesticide-use decision
is the relocation of agricultural production. As was shown in Chapter 3 in
the example of aldrin used on corn, even when total production is virtually
constant, certain regions may have significant increases or decreases in the
production of a particular crop which can lead to further important changes
in the economy of the area. If corn production, for example, is signifi-
cantly reduced in an area, the demand for production inputs, such as fertil-
izer and specialized corn production machinery, will be reduced. This may
lead to reduced employment in farm supply firms and perhaps the closing of
some firms. In addition, there will be less corn to be stored or shipped by
grain handling companies; again leading to reduced employment in the area.
If the reduction in corn production continues, it is likely that livestock
feeding activities will be curtailed or discontinued, reducing the employment
in livestock shipping companies, marketing firms (e.g., auction markets) and
packing plants.
*
Principal researcher in this section is Dr. Kenneth P. Wilkinson, Department
of Rural Sociology, The Pennsylvania State University. This section was
prepared under the supervision of Dr. Donald J. Epp, Principal Investigator.
133
-------
In evaluating these possible changes, the analyst must consider the
availability of other employment in the region, the magnitude of the lost
employment relative to total area employment and linkages in the local economy
which may accentuate or mitigate the impact of agricultural production and
employment shifts.
A similar consideration should be given to areas where agricultural pro-
duction is projected to increase. A substantial increase in soybean production,
for example, may lead to changes in farm size and, therefore, the number of
farm families. New businesses, such as feed supplement manufacturing, may come
into the area bringing increased employment and possibly an inflow of new
people to the area. Again, a knowledge of linkages in the local economy will
help the analyst predict the economic ramifications of these changes.
Techniques for measuring linkages in the regional economy and developing
employment and income multipliers are available. Isard (4) describes how
these measures are calculated and possible extensions for analyzing regional
effects of a change in basic employment and production. Input-output analysis
and the resulting multipliers must be used with caution, however, since they
assume .that the basic structure of the economy remains constant. If a pesti-
cide registration decision causes a major change in the agricultural produc-
tion of a region, the multipliers calculated from the previous economic
structure are likely to be wrong.
Significant human health effects or changes in the environment can also
differ from region to region and may lead to economic and social changes. For
instance, the denuding of large areas of forest land, due to discontinued
spraying of insects, may lead to a reduction in tourist visits to the area and
a decrease in the value of seasonal home sites. These changes could cause
loss of seasonal employment and the failure of some tourist oriented businesses,
a reduction in sales tax and real estate tax revenues and a reduction in the
quantity or quality of local governmentally provided services. This could
accelerate the decline in local population. On the other hand, an area that
discontinues the use of potentially harmful pesticides may experience an
inflow of people who appreciate the "safer" environment or the wildlife that
is more abundant without the adverse chemical effects.
A more complete analysis of region impacts will include the likely social
ramifications of altered economic activity. This analysis is described in
the following section.
SOCIAL EFFECTS
Social effects of any policy decision are difficult to foretell, because
of the complexity of forces involved, and difficult, if not impossible, to
evaluate on objective grounds. Attempts to bring about specifically desired
social changes, particularly in complex social systems, rarely succeed.
Indeed, as Jay Forrester (3) among others has pointed out, the complex web of
"feedback loops" in a social system produces such an array of unanticipated
consequences that well-intended policy action often results in changes in
direct opposition to those desired. For a simple example, an economic
134
-------
development program intended to reduce poverty in a community might produce,
in order, new jobs, an influx of poorly trained migrants seeking service
employment, resulting in more poverty. Breaking into such cycles with any
chance of successful policy impact, requires, as a minimum requisite, a sense
of the complexity of social systems. Social changes are difficult to
evaluate objectively because of the inherently subjective character of
evaluation criteria. A given change, for example, from a pattern of family
farms to one of vertically linked corporate structures, might be viewed as
desirable from one value perspective and undesirable from another. There is
no universal basis, grounded in scientific knowledge, for assessing the
relative goodness of any value system; rather value systems must be treated
as integral wholes each with its strengths and weaknesses for meeting human
needs. Among social scientists interested in policy effects, one response
to problems of complexity and value relatively in social impact analysis has
been to attempt to develop social accounting systems rather than programmic,
goal-oriented models of social change.
A Social Accounting System
A social accounting system is based on a taxonomy of aspects of social
organization likely to be affected by a policy decision. At this stage of
the development of social science, such a taxonomy must be conceptually,
rather than empirically generated. Among alternative theories of social
organization there is some consensus as to the basic elements (but much dis-
agreement as to how these elements operate and interrelate in the functioning
and change of social systems). Abstracting commonalities from diverse
theoretical approaches reveals the following categories of elements of social
organization which must be considered in any social impact assessment:
• Human Ecology: Spatial distribution and inter-
relationships of people in adaptation to the
natural environment (e.g., population size,
composition and change; sustenance organization;
technology).
• Culture: Values and norms, particularly institu-
tions, in social life (e.g., value-based norms
governing family life, economic activities, edu-
cation, religious life, the polity, etc.).
• Social Structure: Group relationships expressing
cultural values and norms (e.g., families, volun-
tary organizations, firms, friendship groups,
government agencies, social classes, social
position and role networks).
• Social Processes: Interaction dynamics (e.g.,
cooperation, competition, conflict -- the
dynamics of social structure).
135
-------
• Social Psychology: Relation of the individual
to"society (e.gT, participation and ego-involve-
ment, self-concept and reference-identification,
"mental health").
The question of how these elements of social organization might be
affected by a policy change, such as regulation of a given pesticide, should
be approached with consideration of the following issues:
Social Significance--
Of what social significance (in terms of the above categories) is the
policy variable (e.g., the pesticide practice) in the relevant unit or level
of social organization (community, region, nation, etc.)? A crude approach
to answering this question would be simply to enumerate the number of elements
of social organization likely to be affected directly and obviously by a
policy change. Even a gross estimate, however, would have to be couched in
terms of a description of the major elements of current social organization.
Expected Changes in Social Organization--
What kinds of changes should be expected in those aspects of social
organization likely to be affected directly and obviously by the policy
change, and what subsequent changes should be expected in turn? This, of
course, raises again the issue of complexity in social forecasting. Two
sources of guidance, neither being in any sense definitive, are available in
the sociological literature. One is in the theories of social change which
have been developing over many years and the other is in the body of empirical
findings of sociological research, much of which dates only to the past few
decades.
Theories of social change are many and competitive. For convenience
these have been grouped, more or less satisfactorily in general treatments,
in two main types according to philosophic premises upon which they are based.
One type of theory of social change is based on the assumption that social
life is governed by a set of principles of order which prevade all aspects
of existence in the universe. An example is the theory referred to as
functional ism, which assumes that changes in any part of a social system set
off complementary changes in other parts so as to retain the fundamental
equilibrium of the whole system. A second type of theory of social change is
based on the assumption that order is created by the acts of men, that, in
fact, the natural direction of change in all elements of the universe is
toward random dispersion or entropy. Conflict theory, for example, assumes
that change occurs as the result of the confrontation of opposed interests
of independent groups. Interactional theory, a second example of the latter
type, assumes that social life reflects the collective interests of people
rather than the functional needs of systems. Empirical testing of propositions
derived from these competing theories has proceeded at a very slow rate, and
is in fact at an early stage.
The relevant body of empirical evidence relates to what has been called
"theories of the middle range," that is, to theoretical propositions which
136
-------
do not relate explicitly to the more global, philosophic assertions. A
number of bivariate relationships, such as that between social participation
and socio-economic status, have been reported with such frequency in socio-
logical studies as to give them the status of middle-range principles. The
testing of multivariate models is a recent innovation in sociological research.
Desirability of Expected Changes--
Are the changes anticipated desirable? This of course is a value
question, for which there is no value-free answer. Once certain values have
been accepted, however, weights and directions can be assigned to the changes
which are anticipated. A number of observers have attempted to ascertain the
fundamental values pursued in American Society. Central to most treatments
are the following:
• Efficiency (especially in adaptive or economic activities).
• Equity (especially in distribution of benefits and costs
of change).
• Community (especially in the sense of integration of
behavior and identities.)
• Self-actualization (i.e., the fulfillment of distinc-
tively human potentialities--which apparently requires
at least minimal levels of efficiency, equity and
community).
Community Level Assessment
The most relevant unit for assessment of social effects of pesticide
regulation decisions is the local society or community. For larger units,
such as the nation, economic, environmental and health consequences would be
more clearly discernible. Social effects at the national level would be
diverse and, perhaps, counter-balancing among themselves. One exception at
the national level would be in the effects on concentration of power in the
control of agriculture. While family-type enterprises continue to be the
unit in the agricultural sector, large-scale, vertically linked corporate
structures account for much of the growth in agricultural production in
recent years. In general, larger corporate structures should be expected to
have greater resources to apply in adapting to a regulatory decision; thus
further growth of such structures—with obvious ramifications in social
organizations might be encouraged by such decisions. However, the forces
encouraging the development of such structures are much broader than those
having to do with use of specific pesticides. Increased pesticide use and
complex organizational structure are part of a process, sometimes referred
to as modernization, which has been occurring in accelerated form in
virtually all sectors of Western society for many years. It is highly
unlikely that any action by government, or any set of purposive actions,
could alter the fundamental direction of this process. Social effects which
could be effectively monitored and perhaps affected are best seen in the
variety of circumstances prevailing at the local level.
137
-------
In describing effects at the local level, it is important to have a
realistic concept of the geographic scope of the local society. Areas which
were relatively independent communities in former times might now be sub-units
of a larger community. Sociological studies of communities said to be
"declining," for instance, often reveal that the community is in fact expan-
ding as services and institutional functions formerly associated with a
specific, limited place move to a larger regional center. The general cri-
terion for operational definition of a local society is the presence in it
of the full range of institutional forms available in the larger society.
Functional economic area delineations often provide an adequate starting
place for identifying community boundaries.
At the community level, social effects of the regulation of a given pes-
ticide will depend in part on the characteristics of the community and the
significance of the use of the pesticide in its structure and functioning.
Considerable work has been done in recent years by Brian Berry (2) and other
central-piace geographers in the use of factor analysis and secondary data
sources to classify community types. Such work might be used to estimate
the significance of agriculture in community life. The general indication
of research in this area is that the number of communities organized pri-
marily around agriculture has declined significantly since World War II and
is likely to decline further. The number of such communities in which a
given pesticide is of crucial significance in agriculture is small, though
still large enough in absolute terms to warrant national concern for social
consequences of regulation. In most other communities, i.e., those in which
agriculture is of relatively minor significance or in which a given pesticide
is not crucial, the social effects of a change in policy are likely to be
slight, although they might be intensively experienced by some individuals
and families.
It is the case of the clear significance of pesticide use in community
life that the question as to what effects are likely becomes most relevant.
If, for example, regulation means that the type of agricultural enterprise
in the area must be altered, or replaced with non-agricultural activities,
the effects.-could be far-reaching. The size and age-composition of the
population might be altered, resulting in a new mix of orientations, needs
and demands in community affairs. Tax revenues for support of public ser-
vices and amenities might be curtailed. Status relationships might be
altered. Linkages to the larger society through formal agencies might be
affected. The specific areas of likely impact would depend upon character-
istics of the given local situation.
The "state of the art" and the "nature of the beast" are such that social
effects cannot be foretold with precision. The theories, methods and findings
of sociology could be used, however, to identify significant issues and social
parameters to be considered in the making of policy decisions.
BALANCE OF PAYMENTS EFFECTS
The purpose of this section is to investigate the impact that a pesti-
cide-use decision might have on the international trade position of the
133
-------
United States. This section does not contribute to a benefit-cost ratio
directly but is considered under the heading of distribution. As such, the
results of this section are considered as auxiliary information available to
the decision-maker.
Impact of Foreign Trade
Foreign trade allows a country to improve its efficiency by purchasing
from abroad those products which can be produced more efficiently abroad.
Payment for these products must ultimately be made by sales of domestic goods
to foreigners. It can be said that a country exports in order to pay for its
imports.
U.S. agriculture makes a significant contribution to a positive trade
balance. In fiscal year 1975, the value of agricultural exports was 21 percent
of the total value of U.S. exports. The value of agricultural imports was 9
percent of the total value of U.S. imports. This resulted in a positive
agricultural trade balance of over $12 billion which was able to cover a
negative non-agricultural trade balance of over $1.8 billion. Table 15 shows
that the export value of just seven crops accounted for over 71 oercent of
total U.S. agricultural export value in fiscal year 1974 and 1975. Percentage
of total domestic crop which entered into foreign trade is given in Table 16.
These tables demonstrate the importance of major agricultural products to
U.S. trade balance.
In Chapter 3 and in U.S. Environmental Protection Agency (7), the
quantitative analysis of production effects of restricting the use of aldrin
on corn assumed that any changes in production caused by a decision in pesti-
cide use would not significantly affect the quantity of product consumed
domestically. It is possible, however, that a relatively small change in the
amount produced may have a significant effect on the amount of produce ex-
ported. The difference in effect on domestic and foreign markets is primarily
due to the greater price elasticity of demand in the foreign market than in
the domestic market for most products dependent on pesticides for their
production. Another factor which may reduce the amount of export sales,
particularly of agricultural products, is that a reduced supply limits the
opportunities for special sale efforts tied to foreign policy considerations.
Both of these factors, a slightly increased price and reduced foreign policy
arrangements, may reduce the amount of foreign exchange earnings from the
sale of agricultural products.
The method for calculating the foreign exchange effects utilizes price
changes from the simulation model in the economic effects section along with
estimates of the elasticity of demand in the foreign market. These calcula-
tions indicate changes in foreign demand for the products under consideration
and iterate with the simulation model through adjustment of the exogenous
foreign demand components that are added to various regional demand quantities.
If ft appears that this alters the results of the simulation analysis, the
procedure is rerun using the new foreign demand quantities. After a final
foreign demand effect is calculated, it is translated through current prices
or projected prices into a foreign exchange earnings adjustment.
139
-------
TABLE 15. MAJOR U.S. EXPORT CROPS
1974 - 1975
Crop
Wheat
Corn
Soybeans
Cotton
Tobacco
Ri ce
Grain sorghums
Fiscal Year 19741
$ 4,556,405,000
3,729,142,000
3,272,832,000
1,293,999,000
813,601,000
752,282,000
631,127,000
$15,049,388,000
Fiscal Year 1975
S 4,797,015,000
3,988,207,000
2,951,232,000
1,016,953,000
910,088,000
1,002,185,000
616,308,000
$15,281,988,000
Total U.S.
Agricultural
Exports
$21,293,000,000
$21,584,000,000
SOURCES: 1) U.S. Foreign Agricultural Trade Statistical Report,
Fiscal Year 1974. Economic Research Service, U.S.
Department of Agriculture, October 1974.
2) U.S. Foreign Agricultural Trade Statistical Report,
Fiscal Year 1975. Economic Research Service, U.S.
Department of Agriculture, December 1975.
140
-------
TABLE 16. UNITED STATES EXPORTS OF MAJOR CROPS AS A
PERCENTAGE OF TOTAL UNITED STATES PRODUCTION
1 2
Quantity Exported Domestic Production
Crops (Fiscal Year 1975) (1974) Percent
Wheat 999,236,000 bu. 1,796,187,000 bu. 55%
Corn 1,122,137,000 bu. 4,663,631 ,000 bu. 24%
Soybeans 404,514,000 bu. 1,214,802,000 bu. 32%
Cotton 3,816,000 bales 11,540,100 bales 33%
Tobacco 638,404,000 Ibs. 1,989,728,000 Ibs. 32%
Rice 5,059,180,000 Ibs. 11,239,400,000 Ibs. 45%
Grain sorghums 190,894,000 bu. 629,222,000 bu. 30%
SOURCE: 1) U.S. Foreign Agricultural Trade Statistical Report,
Fiscal Year 1975. Economic Research Service, U.S.
Department of Agriculture, December 1975.
2) Crop Production. Statistical Reporting Service, U.S.
Department of Agriculture, January 15, 1976.
141
-------
It is entirely possible that a pesticide decision may have a very minor
effect on total production of a particular crop and an insignificant effect
on domestic price but at the same time cause a substantial adjustment in
foreign sales and have a significant effect on foreign exchange earnings.
This could be an important consideration in making a decision since agricul-
tural exports recently have accounted for large portions of the foreign
exchange earnings of the United States.
The Effects on the U.S. Trade Balance of Restricting Aldrin on Corn
Corn is used throughout the world as a feed grain for cattle, hogs and
poultry. A large quantity is also used for human consumption. The U.S.
contribution to this worldwide demand for corn is substantial and the return
to the U.S. in terms of foreign currency earnings is likewise substantial.
As shown in Tables 15 and 16, the U.S. exported approximately 24 percent of
its 1974 corn crop and received nearly $4 billion from this sale. The magni-
tude of this dollar flow indicates that the result of even a small percentage
change in output can be a significant change in the amount of absolute dollar
terms. It is for this reason that the impact on the U.S. trade balance of a
change in the production function of corn (due to the restriction of aldrin
on corn) is examined.
As described in the previous section, the procedure to be followed in
this evaluation of foreign exchange effects is to estimate the domestic price
change expected from the registration action, and, using estimates of the
elasticity of foreign import demand, calculate the new quantity demanded by
foreign buyers. The domestic price change is derived from the simulation
model described in Chapter 3. The estimate of foreign import demand is
obtained from a U.S.D.A. simulation model titled Commodity Projections for
1985 (6).
Among the commodities which are entered into the U.S.D.A. model are
wheat, coarse grains (corn, barley, oats, sorghum), oil seeds, beef and dairy
products. This model breaks the world into developed and less-developed areas
since it has been found that these two groupings have significantly different
reactions to price and quantity changes. The estimate of the elasticity of
foreign import demand for corn obtained from this model is .3 for the
developed area and .65 for the less-developed. This means that for a 10 per-
cent increase in the price of corn, foreigners from developed countries will
demand a 3 percent lower quantity and foreigners in less-developed countries
will demand a 6.5 percent lower quantity. Both of these elasticity estimates
are inelastic.
The results of the production-section modeling indicate that if aldrin
is restricted from use on corn soil insects, the response of the farm sector
is one of increasing the acreage devoted to corn as well as some shifts in
cropping location. This results in a maintainence of the quantity of corn
produced and a stable price of corn. With no change in domestic price,
there can be no predicted change in the quantity exported due to a domestic
price change.
142
-------
BIBLIOGRAPHY FOR DISTRIBUTION EFFECTS CHAPTER
1. Abel, Martin E. and Frederick V. Waugh. Measuring Changes in
International Trade. Journal of Farm Economics, 48, Nov. 1966.
2. Berry, Brian, ed. City Classification Handbook. Wiley, New York, 1972.
3. Forrester, Jay W. World Dynamics. Wright-Allen Press, Cambridge,
Mass., 1971.
4. Isard, Walter. Methods of Regional Analysis: An Introduction to
Regional Science. MIT Press, Cambridge, Mass., 1960.
5. Rojko, Anthony S., Francis S. Urban and James J. Native. World Demand
Prospects for Grain in 1980. Foregin Agricultural Economic
Report No. 75, USDA, ERS, Dec. 1971.
6. U.S. Department of Agriculture. Commodity Projections for 1985.
Economic Research Service, undated.
7. U.S. Environmental Protection Agency, Economic Analysis Branch, Criteria
and Evaluation Division, Office of Pesticide Programs. Detailed
Information on Linear Programming Analysis of Organochlorine
Suspension of Corn Use in the United States. September 1975.
8. Waugh, Frederick V. Demand and Price Analysis, Some Examples from
Agriculture. Technical Bulletin 1316, USDA, ERS, Nov. 1964.
143
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/5-77-012
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
IDENTIFICATION AND SPECIFICATION OF INPUTS FOR
BENEFIT-COST MODELING OF PESTICIDE USE
5. REPORT DATE
August 1977 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
D. J. Epp, F. R. Tellefsen, G. A. Shute, R. M. Bear,
K. P. Wilkinson
9. PERFORMING ORGANIZATION NAME AND ADDRESS
The Pennsylvania State University
University Park, Pennsylvania 16802
10. PROGRAM ELEMENT NO.
1HB6I7
11. CONTRACT/GRANT NO.
R863247-01-1
12. SPONSORING AGENCY NAME AND ADDRESS
Office of Air, Land and Water Use - Wash., DC
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/5no/16
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Benefit-Cost (B/C) analysis requires inputs which are inclusive, valid, quantifi-
able and reliable. Proper attention to the procedures outlined in the six sections of
this report will give a broad, although not necessarily complete, review of the effects
of a pesticide-use decision. The six major sections are:
(1) A taxonomy of pesticide use effects. (2) Economic production -- Improvement
in production is the major benefit of pesticide use. Although alternative cost assumpp
tions are commonly used in B/C modeling, the opportunity cost approach is argued to be
a superior procedure. (3) Human health — Statistical methods are identified to
measure the health costs of pesticide use. Chronic health effects are the most diffi-
cult to analyze but methods are shown for measuring both dollar and utility values.
(4) Environmental impact -- Impacted organism/effects matrices are developed to insure
comprehensive consideration of environmental impacts even though the state-of-the-art I
does not permit direct incorporation of these effects into the B/C ratio. (5) Aesthetic]
impacts -- Organism/effects matrices are developed for aesthetic impact assessment.
(6) Distribution effects -- Social class, income class and international trade effects
are presented as important considerations exogeneous to the B/C ratio.
A bibliography is presented with each major section.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Pesticide
Economic Analysis
Benefit/Cost Analysis
Aesthetics
Equity
Econometric
Statistical Analysis
I).IDENTIFIERS/OPEN ENDED TERMS
Environmental Economics
Environmental f'anaqerrent
Human Health Effects
Environmental Impacts
rtisk/Benefit Analysis
c. COSATI l-'iclil/tiroup
05A
05r
13B
18. DISTRIBUTION STATEMENT
UNLIMITED
19. SECURITY CLASS (This Report)
UNEI ASSIFIER
21. NO. OF PAGES
154
20. SECURITY CLASS (Timpage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
144
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Technical Information Staff
Cincinnati, Ohio 45268
OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE. S30O
AN EQUAL OPPORTUNITY EMPLOYER
POSTAGE AND FEES PAID
US ENVIRONMENTAL PROTECTION AGENCY
EPA-335
Special Fourth-C lass Rate
Book
I
\
\
ui
o
// your address is incorrect, please change on the above label;
tear off; and return to the above address.
If you do not desire to continue receiving this technical report
series. CHECK HERE Q; tear off label, and return it to the
above address.
------- |