EPA-600/5-75 010
March 1975
Socioeconomic Environmental Studies Series
Measuring External Effects of
Solid Waste Management
Office of Research and Development
U.S. Environmental Protection Agency
Washington. DC 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and -'Development, Environmental
Protection Agency, have been grouped into five series. These five broad
categorieswere ".established to facilitate farther development and appli-
cation of environmental technology. Elimination of traditional grouping
was consciously planned to foster technology transfer and a maximum inter-
face in related fields. The five series are:
1. Environmental Health Effects Research
2. Enviromtjental Protection Technology
3, Ecological Research
4. Environmental Monitoring
5, Socioeconomic Environmental Studies
This report has been assigned to the.-SOCIOECONOMIC ENVIRONMENTAL STUDIES
series. This series includes research on environmental management,
economic analysis,ecological Impacts, comprehensive planning and fore-
casting and analysis Hjethodologies. Included are tools for determining
varyfrtgip^acts of alternative policies, analyses of environmental plan-
ning techniques at the regional, state and local levels, and approaches
to measuring environmental quality perceptions, as well as analysis of
ecological and economic impacts of environmental protection measures.
Such topics as urban form, industrial mix, growth policies, control and
organizational structure are discussed in terms of optimal environmental
performance. These interdisciplinary studies and systems analyses are
Resented in forms varying from quantitative relational analyses to manage-
ment and poli cy-oriented reports-
EPA REVIEW NOTICE
This report has been reviewed by the Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or reconfRendation for use.
Document is available'to the public fhrpugh the'Ua-ticnal Technical
.Isaformation Service, Springfield, Virginia. 22151.
-------
EPA-600/5-75-010
March 1975
MEASURING EXTERNAL EFFECTS
OF SOLID WASTE MANAGEMENT
By
Dr. Richard Schmalensee
Dr. Ramachandra Ramanathan
Dr. Wolfhard Ramm
Dr. Dennis Smallwood
Grant No. R-801673
Program Element 1DB314
ROAP 02AAE, Task 06
Project Officer
Dr. Fred H. Abel
Resource Analysis Staff
Washington Environmental Research Center
U.S. Environmental Protection Agency
Washington, D.C. 20460
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
-------
ABSTRACT
This study analyzes the environmental impact of
using sanitary landfills for disposing of solid waste.
The relevant economic theory concerned with the measure-
ment and valuation of external effects is developed and
previous empirical studies undertaken to estimate
environmental impacts of various activities are reviewed.
Both the theoretical and empirical evidence suggests
that the costs and benefits of external effects are
extremely difficult to measure directly but under cer-
tain circumstances property value studies can be used
to obtain indirect estimates.
A survey of the technology of sanitary landfills
suggests that a properly designed fill will cause very
little air and water pollution, but may impose visual
and noise pollution on nearby residents. These hypoth-
eses are tested with data on property surrounding four
sanitary landfills in Los Angeles County. A model of
the determinants of residential property values is
formulated and estimated. The model includes variables
which describe the characteristics of the property and
variables which describe the relationship of the prop-
erty to the fill. Statistical estimates of the param-
eters of the model indicate that proximity to and
111
-------
view of a sanitary landfill do not significantly reduce
the market or assessed value of surrounding property.
This report was submitted in fulfillment of Grant
Number R-801673, by the Institute for Policy Analysis under
the sponsorship of the Environmental Protection Agency.
Work was completed as of January, 1974.
iv
-------
CONTENTS
Page
Abstract ill
List of Figures vii
List of Tables ix
Acknowledgements xi
Conclusions 1
Recommendations 3
Chapters
Chapter I: The Price System, Market Imper- 4
fections, and Resource Allocation
Chapter II: Measuring Externalities in 39
Theory
Chapter III: The Valuation of External 115
Effects: Empirical Studies
Chapter IV: A Survey' of the Technical 174
Aspects of Solid Waste Externalities
Chapter V: Problems in Measuring the 301
Externalities of Solid Waste Disposal
Chapter VI: An Empirical Study of the Impact 338
of Sanitary Landfills on Residential
Property Values
-------
FIGURES
No. Page
1.1 Representative Indifference Curves 9
from Two Preference Maps
1.2 The Production Possibilities Frontier 13
for Apples and Beef and a Representative
Indifference Curve
1.3 Competitive Equilibrium in the A and B 16
Industries
1.4 The Supply and Demand Equilibrium with 25
an External Cost Present
1.5 A Pigouvian Tax Set Equal to Marginal 27
Damage in Equilibrium
1.6 A Pigouvian'Tax Set Equal to Marginal 27
Damage in the Original Competitive
Equilibrium
2.1 Compensating Variation and Equivalent 49
Variation Measures of the Benefits from
a Reduction in the Price of x
2.2 Demand Curves and Measures of the Benefits 55
of a Reduction in the Price of x
2.3 Compensational Variation and Equivalent 63
Variation Measures of the Benefits of a
Reduction in Air Pollution Control
2.4 Compensated Demand Curves on the Benefits 67
of Air Pollution Control
3.1 The Benefits of an Externality in 126
Hydroelectricity Generation
4.1 Aerobic Decomposition — Nitrogen, 228
Carbon, and Sulfur Cycles
4.2 Anaerobic Decomposition — Nitrogen, 230
Carbon, and Sulfur Cycles
vii
-------
No. Page
4.3 Variation in Gas Composition with Time 232
in Cell A from Inverted Collection Can
at 13-ft Depth
4.4 Gas Concentration over Time at a Test 233
Sanitary Landfill
4.5 Total Water Withdrawals by States from 242
Groundwater Resources and from Surface
Water
4.6 Average Potential Infiltration for the 247
U.S.
5.1 Three Possible Relationships between 324
Damage (Social Cost) and the Level of
Some Causal Variable x
6.1 Existing Landfills in Metropolitan 349
Los Angeles County
6.2 Palos Verdes, Landfill No. 1 350
6.3 Spadra, Landfill No. 2 351
6.4 Mission Canyon-Sepulveda, Landfill No. 3 352
6.5 Map of Mission Canyon Landfill 353
6.6 Scholl Canyon, Landfill No. 4 354
61.? Map of Scholl Canyon Landfill 356
6.8 Calabasas, Landfill No. 5 357
6.9 Puente Hills, Landfill No. 6 357
6.10 Map of Puente Hills Landfill 358
viir
-------
TABLES
No. Page
4.1 General Classification of Solid Waste 178
Materials
4.2 Estimated Annual Average Composition of 191
1968 Municipal Refuse on "As-discarded"
Basis
4.3 Projected Refuse Composition 192
4.4 Recycling of Major Materials: 1967 197
4.5 Percentages of Weekly Collections That 203
Contained 50 or More Fly Larvae During
the Eight Weeks of Study
4.6 Number of Garbage Containers From Which 204
Fly Larvae Migrating Exceeded 50
4.7 Composite Analysis of Average Municipal 206
Refuse, As-received Basis
4.8 The Size Distribution of Particulates From 208
European Incinerators
4.9 Particulate Emission From Municipal 210
Incinerators at Furnace Outlet
4.10 Chemical Analysis of Incinerator Flyash 211
Samples
4.11 Particulate Loading of Incinerator Stack 215
Gases
4.12 Particulate Loading of Incinerator Stack 216
Gases
4.13 Approximate Hydraulic Conductivities 234
4.14 Gas Production From Organic Substances 237
4.15 Comparison Refuse Leachates with U.S. 249
Public Health Service Standards
IX
-------
No. Page
4.16 Incinerator Residue Composition 262
4.17 Incinerator Residue Composition 262
4.18 Emission Factors for Open Burning of 275
Municipal Refuse
6.1 Operating Landfill Sites in Los Angeles 347
County
6.2 List of Variables and Their Definitions 361
6.3 Means and Standard Deviations of Selected 396
Variables
6.4 Name of Landfill: Palos Verdes 397
6.5 Name of Landfill: Palos Verdes 398
6.6 Name of Landfill: Palos Verdes 399
6.7 Name of Landfill: Tujunga 408
6.8 Name of Landfill: Tujunga 409
6.9 Name of Landfill: Tujunga 410
6.10 Name of Landfill: Sheldon-Arleta 418
6.11 Name of Landfill: Sheldon-Arleta 419
6.12 Name of Landfill: Sheldon-Arleta 420
6.13 Name of Landfill: Calabasas 427
6.14 Name of Landfill: Calabasas 428
6.15 Name of Landfill: Calabasas 429
6.16 Elasticity of Price with Respect to Fill 430
Variables
-------
ACKNOWLEDGMENTS
The principal researchers on this project were
Dr. Ramachandra Ramanathan, Dr. Wolfhard Ramm, Dr. Richard
Schmalensee, and Dr. Dennis Smallwood of the Institute for
Policy Analysis, La Jolla, California 90237. Substantial
research assistance was provided by Mr. Harold Nelson,
Mr. Richard O'Toole and Mr. Stephen Thompson.
Data on the assessed values and other property
characteristics were provided by the Office of Assessor,
County of Los Angeles. Data on sales price and other
property characteristics were obtained from SREA Market
Data Center, Inc.
We are indepted to the EPA reviewers, whose comments
on an earlier draft of this report were most helpful.
XI
-------
CONCLUSIONS
1. Direct measurement of external costs and benefits
is in principle feasible, but depends on having infor-
mation which is typically unobtainable. Since perceived
external effects will be reflected in property values/
the analysis of the determinants of property values pro-
vides a method for approximating external costs and benefits,
This indirect method has the advantage of requiring less
extensive data.
2. Past empirical studies have demonstrated that pol-
lution costs are in fact reflected in property values.
A properly specified property value study can be ex-
pected to reveal the existence of significant environ-
mental impacts that may be associated with solid waste
disposal in sanitary landfills.
3. An important assumption of the indirect method of esti-
mating the costs and benefits of external effects is that
property owners fully comprehend the impact of such effects
on them. The technology of sanitary landfills is such that
little if any water or air pollution would result from
a properly designed fill. The primary external effect,
-------
if any, associated with a fill would take the form of
noise and visual pollution. Since these effects would
be readily perceived by property owners, a study of
property values is an appropriate technique for measuring
their costs.
4. A study of values of residential property adjacent
to four sanitary landfills in Los Angeles County re-
vealed a general lack of significant detrimental environ-
mental effects. In the one case where truck noise
associated with the fill had a negative effect on property
values, both proximity and view of the fill had positive
effects due presumably to the anticipated transformation
of the fill to a recreational site.
5. The precision of the estimates of the costs and
benefits of external effects is limited by the adequacy
of information on the relationships between property
values and non-environmental variables, such as the
characteristics of structures. Further work on
specification and estimation of these relationships is
needed.
-------
RECOMMENDATIONS
1. Since this study suggests that the social costs of
properly designed sanitary landfills do not significantly
exceed the direct costs to the municipality, there
does not appear to be a case for discouraging the use
of such fills.
2. Property value studies may be used as a tech-
nique for estimating the costs of environmental effects
in those instances where property owners can be ex-
pected to have knowledge of the existence and conse-
quences of such effects.
3. In order to improve the reliability with which the
costs and benefits of environmental effects can be
measured, further work on the specification and esti-
mation of models of real estate values is desirable.
-------
CHAPTER I
THE PRICE SYSTEM, MARKET IMPERFECTIONS,
AND BENEFIT-COST ANALYSIS
A. INTRODUCTION AND OVERVIEW
It is not necessary to document the increasing
importance to municipalities of the problem of disposing,
or of otherwise processing, the solid waste generated within
their boundaries. Governments, particularly at the local
level, are faced with the choice of techniques to process
and dispose of growing quantities of waste. In addition,
they have the option of implementing policies to alter the
flow or composition of waste, such as laws prohibiting
certain non-return containers, or a law making residential
garbage grinders compulsory, or even a fee on the quantity
of waste generated by households. In deciding which policies
and techniques to adopt as part of a waste management system,
it is important that the policymakers understand the nature
of an optimal decision, and particularly, the appropriate
criteria by which to compare alternatives to make a socially
optimal decision.
The purpose of economic theory is to analyze and clarify
the nature of socially optimal decisions in the context of
scarce resources. While the context is generally taken to
-------
be the supply of products which consumers buy in the market-
place, the theory is also applicable to such increasingly
scarce goods as clean air, unpolluted streams, and pleasing
views. In fact, the existence of such goods as these provides
the general basis for this report. It is because the dif-
ferent techniques for disposing of solid waste require the
"input" of such goods, in differing mixtures, that one must
address the problem of how to weight these social costs, which
arc not "internalized" to the municipality, along with the
ordinary costs, which are.
In this chapter, we briefly review the theory of effi-
cient resource allocation; it is clearly necessary to under-
stand the basic problem of externalities, which are, roughly
speaking, costs of carrying on an activity which are not
borne by the agent performing the activity. The problem of
modifying benefit-cost analysis to account for externalities
is then discussed, in particular the question of how to
quantify such external costs, in units which are commensurate
with the other benefit and cost measures. We also consider
the nature of the equity considerations which are often
glossed over in most benefit-cost analyses.
Chapter II provides a more rigorous review of several
of these issues. Various approaches to the measurement and
valuation of external effects are treated. The remainder of
-------
this chapter provides an introduction to that analysis.
Chapter III examines attempts that have been made to
obtain actual estimates of the money value of externalities.
Studies using both direct and indirect methods are discussed.
Particular attention is paid to studies that employ data on
property values or rents.
Chapter IV summarizes what is known about the physical
manifestations of the externalities associated with solid
waste management. It provides a survey of the literature
with respect to both the qualitative nature and the quan-
titative importance of these effects. This material is
important because in general, the better an economist under-
stands the relevant technology, the more likely he is to
base his analysis on defensible assumptions.
Chapter V relates the general principles of externality
valuation presented in Chapters II and III to the realities
of the solid waste problem, as outlined in Chapter IV. Some
of the important practical problems of externality valuation
in the solid waste context are considered.
Finally, Chapter VI presents the results of an empirical
analysis of the external effects of four sanitary landfills
in the Los Angeles area. The indirect method is used, and
disaggregated data on property values are employed.
-------
B. BASIC WELFARE ECONOMICS
Economic theorists have developed a strong justification
for allocating most economic resources through free markets
in which goods are traded at prices determined by competitive
processes. This justification is not based on presumptions
about the implicit desirability of capitalism or the insti-
tution of private property. Rather, it follows from a demon-
stration that the price system, together with competitive
markets, leads to "economic efficiency." As defined below,
economic efficiency can reasonably be considered as a desi-
rable attribute of an economic system without regard to one's
value judgments on most issues. The increasing, although
somewhat begrudging, use of prices to reduce misallocation in
the Soviet Union provides confirming evidence for most
economists1 belief that their advocacy of the price system
is not ideological in nature.
The concept of economic efficiency was developed as a
tool to be used to finesse the problem of weighing one man's
welfare against that of another. In a strikingly deft solu-
tion to this problem, Vilfredo Pareto, in his treatise
Cours d'Economie Politique published in 1897, suggested a
criterion for judging the relative efficiency of alternative
situations. The Pareto criterion states that: Situation A
-------
is considered to be preferable to Situation B if no one is
worse off and at least one person is better off in Situation
A, by their own evaluation. An allocation is then called
Pareto optimal if there are no reallocations which meet
the Pareto criterion. While this definition of efficiency
may seem almost vacuous, it has surprisingly strong impli-
cations about the nature of an optimal allocation of resources.
It is important to note that the welfare of an individual
is to be judged by the individual; paternalism is specifically
rejected. In order to specify the individual's own evaluations
of his welfare, the concept of a "preference map" and of the
related "indifference curves" are introduced. Taking all
possible combinations of all commodities (usually expressed
in terms of rates of consumption), indifference curves re-
present the classes of commodity bundles among which a par-
ticular consumer is indifferent. If commodity bundle 1
is preferred by the consumer to commodity bundle 2, it must
be on a "higher" indifference curve. For diagrammatic pur-
poses, consider the preference structure of two individuals,
John and George, in the context of just two goods, which we
will think of as apples and beef. The indifference curves
could then be depicted as in Figure 1.1. Since points 2 and
3 are on the same indifference curve, they represent alter-
native combinations of apples and beef (measured, for instance,
8
-------
Figure 1.1 REPRESENTATIVE INDIFFERENCE CURVES FROM TWO
PREFERENCE MAPS
apples
John
apples
beef
beef
George
in terms of pounds per week) between which John is indifferent,
Since points 2 and 3 are on a higher indifference curve than
point 1, they are bundles which are preferred by John to the
commodity bundles represented by point 1. For reasons which
we will not explore here, indifference curves are generally
assumed to be shaped as those in Figure 1.1, so that the
set of all bundles preferred to those on a particular indif-
ference curve forms a convex set.
The marginal rate of substitution of one commodity
for another refers to the rate at which an individual is
willing to trade one commodity"for another, for small
changes around some point. Clearly this amounts to the
-------
condition that the individual remains on the same indifference
curve, and as the small change is allowed to become infini-
tesimal, the marginal rate of substitution becomes the slope
of the indifference curve at the point. Thus for instance,
the slope of the indifference curve through point 2 in
Figure 1.1 represents (in absolute terms) the rate at which
John is willing to accept more apples in return for less
beef, for a small change in the consumption rates of those
two commodities.
Then the first implication one can derive from the
Pareto criterion is that an allocation is inefficient unless
the marginal rate of substitution between any two pairs of
commodities is equal for all consumers. If it is not, then
it will be possible for one or more consumers to be made
better off without making other consumers worse off by
exchanging, or trading, goods. For example, if two consumers
had different marginal rates of substitution for apples
and oranges, then one would be willing to give up more apples
to obtain an orange than the other would require to give
up an orange. In this situation, an exchange on terms that
lie in between the terms on which each is willing to trade
would make both consumers better off. As long as there are
differences in the marginal rates of substitution of con-
sumers, a preferred allocation in terms of the Pareto criterion
10
-------
can be derived. Thus only an allocation in which the marginal
rates of substitution are equal across all consumers will be
Paretc optimal.
This conclusion can be illustrated in the context of
Figure 1.1. Suppose, for instance, that John is at point 3
while George is at point 4. Since John's indifference
curve at point 3 is steeper than George's indifference
curve at point 4, it follows that George requires fewer
additional apples to compensate him (that is, keep him in-
different) for the loss of a unit of beef than does John,
at the margin. Thus consider points 2 and 5, which are
drawn such that the same increase in beef and decrease in
apples is represented by either the move from point 3 to
point 2 or by the move from point 5 to point 4. It then fol-
lows that a trade exists in which George gives some beef
to John and John gives some apples to George which results
in John's moving from point 3 to point 2 while George moves
from point 4 to point 5. But points 2 and 3 are on the
same indifference curve for John, so the trade makes him
just as well off as before, while the trade moves George
onto a higher indifference curve. Whenever such a possible
trade exists, the given allocation does not meet the Pareto
criterion for efficiency. But consideration of the diagrams
indicates that such a trade can be found whenever the slopes
11
-------
of the indifference curves of two individuals are not equal;
that is, whenever their marginal rates of substitution are
not equal.
With exactly the same argument, one can analogously
demonstrate that the marginal rate of substitution between
two inputs must be equal across all firms, where the mar-
ginal rate of substitution refers in this context to the
rate at which the firm can substitute one input for another
at the margin while holding output constant. The third
general conclusion which can be obtained from the Pareto
criterion is that the marginal rate of substitution of each
consumer between two commodities must equal the marginal
rate of transformation between the commodities, where the
latter refers to the rate at which the output of one commo-
dity can be increased when the output of another is decreased,
holding the output of all other commodities constant.
The last conclusion can be understood by superimposing
the indifference curve an individual is on onto the produc-
tion possibilities frontier for the commodities, where the
frontier defines all combinations of the two commodities
which are possible to produce, given the output level of
all other commodities. In Figure 1.2 (next page) we draw
the production possibilities frontier for apples and beef
and superimpose an indifference curve. Suppose that society
12
-------
Figure 1.2 THE PRODUCTION POSSIBILITIES FRONTIER FOR APPLES
AND BEEF AND A REPRESENTATIVE INDIFFERENCE CURVE
apples
production possibilities
frontier
beef
is producing efficiently — that is, is on the production
possibilities frontier — 'but the marginal rate of substi-
tution between the two commodities is not equal to society's
marginal rate of transformation. Such a situation is de-
picted as point 1 in Figure 1.2. Then since it is possible
for society to move to a point such as point 2, this individual
can clearly be moved to a higher indifference curve main-
taining the consumption bundles of all other consumers. Thus,
economic efficiency requires that the slope of the production
possibilities frontier — that is, the marginal rate of
transformation — be equal to the marginal rate of substi-
13
-------
tution, for each pair of goods. This last condition means
that for society's resources to be allocated efficiently,
each consumer must be just willing to trade goods on the
same terms as society is capable of producing more of
one by reducing the production of the other.
The connection between these -conclusions and competi-
tive markets is that the three general conditions, which
we have argued are necessary conditions for an allocation
to be Pareto optimal, would be fulfilled if resources were
allocated by freely operating competitive markets in which
no single consumer or firm can affect prices by his own
actions. The first condition, that the marginal rate of
substitution between all pairs of commodities be equal for
all consumers, will be met as long as all consumers face
the same prices for all commodities. Each consumer who
uses his income to buy the best possible bundle of commo-
dities, will choose a bundle for which his marginal rates
of substitution equal the ratios of prices for all pairs
of goods. If all consumers face the same prices, then they
will all choose bundles which yield the same marginal
rates of substitution. Similarly, the second condition
will be met if firms face the same prices for all inputs.
In their desire to minimize costs for any level of produc-
tion, firms will choose combinations of inputs where the
14
-------
marginal rate of substitution between any two inputs is
equal to the ratio of the prices of the two inputs. If
all firms face the same prices for inputs/ then they will
all use inputs in amounts which yield the same marginal
rates of substitution.
Whereas the first two conditions for Pareto optimality
insure that the bundle of commodities produced by society
as a whole is allocated efficiently among consumers and
produced efficiently by firms, the third condition, that the
marginal rate of substitution of consumers is equal to the
marginal rate of transformation for any pair of goods, in-
sures that the proper mixture of commodities is produced.
This last requirement is attained by the movement of resources
among industries until the value of the marginal product
(the extra product which can be produced with one extra
unit of some input) of each input is equalized across all
industries. To see how competitive markets will ensure
that this condition is met/ consider an economy with
two commodities, A and B, and two inputs of production,
X and Y. We will use MP^ to denote the marginal product
^~ ^™ *L
of X in the production of A; thus it represents the extra
quantity of A which can be produced by shifting one extra
unit of X into the A industry. A competitive firm
will hire inputs up to the point where the value of the
15
-------
marginal product is equal to marginal cost; thus X will
by used in the A and B industries to the point where:
MCA = P,
'MPA = P,
MCB = P
Dividing these two equations implies that the ratio of the
prices of A and B is equal. to the ratio of the marginal
products of X in the production of B and A, which is society's
marginal rate of transformation. This latter equality
follows from the fact that if one unit of X is taken from
the prpduction of A and used to produce more B, the decrease
in A will be MPA while the increase in B will be MP® .
A
It is useful, for the discussion which follows, to
translate this last efficiency condition into terms of
supply and demand. Consider the equilibrium of supply and
demand in the markets for A and for B:
Figure 1.3 COMPETITIVE EQUILIBRIUM IN THE A AND B INDUSTRIES
^B
16
-------
where P and Q represent the equilibrium price and quanti-
A A
ty for commodity A and similarly for B. Since a firm opera-
ting in a competitive industry will expand output out to the
point where price equals marginal cost, the competitive
supply curve for an industry indicates, for each level of
industry output, the marginal cost of each producer in
the industry. Since individual consumers cannot affect
Pe and P® , each consumer will purchase A and B up to a
A B
point at which the marginal rate of substitution equals
the ratio Pe/Pe • Thus if the prices of the commodities
are set by competition and consumers are free to buy as
much or as little as they like, the marginal rate of substi-
tution of each consumer will equal MC /MC , the ratio of
O A
the marginal costs of the two commodities, which we saw
above equals the marginal rate of transformation between
the two commodities. Thus determining the rate of production
of different commodities by the test of whether firms can
survive leads to the socially optimal mixture of outputs,
and leads to the point on the production possibilities fron-
tier which is socially efficient.
Thus far we have suggested a definition for efficiency -•
Pareto optimality — and sketched the theoretical arguments
which lead to the conclusion that economic efficiency will
be attained if the decisions on how much of each commodity
17
-------
to produce, how to produce it, and who will get to consume
it are all left to the decentralized mechanisms of competi-
tive markets. Thus, a competitive equilibrium has the
economic efficiency, and we know that no change can be made
in a competitive allocation which does not move at least
one consumer to a lower indifference curve; in terms of
the Pareto criterion, there can be no improvements over a
competitive equilibrium.
It does not follow, however, that a competitive equili-
brium is necessarily a good allocation, if questions of
equity or fairness are considered. The competitive process
simultaneously determines the prices of factor inputs,
including labor of all types and all physical assets such
as land, capital goods or raw materials. Thus, an income
distribution is determined along with the allocation of
inputs to the production of alternative output. Far from
being optimal, the distribution may seem completely unrea-
sonable to some. There are, in fact, generally an
infinity of different Pareto optimal allocations, each
implying a different income distribution. That is, there
are many allocations of resources for which it is true that
a reallocation can make some one better off only by making
someone elso worse off. But, each such allocation has a
different distribution of income corresponding to it. This
18
-------
might seem to destroy the case for allocation by competitive
markets in so far as one is concerned with income distribu-
tion and dissatisfied with the resulting distribution. However,
there is a converse theorem which suggests why it may still
be desirable to use competitive markets. One can also prove
that, under fairly general conditions, any Pareto optimal
allocation can be attained as a-competitive equilibrium
for some particular distribution of endowments of ownership
rights to assets. Thus if one is unhappy with the existing
distribution of income, that "problem" can be corrected by
redistribution while maintaining the efficiency properties
of the competitive process.
Though it ignores some difficult technical problems/ the
foregoing discussion summarizes the theoretical underpinnings
of the presumption, shared by most American economists, that
competitive markets provide a desirable resource allocation
mechanism. It follows that if the income distribution is
unsatisfactory, the competitive mechanism should be retained
and direct transfers of income should be employed to bring
about distributional equity.
C. APPLICATIONS AND PROBLEMS
While it may not be readily apparent, these issues
are directly relevant to the questions at hand. First, the
19
-------
basic benefit-cost methodology, which we describe below,
depends crucially on the theoretical underpinnings just
discussed. We are concerned here with analyzing the benefits
and costs of solid waste disposal from a point of view
which should, at least, have the welfare of society as a
whole as the ultimate criterion. The question of how the
welfare of society should be evaluated thus arises, and the
Pareto criterion provides a tool by which rankings can be
made which are as free from one's own political, social,
and moral values as seems possible.
Consider then a government project which "takes" certain
commodities as inputs, processes them in some way, and then
"gives" the resulting outputs back to members of society.
Assume that we have an exhaustive list of all the costs
and benefits of the project. Assume also that the purpose
of the project is to supply those outputs, not to redistri-
bute income or some other purpose. How should these inputs
and outputs be weighed? In particular, how can it be
determined whether the overall value of the outputs is
greater than the overall value of the inputs to society as
a whole? The economist's answer is to weigh the outputs
and inputs at their competitive market prices. The value
of the project is then completely determined by the dif-
ferential, if any, between the total market value of inputs
20
-------
and outputs. Note that in this situation the criterion
used by the government is identical to that which would
be used by a private firm: is the project "profitable"?
There are two classes of complications which arise
which make it difficult if not impossible to use this simple
benefit-cost procedure. One class of problems arises where
the output is not sold in a competitive market. Indeed/ the
typical case is where the government is the sole producer
of the commodity or commodities. In that case, the govern-
ment sets the price itself, and indeed the relevant policy
question becomes not just whether to produce, but also at
what price and at what quantity. Furthermore, this situation
in which the government is a monopolist arises not as a
coincidence, but usually because the government preempts
the field for some compelling reason. One such reason is
the case of "economies of scale" in which higher rates of
output lead to lower average cost (holding input prices
fixed). In such a situation, which is often termed a "natural
monopoly", it can be shown that the socially optimal level
of output occurs where the price of the output is set equal
to marginal cost, if it is socially optimal to produce the
commodity at all. However, it is important to realize that
those conclusions about the optimal output level and the
optimal pricing scheme for the "natural monopoly" are derived
21
-------
in a context where it is assumed that the government can
apply "lump sum" taxes to individuals in order to finance
the deficit which such a project will necessarily incur,
since marginal cost will be below average cost if average
cost is falling. In order to meet the strict Pareto cri-
terion of making everyone at least as well off as previously
(when no output of this commodity was produced), it is
necessary to assume that the government knows the preferences
of all citizens and can cover the surplus entirely from
lump sura taxes which are levied on consumers of the product
who have positive "consumer surplus"; that is, who benefit
by being able to buy the good in question at marginal cost
rather than not having the commodity available at any price.
The next chapter introduces the concept of consumer surplus
more rigorously and discusses some of the complications that
arise in trying to measure it.
The second class of problems which make it difficult
to compute social "profit" as a guide to the desirability
of a project arises when inputs are not purchased in com-
petitive markets. Here the typical case is the appropri-
ation of certain inputs as privately "free" goods, such as
the use of clean air or water to dispose of pollutants,
or the "consumption" of a quiet neighborhood by the produc-
tion of noise. Historically, such inputs have been freely
22
-------
used by firms (and consumers), although their use has often
been regulated. These are examples of what are termed
"externalities", in which the production or consumption of
one party affects the welfare of others through effects
which are not already reflected directly through the market
mechanism.
The term externality denotes the distinction between
those costs imposed on society for which the firm does not
pay as opposed to those which are "internalized" by the firm.
Although it may seem artificial to construe external effects
as the use of inputs/ that interpretation provides insight.
Suppose one thinks of an input, say coal, which the govern-
ment decides to buy from producers and supply free to all
those who request it. Obviously firms would have no incen-
tive to use the coal efficiently. Coal would replace other
fuels in many uses, and generators would be adopted which
used coal wastefully but which were relatively cheap to
construct. The amount of coal demanded would be constrained
only by the costs of storing it and the limits of substituting
it for other fuels. Firms would use it out to the point
where its marginal value was zero. Such an analogy can
be applied to the problems of air and water pollution and
other similar externalities. Clean air can be considered
an input, one which happens to appear directly in the pre-
23
-------
ference structures of consumers, but can also be used as
an input by firms. The input "clean air" can be substituted
for filtering devices and other pollution control equipment.
When firms are free to use as much clean air as they wish,
without being charged for its opportunity cost, there is
no mechanism to lead them to a socially optimal level of
"consumption" of clean air.
This distortion in allocation can be usefully represented
in the competitive equilibrium between supply and demand,
where we represent the externality simply as a cost imposed
on society which is in addition to the opportunity costs
of the resources purchased by the industry which are reflec-
ted in the supply curve. Thus in Figure 1.4, we illustrate
the competitive equilibrium for an industry which produces
commodity G and which imposes external costs on society.
We use CE to represent external costs at each level of out-
put. That is to say, CV(Q) is a function which indicates
r. G
at each level of QG the extra social cost imposed for an
additional unit of output which is not internalized to the
firms in the industry. Thus €„ represents the difference
Ci
between the industry supply curve which would result if
the firms were charged for all costs resulting from their
production and the existing supply curve which reflects only
private costs which are internal to the firms.
24
-------
Figure 1.4 THE SUPPLY AND DEMAND EQUILIBRIUM WITH AN EXTERNAL
COST PRESENT
We denote the competitive equilibrium by Qe and P®.
G G
At the competitive equilibrium, the opportunity cost of
the marginal unit of G, which is SG + CE , is higher than
the value to consumers of the marginal unit.
Thus the presence of an external cost implies that
the competitive equilibrium does not have the optimal pro-
perties discussed above. There are several corrective
measures which have been suggested. The first, suggested
by A. C. Pigou [2 ], is to impose a tax on those firms
imposing the external cost on others, with the tax equal
to the level of external cost at the margin. That is,
25
-------
referring to the external cost as "damage" not accounted
for by market, the level of the tax would equal the marginal
damage rate — the extra damage produced by an extra unit
of output — in equilibrium. The purpose of the tax is
to raise the supply curve sufficiently to generate the social-
ly optimal level of output, by making the cost of the product
to the consumer equal to its social marginal cost; such a
tax is often referred to as a "Pigouvian tax", the purpose
of which is not to collect revenue but to bring about effi-
ciency .
It is important to emphasize the point that the Pigouvian
tax is set equal to marginal damage in the post-tax equilibrium.
Consider the case where marginal damage increases as output
is increased, as in Figures 1.5 and 1.6 (next page). Where
a tax is imposed, we must distinguish the price paid by
2
consumers (designated Pr) and that received by producers
(designated P_). In Figure 1.5, the imposition of a pro-
G
perly computed Pigouvian tax is illustrated; the level of
2 1
the tax, T = PQ - PQ , is equal to C at the equilibrium
T
output level Q . In Figure 1.6, we illustrate the equili-
G
brium which results when the Pigouvian tax is set equal
to marginal damages CE measured at the original, competitive
equilibrium. Since C is assumed to be increasing, the
£
resulting equilibrium is at an output level which is below
26
-------
Figure 1.5 A PIGOUVIAN TAX SET EQUAL TO MARGINAL DAMAGE
IN EQUILIBRIUM
PI >^ —-^
SG+CE
Figure 1.6
A PIGOUVIAN TAX SET EQUAL TO MARGINAL DAMAGE
IN THE ORIGINAL COMPETITIVE EQUILIBRIUM
SG+CE
27
-------
T
the optimal level Q and at which marginal damages are
\~t
less than the imposed tax. In this case, the measurement
of damages at the original competitive equilibrium rather
than at the new equilibrium results in overkill; the social
value of an extra unit of C is new greater than its oppor-
tunity cost. The reader can convince himself that output
will be greater than the optimal level, in the case of
increasing marginal damages, if the tax is set equal to
average damage in equilibrium rather than equal to marginal
damage. Similarly if marginal damage decreases as output
is increased, the imposition of a tax equal to average
damage in equilibrium will result in an output level below
the socially optimal level.
There is no presumption that the receipts of a Pigouvian
tax must be used to compensate those who are damaged by
the production of the product. Indeed, if marginal damage
is increasing, the total tax receipts will be greater than
total damage (in equilibrium) while if marginal damage is
decreasing, the total tax receipts will necessarily fall
short of total damage. In fact, Pigou was not concerned
with questions of distribution, and the notion of Pareto
optimality was not yet posed when he wrote. Pigou simply
noted that when all other prices are assumed fixed, the
maximum social value of all output is maximized when price
28
-------
is set equal to social marginal cost in all industries.
Two other schemes exist for "internalizing" the external
cost of some activity. The first is to make the firm legally
liable for the damage imposed on others, so that those harmed
can sue the firms for their damages. Note that the imposi-
tion of liability automatically provides compensation for
those damaged. However, there are two significant problems.
The first, obvious problem is the administrative costs which
arise in the frequent situation where the damages are im-
posed on many — possibly millions — of persons while the
damage to each may be rather"small, particularly in comparison
to the social (and private) costs of adjudicating the claim.
Second, legal procedures may be too inflexible to provide
an appropriate response where the marginal damage produced
by the output of each firm does not add up to the total
damage incurred. That is to say, consider the case where
marginal damage increases. In that case, the damage imposed
by the activity of a particular firm is less, the smaller
the production of other firms. Legal thinkers may accept
the defense that "our firm cannot be held legally liable
for the extra damage it imposes since such damage arises
only because of the activities of other firms". While it
is true that the movement to the optimal output level requires
that each firm should be treated as the "marginal firm", as
29
-------
the Pigouvian tax implicitly does, this notion undoubtably
contradicts what seems to be common sense.
The third reorganization which accomplishes the objective
of attaining an efficient allocation of resources is to
bring the production resources of the offending industry
(or firm) and those assets of the parties damaged under
common ownership. If the external cost is based on a local
effect of some sort, for instance, the joint ownership of
the land on which the offending activity takes place and
that on which the damage occurs will automatically lead
the owner to restrict the offending activity to its opti-
mal level. The reason, of course, is that the damage is
automatically "internalized". While joint ownership may
be an effective solution to some cases of external costs,
it is also a useful device for purposes of analysis. For
instance, Baxter and Altree [ 1 ], provide considerable
insight into the problem of optimally controlling airport
noise by repeatedly asking "what would a joint owner of
the airport and of the surrounding land affected by the
noise do?" Clearly such a solution may produce problems
of local land monopolization and problems of inefficient
land management, produced by insufficient specialization,
in most cases. Furthermore, such a solution would not
be politically feasible in general, since it would typically
30
-------
necessitate increasingly centralized control of land use.
Nevertheless, joint ownership should be recognized as a
possible solution for extreme cases.
D. SOLID WASTE MANAGEMENT
It will be quite instructive to see how the solution
of joint ownership would work in the solid waste context. Let
us assume that the externalities associated with landfills
extend only within a specified area of reasonable size
(this may not be appropriate when the externalities associ-
ated with transportation to and from the landfill are con-
sidered) . Suppose that all the land within that area were
purchased, at market value, from its owners before the
possibility of a nearby landfill became apparent (also,
before the possibility of profits from a government land
acquisitition became apparent). That is, the land value would
have to be determined in the absence of any affect from the
prospect of government action.
Suppose, then, that all of the land were transferred
to a "sanitary landfill authority." Suppose that this
authority were to be paid for disposing of solid waste,
the payment coming from the municipality, and representing
the opportunity cost to the municipality of otherwise dis-
posing of the solid waste; that is, the total cost to the
31
-------
municipality of disposing of the solid waste in the next
best alternative. Suppose the municipality were then told
to operate as if it were a private firm, with an objective
of maximizing the value of the firm's assets, namely its
holdings of land. The value of each parcel of land is equal
to the present discounted value of its potential rents in
all future periods.
In deciding how much waste to accept — and whether
to accept it at all — the authority would be led to consider
the effect of using part of the land for a landfill on the
value of the surrounding land. Given the assumptions that
the external costs are contained within the region owned
by the authority, and also the assumption that all external
costs are perceived and understood by potential users of
the land, the possible decrease in rents in the surrounding
land provides a precise evaluation of these external costs.
The rental value of the adjoining land thus reflects the
social evaluation of the external costs for a particular
year; the authority must weigh the possible decreases in rent
over all time. This is accomplished by considering the dis-
counted value of rents, which is precisely the market value
of the land.
The conceptual scheme by which the authority would
arrive at the optimal use of the land and the optimal size
32
-------
and design of a landfill on part of the land would then be
the following. The authority would announce a proposed
plan for the use of part of the land as a landfill. The
proposed plan would include plans to "buy" a certain amount
of solid waste from the municipality for each year until
the landfill site were exhausted. The payments for "purcha-
sing" the solid waste would then represent revenues for the
authority; direct costs of disposal, such as equipment and
labor costs, would be subtracted from these revenues to give
yearly net "profit". These "profits" would then be discounted
back to the present using the applicable rate of interest
(the correct computation of which will not be discussed here).
If there were no external costs to be considered, this would
complete the conceptual scheme. The municipality could
set up many different authorities, giving each the mandate
to maximize the "value" of the authority — namely, the
present discounted value of its net receipts — and letting
the various authorities bid on the solid waste to be disposed.
However, since there may be external costs, which may
decrease the value of the land adjoining the landfills,
these costs must also be considered in arriving at a socially
optimal system for solid waste disposal. Within the scheme
suggested above, however, this can be accomplished very
simply. The authority must look not simply at the value of
33
-------
the land to be used as a landfill but also at the value
of the surrounding land which is affected by the landfill.
Thus, any decrease in the value of the surrounding land must
be subtracted from the "present value of the authority"
to arrive at the net value of the authority. Of course, if
the landfill were to increase the value of surrounding land,
this would be added to the authority's balance sheet.
Within the assumptions mentioned, an authority which
maximized the total value of the land it controls would
naturally choose the plan which was socially optimal. This
is predicated, of course, on the assumption that the "price"
it receives to dispose of solid waste reflects the opportunity
cost of disposing of the solid waste by some other means,
either by another authority or in some other fashion.
If that price is appropriate, however, the decisions of the
authority would be socially optimal. If the "net profits"
from the landfill activity were not positive at that price,
it would be a signal that this landfill site should not be
used at all, and a profit maximizing authority would choose
not to accept any solid waste. If profit from the landfill
activity were positive but were not large enough to balance
the losses the authority would experience on the "land
holding" activity, this would correspondingly signal that
from the social point of view, the landfill site should not
34
-------
be used. If, however, the landfill site should be used
intensively because the next best alternatives have very
high opportunity costs, the authority would naturally be
led to increase its landfill activity in the face of a "high
price" for disposing of solid waste.
The viability of such a scheme depends very crucially
on the assumption that such an authority would work efficiently
and would be capable of predicting the effect of the landfill
activities on the present and future rental value of the
surrounding land, of course. It also assumes that by simply
establishing such an authority with the mandate to maximize
the value of all of its land, an organization would arise
which would actually pursue that objective and not be diverted
to an alternative objective dictated by bureaucratic forces.
Nevertheless, the significant point is that our conceptualized
authority would make socially optimal decisions if it pursued
the wealth maximization objective. This provides the standard
by which to evaluate the plan of an actual municipality.
The crucial aspect of our conceptual authority, of
course, is that the external costs have been "internalized"
by giving the authority the ownership of all adjoining land
which is affected by the landfill activity. If such a scheme
is infeasible, the socially optimal method of deciding how
and where to dispose of solid waste is to mimic the behavior
35
-------
of such an authority. Since such an authority "pays" for
all of its "inputs", the efficiency criteria described above
are automatically met.
In order for a municipality to duplicate the decisions
of such an authority, it must be able to predict the effects
on surrounding land values of the landfill activities. This
report is directed toward advancing the state of the art in
that area. Of course, the conclusion that changes in land
values provide an appropriate index of external costs depends
on a series of assumptions, some of which appear partially
invalid. Thus we also consider dix-ect methods of evaluating
such external costs in the next four chapters.
A proper benefit-cost analysis of a municipal project
such as a sanitary landfill must include external costs in
the costs of the project. However, it must be emphasized
that simply including such external costs does not mean that
the project will meet the strict Pareto criterion. If the
benefit-cost ratio is greater than 1 with the inclusion of
all costs and benefits, it means that there is a conceptually
feasible tax scheme to pay for the project and to compensate
all those who bear costs of the project which will insure
that all members of society are better off with the project
than without. However, the standard benefit-cost analysis
blatantly adds costs and benefits across different citizens.
36
-------
A cost to one citizen and an equal benefit to another
are assumed to "cancel out," whereas the strict Pareto
criterion does not allow these two effects to be compared.
In this introductory chapter/ we have provided a super-
ficial introduction to the theoretical basis of benefit-cost
analysis and to the treatment of external costs. The next
chapter analyzes several of the crucial theoretical questions
in more detail and with greater rigor.
37
-------
REFERENCES
1. Baxter, W.F. and L.R. Altree. Legal Aspects of
Airport Noise. Journal of Law and Economics.
Vol. XV, No. 1 (April 1972)".
2. Pigou, A.C. The Economics of Welfare, 4th edition.
Macmillan & Co., Ltd., London, 1932.
38
-------
CHAPTER II
MEASURING EXTERNALITIES IN THEORY
A. INTRODUCTION
This chapter surveys and evaluates various approaches
to the measurement and valuation of external effects.
The chapter deals primarily with the problem of defining
an ideal measure and discusses some of the problems
which can arise when one attempts to obtain estimates
of the proposed measures. Empirical work on the
measurement of externalities is discussed in Chapter III.
An externality can be defined as follows. Let
W1, the welfare or production variable of the ith
unit in the economy/ be given by the welfare (produc-
tion) function
i _ i. 1 j N 1 j M
> *~- i \X /•••/ X /•»•/ X j 2 /•••/ Z j * • * f Z
where x-5 = x^ / . . . , x^, is the consumption bundle which
1 -K
is under the control of the jth consumer and z = z ..... z-3
J — 1 L
is the input bundle which is under the control of the
jth production unit. Then the jth unit's consumption
of commodity k (or use of input k) imposes an exter-
nality on unit i if and only if
39
-------
AW AW1
1
0, or - r X 0.
AzP
K-
In words, unit j imposes an externality on unit i if
and only if a change in a variable under the control
of unit j (commodity or production input) results in
a change in unit i's utility or output level.
The above definition is formulated in terms of
finite differences since a small (marginal) change
in a variable under the control of unijb j may not have
an effect on unit i, while larger discrete (infra-
marginal) changes may have an effect. A good discus-
sion of this and related issues in the definition of
externality can be found in Buchanan and Stubblebine
[9] and Arrow [2].
This chapter will thus be concerned with theore-
tical issues that arise when one tries to value changes
* •
in W1 which are attributable to changes in a x£ or
JC
4-
Approaches to the Valuation of External Effects
A number of approaches to the valuation of ex-
ternalities have been used or suggested since the pro-
blems encountered vary with the type, source, and loca-
40
-------
tion of the external effect being considered. No one
approach appears to be best in all situations. Two
general approaches are discussed in this chapter. In
Section B we consider the possibility of obtaining more
or less direct valuations of external effects based
on individuals' willingness to pay for the elimination
or prevention of an external effect. A distinct dis-
advantage of the direct measurement approach is that
it requires more information than is available in most
cases where external effects are present. In parti-
cular/market prices, or some good approximations thereof,
are required. Such prices are, however, extremely
difficult to obtain for most externalities associated
with solid waste disposal activities. The theoretical
considerations which lead to this information problem
are discussed in Section C.
Where externalities are associated with parti-
cular sites or locations, as is the case with solid
waste disposal activities, it may, however, be possible
to infer measures of willingness to pay for the removal
of most of the externalities associated with a particular
activity in a somewhat less direct manner from differences
in equilibrium land rents between two sites identical
in every respect except the activity. This approach
41
-------
to measurement is discussed in Section D.
Criteria for Ideal Measures of External Effects
Since, as will be shown in Section B, it is in
general possible to define individual willingness to
pay for a particular action in a large number of ways,
we devote the remainder of this section to a discussion
of criteria which can be used to evaluate the desira-
bility and usefulness of alternative definitions. Such
criteria are useful since there may be more agreement
among economists on the appropriateness of certain
criteria than on the usefulness of any one particular
definition. This approach has recently been used by
both Harberger [23] and Mohring [42] to defend parti-
cular measures of willingness to pay.
Harberger [23, p.785] suggests that economic
measurements be based on the following set of postulates;
(a) the competitive demand price for a particular
unit of any commodity measures the value of that unit
to the demander;
(b) the competitive supply price for a particular
unit of any commodity measures the value of that unit
*It is, of course, possible that no definition or
measure exists which satisfies a given set of cri-
teria, or that a large number of measures satisfy the
criteria.
42
-------
to the supplier;
(c) when evaluating the net benefits or costs of
a given action (project/ program, or policy), the costs
and benefits accruing to each member of the relevant
group (e.g./ a nation) should normally be added without
regard to the individual (s) to whom they accrue.
These postulates can be rationalized as follows.
The first two postulates are based on the efficiency
of market allocations under competitive conditions/
which was discussed in Chapter I. Measurements based
on these postulates would thus extend hypothetical
competitive market valuation to non-market activities.
It can thus be argued that the resulting allocation of
resources will be efficient since the attributes of the
competitive market mechanism are extended to the non-
market activity (e.g., the marginal benefits of the acti-
vity are equated with its marginal costs). It should
be noted, however, that if imperfections exist in the market
sector second best issues may arise (e.g., equating
marginal costs to marginal benefits in the non-market
sector does not assure efficiency). This issue is dis-
cussed in greater detail in Section B. The third pos-
tulate involves a distributional value judgment and cannot
43
-------
be evaluated on purely scientific grounds. It can, however/
be interpreted as representing a politically approved
value judgment in the U.S., since the Congress in the
Flood Control Act of 1936 required that projects be under-
taken if and only if "the benefits to whomsoever they may
accrue [are] in excess of the estimated costs." (Quoted
in Marglin [34, p.16]).
Harberger uses these postulates to defend measures
of welfare effects (effects of activities which distort
resource allocation) based on Marshallian consumers surplus,
which we define below. These postulates do not, however,
define a unique measure of welfare effects since they may
be consistent with more than one measure if prices change
due to the activity whose effect we are attempting to
evaluate. This, of course, is a problem with all measures
based on consumers surplus and is discussed in detail in
Section B. It may also be difficult to evaluate measures
which are not based directly on market valuations, such as
measures based on differences in land rents or direct sur-
veys of willingness to pay, with these postulates. The
*It should also be noted that properties of welfare mea-
sures which hold for the individual need not, and in fact
do not, hold when summed, or aggregated in some other
manner, over individuals. This issue is discussed in
greater detail at the end of Section B.
44
-------
first two postulates appear to be directed at restricting
the techniques used to construct a particular measure/
rather than restricting the properties an acceptable
measure may possess.
A more general set of criteria, without these short-
comings, has been suggested by Mohring [42, p.349]. He
would require:
(1) If an economic activity results in several
effects (say externalities) the measured change in wel-
fare attributed to the activity should be independent
of the order in which the effects are regarded as having
occured.
(2) The maximization of net benefits from an acti-
vity, including measured external benefits and costs, should
yield necessary conditions which are consistent with
Pareto Optimal resource allocations.
(3) In comparing the value of two alternative acti-
vities, say A and B, to an individual the measure should
value A more highly than B if and only if the individual
prefers A to B.
*Mohring is concerned primarily with the welfare effects
of price changes and taxes. We have therefore modified
his language to include other welfare effects. This does
not materially modify the criteria, however, since most
welfare effects can be interpreted as a change in the
price of some appropriately defined commodity.
45
-------
(4) In aggregating the measure over individuals,
the weights attached to individuals should depend only
on their utility levels, and not on their incomes and
the market prices which they face.
The first criterion assures that a unique measure
will be assigned to any activity generating several ex-
ternal effects. The second criterion ensures that the
measure will embody information relevant to efficient
(in the sense discussed in Chapter I) decision making by
a policy maker. Criterion (3) ensures the Pareto Opti-
mality of decisions involving discrete or distinct alter-
natives. For example, if the measure assigns values to
alternative's A and B relative to a third alternative C/
the measure should assign a higher value to A if and only
if A is preferred to B. Most conceptual measures of
welfare effects are defined with respect to particular
individuals. Reaching a decision about a particular
activity will in general require some form of aggregation
of such measures for the affected group. Only if such
aggregates satisfy Criterion (4), and if individual pre-
ferences satisfy some additional restrictions/ is it pos-
sible to state that an increase in aggregate measured
welfare also implies an increase in a well defined, indi-
vidualistic, social welfare function. This issue is
46
-------
discussed in greater detail in Section B. The conditions
and Criterion (4) turn out to be extremely restrictive,
implying a lack of generality in commonly used measures
of welfare effects.
VThich criteria are most important must, however, be
determined by the judgment of the user of the welfare
measure, and the context in which it is to be applied.
The criteria do, however, provide a set of standards
against which alternative measures can be judged, and
should aid in eliminating "weak" measures. The situa-
tion is similar to that encountered by producers of index
numbers. An "ideal" index does not in general exist.
Alternative indices are therefore judged on how well they
perform relative to a limited set of criteria.
B. WILLINGNESS TO PAY: TRADITIONAL APPROACHES TO THE
MEASUREMENT OF WELFARE CHANGES.
Economists have been concerned with the problem of
valuing economic activities for which no market exists,
and the closely related problem of valuing the welfare
gain (or loss) a consumer obtains when he is given the
opportunity to purchase some commodity at a lower (or
higher) price than some reference price, since the middle
A
of the 19th century. In this section we discuss this
*A problem of this type was first discussed by Depuit [16]
in 1844. He suggested the use of a consumer's surplus
47
-------
literature and its relevance to the problem of valuing
externalities associated with an economic activity. We
consider first the problem of estimating welfare costs
or gains associated with price changes. We then discuss
the relevance of this problem to the measurement of ex-
ternal effects.
Measuring the Effect of Price Changes on Individual Welfare
The typical problem addressed in this literature can
be stated as follows: What is the welfare gain or loss
to an individual when the price of a commodity, say com-
modity x, changes from p to p*? This problem is of general
••* JV
interest since a broad range of economic changes can be
analyzed in this framework. For example/ what is the
welfare cost of an excise tax on commodity x? Or what is
the social value of a new bridge for which no tolls are
charged? (The price of its services may be considered to
be arbitrarily large before it is opened and to be zero
afterward.}
The problems inherent in assigning a value to the
changes considered above can be readily illustrated using
indifference curve analysis. Assume we are given a consumer
measure as a means of determining the social value of the
services of a bridge. After Depuit the problem did not,
however/ receive further notice ur '1 the more recent work
of Marshall 136] and Retelling 12; .
48
-------
w-,
. t t
q1'
Figure 2.1 COMPENSATING VARIATION AND
EQUIVALENT VARIATION MEASURES OF THE BENEFITS
FROM A REDUCTION IN THE PRICE OF x
49
-------
who has a money income of I which he can use to purchase
two goods, x and y. The consumer's budget constraint,
expressed in terms of good y, is thus given by
qy = i/py - px/py • qx
where q and q are the quantities of x and y respectively.
« Ji
The consumer's utility is assumed to be given by the
(quasi concave) utility function
W = W(q , q ) .
The consumer's budget constraint is given, for relative
prices P = p /p and money income I , by the line
I /p I /p in Figure 1. If the consumer maximizes
his utility subject to this budget constraint he will be
in equilibrium at point A. At point A we can assign an
arbitrary number, W , to the consumer's utility or wel-
fare. This level is constant along the curve labeled
W by the definition of indifference curver. Now sup-
pose that the consumer is given the opportunity to purchase
x at a lower price, p'/p = P'. His budget constraint
x y
is now given by the line I /p I /p', and there is a
o y ox
new equilibrium at B with a new welfare level, to which
50
-------
we can assign any arbitrary number VL > W . Since the
choice of W and W, is inherently arbitrary, so is their
difference. A measure of the change in welfare based on
the difference W, - WQ would thus have little meaning.
We must employ instead some meaningful unit of measure-
ment, such as units of y or money income.
There are, however, a large number of ways in which
a dollar value can be assigned to this change in welfare.
The "conventional" approach, first suggested by Depuit [16]
and Marshall [36], is to use the change in consumer's
surplus attributable to the price change. Depuit measured
consumer's surplus by the relevant area under the con-
sumer 's ordinary demand curve. Marshall's measure has
been given a similar interpretation. There is, however,
controversy about his interpretation (see, for example
Hicks [26] and Patinkin [47]). The relationship between
areas under demand curves and measures of willingness to
pay is discussed in greater detail below.
Much of the recent theoretical discussion of wel-
fare measures has, however, centered on two definitions
which were suggested by Hicks [27]. (A more recent and
extensive discussion is given in Hicks [26]). The mea-
sures are:
(1) The compensating variation, defined as the
51
-------
maximum amount of money that the consumer would be willing
to pay in his initial position for the opportunity of pur-
chasing commodity x at relative price P1.
(2) The equivalent variation, defined as the mini-
mum amount of money the consumer would be willing to accept,
after the price has been changed, as compensation for
returning to the original price.
The compensating variation is given by I - I in
o c
Figure 1, since with an income of I and relative price
C
P1 the consumer will be in equilibrium at B' on indifference
curve W . The consumer is thus indifferent between facing
the relative price P with income I and facing the relative
o
price P1 with income I . He is therefore willing to
c
pay I - I for the privilege of facing prices P1.
o c
The equivalent variation is given by I - I in
Figure 1, since the consumer would require an income of I
e
*These two definitions of willingness to pay can also be
defined in terms of a consumer's indirect utility function
(which gives a consumer's utility in terms of his income and
the vector of prices he faces). Let U(I,P) give the maxi-
mum utility the consumer receives when he has money income
of I and faces a price vector P. Then, the compensating
variation measure of willingness to pay is given by the
quantity CV which satisfies tKI-CV,?1) = U(I,P), where P is
the initial price vector and P' is the price vector which
prevails after the effect, which we are trying to measure,
has occurred. Similarly, the equivalent variation measure
is equal to the quantity EV which satisfies
U(I,P') = U(I+EV,P).
52
-------
with prices P to reach a point, A1 , at which he is at
least as well off as he was with income I and prices P'.
c
Unfortunately, the compensating variation measure
of any change in welfare will not in general equal the
equivalent variation measure of the same change. The
reason is that the respective variations are calculated
for two different relative prices for x and y. One can
define willingness to pay for a change in economic condi-
tions which change an individual's welfare with respect
to any arbitrary set of relative prices; Hicks merely
selected two sets of prices at. which the changes were
valued. This implies that willingness to pay can be de-
fined in an infinite number of ways, one for each relative
price, each of which may assign a different monetary value
to the welfare change. Criteria, such as those discussed
in Section A, are therefore required for the selection of
"good" measures from this large number of potential mea-
sures.
It has been argued that in any specific application,
only one measure of willingness to pay will be relevant
(see Henderson [25]). For example, if in Figure 1 we
want to know how much an individual with an initial welfare
level of W and income of I would be willing to pay for
o o
a change which resulted in a new equilibrium price ratio
53
-------
of P'f only the compensating variation measure is relevant.
"Mishan [40] has suggested that only the compensating
variation and equivalent variation measures are relevant.
Patinkin [47] points out, however/ that this argument
assumes that the consumer faces a linear budget constraint/
as in the case when prices are determined in competitive
markets. However, if the consumer buys from a discrimi-
nating monopolist, the relative prices he faces will no
longer be independent of his consumption choice, making
other measures of welfare change relevant.
If the competitive pricing assumption does hold,
the compensating variation (or, equivalent variation)
measure has several attractive properties. It is based
on intuitive notions of willingness to pay. It is consistent
with Harberger's first two postulates and Mohring's first
three criteria. The compensating variation can also be
related to the compensated demand curve for x, constructed
by adjusting the consumer's income so as to keep his wel-
fare constant at its level before the price change. This
compensated demand curve is shown in Figure 2 by D D1 .
It can be derived from the indifference curve analysis of
Figure 1 as follows. With an income of I and relative
price P, the consumer maximizes his welfare at A in
Figure 1. This equilibrium is shown as point a in Figure 2.
54
-------
Px J
P'
q
M
I I I
Figure 2.2 DEMAND CURVES AND MEASURES OF THE
BENEFITS OF A REDUCTION IN THE PRICE OF x.
55
-------
When the relative price of x is reduced to P', the consumer
can increase his welfare by increasing his consumption of x
from q'x to q' tx- This new welfare maximum is shown as point
B in Figure 1 and d in Figure 2. By considering the con-
sumer^ consumption of x at all other relative prices
for x in a similar manner, the ordinary demand curve, DD1,
for x can be derived. If, as the price changes, income
is adjusted to keep his welfare at level W , the resultant
demands trace out the consumerls compensated demand curve,
D D' . Thus if the price of x is reduced to P' and income is
oo *
lowered to I (to keep welfare at WQ), an equilibrium consump-
tion q111 results. This is shown as point B1 in Figure 1
J^
and point b in Figure 2.
It can be shown that the compensating variation,
I -Ic, is given by the area under this compensated
demand curve between the prices P and P1 (see Patinkin
147, pp. 87, 88J for a detailed derivation of this result).
This result can be intuitively motivated as follows. The
demand curve D D1 can be interpreted as a function relating
o o
the consumer's willingness to pay for marginal units of
x to his consumption of x (when his income is adjusted
so as to keep his welfare constant at level W).
Therefore, the area under the demand curve up to
q*, for example, represents the consumer's total
willingness to pay for ql units of x (area
56
-------
Oq'ae in Figure 2). The difference between what the
ji
consumer actually pays, i.e., the area under the price
line from 0 to q1 (area OPaq1 in Figure 2) and his
rfC X
total willingness to pay is thus a measure of what the
consumer would be willing to pay for the privilege of
purchasing x at price P when his welfare is maintained
at W . This difference increases by the amount P'Pab
when the relative price at which x may by purchased is
reduced to P'. The consumer would thus be willing to pay
at most an amount equal to area P'Pab for the privilege
of buying x at relative price P1 rather than P, if his
welfare is held constant at W . This, of course, is
the definition of a compensating variation.
Likewise, we can construct through point d in Figure
2 the compensated demand curve, D D.!, which holds constant
the consumer's welfare at level W / the welfare level the
consumer attains when he maximizes his welfare after the
price change. Since D.^ D' can be given an interpretation
analogous to that given D D' , areas under this curve are
o o
measures of the consumer's willingness to pay for x when
his welfare is held constant at W.. . Area P'Pcd thus
represents the minimum compensation a consumer would demand
for returning to P, after the price had changed to P1.
This is, of course, the equivalent variation measure of
57
-------
of the welfare change.
The "conventional" measure of the welfare change is
given by the change in the {Marshallian) consumer's
surplus, area P'Pad. The conventional measure of the
welfare change corresponding to a decrease in the price
of x thus exceeds the compensating variation measure, but
is less than the equivalent variation measure, if x is a
normal good. (In the case of a price increase, the con-
ventional measure is less than compensating variation and
greater than the equivalent variation.)
An important result of Hicks [26] is the proposition that
if the demand for x is independent of income (i.e., if the
income elasticity of x is equal .to zero, implying that
there are no income effects when prices change), the com-
pensating variation measure, the equivalent variation mea-
sure, and the Marshallian consumer's surplus measure are
identical. To see this, consider the reduction in the
relative price of x from P to P' in Figure 2. This change
results in a movement from a to d along the ordinary de-
mand curve, D D1. We then adjust the consumer's income,
so as to maintain his welfare at level WQ/ in order to get
a point on the compensated demand curve. This adjustment
in income will not, however, alter the consumer's consump-
tion of x, if his demand for x is income inelastic. Point
58
-------
d is therefore also on the compensated demand curve.
The ordinary demand curve thus coincides with the compen-
sated demand curves in this case. It follows that wel-
fare measures which are based on areas under these curves
will also coincide.
Thus, if income effects in the demand for x can
reasonably be ignored, willingness to pay for welfare
changes can readily be computed from the areas under ordi-
nary demand curves. Such demand curves can be estimated,
in principle, from observable price and quantity infor-
mation. In cases where income effects are present, the
determination of a consumer's willingness to pay is more
difficult since it requires knowledge of the consumer's
preferences at consumption levels which are not ususally
observed (i.e., the shape of the indifference curve be-
tween points A1 and B, and A and B1 in Figure 1), or know-
ledge of the consumer's compensated demand curve. A
possible approach in this case would be to postulate a
particular form of the utility function and to estimate
the parameters of this function, thereby obtaining esti-
mates of the income effects (see Mohring [42] for a dis-
*It should be emphasized that we are dealing with individual
demand curves here. The use of market demand curves
involves aggregation problems, which are discussed at the
end of this section.
59
-------
cussion of this possibility). This approach appears to
require, however, either very restrictive assumptions
about consumer preferences, or more data than is usually
readily observable.
It is thus apparent that willingness to pay for
a change in prices cannot be sensibly defined in terms
of observable economic variables except in very special
cases. How it should be defined must depend both on the
specific use to which the measure is put, and on the condi-
tions facing the consumer whose willingness to pay we are
attempting to estimate. Only if competitive conditions
prevail and income effects are absent will the Marshallian
measure, based on areas under the ordinary demand curve,
yield unique measures of willingness to pay from readily
observable data.
Measuring the Effect of Externalities on Individual Welfare
We consider next the special problems one encounters
when one attempts to measure and value external costs.
*Welfare measures based on the Marshallian and other mea-
sures of consumer's surplus have, however, been broadly
applied in spite of these conceptual difficulties. Appli-
cations include studies of the welfare effects of various
taxes (see Harberger {21, 22] for example), and the measure-
ment of benefits from government programs (see Prest and
Turvey [49] and the articles in Dorfman [17] for example).
Numerous other applications of consumers surplus to the
measurement of other types of distortion, such as tariffs
and monopoly, are surveyed by Currie, Martin, Murphy and
Schmitz [12].
60
-------
These problems can be readily analyzed in the framework
which we developed above if the externalities are defined
appropriately. In most cases/ an external "bad" can be
considered the absence or unavailability of a commodity
which the consumer desires. For example, if a burning
dump pollutes a neighborhood, the residents of the neigh-
borhood experience a lack of clean air. Thus the resi-
dents' willingness to pay for clean air, given the current
level of air pollution, provides a measure of the external
costs of the pollution. In general, we assume:
EXTERNAL COSTS TO
INDIVIDUAL i OF
ACTIVITY j
EXTERNAL BENEFITS RECEIVED
BY THE INDIVIDUAL WHEN
ACTIVITY j IS CURTAILED
For example, in terms of Figure 1, a movement from, say
B to A would represent a cost to the individual. A move-
ment from A to B would therefore be defined as the corre-
sponding benefit. Our problem is to place a monetary
value on the difference between situations A and B. This
value should ideally be the same whether we view the dif-
ference as a cost (moving from B to A) or a benefit (moving
from A to B).
Let us now carry the analogy to market commodities
61
-------
a step further. When air pollution exists, clear, air
is available to the residents only at a positive price.
On the other hand, if the pollution is eliminated, clean
air becomes available at a price of zero. The problem of
estimating the external costs of, say, pollution, can
thus be formulated in the same framework as the one used
to measure the welfare effects of price changes.
This conceptual approach to the measurement of ex-
ternal effects is illustrated in Figure 3, which shows
the opportunities and preferences of an individual who
is affected by an externality, say air pollution. The
individual's dollar consumption is shown on the vertical
axis and the level of pollution on the horizontal axis.
We assume the consumer's initial income is equal to 01
o
and that the initial level of pollution is equal to OE .
The consumer is thus initially at point A in consumption-
pollution space on indifference curve W . The indifference
curves slope downward to the left since air pollution is
considered a "bad" rather than a "good". The good which
we are really considering in this case is clean air which
increases to the left from E . Thus, for example, point
E would represent a mixture of air consisting of 50% air
^
polluted at level EQ and 50% clean air.
We assume the individual purchases the mixture of
62
-------
Consumption
(Air
Pollution)
Figure 2.3 COMPENSATIONS VARIATION AND EQUIVALENT
VARIATION MEASURES OF THE BENEFITS OF A REDUCTION
IN AIR POLLUTION CONTROL
63
-------
clean air and other consumption goods which maximizes
his welfare subject to his budget constraint. His budget
constraint is/ however, unusual in that one would expect
it to have a kink at point A. Consider first the possibi-
lities for purchasing clean air. In general, the most
efficient point to stop air pollution is at the source.
For any one individual this is, however, likely to be ex-
tremely costly. Therefore, we represent the individual's
budget line for purchases of clean air by AE in Figure 3.
The individual cannot, however, readily sell clean air
at any positive price (i.e., receive payment for accepting
higher levels of pollution), given present institutional
and technological conditions. Thus, the individual's
market opportunities for voluntarily accepting higher
levels of pollution would typically be given by a budget
line such as ATT£.
It should be remarked that the two segments of the
budget constraint shown in Figure 3 need not be linear
since there are no markets in which clean air can be tra-
ded at a constant price. For example, a one per cent
reduction in pollution from level E may be cheaper at
point EQ than at point E. or E . This does not, however,
effect the analysis which follows since it requires only
that a kink exist at point A.
64
-------
The situation in which all pollution is eliminated
at zero cost to the pollutee is represented by point I
in Figure 3. At this point the individual is on indif-
ference curve W,. The consumer's loss in welfare from
pollution is thus given by the difference between indif-
ference curves W, and W . As was the case with the wel-
1 o
fare effects of price changes, there are many ways in which
this difference can be measured in money terms. However,
in any specific application, one particular measure will
usually be indicated. In the present case, two measures
are of particular interest.
One measure, which is particularly relevant when the
elimination of external effect is being considered, is the
maximum amount the individual would be willing to pay for
the elimination of the pollution. If the pollution is to
be eliminated at zero cost to the consumer (the price
of clean air falls to zero), the maximum amount he would
be willing to pay is I - I , since with income I he
would be indifferent between living without pollution and
living with pollution OE and income I .
This measure is related to the compensating variation
measure discussed above. It should be noted, however,
that the area under the compensated demand curve for clean
air, derived by holding welfare constant along indifference
65
-------
curve W , no longer is necessarily a measure of the com-
pensating variation. This can be seen more readily if we
decompose the elimination of pollution into two steps.
First assume that the consumer is confronted by a price
for clean air just low enough so that he would purchase
enough clean air to eliminate all air pollution (repre-
sented by iT-i't-j in Figure 3) . The income of the consumer
measured in terms of consumption is thus I after the price
reduction. The compensating variation measure of the
improvement in welfare experienced by the individual
due to this price change can be found by constructing
ir_Tr' parallel to TT.TT' and tangent to W . The consumer is
22 11 o
thus as well off with income I and the possibility of
purchasing clean air along the budget constraint n^'
£* 4+
as he was in his original position (point A). The compen-
sating variation measure of the welfare gain due to the
price change is thus I - I . The compensated demand curve,
relevant for this price change is shown in Figure 4. The
compensating variation is therefore given by area edc.
Consider now a further decline in the price of clean air
below IT . Since the consumer has already purchased enough
clean air to eliminate all pollution/ he cannot purchase
"j,
more. Further price reduction (to, say, zero) will there-
fore represent pure income gains for the consumer and are
66
-------
D
TT
ir-
Price of
Clean Air
Pollution
Clean Air
Figure 2.4 COMPENSATED DEMAND CURVES ON THE
BENEFITS OF AIR POLLUTION CONTROL
67
-------
represented simply by the income change IQ-II in Figure 3.
An attempt to measure the costs of pollution based on the
area under ~the compensated demand curve would thus over
estimate the costs of air pollution by amount I - I in
c 2
Figure 3. The reason for this apparent discrepancy is
that when the price falls below IT , the consumer is in
a no-trade situation characterized by already pollution-
free air, and the prices/ in terms of which the goods
are being evaluated, no longer reflect the consumer's
relative marginal valuations.*
The second measure corresponds t*o the equivalent
variation measure discussed earlier, and is relevant if
we want to know now much a receptor of pollution should
be compensated if the pollution is not eliminated. In
this case we are interested in determining how much addi-
tional income would make the individual indifferent between
living with zero pollution and income I / and living with
o
a pollution level of P and income level of I plus the
income compensation.
*Hicks [27] suggested two additional measures of welfare
change, the compensating surplus and the equivalent surplus,
which constrained the individual to consume the same
amount of the commodity whose price had changed before and
after any compensation. In terms of Figure 3, I - Ic would
be the compensating surplus and Ie - Io the equivalent sur-
plus. The welfare measure defined above is thus the same
as Hicks' compensating surplus. The compensating and
equivalent surplus measures have not been commonly used
since in most applications there is no reason why the
68
-------
The required income compensation is obviously I - I .
As was the case with the compensating variation, a measure
of this change based on the area under the compensated
demand curve, holding the individual's welfare constant
at level W , need not equal I - I . A more important
point is the fact that I - I need not equal I - I . In
e o c c
fact, while the size of compensating variation measure is
bounded by the size of the individual's income, there is
no such bound on the equivalent measure.* It should also
be remarked, that if income effects are absent (so that
the indifference curves in Figure 3 are equidistant at all
pollution levels), the compensating variation measure
will be equal to the equivalent variation measure. All
of these measures are therefore given by relevant area
under the compensated demand curve since ordinary and
compensated demand curves will coincide in this case.
Assuming the absence of income effects is, however, equi-
valent to assuming that the income elasticity of demand
consumption of the affected commodity should be constrained
at any particular level. In the present application, the
kink in the budget constraint insures that the consumer
will behave as if his consumption were constrained.
*The assertion that the compensating variation is bounded
by the consumer's income assumes that the consumer's wil-
lingness to pay is constrained by his ability to pay, and
that the negative consumption of goods is impossible.
69
-------
for the external effect is equal to zero. This probably
is not the case for many external effects. For example,
most people would consider the emlimination of external
effects such as noise and pollution to be non-inferior
goods (see, for example, Paul [48] and Anderson and
Crocker [1]). The definition of, and magnitude of, an
external cost is thus not generally well defined or readily
measurable. The appropriate measure will depend on the
context in which it is to be used.
Extension to Many Individuals and Commodities
Up until now, we have discussed the measurement of
external effects in a very special setting, consisting of
one individual, an external effect, and a general composite
commodity. We consider next the implications of aggregating
the measures discussed above across more than one individual,
along with the implications of considering the same problem
in a general equilibrium setting with more than two com-
modities and distortions, taking into account the interaction
between markets and individuals and the existence of
market imperfections.
The theoretical aggregation problem can be stated
as follows: Given some social welfare function,
SW(W^,..., Wn), which depends only on the welfare (or
utility levels), W ,..., W11, of the individuals in the
70
-------
society, is there some way of aggregating net benefit or
cost measures over these individuals such that a change
in the state of the economy resulting in a measured ag-
gregate cost will always yield a lower value of the social
welfare function? For example, assume that a change
occurs which benefits some individuals and imposes costs
on others. The question is then whether an aggregate
based on measured (in terms of money) individual net bene-
fits or costs exists which necessarily implies that social
welfare has increased if the aggregate is positive and
decreased if the aggregate is negative. The answer to
this last question may be no if the change results in a
change in the distribution of income, or if the initial
state was not efficient. The importance of this question
will be illustrated below. The problems associated with
aggregating welfare measures are discussed, in greater detail
in Mohring [42] and Milleron [38] . Problems associated
with redistribution effects are discussed by Weisbrod [57] .
It can be shown that this is possible only for spe-
cial cases. For example, national income, evaluated at
current prices, does not satisfy the above requirement.
Sarauelson [52] showed that there does not exist any social
welfare function, depending only on individuals' welfare,
which always increases when national income increases
71
-------
(see Graaf [19/ Ch.ll] for a good survey of the literature
on the interpretation of aggregate income measures). Simi-
lar problems arise when one attempts to aggregate individual
consumers' surplus measures under general conditions.
Some exceptions do, however, exist. In particular, Milleron
{38] has shown that if individual utility functions are
homogeneous, and if the social welfare function is a posi-
tive weighted sum of individual utilities, then it is in
fact meaningful to speak of a social welfare function. He
shows in particular that, under the specified conditions,
a positive change in consumers' surplus is equivalent to
an increase in the social welfare function.
These special conditions do not, however, hold in
general (homogeneous utility functions imply unitary income
elasticities of demand for all goods). The general approach
in applied welfare economics has been to ignore these
distributional and aggregation problems. In actual appli-
cation, welfare measures have often been based on consumers1
surplus computed from areas under aggregate (or market)
demand curves instead of individual demand curves. This
implies that aggregate measured welfare changes are computed
*A number of applications which base welfare measures on
areas under aggregate demand curves are surveyed in Prest
and Turvey [49]. Also see Burt and Brewer [10] and
Harberger [22] for examples.
72
-------
as the arithmetic sum of individual welfare changes.
This will generally imply a welfare function in which
the weights assigned to individuals depend on money
incomes and prices since these determine the height of
individuals1 demand curves. Thus, cases may arise (see next
paragraph) where a change results in a net measured
gain if computed in terms of initial prices and a loss if
measured in terms of post-change prices. There does not
appear to be any ready way of avoiding such potential
ambiguities/ since we do not know which form of the social
welfare function to employ.
The relevance of aggregation problems can be illustrated
by the following possibility. Consider a city which disposes
of its solid wastes in landfills. It computes the external
costs of operating these land fills under the currently
prevailing distribution of income and set of prices and
finds that these costs would exceed the cost of operating
an incinerator under the same conditions. It may find/
however/ if it closes the land fills and installs the
incinerator, that under the new conditions the external
costs of the incinerator exceed the external costs of the
landfills. This would be possible if, for example, the
individuals who were affected by the land fill assign a
lower cost to the externalities after the incinerator is
73
-------
built (due to changes in the distribution of income and
prices) than they assigned to the same externalities before
the incinerator was built.
Further complications arise when the existence of
many commodities, external effects/ and the possibility
of interactions between different markets is explicitly
recognized. We will discuss two major complications here/
both of which should be treated in the context of a general
equilibrium model.
A major criticism of the measures we have discussed
above is that they are developed in a partial equilibrium
framework/ which does not adequately allow for the type
of substitution possibilities and interactions which occur
in an economy. For example, a waste disposal facility
may pollute the water used by a manufacturing firm/ in-
creasing the cost of clean water to the firm. This in
turn/ may lead the firm to use less water-intensive pro-
duction techniques/ leading people to buy less water-
intensive goods. At the same time, one would expect the
prices of other factors and goods to change in response
to the waste disposal facility. Only by comparing the
value of the output (or output plus total consumers'
surplus) of the economy before the change with the output
after the change can a correct measure of the costs imposed
74
-------
by the pollution be obtained.
This is especially important if the effect is expected
to have a significant impact on the pattern of economic
activity within the relevant region or country. Thus, an
analysis of the external effects produced by the Interstate
Highway System should consider the possibility of country-
wide changes. An analysis of a dump should consider the
adjustment of economic activity within the local community.
That a general equilibrium approach can significantly
affect measurements of welfare effects has been demonstrated
in a number of contexts. Early efforts in the field, deve-
loped by Tinbergen [56] and Bos and Koyck [7], were schemes
to estimate the benefits associated with new roads. They
showed that the direct savings in transport costs attri-
butable to a new road network could significantly under-
state the total benefits of the roads since the improved
road system also allowed more efficient patterns of pro-
duction. Similar results were obtained by Harberger [20]
in his analysis of the welfare costs of the distortions
produced by the corporate income tax. A more general
framework for computing equilibria in models with and
without distortions has recently been developed by Shoven
and Whalley [54 j . The theoretical work of Debreu [14]
in developing measures of welfare costs in a general equi-
75
-------
librium model should also be cited in this context.
General equilibrium analysis has also been used to analyze
the geographic relocation of economic activities due to
external effects. This is discussed in greater detail in
Section D, where the relation between external effects
and land rents is discussed.
The second major complication one encounters when one
considers a more general framework is the possible existence
of other external effects, distortions or imperfections in
the economy. The implications of this possibility have
been discussed in the welfare economics literature by
Lancaster and Lipsey [31] under the heading of general
theory of the second best (for further references and
a correction see Santoni and Church [53]).
Complications arise since when more than one distortion
or externality exists in an economy, the removal of a
single distortion may actually reduce economic welfare.
For example, taken by itself the excise tax on gasoline
is a distortion, which reduces individuals' welfare more
than is necessary for the amount of revenue it produces.
Thus, by eliminating the gasoline tax and substituting an
income tax of equal yield one would improve gasoline buyers'
welfare (if the income compensated wage elasticity of supply
of labor is zero). However, an increase in gasoline con-
76
-------
sumption and hence pollution may also result from this
change. The net result of eliminating the gas tax may
thus be a reduction in overall social welfare. The effects
need not, however, necessarily be offsetting. For example,
an increase in air pollution may increase the demand for
air conditioners. If there is a further distortion in the
economy in the form of an excise tax on air conditioners,
the additional excess burden on purchasers of air conditioners
should also be considered as a cost.
This again suggests that externalities should be
valued in a general equilibrium framework. It should,
however, be recognized that the theoretical and practical
problems associated with a complete general equilibrium
analysis may make the cost of such an approach prohibitive
in many applications. An alternative strongly advocated
by Harberger [21, 22, 23] is to identify all the markets
in which a given action is likely to have a significant
effect, compute the welfare gain or loss which results from
the action in each of these markets allowing for relevant
interactions, and then sum across all the affected markets
to obtain a measure of the total welfare cost of the
action.
In particular Harberger [23] suggests the following
measure:
77
-------
z*
AW = / I D. (x.) dz
z=0 i 1 3z
where D. represents net external benefits or costs in
activity i, x represents the level of activity i and z
i
is the policy variable. The summation is over all acti-
vities affected by the policy variable. This measure
would improve on the standard Marshallian measure and
deal with some of the general equilibrium problems dis-
cussed above. Harberger also suggests the measure can
be readily applied in practice since the number of acti-
vities which, both are affected by any one policy and
generate significant external costs or benefits, are likely
to be small in most cases.
It may, however, be more difficult than Harberger
suggests to correctly implement such a measure. It may,
for example, be difficult to ascertain which activities
will be affected by a policy without first performing a
general equilibrium analysis of the policy being considered.
If such a general equilibrium analysis is in fact performed,
measures of external costs can be obtained more directly
(see, for example, Shoven and Whalley [54]). In other
words, it may be impossible to correctly assess the magni-
tude of the terms 8x^/8z without a prior general equili-
brium analysis of the problem. The Harberger measure may
78
-------
thus be fairly arbitrary if the affected activities are
selected without a rigorous general equilibrium analysis
or fairly costly if affected industries are selected
correctly. Other, partial equilibrium measures which
totally ignore general equilibrium problems are, of course,
even more arbitrary.
Summary
We can summarize the results of this section as fol-
lows. If one abstracts from general equilibrium and
aggregation problems external cost measurement based on
either a compensating variation or equivalent variation
concept is defensible in most cases. These measures can
be computed directly from information about the individual's
welfare function, or from areas under the appropriate
compensated demand curves. The standard Marshallian
measure, which is the one which has generally been used
in attempts to measure non-market benefits or costs, will
coincide with the above measures only if income effects
equal zero.
The information requirements of both the equivalent
variation measure and the compensating variation measure
are, however, rather prohibitive. Estimates can be readi-
ly obtained if competitive price and quantity data are
available and income effects are zero. If income effects
-------
are significant, these must also be estimated if compen-
sated demand curves are to be obtained. Unfortunately,
competitive price data are frequently unobtainable in the
type of situation where measures of welfare changes are
required. In some cases where competitive price data are
unavailable due to distortions such as market imperfections
and taxes, shadow prices which approximate competitive
values can be computed (see, for example, Margolis [35] and
McKean [37]). However, in cases where externalitites are
important, markets may not exist in any form and the direct
determination of competitive valuations or willingness to
pay becomes very difficult for reasons which are discussed
in detail in Section C.
Even in cases where measurement is feasible, care
must also be taken to account for changes in other sectors
or activities if distortions exist in these sectors.
C. INFORMATION PROBLEMS IN THE VALUATION OF EXTERNAL EFFECTS
It was shown in Section B that the valuation of an
external effect requires fairly detailed information about
the preferences of the receptor of the external effect.
In this section we discuss some of the problems which
arise when direct approaches are used to obtain such
information. The discussion will relate, in particular,
to cases where attempts are made to obtain the information
80
-------
directly by asking individuals how much they would be
willing to pay to eliminate the externality (the compen-
sating variation measure) or how much compensation would
make them as well off with the externality as they were
without it (the equivalent variation measure) . (See,
for example, Ridker [51, ch.5]). The discussion also
applies to the somewhat less direct measurement approaches
based on areas under compensated demand curves, since
information about individuals ' valuation of the externa-
lity ip also required for estimates of these demand curves.
In this section we adopt Arrow's [2, pp. 13-16]
interpretation of an externality and assume that externa-
lities are commodities which are exchanged among individuals,
firms or agencies in the economy, but not traded in markets.
In terms of the notation of Section A, if
Aw1
> fe
we can define a new set of commodities x-*1 which satisfy
JC
the constraint
= xj.
k
* •
The variable x?1 thus represents the effect on individual i
81
-------
of commodity k when it is under the control of individual j.
We also define the numbers N3 which represents the number
k
of individuals, in addition to individual j, who are
influenced (i.e., AwVAx^ jt o) by j's consumption of
JC
commodity k.
Most externalities differ, however, in important ways
from ordinary private goods. In the case of an externality,
one or more of the parties to the exchange is usually
an involuntary partner. For example, in the case where the
» *
commodity x^1 is a good, it is usually difficult for the
JC
producer, j, of the good to exclude others from the benefits
of the good. In the case where the commodity is a "bad",
any reduction in the person's consumption often reduces
other's as well; they cannot readily be excluded from
benefits of a reduction in the bad. This problem of ex-
cluding individuals from the benefits of increases in
external goods, or decreases in external bads is one
reason why markets, which would provide evaluations of
these commodities, fail to exist.
The possibility of exclusion is crucial for the
direct measurement of externality since, whenever exclu-
sion possibilities do not exist, individuals will not have
an incentive to reveal correct information about their
valuation of the commodity being considered. We should note,
82
-------
however/ that the possibility of exclusion is not an
absolute characteristic which a commodity has or lacks.
In the case of some commodities the process of consump-
tion or use automatically excludes all others from consump-
tion. For example, the consumption of an ice cream cone
(commodity k) by individual j excludes all others from its
benefits (i.e., N-J = 0). In this case, the good is said
J£
to be rival in consumption and exclusion is not a signi-
ficant problem (see Musgrave [46]). in other cases, however,
exclusion problems do not prevent market allocation only
because some agent or agency has incurred costs in the
process of excluding people from their consumption. For
example, amusement parks incur exclusion costs in the
form of fences, security guards and other security devices.
Exclusion is also supported by society through the expendi-
ture of resources on police and institutions such as courts.
In certain cases where exclusion would be impossible or
too costly for the individual, society creates special
institutions, such as patent rights in new inventions, in
order to reduce exclusion costs. In general, a commodity
lacks exclusion problems whenever it is rival in use or
exclusion costs are low relative to the benefits the owner
of the good could receive by excluding others from the
consumption or use of the good.
83
-------
On the other hand, in the case of commodities for which
exclusion seems to be a problem, exclusion is often possible if
the relevant agent is willing to incur sufficient costs. For
example, a family living next to an open dump can reduce
the visual offense by building a fence, reduce air pollution
and odors by installing air conditioning and filtering
devices, and reduce sound pollution by sound insulation.
The exclusion problem thus arises when exclusion costs be-
come large relative to the value of the commodity being
traded. This would seem to imply that the exclusion pro-
blem is merely a matter of degree, and that if the exclu-
sion problem is the only difficulty, then measurements
and valuations of external effects can be obtained by
incurring the necessary exclusion costs on an experimental
basis. This may in fact be possible in certain cases and
is discussed briefly at the end of this section.
However, even if an agency is willing to incur the
required exclusion costs, the valuation process is often
complicated by further public goods aspects of many ex-
•
ternalities when many agents are involved (i.e., if N^ is
k
large) and by bargaining problems when only a few agents
are involved (i.e., if N^ is small). These problems can
be illustrated by considering, for example, the problem
of estimating the external costs that air pollution from a
84
-------
dump imposes on a community. We consider this problem in
three special cases, the last two of which illustrate
complications cited above.
Case I. Assume:
A.l) There are many individuals in the community
who are affected by the air pollution.
A.2) Individual air pollution control is possible.
By this we mean that it is possible/ for finite cost, to
eliminate the effects of air pollution for single individuals
in the community without eliminating them for all individuals
in. the community at the same time.
We will also assume that the cost of providing indi-
vidual pollution control is above most individuals' wil-
lingness to pay for pollution control. If it were less,
there would presumably be a well-developed private market
for such devices and a serious valuation problem would not
exist. Under the assumed conditions the government, or
the agency attempting to estimate the external costs, can,
however, create an experimental market which would provide
the information required to value the external costs.
The government could conduct the following experiment.
It would offer to make local pollution control available
at a price where just those individuals with the highest
willingness to pay for pollution control would purchase the
85
-------
service. Once it has been determined which individuals
are willing to pay this price, and what the economic char-
acteristics of these individuals are/ it would repeat the
experiment at a somewhat lower price. By running a suc-
cession of such experiments in otherwise identical communi-
ties, the government can obtain all the information it needs
to estimate the demand function for pollution control,
which can in turn be used to estimate the external cost of
the pollution. A number of communities (or a number of
groups randomly chosen within the same community) would be
required since, if successively lower prices were announced
in a single community (or group), a serious expectations
problem could arise. People with a high willingness to
pay would have an incentive to conceal their true marginal
evaluation in order to hold out for lower prices.
We can summarize Case I as follows. If assumptions
1 and 2 hold, there exists a purely private good called
local pollution control. There is not a functioning
market for this good because the price at which private
firms would be willing to supply the good exceeds the price
individuals are willing to pay. There is therefore no
*Normally, the fact that the supply price of a private good
exceeds the demand price (what individuals are willing to
pay) can be interpreted as an indication that the good should
not be produced. We are interested in providing the good
under these conditions in this case only as a means of ob-
taining information about peoples' willingness to pay for
cheaper, but also less information producing, control strategies,
86
-------
market information available. Such information can, however,
be obtained if the government makes the good available
at a price which is less than 'the private supply price,
either by producing and supplying the good directly, or
by offering potential private suppliers a subsidy.
Case II. We replace assumption A.I by:
A.3) There is only one agent who is damaged by the
pollution (i.e., N-j = 1).
Jt
We begin by noting that assumption A.2 is now no
longer necessary. When only one agent is involved it does
not matter in the valuation process whether the pollution
control is at the source of the pollution or at the site
of the receptor since exclusion of other agents from the
potential benefits of pollution control is no longer an
issue. As Arrow [2] has emphasized, the problem of valua-
tion in this case is related to the general indeterminacy
of bargaining processes rather than exclusion problems.
In this case we are essentially dealing with a bilateral
monopoly situation, in that there is one agency, the govern-
ment, "which is willing to provide pollution control and one
potential purchaser of pollution control. Consider now an
experiment similar to the one outlined under Case I. In
particular, assume that the government offers to make
pollution control available at less than cost in order to
87
-------
obtain information about the receptors' willingness to pay
for pollution control.
The government wants to know the maximum amount the
receptor would be willing to pay. It must therefore try
to extract the maximum possible payment, from the receptor.
The receptor on the other hand wants to minimize his pay-
ments for pollution control. The outcome of the bargaining
process will thus in general be indeterminate and bias the
valuation downward, since the receptor will have an incentive
to understate the damages he suffers from the pollution.
In no case would a receptor offer to pay more for pollution
control than his valuation of the damages imposed by the
pollution.
An alternative would be to make pollution control
available without requiring payment, but asking the receptors'
willingness to pay for pollution control. In this case the
receptor would have a strong incentive to overstate the
damages suffered, in order to maximize pollution control.
Thus, whenever small numbers of agents are involved,
it will in general be difficult to get information about
external costs from market observations, surveys or experi-
ments .
Situations arise, however, in which many
individuals are involved, but local pollution control is
88
-------
not a relevant possibility. We therefore consider next:
Case III. We reinstate assumption A.I and replace
A.2 by A.4:
A.I) There are many individuals who are affected by
the air pollution (i.e., N3 is a large number).
JC
A.4) Pollution control is possible only at the source
of the pollution.
In this case, pollution control becomes a public good
for the community affected by the pollution and market
experiments of the type discussed under Cases I and II are
no longer possible. Assumption A.I means that if the amount
of pollution control is based on individual valuations, the
information any one individual provides will have, at most,
a very small impact on the level of pollution control which
is provided. Assumption A.4 assures the individual that
he will benefit from the air pollution control which is
eventually provided, independently of information or
misinformation which he may have provided. An individual
will therefore have an incentive to frustrate the valuation
of the pollution damage he incurs, since one of the fol-
lowing two situations will generally apply.
Consider first the case where the cost to the individual
of pollution control is not related to the valuation the
government places on the damages the individual incurs.
89
-------
The individual could benefit in this case by overstating
the damages he suffers, since his overstatement may en-
courage the government to provide additional pollution
control at no cost to the individual. (This assumes/ of
course, that the marginal benefits of additional pollution
control are always positive.)
The government may therefore decide to relate charges
for pollution control to the benefits received from pol-
lution control. In this case, the government will, however,
run into the "free rider" problem, since understatement by
any one individual of his benefits would have only a small
effect on the level of pollution control provided, but
would presumably result in a significant reduction in
marginal charges insofar as they are set equal to marginal
stated benefits. In other words, the reduction in the
marginal benefits the individual receives would be less than
the reduction in his marginal charges. We would thus
expect any direct measurement techniques to encounter
serious measurement difficulties if actual conditions
are approximated by the assumption of Case III.
*See Buchanan [8] for a good discussion of public goods
issues. The "free rider" problem is disscussed in Chapter
5. For a more rigorous discussion of public goods issues
see Milleron [39].
90
-------
We should also note/ however, that considerably more
sophisticated non-market experiments than those discussed
above are possible. For example, one could announce that
pollution control will be provided only if a certain amount
in money payments is subscribed. In this case, if the
individual does not know the target amount, or how much
money other individuals have subscribed, the individual
will realize that there is a possibility that a small
understatement can result in no pollution control at all.
Bohm [6] reports on a number of such experiments which
attempt to determine the willingness to pay for broadcasting
that obtain some apparently encouraging results. Theoretical
issues in the design of such experiments have also been
studied by Kamien and Schwartz [29, 30]. In general,
however, very little is known about the properties of
such experiments and further research, involving both the
running of actual experiments, and the analysis of their
general theoretical properties is necessary before any
reliability can be attached to these estimates.
This section can be summarized as follows. Only under
special conditions is it possible to obtain direct estimates
of an individual's willingness to pay for an externality
(or to pay for its removal). These conditions are that
the externality have the characteristics of a private good,
91
-------
as in Case I, so that exclusion and bargaining problems are
minimized. Unfortunately, situations similar to Cases II
or III are much more likely to occur in the presence of
external effects. For example, a dump may result in air
pollution, an externality which usually results in direct
measurement problems of the type discussed under Case III,
and also in the pollution of ground water. This will give
rise to a measurement problem of the type considered under
Case II if water is distributed centrally through a muni-
cipal or privately owned water company. Direct measurement
approaches of the type considered in Section B therefore
do not appear to be promising without considerable basic
research on experimental techniques. We therefore consider
next a somewhat less direct approach to the valuation of
external effects based on land rents.
D. MEASUREMENT OF EXTERNAL EFFECTS BASED ON LAND RENTS
Introduction
When the information required for the direct evalua-
tion of the compensating variation measure (or related
consumers' surplus measures) cannot be readily obtained,
measures of external effects can still be obtained if the
source of the external effect is specific to particular
sites or locations, and the external effect is not uni-
formly distributed over the region. Under these conditions
92
-------
an individual or productive activity can avoid the external
cost by relocating. The market for alternative locations
will thus also be an indirect market for units of the
external effect, and it should be possible to infer infor-
mation about willingness to pay for the elimination of
pollution from agents' transactions in this market.
Ideally, information on differences in land rents
attributable to an external effect should measure the cur-
rent cost or benefit of the effect. In cases where infor-
mation on land rents is not available, land values, which
are capitalized land rents, serve a similar purpose. How-
ever, in this case the linkage is less direct since the land
values will reflect expected future (discounted) rents
as well as current rents. The implications of this ob-
servation are discussed in greater detail at the end of
this section.
The use of changes in land rents as a measure may
also be advantageous when one wants to estimate the social
costs of an activity producing a number of external effects
since a single measurement may capture all of the external
costs produced by the activity. For example, a dump that
produces externalities in the form of noise, air pollution
and a health problem due to vectors, will effect land values
due to all of these externalities. The change in land
93
-------
rents attributable to the dump will thus be a measure of
all three external costs.
The Theoretical Framework
In a partial equilibrium setting the use of differences
in land rents as a measure of external effects can be
justified by reference to the classical theory of rents.*
This theory assumes that no pure profits remain in a com-
petitive equilibrium. Thus, if a piece of land-is more
productive in farming than adjacent land, potential far-
mers will bid up the rent of the land until it is no more
profitable to farm it than the surrounding land. Similarly,
if a piece of property suffers pollution damage, its
rent will fall relative to similar land until it is as
profitable to utilize as the surrounding land. Thus the
full cost of the external effect should be reflected in
the land rent. This relationship between external effects
and land values was first investigated in a partial equi-
librium setting. Mohring [43, 44], for example, performed
some detailed studies of the relationship between land
values and transportation improvements.
*See Mishan [41] for a discussion of the relation between
rents and other welfare measures. He concludes, in a
partial equilibrium setting, that economic rent and con-
sumer's surplus are formally equivalent (measure the
same thing), the first being computed with reference to
supply prices, the latter with reference to demand prices.
94
-------
The approach described above is, however, in general
deficient since it does not allow for changes in the
pattern of land use. The problem becomes considerably more
complicated when such changes are considered and a general
equilibrium approach is required. The problem was first
considered in this setting by Strotz [55) who considered
a simple general equilibrium model with two types of
land, which are identical except that one was affected
by pollution. He found that a first order approximation
to the compensating variation measure of pollution costs
is equal to the sum of the absolute value of the changes
in all land rents in the system.
This work was extended by Lind [32,33] who considered
the problem as an optimal assignment problem in a
general equilibrium setting.* He assumes that there
are m parcels of land and n activities in the economy.
An activity is either a production activity, such as
farming, factories etc., or a consumption activity, such
as residential housing. Let a.. be the maximum amount
ID
activity i is willing to pay for parcel j. The value
of a.. will in general depend jointly on the activity
and the parcel.
* For a discussion of the assignment problem see
Beckmann and Koopmans [5] or Gale [18] .
95
-------
In terms of the classical theory of rents, it is the
maximum rent the owner of parcel j can extract from a
renter engaged in activity i.
The optimal assignment problem is now one of assigning
the n activities to the m parcels in such a manner that the
sum of the a is maximized. This can be transformed
JO
into a linear programming problem by defining a set of choice
variables s i=l,..., n; j=l,..., m, such that s.. = 1
iD -^
if activity i locates on parcel j and is zero otherwise.
The problem can then be stated as follows:
n m
MAX T T a.. s..
I 5 ID ID
{s..}
SUBJECT TO:
si-i = 1 = 1,..., n
<_ 1 I j = l,...,m.
D ^
The constraints ensure that at most, one activity will
occupy each parcel of land and that any given activity
is assigned to only one parcel.
Let p. be the rent on a parcel j arising from a
96
-------
competitive equilibrium. It can then be shown that there
exists a set of equilibrium rents such that if parcel j
is used for activity i in the optimal assignment:
a.. > p. (2.1)
ij = j
and
a.. - p. £ a.. - p k = 1,..., n. (2.2)
ID D ID *
Note that these conditions ensure that activity i is not
loosing money when it locates on parcel j (2.1) and that
there is no better location for it (2.2). These rents
will in general not be unique, but Lind proves that any
equilibrium rent vector, p, will lie in an interval
P <, P <, Pi where both p and p satisfy ,(2.1) and (2.2). <-The
actual rent vector obtained in a given situation will
depend on factors such as actual market adjustment mechanisms,
bargaining strengths, etc.
Consider now the relationship between land rents and
net benefits (or costs) from changes in an external effect.
Define c^ = s. . (a. . - p.) for each i (where s. . represents
the values of the variables s,. chosen in the optimal
^
assignment). For a consumer, c. represents the consumers
surplus consumer i receives from his (optimal) selection
97
-------
of parcel j. For firms, c^ Is the profit of firm i when
it selects parcel j in the optimal assignment. Thus, it
is obvious that the net social benefits from the use of
land are
I I s*. a = I c± + I p. .
i j *3 *3 i j ^
Consider now, for example, the elimination of an
external cost in the system; this will enhance, by definition,
the value of at least a subset, Q, of the parcels of land
in the system. Let a1.. be the values of land parcels
associated with this new state of affairs. By definition
a *.. > a.. for some i = l,...,n jeQ,
and
a*,. = a.. for all i = 1,..., n j e N-Q .
Now, for simplicity, renumber activities such that acti-
vity j is assigned to parcel j in the old optimal assign-
ment. Let TT(J) be the new optimal assignment of activity
j. Thus, obviously, total net benefits from elimination
of the external effects are give by
98
-------
Thus, land rent changes equal the change in total social
net benefits only if the final term of (2.3) is zero. This
means that all firms must earn zero profit, or that the
sum of the profits specific to parcels are unchanged in
the new assignment, and likewise for the consumers' sur-
plus of households. This is an extremely strong require-
ment/ especially with respect to households; it requires
that the consumers' surplus of the households be, in sum,
the same in the old and new assignment. There do not appear
to be any economic reasons why this sum should in general
increase/ decrease, or remain constant.
The problem as stated also requires evaluation of the
changes in the rents of a large number of parcels in the
system. Lind makes a major contribution here by showing
that changes in the rents and surpluses of activities which
do not locate on sites directly affected by the external
effect will sum to zero.* We thus need be concerned only
with changes in the productivity or surpluses of those
activities which are directly affected. This reduces
the computational problem significantly for those projects
or activities, such as a solid waste disposal site, which
affect only a limited number of locations.
* The proof of this proposition uses properties of permu-
tations and the equilibrium conditions discussed above.
For a more detailed discussion see Lind [33, pp. 194-201].
99
-------
Limitations of Measures Based on Land Rents
The theory developed above also points to some defi-
ciencies most land rent studies are likely to have. First,
given the assumptions of the assignment problem model, the
correct (compensating variation) measure of external costs
or benefits is given by (2.3). As we noted, this measure
will equal the change in land rents only if the sum of the
surpluses earned on all parcels remains unchanged. True
costs may thus be either greater or less than the measured
changes in land rents, depending on whether the sum of
the surpluses increased or decreased after the change.
Further complications arise when some of the assump-
tions of the model are relaxed. In particular, the model
takes the willingness of each activity to pay for a given
parcel of land as given. It determines an optimal assign-
ment and equilibrium rent vector assuming all other prices
for factors and goods remain constant. If this is the case,
any benefit or cost from an externality affecting a parcel
of land will accrue to that parcel in the form of a change
in rents to its owner, or a change in the surplus or profit
of the activity occupying the parcel. If, however, the
product or factor prices of the activity occupying the
parcel also change because of the external effect, it is
no longer true that all the benefits or costs will be
100
-------
reflected by (2.3). In this case, the compensating variation
measure would require that the a] . be evaluated on the
ij
basis of prices existing before the external effect occurred.
The relevance of the Lind model is thus limited to
the effects of activities which do not disturb the equilibrium
prices in markets other than the market for land. It is
thus probably relevant for the evaluation of external costs
associated with particular solid waste disposal facilities.
It may not, however, be appropriate for estimating the
benefits or costs of any national policy, altering the
pattern of waste disposal. For example, a national policy
closing all dumps and landfills may significantly affect
aggregate waste disposal costs and hence prices in other
markets.
Lind notes, however, that in the case of the housing
market, even relatively small and local projects may have
significant price effects, since the price of housing
services is determined in part by the local supply of land
suitable for housing. In this case, a more general model
relating rents to the external effect which also takes
price determination in the housing market as well as land
rents into account would be appropriate. However, such
models turn out to be exceedingly complex, especially if
analytical solutions are desired, as the recent work by
101
-------
Beckman [4], Delson [15] and Montesano [45] shows.
Further complications may arise when one tries to
estimate the relevant rents, p! and p., in(2.3) since in
most cases available data relate to land values, rather
than land rents, and land is typically improved.* We
consider first the. problems of converting changes in
land values to changes in land rents. In general, the
relation between the value, v, of a parcel of land (valued
at time t = 0) and its rent is given by
00
v = / p(t)e"rtdt (2.4)
o
where r is the purchaser's discount rate. Equation (2.4)
integrates to
P
v = - , (2.5)
r
if p(t) = p for all t (i.e., if the rent does not change
*In certain cases measures based on land values may be
preferable. This possibility is discussed briefly at
the end of this section.
+The determination of the consumers' discount rate can also
present serious problems, since there appears to be no
concensus on what this rate should be. Estimates range
from 5% to over 30%. See Arrow and Kurz [3, pp. 117-119]
for a brief discussion.
102
-------
over time).
Now assume that a dump is opened next to the land
parcel in question, causing the equilibrium rent of the
parcel to fall to p' (t) = p'/ and that the market expects
the dump to continue in operation forever. Then clearly,
p« - p = r(v' - v)
where v1 is the value of the parcel after the dump is
opened. Thus, the changes in rents can be obtained from
changes in land values if the appropriate discount rate
is known and if the externality is expected to last forever,
If, however, the dump is expected to last only a
finite period of time, additional information is required.
Assume that the market expects the dump to close after a
period of T years. Then
T °° j_
v' « / p'e"rtdt + / pe dt
o T
which becomes on integration:
= p' - + - - . (2.6)
Subtracting equation (2.5) from (2.6) and solving for p1 - p,
103
-------
we obtain:
-1
r(vf - v). (2.7)
Thus knowledge of how long, T, the externality is expected
to continue in effect is required. More complicated time
paths of the rents are of course possible; this example
merely illustrates the existence of a problem.
Another problem is that buildings or other improve-
ments are generally situated on parcels of land where the
measurement of externalities is considered important. The
assignment model developed above indicates that it is only
the value of the land itself which is affected by the
externality. However, this is true only when all activi-
ties which find it profitable to relocate have done so.
Where improvements are durable, such as in the case of
houses and factories, the adjustment to a new equilibrium
may require a relatively long period of time. If this
is the case, the value of the improvements on a land parcel
may be affected by an externality while the parcel itself
is not affected. For example, a home site exposed to
noise pollution may be equally valuable as a junkyard
location. In this case, the value of the home will
decline due to the noise, but the land will not decline
104
-------
in value. Thus, the external costs of the noise measured
in terms of changes in equilibrium land rents will be zero.
The value of the home, however, will decline as a result
of the noise, the amount of the decline depending on both
the amount of noise and the expected life of the house
(the length of time until a new locational equilibrium is
attained).
A problem thus exists in studies where the available
price data are for both the land and any improvements.
Assume in the example above that the noise is from a source
which is expected to continue in operation for an addi-
tional 20 years, but that the house has an expected life of
10 years. Then, if the noise affects the productivity of
the land, the change in value of the land would be converted
into rents based on a 20 year horizon (by solving for
p* and using T=20). On the other hand, if it is assumed that
the noise affects the improvement on the land only in its
current use, the appropriate horizon becomes 10 years,
implying a higher current annual external cost from the
noise.
Whether or not the problem of going from land values
to rents is a significant one will, of course, depend
on the context in which the measurements are to be used.
If compensation for damages is being considered, the land
105
-------
value figure is probably the most important. If, on the
other hand, one is trying to determine whether a given
annual expenditure to alleviate the effects of the exter-
nality is justified/ the land rent figures would be pre-
ferred. One possible alternative in the latter case would
be to discount future expenditures on externality control
by the public discount rate and compare the capitalized
cost with the land values. This procedure would still
require knowledge of the time path of the externality,
however, plus knowlege of the appropriate public discount
rate.
Measures of external costs based on changes in land
rents may also result in underestimates of social costs
since disequilibrium conditions may persist for long periods
of time due to adjustment costs. Paul [48], for example,
notes that in the case of residential housing moving costs
may be substantial. Thus, a family which is adversely
affected by the noise generated by a new dump may not relo-
cate if moving costs exceed the costs imposed by the noise.
The value of their house and land may not, however, be
affected if there exist potential buyers who are unaffected
by the noise. The noise costs incurred by this family
would thus be omitted in a study of the social costs of
the noise nuisance which is based on changes in land
106
-------
rents. In the case where the noise costs exceed moving
costs, and an optimal assignment of families to residence
sites occurs, the measure based on land rents would still
understate total social costs by the amount of the moving
costs.
A number of additional problems also arise when one
attempts to obtain empirical estimates of the impact of an
activity producing external effects on a set of heterogeneous
sites. These problems, which relate to the correct speci-
fication of an econometric model which would explain dif-
ferences in equilibrium'land values, are discussed in
detail in Chapter III.
E. SUMMARY AND CONCLUSIONS
We can summarize the principal conclusions of this
chapter as follows. There are serious conceptual problems
*
associated with even defining "the cost" of an external
effect, since the appropriateness of any given definition
depends on the specific application of it. Given an ap-
plication, however, the external cost can be defined and
measured directly given sufficient information about
the preferences of the receptors of the external effect.
Such information is however in general unobtainable, making
alternative indirect measurement approaches, such as
property value studies, look appealing. The principal
107
-------
advantage of these approaches is their less stringent
data requirements. A further advantage is their ability
to capture the effects of any number of externalities,
produced by a given activity, in a single study.
There are, however, also a number of problems asso-
ciated with property value studies. In many applications
they may only be suggestive of costs, instead of being
correct measures of social costs. Problems arise when the
total surplus (profits and consumers' surplus) change due
to an activity, when the activity will occur for only a
limited time and when there are adjustment costs in moving
from one optimal assignment to another. Further problems,
which are discussed in Chapter III, arise when one attempts
to obtain empirical estimates of changes in land values
attributable to an externality producing acitivty.
108
-------
REFERENCES
1. Anderson, R. J. , and T.D. Crocker. The
Economics of Air Pollution: A Literature
Assessment. In: Air Pollution and the Social
Sciences, Downing, P.B. (ed.). New York,
Praeger Publishers, 1971. pp. 133-166.
2. Arrow, K.J. Political and Economic Evaluation
of Social Effects and Externalities. In: The
Analysis of- Public Output, Margolis, J. (ed.).
New York, Columbia University Press, 1970.
pp. 1-23.
3. Arrow, K.J., and M.Kurz. Public Investment,
The Rate of Return_, and Optimal Fiscal^ jPolicy.
Baltimore, Johns Hopkins Press, 1970.
4. Beckmann,M.J. On the Distribution of Urban
Rent and Residential Density. Journal of
Economic Theory. Vol. 1 (June, 1969): pp. 60-67.
5. Beckmann,M.J., and T.C. Koopmans. Assignment
Problems and the Location of Economic Activities.
Econometrica. Vol. 25 (1957): pp. 53-76.
6. Bohm, P. Estimating Demand for Public Goods:
An Experiment. European Economic Review. Vol. 3,
No. 2 (1972): pp. 111-130.
7. Bos, H.C., and L.M. Koyck. The Appraisal of
Road Construction Projects: A Practical Example.
Review of Economics and Statistics. Vol. 42
(February 1961): pp. 13-20.
8. Buchanan, J.M. The Demand and Supply of Public
Goods. Chicago, Rand McNally & Co., 1968.
9. Buchanan, J.M., and W.C. Stubblebine. Externality.
Economica, N.S. Vol. 29 (1962): pp. 371-84.
10. Burt, O.R., and D. Brewer. Estimation of Net
Social Benefits from Outdoor Recreation. Econ-
ometrica. Vol. 39 (September 1971): pp. 813-828.
11. Chase, S.B. (ed.). Problems of Public Expenditure
Analysis. Washington, D.C., The Brookings Inst., 1968,
109
-------
12. Currie, J.M., J.A. Murphy, and A. Schmitz.
The Concept of Economic Surplus and its Use
in Economic Analysis. The Economic Journal.
Vol. 81 (December 1971): pp. 741-799.
13. Davidson, P. The Valuation of Public Goods.
In: Social Sciences and the Environment, M.E..
Garnsey and J.R- Hibles, (eds.). Boulder,
University of Colorado Press, 1967. pp. 125-164.
14. Debreu, G. A Classical Tax-Subsidy Problem.
Econometrica. Vol. 22 (January 1954): pp. 14-22.
15. Delson, J.K. Correction on the Boundary Con-
ditions in Beckmann's Model on Urban Rent and
Residential Density. Journal of Economic Theory.
Vol. 2 (September 1970): pp. 314-318.
16. Depuit, J. On the Measurement of the Utility of
Public Works. Reprinted in: Arrow and Scitovsky,
(eds.). Readings in Welfare Economics. Homewood,
R.D. Irwin, Inc., 1968. pp. 255-283.
17. Dorfman, R. (ed.). Measuring Benefits of Govern-
ment Investments. Washington, D.C., The Brookings
Inst., 1963.
18. Gale, D. The Theory of Linear Economic Models.
New York, McGraw-Hill, 1960.
19. Graaff, J. de V. Theoretical Welfare Economics.
Cambridge, The University Press, 1967.
20. Barberger, A.C. The Incidence of the Corporation
Income Tax. Journal of Political Economy.
Vol. 70 (June 1972): pp. 215-240.
21. Harberger, A.C. Principles of Efficiency: The
Measurement of Waste. American Economic Review.
Vol. 54 (May 1964): pp. 58-76.
22. Harberger, A.C. Taxation, Resource Allocation,
and Welfare. In: The Role of Direct and Indirect
Taxes in the Federal Revenue System, Due, J.
(ed.).Princeton, Princeton University Press,
1964. pp. 25-80.
110
-------
23. Harberger, A.C. Three Basic Postulates for
Applied Welfare Economics: An Interpretative
Essay. Journal of Economic Literature. Vol. 9
(September 1971): pp. 785-797.
24. Haveman, R.H., and J. Margolis. Public Expenditures
and Policy Analysis. Chicago, Markham Publishing
Co., 1970.
25. Henderson, A. Consumer's Surplus and the Com-
pensating Variation. Review of Economic Studies.
Vol. 8 (February 1940-1941): pp. 117-121.
26. Hicks, J.R. A Revision of Demand Theory. London,
Oxford University Press, 1956. Chs. 7-10.
27. Hicks, J.R. The Four Consumers' Surpluses.
Review of Economic Studies. Vol. 11 (Winter 1943):
p. 126.
28. Hotelling, H. The General Welfare in Relation to
Problems of Taxation and of Railway and Utility
Rates. Econometrica. Vol. 6 (1938): pp. 242-269.
29. Kamien, M.I., and N.L. Schwartz. Exclusion Costs
and the Provision of Public Goods. Public Choice.
Vol. 12 (Spring 1972): pp. 43-55.
30. Kamien, M.I., and N.L. Schwartz. Revelation of
Preference for a Public Good with Imperfect
Exclusion. Public Choice. Vol. 9 (Fall 1970):
pp. 19-30.
31. Lancaster, K., and R.G. Lipsey. The General
Theory of Second Best. Review of Economic
Studies. Vol. 24 (January 1957): pp. 11-32.
32. Lind, R.C. Flood Control Alternatives and the
Economics of Flood Protection. Water Resources
Research. Vol. 3 (1967): pp. 345-357.
33. Lind, R.C. Spatial Equilibrium, Rents, and
Public Program Benefits. Quarterly Journal of
Economics. Vol. 87 (May 1973): pp. 188-207.
34. Marglin, S.A. Public Investment Criteria.
Cambridge, M.I.T. Press, 1967.
Ill
-------
35. Margolis, J. Shadow Prices for Incorrect or
Nonexistent Market Values. In: Haveman and
Margolis, Public Expenditures and Policy
Analysis. Chicago, Markham Publishing Co., 1970.
pp. 314-329.
36. Marshall, A. Principles of Economics, Eighth
Edition. New York, Macmillan, 1920.
37. McKean, R.N. The Use of Shadow Prices. In:
Chase, S.B., Problems of Public Expenditure
Analysis. Washington, D.C., The Brookings Inst.,
1968. pp. 34-64.
38. Milleron, J.C. Distribution of Income, Social
Welfare Functions, and the Criterion of Consumer
Surplus. European Economic Review. Vol. 2
(Pall 1970): pp. 45-77.
39. Milleron, J.C. Theory of Value with Public Goods:
A Survey Article. Journal of Economic Theory.
Vol. 5 (December 1972): pp. 419-477.
40. Mishan, E.J. Realism and Relevance in Consumer's
Surplus. Review of Economic Studies. Vol. 15
(1947-48):
41. Mishan, E.J. Rent as a Measure of Welfare
Change. American Economic Review. Vol. 49
(1959): pp. 386-395.
42. Mohring, H. Alternative Welfare Gain and Loss
Measures. Western Economic Journal. Vol. 9
(December 1971): pp. 349-368.
43. Mohring, H., and M. Harwitz. Highway Benefits;
An Analytical Framework. Evanston, Northwestern
University Press, 1962.
44. Mohring, H. Land Values and the Measurement of
Highway Benefits. Journal of Political Economy.
Vol. 69 (June 1961): pp. 236-249.
45. Montesano, A. A Restatement of Beckmann's Model
on the Distribution of Urban Rent and Residential
Density. Journal of Economic Theory. Vol. 4
(April 1972): pp. 329-354.
112
-------
46. Musgrave, R.A. Provision for Social Goods.
In: J. Margolis and H. Guitton (eds.), Public
Economics. New York, St. Martin's Press, 1969.
pp. 124-144.
47. Patinkin,D. Demand Curves and Consumer's Surplus.
In: Measurement in Economics: Studies in Mathematical
Economics and Econometrics in Memory of Yehuda
Grunfeld, Christ et al.Stanford, Stanford"
University Press, 1963. pp. 83-112.
48. Paul, M.E. Can Aircraft Nuisance be Measured
in Money? Oxford Economic Papers. Vol. 23
(November 1971): pp. 297-322.
49. Prest, A.R., and R. Turvey. Cost-Benefit Analysis:
A Survey. Economic Journal. Vol. 75 (December
1965): pp. 683-735.
50. Randall, A. Market Solutions to Externality
Problems: Theory and Practice. American Journal
of Agricultural Economics. Vol. 54 (May 1972):
pp. 175-183.
51. Ridker, R.G. Economic Costs of Air Pollution.;
Studies in Measurement. New York, Praeger
Publishers, 1967.
52. Samuelson, P.A. Evaluation of Real National
Income. Oxford Economic Papers, New Series.
Vol. II (1950): pp. 1-29.
53. Satoni, G. , and A. Church. A Comment on the
General Theorem of Second Best. Review of
Economic Studies. Vol. 39 (October 1972): pp.
527-530.
54. Shoven, J.B., and J. Whalley. A General Equil-
ibrium Calculation of the Effects of Differential
Taxation of Income*from Capital in the U.S.
Journal of Public Economics. Vol. 1 (November
1972): pp. 281-322.
55. Strotz, R.H. The Use of Land Rents to Measure
the Welfare Benefits of Land Improvement.
Processed, 1966.
113
-------
56. Tinbergen, J. The Appraisal of Road Construction:
Two Calculation Schemes. Review of Economics
and Statistics. Vol. 48 (August 1957): pp. 241-
249.
57. Weisbrod, B.A. Income Distribution Effects and
Benefit-Cost Analysis. In: Problems of Public
Expenditure Analysis, S. Chase (ed.>. Washington,
B.C., The Brookings Inst., 1968. pp. 177-208.
114
-------
CHAPTER III
THE VALUATION OF EXTERNAL EFFECTS:
EMPIRICAL STUDIES
A. INTRODUCTION
This chapter surveys and evaluates attempts
which have been made to obtain empirical estimates of the
money value of externalities produced by an activity,
such as a dump or landfill. Studies related to two
alternative valuation strategies are discussed. One
approach to measurement and valuation employs a two
stage procedure. In the first stage one would obtain
estimates of all of the physical effects produced by
an activity which are not valued, or incorrectly valued
in competitive markets. In the second stage a net
dollar value is assigned to each of these physical
effects* The basic objects of analysis are thus the
physical effects of an activity. Studies related to
the evaluation of individual physical effects are
surveyed in Section B. The alternative approach takes
as the basic object of analysis the activity itself,
and attempts to obtain estimates of the non-market
benefits or costs of the activity taken as a whole.
115
-------
This approach is implicitly used in most studies
which relate differences in equilibrium land rents to
externalities. These studies are surveyed in Section
C. Both of these approaches to measurement and val-
uation are discussed in greater detail below.*
The first approach outlined above has been advo-
cated by analysts who believe (taking a study of the
benefits of air pollution abatement as an example):
"Consumers are dreadfully ignorant of the benefits of
abatement" [ 42, p. 214], Lave [ 42] thus believes
that attempts to ascertain directly an individual's
willingness to pay (either explicitly by asking the in-
dividual in a way that will elicit an honest answer or
implicitly by observing his behavior in a related market
such as the land market) for the elimination of an external
* It should be noted that other taxonomies of approaches
to the measurement of external effects have been pro-
posed. Ridker [ 65] suggests three major classifi-
cations: (1) calculate the physical effects and then
calculate the value of these effects in terms of money.
(2) Estimate the expenditures an individual incurs in
adjusting to the effects of an activity. (3) Evaluate
the change in consumer and producer surplus in the
markets affected by an activity. This is similar to
the taxonomy of approaches to the valuation of physical
effects used in Section B.
116
-------
cost producing activity are likely to yield
estimates of benefits which are biased downward.
The first approach attempts to avoid this
problem of possible consumer ignorance by first
determining the physical effects of an activity.
For example, air pollution may affect health, soil
clothes and materials, corrode materials, and have
aesthetic effects (a detailed analysis of the
physical effects of landfills is presented in
Chapter IV). These effects can then be valued
in terms of money using the type of measures which
were discussed in Section B of Chapter II. If,
for example, it is determined in the first stage
that the elimination of pollution reduces the
probability of contracting bronchitis by 10
per cent, the second stage must determine what
affected individuals are willing to pay for such
a reduction in the probability of contracting
the disease.
The second approach ignores the possibility
of ignorance on the part of individuals affected
by an activity and assumes that individuals'
117
-------
money valuation of the activity accurately reflects
its non-market costs and benefits. The only
problem facing an analyst employing this approach
is thus the determination of these money
valuations. Since the direct determination of
these money values through the questioning
of affected individuals is likely to give
biased results for the reasons discussed in Chapter
II, Section C, the information required to value
the activity is usually inferred from individuals'
behavior in related markets, such as the market
for land.
Which of the two measurement approaches-
outlined above will yield more meaningful
estimates of the dollar value of external effects
will depend both on the assumptions one makes
about consumer ignorance and the quality of the
information available for each approach. If
consumers are unaware of the external physical
effects of an activity, but an accurate
evaluation of these effects can be obtained, the
first approach should be preferred. If, however,
118
-------
individuals are aware of the physical effects
of an activity, the quality of the estimates
produced by each approach will depend on the
quality of the data used in each case. For
example, much better data may be available on
willingness to pay to reduce the probability
of contracting a disease than is available on
willingness to pay to avoid air pollution.
It may, however, also be difficult to
obtain accurate estimates of the physical effects
of an activity, or to even identify all of
the physical effects, and to value them pre-
cisely '(See Lave [ 42] for a discussion
of such problems in the case of air pollution).
If this is the case, and adequate data are
available on individuals' valuations of the
activity as a whole, the second approach may
be preferrable. The second approach may also
be preferrable since it involves only the
valuation of a single activity, whereas the
first approach involves the estimation (of
possibly a large number) of physical effects
119
-------
and their dollar valuation. In particular if
a large number of effects are involved in the
first approach one would expect it to be the
more costly approach in terms of computation
and data requirements. There are also more
possibilities for error in the first approach
since significant physical effects may be
omitted, inaccurately estimated, or incorrectly
valued.
It should also be noted that the distinction
between an activity and the physical effects of
the activity may be fairly arbitrary. For example,
one of the physical effects of a burning dump
may be the release of particulate matter into the
air. Given this pollution one can, however, con-
sider the new activity of pollution control, which
may have as its physical effects a reduction in
the incidence of disease and soiling of materials.
The distinction between the two approaches to the
evaluation of the external effects of an activity
is thus not very sharp. Similar methods of
determining willingness to pay (for the elimination
of an activity or the elimination of a particular
120
-------
physical effect of an activity) are often applicable
with both ^approaches. There may thus be as many
problems involved in the valuation of a single
physical effect (such as a decrease in the incidence
of a disease) as are involved in the valuation of an
activity as a whole (such as a program to decrease
pollution levels).
A further problem may arise in the first
approach if there are significant interactions between
individual effects. For example, if air pollution
affects an individual's enjoyment of outdoor living
(due to respiratory problems) and also damages his
flower garden, the sum of the damage to the flowers
(evaluated without taking into account the effect of
pollution on the individual's enjoyment of outdoor
living) plus the damage to the individual in terms
of his enjoyments (evaluated without taking into
account the diminished beauty of his garden) may
overstate total pollution damages. The problem of
interactions need not arise with the first approach
if care is taken in evaluating each physical effect
at the equilibrium levels of the remaining effects.
This may, however, be difficult in practice due to
a lack of information or awareness about interaction
121
-------
effects. The problem will not, of course, arise
in the second approach since it measures only the
net effect of the activity overall.
Examples of both approaches are surveyed in the
remainder of this chapter. The survey of approaches
to the valuation of physical effects is highly selective
since most cost-benefit studies can be interpreted
as employing such approaches, and the cost-benefit
literature is voluminous. The survey of the literature
relating externalities to land rents or values is
more comprehensive since this literature is shorter.
It is also the approach employed to evaluate the
external effects of landfills in this study.
B. THE EVALUATION OF EXTERNALITY PRODUCING ACTIVITIES
THROUGH THE ESTIMATION AND EVALUATION OF PHYSICAL EFFECTS
Most of the attempts to obtain estimates of the
dollar value of the external effects (and other non-
marketable outputs) have occured in the context of
cost-benefit analyses of economic activities such as
water resources development, transportation, recreation,
education and pollution control projects. The most
common approach taken in these cost-benefit analyses has
been to enumerate and estimate the magnitude of all
of the physical effects, and then to evaluate these
122
-------
effects (i.e. the first approach discussed in the
introduction).
Two measurement problems thus arise in these
cost-benefit studies: the estimation of physical
effects and the evaluation of these effects in terms
of money. While the difficulties involved in obtaining
measures of physical effects may be substantial in
many cases, they are not discussed in detail in this
chapter since they are primarily technical, rather
than economic. The problems involved in determining the
physical effects of landfills are discussed in detail
in Chapter IV. In this section we emphasize the dollar
valuation of physical effects.
The conceptual issues involved are discussed in
detail in Section B of Chapter II. We survey here
the approaches which have been used to obtain empir-
ical estimates. Most of these studies attempt
to estimate the Marshallian consumer's surplus for the
physical effect which is consumed. Since these
effects are not in general marketed, the major problem
faced in these studies is acquisition of inform-
ation on individual valuations of the physical effects
which can be used to estimate the demand curves.
In the remainder of this section we survey selected
123
-------
applications of various approaches which have been
used to abtain such information.
Direct Market Information
In some cases the physical effect produced by
an activity is also marketed. A good example of such
an externality arises in the construction of many
hydroelectric generation projects. If a new dam is
constructed upstream from existing dams, the generating
capacity of existing downstream facilities may be
increased significantly through improved water storage
and flow control. It was estimated for example that
the construction of the Hungry Horse dam on the
Columbia river system would provide 212,000 kilowatts
of on-site generating capacity and increase the capacity
of downstream sites (controlled by a number of different
utilities and the U.S. government) by 628,000 kilowatts
(See Krutilla and Eckstein [ 39, p. 62].
In this case accurate measures (in terms of
kilowatts) of the externality are readily available.
Since market prices for electricity are usually available
the valuation problem is trivial if the project does
not affect this market price. If the project does
affect the market price, the proper evaluation would
require the estimation of this price effect, and the
124
-------
computation of the consumer surplus generated by the
incremental electricity supplied. In Figure 3.1 let
x represent the total generating capacity in the mar-
ket, D(x) represent the market demand for electricity,
and g1 - q the increment in generating capacity due
x x
to the externality. The Marshallian measure of the
value of the electricity is then given by area g beg*
x *£
since this generating capacity is produced at zero mar-
ginal cost to the facilities where capacity has been
increased.
Market data have also been used in a number of
other studies. One of the major effects of many
transportation projects is the time saved by affected
individuals. The price at which this time is valued
is often based on market wage information (See for
example Foster and Beesely [ 19], Moses and Williamson
[ 53], and Gronau [ 24]. Another example is the use
of information about consumer expenditures on marketed
recreation activities to value the output of recreation
facilities whose output is not marketed (See Mack and
Myers [ 44, pp. 78-88]).
In cases where such comparable private commodities
exist and are marketed the approach discussed above
should produce good estimates of the dollar value of
125
-------
x /
D(x)
Figure 3.1 THE BENEFITS OF AN EXTERNALITY
IN HYDROELECTRICITY GENERATION
126
-------
the physical effects, subject to the theoretical
qualifications about consumers' surplus measures
discussed in Chapter II. In many cases, however,
such close substitutes do not exist. Caution also
should be exercised since valuations may be seriously
distorted if the prices used are not competitive
valuations (See the sub-section on the intermediate
input approach for a further discussion of this point).
Direct Interviews of Recipients
In cases where market information is unavailable
it is tempting to ask individuals what they would be
willing to pay for a physical unit of the external
cost or benefit, and to estimate demand functions from
these data. This approach has been used extensively
to value recreation benefits. Recent applications are
studies of a Maine forest area by Knetsch and Davis
I 38J; crowding in wilderness areas by Fisher and
Krutilla [ ISJand water fowl related recreation by
Brown and Hammock [ 7J . The approach has also been
used to obtain estimates of pollution damage and
willingness to pay for air quality. See for example
the study of a pollution episode in Ridker [ 65, Ch. 5]
and the studies cited in Anderson and Crocker [ 3, p. 145]
Schelling [ 70] has suggested a similar approach to
127
-------
the valuation of programs having health effects.
He suggests that an activity which changes the
probability of illness, injury or death can be
evaluated ex ante by asking the affected individ-
uals their willingness to pay for changes in such
probabilities. Schelling notes, however, that
problems exist in the valuation of events having
very small probabilities and very high associated
gains or losses (See also Mishan [ 49, Chs. 22,23]).
The Intermediate Input Approach
It has been noted that many non-marketed goods
such as public goods or externalities enter as inputs
(positive or negative) into the production process
of commodities which are sold in competitive markets
(See Musgrave [54 ] or Margolis [ 46] for example).
In this case it is often possible to estimate the
dollar value of an externality of public good by
estimating the value of its marginal product in
further production.
For example, the dollar value of irrigation water*
*The output of irrigation water from a project is
not usually an external effect as defined in Chapter
II. Evaluation difficulties arise, however, since
irrigation water is often supplied at non-market prices
and significant distortions exist in related markets.
Krutilla arid Eckstein [40, pp. 56,57], however, also
128
-------
produced by water resources development projects is
often calculated by estimating the increase in
the income of farmers using the water, which is
attributable to their increased use of irrigation
water (See Bain [ 4 ] for example). The intermediate
input approach has also been used to estimate the
value of public education and public health programs.
Education is valued by estimating the increase in
the future earnings which can be attributed to
education (See for example Becker [ 5 ] or Weisbrod
[ 75]). Health and safety programs are evaluated
by computing the increased productivity of affected
individuals. This may take the form of increased
output while working, increased output
due to fewer sick days or increased output due to
lengthened life spans (See Mushkin [ 55], Weisbrod
[ 73, 74 ] or Rice [ 64]). For a general equilibrium
analysis of transportation projects viewed as an
intermediate input see Bos and Koyck [ 6 ].
A shortcoming of the intermediate input approach
cites a case where irrigation has improved aquifers,
thereby increasing the output of irrigation water
.in other areas. This would be an externality of
the type discussed in Chapter II.
129
-------
is the fact that it must be regarded as only a
partial evaluation in many cases. Education, for
example, produces benefits other than increased
earnings capability in the form of direct consumption
benefits to the student and "citizenship11 benefits
to the community (See Weisbrod [ 75], Paul [ 60]).
Health produces individual benefits in the form of
reduced pain and suffering, increased comfort and
lengthened life expectancies among unemployed and
retired workers which are not reflected in
increased earnings. The treatment of a communicable
disease may also produce additional community
benefits if the probability of further transmission
is reduced (See Klarman [ 34]).
A further potential problem arises in the
intermediate input approach (and in other approaches
which use market data) if the final output is not
sold on a competitive market. In this case a social
shadow price which reflects individuals1 willingness
to pay for incremental units of the final output
should be used to value the output, instead of the
observed market price. For example, to the extent
that farm output is subsidized, a downward adjustment
should be made to valuations based on changes in
130
-------
the value of farm output (See Bain [4 ]).
Shadow prices have also been used to allow for tax
distortions, monopolistic conditions, unemployed
resources and uncertainty. The problems that arise
in estimating shadow prices are very similar to those
that exist for externalities. For a discussion see
McKean [ 47], Margolis [ 46], Mishan [ 49,Ch. 12] and
Haveman and Krutilla [ 27].
Indirect Market Information
It is sometimes possible to obtain information
about the value of an external effect which is not
marketed from individual responses in markets for
related goods. For example, the travel costs an
individual incurs in order to reach a recreation
facility can be interpreted as a part of the price
an individual pays for the use of the facility. Since
people generally live different distances from
recreation facilities, and incur different costs to
consume recreation benefits this information can be
used to estimate a function relating the average
use of a facility per person (visits per 1,000
population per year for example) to the price (i.e.
costs incurred). This relationship and data on
the distribution of people around a facility can
131
-------
then be used to construct a demand curve indicating
how many individuals would use the facility if a
price, in addition to travel costs were imposed for the
use of the facility. The area under this demand
curve is then a consumers1 surplus
measure of net benefits produced by the facility.
This procedure is proposed and discussed, but not
implemented empirically in Clawson and Knetsch
[ 10, Chs. 5,11]. For a good formal discussion
of the theoretical issues and an empirical application
to water based recreation in Missouri see Burt and
Brewer [ 8 ]. For another recent application see
Mansfield { 45J .
In many cases where it is not possible to
readily ascertain the demand for an activity or
physical effect from the demand for related goods,
as was the case in recreation, these markets may
still provide useful information in the form of
bounds on possible costs or benefits. For example,
if households who live near a dump purchase more rat
poison than other households in otherwise similar
circumstances, these expenditures can be interpreted
as a lower bound on the externality vectors from the
dump impose on the community. Similarly, added
132
-------
expenditures on cough medicine may provide a partial
measure of air pollution costs.
This approach has been applied extensively
in the evaluation of pollution and health effects.
Ridker [65 , Ch. 4] estimates some of the costs
of air pollution by examining cleaning, maintenance,
and painting expenditures in polluted and unpolluted
areas. In a similar vein Lave I 42] suggests
that consumer expenditures (including power)
on air filters and precipitators which reduce indoor
pollution would provide a measure of air pollution
costs. Costs currently incurred in the treatment
of a disease will, of course, provide a lower bound
on the benefits the community would receive from its
eradication (See Klarman [ 34]).
The value of individual physical effects,
and of activities taken as a whole, can also be
estimated from individuals' market responses in the
land market. These studies are discussed in detail
in the next section.
C. EMPIRICAL STUDIES RELATING LAND VALUES TO EX-
TERNALITIES
This section has two major parts. In the first
part we discuss problems and issues which arise in
133
-------
the specification of econometric land value studies.
The second part surveys the important features of
a number of major land value studies.
Problems and Issues in Specification
We consider first the empirical problems that
arise when one attempts to measure externalities using
land values or rents. Problems arise because the
parcels of land, and their improvements, are hetero-
geneous commodities. Each parcel of land differs from
all others with respect to its location by definition.
It may also differ in terms of topography, soil
conditions, the accessibility of utilities, zoning
regulations, etc. Improvements may also differ in
terms of a large number of characteristics, all of
which affect the value of the land or its improvements.
The determination pf the effects of an externality on
land values thus becomes a major problem. For example, is
lot A, which is closer to a dump than lot B, less
valuable because of the odors of the dump, because
it has a poorer view, or because it is further from
a local shopping center? The answer is, of course,
possibly all three. In this case, the problem becomes
one of determining how much of any difference in land
values to attribute to each of the characteristics.
134
-------
The usual approach to this problem has been to
postulate that the value of a site i, v., is a function
of the characteristics of the site, zil/...fz. ,
r A ** » *V1^* •* «*•« V»j»fc **».£ &~\+. *•* *i >*l* t. W»*ki «v^. M.
-------
land values by Anderson and Crocker [Ir2], solid
waste disposal sites by Havlicek, Richardson, and
Davies [ 28], and urban renewal by Rothenberg [68,69].
There are a number of methodological issues and
questions which arise in these land value studies.
A major problem is the fact that most of the literature
looks at the value of land and improvements, with
obvious problems. The effects of these improvements
must be controlled for by including their character-
istics in equation (3.1). Many issues arise, however,
in the selection of such variables. It must be deter-
mined which variables, in addition to a variable
representing the effect being studied, determine prop-
erty values and should be included in a regression
analysis. There have been several studies, in
addition to those cited above, exploring this question.
The studies by Kain and Quigley [ 30] and Grether and
Mieszkcwski [ 22] should be noted in particular. These
studies attempted to identify and examine the sig-
nificance of many different characteristics. Kain
and Quigley, using a sample of residential homes,
subjected a total of 39 variables representing the
visual and physical quality of the homes and their
neighborhoods to a factor analysis. They found that
136
-------
a five factor solution "accounts for 60% of the
variance among the 39 original variables and seems
to provide the most meaningful description of the
quality dimensions of residential services" (Kain
and Quigley [30 , p. 534]). The five factors were
identified with: (1) the overall condition of the
structure, landscaping and neighborhood (i.e., the
overall exterior physical environment), (2) the
interior quality of the structure, (3) the condition
of neighboring properties, (4) the effect of
commercial and industrial land use in the immediate
vicinity and (5) the average structural quality of
the neighborhood. The factor analysis does not
provide any information about the economic importance
of any of the factors in determining the price of a
property. The factors do, however, give some
indication of the independent dimensions of properties,
all of which should be considered when one investigates
the determinants of property values.
The paper by Grether and Mieszkowski employs
an even larger set of independent variables. They
explicitly consider three types of variables;
structural, lot and neighborhood characteristics.
Variables in each category proved to be significant
137
-------
in explaining price variations among differing
properties. Overall, a total of 47 variables in
the study explained 79% of the variance in the
prices of a large sample of properties for which
sales data were available in New Haven, Connecticut.
Similar results have been obtained in a number of
other studies, such as Ridker and Henning I 66].
This study is, however, somewhat less
satisfactory in that data aggregated across
census tracts was used. There thus appears to be
fairly wide agreement on the necessity of including
the types of variables discussed above as controls
in property value studies, data permitting.
The identification of all variables which
determine property values is extremely important
since the omission of a relevant variable can have
serious consequences if it is correlated with an
external effect. If, for example, the correct
specification of equation (2.1) is
Vi - - + 6lZil * B2Zi2 + Ei (3'2)
but we estimate instead
v = a + ^z + e* (3.3)
(i.e. the relevant variable z«2 is omitted), then
138
-------
it can be shown that
E(V =6l+
where E (b, ) is the expected value of the ordinary
least squares estimate of £.. from specification
(3.3) and d is the least squares estimate of
z in the descriptive regression equation
Zi2 ' d20 + d21Zil + "i
(For a good accessible discussion of omitted variables
see Kmenta { 35, pp. 792-96]). Thus if variable z^
measures the external effect and z is omitted,
the measured effect of the externality on property
values will be biased unless d = 0. The direction
of the bias will depend on the sign of d and (5 .
&
This result also generalizes to the case of more than
two relevant independent variables.
A more controversial issue has been the inclusion
of buyer or resident characteristics in the regression
equations , by some investigators. In particular,
Anderson and Crocker [ 2 ] have included a resident
income variable in their regressions. They justify
this by deriving the potential purchaser's offer
function for housing, which gives the maximum price
an individual would be willing to pay for a given
139
-------
property and obviously includes both the characteristics
of the property and the income of the individual.
They then note that this function determines the
value of the property "...provided only that each
household occupies the most expensive available
property it is willing to buy, and that each property
is occupied by that household willing to pay the
highest price". They thus conclude that "... if the
(regression) model is to be properly specified, in-
come must be included as an explanatory variable in
its own right" (Anderson and Crocker I 2 , p. 173]).
This is, however, true only if the seller of a
property always knows and receives the maximum
price an individual is willing to pay. The purchaser
of a property therefore never receives any consumer's
surplus from the purchase. This would, however, be
unlikely if a large number of essentially similar
properties are available in the market, as is
usually the case, or if the buyer is as able a
bargainer as the seller. In this case it is the
valuation of the marginal buyers and sellers which
determines the common price for the set of essentially
similar properties.
The issue is thus obviously empirical and also
140
-------
very important, since the willingness to pay for the
elimination of external effects is presumably also
positively related to income. One would thus
expect high income people to locate further from the
source of the externality and pay more for their
property due to, say, lower pollution levels, ceteris
paribus. If the Anderson-Crocker model is correct,
these properties would also be worth more since their
owners had a higher income. The exclusion of an
income variable would thus result in a upward bias
in estimated pollution costs (see the discussion of
omitted variables above).
If, on the other hand, the assumptions of the Anderson-
Crocker model are incorrect, but income is nevertheless
included, the estimate of pollution cost will not be efficient.
more precisely Csee Kiaenta 135, pp. 396-3993), if equation
C3.31 is correct but 03.2) is estimated instead, the sampling
variance of the estimate of &.. will be larger than it would
have been with C3.3) , unless z^ and z.^ are uncorrelated.
.Here empirical work on the role of individual income would
thus be useful in the design of property value studies.
These same i :.3ues arise whenever variables that might
or might not influence property values are considered.
Ey.-lusion of variables that should be included will result
141
-------
in biased estimates of pollution costs, while inclusion of
variables that do not affect property values will lead to
inefficient estimates of these costs. The latter problem
is more acute in small samples than in large ones/ where
the loss of a few degrees of freedom is less serious.
While the role of individual income in the determination
of property values is unclear, it may be sensible to employ
neighborhood income as a proxy for the general amenity of
a neighborhood. As we noted above, the general
characteristics of other properties in a neighborhood
may influence property values. For example, a mansion
surrounded by other mansions is likely to be more
valuable than an identical mansion surrounded by
tract homes. If information about neighborhood
characteristics is not available, neighborhood income
may provide a reasonable proxy.
The use of property values as a price variable
has also been questioned. Wieand I76J has suggested
that the price of residential property is actually
the product of the price of housing and the quantity
of housing, both of which are unobservable. He
then notes that the price of housing, and individuals*
housing expenditures, need not be systematically related,
suggesting that property values are inappropriate.
142
-------
dependent variables in pollution studies. However,
as Crocker I 12] notes the issue turns
on how the commodity housing is defined. If all
properties are regarded as being the same homogeneous
commodity, differing only in certain dimensions,
the approach outlined in (3.1) seems acceptable.
There is also a problem in the Wieand approach in
that his housing quantity variable is unobservable ,
and it is not obvious how a reasonable measure of
his housing price is to be obtained. His use of land
area as a proxy for housing quantity is not convincing,
since the relationship between lot size and the
quantity of housing (exclusive of the lot) may
depend on a number of factors such as relative land
and construction prices and tastes as well as total
housing expenditures.
A functional form must also be specified for
the relationship (3.1). Equations which are linear
in the variables (See Havelicek, Richardson and Davies
I 28J for example) , i.e.
+ u (3-4)
and linear in the logs of the variables (See Anderson
and Crocker I 2]), i.e.
143
-------
log v± = B0 + Bx log zn + + &k log z^ + e± (3.5)
have both been estimated. The model underlying
specification (3.5) can also be written:
vi = e3° zil3:L zi.2>2 '- Zik3k &&i (3>6)
The double log specification (3.5) would appear
to be preferrable on the basis of a priori consider-
ations. The linear specification (3.4) constrains
the marginal effects of a change in characteristics
to be the same for all properties in the sample,
regardless of age or other characteristics. For
example, the imputed price of an additional bedroom
(i.e. the coefficient of the bedroom variable) would
have to be the same for a new, $20,000, two bedroom
house or a 40 year old four bedroom house selling
for $10,000. The double log specification/ on the
other hand, constrains the percentage difference in
prices to be the same for identical percentage
differences in characteristics. In this case all
properties having 50 percent more bedrooms than other
properties would carry the same percentage premium
in price. The linear specification may be particularly
inappropriate for certain measures of external
effects. For example, if the externality variable
144
-------
is distance from a landfill, the linear specification
would imply that the dollar cost of being 100 feet
closer to the landfill is the same for a $10,000
house and a $100,000 house. The double log formu-
lation also allows for the possibility of non-linear-
ities in the way in which units of a characteristic
enter into the value of a property. For example,
the marginal effect of an additional bedroom on
house value will decline as the number of bedrooms
is increased if the coefficient of the bedroom
variable is between 0 and 1. The double log
functional form also allows some interaction between
characteristics since the marginal value of char-
acteristic k, for example, will depend on the levels
of all other characteristics. In other words,
differentiating v. in (3.6) with respect to z we
1 J.JC
obtain :
zik
This last result may, however, also be a short-
coming if there is considerable variation in the
price of properties within the sample. It may, for
example, be unreasonable to allow lot size to affect
the marginal valuations of a garage. The double log
145
-------
specification should thus be used with care in cases
where there are large variations in prices or
characteristics within a sample. For a good discussion
of ways in which individual characteristics should
enter this model see Emerson [17]. Details on the
specification used in particular studies are discussed
in greater detail in the next part of this section.
The quality of some data available for land value
studies has also been questioned. A study of Pittsburg by
Davis and Wertz J13J concludes that assessed property values
"... are not accurate proxies for market values." While
this conclusion may or may not be valid for other jurisdictions,
their results imply that care must be taken in the inter-
pretation of studies using assessed property value as the
dependent variable. In Chapter 6, below, both assessed and
market values are employed, and some evidence of this
problem is encountered.
Census data based on ownerls valuations of their
property have also been used in a number of studies,
The reliability of such data has also been questioned.
A 1954 study by Kish and Lansing 133] indicates that
owner's valuations in the 1950 Survey of Consumer
Finances varied substantially from those of professional
appraisers. The errors tended to be offsetting/ however,
146
-------
with a small ($350) upward bias in the owner's valua-
tion. There is, however, no evidence on how the owner's
or appraiser's valuations relate to actual market values.
There is thus a substantial amount of uncertainty
about the quality of estimated property value data. This
suggests that actual market transactions data, which
is compiled by multiple listings services in many
areas, should be used when it is available.
147
-------
A Survey of Selected Empirical Studies
We will consider next a number of studies of
the determinants of property values, with special
reference to the types of data used, special
specification problems, and a summary of important
results. We consider first studies which attempt to
measure the impact of particular externalities on
land values. We also discuss briefly a number of
related studies on the determinants of the property
values.
The impact of solid waste disposal sites on
property values has been studied by Goldberg et. al.
[ 21] and Havelicek, Richardson and Davies [ 28]. The
study by Havelicek, Richardson and Davies was based on
records from 182 single family dwellings sold during
1962-1970, obtained from the multiple listing
service in Fort Wayne, Indiana. The homes are
situated in the neighborhood of five solid waste
disposal sites. The dependent variable used in
their study was the sales price of the dwelling.
The independent variables used were size of house in
square feet, number of bedrooms and bathrooms, age
of property, lot size, the outstanding amount of
loan, a dummy variable for owner or tenant, the sale
148
-------
date, the number of degrees property is away from
prevailing downwind direction, distance from nearest
disposal site, and dummy variables for the disposal
sites. A linear relationship between price and the
variables mentioned above was used. The shortcomings
of this specification were discussed above. Their
empirical results show that the distance from the fill
and the number of degrees away from the downwind
have significant positive effects on the sale price.
In other words, other things being equal, the further
away from the fill the higher the value the property.
Similarly, the further away from the downwind direction
the higher the value of the property. Their models
explain between 78 and 79 percent of the total
variation in residential property prices.
The regression analysis in Goldberg et. al.
analyzed residential property values in the vicinity
of eight solid waste disposal sites - seven landfills
and one open dump - located in California and Vir-
ginia. The property value data are based on.
assessors records. Other data were obtained from
on-site visits and local records. The dependent
variables used in the regressions were the increase
in property values, with and without improvements,
149
-------
from 1964 to 1970. The price data were obtained
for housing near the disposal sites and in comparison
areas having similar characteristics. Independent
variables were dummy variables representing the
distance from disposal site, visibility of disposal
site, ethnic composition of the area, elevation/
type of disposal site, proximity of truck routes,
and location (California or Virginia). A linear and
log-linear relationship appears to have been estimated.
The elevation, location, type of site, and truck
route variables were significant. The distance
and visibility variables were insignificant. Between
25 to 81 percent of the variation in the independent
variables was explained by the regressions.
The interpretation of this study is, however,
rather unclear. It is not obvious why changes in
land value were chosen as the dependent variable,
since there exists no theory which indicates that
externalities should influence changes or rates of
change in property values. For example, if the
effect of an externality is to depress property
values a constant percentage amount relative to the value of
comparison properties, rates of change in property
values will remain unaffected. Their results are
150
-------
thus not necessarily indicative of the presence or
absence of external costs.
A number of studies have attempted to measure
the effects of air pollution on property values.
The study by Ridker and Kenning I 66] used median
property values in census tracts in the St. Louis
metropolitan area for the year 1960. The use of
census tracts ignores the variation in property
values and characteristics of a property within a
tract. Furthermore, the values are owners' estimates
of the property and may substantially differ from
the market price. However, in the case of air
pollution, the relevant data are generally available
for neighborhoods rather than for individual houses.
The authors therefore, chose to use average census
tract data. The explanatory variables used were:
median number of rooms in the census tract,
percentage recently built, number of houses per mile
(as a proxy for lot size), time zones depending on
express bus travel time to the downtown area, dummy
variables defining accessibility to highways,
shopping centers and industrial areas, school quality,
crime rates, number of persons per unit, the
occupation raio giving the ratio of craftsmen, foremen,
151
-------
operatives and laborers to the total number of
employed persons, percentage of non-white population,
median farm income, and mean sulfur levels as a
measure of air pollution. Ridker and Henning
justified the use of occupation ratio as a measure
of the homogeniety of the neighborhood, the hypothesis
being that in general, people like to live in
neighborhoods that are homogenous with respect to
occupation and social classes. The functional forms
used are linear for all variables except houses per
mile, occupation ratio and percentage of non-white,
all of which appear in a quadratic form also. The
empirical results show that the air pollution variable
was statistically significant in explaining property
values. Higher sulfur concentration implies lower
median value of properties. The independent variables
together explained 87 to 94 percent of the variation
in the median property values.
Anderson and Crocker I 2 3 carried out a study
of the effect of air pollution on residential
property values in the St. Louis, Kansas City and
Washington D.C. standard metropolitan .statistical
areas. Three alternative dependent variables were
used in their study: median property values, median
152
-------
gross rent and median contract grants from census
tracts. The independent variables were: percentage of
non-whites, median family income, percentage of
each tract*s units more than 20 years old, percentage
classified as dilapidated, distance from the central
business district, median number of rooms per housing
unit, sulfur dioxide concentrations in the air and
suspended particulates. At least one of the pollution
variables was significant in each of the regressions
considered and of the right sign, i.e. high levels
of air pollution implied lower property values,
ceteris paribus. There were no significant co-
efficients of the wrong sign. The functional form
used was double log. The models used explained
70 to 79 percent of the variation in the logarithm
of median property values.
A 1971 study by Crocker [12] on property
damage due to air pollution used data from 1,288
FHA insured single family residences in Chicago.
He used several alternative dependent variables,
namely sale price of the residence, FHA estimated
market price, FHA evaluation price as defined by
"the price the typical buyers would be warranted
in paying for the property for long term use or
153
-------
investment", FHA estimated average yearly cost of
maintaining the property, and annual property taxes
as estimated by the FHA. In regressions explaining
the median property values in census tracts the in-
dependent variables used were: median family income/
average income, percentage black, crime rate,
percentage of houses dilapidated, percentage non-
white population, percentage of older houses, school
quality, minimum, maximum and average particulate
concentration and sulfur dioxide concentration. In
regressions using individual house characteristics,
such as the estimated price or annual maintainance
cost,as the dependent variable the independent
variables used were: lot size, area, number of
stories, mortgage term, a dummy variable taking the
value of one if the house had aluminum siding,
distance to the central business district, distance
to the lake, a dummy indicating whether the house
is a frame house or not, age of the house, minimum,
maximum and average particulates and sulfur dioxide
concentrations. The functional form used was double
log except for the dummy variables. The variation
in the dependent variable explained by the regressions
varied from 57 percent for maintainance cost to 80
154
-------
percent for the value of the house. This study also
found that there is a significant negative relation
between price and the other dependent variables and
concentration levels of air pollution.
Other studies, which are not described in
detail here in order to conserve space, relating air
pollution levels to property values include Wieand
I 76], Ridker [65 , Chs. 6,7], Peckham [ 61]. The
results of Ridker [ 65] are- also summarized in
Nourse[ 56]. The only study which differs signifi-
cantly from the studies discussed above is Ridker [65,
Ch. 7]. He constructs two time series indexes of
property values; one for a neighborhood in St. Louis
which experienced a significant increase in pollution
in 1962 and one for a comparable comparison
neighborhood. He finds that the two indexes do appear
to diverge after 1962, but no tests of significance
are presented. Data for the period 1957-1965 were
used in the comparison.
The empirical work relating levels of external
effects to property values also includes studies of
the impact of aircraft noise on property values. The
most extensive study in this area is the one by
Emerson I 17]. It also summarizes other studies in
155
-------
this area. Emerson's study is based on individual
data for approximately 250 residences in the south
Minneapolis area, as obtained from the multiple
listings service. The dependent variable is the
sale price of the residences. A long list of
independent variables includes lot size, area, garage
space, number of bathrooms, number of floors, the
number of built-in appliances, the number of fireplaces,
the age of the house, distance to the commercial
center, distance to school, dummy variables for
corner lots, location on. main street, proximity to
bus routes, siding, carpeting and draperies, and view,
neighborhood compatibility of residence as measured
by the ratio of the price of the given residence to
the average price of residences on the block, access-
ibility and proximity to freeways, the distance to
the nearest open green space and finally a composite
aircraft noise rating. The functional form used is
generally double log except for the dummy variables.
The results indicate that the aircraft nuisance
variable was statistically significant, implying that
the higher the nuisance level the lower the value
of the property.
A less extensive study of aircraft nuisance
156
-------
is the one by Paik [ 58]. His sample consists of
162 cross-section observations, by census blocks,
from the vicinity of Kennedy airport in New York in
1960. The median values of single family dwelling
units was used as the dependent variable. The inde-
pendent variables were the number of"people, average
number of houses, average number of single houses/ num-
ber of non-white" houses/ number of persons per room/
median income, accessibility/ and finally, noise
levels obtained from noise contours based on 1965
flight and operations information. In this study
also, the noise coefficient was highly significant
with a negative coefficient/ indicating that the
higher the levels of noise, the lower the median
value of properties in that area. The only functional
form used was a straight linear relation.
The effect of freeway traffic noise on residential
property values and apartment rents has also been
studied. The study by Colony [11 ] was based on 91
parcels of land in Toledo/ Ohio. The dependent
variable used was the 1965 estimated sale price
expressed as a percentage of the 1951 estimated
sale price. The independent variables were hori-
zontal distance from the expressway center line to
157
-------
the nearest building corner, vertical distance from
expressway pavement to lot grade, and sound levels
interpolated from sound level contour maps. Since
the study did not present any standard error for the
regression coefficients, it's not possible to conclude
whether or not the variables included in the re-
gression are statistically significant. It should
also be noted that their specification of the
dependent variable is subject to the same qualifica-
tions as the specification of the dependent variables
in the Goldberg et.al.I 21] waste disposal site
study discussed above.
The study by Towne and Associates [ 72]
investigates the effect of freeway traffic noise on
apartment rents. The independent variables used
in their study were floor area, bedrooms, vacancy
rate, noise during rush hour, non-rush hour and at
night, distance to freeway route access as well as
to the central business district and shopping center,
the type of shopping center, distance to school and
recreation areas, quality of the apartments, number
of units in the building, air-conditioning, type of
floor, and neighborhood characteristics. The study
concluded that the effect of freeway noise on
158
-------
apartment rents was insignificant.
The impact of reservoirs on land values has also
been investigated. Reservoirs presumably produce
positive externalities in the form of recreation
possibilities. A study by Knetsch I 37] used the
price per acre of individual land parcels near 11
of the Tennessee Valley Authority reservoirs, as
the dependent variable. The list of independent
variables includes the distance from the reservoir/
the square of the distance/ a topography index, a
dummy variable which took the value one if the lot was
directly overlooking the reservoir, the value of
improvements, cost of development per acre, the
average work week as obtained from U.S. Department
of Labor data, driving time from nearby urban centers
and total retail sales of nearby population centers.
Total sales was used rather than population to reflect
the differences in the standard of living that might
be exhibited by different cities having the same
population. The relationship estimated was linear.
The results indicate that tracts with a lake frontage
were generally more expensive than similar tracts
without lake frontage. It was also found that the
value of a property declined as the distance from the
159
-------
lake increased, but at a much lower rate.
Williams and Daniel [ 77] also investigated the
relationship between land values and reservoirs.
They studied properties located near the outskirts
of Jackson City in Mississippi. As in the Knetsch
study, the dependent variable was price per acre.
The explanatory variables were the number of acres
involved in the transaction, the percentage of open
land, the accessibility of the land as measured by
the type of road frontage, estimated distance of land
from interstate highway, an index of farm real
estate values in Mississippi, approximate distance
of the land from the reservoir and a measure of the
effect of growth in the northeast Jackson area.
The results show that the distance to the reservoir
has a significant negative effect on the price per
acre. In other words, the further away from the
reservoir, the less the value of the property.
The effect of local public school expenditures
and their effects on house values has been studied
by Pashigian [59 ]. He used the natural logarithm
of the median value of an owner occupied house in
1970 as the dependent variable. The explanatory
variables were 1970 assessed value per average
160
-------
daily attendance in school divided by average
number of rooms, median number of rooms in owner
occupied units in 1970, median family income in
1960, median years of education for persons older
than 24 years in I960, percent non-white in munic-
ipality in 1970, ratio of 1970 average daily attendance
to the 1960 average daily attendance, a dummy var-
iable for house value, current expenditures per
average daily attendance in 1970, percentage of
teachers with masters degree, student-teacher ratio,
and average number of total years of experience of
teachers. Except for the house value and current
expenditure, the remaining independent variables
were expressed in their natural logarithms. The
study indicates that house value increases with
rising expenditure per average daily attendances,
is not significantly affected by the ratio of
teachers with masters degrees to teachers with
bachelors degrees, rises with decreases in the
student-teacher ratio, and falls with increases
in the number of years of experience of a given
teacher. While the results of the study may be
interesting, the formulation implicitly assumes that
the direction of causation is from expenditure per
student to property values. It is more plausible,
however, to argue that expenditure per
161
-------
student should depend on the average property value,
because it's the latter that determines the revenue
to the government in the form of property taxes.
In a similar study, Gates [ 57] studied the
effect of local taxes and expenditures on the median
value of owner occupied dwellings in 53 New Jersey
municipalities. Independent variables used were
percentage of houses constructed since 1950, median
family income, distance from Manhattan, annual
expenditures per pupil in public schools, the
effective property tax rate and percentage of poor
families. The tax, expenditure, and distance from
Manhattan variables were expressed in their natural
logarithms. The remaining variables entered linearly.
The equations were estimated by ordinary least squares
and two stage least squares. The latter estimation
procedure was used to allow for the possibility
of simultaneous equation bias since, as was noted
in our discussion of Pashigian [59 ], property
values may also have an effect on expenditures and
taxes. These variables are therefore regarded as
being endogenous. The tax variable is significantly
negative and the expenditure variable significantly
positive in both regressions, indicating that higher
162
-------
taxes tend to depress property values and high public
expenditures tend to increase property values,
ceteris paribus. The regressions explained 93
percent of the variance in property values.
The use of land value or rent differentials as
a measure of non-market effects has also been sug-
gested, and in some cases applied, in a number of
other areas. Mohring [50,51,52 ] suggests the use of
and uses differentials in land values as a measure
of the value of time saved due to transportation
improvements. Other suggested applications include
estimating the value of irrigation water (Milliman
I 48J) , the value of parks (Hendon [29 ]), the
benefits of urban renewal (Rothenberg [ 68, 69]),
flood control benefits (Lind [ 43]), -and the effects
of one-way streets (Kennedy [ 313)• These studies
are not discussed since they do not contain empir-
ical analyses as detailed as those discussed above,
or are entirely conceptual.
There are also a number of studies on land values
which are not directly concerned with the measure-
ment of particular externalities or program benefits.
In addition to the study by Kain and Quigley [ 30]
discussed in the first part of this section, there
163
-------
is a study by Downing [ 16] of factors affecting
commercial land values. The independent variables
used in the study are distance to the central
business district, distance to the shopping center,
traffic level on main street, the corner influence,
population, median family income in the census, tract,
percentage of non-white, the average size of the
lot, and a time trend. The functional form used was
generally linear except for distance to the central
business district and distance to the nearest regional
shopping center. The distance, corner influence, median
income, and zone 13 dummy variables were not
statistically significant. All the other variables
had statistically significant regression coefficients.
The regressions explained about 44 percent of the
variation in commercial land values.
Grether and Mieszkowski I 22J studied the
determinants of real estate values in the New Haven
area using multiple listing data. They used
a large number of detailed characteristics of a
property. Their formulation includes age of the property
in quadratic form. in a similar way, the lot size
was entered the formulation in quadratic form. To
capture the effect of time on property values,
164
-------
a time trend was expressed in an exponential form.
The study also indicated the presence of hetero-
scadesticity, indicating larger variation associated
with larger transactions. To correct for this, they
divided each observation by the size of the house/
thus transforming the independent variable to price
per square foot.
There is also a study by King and Mieszkowksi
I 32J of rents in the New Haven area which includes
the racial and sex characteristics of renters and
the racial characteristics of the neighborhood in
a large list of independent variables. These
variables were significant in some cases, but it
was not clear whether supply or demand effects were
being estimated. The regressions explained 77-78
percent of the variation in rents in the sample.
Gillingham [ 20] studied the determinants of
apartment rents in 10 major U.S. cities using up
to 29 independent variables, relating to apartment,
building and neighborhood characteristics. The
dependent variable was natural logarithm of rents.
Independent variables entered linearly. The re-
gression explained 53-82 percent of the variation in
rents.
165
-------
The recent book by Harris [ 25] should also
be cited. It appears to contain a number of relevant
papers, but was received too late to be included here.
166
-------
REFERENCES
1. Anderson, R. J., and T. D. Crocker. Air Pollution and
Property Values: A Reply. Review of Economics and
Statistics. Vol. 54 (November 1972): p.470-473.
2. Anderson, R. J. , and R. D. Crocker. Air Pollution and
Residential Property Values. Urban Studies. Vol. 8
(October 1971): p. 171-180.
3. Anderson, R. J., and R. D. Crocker. The Economics
of Air Pollution: A Literature Assessment. In: Air
Pollution and the Social Sciences, Downing, P. B.
(ed.).New York, Praeger Publishers, 1971. p. 133-166.
4. Bain, J. S. Water Resources Development in California.
In: Water Research, Kneese, A. V. and S. C. Smith (eds.).
Washington, Resources .for the Future, 1965. p. 51-67.
5. Becker, G. S. Human Capital. New York, N.B.E.R.,
1964.
6. Bos, H. C., and L. M. Koyck. The Appraisal of Road
Construction Projects: A Practical Example. Review
of Economics and Statistics. Vol. 42 (February 1961):
p. 13-20.
7. Brown, G. M., and J. Hammack. A Preliminary Investigation
of the Economics of Migratory Waterfowl. In: Natural
Environments: Studies in Theoretical and Applied
Analysis, Krutilla, J. V. (ed.). Baltimore, Resources
for the Future, 1972. p. 171-204.
8. Burt, O. R. and D. Brewer. Estimation of Net Social
Benefits from Outdoor Recreation. Econometrica.
Vol. 39 (September 1971): p. 813-828.
9. Chase, S. B. (ed.). Problems of Public Expenditure
Analysis. Washington, D.C., The Brookings Institute,
1968.
10. Clawson, M. and J. L. Knetsch. Economics of Outdoor
Recreation. Baltimore, Johns Hopkins Press, 1966.
167
-------
11. Colony, D. C. Study of the Effect, If Any, of an
Urban Freeway Upon Residential Properties Contiguous
to the Right of Way. Toledo University (For State
of Ohio, Department of Highways) 1968.
12. Crocker, Thomas D. Urban Air Pollution Damage Functions:
Theory and Measurement. (Prepared for the Environmental
Protection Agency, Office of Air Programs. June 1971.)
13. Davis, O. A., and K. L. Wertz. The Consistency of
the Assessment of Property: Some Empirical Results
and Managerial Suggestions. Applied Economics.
Vol. 1 (1969): p. 151-157.
14. Dorfman, R. (ed.). Measuring Benefits of Government
Investments. Washington, D.C., The Brookings Inst.,
1963.
15. Downing, Paul B, Estimating Residential Land Value
by Multivariate Analysis. In: The Assessment of Land
Value, Holland, Daniel M. (ed.).Madison,University
of Wisconsin Press, 1970. p. 101-125.
16. Downing, Paul B. Factors Affecting Commercial Land
Values: An Empirical Study of Milwaukee, Wisconsin.
(Working Paper #13. Program in Environmental Econo-
mics. University of California, Riverside. 1371.)
17. Emerson, F. C. The Determinants of Residential Value
with Special Reference to the Effects of Aircraft
Nuisance and Other Environmental Features. (Ph.D.
Thesis. University of Minnesota. 1969.)
18. Fisher, A. C. , and J. V. Krutilla. Determination of
Optimal Capacity of Resource-Based Recreation Facili-
ties. In: Natural Environments: Studies in Theoretical
and Applied Analysis, Krutilla, J. V.(ed.).Baltimore,
Resources for the Future, 1972. p. 115-141.
19. Foster, C. D., and M. E. Beesley. Estimating the
Social Benefit of Constructing an Underground Railway
in London. Journal of the Royal Statistical Society.
Series A, Vol. 126 (1963): p. 46-92.
20. Gillingham, R. F. Place to Place Comparisons Using
Hedonic Quality Adjustment Techniques. Research
Division, Office of Prices and Living Conditions,
U.S. Bureau of Labor Statistics. Research Discussion
Paper No. 7. March 1973.
168
-------
21. Goldberg, L., et. al. The Effect of Solid Waste
Disposal Sites on Property Values. Environmental
Protection Agency, Washington, D.C. 1972.
22. Grether, D. M. and P. Mieszkowski. Determinants
of Real Estate Values. (Unpublished, c. 1972.)
23. Griliches, Zvi. Hedonic Price Indexes Revisited.
In: Price Indexes and Quality Change: Studies
in New Methods of Measurement, Griliches, Z. (ed.).
Cambridge, Harvard University Press, 1971. p. 3-15.
24. Gronau, R. The Value of Time in Passenger Travel:
The Demand for Air Travel. New York, N. B. E. R.
N. B. E. R. Occasional Paper 109. 1970.
25. Harris, C. L. Government Spending and Land Values.
Madison, The University of Wisconsin Press, 1973.
26. Haveman, R. H., and J. V. Krutilla. Unemployment,
Idle Capacity, and the Evaluation of Public Expenditures.
Baltimore, The Johns Hopkins Press, 1968.
27. Haveman, R. H., and J. Margolis. Public Expenditures
and Policy Analysis. Chicago, Markham Publishing
Co., 1970.
28. Havlicek, J., R.Richardson, and L. Davies. Measuring
the Impacts of Solid Waste Disposal Site Location
on Property Values. (Unpublished, c. 1972.)
29. Hendon, W. S. The Park as a Determinant of Property
Values. American Journal of Economics and Sociology.
Vol. 30 (July 1971): p. 289-300.
30. Kain, J. F. , and J. M. Quigley. Measuring the Value
of Housing Quality. Journal of the American Statistical
Association. Vol. 65 (June 1970): p. 532-548.
31. Kennedy, John, and Dennis Hull. The Economic and
Environmental Effects of One-Way Streets in Residential
Areas. Michigan Dept. of State Highways, 1969.
32. King, A. T., and P. Mieszkinski. Racial Discrimination,
Segregation, and the Price of Housing. Journal of
Political Economy. Vol. 81 (May/June 1973): p. 590-606
169
-------
33. Kish, L., and J. B. Lansing. Response Errors in Esti-
mating the Value of Homes. Journal of the American
Statistical Association. Vol. 49 (September 1954).
34. Klarman, H. E. Syphilis Control Programs. In:
Measuring Benefits of Government Investments, Dorfman,
R^(ed.).Washington, D.C., The Brookings Inst., 1963.
p. 367-409.
35. Kmenta, J. Elements of Econometrics. New York, The
McMillan Co., 1971.
36. Kneese/ A. V., and S. C. Smith (eds.). Water Research.
Washington, Resources for the Future, 1965.
37. Knetsch, J. L. The Influence of Reservoir Projects
on Land Values. Journal of Farm Economics. (February
1964): p. 231-243.
38. Knetsch, J. L. and R. R. Davis. Comparison of Methods
for Recreation Evaluation. In: Water Research, Kneese,
A. V., and S. C. Smith (eds.). Washington, Resources
for the Future, 1965. p. 125-143.
39. Krutilla, J. V. (ed.). Natural Environments: Studies
in Theoretical and Applied" Analysis.Baltimore,
Resources for the Future, 1972.
40. Krutilla, John V., and Otto Eckstein. Multiple Purpose
River Development. Baltimore, Johns Hopkins Press, 1958,
41. Lancaster, K. A New Approach to Consumer Theory.
Journal of Political Economy. Vol. 74 (April 1966):
p. 132-157.
42. Lave, L. B. Air Pollution Damage: Some Difficulties
in Estimating the Value of Abatement. In: Environ-
mental Quality Analysis, A. V. Kneese and B. T. Bower,
(eds.).Baltimore, The Johns Hopkins Press, 1972.
p. 213-242.
43. Lind, R. C. Flood Control Alternatives and the Eco-
nomics of Flood Protection. Water Resources Research.
Vol. 3 (1967): p. 345-357.
44. Mack, R. P., and S. Myers. Outdoor Recreation. In:
Measuring Benefits of Government Investments, Dorfman,
R. (ed.). Washington, D.C., The Brookings Inst., 1963.
p. 71-116.
170
-------
45. Mansfield, N. W. The Estimation of Benefits from
Recreation Sites and the Provision of a New Recreation
Facility. Regional Studies. Vol. 5 (1971): p. 55-69.
46. Margolis/ J. Shadow Prices for Incorrect or Nonexis-
tent Market Values. In: Haveman and Margolis,
Public Expenditures and Policy Analysis. Chicago,
Markham Publishing Co., 1970. p. 314-329.
47. McKean, R. N. The Use of Shadow Prices. In: Chase,
S. B., Problems of Public Expenditure Analysis.
Washington, D.C., The Brookings Inst., 1968. p. 34-64.
48. Milliman, J. W. Land Values as Measures of Primary
Irrigation Benefits. Journal of Farm Economics.
(May 1959): p. 236-249.
49. Mishan, E. J. Cost-Benefit Analysis: An Informal
Introduction. London, George Allen & Unwin Ltd., 1971.
50. Mohring, H. Land Values and the Measurement of Highway
Benefits. Journal of Political Economy. Vol. 69
(June 1961): p. 236-249.
51. Mohring, H. Urban Highway Investments. In: Measuring
Benefits of Government Investments, Dorfman, R~T(ed.).
Washington, D.C., The Brookings Inst., 1963. p.231-274.
52. Mohring, H., and M. Harwitz. Highway Benefits; An
Analytical Framework. Evanston, Northwestern University
Press, 1962.
53. Moses, L. N., and H. F. Williamson Jr. Value of Time,
Choice of Mode, and the Subsidy Issue in Urban Trans-
portation. Journal of Political Economy. Vol. 71
(June 1963): p. 247-264.
54. Musgrave, R. A. Cost-Benefit Analysis and the Theory
of Public Finance. Journal of Economic Literature.
(September 1969): p. 797-806.
55. Mushkin, S. J. Health as an Investment. Journal of
Political Economy. Vol. 70, Supplement (October 1962):
p. 106-123.
56. Nourse, Hugh 0. The Effect of Air Pollution on House
Values. Land Economics. Vol. 43 (May 1967): p. 181-189,
171
-------
57. Dates/ Wallace E. The Effects of Property Taxes and
Local Public Spending on Property Values: An Empirical
Study of Tax Capitalization and the Tiebout Hypothesis.
Journal of Political Economy. Vol. 77 (December 1969):
p. 957-971.
58. Paik, Inja K. Impact of Transportation Noise on Urban
Residential Property Values with Special Reference
to Aircraft Noise. (Consortium of Universities, Urban
Transportation Center, Washington D.C. August 1970.)
59. Pashigan, B. P. Local Public School Expenditures and
their Effects on House Values. Center for Mathematical
Studies in Business and Economics, University of Chicago,
Center Report 7238. 1972.
60. Paul, S. An Application of Cost-Benefit Analysis to
Management Education. Journal of Political Economy.
Vol. 80 (March/April 1972): p. 328-346.
61. Peckham, B. W. Air Pollution Damages to Residential
Properties in Philadelphia. (Unpublished, 1970).
62. Prest, A. R., and R. Turvey. Cost-Benefit Analysis:
A Survey. Economic Journal. Vol. 75 (December 1965):
p. 683-735.
63. Quandt, Richard E., and William J. Baumol. The Demand
for Abstract Transport Modes: Theory and Measurement.
Journal of Regional Science. Vol. 6 (1966): p. 13-26.
64. Rice, D. P. Estimating the Cost of Illness. U. S.
Government Printing Office, Washington D. C. U. S.
Public Health Service Publication No. 947-6. 1966.
65. Ridker, R. G. Economic Costs of Air Pollution;
Studies in Measurement. New York, Praeger Publishers,
1967.
66. Ridker, R. G., and J. A. Henning. The Determinants
of Residential Property Values with Special Reference
to Air Pollution. Review of Economic Statistics.
(May 1967): p. 246-257.
67. Rosen, S. Hedonic Prices and Implicit Markets:
Product Differentiation in Pure Competition. Harvard
Institute of Economic Research Discussion Paper No.
296. May 1973.
172
-------
68. Rothenberg, Jerome. Economic Evaluation of Urban
Renewal. Washington/ D.C., Brookings Institution,
1967.
69. Rothenberg, Jerome. Urban Renewal Programs. In: Measu-
ring Benefits of Government Investments, Dorfman, R. (ed.)
Washington^ b.C.,The Brookings Institution, 1963.
p. 292-341.
70. Schelling, T. C. The Life You Save May Be Your Own.
In: " Problems^of Public Expenditure Analysis, Chase,
S. B. (ed.). Washington, D.C., The Brookings Institute,
1968. p. 127-161.
71. Tinbergen, J. The Appraisal of Road Construction:
Two Calculation Schemes. Review of Economics and
Statistics. Vol. 48 (August 1957): p. 241-249.
72. Towne, Roben M., and Associates. An Investigation
of the Effect of Freeway Traffic Noise on Apartment
Rents. Seattle, October 1966.
73. Weisbrod, B. A. Costs and Benefits of Medical Research:
A Case Study of Poliomyelitis. Journal of Political
Economy. Vol. 79 (May/June 1971): p. 527-44.
74. Weisbrod, B. A. Economics of Public Health: Measuring
the Economic Impact of Diseases. Philadelphia, Uni-
versity of Pennsylvania Press, 1960.
75. Weisbrod, B. A. Education and Investment in Human
Capital. Journal of Political Economy. Vol. 70,
Supplement (October 1962): p. 106-123.
76. Wieand, K. F. Property Values and the Demand for
Clean Air: Cross-Section Study for St. Louis.
Appendix A In: Crocker, Thomas D. Urban Air Pollution
Damage Functions: Theory and Measurement. (Prepared
for the Environmental Protection Agency, Office of
Air Programs. June 1971.) p. A1-A18.
77. Williams, D. C. , and L. Daniel. The Impact of Reser-
voirs on Land Values: A Case Study. Mississippi
State University State College. June 1969.
173
-------
CHAPTER IV
A SURVEY OF THE TECHNICAL ASPECTS
OF SOLID WASTE EXTERNALITIES
A. INTRODUCTION
Although our report focuses on the measurement of
externalities associated with solid waste through their
effect on economic variables, this chapter summarizes
what is known about their physical manifestations. We
do not consider here those aspects which may affect
economic variables but do not have measurable physical
characteristics, such as the unsightliness of dumps.
We do survey the literature with respect to both the
qualitative nature and the magnitudes of those external-
ities connected with the transfer, processing, and disposal
of solid wastes. We have attempted to be particularly
thorough in establishing what is known about the extern-
alities associated with sanitary landfills, and dumps,
particularly the adverse effects related to gas production,
surface and groundwater pollution, and the propagation of
vectors and vermin. We also note the apparent effects on
these externalities of not meeting the requirements of
a proper sanitary landfill. However, we discuss, at least
briefly, all of the externalities of which we are aware.
An important reason for including this survey is that
174
-------
the more clearly the relevant "technology" is under-
stood by an economic analyst, the more likely that the
assumptions of a related model will be appropriate.
This is true of not only the qualitative nature of the
assumptions but also of the implicit assumptions which
must be made, in any empirical study, about the quan-
titative importance of different factors. A second
important reason is that the indirect methods of valuing
externalities (using, for instance, property value data)
discussed in the last two chapters rely heavily on the
assumption of perfect knowledge. If an individual mis-
perceives his environment, his economic activity will not
correctly reflect changes in this environment. In the
present context, the assumption of perfect knowledge
requires all traders to be fully aware of all implications
of all relevant physical externalities. For instance,
.both the disagreeable aspects of air pollution and its
specific effects on health must be known by all traders.
As was discussed earlier, if this assumption is clearly
inappropriate, valuation of a particular externality must
be done directly, and the investigator must have detailed
knowledge of its physical effects.
Recognizing the multi-disciplinary nature of prob-
175
-------
lems associated with solid waste disposal, we had
occasion to consult diverse sources in our literature
survey. We would like to acknowledge here the prompt
and courteous assistance we received from MEDLINE lit-
erature searches obtained through the reference depart-
ment of the UCSD Biomedical Library and from BIOLOGICAL
ABSTRACTS, BIORESEARCH INDEX, and CHEMICAL ABSTRACTS
searches provided via the Center for Information Services
and the reference department of the UCSD Science and
Engineering Library.
On most of the topics we discuss, one or more
surveys of the literature already exist. This survey
clearly does not attempt to comprehensively reproduce the
extensive detail found in some of these. Rather, we
have attempted to provide a survey which will be use-
ful to economists or others who are not experts in the
area. We go into greater detail on those questions
which seem directly relevant to the economic analysis
contained in the other chapters of the report.
B. OVERVIEW
Definition of Solid Waste
176
-------
Solid waste may be defined as any solid matter that
is discarded or thrown away. The U.S. Environmental
Protection Agency feels that the term solid waste should
be defined to include: "garbage, refuse, and other dis-
carded solid materials, including solid wastes, materials
resulting from industrial, commercial, and agricultural
activities, and from community activities" and should
not include: "solids or dissolved material in domestic
sewage or other significant pollutants in water resources,
such as silt, dissolved or suspended solids in industrial
wastewater effluents, dissolved materials in irrigation
return flows or other common water pollutants" [135,p.20]
Table 4.1 [135, adapted from 44] presents a general
classification of solid wastes. Although such a general
classification includes dead animals, abandoned vehicles,
and special wastes of many types, these account for only
a minor portion of all solid wastes and generally must
be handled with specialized procedures. Also, various
types and quantities of industrial and agricultural wastes
are managed by the waste producers themselves and do not
enter general municipal solid waste management systems.
In this report we shall concentrate on the bulk of solid
wastes, namely the general refuse, garbage, and so forth
which is usually collected by municipal agencies.
177
-------
Table 4.1 GENERAL CLASSIFICATION OF
SOLID WASTE MATERIALS
Garbage
Rubbish
Ashes
Bulky
wastes
Street
refuse
Dead
animals
Abandoned
vehicles
Construction
& demolition
wastes
Industrial
refuse
Special
wastes
Animal and
agricultural
wastes
Sewage
treatment
residues
Wastes from the preparation, cooking, and serving
of food
Market refuse, waste from the handling, storage,
and sale of produce and meats
Combustible
(primarily
organic)
Noncombustible
(primarily
inorganic)
Paper, cardboard, cartons
Wood, boxes, excelsior
Plastics
Rags, cloth, bedding
Leather, rubber
Grass, leaves, yard trimmings
Metals, tin cans, metal foils
Dirt
Stones, bricks, ceramics,
crockery
Glass, bottles
Other mineral refuse
Residue from fires used for cooking and for heat-
ing buildings, cinders
Large auto parts, tires
Stoves, refrigerators, other large appliances
Furniture, large crates
Trees, branches, palm fronds, stumps, flotage
Street sweepings, dirt
Leaves
Catch basin dirt
Contents of litter receptacles
HSmall animals: cats, dogs, poultry, etc.
Large animals: horses, cows, etc.
Automobiles, trucks
Lumber, roofing, and sheathing scraps
Rubble, broken concrete, plaster, etc.
Conduit, pipe, wire, insulation, etc.
Solid wastes resulting from industrial processes
and manufacturing operations, such as food-
processing wastes, boiler house cinders, wood,
plastic, and metal scraps and shavings, etc.
Hazardous wastes: pathlogical wastes, explosives,
radioactive materials
Security wastes: confidential documents, negotiable
papers, etc.
Manures, crop residues
Coarse screenings, grit, septic tank sludge, de-
watered sludge
From
households,
institutions,
and commercial
concerns such
as:
hotels,
stores,
restaurants,
markets, etc.
\
From
streets,
sidewalks,
alleys,
vacant lots, etc.
From
factories,
power plants,
etc.
Households,
hospitals,
institutions,
stores,
industry, etc.
Farms,
feed lots
Sewage treat-
ment plants,
septic tanks
Source: [135, p. 42]
178
-------
Quantity
The quantity of municipal refuse generated annually
in the U.S. is enormous. According to available estimates
[16,144] a national average of 5.32 pounds of solid wastes
per person per day was collected in 1967. Of this,
approximately 3 Ibs. is estimated to be domestic in
origin, 1 Ib. commercial, .59 Ib. industrial, .18 Ib.
demolition and construction, and .55 Ib. other [16].
Applying a conservative 4% annual increase in per capita
generation to this figure yields a daily per capita
collection figure of 7.0 pounds for 1974. Assuming a
national population of about 210 million in 1974, this means
that roughly 268 million tons of solid waste are collected
annually. In an uncojnpacted state (density of"8-10 Ibs/
cubic foot) this quantity of wastes could easily cover
the state of Delaware to a depth of one foot. But these
figures exclude 10-15% of household and commercial wastes
and 30-40% of industrail wastes that are self-collected
and transported, as well as 3.1-3.7 billion tons of
agricultural waste and crop residues, animal wastes, and
mineral wastes [16,46].
Much less domestic refuse is generated in Europe
than in the U.S. European domestic refuse generation
has recently been estimated at 1.1-2.64 Ibs/person/day
179
-------
(.5-1.2 kg/person/day) [176,ref (3);93]. Solid waste
authorities there, however, feel that the evolution of
solid waste characteristics in Europe is similar to
that of the U.S. but lags about 10 years behind.
Waste Management
This huge quantity of municipal refuse is subject
to four stages in the management process: storage,
collection and transportation, processing and final
disposal. Many different techniques can be employed at
each stage. For example, refuse in New York City may be
stored in plastic bags, collected at the street curb,
transported by compactor trucks to large multi-chamber
incinerators for processing, and the incinerator residues
finally dumped at sea. In rural areas, refuse may be
stored in paper bags or metal cans, collected by an open
truck, and transported to an open dump. Consequently,
the environmental impact of the overall management sys-
tem is radically affected by the particular management
techniques which are employed.
It is also important to recognize that altering
one or more techniques within the solid waste management
system can produce effects not only internally—that is,
on other components of the solid waste management system-
bat also on the related problems of managing our air
180
-------
and water resources. Encouraging the use of home garbage
disposals may reduce the solid waste management system
but increase the burden on the sewage system. Putting
a tax on the solid waste generated by households may
encourage people to burn much of their trash, which may
be illegal but effectively impossible to control outside
the cities. In many ways, there are clear connections among
the management systems for solid, liquid, and gaseous
wastes, and ideally all of these should be analyzed as
one interrelated problem. In this report, we consider
the direct external effects on water and air quality
of solid waste techniques, but not the possible
indirect effects.
General studies have been completed which survey
present or planned solid waste management practices in the
U.S. For the following areas see for example: California
[2,31,159,160]? Fresno [41; Kentucky [100]; New Orleans
[187]; Oregon [139]; Pennsylvania [143]; San Francisco
[35,158]; United States (national) [16,25,42,120,129,
130,131,132,135].
Storage
Refuse is usually stored in containers at its source
prior to collection arid,transportation for various
periods, depending on collection frequency. A substantial
181
-------
and increasing portion of household refuse, however, is
being processed through garbage grinders, decreasing
the amount of refuse stored at domestic sources and
increasing the treatment burden of wastewater treatment
facilities.
A variety of containers are used to store refuse.
The choice of containers is influenced by factors of
economy, socio-economic conditions, convenience, climate,
and municipal regulations. The most common refuse
container is undoubtably the familiar metal can, used
either singly or in conjunction with paper or plastic
bags [14,44]. Other containers receiving increased use
in recent years, due to their advantages 'of convenience,
increased collection efficiency, or improved sanitation
are: paper bags, plastic bags, large metal bins, and
special containers such as sealed paper bags prepared by
home refuse compactors [161].
Collection and Transportation
In terms of employment and capital expenditures,
refuse collection is by far the most important phase
of solid waste management. It has been reported [16]
that of the total management budget, 80% was spent on
collection and 20% on processing and disposal. About
90% of municipal refuse is collected by crews of men work-
182
-------
ing in conjunction with trucks. In 1968, 337,000 men
were employed in the United States in collecting and
transporting refuse, using 93,000 compactor trucks and
179,000 other collection vehicles to transport the
refuse to disposal facilities. Both public and private
agencies are involved in collection activities, with
public services dominating in the collection of household
wastes (51%) and private services dominating in commercial
and industrial waste collection (62% and 57% respectively).
Approximately 33% of the population lives in communities
with separate collection systems in which garbage or
rubbish must be separated for collection; the remaining
66% live in communities with combined collection systems
or with both combined and separate systems. For combined
collection systems 48% is collected once per week, 32%
twice per week, and 20% at other frequencies. For separate
systems, 61% of the material is collected once per week,
29% twice per week, 3% at other frequencies, and the
remaining 7% of the garbage, although separated, is not
collected.
Processing
After collection more than 90% of the municipal
refuse is transported to final disposal facilities,
primarily dumps and sanitary landfills; the other 10%
183
-------
is processed. The goals of waste processing are volume
reduction or the separation of waste components.
Although commonly considered a final disposal method,
incineration is actually a volume reduction technique.
Typically, municipal refuse is reduced to 5-10% of its
original volume and 20-25% of its original weight through
incineration [45,50]. There were 289 community incin-
erators built between 1945 and 1965 in the U.S. As of
December 31, 1966, there was a total installed daily
refuse incineration capacity of 74,600 tons [42]. In
contrast to other solid waste management facilities,
96% of these are publically owned.
Operating costs for refuse incinerators are quite
high: $4-$6 per ton or higher compared to $.50-$1.25
per ton for most landfill disposal methods [1,16,26,129,
173], Despite these high costs, incineration is quite
common in metropolitan and other heavily populated areas
because it requires much less land than dumping or sanitary
landfilling of unprocessed refuse.
Many other waste processing techniques have been
tried or are currently being used for demonstration
projects of waste processing methods for the purpose
of volume reduction or resource recovery. These methods
include: various types of separation processes, chemical
184
-------
processing methods, mechanical densification and size
reduction techniques, composting, and experimental
resource recovery methods for metals, glass, and paper.
Most of these processes have been rejected as financially
or technically infeasible for widespread use, but re-
search and development work continues on many techniques,
notably pyrolysis, composting, baling, air classification,
and gravity, mechanical, or magnetic separation. For
information and further references on processing techniques
the reader may consult, for example, Boettcher [18],
Breidenbach [23], City of San Diego [165], Engdahl [64],
Public Works [149], Wolf et al [198],
Final Disposal
The last step in solid waste management is final
disposal of waste material, usually by sanitary land-
filling or open dumping.
Sanitary landfilling is a specialized method of land
disposal of solid wastes. A sanitary landfill is defined
by the American Society of Civil Engineers as: "A method
of disposing of refuse on land without creating nuisances
or hazards to public health or safety, by utilizing the
principles of engineering to confine the refuse to the
smallest practical area, to reduce it to the smallest
practical volume, and to cover it with a layer of earth
185
-------
at the conclusion of each day's operation, or at such
more frequent intervals as may be necessary" [43].
Shorter and more explicit working definitions of a sanitary
landfill include: "criteria for a site to be termed
'sanitary': (1) the site must have daily cover; (2) the
site must not practice open burning; (3) the site must
not have water pollution problems" [16]; and"...a process
of dumping and compacting the solid wastes to the smallest
practical volume and covering them daily with compacted
earth in a systematic and sanitary manner" [159]. In
any case, the basic operational characteristic of a sanitary
landfill is the complete daily covering of all refuse
with compacted earth. Specific operational and engineering
techniques vary from site to site depending on such
factors as site topography and hydrogeology, climate,
available equipment, planned uses, and so forth. Generally
speaking, refuse is dumped in a prepared portion of the site;
it is spread in thin layers and compacted by heavy
equipment, usually specially equipped buldozers or
tractors. The compacted refuse is covered daily or more
frequently, if necessary, with earth. The earth layer
is compacted daily.
Land disposal sites which do not meet all of the
criteria for sanitary landfills are classified as open
186
-------
dumps. Naturally, this subsumes a wide variety of
land disposal practices, ranging from uncontrolled,
open burning dumps to landfill sites which do not receive
a daily earth cover but otherwise meet the requirements
for a sanitary landfill or which have minor water
pollution problems. In discussing dumping we will use
the term "open dump" as described below and will ex-
plicitly consider those sanitary landfills which do not
meet all criteria.
Open dumps may be easily distinguished from san-
itary landfills. An open dump is simply a plot of land,
frequently a ravine, natural depression, marshland
or other site, either privately or publicly owned,
where refuse is dumped in either a controlled or un-
controlled manner. At an open dump, no refuse compaction
or earth cover is provided by bulldozers or tractors.
Refuse is left above ground to decompose indefinitely;
pollution control and public health programs are often
nonexistent or unenforced. In addition, open burning
of refuse, an on-site volume reduction processing
technique, is practiced at 70-80% of these sites [129],
The total number of authorized land disposal sites
in the U.S. is variously estimated at 12,000-25,000 [16,
25]. There may be ten times as many unauthorized or
187
-------
"promiscuous" sites [126], Of the authorized sites, only
6% meet the requirements of a sanitary landfill. About
79% of all land disposal sites are publicly operated
and approximately 63% are publicly owned [16],
Ocean dumping is a final disposal method which
is also used for a wide assortment of solid wastes.
However, ocean dumping of raw municipal refuse from
U.S. coastal cities is no longer practiced. Although
ocean disposal of municipal incinerator residues has been
discussed, the quantity of this municipal waste, if any,
which is currently being dumped in the sea is difficult
to determine from available information. It is import-
ant to note that a substantial proportion of solid wastes
bypasses some or all of these phases. This proportion
includes wastes which are transported to disposal sites
by individuals, industrial wastes disposed of in sep-
arate company owned facilities, and domestic refuse that
is burned in backyard incinerators. It also includes
litter which is carelessly discarded along highways
or in vacant lots.
As mentioned above, the techniques which are
employed for solid waste management have a radical
influence on the environmental impact of the management
system. There is considerable variation in the sequence
188
-------
of storage, collection, processing, and final disposal
techniques used in the U.S., both among regions and states
and even parts of the same state. For example, in the
1968 National Survey of Community Solid Waste Practices
[129], region II (New York, New Jersey, Pennsylvania)
reported a total of 226 incinerators, while region VIII
(Montana, Idaho, Wyoming, Colorado, and Utah) reported
no incinerators at all.
In region I (Connecticut, Rhode Island, Massachusetts,
Vermont, New Hampshire, Maine), routine uncontrolled
burning was reported at 6% of its land disposal sites,
while in region VI (North Dakota, South Dakota, Nebraska,
Kansas, Missouri, Iowa, Minnesota), routine uncontrolled
burning was reported at 65% of its land disposal sites
[129],
In region I, 48% of all land disposal sites were
classified as sanitary landfills, while in region VI,
only 28% of all land disposal sites were classified as
sanitary landfills [129].
Refuse Composition
The composition of municipal refuse is also an im-
portant factor in determining the environmental effects
of solid waste disposal. National figures indicate
that refuse composition has changed substantially over
189
-------
recent years and will continue to change. Table 4.2
[136] provides an estimate of the current composition on
an "as-discarded" basis for seasonal, semi-seasonal,
and non-seasonal states. Table 4.3 [136] shows projected
refuse compositions up to the year 2000. These projections
indicate several important trends.
The fraction of glass in refuse is not expected to
change significantly over the next 30 years. This pro-
jection may be high if low-cost food and beverage
plastic containers are developed, especially if these
containers are degradable.
The metal content of refuse may decline slightly
in the next 30 years. As with glass, changes in pack-
aging technology could also cause a much stronger decline
than anticipated.
The percentage of paper, cardboard, and other
wood-fiber components in refuse will continue to in-
crease by the year 2000. Paper products may account
for more than one-half of refuse by weight on an "as-
discarded" basis. Since the per capita rate of refuse
generatio^expressed in pounds/person/day, is steadily
rising while the paper content of refuse, which decreases
its bulk density, is increasing, the storage, collection,
and disposal problems will become increasingly severe.
190
-------
H
to
Table 4.2 ESTIMATED ANNUAL AVERAGE COMPOSITION OF 1968 MUNICIPAL
REFUSE ON "AS-DISCARDED" BASIS (by percent)
Category
Glass
Metal
Paper
Plastics
Leather , rubber
Textiles
Wood
Food wastes
Yard wastes
Miscellaneous
Un-Seasonal
State (e.g. ,
Florida)
7.6
7.5
32.6
1.0
1.3
1.8
2.3
18.2
26.1
1.6
Semi-Seasonal
State (e.g.,
Alabama)
8.1
8.1
35.1
1.1
1.4
1.9
2.4
19.5
20.7
1.7
Seasonal
State (e.g. ,
Massachusetts )
8.8
8.7,
38.2
1.1
1.5
2.0
2.7
21.1
14.1
1.8
Source: [136,p.13]
-------
Table 4.3 PROJECTED REFUSE COMPOSITION
(by percent)*
Refuse
Category
Glass
Metal
H Paper
10 Plastics
Leather , rubber
Textiles
Wood
Food wastes
Miscellaneous
Yard Wastes
Total
Sea-
son-
al
8.8
8.7
38.2
1.1
1.5
2.0
2.7
21.1
1.8
14.1
100.0
" 1968
Serai-
Sea-
son-
al
8.1
8.1
35.1
1.1
1.4
1*9
2.4
19.5
1.7
20.7
100.0
Non
Sea-
son-
al
7.6
7.5
32.6
1.0
1.3
1.8
2.3
18.2
1.6
26.1
100.0
Sea-
son-
al
9.
8.
39.
1.
1.
2.
2.
20.
1.
13.
100.
1
8
1
3
5
0
5
0
7
8
0
1970
Semi-
Sea-
son
al
8.4
8.2
35.8
1.3
1.4
1.9
2.3
18.7
1.6
20.4
100.0
Non
Sea-
son-
al
7.9
7.6
33.5
1.1
1.3
1.8
2.2
17.4
1.5
25.7
100.0
Sea-
son-
al
9.9
9.0
40.8
1.9
1.5
2.1
2.2
17.9
1.5
13.2
100.0
1975
Semi-
Sea-
son-
al
9.2
8.4
37.6
1.8
1.4
2.0
2.0
16.6
1.4
19.6
100.0
Non
Sea-
son-
al
8.
7.
35.
1.
1.
1.
1.
15.
1.
24.
100.
7
8
2
7
3
9
9
5
3
7
0
*Percentages shown are on an "as-discarded" basis.
-------
Table 4.3(continued). PROJECTED REFUSE COMPOSITION
(by percent)*
1980
Refuse
Category
Glass
Metal
Paper
Plastics
Leather / rubber
Textiles
Wood
Pood wastes
Miscellaneous
Yard wastes
Total
Sea-
son-
al
10.3
9.4
41.5
2.8
1.5
2.1
2.0
16.2
1.4
12.9
100.0
Semi-
Sea-
son-
al
9.
8.
38.
2.
1.
2.
1.
15.
1.
19.
100.
6
7
4
7
4
0
8
0
3
2
0
Non
Sea-
son-
al
9.0
8.1
36U
2.5
1.3
1.9
1.7
14.1
1.2
24.1
100.0
Sea-
son-
al
9.5
9.0
45.0
3.5
1.5
2.5
1.6
14.0
1.2
12.2
100.0
1990
Semi-
Sea-
son-
al
8.
8.
41.
3.
1.
2.
1.
13.
1.
18.
100.
9
4
7
5
4
3
5
1
1
1
0
Non
Sea-
son-
al
8.4
7.9
39.3
3.1
1.3
2.2
1.4
12.3
1.1
23.0
100.0
Sea-
son-
al
8.1
7.4
49.7
4.2
1.6
2.8
1.3
12.1
1.0
11.8
100.0
2000
Semi- Non
Sea- Sea-
son- son-
al al
7.6
6.9
46.0
4.2
1.5
2.6
1.2
11.4
1.0
17.6
100.0
7.2
6.5
43.5
3.8
1.4
2.5
1.2
10.7
0.9
22.3
100,0
*Percentages shown are on an "as-discarded" basis.
Source: [136,p.13]
-------
Plastic concentrations in refuse are expected to
increase over 400%. The major categories of thermo-
plastics which are expected to dominate this growth are
polyolefins (primarily polyethylene), polyvinyl chloride,
and polystyrene [115]. Due to their high molecular weight
and hydrocarbon structure, combustion of these polymers
in conventional incinerators is unsatisfactory when
their concentration in refuse exceeds the current 1-2%
level [83]. Such combustion results in clogged or melted
firing grates, excessive heat generation, and evolution
of noxious and toxic, incomplete and complete combustion
products, such as HC1 in the case of polyvinyl chloride
[83]. The morphology (degree of crystallinity) and
chemical structure of these polymers also renders them
almost totally immune to biodegradation. The non-bio-
degradability of plastic wastes, and their relatively
bulky nature and poor compactability, combine to make
sanitary landfilling of these wastes less attractive as
a final disposal technique [115]. Further more, the
fire hazards of exposed land disposal of plastic wastes
precludes open dumping as a safe method of final disposal.
In view of these special disposal problems, such a
dramatic increase in the plastic content of refuse may
have a serious impact on the externalities of the manage-
194
-------
ment system.
Resource Recovery
As the total volume of solid wastes increases and
its composition changes, traditional waste management
techniques and facilities appear increasingly inadequate,
especially in urban areas. An attractive way of "disposing"
of solid waste materials is to recover its components
for reuse or to find alternative uses for the waste mater-
ials themselves. Recovery of materials from wastes is
being practiced in the U.S. on a substantial, but
decreasing, scale, Table 4.4 presents the quantity of
recovered materials in relation to total comsumption of
those materials for nine selected categories in 1967
[149]. These figures refer both to materials recovered
during production and materials recovered from wastes.
The great bulk of the materials which are recovered from
wastes are taken from highly concentrated material sources
such as steel from scrap auto bodies and appliances,
or newprint from old newspapers. In general, the portion
of input materias derived from waste sources has been
declining. For example, the recovery rate (percentage
of a material's total supply derived from wastes) for
paper fiber has declined from 23.1% in 1960 to 17.8% in
1969. Similarly, the recovery rate for rubber has dropped
195
-------
from 18% in 1968 to 9.5% in 1970; and the rate for iron
and steel has fallen from 44.9% in 1949-53 period to
40.0% in the 1964-68 period [149], The gradual decline
in resource recovery reflects primarily the economically
non-viability of recycling with present technology.
Although resource recovery from specific wastes
is being practiced, economically feasible technology for the
recovery of metal, glass, and wood-fiber components from
municipal refuse is currently unavailable [48], The
biggest technical obstacles to recovery of refuse com-
ponents are problems of classification and separation of
components and of purification of the separate materials
to acceptable levels. Naturally, the composition of
refuse affects the possibilities of effective recovery.
Projections which indicate increased paper content seem
favorable to eventual development of fiber recovery
processes, while decreasing metal content will certainly
make metal recovery more difficult [136,48,74],
C. COLLECTION AND STORAGE OF MUNICIPAL REFUSE
The first stages in the management of municipal
refuse are storage and collection. In the U.S. a
variety of storage containers and collection systems
are employed. Although comprehensive data on national
collection practices is lacking, it appears that the
196
-------
Table 4.4 RECYCLING OF MAJOR MATERIALS: 1967.
Material
Paper
Iron and Steel
Aluminum
Copper
Lead
Zinc
Glass
Textiles
Rubber
Total
Total Consumption
(million tons)
53.110
105.900
4.009
2.913
1.261
1.592
12.820
5.672
3.943
191.298
Total Recycled
(million tons)
10.124
33.100
.733
1.447
.625
.201
.600
.246
1.032
48.108
Recycling as Percent
of Consumption
19.0
31.2
18.3
49.7
49.6
12.6
4.2
4.3
26.2
25.1
Source: [149]
-------
most common storage and collection system is the
familiar metal garbage can and the compactor truck
with a muti-man crew [135,44].
Refuse storage and collection create both direct and
indirect adverse effects. The direct adverse effects
are the unsightliness and odor of storage containers
and their contents and the noise, road congestion, and
air pollution from collection activities. The prop-
agation of various species of flies and rats which act
as disease vectors and nuisances are indirect adverse
effects.
Very little information or data is available on the
extent of the direct adverse effects of storage and
collection on a national basis. It is evident, however,
that innovations in specific systems can substantially
alter the extent and composition of these externalities.
Recent innovations in collection systems have been
largely confined to increasing the capacity, compaction
power and loading efficiency of collection trucks. The
major impact of these improvements has been to lower
collection time and the number of trips necessary on a
particular route [145]. In turn, these have probably
decreased externalities due to air pollution and con-
gestion while they may have increased noise.
198
-------
Another interesting innovation in collection
systems is the development of satellite vehicle coll-
ection. These are small, 3 or 4 wheeled vehicles that
shuttle between domestic refuse storage points and a
packer truck which subsequently takes the refuse to a
disposal site. These satellite vehicle systems are
being used in 50 U.S. communities in areas where pick-up
points are not at a curb or in an alley. One study
[49] which evaluated these systems noted that packer
trucks remained on main roads and that their mileage
was reduced. This indicates decreased air pollution
and road congestion. On the other hand, the study iden-
tified several potential hazards of satellite vehicle
operation, including increased littering of collection
areas, damage to dwelling unit property and excessive
noise.
In the area of refuse storage, innovation in con-
tainer design and operation can also reduce external
effects. Plastic cans, plastic bags, and paper bags
can reduce collection noise, speed collection operations
and make routes more sightly. Home trash compactors
reduce refuse volume and seal compacted refuse in odor
tight bags.
The indirect adverse effects of storage and collection
199
-------
practices are the propagation and harborage of flies
and rats in and around storage containers. The
association of fly and rat populations with refuse
storage and collection practices is well documented
[13,14,24,37 refs (1-7,9-16,18),61,62,85 refs (150,
152,153,391), 114,119,141,164],
Coomensual species of rats and flies are known
vectors of a variety of agents pathegenic to humans.
(For a thorough discussion of the vector activities of
these species see "Vectors and Vermin" in Section F,
below). Besides imposing external costs through disease
transmission, these species, especially the house fly,
Musca domestica, are a ubiquitous nuisance in domestic
environments, particularly during the summer months, the
peak fly breeding season.
The process of fly propagation in refuse containers
is straightforward. Female flies enter the container and
lay their eggs on a suitable organic medium. The larvae
then feed on this material for several days and then
crawl out of the container to burrow into the adjacent
dry soil to pupate. After pupation is completed, the
adult females emerge from the soil. A number of factors
are important in regulating this process: the quantity of
putrescible materials in the container, the ease of access
200
-------
into the container for egg-laying adults and the ease
of departure of larvae, collection frequencies, and
temperature.
Research on the relationship between refuse storage
and collection and fly propagation has shown that:
1. If garbage grinders are used to reduce the
garbage content of domestic refuse or if garbage is
drained and securely wrapped in at least three thickness*
of newspaper, that flies have very few suitable breeding
media and that fly production is thus greatly reduced
[37,114].
2. Containers which afford free movement of egg-
laying adults and migrating larvae allow much larger
fly populations to develop. An increase of four times
or more can be expected [14,24,37,61,62,114].
3. Twice a week collection of refuse result in much
lower fly migration and production at the refuse source
114,24,37,61,62].
4. During periods of high temperature, fly life
cycles are considerably accelerated, resulting in
higher fly production from all storage and collection
systems [37,114],
Some quantitative idea of the impact of various
collection and storage systems can be gained from Tables
201
-------
4.5 and 4.6 [61,62].
Although many publications refer to the association of
rats with the storage and collection of refuse, little
qualitative work has been completed [See 119 and 13 for
references]. Although it appears that refuse—particulary
mis-managed, infrequently collected and inadequately
stored refuse—is a major source of food for commensual
rat species, it is unclear that it is a vital source.
In other words, better management of refuse storage and
collection is certainly necessary for elimination of
dangerous rat population but may not be a sufficient
condition. Even the elimination of human refuse and
garbage would not necessarily have a significant effect
on the rat population except possibly in inner-city areas.
D. INCINERATION
Refuse incineration is a volume reduction, waste
processing technique that is widely practiced in many
areas of the U.S. Incineration is the controlled
combustion of waste materials with the oxygen in air
to convert refuse into gaseous products and solid residues.
Incinerators are generally designed to allow rapid and
complete refuse combustion and efficient handling of all
residues and combustion products.
The combustion within a refuse incinerator is a
202
-------
Table 4.5 PERCENTAGES OF WEEKLY COLLECTIONS THAT CONTAINED 50
OR MORE FLY LARVAE DURING EIGHT WEEKS OF STUDY
10
0
CJ
Type of
System 8/2
once -a- week
65
metal cans
once-a-week
25
paper bags
twice-a-week
25
metal cans
twice-a-week
12
paper bags
Week Ending:
Total per-
8/9 8/16 8/23 8/30 9/6 9/13 9/20 iod of
study
60 65 50 35 50 30 42 50
20 22 15 22 25 2 17 19
12 15 12 7 15 0 2 11
5 12 5 2 2 2 2 5
Source: [61J
-------
o
*«.
a
Table 4.6 NUMBER OP GARBAGE CONTAINERS FROM WHICH
PtY IARVAE MIGRATING EXCEEDED 50
Type Unit
b
Twice-a-week
a
Once-a~week
a
Exposed bag
a
Enclosed bag
1
3
14
10
7
2
5
16
3
6
Week
3
9
20
5
16
of
4
3
19
9
5
Study
5
6
18
7
10
6
3
19
8
7
7
7
17
9
8
8
1
9
3
3
9
3
11
3
5
10
2
9
1
4
Garbage collected once-a-week
b
Garbage collected twice-a-week
Source: [62]
-------
high temperature process. Heat released by combustion
is partly stored in the combustion products and partly
transferred to the incinerator walls and incoming fuel.
If prompt fuel ignition and combustion occur with the
correct proportion of fuel and air, combustion is rel-
atively complete [45],
The major products of complete combustion of organic
fuels are carbon dioxide, water, sulfur dioxide, and
nitrogen. Some nitrogen is converted to its oxides,
particularly nitric oxide, and some of the sulfur di-
oxide is converted to sulfur trioxide. A small portion
of the metals present, such as iron and aluminum, may also
be converted to their oxides [45], Table 4.7 presents
a composite analysis of average municipal refuse 1176
from 96], Considering the very low sulfur content of
refuse and the low conversion rate of nitrogen to ni-
trogen oxides, the gaseous products of complete combustion
seem to pose minor air pollution problems, if any.
Furthermore, residues from these ideal combustion con-
ditions would be completely sterile and totally free of
combustibles, thus presenting minimal final disposal
problems. However, the combustion conditions in actual
refuse incinerators may be far from ideal.
205
-------
Table 4.7 COMPOSITE ANALYSIS OF AVERAGE MUNICIPAL
REFUSE, AS-RECEIVED BASIS
o
cr»
Moisture . .
Carbon . .
Hydrogen .
Oxygen .
Nitrogen . . .
Sulfur ...
Noncombustibleb. .
Total
Percent
. 20.73
. . 28.00
3.50(0.71)a
. 22.35
0.33
0.16
. 24.93
100.00
Theoretical Combustion
air, Ib/lb refuse
xll.53 *» 3.2284
X34.34 • 0.2438
x 4.29 m 0.0069
3.4791
Caloric value, Btu/lbj 4,917 as fired; 6,203 dry basis; 9,048 dry-
ash-free basis
aThe net hydrogen available for combustion (0.71 percent) equals
the total hydrogen (3.50)
bNoncumbustibles: Ash, glass, ceramics, metals.
Source: [176 from 96]
-------
The nature of refuse as a fuel is the primary
reason for poor combustion conditions during inciner-
ation. Because of the physical and chemical variations
in refuse, combustion proceeds unevenly in the fuel bed.
Combustion air flows through the more permeable portions
of the bed, while other portions are starved for air.
Moisture content variations in refuse components exac-
erbate this problem. Steam flowing from high moisture
content regions also tend to disturb the air flow. In
addition, variations in heating values of the refuse
can cause large temperature differences and difficulities
in temperature control. These and other factors result
in incomplete refuse combustion.
Incomplete combustion of refuse has substantial
adverse effects: a greatly increased volume of noxious
effluents, especially particulates, and undesirable
residue*composition. The emissions from refuse incin-
erators have been measured in field situations to deter-
mine the components of incinerator effluents and to in-
vestigate the effect of operating and design character-
istics on emissions [21,69,73,103,137,138,146,148,162,
175,180,181,182,194,196,201]. See also the literature
reviews by Wikstrom and Carotti [172] and the Air
Pollution Control Office [7],
207
-------
Particulate matter in incinerator effluents con-
sist of smoke, soot, flyash, grit, dirt, carbonaceous
flakes, aldehydes, organic acids, esters, fats, fatty
materials, phenols, hydrocarbons, polynuclear hydro-
carbons, and other organics. Particle size ranges from
less than 5 microns to 200 microns and larger. Table
4.8 shows the size distribution of particles from
large European refuse incinerators [176 taken from 27],
Table 4.8 THE SIZE DISTRIBUTION OF
PARTICULATES FROM EUROPEAN INCINERATORS
Particle diameter,
microns
< 5 ....
< 7 ....
<27
<39 ....
<59 ....
>59 ....
Percent
• . • • • J.J. • .L
..... * J. . /
. . 30.1
40.2
. . • . .46. 9
..... D j . .L
..... 67. 5
90.5
* . . . . "• D
208
-------
Actual particulate emissions at the furnace outlet
from incinerators of various types and designs are
presented in Table 4.9 [176 taken from 133]. Published
data on the chemical composition of particulates is
scarce. Table 4.10 gives the results of a chemical
analysis of flyash samples from a New York incinerator
[176 taken from 133].
Gaseous incinerator emissions, unlike particulates,
are of minor importance as a source of air pollution at
present. Major toxic components include: oxides of
nitrogen; oxides of sulfur; aldehydes; hydrocarbons;
and ammonia. Concentrations of these effluents are
primarily affected by excess air and underfire air
during combustion. Excess air is the air supplied to
burn the refuse in addition to that theoretically
(stoichiometrically) necessary for complete combustion.
Excess air is usually expressed as a percentage of the
amount of air theoretically necessary for complete
combustion, as "130 percent excess air". The furnace
temperature of a refuse incinerator is usually con-
trolled by adding excess air; the more excess air the
lower the furnace temperature [176]. Underfire air is
any air, controlled with respect to quantity and direction,
that is supplied beneath the firing grate and that passes
209
-------
Table 4.9 PARTICULATE EMISSIONS PROM MUNICIPAL INCINERATORS AT
FURNACE OUTLET (corrected to 50 percent excess air)
Furnace type
50-ton-per-day batch
50-ton-per-day batch
50-ton-per-day batch
250-ton-per-day continuous
250-ton-per-day continuous
250-ton-per-day continuous
250-ton-per-day continuous
(traveling grate)
250-ton-per-day continuous
(reciprocating grate)
120-ton-per-day continuous
(rocking grate)
150-ton-per-day continuous
(rocking grate)
Excess air
percent
235
110
100
190
180
150
(6.0% CO )
(5.0% C02)
(7.0% C02)
—
Underfire air,
percent
20
50
70
20
50
100
41.8 scfm/sq ft
grate area
105 scfm/sq ft
grate area
17.5 scfm/sq ft
grate area
mm
Average
dust loading
Ib/ton
of charge
0.78
1.04
1.79
3.8
2.8
4.6
12.4
25.1
9.1
30.8
to
H
o
Source: [176 taken from 133, p. 83]
-------
Table 4.10 CHEMICAL ANALYSIS OP INCINERATOR
PLYASH SAMPLES (percent by weight)
Component
Organic
Inorganic
Silica as SiO-
Iron as Fe2^3
Alumina as A^O^
Calcium as CaO
Magnesium as MgO
Sulfur as SO-
Sodium and potassium oxides
Source of Sample
Upper flue
0.5
99.5
50.1
5.3
22.5
7.9
1.8
4.3
8.1
Expansion Chamber
0.6
99.4
54.6
6.0
20.4
7.8
1.9
2.3
7.0
Emitted
10.4
89.6
36.1
4.2
22.4
8.6
2.1
7.6
19.0
to
Source: [176, taken from 133, p. 84]
-------
through the fuel bed [176].
Formaldehyde is only produced in minute amounts in
municipal incinerators. For a 250-ton-per-day contin-
uous feed incinerator operating at 185% excess air,
.0014 pound of formaldehyde per ton of refuse was
generated; for a 50-ton-per-day batch feed incinerator
operating at 108% excess air, no formaldehyde was pro-
duced [84,179], Hydrocarbon emissions are usually below
the detectable level of measuring instruments. A 250-
ton-per-day continuous feed incinerator using 150-190%
excess air formed less than .08 pound hydrocarbon per
pound of dry flue gas [179],
Because of the relative importance of particulate
emissions from an air pollution standpoint, air poll-
ution control equipment for incinerators has been
designed primarily to remove particulate matter. A large
variety of control devices have been developed toward
this end. They include: settling chambers, in which
large particles of flyash settle out of the flue gases
immediately after leaving the combustion chamber;
baffled collectors, in which large particles of flyash
are removed changing the direction and/or reducing the
velocity of the flue gases; scrubbers, in which flyash
particles are removed by sprays or wet baffles? cyclone
212
-------
collectors which set up an intense vortex causing
particles to be removed centrifugally; fabric filter
collectors, in which particles are removed by filtering
through long, tubular fabric bags; and electrostatic
precipitators which collect particles by placing a charge
on them by means of gaseous ions or electrons and then
removing the particles at collecting electrodes of the
opposite charge. These devices vary considerably in
collection efficiency. Under one particular set of test
conditions, the fabric filter collected from 97 to 99.9%
of the dust measured by weight, the electrostatic pre-
cipitation collected from 90 to 97%, the scrubber collected
from 80 to 96%, the mechanical collector from 30 to 80%,
and the settling chamber up to 30%.
Nearly all municipal incinerators, constructed since
1945 have some kind of pollution control equipment.
A vast technical literature exists on the design and
operating techniques of refuse incinerators and aux-
iliary air pollution control equipment. No attempt will
be made here to survey these individual studies. The
reader is referred to Corey [45] and Stear [176], for
excellent treatments of these topics and further ref-
erences. Stack emissions of particulates from incin-
erators with various control devices are listed in
213
-------
Table 4.11 [176 taken from 92,179,192] and Table 4.12
[adapted from 1].
On a nationwide basis, municipal incineration is
certainly a minor source of particulate emissions. If
we assume that in 1968 approximately 200 million tons of
municipal refuse were generated and that approximately
10% of this total was processed in municipal incinerators
with an emission factor of 10 Ibs. particulates per ton
of refuse charged, then approximately 100,000 tons of
particulates would have been emitted. For the same year,
total particulate emissions in the U.S. were reported
at 28.3 million tons per year [52], This means mun-
icipal incineration contributed about .35% of total
particulate emissions. Of course, its relative signif-
icance for a particular city might be much higher.
The same type of calculation can be applied to the
other major categories of air pollutants sulfur dioxide,
nitrogen oxides, and hydrocarbons. Assuming again that
20 million tons of refuse were incinerated and that
SO2» NOX, and hydrocarbon emission factors are 2 Ibs.
per ton charged, 2 Ibs/ton charged, and .3 Ibs/ton
charged respectively [60], total municipal incinerator
emissions would have been 20,000 tons SO2 per year,
20,000 tons NOX per year, and 3000 tons hydrocarbons
214
-------
Table 4.11 PARTICULATE LOADING OF INCINERATOR STACK GASES
Furnace type
50-ton-per-day batch
50-ton-per-day batch
50-ton-per-day batch
250-ton-per-day continuous
(reciprocating grate)
120-ton-per-day continuous
(rocking grate)
1 50-ton-per-day continuous
(rocking grate)
Control
Equipment
Scrubber
Scrubber
Scrubber
Settling chamber
and wet baffle
Settling chamber
and wet baffle
Wet baffle
Excess
air,
percent
235
110
100
-
-
••
Underfire
air,
percent
20
50
70
—
-
"*
Average
dust loading,
Ib/ton charge
in stack gases
0.57
0.55
0.61
11.8
8.2
8.24
to
H
Ul
Source: [176 taken from 133, p. 84]
-------
Table 4.12 PARTICULATE LOADING OF INCINERATOR STACK GASES
Furnace type
300 TPD-continuous
(reciprocating grate)
300 TPD-continuous
(rocking grate)
5-6 TPD-batch
500 TPD-continuous
(traveling grate)
500 TPD-continuous
(reciprocating grate)
600 TPD-continuous
(reciprocating grate)
400 TPD-continuous
(reciprocating grate)
Control
equipment
scrubber
scrubber
scrubber and
electrostatic
precipitator
scrubber
scrubber
scrubber
multi -cyclone
collectors and
wet baffles
Excess
air,
percent
270
-
370
260
220
320
500
Average
dust loading,
Ib/ton charge
in stack gases
10.4
14.5
4.1
8.8
8.6
12.5
20.4
e\
Source: [taken from 1]
-------
per year. Nationwide emissions of SO,, NO, and
2 X
hydrocarbons for 1968 have been reported at 33.2 million
tons, 20.6 million tons, and 32.0 million tons, respect-
ively. Thus municipal incineration accounted for .06%
of all SO2 emissions, .1% of all NO emissions, and
.009% of all hydrocarbon emissions.
The adverse effects of air pollution are well known.
Air pollutants cause deleterious physiological effects
in humans and animals, structural damage and soiling
of affected vegetation, buildings and clothing, and
visibility reduction. However, considering the minor
role of municipal incinerators as air pollution sources,
the incinerator externalities due to air pollution appear
to be relatively small, unless marginal damage rates are
considerably higher than average damage rates.
Further potential externalities due to refuse
incineration are water pollution caused by discharge of
process waters and leachate problems from land disposal
of residue materials. Leaching problems associated with
residue disposal will be treated in the following section.
Nearly all municipal incinerators utilize water for
expansion chambers, cooling charging chutes, ash sluicing,
quenching, conveying residues, and air pollution control
[50]. The quantity of water required varies over a
217
-------
wide range according to incinerator design, but most
plants use 1000-2000 gal. per ton of solid waste
processed. However, if water is treated and recircu-
lated, water consumption can often be reduced by 50 to
80%. Studies of incinerator wastewater show that it
contains suspended solids, inorganic materials in
solution, and occasional small concentrations of
organic materials. Incinerator wastewaters are suffic-
iently polluted to pose a potential water pollution
hazard if discharged indiscriminately to surface waters
without prior treatment [116]. Suitable treatment
methods for incinerator wastewaters include primary
clarification, pH adjustment, and biological treatment
[50]. Of course a convenient alternative, in some cases,
to on-site treatment is discharge of wastewaters to a
municipal sewage system. The 1968 National Survey of
Community Solid Waste Practices [129] reports that
30.95% of incinerators surveyed needed wastewater
effluent treatment and 27.43% provided such treatment.
E. SANITARY LANDFILLS
Sanitary landfilling, although perhaps the most
apparently innocuous disposal method for solid wastes,
may create a number of different externalities: noise,
odor, gas generation, and ground and surface water
218
-------
pollution.
Noise
An almost totally neglected externality of sanitary
landfilling is noise. Noise at sanitary landfill sites
is generated by the unloading of refuse, including the
operation of compactor and transport trucks as they
disgorge their cargoes and the constant operation of the
heavy equipment used for spreading, compacting, and
covering the refuse. In order to achieve efficient
operation and optimal compaction at medium to large
capacity landfills, the heaviest and most powerful
tractors, crawler tractors, steel wheel compactors,
and draglines are employed. When all the heavy equip-
ment at a medium to large sanitary landfill is operating,
noise generation will be in the 80-120 dBa range [19,40].
Noise has both auditory and non-auditory adverse
effects on exposed populations. Noise can interfere
with speech communication [11,22,40]. Noise can cause a
partial or permanent loss of auditory sensitivity [11,
22,40,41]. The non-auditory effects of noise exposure
include: neural-hormonal stress response [11,22,79];
sleep interference [11,22]; task interference [11,22,79];
and mental stress and annoyance [11,22,40,79].
Noise from ground sources, such as landfill equip-
219
-------
ment, is attenuated by many factors before it reaches
receivers. Foremost among these is wave divergence.
The sound-pressure level generated by a sound source
at a receiver decreases as a function of the distance
from the source due to wave divergence. For hemispher-
ical wave divergence over loss-free ground free of
barriers, through a homogenous loss-free atmosphere,
the sound-pressure level drops off 6 dB for each
doubling of distance [105].
Other sources of attenuation due to environment
conditions are:
1) attenuation due to ambient temperature and
barometric pressure;
2) attenuation by absorption in the air;
3) attenuation by rain, sleet, snow or fog;
4) attenuation by barriers;
5) attenuation by grass, shrubbery, and trees;
€) attenuation due to wind and temperature grad-
ients, atmospheric turbulence, and to the
characteristics of the ground [105].
Any or all of these factors may affect the level of noise
experienced by those near landfills.
Odor
As with noise and visual pollution, odor from land
220
-------
disposal of solid wastes has physical characteristics
which are difficult to measure objectively. The two
major obstacles to accurate measurement are specifica-
tion of observer reaction to odors, and determination
of odor emission factors and odor dispersion patterns.
The problem of odor perception is quite complex. Odor
is defined by the American Heritage Dictionary as "any
sensation, stimulation, or perception of the sense of
smell". The perception of an odor involves the olfact-
ory detection of specific types or combinations of
molecules present in the air. The exact mechanism of
odor perception is a matter of considerable controversy.
For further information concerning the mechanism of
odor perception, the reader is referred to Amoore [8,9],
Dravnieks [54], Friedman and Miller [75], McCartney
[117], Russell and Hills [163], Theimer and McDaniel
[190], and Wright [199], One clear finding is that there
is a pronounced variation in individual perception of
odor thresholds and qualities. The authors of one paper
[3 ref (100)] determined experimentally that materials
in concentrations of 2 parts per billion or less have
frequently yielded detectable odors. They state: "There
is no such thing as a threshold odor for a community.
Measuring the threshold odor for one man gives a number.
221
-------
Measuring it for 5 men gives a range. If 200,000
people are used, a very wide range is obtained." From
experimental evidence they reported that the most sen-
sitive 5% of the population was at least 17 times as
sensitive as the least sensitive 5%, and that the most
sensitive observer in the 20-man panel in their exper-
iment was on the average 11 times as sensitive as the
least sensitive man.
A study of community odor surveillance by an
observer corps of 120 high school students [89] con-
cluded that: "The problem of precisely defining an
offensive odor is difficult. An odor may be pleasant
to one and offensive to another because of psychological
associations with the odor. An objectionable odor can
not be defined as one that is offensive to all individ-
uals exposed to it since there will always be a few
who will be blissfully unaware of even the most fetid
stench." The authors go on to point out that psycho-
logical responses to a given odor vary among observers.
Intensity is only one factor of odor perception.
A report on odor nuisance [6] states that " a strong
chemical odor offends because of its intensity. Obnox-
ious or malodorous odors offend because of their
'quality1. The latter causes more complaints and tend
222
-------
to be odors originating from the handling and processing
of organic compounds containing sulfur and nitrogen."
The authors warn that odor emissions can cause "great
personal discomfort, nausea, loss of appetite, loss of
sleep, skin rash, psychic disorders, and eye, ear, nose,
and throat irritation."
In addition to variation in odor perception, there
are many environmental conditions which influence odor
evaluation, dispersion, and persistence. Particularly
important factors are wind-speed and direction, vertical
temperature gradients, temperature, humidity, and top-
ography [89]. Naturally, the condition of the odor source
itself, namely the putrescible and odorous portions of the
refuse, will have a major influence on odor emissions
and quality. The situation and exposed surface area
of the odor source are primary determinants of odor emissions
[87].
To obtain even an approximate measure of the quan-
titative and qualitative aspects of the odor nuisance
caused by a surface source such as a landfill requires
conbining analytic techniques related to odor evolution,
transmission, and perception. This has proved to be a
difficult undertaking. Proposals of methods to evaluate
community odor problems have focused on survey techniques
223
-------
[6,87r89]. Generally, such odor surveys employ a panel
of odor evaluators or a corps of field observers to
appraise odor samples in the laboratory or to make
field observations in the community. Olfactory sen-
sations are then classified according to simple verbal
or numerical scales. These surveys provide valuable
information on odor incidence but can not provide a
satisfactory measure of the overall community impact of
odor sources such as landfills. The importance of expo-
sure time, odor fatigue, seasonal variation, and psycho-
logical response make long-term generalizations of
survey results questionable.
Decay and putrefication of the putrescible por-
tions of refuse and the exposure of odorous wastes
releases objectionable air pollutants [6]. The extent
to which this is true for municipal refuse placed in
sanitary landfills is largely undocumented. Prompt
covering of putrescible and odorous wastes is said to
prevent the release of objectionable odors to the atmo-
sphere [6], However, a preliminary analysis of the land
disposal investigation site data for regions 1 and 2
of the 1968 National Survey of Community Solid Waste
Practices [130,132] indicates that approximately 10%
of those sites with daily compacted earth covers
224
-------
required odor control programs compared to 24% of
those sites which did not receive daily compacted earth
covers, although the criterion of when a program is
"required" is not clear. These statistics show that
although odor problems are diminished by sanitary land-
filling, they are not eliminated.
Gas Production
Another major type of external effect associated
with sanitary landfills is the production of gases in
the landfill. Many publications have dealt with gas
production in sanitary landfills. See [30,76,123,124,
125,186,147] for comprehensive analyses of gas production
from laboratory lysimeters and test cells; [12,26,58,59,
90,122] for descriptions of gas production problems at
landfill sites; and [65,66,67,70,91] for particularly
thorough theoretical treatments of gas generation and
diffusion as well as measurements of gas compositions
within landfills. Gas production by landfills can impose
externalities in two ways:
1) causing fires or explosions in areas adjacent
to or above the landfill due to a methane build-
up ; and
2) the pollution of groundwater due to solution
of
225
-------
Groundwater pollution problems will be discussed in the
following section. Before examining the problem of
methane buildup a brief review of the mechanism of gas
evolution and dispersion in a sanitary landfill is useful,
Gas production in a sanitary landfill is the result
of the microbial decomposition of organic fill materials.
The compostion of gases produced in the fill depends
upon the composition of the organic materials and the
quantity of oxygen present. If sufficient oxygen is
present decomposition will generally be aerobic; if not,
decomposition will generally be anaerobic? both may take
place within a landfill simultaneously. Since oxygen
is present immediately after sanitary landfills are
completed all decomposition is initially aerobic.
The agents of aerobic decomposition are microorgan-
isms known as aerobes. Aerobes are endemic to many
aerobic environments such as soils, and comprise meso-
philic and thermophilic bacteria, fungi, actinomycetes,
and protozoa of many species [186], In humid climates
fungi are usually the primary agents while bacteria are
more important in semi-arid locations. The addition of
organic matter and oxygen to an aerobe environment
increases the number of microorganisms dramatically and
decomposition begins. Aerobic decomposition involves
226
-------
the progressive oxidation of organic matter by aerobes
to produce CC>2 and HO and the oxidation of nitrogen
and sulfur to nitrates and sulfates. Nitrogenous com-
pounds are first oxidized to ammonia, then to nitrites,
and finally to nitrates. Some production of nitrogen
gas may occur. Sulfurous compounds are similarly
oxidized to sulfates. Carbonaceous materials are com-
pletely broken down to C02 and H-0 [32],
Figure 4.1 depicts the carbon, nitrogen, and sulfur
cycles for aerobic decomposition of organic materials
[32],
Organic materials differ widely in their suscept-
ibility to decomposition; sugars and starches are easily
digested while cellulose and other complex organic
compounds decompose very slowly. Furthermore, aerobes
do not break down complex molecules directly; extra-
cellular enzymes are excreted which convert complex
molecules into simpler, intermediate products. These
are finally reduced to C02, H20, N , nitrates, and sul-
fates [66], When the oxygen supply becomes depleted
from the demands of aerobic organisms and is not renewed
by atmospheric infusion or dissolved O2, aerobic
bacteria cease to function and anaerobic processes take
over. Anaerobic chemical reactions consist of the
227
-------
Figure 4.1 AEROBIC DECOMPOSITION -
NITROGEN, CARBON, & SULFUR CYCLES
OEAO
ORGANIC
MATTER
INITIAL
PRODUCT OF
DECOMPOSITION
carbonaceous
nitrogenous
sulfurous
LIVING
ANIMAL
MATTER
hydrogen sulfid*
ammonia
CO,
AIR
AND WATER
RESERVOIR
BIOLOGICAL
OXIDATION
(N.CU.CO, )
TERMEDIAT
PRODUCT OF
DECOMPOSITION
Nitrite (N
sulfur
CO,
LIVING
PLANT
MATTER
carbohydrates
proteins
fats
PRODUCT OF
DECOMPOSITION
nitrate (N)
sulfata
CO
Source: [32, p. 20]
228
-------
initial decomposition of organic compounds to organic
acids, acid carbonates, COj and 1^0. The organic
compounds are further broken down to ammonia, nitrogen
humus, carbon dioxide, methane, and sulfides. Figure
4.2 depicts the carbon, nitrogen, and sulfur cycles for
anaerobic decomposition of organic matter.
A comparison of aerobic and anaerobic respiration
shows that relatively more carbon dioxide is produced
under aerobic conditions while methane, CH., is a major
product of anaerobic conditions [32], For excellent
summaries of refuse decomposition the reader is referred
to Hughes, Landon and Farvolden [91] and Engineering
Science [65].
The volume and composition of gases produced in
actual landfill situations has been a topic of consid-
erable study. Over 90% of the gas produced by refuse
decomposition in large landfills is CO2 and CH [67],
A typical pattern of gas generation usually develops
for a newly completed landfill. During the initial
aerobic stage, CO2 production is very rapid. Peak
concentrations of C02 are normally 75-98%, occurring one
to seven weeks after the fill is completed [67,122,125],
As molecular oxygen supplies are exhausted, aerobic
processes taper off, and CO2 concentrations decline;
229
-------
Figure 4.2 ANAEROBIC DECOMPOSITION -
NITROGEN, CARBON, & SULFUR CYCLES
DEAD
ORGANIC
MATTER
INITIAL
PRODUCT OF
DECOMPOSITION
(.carbonaceous
2. nitrogenous
j.sulfurous
i«2.organic ccidi,C02
acid carbonates. 3.
hydrogen sulfide
AIR
AND WATER
RESERVOIR
-------
methane concentrations gradually begin to rise.
Figure 4.3 [125] shows the gas concentration over
time in a test landfill cell; Figure 4.4 [67] shows
the gas concentration over time in a test landfill.
From Figures 4.3 and 4.4 both the typical CO "bloom"
and the gradual CH. buildup can be noted.
Gas movements through refuse and adjacent soil
depends on gas pressure and density (concentration)
gradients. The laminar movement of gases under pressure
through a porous medium is governed by Darcy's law.
Since the transfer area through which gases can flow
from landfills into surrounding soils is relatively
large and soils are usually quite permeable to gases,
no large pressure differences are likely to develop.
It is therefore permissable to treat gases as incom-
pressible fluids when calculating flow rates. The
proper form of Darcy's law is then:
v = - 51 (grad p)
where kg equals hydraulic conductivity, which depends
on the medium through which the gas is flowing and the
viscosity of the gas; wg equals the unit weight of the
gas; and grad p equals the pressure gradient.
Because of the profound effect of the particle
size distribution of a porous medium on its hydraulic
231
-------
Figure 4.3 VARIATION IN GAS COMPOSITION WITH TIME IN CELL A
PROM INVERTED COLLECTION CAN AT 13-PT DEPTH
lo
u
to
100
90
80
I TO
S W
M
I"
I «o
3 30
10
10
0
20 *0 *0 »0 100 120 1*0 160 160 800 220 2*0 2tO 280 300 320 3*0 360 380 *00 MO **0 *60 *80 SOO 320 3*0
EUM1D TIME SINCX CELL COMPLETION (DATS)
Source: [125]
-------
SANITARY LANDFILL (bottom layer) Source: [671
10
U)
GAS CONCENTRATION VS TIME FOR PROBE I03A IN REFUSE (BOTTOM LAYER)
Not*: 02 concentration was generally
less than 2.0 percent
100
200 300 400 500 600 700
ELAPSED TIME IN DAYS SINCE PROBE WAS PLACED IN REFUSE
800
900
-------
conductivity, a wide range of conductivities is commonly
observed in various types of soil Table 4.13 lists
approximate values for the conductivities of some
natural granular media. The conductivity of porous
media to gases is sharply reduced when moisture is
added to the medium. This reduction is greatest in the
finer textured media which have high moisture holding
capacities [65].
Table 4.13 APPROXIMATE HYDRAULIC CONDUCTIVITIES
Medium
Alluvial Sand
Silt Loam
Clay Loam
k (Water)
ft/day
250
1.0
0.2
kg/wg(Gas)*
ft4/lb-day
267
1.07
0.21
* Viscosity = .015 cp.
Source: [65]
Gas pressure gradients in a landfill can arise in
two ways. First, a net production or uptake of gas
by decomposition processes will increase or decrease
the pressure inside the fill with respect to the outside
air, causing a displacement pressure. Second, if there
is a difference between «the density of the fill gases and
234
-------
the outside air, a buoyancy pressure will develop due
to gravity. The development of either displacement
or buoyancy pressure will depend on the hydraulic conduct-
ivity of the bottom, sides, and cover of the landfill.
It is assumed that the refuse itself will have a very
high permeability and hence will not influence local
pressure gradients, although this may not be the case if
the landfill contains large amounts of fine material.
It is apparent that the relative conductivities of the
landfill boundaries are a major factor in determining
the course of gas flow within a landfill. In addition, the
rate of gas production may determine what portion of
the gas is vented to the atmosphere or to the underlying
strata.
During initial CO- blooms, experimental data has
indicated horizontal diffusion velocities of 1-1.4 ft/day
and downward vertical velocities of .8-1 ft/day. As
the pressure gradient diminished, horizontal and vertical
velocities decreased to .24 and .22 ft/day respectively
[32,65,67].
The quantity of CO2 and CH. produced by a landfill
and, perhaps more important, the rate of gas production
are closely related to several variables: the quantity,
composition, and compaction of the refuse; the temperature
235
-------
and moisture content of the fill; the type of decom-
position taking place; and the hydraulic characteristics
of the fill boundaries [65,66,67,111,125,183], Broadly
speaking, the greater the quantity, putrescibility,
moisture content and temperature of the refuse; the more
loosely it is compacted; and the more porous the fill
boundaries-, the higher the rate of gas production within
the landfill. It is important to note here, that although
environmental or engineering techniques, such as the
choice of a clayey cover material of low conductivity
or the installation of gas underdrains, may effectively
control the production rate and escape routes of land-
fill gases, the production of gases within the landfill
after refuse emplacement is inevitable.
Table 4.14 [65] shows the variation in gas pro-
duction from the decomposition of the three main types
of organic materials found in refuse, carbohydrates,
proteins, and fats, under aerobic and anerobic con-
ditions at constant temperature and pressure. Though
theoretical predictions of gas generation from landfills
in the field are seldom accurate, the theoretical data
in Table 4.14 serves to corroborate the general state-
ments made earlier regarding the typical pattern of
landfill gas production. Aerobic decomposition, which
236
-------
Table 4.14 GAS PRODUCTION FROM ORGANIC SUBSTANCES
Type of Compound Carbohydrate
Composition, % by weight
Carbon 45
Hydrogen 6
Oxygen 49
Nitrogen 0
Composition, atoms/atom C
Carbon 1
Hydrogen 1.67
Oxygen 0.83
Nitrogen 0
Aerobic Respiration
Production & Utilization/lb Compound
Ib cu ft Ib
02Uptake 1.20 14.5 1.46
C09 Produced 1.65 14.5 1.95
Net Uptake 0
Production & Utilization/lb Carbon
0, Uptake 2.67 32.2 2.75
C02Produced 3.67 32.2 3.67
Composition Exhaust Gas, % by Volume
N2 79
CO, 21
Av. Molecular Weight 31.3
Anaerobic Respiration
Gas Productibn/lb Compound
C02 0.82 77.33 0.95
CHa 0.30 7.3 0.36
Total Produced 14 . 6
Gas Production/lb Carbon
C02 1.83 16.1 1.78
CH| 0.67 16.1 0.69
Composition Exhaust Gas, % by Volume
CO 50
CH| 50
Av. Molecular Weight 30.0
Protein
53
7
23
17
1
1.58
0.33
0.27
cu ft
17.6
17.1
0.5
33.2
32.2
80
20
31.2
8.3
8.8
17.1
15.6
16.6
49
51
29.7
Fat
77
12
11
0
1
1.89
0.11
0
Ib
2.92
2.82
3.79
3.67
0.82
0.72
1.07
0.94
cu ft
35.2
24.8
10.4
45.8
35.2
84
16
30.5
7.2
17.6
24.8
9.4
22.8
29
71
24.2
Gas volumes at 1 atmosphere, 70° F.
Source: [65, p.50]
237
-------
always predominates at first, results in large
quantities of CO produced per pound of organic mater-
ial. Anaerobic decomposition processes, on the other
hand, which take over after a period of weeks or months,
lead to a reduction in CO production and to the for-
mation of CH., especially from fat decomposition. Since
anaerobic decomposition usually proceeds more slowly
than aerobic processes, CH. production can be expected
over a considerable period of time.
The active life of a sanitary landfill is difficult
to predict due to the many factors influencing the speed
of decomposition. In colder climates, a landfill with
a large non-putrescible content may decompose very
slowly. Excavations through a completed landfill in
North Dakota which had been undisturbed for more than
25 years revealed refuse items such as pictures in old
catalogs and grocery sales slips that were still legible
[20J. Using the total amount of carbon removed as a
measure of the active life of a landfill, one study
calculated that 50% of the initial carbon content of
a test landfill in California would have been removed in
57 years and that 90% would have been removed only after
950 years [66].
The problems of fire and explosion hazards from
238
-------
landfills occur when methane gas, generated in the fill,
seeps into enclosed areas such as adjacent sewer lines
or the foundations of buildings located above the fill.
The lower explosive limit (LEL) of a methane-air mixture
is 5% methane. Methane seepage from completed landfills
is both inevitable and sufficiently abundant to produce
a significant hazard. At an experimental landfill it
was calculated from field data that 24 times as much
methane passed through the fill cover as went into the
surrounding soil [66].
A long-term field study was conducted on gas seep-
age at a 29-building public housing development con-
structed over a completed sanitary landfill (70]. The
composition and age of the fill varied considerably
in different locations. In some areas filling had been
recently completed with putrescible refuse while in
other areas the fill contained relatively inert materials
which had been filled many years earlier. A survey
of the crawlspace atmosphere was begun 15 months after
the foundations were poured. Combustible gases between
20 and 24% of the LEL were found in three of the
buildings and concentrations between 10 and 20% were
found in three others. Six months later, two buildings
showed explosive mixtures, five showed concentrations
239
-------
in excess of 50% LEL and six others concentrations
between 10 and 50% LEL. At this point mechanical
ventilation with portable suction pumps of all the
dangerous crawlspace was begun. During the second and
third years methane seepage become so rapid in 5 of
the buildings that automatic ventilation equipment was
installed. Methane generation and seepage showed a
definite seasonal variation, increasing in the early
spring, peaking in July and August and tapering off during
the fall and winter. It was postulated that higher under-
ground temperatures stimulated microbial activity which
increased gas production and the pressure of existing
gas reservoirs against the floor slab. After 5 years
gas seepage had ceased. Two explanations of this rel-
atively short period of gas seepage were suggested.
First, the age of the fill material may have been greater
than assumed. Second, heat transmitted to the fill
materials through the heavy underground concrete founda-^
tions of the tall apartment buildings may have speeded up
bacterial action, especially during the colder months
of the year, causing a large portion of the decompo-
sition to be completed in a relatively short period of
time [70],
Although very few instances of fires or explosions
240
-------
have been traced to methane seepage from landfills,
the potential danger of such explosions is clear.
Attempts to seal off the foundation slabs of the apart-
ment building with asphalt and sodium silicate coatings
proved unsuccessful. At other sites construction of
special gas vents and underdrains have allowed controlled
escape and burning of methane [58],
Groundwater and Surface Water Pollution
In 1965, total water use in the United States for
public supply, rural domestic, livestock, irrigation,
and industrial purposes (including thermoelectric power
but excluding hydroeclectric power generation) was
estimated at 310 billion gallons per day or approximately
1600 gallons per capita per day [134]. About 25% of this
total was supplied by groundwater sources on a nation-wide
basis. In certain key catchment areas, however, ground-
water sources account for 40-80% of the total water supply,
Moreover, in some river basins all available surface
supplies are frequently utilized and groundwater must be
withdrawn faster than it is recharged. Figure 4.5
shows total water withdrawal by states (excluding hydro-
electric power) from groundwater sources and from sur-
face water sources in million gallons per day in 1965
[134]. From Figure 4.5 it is evident that certain
241
-------
Figure 4.5 TOTAL WATER WITHDRAWALS BY STATES
(excluding hydroelectric power) FROM GROUNDWATER
RESOURCES (upper number) AND FROM SURFACE WATER
RESOURCES (lower number)
(in million gallons per day, 1965)
Source: 1134]
242
-------
states, particularly in the western U.S. rely heavily
on groundwater supplies. By 1980, average daily with-
drawals are expected to reach 600 billion gallons per
day. Although a large portion of this increased demand
will be met by re-use and reclammation of water, up
to a 100% increase in the demand upon groundwater
resources can be anticipated [72].
Ground and surface water pollution from sanitary
landfills occurs through two different leaching processes,
The first process is the solution of solid and liquid
pollutants present in a landfill by excess water and
their subsequent transportation out of the fill into
surface or groundwater reservoirs. The second process
is the solution of CO- in the groundwater adjacent to
the fill and the consequent solution of minerals, mainly
carbonates and bicarbonates resulting in excessive
alkalinity and hardness.
Water quality can be defined as all the character-
istics of water which affect its suitability for use.
These include chemical, physical, and biological prop-
erties, specifically its organic, inorganic, and bacter- .
ial chemical content, taste, and odor [28], Impairment
of surface of groundwater quality due to leaching could
be a major externality imposed by sanitary landfills.
243
-------
These adverse effects take the form of decreased or
total disutility of water for domestic, commercial,
recreational, or public uses; or of direct and indirect
deleterious effects on public health.
The generation and movements of contaminants in a
sanitary landfill are dependent upon the content,
spatial distribution and time variation of moisture in
the fill materials [154]. This process of leachate
generation and movement within the landfill takes place
according to a water budget and the laws which govern
moisture routing in porous materials [32,154]. In the
soil cover, the uppermost layer of a landfill, moisture
is usually added by precipitation and removed by runoff,
evapotranspiration, and drainage. In the underlying
areas, moisture is added by drainage from overlying layers
or removed by evapotranspiration if roots penetrate
to that depth. The hydraulic characteristics of the
refuse layers in a sanitary landfill are similar to those
of unsaturated, permeable soils. From the saturation
point the soil drains rapidly at first and then at con-
tinually decreasing rates until, after two days, the
drainage rate is small. The moisture content of the
soil after the drainage rate has become small is known
as the field capacity. When-water is applied to a layer
244
-------
of soil at field capacity it drains rapidly to the
underlying layers. After the moisture content has
decreased to field capacity, the soil remains at
approximately that moisture content, unless moisture
is removed in other ways.
The moisture content of soil layers within the root
zone of surface vegetation can be reduced considerably
below field capacity by evapotranspiration. The moisture
content of the soil at which no further water can be
withdrawn by plant roots is known as the permanent
wilting percentage.
As moisture is added to a soil below field capacity
it does not distribute itself uniformly throughout the
soil. Each layer must reach field capacity before sig-
nificant quantities can drain to the underlying material.
The mass of percolating water is preceeded by a wetting
front or region of steep moisutre content gradient.
The difference between a soil's field capacity and
permanent wilting percentage is referred to as available
water. After determining the available water storage
capacity of each soil and refuse layer it is possible
to apply the principle of continuity to moisture routing
through sanitary landfills and associated materials [154].
The production of a leachate from a sanitary landfill
245
-------
is entirely dependent on sufficient net addition of
water to the fill. The moisture required to bring
various refuse samples to field capacity has been cal-
culated at between 1.3-2.98 inches per foot of compacted
refuse [154,65]. Thus for a 20' deep landfill, approx-
imately 26"-60" of added moisture would be needed to
bring the fill to field capacity. Figure 4.6 [76] is
a map which shows areas of average annual potential
moisture infiltration in the U.S. Prom the map it is
evident that there is potential infiltration of water
into landfills over a large part of the United States.
As long as there is net infiltration leachate will
eventually begin to be produced by a landfill. Indeed,
by taking the infiltration and moisture storage figures
from above it may be concluded that leachate production
from a 20' thick landfill could begin in less than a year,
in some areas.
The composition of leachate is primarily dependent
on: the types and amounts of material in the fill mater-
ial; the numerous supplemental variables controlling
the type and speed of decomposition? leachate flow rate;
and a number of interactions among the chemical processes
taking place in the landfill. Solution of C0_ and inter-
mediate decomposition products and leachate flow rate
246
-------
Figure 4.6 AVERAGE POTENTIAL INFILTRATION FOR THE U.S,
(inches)
0-10
10-20
20 - 30
30
Source: [76, p. 125]
-------
are major factors in determining leachate pH- which
has a profound effect on the corrosive and solvent
properties of the leachate. Leachates of low pH (below
5.5) are capable of greatly increased solubility of iron
and iron salts, zinc, and lead.
Several comprehensive empirical studies have been
completed on the quantity and quality of leachate
produced, both in experimental and laboratory landfills,
and actual field situations [30,32,65,76,91,111,150],
One summary of several of these chemical analyses is
presented in Table 4.15. One study in which measurements
were made from a laboratory sanitary landfill showed that
most leachate components were being produced steadily
or in increasing amounts after 600 days [76]. The
analytic results show that leachate is a grossly polluted
liquid which compares in dissolved inorganics with
chemical plant wastes, compares in organic content with
food-processing plant effluent, and may contain hazardous
concentrations of trace elements and heavy metals [91].
Naturally, many other deleterious substances are undoubt-
edly present in landfills and can be leached into ground-
water. These can include medicines, inflammable liquids,
pesticides, insecticides, weed killers, and other toxic
chemicals for domestic use as well as inadvertent or
248
-------
Table 4.15 COMPARISON OF REFUSE LEACHATES WITH U.S. PUBLIC
HEALTH SERVICE STANDARDS (in parts per million)
to
*»
vo
U.S. Public Health
Service Standards1
Substance
Alkyl benzene
sulfonate
Arsenic
Chloride
Copper
Carbon chloroform
extract
Cyanide
Fluoride
Iron
Manganese
Nitrate
Phenols
Sulfate
Total dissolved
solids
Zinc
Barium
Cadmium
Chromium (Cr+6)
Lead
Selenium
Silver
Group I2'*
0.5
0.01
250
1
0.2
0.01
0.3
0.05
45
0.001
250
500
5
Group 11* >s
0.05
0.2
3.4
1
0.01
0.05
0.05
0.01
0.05
Blackwell6
4.31
1,697
0.05
0.024
5,500
1.66
1.70
680
19,144
8.5
<0.05
0.20
2.7
<0.1
LW5B
Dupage7
0.72
<0.10
1,3315
< 0.05
< 0.005
2
6.3
0.06
0.70
2
6,794
0.13
0.80
< 0.05
0.15
0.50
< 0.10
< 0.1
LW6B
Dupage*
0.30
4.6
135—
< 0.05
0.02
0.31
0.6
0.06
1.60
2
1,198
< 0.10
,0.30
< 0.05
< 0.05
0.50
< 0.10
< 0.1
-------
K>
cn
o
Table 4.15 (cont.) COMPARISON OP REFUSE LEACHATES WITH U.S. PUBLIC
HEALTH SERVICE STANDARDS (in parts per million)
Substance
U.S. Public Health
Service Standards *
•»!••«, 1 2 "3 S-<*-*N.IIVX 1 1 '
Blackwell1
LW5B LW6B
Dupage7 Dupage*
Ammonium
Alkalinity (as CaCO-)
Hardness (as CaCO*)
Phosphate
Titanium
Aluminum
Sodium
Hexane solubles
Biological oxygen
dema'nd 13
Chemical oxygen
demand pH
3,255
7,830
6
2.20
900
350
54,610
39,680
4,159
2,200
1.20
0.10
810
18
14,080
s;ooo
6.3
1,011
540
8.90
0.90
74
7
225
40
7.0
*U.S. Department of Health, Education, and Welfare (1962).
2Nitrates exceeding 45 ppm dangerous for infants.
'Should not be used if more suitable supplies available.
11 Larger concentrations should be rejected.
5Fluoride is temperature dependent.
6Probably represents leachate from compaction and infiltration.
7Leachate from refuse about 6 years old.
'Leachate from refuse about 17 years old.
'Data provided by Metropolitan Sanitary District of Greater Chicago.
10Data from the files of the Illinois Department of Public Health.
^Rare earth and thorium production (Butler, (1965, p. 63)].
1 Questionable values underlined.
13 20-day biological oxygen demand for leachate. Other values are 5-day BOD.
Source: [91, p. 1141
-------
illegal deposits of industrial wastes, such as hex-
avalent chromium or cyanide. Normally one would ex-
pect to find only small quantities of these latter
substances in leachate. However, if large, truckload
size, deposits of these materials were placed in a
landfill, special pollution hazards to groundwater
supplies would develop [65]. The studies also show that
leachate composition is highly variable across sites.,
and any generalizations may be unreliable.
The production of leachate is only the first step
in groundwater pollution; several factors may mitigate
or aggravate the pollution potential of a leachate source,
As refuse leachate enters the ground and migrates
through the underlying soil and rock layers, it is
attenuated by ion exchange, dilution, dispersion, com-
plexing, and filtration, and oxidation[91]. The extent
of attenuation due to ion exchange is largely dependent
on the type and texture of the subjacent materials and
the chemical composition of the leachate and its pH.
Clays are particularly active in ion exchange; the
amount of exchange of a particular ion depending on
1) the type of clay involved; 2) the ions already pre-
sent on the surface of the clays; 3) the other elements
in solution and their concentration. The generally acid
251
-------
nature of leachate compounds the potential pollution
problem because low pH values tend to reduce exchange
capacities of renovating soils [76], For a complete
treatment of ion exchange on clay minerals the reader
is referred to Grim [81,82].
The filtering action of subjacent materials is
primarily a function of their texture. Fine-textured
materials have a high capacity to retain dissolved
solids and permit a low rate of groundwater movement
due to their permeability. Coarser sands and gravels
have less filtering capacity and allow higher rates of
groundwater movement. British studies [see 91] estab-
lished that by filtering refuse leachate through sand
and gravel "general purification from organic matter
can be effected under anaerobic conditions". Purifi-
cation from chlorides, sulfates, and ammonia was found
to be much less complete. For sites in northeast
Illinois it was estimated that for leachate migrating
through silty and clayey tills (heterogeneous mixtures
of clay, sand, gravel and boulders), unfractured shales,
and clays the total dissolved content would be reduced
by one to two orders of magnitude in travelling a dis-
tance of 5 ft. An order of magnitude is a range of
values between a designated lower value and an upper value
252
-------
ten times as large. Sands and silts would reduce
total dissolved solids by one order of magnitude in
500 ft. and gravel and fractured rocks would be con-
siderably less efficient [91], Filtration of bacteria
and other microorganisms in the subsurface is also
closely related to texture; granular surfaces accomplish
complete filtration in short distances while cracked
or fissured material may permit greater travel.
The evaluation of the groundwater pollution hazard
from any landfill site must take account of all the factors
which influence hydrogeologic conditions in the sub-
surface; depth of the water table; permeability; flow
gradients; distance to wells; geologic characteristics
of the aquifer network, etc. [110], At sites with rel-
atively low capacities for ion exchange, complexing
and filtering, pollution control must rely on dilution
and dispersion of pollutants to acceptable levels before
groundwater supplies reach points of use. Theory and
empirical work suggest that for refuse leachate flowing
in a uniform aquifer, longitudinal dispersion has little
effect on the concentration of pollutants appearing
in downstream wells, except possibly at very great
distances from the source. In non-uniform aquifers
both longitudinal and transverse dispersion will be
253
-------
effective in diluting the pollutants. However, any
calculation on the probable degradation of groundwater
supplies must consider the particular characteristics
of the site, including the cones of influence of pumping
wells [65].
Low travel velocities and diffusion rates can
produce serious pollution problems. Pollution may not
be noticed for years or decades. After the pollution
is discovered the damage can not be repaired by simply
stopping the source of contamination. With careful
planning, however, groundwater pollution need not occur.
For controlling pollution there are four possible objectives
in landfill design: 1) elimination of leachate production;
2) migration of leachate under acceptable conditions;
3) recovery of leachate after migration; 4) retention
and recovery of leachate. These objectives can be
accomplished using standard engineering techniques such
as grading, drainage construction and pumping, etc.,
if the hydrogeologic conditions of the site are properly
understood.
Field study results of groundwater quality im-
pairment due to leaching from sanitary landfills are also
available for sites in California, Pennsylvania, Illinois,
Wisconsin, South Dakota, and Texas £65,34,76,63,142,91,
254
-------
193,10,166,118,99,72]. In addition to these, many
publications have presented a more general discussion
of various aspects of the problem of leaching from
sanitary landfills [16,26,32,35,38,39,47,58,85,86,88,
90,107,108,109,110,113,129,130,132,135,140,151,153,
156,170,185,188,195,197,200,202].
Typically, the most concentrated pollutants ob-
served at these sites are iron, nitrate (oxidized from
ammonia and organic nitrogen), and hardness (an unde-
sirable larger mineral content in water, particularly
of carbonates and bicarbonates).
The second and perhaps more prevalent cause of
groundwater pollution by sanitary landfills is carbon
dioxide* As outlined above, CO. is a major product
of both aerobic and anaerobic decomposition processes,
and is produced in great quantities. Although the
vast majority of the CO produced is usually vented
^
through the fill cover, a substantial quantity may
migrate downwards until it contacts and dissolves in the
groundwater reservoir. Recent evidence from sites
in California indicates that for a 20* section of fill,
vertical downward migration of CO. varied from 22,400-
13,500 Ibs/acre/yr. for several years. These quantities
represented approximately 4-5% of the total CO production
255
-------
in the fill [66,67], Deep sampling wells revealed
the presence of CO production at depths up to 100 ft.
below the refuse in homogenous sand and gravel soil.
Moreover, if excess water is present within the fill
C02 will dissolve there and aid in the further solution
of pollutants by the leachate [65,66]. The three prin-
ciple effects of CO on groundwater are increased hard-
ness, corrosivity, and acidity. These effects are
functions of the relative quantities of CO2, carbonates,
and bicarbonates present in the groundwater and the
water's pH.
CO2 dissolves in water to form carbonic acid,
H CO , which, in turn, reacts with otherwise insoluble
or slightly soluble carbonates in soil and rock deposits
to form soluble bicarbonates. As CO. is removed by
the reaction more C02 is free to dissolve. This pro-
cess continues until equilibrium is reached with the
partial pressure of C02 in the atmosphere right above
the water. This entire process of mineral solution
can be reversed if the affected groundwater flows into
a region where the soil atmosphere is low in CO,,.
However, this process of reducing the mineral content in
groundwater can take many years to complete [651. Cal-
ifornia studies concluded that a sizable CO- concentra-
256
-------
tion in the bottom layer of refuse can be expected
for many years, with CO. passing into the ground. As
usual, the pollution potential of any site depends on
the particular environmental factors prevailing there.
However, in general, it would seem that C0_ poses the most
serious threat to groundwater from landfills, since it
is an inevitable product of refuse decomposition.
Excessive concentrations of carbonates, bicarbonates,
and hydroxides cause alkalinity and hardness. Alka-
linity, while not considered deleterious to human health,
imparts an objectionable and unpalatable taste to water
for domestic use. Furthermore, alkalinity is detrimental
to many industrial processes, especially food and bev-
erage production, which have recommended threshold
values of 30 to 250 ppm [65,36], Typical alkalinity
values drawn from wells effected by the CO. production of
nearby landfills range from 200 to 600 ppm.
Water hardness is caused by bivalent metallic ions
(calcium, magnesium, and ferrous iron) which react with
soap to form precipitates and impair cleaning action.
In addition, water containing dissolved calcium and
magnesium carbonates and bicarbonates will form a scale
of calcium carbonates and magnesium hydroxides (tem-
porary hardness) when evaporated or heated. Such
257
-------
scaling eventually deteriorates plumbing systems and
their related components, such as hot water heaters.
The maximum recommended hardness limits for domestic
sources is 80 ppm [32], Hardness values for affected
wells may range up to 800 ppm.
Prominent examples of increased alkalinity and
hardness in groundwater supplies have been documented
[32 ref (4), 65,66,67] for sites in California.
As with pollution problems due to leaching, gas
movement into groundwater supplies can be controlled
by the application of the proper engineering techniques.
Several types of drainage/ventilation schemes have been
used successfully to control gas movement [32,59,66,67,
70], Gas barriers and liners have also been developed,
which can arrest or retard downward gas movement [32,
66,67],
Settlement
As mentioned above, landfills settle gradually
as the fill materials undergo decomposition. A land-
fill may settle due to its own weight, the weight of
the cover materials, the weight of any structure placed
on it, or weight added to the fill by precipitation.
The effect of settlement on the usefulness of a completed
landfill site would not generally by considered an
258
-------
externality although, as we discuss in Chapter 5,
its economic effects may be similar in some circum-
stances. Furthermore, settlement influences other
externalities associated with landfills, particularly
water pollution. We thus include a brief discussion
of the problem of settlement.
There is little information concerning the engin-
eering characteristics of sanitary landfills. From
the few titles in the literature [101,112,121,174,183]
several important facts are evident. Because of the
extremely diverse nature of fill materials and the un-
predictable extent of settlement, landfills are a
particularly poor foundation material for construction.
Fill materials have been noted for their erratic com-
position and extremely erratic, but low, densities [174].
This heterogenous composition and the low density and
stability of fill materials creates severe foundation
engineering problems, related to bearing capacity. If
a building foundation is relatively small compared to
the thickness of the soil cover of a landfill, the
foundation may punch through the cover into the fill
materials. Alternatively, if the foundation is somewhat
larger and if the fill material is not greatly weaker than
the soil cover, bearing failure may occur by rotation
259
-------
of a segment of the soil cover and fill materials.
Estimates put the bearing capacity of a sanitary
landfill at 500-800 Ibs. per square ft. [174], Such
low bearing capacities may be circumvented, however,
through several engineering techniques; foundation
design; additional fill covers; and support piles or
piers which extend through the entire fill. Naturally,
foundations laid on support piles will be independent of
landfill characteristics but their construction may
be difficult.
Settlement problems are more persistent and
possibly more serious than those of low bearing capacity.
In general, landfills are subject-to very irregular
patterns of settlement. Different sectors or components
of the fill may settle at very different rates. Although
accurate and general data are lacking/ limited informa-
tion [121,125,189] suggests that sanitary landfills
typically settle 10-30% or more of their original
thickness. Since this settlement is not uniform, large
surface depressions may appear in the fill cover. When
water drains into these depressions, the rate of leachate
production below them can be greatly increased. The
water pollution problem may thus be much more severe than
260
-------
if settlement were uniform.
Leaching From Ash Landfills
Solid residues from municipal refuse incinerators
are generally disposed of by landfilling. Like refuse
landfills, residue landfills are potential groundwater
and surface water pollution hazards. The same physical
and environmental factors described above for refuse
landfills govern leaching processes in residue land-
fills; the only major difference being the dissimilar
composition of the fill materials. Very little research
has been completed on the chemical, physical, and
hydraulic characteristics of incinerator residues [98,
167], Table 4.16 [176] presents the typical range of
values for the various components of residue. Table
4.17 [1] presents the physical characteristics of res-
idues from several types of incinerators. It is appar-
ent from Tables 4.16 and 4.17 that incinerator residues
and refuse have quite different physical compositions.
One can only speculate how these differences affect
the hydraulic characteristics of residues compared
to refuse. Residues have quite low moisture contents,
which might indicated high field capacities. On the
other hand, the uniform, dense nature of residues and
the large proportion of inert, non-absorbing components
261
-------
Table 4.16 INCINERATOR RESIDUE COMPOSITION
(percent)
Material
Range
Metals
Glass
Ceramics, stones
Clinkers .
Asha
Organic . .
. 19 to 30
. 9 to 44
. 1 to 5
. 17 to 24
. 14 to 16
. 1.5 to 9
aExclusive of other material listed.
Source: [176,p. 98]
Table 4.17 INCINERATOR RESIDUE COMPOSITION
Incin-
erator
A
B
C
D±
E
P
6
Moisture,
as
Sampled
(%)
15.0*
24.5
0.3
-
21.8
24.8
10.5
Heat,
Dry
Basis
(Btu/lb)
170
200
180
-
520
940
70
Ash, Volatiles, Density,
Dry Dry as
Basis Basis Sampled
(%) (%) db/yd3)
97.4
98.4
98.0
. -
97.0
92.7
99.4
2.6
2.0
2.0
-
3.0
7.3
0.6
(t)
(t)
(t)
-
1490
1620
1600
* Assumed.
t No measurement made.
± No laboratory analysis performed,
Source: [1]
262
-------
point to extremely low field capacities. Because the
exact hydraulic features of residue landfills remain
unknown, it is difficult to predict leachate generation
in field situations.
Preliminary information on the chemical composition
of incinerator residues indicates that they contain
about the same percentage of soluble material by weight
as refuse but that most of the residue solubles are
inorganic rather than organic. Leachates from residue
landfills will reflect this predominance of inorganic
constituents; in fact, one expects residue leachate to
be similar to incinerator process water, with its high
inorganic and low organic pollution loading. If the
residue materials have a significant organic content,
aerobic and anaerobic decomposition of these materials
may occur, resulting in CO generation [33]. Solution
of this CO in infiltrating waters, resulting in acidity,
could increase the alkalinity and hardness value of the
residue leachate. However, since the organic content
of residues is usually quite low, CO- related pollution
problems are much less severe for residue landfills than
for refuse landfills.
263
-------
F. OPEN DUMPS
Open clumping is certainly the most noisome method
of solid waste disposal. Open dumping leads to:
groundwater pollution; vermin and vector propagation;
and air pollution due to open burning,
Noise
In general, noise problems due to dumps are less
severe than those due to sanitary landfills since heavy
machinery is not used in dumping operations. The
primary noise sources at dumps .are collection trucks
and private vehicles discharging refuse. Thus the
noise level at a dump will rarely exceed the 90-95 dBa
level which is characteristic of large collection trucks
operating at full power [19],
Odor
Malodor emissions are clearly much more severe for
open dumps than for sanitary landfills. The source of
malodors, the putrescible portion of municipal refuse,
remains uncovered indefinitely at an open dump, thereby
presenting the greatest opportunity for odor emissions.
A survey of the site data from regions 1 and 2 [130,
132] of the 1968 National Survey of Community Solid
Waste Practices revealed that about 50% of land disposal
264
-------
sites which received no cover were classified as
needing an odor control program. As mentioned pre-
viously, only about 10% of all landfill sites in region
1 and 2 which received daily compacted earth cover were
listed as needing an odor control program.
As cited in the earlier section on odors from
landfills, the same factors of wind speed and direction,
temperature* vertical temperature gradients, humidity,
topography, and the condition of the odor source will
be important in determining the emission, dispersion,
and persistance of odors from dumps.
Unsightliness
Probably the most subjective and elusive exter-
nality imposed by open dumping of solid wastes is un-
sightliness. Unlike air, noise, or water pollution,
unsightliness seems to have no readily measurable
physical characteristics or parameters. Attempts to
quantify unsightliness at land disposal sites have been
made using surveys with small numbers of observers
[4,130,132], In the 1968 National Survey of Community
Solid Waste Practices land disposal sites were
classified as "sightly" or "unsightly" by field in-
vestigators. A tabulation of these site classifications
showed that about 24% of those sites receiving a daily
265
-------
compacted earth cover were classified as unsightly.
About 60% of those sites with other than a daily com-
pacted earth cover were found unsightly, while 91%
of those sites receiving no cover of any kind were
judged unsightly. From these statistics, it is evident
that the appearance of land disposal sites is greatly
improved through the use of sanitary landfilling
techniques.
Surface and Groundwater Pollution
Open dumps cause many of the same surface and
groundwater pollution problems as sanitary landfills.
Several instances of surface and groundwater impair-
ment have been traced to leaching from open dump sites
in California, Illinois, New York, and Great Britain
[See 65 refs (32,33,46,66), 193], Unfortunately, there
are no comprehensive quantitative studies which deal
specifically with water pollution problems from open
dumps. However, the physical differences between dumps
and landfills suggest that the water pollution effects
may also differ.
At a typica.1 open dump, refuse is simply dumped
on top of the ground. No compacted earth cover is
provided and vegetation rarely grows over the refuse.
266
-------
These differences affect decomposition processes, gas
production, and leaching.
Although decomposition becomes anaerobic in a
sanitary landfill after several weeks, aerobic de-
composition may continue indefinitely at an open dump,
because of the opportunity for oxygen infiltration into
the loosely piled, uncovered refuse. If refuse decom-
position is aerobic, relatively more CO^ and no CH. will
be produced from the same quantity of refuse. In addi-
tion, since aerobic decomposition of refuse proceeds
faster than anaerobic decomposition, the rate of CO2
generation is greater. These factors will increase the
hazards of CO- related groundwater pollution. Conversely,
since dump refuse has no cover and a greater surface area
than landfill refuse, gas evolution from dump refuse
will be proportionally greater, thereby offsetting this
increased gas production.
As in sanitary landfills, leaching from dumps
depends on sufficient net addition of moisture. Since
dump refuse is piled openly above the ground net moisture
infiltration into dumps may be higher than landfill
infiltration, since no moisture is removed by evapo-
transpiration from a soil cover. This may be offset
267
-------
by evaporation from the refuse itself. Leachate
generation and quality are also affected by dump
conditions. First, since dump refuse is not graded
to afford good surface runoff, moisture infiltration
in low-lying areas will be high. Second, little or no
compaction of dump refuse means that the refuse will
have a low field capacity, allowing high leachate flow.
Leachate which percolates quickly through refuse can
become more acidic than a slowly moving leachate. Acidic
leachate has several corresponding adverse effects:
increased leachate corrosivity, mineral solubility,
and impairment of ion exchange capacities of renovating
soils. Third, as leachates exit from the dump refuse
only a proportion will enter the underlying soil layers.
Because leachate production may be heavy during periods
of heavy precipitation, surface water pollution may be
cause by the resulting leachate runoff. Fourth, leachate
which does enter subjacent soils may have a lower
pollution load due to dillutioh. Finally, dumping does
not require excavating large pits for refuse emplacement.
Consequently, leachate pollutants may be more thoroughly
attenuated in the thicker renovating soil layers which
remain. This effect is particularly important at sites
268
-------
where the water table lies close to the surface.
In sum, then, it seems probable that groundwater
pollution is decreased while surface water pollution
is increased by using open dumping techniques.
Vectors and Vermin
A major externality of open dumping is the propa-
gation of vectors and vermin at dumping sites. The
principle vectors are various species of flies and
rats. Dumps are important to these vectors as a food
supply, breeding ground, and place of shelter.
The unequivocal relation of human disease to
solid waste disposal methods, in this case open dumps,
is not an easy task, however. In his excellent study
of solid waste/disease relationships. Hanks [85] has
stated five steps which must be satisfied to establish
a firm solid waste/disease link:
1) discovery of harmful agents in the waste, or
2) demonstration that such agents develop within,
or in association with, the wastes;
3) discovery of disease or other effects among
the population which may reasonably be
associated with these agents;
4) demonstration of the pathways by which the
269
-------
effects are accomplished;
5) demonstration of the absence of effects
following interruption of these pathways by
one means or another, or their absence in
populations not similarly exposed.
Flies represent a significant externality assoc-
iated with open dumps, both as disease vectors and as
nuisances. Sacca [164, p. 346] states:
The dump constitutes a place of attraction for
masses of adult flies which can permanently live,
feed, and reproduce themselves on it. The pro-
duction of new individuals is often huge there,
owing to the abundance of larval food, made up
of organic residues of vegetal and animal origin
which have long fermented and are mixed with inert
material (rags, waste paper, etc.) that help its
aeration. During the favorable season, i.e.,
when the temperature allow the winged insects to
fly, the dumps are a source from which masses of
adult flies move toward human dwellings. The
house flies often gather in the domestic environ-
ment in large numbers and are therefore important
from the epidemiological viewpoint.
The literature contains many references on the
importance of dumps for the production and maintenance
of fly populations. See Hanks [85] for an excellent
literature review. One source [169] estimates that
70,000 flies may be expected to propagate and emerge
from one cubic foot of garbage.
Flies are firmly established as mechanical carriers
270
-------
of a large number of agents pathogenic to man [85].
The most important of these agents are viruses and
bacteria. Some types of protozoa and worms, tri-
trichomonas foetus, arthropods, and myasis are also
carried by flies [164]. The following diseases have
been linked to fly transmission: enteric diseases
(typhoid, bacillary and amoebic dysentery, diarrheas,
Asiatic cholera, helminth infections); myiasis, loiasis,
onchocerciasis; Ozzard's filariasis; leichmaniasis;
African sleeping sickness (trypanosomiasis); yaws;
tylaremia; bartonellosis; catarrhal conjunctivitis;
sandfly fever. Tentative additions to this list are:
anthrax, salmonellosis, protozoal infestations, trachoma;
poliomyelitis; tuberculosis; and hepatitis [85]. It is
very likely that in areas in which flies have wide and
prolonged access to concentrated sources of disease
agents, especially human and animal feces, they will
contribute significantly to human disease and mortality.
However, in areas in which sanitation and personal hygiene
standards are high in this regard, flies have little
opportunity to transmit disease. Hanks concludes:
Although much research is needed to clarify the
role of flies even in those few diseases in which
flies associated with solid wastes have been
271
-------
determined to be transmitters, the evidence
overwhelmingly demonstrates that control of
solid waste against breeding of domestic flies
greatly limits their population. Present know-
ledge of the dynamics of disease transmission,
exposure, dosage, etc., permits the conclusion
that such limitation can contribute to the pre-
vention of fly-borne disease [85, p.49].
Empirical studies have shown that a 6" daily
compacted earth cover of the kind used to cover refuse
in sanitary landfills can effectively prevent the
emergence of flies [15,53], On the other hand, effective
fly control at open dumps is "difficult if not impossible"
since fly larvae can be expected to emerge from all
parts of the dump in enormous number, especially in the
summer months [80]. While the exact proportion of the
fly population which depends on open dumps is unknown,
many studies speak of dumps as the major non-rural
production source [85]. Consequently, replacing open
dumps by sanitary landfills would greatly reduce the
size of fly populations in human environments.
The other major disease vector associated with
dumps is the rat. As with flies, rats have been im-
plicated in the transmission of a number of diseases.
These include: echinostomiasis; hemorrhagic septicemica;
histoplasmosis; lymphocytic choriomeningitis; plague;
rat-mite dermatitis; rat tapeworm infection; Rocky
272
-------
Mountain spotted fever? salivary gland virus infection;
salmonellosis; schistosomiasis; bilharziasis; sporo-
thrichosis; swine erysipelas; trichinosis; leptospirosis;
leichmaniasis; relapsing fever; tularemia; rickettsial
pox; murine typhus; and perhaps other diseases [85],
However, a positive relation between dump maintained
rat populations and increased incidence of any of the above
diseases is difficult to establish. It is known that
rats are extremely adaptable and have a prolific
breeding capacity [85,119], They can propagate in large
numbers whereever food- and harborage are available.
It is also well known that rats are frequently attracted
to refuse and are often found at open dumps and poorly
operated landfills, yet open dumps and food wastes in
general may have a very low marginal effect on the size
of rat populations. Indeed, if open dumps were com-
pletely eliminated it is unlikely that this would
significantly reduce rat populations, owing to the
adaptability of the rat to other food sources and places
of shelter, and the ready availability of these
necessities in many domestic environments. The control
of rat populations would require a coordinated program
to systematically eliminate all possible food sources
273
-------
and shelter. Only in the context of such a compre-
hensive program would better solid waste management
and the elimination of open dumping play a significant
role in eliminating the externalities imposed by the
commensual rat.
Open Burning
A final externality of open dumping relates to
open burning. Open burning refers to the open combustion
of refuse at a dump site, usually for the purpose of
volume reduction. The refuse may be piled in a special
area prior to combustion r or it may be ignited or catch
fire spontaneously where it is first dumped. Of the
sites sampled in the 1968 National Survey of Community
Solid Waste Practices 1129], it was found that 28%
practiced "planned and limited" burning, and 48%
practiced "uncontrolled" burning. The characteristics
of open burning operations have been the subject of
considerable discussion [ See, for example 5,17,57,
128]. Open burning of refuse has a number of adverse
effects: impairment of public health - especially dele-
terious effects on respiratory functions and eye
irritation; visible air pollution plumes, decreased
visibility, and esthetic complaints; soiling of the
274
-------
surrounding environment by fallout of partially
burned carbonized organic matter; and local vegetation
and structural damage due to aerosols from the com-
bustion process [68]. The extent of these adverse
effects is controlled by factors related to combustion,
atmospheric conditions, length, and extent of human
exposure, and the susceptability of the surrounding
vegetation and structure to damage.
Open burning emissions are similar to those of
municipal incinerators with no air pollution control
devices. Table 4.18 gives emission factors for open
burning of municipal refuse [52]:
Table 4.18 EMISSION FACTORS FOR OPEN
BURNING OF MUNICIPAL REFUSE
(pounds per ton of refuse)
Pollutant
Particulate 16
Aldehydes (HCHO) 0.1
Hydrocarbons (hexane) 2
Nitrogen Oxides (NO2) 2
Organic Acids (acetic) 13
275
-------
Comparison of open burning emissions with these given
earlier for municipal incineration shows that open
burning results in significantly higher emissions of
hydrocarbons, nitrogen oxides, and organic acids, and
lower emissions of particulates and aldehydes. These
emissions can be expected to vary, however, according
to combustion conditions [52,56,68,78,94,104].
About 30% of all municipal refuse is processed
by open burning [129], Given the higher emission
factors for open burning, and the impossibility of
controlling these emissions, the absolute quantity of
air pollutant emissions from open burning is much
larger than total emissions from municipal incineration,
Nevertheless, the contribution of open burning to
national air pollutant emissions remains quite small.
Hoffman [52] has estimated that in 1968 all solid waste
disposal activities accounted for 1.1% of particulate
emissions; .1% of sulfur oxide emissions; 5% of hydro-
carbon emissions; and 1% of nitrogen oxide emissions
on a national basis.
G. OCEAN DISPOSAL
A wide variety of solid wastes are disposed of in
the oceans adjacent to the U.S. by surface dumping and
276
-------
through submarine outfalls. Information concerning
current practices of refuse disposal in the ocean is
quite limited [ See 171 and 177]. Small scale ocean
disposal of refuse takes place off the Pacific coast
of the U.S. This refuse originates at military in-
stallations in Long Beach and San Diego and from a
cannery in San Francisco. However, in the last 25 years
there have been no sizable refuse disposal operations
at sea. Several U.S. coastal cities had disposed of
municipal refuse at sea prior to 1953 but this
practice was discontinued because floating debris caused
fouling of beaches which resulted in preventive legal
action [171].
Aside from beach fouling, the adverse effects of
ocean disposal of large amounts of raw refuse have
yet to be investigated. It is probable that large
scale ocean dumping of loose refuse causes severe
ecological disruptions in the dumping zone, but these
effects are merely speculative [171]. Indeed, it is
possible that marine disposal of baled refuse may be
favorable to the repopulation of overfished coastal
zones by many species [191]. Research is currently
underway to determine if artificial reefs made of
277
-------
baled refuse can provide beneficial habitats and food
supplies for a variety of marine species [ See 171].
In order to obtain an accurate understanding of
the effects of ocean disposal of solid wastes, com-
prehensive research is needed in three areas:
1) baseline environmental data; 2) the fate of waste
materials in the ocean ? and 3) the overall ecological
effects of waste materials on the marine biota.
Before one can properly assess the effects of
waste materials in the marine environment, it is essential
to have an adequate understanding of the natural
background biota fluctuations and the water mass char-
acteristics that are normal for 'a particular region.
Without this understanding it is impossible to dis-
tinguish normal variations from the effects of pollution.
Broad-spectrum studies are also required to gather
physical and chemical data on parameters such as
temperature, salinity, etc., which are important deter-
minants of marine biological processes [171].
At present, practical knowledge of how waste
materials mix, disperse, degrade, and accumulate in
resident species is very limited. Improved equipment
and sampling procedures need to be developed and
278
-------
improved so that accurate predictive models of waste
evolution in the sea can be obtained.
Finally, major research must be initiated to
determine the long run effects of chronic pollutant
concentrations on important species. Specifically,
survival, reproduction, and behavior of these species
must be observed in order to comprehend the probable
long-term effects of pollutants on the dynamics of
marine populations.
279
-------
REFERENCES
1. Achinger, W.C., and L.E. Daniels. An Evaluation
of Seven Incinerators. Environmental Protection
Agency, Washington, D.C. Publication Number
(SW-51ts. Ij) 1970.
2. Aerojet-General Corp. California Waste Management
Study. California Dept. of Public Health,
Sacramento. 1965.
3. Aerojet-General Corp. Fresno Region Solid Waste
Management Study - Vol. II, Appendixes A,B, and
C. California Dept. of Public Health, Sacramento.
1968.
4. Aerojet-General Corp., and Engineering-Science
Inc. A Systems Study of Solid Waste Management
in the Fresno Area. U.S. Public Health Service,
Washington, D.C. Publication Number (SW-4d).
1969.
5. "Air Pollution Aspects of Waste Disposal Part 1 -
Land Burial". Journal of the Air Pollution
Control Association. Vol. 10, No. 4 (August 1960) :
297.
6. Air Pollution Control Association. Proceedings
of the Mid-Atlantic States Section Semi-Annual
Technical Conference on Odors. Pittsburgh. 1965.
7. Air Pollution Control Office. Air Pollution
Aspects of Emission Sources: Municipal Incineration
A Bibliography with Abstracts. Environmental
Protection Agency, Washington, D.C. 1971.
8. Amoore, J.E. Chemistry and Physiology of Flavors.
Westport, Avi, 1965. Chapter 5.
9. Amoore, J.E. Molecular Basis of Odor. Spring-
field, Thomas, 1970.
10. Andersen, J.R., and J.N. Dornbush. "Influence
of Sanitary Landfill on Ground Water Quality".
Journal of the American Water Works Association.
Vol. 59, No. 4 (April 1967): 457.
280
-------
11. Beranek, L.L. "Criteria for Noise and Vibration
in Communities, Buildings, and Vehicles". In:
Noise and Vibration Control. Beranek, L.L. (ed.).
New York, McGraw-Hill, 1971.
12. Bishop, W.D., R.C. Carter, and H.F. Ludwig.
"Water Pollution Hazards from Refuse Produced
Carbon Dioxide". Journal of the Water Pollution
Control Federation. Vol. 38, No.3 (March 1966):328.
13. Bjornson, B.F., H.D. Pratt, and K.S. Littig.
Control of Domestic Rats and Mice. U.S. Public
Health Service, Washington, D.C. Publication
Number 583. 1970.
14. Black, R.J. "Storage in Paper Sacks Avoids Rats
and Flies". Refuse Removal Journal. Vol. 7, No. 12
(December 1964): 16.
15. Black, R.J., and A.M. Barnes. "Effect of Earth
Cover on Fly Emergence from Sanitary Landfills".
Public Works. Vol. 89, No. 2 (February 1958): 91.
16. Black, R.J., A.J. Muhich, A.J. Klee, H.L. Hickman,
Jr., and R.D. Vaughan. The National Solid Wastes
Survey - An Interim Report. U.S. Public Health
Service, Washington, D.C. 1968.
17. Black, R.J., and L.B. Near. "Controlled Burning
on Refuse Dumps". California Vector Views.
Vol. 6, No. 9 (November 1959): 55.
18. Boettcher, R.A. Air Classification of Solid
Wastes. Environmental Protection Agency, Wash-
ington, D.C. Publication Number (SW-30c). 1972.
19. Bolt, Beranek, and Newman. Noise From Construction
Equipment and Operation, Building Equipment, and
Home Appliances. Environmental Protection Agency,
Washington, D.C. 1971.
20. Booth, E.J. "Buried 25 Years and Still Legible".
American City. Vol. 80, No. 7 (July 1965): 26.
21. Boubel,R.W. Wood Residue Incineration in Tepee
Burners. Corvallis, Qregon State University.1965.
281
-------
22. Bragdon, C.R. Noise Pollution. Philadelphia,
University of Pennsylvania Press. 1970.
23. Breidenbach, A.W. Composting of Municipal Solid
Wastes in the United States. Environmental Pro-
tection Agency, Washington, D.C. 1971.
24. Brown, P., W. Wong, and I. Jelenfy. "A Survey
of the Fly Production from Household Refuse
Containers in the City of Salinas, California.
California Vector Views. Vol. 17, No, 4 (April 1970):
19.
25. Brunner, D.R., S.J. Hubbard, D.J. Keller, and
J.L. Newton. Closing Open Dumps. Environmental
Protection Agency, Washington, D.C. Publication
Number (SW-61ts). 1971.
26. Brunner,D.R., and D.J. Keller. Sanitary Landfill
Design and Operation. Environmental Protection
Agency, Washington, D.C. Publication Number (SW-
65ts). 1972.
27. Bumb, R.L. "The Use of Electrostatic Precipitators
for Incinerator Gas Cleaning in Europe". In: 1966
National Incinerator Conference. New York, American
Society of Mechanical Engineers, Incinerator Division.
1966.
28. Burch, L.A. "Solid Waste Disposal and its Effects
on Water Quality". California Vector Views. Vol. 16,
No. 11 (November 1969): 99.
29. Burchinal, J.C. Microbiology and Acid Production
in Sanitary Landfills - A Summary Report. U.S.
Public Health Service, Washington, D.C. 1970.
30. Burchinal, J.C., and H.A. Wilson. Sanitary Land-
fill Investigation - Final Report. U.S. Public
Health Service, Washington, D.C. (Project EF-
00040-01; 02; and 03). 1966.
31. California State Department of Public Health.
California Solid Waste Management Studv (1968)
and Plan (1970). Environmental Protection Agency,
Washington, D.C. Publication Number (SE-2tsg). 1971.
282
-------
32. California State Department of Water Resources.
Sanitary Landfill Studies, Appendix A: Summary of
Selected Previous Investigations. (Bulletin No.
147-5). Sacramento. 1969.
33, California State Water Pollution Control Board.
Report of the Investigation of Leaching of Ash
Dumps. Sacramento. (Publication Number 2). 1952.
34. California State Water Pollution Control Board.
Report on the Investigation of Leaching of a
Sanitary Landfill. Sacramento. Publication
Number 10. 1954.
35. California State Water Resources Control Board.
Special Study - Solid Waste Disposal. Task VII -
1C, San Francisco Bay - Delta Water Quality Control
Program. Sacramento. 1968.
36. California Water Quality Control Board. Water
Quality Criteria.- Sacramento. Publication Number
3>-A. 1963.
37. Campbell, E.r and R.J. Black. "The Problem of
Migration of Mature Fly Larvai from Refuse Containers
and its Implications on the Frequency of Refuse
Collection". California Vector Views. Vol. 7,
No. 2 (February 1960): 9.
38. Cartwright, K., and R. McComas. "Geophysical
Surveys in the Vicinity of Sanitary Landfills in
Northeastern Illinois". Ground Water. Vol. 6,
No. 5 (September-October 1968): 23.
39. Cartwright, K., and F.B. Sherman. "Evaluating
Sanitary Landfill Sites in Illinois". Illinois
State Geological Survey Environmental Geology Notes.
No. 27. 1969.
40. Chalupnik, J.D. Transportation Noises. Seattle,
University of Washington Press.1970.
41. Cohen, A., J. Anticaglia, and H.H. Jones. "Socio-
cusis - Hearing Loss from Non-occupational Noise
Exposure". Sound and Vibration. Vol. 4, No. 11
(November 1970) : 12.
283
-------
42. Combustion Engineering, Inc. Technical-Economic
Study of Solid Waste Disposal Needs and Practices.
U.S. Public Health Service, Washington, D.C.
Publication Number (SW-7c). 1969.
43. Committee on Sanitary Landfill Practice of the
Sanitary Engineering Division. Sanitary Landfill,
ASCE - Manual of Engineering Practice No. 39.
New York, American Society of Civil Engineers. 1959.
44. Committee of Solid Wastes, American Public Works
Association. Refuse Collection Practice. Chicago,
Public Administration Service. 1966.
45. Corey, R.C. (ed.). Principles and Practices of
Incineration. New York, Wiley-Interscience.T969.
46. Court, J. Section 205 - Resource Recovery
Report. Environmental Protection Agency, Washington,
D.C. 1972.
47. Cummins, R.L. Effects of Land Disposal of Solid
Wastes on Water Quality. U.S. Public Health Service,
Washington, D.C. Publication Number (SW-2ts). 1968.
48. Darnay,A. and W.E. Franklin. The Role of Packaging
in Solid Waste Management 1966 to 1976. U.S.
Public Health Service, Washington, D.C. Publication
Number (SW-5c). 1969.
49. Delaney, J.E. Satellite Vehicle Waste Collection
Systems. Environmental Protection Agency, Wash-
ington, D.C. Publication Number (SW-82ts. 1). 1972.
50. DeMarco,J., D.J. Keller, J. Leckman, and J.L.
Newton. Incinerator Guidelines - 1969, Municipal
Incineration - A State of the Art Report. U.S.
Public Health Service, Washington, D.C. 1969.
51. District of Columbia. District of Columbia Solid
Waste Management Plan Status REport 1970. Environ-
mental Protection Agency, Washington, D.C.
Publication Number (SW-4tsg). 1971
284
-------
52, Division of Air Quality and Emission Data,
Bureau of Criteria and Standards, National Air
Pollution Control Administration. Nationwide
Inventory of Air Pollutants Emissions. U.S.
Public Health Service, Washington, D.C. 1968.
53. Doty, R.E., and A,M. Barnes. "Effects of Earth
Cover on Emergence of the Vinegar Fly, Drosoohila
melanogaster Meigen". California Vector Views.
Vol, 8, No. 2 (February 1961): 7.
54. Dravnieks,A. Flavor Chemistry. Washington, D.C.,
American Chemical Society. 19~66. Chapter 2.
55. Drobny, N.L., H.E. Hull, and R.F. Testin.
Recovery and Utilization of Municipal Solid Waste.
EPA, Washington, D.C. Publication Number (SW-lOc).
1971.
56. Duckworth,S. "The Meteorologically Schedules Open
Burn". Journal of the Air Pollution Control
Association. Vol. 15, No. 6 (June 1965): 274.
57. "Dump Smoke Causes Asthmatic Outbreak". Refuse
Removal Journal. Vol. 8, No. 1 (January 1965): 28.
58. Dunn, W.L. "Refuse Filling of a Swamo in a Fresh
Water Lake". Civil Engineering. Vol. 38, No.l
(January 1968) : 60.
59. Dunn, W.L. "Storm Drainage and Gas Burning at a
Refuse Disposal Site". Civil Engineering. Vol. 30,
No. 9 (August 1960): 68.
60. Duprey, R.L. Compilation of Air Pollutant Emission
Factors. U.S. Public Health Service, Washington,
D.C. 1968.
61. Ecke,D.H., and D.D. Linsdale. "Fly and Economic
Evaluation of Urban Systems Part 1: Control of
Green Blow Flies (Phaenicia) by Improved Methods
of Residential Refuse Storage and Collection".
California Vector Views. Vol. 14, No. 4 (April
1967): 19.
285
-------
62. Ecke, D.H.r D.D. Linsdale, and K.E. White.
"Migration of Green Blow Ply Larvae from Six
Refuse Container Systems". California Vector
Views. Vol. 12, No. 8 {August 1965): 35.
63. Erarich, G.H. "Generation of Leachate from Land-
fills and its Subsurface Movement". Lecture
Delivered at The Annual Northeastern Regional
Anti-pollution Conference. University of Rhode
Island. July 1969.
64. Engdahl, R.B. Solid Waste Processing. U.S. Public
Health Service, Washington, D.C. Publication
Number (SW-4c). 1969.
65. Engineering Science, Inc. Effects of Refuse Dumps
on Ground Water Quality. California State Water
Pollution Control Board, Sacramento. Publication
Number 24. 1961.
66. Engineering Science, Inc. In Situ Investigation
of Movements of Gases Produced from Decomposing
Refuse. California State Water Quality Control
Board, Sacramento. Publication Number 31. 1965.
67. Engineering Science, Inc. In Situ Investigation
of Movements of Gases Produced from Decomposing
Refuse - Final Report. California State Water
Quality Control Board, Sacramento, Publication
Number 35. 1965.
68. Feldstein, M., S. Duckworth, H.C. Wohlers, and
B. Linsky. "The Contribution of the Open Burning
of Land Clearing Debris to Air Pollution".
Journal of the Air Pollution Control Association.
Vol. 13, No. 11 (November 1963): 542.
69. Fernandes, J.H. "Incinerator Air Pollution
Control". In: the 1968 National Incinerator
Conference. New York, American Society of
Mechanical Engineers Incinerator Division. 1968.
70. First, M.W., F.J. Viles, and S. Levin. "Control
of Toxic and Explosive Hazards in Buildings Erected
on Landfills". Public Health REports. Vol. 81,
No. 5 (May 1966): 419.
286
-------
71. Flawn, P.R., L.J. Turk, and C.H. Leach.
Geological Considerations in Disposal of Solid
Municipal Wastes in Texas.Austin, Bureau of
Economic Geology, University of Texas. 1970.
(Geological Circular 70-2).
72. Fleming, J.R. "Environmental Management - A
Twentieth Century Public Health Challange".
Canadian Journal of Public Health. Vol. 56, No. 1
(January 1965): 1.
73. Flood, L.P. "Air Pollution form Incinerators -
Causes and Cures". Civil Engineering. Vol. 35,
No. 1 (December 1965): 44.
74. Franklin, W.E., and A. Darnay. The Role of Non-
packaging Paper in Solid Waste Management, 1966
to 1976. EPAf Washington, D.C. Publication Number
(Sw-26c). 1971.
75. Friedman, L. and J.G. Miller. "Odor Incongruity
and Chiralty". Science. Vol. 172 (June 1971): 1044,
76. Fungaroli, A.A. Pollution of Subsurface Water by
Sanitary Landfills, Volume 1. EPA, Washington, D.C.
1971.
77. Furness,J.F. "Disposal of Household Refuse in
Wet Gravel Pits". Public Cleansing. Vol. 57,
No. 5 (May 1967) : 255.
78. Gerstle, R.W., and D.A. Kemnitz, "Atmospheric
Emissions from Open Burning". Journal of the Air
Pollution Control Association. Vol. 17, No. 5
(May 1967): 324.
79. Glorig, A. "Non-Auditory Effects of Noise Exposure".
Sound and Vibration. Vol. 5, No. 5 (May 1971): 28.
80. Gotaas, J.D. Composting: Sanitary Disposal and
Reclamation of Organic Wastes. (W.H.O. Monograph
No. 31). 1956.
81. Grim, R.E. Applied Clay Mineralogy. New York,
McGraw-Hill. 1962.
287
-------
82. Grim, R.E. Clay Mineralogy. New York, McGraw-
Hill. 1953.
83. Gutfreund, K. Feasibility Study of the Disposal
of Polyethylene Plastic Waste. EPA, Washington,
D.C. 1971.
84. Hagenbrauck, R.P., D.J. von Lehmden, and J.E.
Meeker. "Emission of Polynuclear Hydrocarbons
and Other Pollutants from Heat-Generation and
Incineration Processes". Journal of the Air
Pollution Control Association. Vol. 14, No. 7
(July 1964): 267.
85. Hanks, T.G. Solid Waste/Disease Relationships -
A Literature Survey. U.S. Public Health Service,
Washington, D.C. 1967.
86. Hart, S.A. Solid Waste Management in Germany;
Report of the U.S. Study Team Visit, June 25-
July 8. Government Printing Office, Washington,
D.C. Public Health Service Publication Number
1812. 1967.
87. Hemeon, W.C.L. "Malodors - A Basis for Regulations".
Journal of the Air Pollution Control Association.
Vol. 21, No. 12 (December 1971): 770.
88. Hopkins, G.J., and J.R. Popalisky. "Influence of
an Industrial Waste Landfill Operation on a
Public Water Supply". Journal of Water Pollution
Control Federation. Vol. 42, No. 3 (March 1970): 431,
89. Horstman, S.W., R.F. Wromble, and A.N. Heller.
"Identification of Community Odor Problems by Use
of an Observer Corps". Journal of the Air
Pollution Control Association. Vol. 15, No. 6
(June 1965): 261.
90. "How to Use Your Completed Landfills". American
City. Vol. 80, No. 8 (August 1965): 91.
91. Hughes, G.M., R.A. Landon, and R.N. Farvolden.
Hydrogeology of Solid Waste Disposal Sites in
Northeastern Illinois. EPA, Washington, D.C. 1971.
288
-------
92. Jens/ W. , and F.R. Rehm. "Municipal Incineration
and Air Pollution Control". In: 1966 National
Incinerator Conference. New York, American Society
of Mechanical Engineers, Incinerator Division. 1966.
93. Jensen,M.E. Observations of Continental European
Solid Waste Management Practices. U.S. Public
Health Service, Washington, D.C. Publication
Number 1880. 1969.
94. Johnson, H.C., and H.A. James. "Controlled Open
Burning in the San Francisco Bay Area". Journal
of the Air Pollution Control Association. Vol. 20,
No. 8 (August 1970): 530.
95. Kaiser, E.R. "Chemical Analyses of Refuse
Components". In: 1966 National Incinerator
Conference. New York, American Society of
Mechanical Engineers, Incinerator Division. 1966.
96. Kaiser, E.R. "Refuse Composition and Flue-Gas
Analyses From Municipal Incinerators". In: 1964
National Incinerator Conference. New York, American
Society of Mechanical Engineers, Incinerator
Division. 1964.
97. Kaiser, E.R. "The Sulfur Balance of Incinerators".
Journal of the Air Pollution Control Association.
Vol. 18, No. 3 (March 1968): 171.
98. Kaiser, E.R., C.D. Zeit, and J.B. McCaffrey.
"Municipal Incinerator Refuse and Residue". In:
1968 National Incinerator Conference. New York,
American Society of Mechanical Engineers, Incin-
erator Division. 1968.
99. Kaufmann, R.F. Hydrogeological Aspects of Solid
Waste Disposal. Madison, University of Wisconsin
Extension, Solid Waste Management. 1969.
100. Kentucky State Department of Health. Kentucky
Solid Waste Management Plan Status Report 1970.
EPA, Washington D.C. Publication Number (SW-6tsg)
1971.
289
-------
101, Klein, S. "New Building Constructed on Sanitary
Landfill". Public Works. Vol. 99, No. 10
(October 1968): 125.
102. Koch, A.S. The Orange County Refuse Disposal
Program. Santa Ana, The Orange Country Road
Department. 1968.
103. Kreichelt, T.E. Air Pollution Aspects of Tepee
Burners Used for Disposal of Municipal Refuse.
Public Health Service, Washington, b.C. 1966/
104. Kurker, C. "Reducing Emissions from Refuse
Disposal". Journal of the Air Pollution Control
Association. Vol. 19, No. 2 (February 1969): 69.
105. Kurze, U. and L.L. Beranek (ed.). "Sound
Propagation Outdoors". In: Noise and Vibration
Control. L.L. Beranek (ed.). Hew York, McGraw-
Hill.1971.
106. Landon, R.A. "Application of Hydrogeology to the
Selection of Refuse Disoosal Sites". Ground Water,
Vol. 7, No. 6 (November-December 1969) : 9.
107. LeGrand, H.E. "Management Aspects of Ground Water
Contamination". Journal of the Water Pollution
Control Federation. Vol. 36, No. 9 (September
1964): 1133.
108. LeGrand, H.E. "Monitoring of Changes in Quality
of Ground Water". Ground Water. Vol. 6, No. 3
(May-June 1968): 14.
109. LeGrand, H.E. "Patterns of Contaminated Zones of
Water in the Ground". Water Resources Research.
Vol. 1, No. 1 (First Quarter 1965): 83.
110. LeGrand, H.E. "System for Evaluation of Con-
tamination Potential of Some Waste Disposal Sites"-
Journal of the American Water Works Association.
Vol. 56, No. 8 (August 1964): 959.
111. Lin, Y.H. "Acid and Gas Production from Sanitary
Landfills". Ph.D. dissertation, West Virginia
University, Morgantown. 1966
290
-------
112. Los Angeles County, Department of County
Engineer, and Engineering-Science, Inc. Develop-
ment of Construction and Use Criteria for Sani-
tary Landfills - An Interim Report. U.S. Public
Health Service, Washington, D.C. 1969.
113. Luce, R. "Sanitary Landfill Pushes Back the River".
American City. Vol. 74, No. 12 (December 1959): 92.
114. Magy, H.I., and R.J. Black. "An Evaluation of the
Migration of Fly Larvae from Garbage Can in
Pasadena, California". California Vector Views.
Vol. 9, No. 11 (November 1962): 55.
115. Marynowski, C.W. Disposal of Polymer Solid Wastes
by Primary Polymer Producers and Plastics Fabri-
cators. U.S. EPA, Washington, D.C. 1972.
116. Matusky, F.E., and R.K. Hampton. "Incinerator
Waste Water". In: 1968 National incinerator
Conference. New York, American Society of
Mechanical Engineers, Incinerator Division. 1968.
117. McCartney, W. Olfaction and Odours. Berlin,
Springer-Verlag"1968,
118. McCormick, J.H. "Chemical Variation of Ground Water
Quality in the Vicinity of a Sanitary Landfill".
M.S. Thesis, South Dakota State University, Brookings,
1966.
119. McDavid, C.G., and E.W. Mood. "Biological Aspects
of Urban Rat Control". HSMHA Health Reports.
Vol. 87, No. 1 (January 1972): 17.
120. McFarland, J.M.f C.R. Glassey, P.H. McGauhey,
D.L. Brink, S.A. Klein, and C.G. Golueke.
Comprehensive Studies of Solid Waste Management -
Final Report. (SERL Report No. 72-3) Berkeley,
Sanitary Engineering Research Laboratory; College
of Engineering and School of Public Health;.
University of California, Berkeley. 1972.
121. Merz, R.C. "Landfill Settlement Rates". Public
Works. Vol. 93, No. 9 (September 1962): 103.
291
-------
122. Merz, R.C., and R. Stone. Factors Controlling
Utilization of a Sanitary Landfill Site. Los
Angeles, University of Southern California. 1963.
123. Merz, R.C., and R. Stone. "Gas Production in a
Sanitary Landfill". Public Works. Vol. 95, No. 2
(February 1964): 84.
124. Merz, R.C., and R. Stone. "Sanitary Landfill
Behavior in an Aerobic Environment". Public Works,
Vol. 97, No. 1 (January 1966): 67.
125. Merz, R.C., and R. Stone. Special Studies of a
Sanitary Landfill. U.S. Public Health Service,
Washington, P.C. 1970.
126. Mission 5000. U.S. EPA, Washington, D.C. 1972.
127. Moncrieff, R.W. The Chemical Senses. Cleveland,
Chemical. Rubber. 1967.
128. Mortensen, E.W. "Description of a Two-Compartment
Trench System of Refuse Disposal". California
Vector Views. Vol. 9, No. 4 (April 1962): 21.
129. Muhich, A.J., A.J. Klee, and P.W. Britton.
Preliminary Data Analysis, 1968 National Survey
of Community Solid Waste Practices. U.S. Public
Health Service, Washingtonr D.C. 1968.
130. Muhich, A.J., A.J. Klee, and C.R. Hampel. 1968
National Survey of Community Solid Waste Practices:
Region 1. U.S. Public Health Service, Washington,
D.C. 1969.
131. Muhich, A. J., A.J. Klee, and C.R. Hampel. 1968
National Survey of Community Solid Waste Practices
Region 2: Volume 1. U.S. Public Health Service,
Washington, D.C. 1969.
132. Muhich, A.J., A.J. Klee, and C.R. Hampel. 1968
National Survey of Community Solid Waste Practices
Region 2: Volume 2. U.S. Public Health Service,
Washington, D.C. 1969.
292
-------
133. Municipal Refuse Disposal. Chicago, American
Public Works. 1966.
134. Murray, C.R. Estimated Use of Water in the United
States, 1965. U.S. Geological Survey, Washington,
D.C. Circular Number 556. 1968.
135. National Association of Counties Research
Foundation. Guidelines for Local Governments
on Solid Waste Management. U.S. EPA, Washington,
D.C. Publication Number (SW-17c). 1971.
136. Niessen, W.R., and S.H. Chansky. "The Nature of
Refuse". In: 1970 National Incinerator Conference.
New York, American Society of Mechanical Engineers,
Incinerator Division. 1970.
137. Niessen, W.R., and A.F. Sarofim. "The Emission
and Control of Air Pollutants from the Incineration
of Municipal Solid Waste". (Paper EN-5c)r presented
at The Second International Clean Air Congress,
Washington, D.C. International Union of Air
Pollution Prevention Associations. 1970.
138. Niessen, W.R., and A.F. Sarofim. "Incinerator
Air Pollution". In: 1970 National Incinerator
Conference. New York, American Society of
Mechanical Engineers, Incinerator Division. 1970.
139. Oregon State Board of Health. Oregon Solid Waste
Management Plan Status Report 1969. U.S. EPA,
Washington, D.C. Publication Number (SW-ltsg). 1971
140. Page, L.M. "Use of the Electrical Resistivity
Method for Investigating Geologic and Hydrologic
Conditions in Santa Clara County, California".
Ground Water. Vol. 6, No. 5 (September-October): 31.
141. "Paper Sacks Cut Fly Larvae Production". American
City. Vol. 80, No. 11 (November 1965): 44.
142. Parizek, R. The Chemical Interaction of Sanitary
Land_f_ill Leachate with Unsaturated Soil in a
Carbonate Rock Terrain"! University Park, Pennsylvania
State University.1971.
293
-------
143. Pennsylvania Department of Health. A Plan for
Solid Waste Management in Pennsylvania. (Solid
Waste Publication 3).Harrisburn, State of
Pennsylvania. 1970.
144. "Per Capita Figure of 5.1 Pounds of Matter Daily
Indicated by National Survey". Solid Waste
Management/ Refuse Removal Journal. Vol. 11,
No. 3 (March 1968): 22.
145. Perkins, R.A. Satellite Vehicle Systems for
Solid Waste Collection, Evaluation and Application.
U.S. EPA, Washington, D.C. Publication Number
(SW-82ts). 1971.
146. Peterson, M.L., and F.J. Stutzenberger.
'Microbiological Evaluation of Incineration
Operations". Applied Microbiology. Vol. 18,
No. 1 (July 1969): 8.
147. Pohland, F.G., and S. Ghosh. "Developments in
Anaerobic Stabilization of Organic Wastes -
The Two-Phase Concept". Environmental Letters.
Vol. 1, No. 4 (April 1971): 255.
148. Powell, T. "Air Pollution and Incineration".
Compost Science. Vol. 1, No. 3 (Autumn 1960): 35.
149. "Pyrolytic Decomposition of Solid Wastes". Public
Works. Vol. 99, No.8 (August 1968): 82.
150. Qasim, S.R. "Chemical Characteristics of Seepage
Water from Simulated Landfills". Ph.D. dissertation.
Morgantown, West Virgina University. 1965.
151. Qasim, S.R., and J.C. Burchinal. "Leaching of
Pollutants from Refuse Beds". Journal of the
Sanitary Engineering Division, Proceedings of the
American Society of Civil Engineers. Vol. 96,
No. SA1 (February 1970): 49.
152. Radcliffe, J.C. et al.."Guidelines for Noise
Exposure Control". Sound and Vibration. Vol. 4,
No.11 (November 1970): 21.
294
-------
153. Refuse-Collection and Disposal - An Annotated
Bibliography, Supplements B,C,D,E, and F. U.S.
Department of HEW, Cincinnati. 1966.
154* Remson, I., A.A. Fungaroli, and A.W. Lawrence.
"Water Movement in an Unsaturated Sanitary
Landfill". Journal of the Sanitary Engineering
Division, Proceedings of the American Society of
Civil Engineers. Vol. 94, No. SA2 (April 1968): 307.
155. Rickles, R. Pollution Control. Pearl River, Noyes
Development Corporation. 1965.
156. Roberts, P., and W.E. Parker. "How to Build a
Sanitary Landfill in Sv/amolands". American City.
Vol. 69, No.5 (May 1954): 112.
157. Roderick, G.F. "Current Ideas on the Chemical
Basis of Olfaction". Journal of Chemical
Education. Vol. 43, No. 10 (October 1966): 510.
158. Rogers, P.A. "San Francisco Bay Area Solid
Waste Management and Planning Problems". California
Vector Views. Vol. 12, No. 10 (October 1965): 51.
159. Rogers, P.A., D.R. Andres, J. Cornelius, L.A. Burch,
S.E. Richardson, Jr., et al. California Solid
Waste Management Study (1968) and Plan (1970).
U.S. EPA, Washington, D.C. 1971.
160. Rogers, P.A., D.R. Andres, J. Cornelius, L.A. Burch,
and S.E. Richardson, Jr. "The Status of Solid
Waste Management in California". California Vector
Views. Vol. 15, No. 12 (December 1968): 123.
161. Rogers, P.A., and G.J. Bellenger. "Fly and
Economic -^Evaluation of Urban Refuse Systems Part
II - An Efficiency Analysis of Paper Bag Containers".
California Vector Views. Vol. 14, No. 5 (May 1967):
29.
162. Rose, A.H. "Air Pollution Effects of Incinerator
Firing Practices and Combustion Air Distribution".
Journal of the Air Pollution Control Association.
Vol. 8, No. 4 (February 1959): 297.
295
-------
163. Russell, G.F., and J.I. Hills. "Odor Differences
Between Enantiomeric Isomers". Science. Vol. 172
(June 1971) : 1043.
164. SaccafG. "Comparative Bionomics in the Genus
Musca". Annual Review of Entomology. Vol. 9.
1964: 341.
165. San Diego, City of; Department of Public Works.
The Feasibility of Baling Municipal Refuse.
Revised 1968.
166. Sawinski, R.J. "Ground Water Quality Variation
at a Sanitary Landfill". M.S. theses. Brookings,
South Dakota State University. 1966.
167. Schoenberger, R.J., and P.W. Purdom. "Classifi-
cation of Incinerator Residue". 1968 National
Incinerator Conference. New York, American
Society of Mechanical Engineers, Incinerator Division.
1968.
168. Secretary of Health, Education and Welfare.
Progress in the Prevention and Control of Air
Pollution. (U.S. Senate Document 92). Washington,
D.C. U.S. Government Printing Office. 1968.
169. "70,000 Flies per Cubic Foot of Garbage".
American City. Vol. 80, No. 2 (February 1965)S 36.
170. Sittig, M. Water Pollution Control and Solid
Wastes Disposal.Park Ridge, Noyes Development
Corporation.T969.
171. Smith, D.C., and R.P. Brown. Ocean Disposal
of Barge Delivered Licruid and Solid Wastes from
U.S. Coastal Cities. U.S. EPA, Washinaton, D.C.
Publication Number (SW-19c). 1971.
172. Smith, R.A., L.L. Wikstrom, and A.A. Carotti.
Air Borne Emissions from Municipal Incinerators.
Summary Report of State of Knowledge, Recommendations
for Further Studies and, an Annotated Bibliography.
University Heights, "New York University. 1967.
296
-------
173. Sorg, T.J., and H.L. Hickmany/ Jr. Sanitary
Landfill Pacts. U.S. Public Health Service,
Washington, D.C. 1970.
174. Sowers, G.F. "Foundation Problems in Sanitary
Landfills". Journal of the Sanitary Engineering
Division, Proceedings of the American Society of
Civil Engineers. Vol. 94, No. SA1 (February 1968):
103.
175. Stabenow, G. "Performance and Design Data for
Large European Refuse Incinerators with Heat
Recovery. American Society of Mechanical Engineers,
Incinerator Division, New York. 1969.
176. Stear, J.R. Municipal Incineration: A Review of
the Literature. U.S. EPA, Washington, D.C.
Publication Number (AP-79). 1971.
177. Steering Committee on Coastal Waste Management.
Wastes Management Concepts for the Coastal Zone.
National Academy of Sciences; National Academy
of Engineering, Washington, D.C. 1970.
178. Steiner, R.L., and R. Kantz. Sanitary Landfill:
A Bibliography. Consumer Protection and
Environmental Health Service, Cincinnati.
Publication Number 1819). 1968.
179. Stenburg, R.L. et al. "Field Evaluation of
Combustion Air Effects on Atmospheric Emissions
from Municipal Incinerators". Journal of the Air
Pollution Control Association. Vol. 12, No. 2
(February 1962): 93.
180. Stern, A.C. (ed.). Air Pollution, Volxaae Ii Air
Pollution and Its Effects.New York, Academic
Press. 1968.
181* Stern, A.C. (ed.). Air Pollution, Volume II:
Analysis, nonitoi-ing, and Surveying. New York,
Academic Press. 1968.
182. Stern, A.C. (ed.). Air Pollution, Volume III;
Sources of Air Pollution and linear control. New
York, Academic Press.1968.
297
-------
183. Ralph Stone and Company, Inc. Solid Waste
Landfill Stabilization, An Interim Report.
Consumer Protection and Environmental Health
Service, Cincinnati. 1968.
184. Ralph Stone and Company, Inc. A Study of Solid
Waste Collection Systems Comparing One-Man
with Multi-Man Crews. U.S. Public Health Service,
Washington, D.C. 1969.
185. Stone, R., and H. Friedland. "National Survey
of Sanitary Landfill Practices". Public Works.
Vol. 100, No. 8 (August 1969): 88.
186. Stone, R, and R.K. Gupta. "Aerobic and Anaerobic
Landfill Stabilization Process". Journal of the
Sanitary Engineering Division, Proceedings of
the American Society of Civil Engineers. Vol. 96,
No. SA 6 (December 1970): 1399.
187. Albert Switzer & Associates, Inc., and Greenleaf/
Telesca. Master Plan for Solid Waste Collection
and Disposal, Tri-Parish Metropolitan Area of
New Orleans. U.S. Public Health Service,
Washington, D.C. 1969.
188. Taft Sanitary Engineering Center. Proceedings
of 1961 Symposium - Ground Water Contamination.
Cincinnati.1961.
189. Taft Sanitary Engineering Center. Soil Waste
Management. U.S. Public Health Service, Cincinnati,
1966.
190. Theimer, E.T., and M.R. McDaniel. "Odor and
Optical Activity". Journal of the Society of
Cosmetic Chemistry. Vol. 22 (1971): 15.
191. Turner, C.H., E.E. Ebert, and R.R. Given. Man-
Made Reef Ecology. California State Department
of Fish and Game, Sacramento. Fish Bulletin 146.
1968.
298
-------
192. Walker, A.B., and F.W. Schmitz. "Characteristics
of Furnace Emissions from Large, Mechanically-
Stoked Municipal Incinerators". In: 1966 National
Incinerator Conference. American Society of
Mechanical Engineers, Incinerator Division, New
York. 1966.
193. Walker, W.H. "Illinois Ground Water Pollution".
Journal of the American Water Works Association.
Vol. 61, No. 1. (January 1969): 31.
194. Warner, A.J., C.H. Parker, and B. Baum. Plastics
Solid Waste Disposal by Incineration or Landfill.
Manufacturing Chemists Association, Washington, D.C,
1971.
195. Warner, D.L. "Preliminary Field' Studies Using
Earth Resistivity Measurements for Delineating
Zones of Contaminated Ground Water". Ground
Water. Vol. 7, No.l (January-February 1969): 9.
196. Weber,E. "Staub-und Abgasanfall bei einer
Hausmuellverbrennungsanlage". Staub. Vol. 24,
No. 6 (June 1964): 210.
197. Williams, J.H. "Can Ground Water Pollution Be
Avoided?" Ground Water. Vol. 7, No. 2 (March-
April 1969) : 21.
198. Wolf, K.W., and C.H. Sosnovsky, et al. High-
Pressure Compaction and Baling of Solid Waste.
U.S. EPA, Washington, D.C. 1972.
199. Wright, R.H. The Science of Smell. New York,
Basic Books. 1964.
200. Wulfinghoff, P.E. (translator). Disposal of
Process Wastes. New York, Chemical Publishing
Co., Inc. 1968.
201. Yocom, J.E. "A Study of the Effluents from
Backyard Incinerators". Journal of the Air
Pollution Control Association. Vol. 6, No. 2
(August 1956): 84.
299
-------
202. Zanoni, A.E. "Ground Water Pollution and
Sanitary Landfills - A Critical Review". Ground
Water. Vol. 10, No. 1 {January-February 1972): 3,
300
-------
CHAPTER V
PROBLEMS IN MEASURING THE EXTERNALITIES
OF SOLID WASTE DISPOSAL
The physical manifestations of external effects
related to solid waste disposal were surveyed in
Chapter IV. The theoretical considerations relevant
to measuring the social cost of these external effects
and a survey of empirical attempts to measure them
were discussed in previous chapters. In this chapter
we consider some of "the difficulties and practical
problems of measurement within the context of solid
waste, although many of these problems have general
relevance.
As noted in Chapter III, there are two general
methods of evaluating the social costs of external
effects of some activity. The first will be termed
the direct method, which actually consists of two
stages. The direct method consists of, first, measur-
ing directly the physical manifestations of the activity
and, second, evaluating the social cost of these
physical manifestations which is not already reflected
in market activities. In measuring the external
social costs of a sanitary landfill, for instance, the
301
-------
direct method would consist of measuring, in physical
terms, the air pollution and groundwater pollution,
if any, the noise generated, the added congestion and
highway noise associated with the transportation
system, and the visual effects of the landfill. (We
discuss below the obvious difficulties in "physically"
measuring such effects). With respect to an open
dump, or to any landfill which does not meet all of
the sanitary landfill standards, all of the effects
mentioned above would have to be measured, but in
addition one would have to physically measure the
possible surfacewater pollution, the breeding of
vermin and vectors, and the probability of fires and
their expected physical damage. The knowledge of physical
effects is not a sufficient basis on which to weigh
their social costs. In order to translate these
physical measurements into commensurate terms, the
value which people place on the absence or extent of
such physical effects must also be measured. Even
if the measurement of the physical manifestations
could be easily made, this translation into equivalent
dollar values presents extreme difficulties.
Both the problems of physical measurement and of
imputing values are finessed by the indirect method,
302
-------
in which the impact of the external effects of the
activity on economic variables is measured. By
measuring the impact of externalities on rents, or
equivalently on the market values of surrounding real
estate, the subjective valuation of the physical
external effects is implicitly measured. Thus the in-
direct method provides a powerful tool by which exten-
sive and difficult measurement and interpretation can
be replaced with the generally manageable measurement
of economic variables. However, there are two general
difficulties with the indirect approach. The first
is that the measurement of the effect of such exter-
nalities on land value requires that one has a model
of the determination of the value of the relevant land
and buildings, a model which is not just conceptual
but operational. Developing and estimating the
parameters of such a model is not a trivial task, as
can be seen in Chapter VI, where such models are
developed for "explaining" the value of land surround-
ing four sanitary landfills in Los Angeles County.
The second general difficulty with the indirect
approach is that the connection between land value
and the evaluation of the social costs of the external
effects is based on all the assumptions of the com-
303
-------
petitive model, some of which may be inappropriate.
The indirect method of measuring the external
costs of sanitary landfills is applied in Chapter VI.
The purposes are both to apply the method to this
particular example and to advance the state of the
art itself. In order to provide perspective on the
relative merits of the indirect method of evaluation
and on its validity, we discuss in this chapter the
problems of applying the direct method and the
limitations of the indirect method.
In applying the direct method of measurement,
it is easy to commit a crucial error which we shall
refer to as the "fallacy of ignoring adjustments".
Suppose, in general, that a particular activity is
imposing external costs on others. Those harmed by the
activity will generally take steps to reduce the costs
imposed upon them, and possibly to "eliminate" the
cost. The proper comparison is obviously with and
without the offending activity, and any costs borne by
others due to the activity are external social costs,
including the possible costs of their adjustments.
Suppose, for instance, that a factory is built in a
residential area which emits fumes of a particularly
obnoxious character. Suppose that the fumes extend
304
-------
for a mile in all directions and that, as a result
of the fumes, all residents choose to move elsewhere,
leaving the affected area vacant. Suppose that out-
side of the affected circular area the fumes are harmless
and disperse to insignificant levels. In such a case,
a superficial application of the direct method of
evaluation could lead to the conclusion that the fumes
from the factory impose no costs on society, since an
inspection of the area in which the fumes are physically
present show that it is uninhabited. This is obviously
fallacious; to an economist, the fallacy can be
described as ignoring the opportunity cost of the land.
With the factory, the land has, let us presume, no
uses at all. Without the factory, it is useful as
residential land. The fact that it cannot be used as
residential land is a social cost of the factory, even
if it is not a private cost. Obviously, if the factory
is there first, before the development of the adjoining
land, the restriction on the uses of the adjoining land
still represents an opportunity cost, although one
which would be less obvious than if vacant houses
surrounded the factory.
In the simplified example above, the fallacy of
inferring that no social cost exists because there is
305
-------
no one to smell the noxious fumes would be fairly
obvious. However, the same fallacy would exist in a
subtler form if, as a result of the factory, the
adjoining land were used for some activity, say agri-
culture, which was not affected by the fumes but which
would not be the best use of the adjoining land in
the absence of the factory. The difference in the
value of the land when used for agriculture as compared
to its use as residential land, in this example, would
still represent an opportunity cost and would represent
the appropriate measure of the external cost involved.
Similarly, if the adjoining land is nevertheless
used as residential land but the residents are forced
to air condition their homes, the cost of the air condi-
tioning represents, at least in part, a social cost
attributable to the factory activity. The complexity
of measuring such costs in a practical situation is
illustrated by this example, for it is likely that the
houses with air conditioning would be preferable to
the houses without air conditioning. Thus it would
not be appropriate to attribute the entire cost of
the air conditioning as an external cost of the factory
operation. The question of how much of the air
conditioning cost to allocate to the factory illustrates
306
-------
the beauty of the indirect approach to measuring
externalities, for the proper answer is that the net
cost should be attributed to the factory, where net
refers to the subtraction of the increase in house
value, attributable to air conditioning, from the cost
of the air conditioning. In order to clarify this point,
consider the following example. Suppose that the houses
surrounding the factory are identical and would be worth
the following amounts with and without the factory
and with and without air conditioning:
Without factory, without air conditioning: $20,000
Without factory, with air conditioning: $22,000
With factory, without air conditioning: $17,000
With factory, with air conditioning: $21,000
Suppose that air conditioning costs $3,000 to install,
purely for purposes of illustration. Then it is not in
the interests of owners to install air conditioning,
since it costs $3,000 and increases the value of the
house by only $2,000. Obviouslyfin a realistic sit-
uation, the residents might differ and some might choose,
nevertheless, to install air conditioning, but we assume
here that the housing market is a perfectly competitive
one and therefore that a resident would never be led
to make an improvement which is not "profitable".
307
-------
Such an assumption is clearly not met exactly in the
"real world"; we discuss the implications of that
discrepancy below in the discussion of the indirect
method of measurement.
With the factory in existence, air conditioning
increases the value of the surrounding houses by
$4,000; by the competitive assumption this implies
that the residents value the change in consumption
opportunities (fresher and cooler air) by $4,000.
Thus the existence of the factory makes air conditioning
preferable. However, it is clearly inappropriate to
charge the total cost of the air conditioning, $3,000,
as an external social cost against the factory; the
hbuse is worth more with the air conditioning and with
the factory than it would be without either. In par-
ticular, it is worth $1,000 more with both than it
was worth with neither, so it seems logical to attribute
the "net cost" of ($3,000 - $1,000) = $2,000 as the
external social cost imposed on the surrounding res-
idences by the factory.
This imputation is not only logical, but is
justified by the economics of the situation. In fact,
under the conditions assumed, the value of the houses,
without air conditioning, would be $18,000, and the
308
-------
decrease in the value of $2,000 reflects the social
external cost. That is to say, the "value" of the
house, without air conditioning, of $17,000, would
not be an equilibrium market value. If the factory
were to suddenly appear on the scene, the market value
of the surrounding houses would not drop to $17,000,
but only to $18,000, given the assumption that
"information is perfect", and in particular the
knowledge that the installation of air conditioning
Cat a cost of $3,000) would produce a house worth
$21,000. The drop in value of $2,000 thus represents
the fall in market value which results when "profitable"
adjustments are perceived, which cost less than they
add in value in the new situation.
The example shows very clearly the power of the
"indirect" method of computing the external costs of
some effect. If one simply looks at the attractiveness
of the houses in the surrounding area before and after
the coming of the factory, the astonishing conclusion
is derived that the existence of the factory increases
their attractiveness (their value, after all, has
increased from $20,000 to $21,000). The residents
are treated to cleaner and cooler air, and to the
greater control of their interior environments, than
309
-------
they had before. With the direct method, one would
ask: how much cleaner is the air than it was before?
and how much cooler is the air? Then one would ask:
what value do the residents put on the cooler, cleaner,
and more uniform air? An intelligent purveyor of the
direct method would recognize immediately that against
the benefits of such better interior air, one must
subtract the cost of the air conditioning (including
capital and amortized operating costs) in order to
arrive at net costs. Thus information must be gathered
on the cost of the air conditioning, and on operating
costs. Note that while the capital costs of the air
conditioning would be fairly directly observable, the
proper computation of operating costs is anything but
simple. Suppose for simplicity that the only operating
costs are charges for electrical power. Then operating
costs refer to the increase in the electrical bill due
to air conditioning. If electricity costs were
constant, one could then compute the average amount of
electrical power used by the air conditioners and
multiply by the rate, to arrive, at least, at the
private increase in operating costs. However, since
electrical power is a "decreasing cost industry",
utilities charge less for the marginal unit than for
310
-------
the average unit; thus the private operating costs
depend upon whether the air conditioners tend to be
used in periods of higher than average or lower than
average usage. In fact, to go beyond private operating
cost to social operating costs for electrical power,
which is not a competitive industry, would require
consideration of even more complex issues related
to whether the air conditioners were used in periods
of "peak load" utilization or not.
Nevertheless, reasonable approximations could
surely be computed for the costs of having and running
the air conditioners. In order to apply the direct
method, one would not want to attribute those costs
completely to the factory as external social costs
for the reasons explained above. The benefits of
the air conditioners would have to be quantified in
commensurate terms and subtracted from total costs to
arrive at the net costs of the adjustments. How
should one value the cleaner, cooler, and more uniform
air? "Direct" methods would probably include surveys
of the residents, asking them "How much is it worth
to you to have clean air rather than being forced to
smell the obnoxious factory fumes?" Such surveys could
also be used to attempt to find out how people "trade
311
-------
off various dimensions of an externality, such as
the frequency of the fumes versus their intensity.
Surveys of that nature, in which people are asked to
rank conjectural situations and to place "values"
on the existence of conjectured changes, are often
the basis for direct evaluations. See, for example,
various studies of noise problems related to airports
done by Bolt, Barenek, and Newman [1,2J.
Unfortunately, such surveys are notoriously un-
reliable predictors of how people will react when
placed in the actual situation conjectured. That
conclusion is particularly well supported by the
chronic difficulties of marketing specialists in
attempting to forecast demand for a new product by
asking people questions such as "Would you buy the
following product described as at a price
of $X?n Compared to the conjectured response which
people are asked to make in the type of survey in which
situations are ranked (e.g., "would you prefer to
experience one very loud plane a day or five fairly
loud planes a day?"), the marketing type questions
are concrete. If marketing analysts find that ex-
trapolating the proportions of people who say they
would buy an item is a very unreliable method of
312
-------
.forecasting demand, it seems plausible to suggest that
people's actual reactions to situations are likely to
be very different from their offhand responses to a
survey questionnaire would suggest.
It is also important to recognize that there may
be substantial incentives for people to exaggerate their
welfare loss or to otherwise disguise their true
feelings, depending on how they suspect the survey
results will be used. In Chapter II, this possibility
was discussed in the context of the problem of deter-
mining a Pareto optimal level of the production of a
public good. However, it obviously also arises in any
case in which those surveyed may modify their answers
where they expect the survey results to be used to
affect policies which concern them.
The power of the indirect method is obviously
that the subjective evaluation of how much the air
conditioning is worth, and of how the improved interior
air is weighed against the degraded exterior air, is
implicitly reflected in the value of the houses.
Furthermore, the decrease in the market value of the
houses (before air conditioning is installed), reflects
precisely how the residents—and potential residents—
weigh all of the advantages and disadvantages of the
313
-------
no-factory/no-air-conditioning situation against the
factory/air-conditioning situation.
The direct method can be applied in its simplest
form—where the external costs of an activity are
assumed to be reflected entirely in its direct physical
effects—only by assuming that those affected make no
adjustments to the external effects. If external
social costs are computed by observing the situation
resulting after adjustments are made, there will be
a systematic -tendency to underestimate the extent
of the external costs by ignoring the opportunity cost
of some resources and by ignoring the direct costs of
adjustment. As the factory example demonstrates/
however, the appropriate modification of the calculations
in the presence of adjustments may be very difficult
and may raise further problems of measurement and
evaluation. Clearly the moral of the example applies
generally; for example, similar problems arise if, in
the face of a sanitary landfill, surrounding houses
erect fences and plant extensive shrubbery in order
to replace a view of the landfill. If the fences and
shrubbery improve the value of the property over
what it would be in the absence of the landfill, the
analogous problem of allocating their costs arises.
314
-------
It must be noted that ignoring the effect of
adjustment will produce the opposite bias—a systematic
overestimate of the external costs—if the direct
method is applied in contemplation of the existence
of an activity with external effects. Thus in the
factory/air-conditioning example/ a survey of the
residents of the surrounding houses would produce the
conclusion, if it were done perfectly/ that the coming
of the factory would result in a decrease in utility
equivalent to $3,000. The fallacy there is obviously
that the possibility of air conditioning is ignored.
A more general adjustment which might easily be
ignored is the possibility that some residents would
move, to be replaced by residents who are less sensi-
tive to the external effect. Suppose, again purely for
purposes of illustration, that residents and potential
residents of the area surrounding the factory are of
two types: those for whom the disutility from the
factory fumes is equivalent to $3,000 and those for
whom the disutility is equivalent to $500. Then
assuming there are large numbers in both groups, the
efficient distribution of homeowners clearly involves
the area surrounding the factory being inhabited
entirely by those who are less sensitive to the fumes.
315
-------
Before the factory is contemplated, the residents
would represent a random selection of the two groups.
An accurate survey would thus uncover some proportion
of residents who would evaluate the conjectured
factory as equivalent to a loss of $2,000. Given a
competitive housing market, however, they would not
actually experience a loss of $2,000 if the factory
comes. Instead, they would experience a loss equal to
moving and searching expenses—including an amount
equivalent to the "wear and tear" involved in moving—
and possibly some loss in consumer surplus due to unique
qualities of their former house and neighborhood.
The costs related to moving are often ignored or under-
estimated by the indirect method, as we discuss below.
However, it is important to note that where residents
choose to move as a result of the appearance of some
activity with external effects, the sum of their losses
attributable to moving necessarily is less than their
evaluation of the disutility due to the external effect.
Even if the presence of external effects of some
activity does not produce adjustments on the part of
those affected, there are often severe problems of
measurement at both stages of the direct approach.
Among the most severe problems is the multidimensional
316
-------
nature of the external effect in many cases. Consider,
for instance, an activity which generates substantial
noise, at levels which produce a significant impact
on the surrounding area. While an airport is the
obvious example, noise is also a potentially signifi-
cant external effect associated with sanitary landfilling,
But noise is not a homogeneous quantity. It varies
in power, as measured in decibels, but a decibel reading
summarizes only the power of the noise at one instant
of time. Furthermore, it does not reflect the partic-
ular shape of the noise wave form; that is to say, it
does not reflect the frequency or other qualities of
the noise, qualities that often have as much or more
to do with the disutility caused by noise as the
power. Even if one is dealing with a homogeneous noise
source, in which each episode consists of noise generated
at a given decibel level, for a given duration, and
with a given quality, there remain the problems of
comparing different combinations of time patterns.
Is a particular noise which occurs once every five
minutes twice as much "total noise" as the same noise
occuring once every ten minutes? Are twenty episodes
within a five minute period worse than the same
twenty episodes strung out over an hour? If worse,
317
-------
how much worse? More generally, is a noise of a given
quality produced at 140 decibels worse than two
episodes of the same quality at 110 decibels? Are
occasional noises of high power worse than a steady
drone at lower power?
In certain dimensions/ such as ear drum damage
these questions involving rankings of different time
patterns and of different qualities of noise may
have objective, "scientific" answers. Furthermore,
the case of ear damage illustrates one situation where
the indirect method of evaluation may be quite inap-
propriate, since it depends crucially on the assumption
that the people affected by the externality are aware
of the damage they are incurring and understand the
nature and extent of the damage. We discuss that
limitation of the indirect method below. But if
we restrict the discussion here to the damage produced
by the noise in terms of discomfort and annoyance, it
is clear that such questions as those listed above do
not have objective answers, and are meaningful only
in terms of the subjective response of the affected
person.
The output of the first phase of the direct method
of evaluating external social costs is a catalogue of
318
-------
the physical manifestations of the external effect,
measured with objective scales. In a case in which the
external effect is multidimensional, as in the case
of noise, the only completely appropriate output of the
first phase must necessarily be a complete description
of all dimensions of the effect. This may be an
infeasible, extremely expensive task. In fact, it is
impossible in the case of noise if the requirement
is interpreted strictly since one dimension of noise
is its precise quality, which is definable only in
terms of combinations of an infinity of different
wave forms.
Obviously substantial compromises have to be made
with the conceptually appropriate procedure, in which
the physical effects are first catalogued and their
social costs are then evaluated. In particular,
indices and summary measures must be substituted for
the complete list of all dimensions and time patterns.
But the only proper criteria by which to rank and
combine different noise patterns, for instance, are
subjective. One is then left with the choice of in-
ferring such criteria from surveys and introspection,
or by deriving them implicitly from observations on
how individuals react to the different combinations
319
-------
of physical effects. Where the weights which individuals
assign to different dimensions are implicitly determined,
as they are by the impact of such external effects
on property values, it seems unwise to ignore that
information.
The multidimensionality of the external effect
creates problems not only of weighing the various
dimensions, but also of monitoring the physical effects.
This latter problem is particularly severe with respect
to the problem of water pollution, in particular the
potential for groundwater pollution at a landfill. In
Chapter IV, the numerous possible components of
leachate from a landfill were described. It was also
noted that the potential for groundwater pollution
through leaching depends in extremely complex ways on
the geological and hydrological characteristics of the
landfill bed, on the distribution of rainfall, on the
nature of the solid waste in the landfill, and on many
other factors such as the vegetation on top of a
completed landfill and the uniformity of settling.
Furthermore, not only does the leachate consist of
many different components whose proportions may vary
enormously, but its relative concentration in the
groundwater and its pattern of dispersion will depend
320
-------
on the particular extent and configuration of the
groundwater flows. Thus/ neither the quantitative
nor the qualitative nature of groundwater pollution
can be generalized from one landfill site to another with
reasonable accuracy. Similar considerations apply to
groundwater pollution attributable to carbon dioxide
production. Since an adequately designed sanitary
landfill is supposed to produce no water pollution
problems at all, and since this objective is alleged
to be feasible in acceptable sites, these considerations
are possibly relevant only to open dumps or to landfills
which do not meet specifications. However, it was
noted in Chapter IV that leachate production and
carbon dioxide production appears to be extremely long
lived, and the groundwater pollution from a landfill
could conceivable appear many ydars after the completion
of a landfill site.
In the case where the damages caused by a partic-
ular external effect depend crucially on specific
characteristics of the landfill site and cannot be
easily predicted, it becomes extremely difficult to
accurately include external social costs in the
relevant policy calculations. Of course, this prob-
lem is not specific to the direct method of calculation;
321
-------
if the external effects are highly variable and
difficult to predict, the method of measurement cannot
compensate for inherent uncertainty. In this situation
it does not follow that such external effects should
be ignored, however. If there is truly no way to
accurately predict the extent of the external effects
which will be produced, it is still preferable to
attribute average external social costs to each
proposed project, where the population of cases on which
the average is calculated should be chosen from examples
as comparable as possible. The attribution of such
an average will counteract a bias which is otherwise
present.
The last difficulty with the direct method of
evaluation which we shall consider is the crucial
importance of measuring the marginal impact of the
activity in question. The external social cost
attributable to an activity relates to the change in
the relevant variables. If an open dump contributes
to air pollution, for instance, the external social
cost attributable to the dump activity roust be
measured as the total social cost of air pollution
with the dump munus the total social cost without the
dump. The important consideration here is that the
322
-------
social cost at the margin may differ considerably
from the average social cost of the effect, and using
an average figure may be very misleading.
One case in which the crucial importance of
measuring the marginal effect is particularly obvious
is where the external effect relates to noise. If
a landfill is placed in a secluded residential area which
was previously tranquil and idyllic, for instance,
even a moderate amount of noise at the landfill and
a moderate amount of heavy traffic to the fill may
be considered extremely damaging by the surrounding
residents. This is particularly likely in that the
residents may well be a self selected group who
have paid a premium, in land price and in inconvenience,
in order to enjoy quiet. Placing a landfill in an
industrial area, however, may be practically un-
noticed by those in the area; the average noise level
may be increased very slightly and the surrounding
buildings may be already soundproofed. Implicit in
these remarks is a noise-level/damage relationship as
pictured in Figure 5.la, where the marginal damage
due to a given increase in the noise level is very
high when the level is low, but decreases as the noise
level increases. In that case, using average damages
323
-------
damage
damag
a.
b.
c.
Figure 5.1 Three Possible Relationships between Damage
(Social Cost) and the Level of Some Causal Variable x.
computed for a particular case as a proxy for marginal
damages always results in overestimating the appropriate
value. It is not inevitable that marginal damage
should be decreasing, however. In some situations,
a very small level of some variable (for instance,
air pollution) might be unnoticeable and cause
essentially no damage. As some "threshold" level is
passed, however, the marginal damages become very great.
Such a situation is depected in Figure 5.1b; in that
case, using average damage as a proxy always produces
an underestimate of the marginal damage. In Figure
5.1c, we show a mixed case where there is a threshold
effect at low levels, but as the level increases a
saturation effect becomes dominant. This latter
case would seem to be a plausible representation of
324
-------
the damage/level relationship, although the'relevant
range for particular variables may be constrained to
a certain section of such a curve. For such a re-
lationship, the marginal damage resulting from a
given increase in the variable becomes much greater
than average damages as the threshold value is
approached and becomes much smaller than average damage .
as the saturation level is approached. When computing
the social cost attributable to some activity, the
use of average damage as a proxy for marginal damage
may result in a considerable bias.
The indirect- method of evaluating the external
social cost attributable to some activity finesses
virtually all of the problems we have examined. Faced
with the presence or with the prospect of an objection-
able activity in the neighborhood, present and potential
landowners are led to investigate the possibilities
of reducing the "damage" by making adjustments such
as air conditioning or fences or staying indoors more,
and calculate the costs of such adjustments. The
market value of surrounding property reflects the
attractiveness of the property after an optimal set
of adjustments is performed and the cost of those
optimal adjustments. In their evaluation of how the
325
-------
external effects of the objectionable activity affect
the attractiveness of the property, the different
dimensions of the external effects are implicitly
weighted. Furthermoref the weights which are implicit
in the change in market value are not necessarily
those of the residents living in the area when the
offending activity is introduced (or first becomes
foreseen), but the subjective weights of those to whom
the property is valued highest in the face of the
externality. The various dimensions of the external
effects are thus combined with weights which implicitly
reflect an economically efficient allocation of land,
which is the appropriate criterion. Finally, since
the indirect method evaluates the social cost of the
external effects by comparing the surrounding property
values with and without the presence of the objection-
able activity, it automatically measures the marginal
Impact of the activity.
These assertions beg the question of how to
estimate the change in property values as a result
of the activity, obviously. Here one could take
the direct approach of asking residents, or possibly
local real estate experts, their opinions on the
conjectural question of how property values would
326
-------
be affected if the objectionable activity were're-
moved. It is the strong belief of most economists,
however, that in such a situation the guesses of
residents, even of real estate specialists, are
likely to be crude and unreliable. Furthermore, as
discussed in Chapter II, there may be strong incentives
for people to misrespresent their true feelings. The
economises alternative approach is to construct an
econometric model of property values in the area, using
economic theory as a guide to the general design and
specification of the model and statistics on property
values and characteristics to calibrate the parameters
of the model. Such an approach, if reasonably
successful, implicitly measures the social costs of
of the externalities in terms of what people do rather
than in terms of what they say.
Obviously the results of the indirect method can
only be as good as the underlying model of the
determination of property values. Various attempts
to construct and estimate such models were surveyed
in Chapter III. Many of these studies are based on
models which appear to include significant methodo-
logical errors. The specification of the model, and
the estimation method used, are crucial to the validity
327
-------
of the final results. In constructing the econometric
model which is estimated in Chapter VI, considerable
effort was expended in attempting to specify a model
which is based on a consistent theoretical analysis
of the determinants of property values and which
avoids apparent errors of previous studies. Never-
theless, it must be acknowledged that the success of
such a model is severely limited by data availability.
This is particularly true where the effects which are
being estimated represent the impact on property values
of secondary effects, such as the impact of a view
or the impact of moderate road noise. Since there
are a limited number of houses within an area which can
be considered to share the same general location
characteristics, the data limitation is inherent
in the problem. This has to be considered an important
disadvantage of the indirect method, although its
significance depends upon the purpose of the study.
If the objective is to estimate the effect of a
nearby landfill rather precisely, say within $100,
it is unrealistic to feel confident that an econometric
model, calibrated with a modest number of observations,
can provide such precision. This would require not
only that the specification of other determinants of
328
-------
house value—such as square footage, number of rooms,
size and shape of lot, et cetera—be virtually "perfect",
but that one also have a large (say 1000) number of
observations on which to estimate the model. The basic
reason that generous data is probably necessary is
that the attributes which make houses more attractive,
such as size, number of bathrooms, quality of kitchen
appliances and built-in fireplaces, quality of con-
struction, location, lot size, and so forth, tend to
be highly intercorrelated. Expensive houses tend to
have more of all the things which make houses more
attractive, and cheaper houses tend to have fewer
and/or less of them. This makes it very difficult
to statistically separate the effects of these
different attributes. While it may not be necessary
to infer precisely how these other attributes, which
are not the effects that are of interest, affect the
value of the house, the less accurate those inferences
are, the more uncertain will be inferences about the
effects of interest, such as the impact of increased
road noise. It is not necessary, however, to worry
independently about these effects on the reliability
of the estimated effects that are of interest; the
method of multiple regression analysis produces
329
-------
statistics useful in judging reliability which im-
plicitly reflect the uncertainty in all of the other
parameter estimates of the model.
It is also important to recognize that changes
in property values ignore certain costs borne by
those owning the surrounding property which are related
to the possible uniqueness of the house or the property
from the owner's point of view. The market value of
a house does not represent what it is worth to its
present owner, but to the potential buyer whose
reservation price (the lowest price at which he is
willing to buy it) is highest. In many cases, the value
of the house to potential buyers becomes higher than
its value to the owner. In that case, it becomes
profitable for the owner to sell and move elsewhere,
if the differential is large enough to compensate for
the costs and headaches of moving and searching. In
most cases, however, the value of a house to its owner
is greater than its market value. This may be due
either to unique aspects of the house and lot which the
owner happens to value more highly than any potential
buyer, but it is more usually due to the fact that
the owner has modified and decorated the house to his
own specific tastes. It may also be due to personal
330
-------
relationships with neighbors or other characteristics
of the neighborhood which are valued by the owner.
It is thus often the case that an owner has
positive consumer surplus in the house that he occupies,
particularly if he has occupied the house for some
time. In Chapter II the difficulties of unambiguously
defining the notion of consumer surplus is analyzed,
but its existence is not in doubt. When an objection-
able activity with external effects appears in a
neighborhood, it may affect not just the market value of
the surrounding property but also wipe out consumer
surplus. In particular, those who are led to move by
the appearance of the activity do not receive com-
pensation for their loss of consumer surplus, or even
for their moving and searching costs. For those
who choose to stay, their consumer surplus may be
reduced more than the reduction in market value, so
that the change in market value underestimates the
cost to the owner. However, it is also possible, in
the case where the resident chooses not to move, that
consumer surplus increases in the face of the fall
in market value, so that the decrease in market value
overestimates the social cost. It must be emphasized
that the consumer surplus referred to here relates
331
-------
to the unique qualities of the particular house, and
not to houses of that general type or in that general
area. While it would be preferable to be able to
refine one's evaluations of social cost to take into
account changes in that consumer surplus, such
measurement is virtually impossible. The only source
for such measurement would be to question the owner,
who would have strong incentives to exaggerate his
loss in welfare.
The roost significant limitation of the indirect method,
however, is that it can be justified in terms of standard
economic theory only under the assumption of perfect
knowledge. This assumption requires all buyers and sellers
to be completely aware not only of the prices of all goods
offered for sale, but also of the complete set of attributes
of each individual product. In particular, the presumed
connection between the market value of a house and the
preferences of its potential owners for the different
dimensions of "housing services" rests directly on the
assumption that all potential buyers fully and completely
appreciate all the "qualities" of each house.
If the housing market were truly characterized
by perfect information, new owners would never be
surprised, either by unforeseen defects or unforeseen
332
-------
virtues in a house. There would be no need for
building codes, except for paternalistic motives.
Obviously, such is not the case. Nevertheless, it
is in the interests of prospective purchasers to
carefully inspect a new house, and building codes
greatly reduce uncertainty. Furthermore, it is
rational for a potential buyer to expect a certain
number of unforeseen problems to reveal themselves
after he occupies the house, and to correspondingly
"discount" his expected satisfaction level. Thus,
even the admission of imperfect information about house
attributes by potential buyers does not imply that
there is a systematic bias in market values, although
it does suggest that observed market values are not
perfect indicators of what market values would be if
information were perfect.
These considerations arise because potential
buyers do not fully comprehend exactly how satisfied
they would feel if they occupied the house. The
perfect knowledge assumption implies that they do,
and is obviously not strictly met in the real housing
market. Nevertheless, this inaccuracy in the com-
petitive model does not seem crucial. The more
bothersome question is whether the satisfaction of
333
-------
of owners who occupy a house, even over a long period,
should be uncritically used as a criterion of their
own welfare.
The primary objection to that presumption—
namely, that a person's satisfaction with his house
should be the criterion of how much his particular
house contributes to his welfare—is the situation
where an individual may be unaware of some of the
effects of his environment on his physical or mental
health, particularly long term cumulative effects.
Consider a house near an airport, for instance. The
airport noise will be considered an irritation by most
people, and their discomfort will be reflected in the
market value of surrounding houses. But the airport
noise may also damage their hearing ability, at a
rate which is so slow as to be virtually unnoticeable
but cumulatively still significant.
If people are unaware of the cumulative, permanent
component of the damage, the market values of surround-
ing houses will reflect only the short run "nuisance"
effect, but not the permanent component of the damage.
Obviously, the reduction in market value may dramatically
understate the external social cost of the external
effect in such a case.
334
-------
Such a bias is even more obvious where the
homeowner is completely unaware of the external effect,
as in the case where the water supply contains
asbestos particles which are unnoticeable but harmful.
Indeed, the discovery rate of toxic effects attributable
to pollutants or chemical additives which were pre-
viously considered harmless appears to be accelerating.
The evaluation of social costs in such cases by the
indirect method is obviously nonsense where the home-
owners are unaware of the hazard. In that situation,
the direct method is the only alternative. Unfortunately,
application of the direct method depends upon statistics
which allow one to predict the incidence and severity
of the relevant health problems on the basis of
contamination rates. Since such data is often unavail-
able, the unsatisfactory procedure of basing policy
on the most pessimistic assumption conceivable appears
to prevail.
The second difficulty in applying the direct method
in this case concerns the problem of evaluating the
welfare loss resulting from the higher incidence of
specific health problems or even death. The approach
economists adopt here is not different in concept from
evaluation problems in general, although complications
335
-------
related to uncertainty and people's attitudes toward
risk arise. We will not attempt to summarize the
extensive literature on the appropriate methods of
doing benefit-cost analysis in the health area. In
general, economists reject the agnostic view, as
reflected in the statement that "it is absurd to believe
that a dollar value can be put on human life", which
is not a basis for policy decisions but for policy
non-decisions. Rather, economists note that people
implicitly trade off risks of death or disease
against other goods continually, and that these
decisions implicitly determine the "disutility" of
death or ill health. Of course, inferences about such
dimensions of people's preferences are different.
336
-------
REFERENCES
1. Bolt, Baranek, and Newman, Inc. Land Use Planning
Relating to Aircraft Noise. United States
Government, Federal Aviation Administration.
October, 1964.
2. Bolt, Baranek, and Newman, Inc. Noise from Aircraft
Operations at Miramar Naval Air Station California
and Land Use Interpretations. Report 2098.
December, 1971.
337
-------
CHAPTER VI
AN EMPIRICAL STUDY OF THE IMPACT OF
SANITARY LANDFILLS ON RESIDENTIAL PROPERTY VALUES
A. INTRODUCTION
In this chapter we present the empirical results
based on an analysis of the impact on property values
of the presence of sanitary landfills in and near residen-
tial areas. Four sanitary landfills in the Los Angeles
area were chosen for the study and detailed information
was obtained about single family dwellings near these
landfills which were sold during the last five years.
The data for the study came from two sources; the Los
Angeles County Assessor's Office and SREA Market Data
Center Incorporated. Only those residential properties
found in both the lists were included in the study, be-
cause much of the information available from one source
was not available from the other and vice versa. For the
selected residences, the data from both the sources were
combined and used in the analysis.
Section B of this chapter describes the solid waste
disposal practices in Los Angeles County. Section C
discusses the determinants of residential property values,
describes the variables used, and discusses the expected
338
-------
nature of their effect on property values. The actual
empirical results are presented in Section D, and a sum-
mary of the conclusions is given in Section E.
B. LOS ANGELES COUNTY SOLID WASTE DISPOSAL PRACTICES
Solid waste disposal facilities in Los Angeles County
are currently provided by three distinct entities: the
County Sanitation Districts, municipalities, and private
operators. Any discussion of solid waste management prac-
tices/ therefore/ must be segregated into these same three
catagories.
At the present time three cities (Los Angeles, Burbank,
and Whittier) operate their own landfill sites. The latter
two cities have only one site each and handle a comparatively
small volume of solid waste (see Table 6.1). In addition,
each of these sites is closed to the public; only city
operated collection vehicles are ..permitted to use their
respective landfill site. The City of Los Angeles {referred
to, as the City from this point on) operates two landfill
sites with a combined volume of over one million tons per
year (see Table 6.1). These two sites receive approxi-
mately two-thirds of the refuse collected by City forces,
:with the remaining one-third disposed of at sites operated
by the County Sanitation Districts [2].
During the past few decades the City has shown a
339
-------
flexible attitude toward solid waste disposal techniques.
Prior to 1955 the householder segregated his refuse, with
garbage collected separately by both City forces and pri-
vate collectors and disposed of under contract to hog
feeders. Other wastes, including ashes and noncombustibles,
were collected and dumped in outlying areas. Combustible
rubbish was burned on the homeowner's property or hauled
away by private collectors [2],
In 1955, a pilot project was started in the Harbor
area for the collection of combustible rubbish as a third
separate service by the City. This change was motivated
by the demand for cleaner air and the elimination of the
backyard incinerator, officially legislated in 1957. In
the early years of the combustible rubbish collection pro-
gram, the City attempted to maintain a salvage and recla-
mation service, but it soon became apparent that this
procedure was simply increasing costs. By 1961 garbage
disposal units had reduced residential food wastes to a
very small percentage of the volume during the 1950's.
Furthermore, as a result of advances in the container
industry, the supply of separately collected glass and
metals had declined to less than 30% of its peak value by
1964. At this time a combined collection program was
instituted with all household refuse to be used as fill
340
-------
material for park and land reclamation, a practice which
has continued until the present [2].
At the present time there are twenty-five landfill
sites being operated in Los Angeles County: six by the
County Sanitation Districts, four by municipalities, and
fifteen by private operators. There is no mode of solid
waste disposal other than that of the sanitary landfill
currently in use in the County [3], Table 6.1 provides
an ennumeration of these sites together with information
on the acreage of and volume handled at each site. Al-
though some of the site names include the word "dump",
there are no open dumps operated in Los Angeles County;
all solid waste has been disposed of by the sanitary land-
fill method since 1964 [3],
The current service given by the Bureau of Sanitation
in the City of Los Angeles is the combined collection of
all household and yard wastes on a one time per week basis,
and the collection of commercial garbage at a minimum of
twice a week and more frequently as required by the cus-
tomer. More than 1,200,000 tons of refuse are collected
and disposed of each year. Of this total, the collection
of commercial garbage from such establishments as restau-
rants and hotels accounts for only 6,000 tons and the col-
lection of animal carcasses another 1,500 tons. The great
341
-------
preponderance of refuse collected, therefore, is that which
comes from the City's residential units. For locations
producing large quantities of refuse, such as high density
apartment complexes, the standard weekly collection service
is increased to three or five times per week. This service
is supported from general tax funds without special fees
being levied against the homeowner [3],
Commercial garbage is collected from restaurants,
hotels, hospitals, and other establishments on a basis of
payment for each collection, with a minimum of two stops
required per week. The collection of animal carcasses is
provided without charge for both animal hospitals and the
homeowner [3] .
The City's Bureau of Sanitation owns and operates
a fleet of approximately 650 vehicles, including: 500
packer-type refuse collection vehicles, as well as earth
movers, tractors, utility collection vehicles, and smaller
service units. Refuse deposited in the rear loading com-
partment: of the packer vehicle is compacted to about one-
half its original volume by a hydraulically activated packer
blade. Trucks with a capacity of 25 cubic yards have been
found to be more suitable for use in the residential col-
lection operation, while smaller units with a 16 cubic yard
capacity are used in areas where more maneuverability
342
-------
is required [3] .
The Los Angeles County Sanitation Districts operate
six landfill sites: Palos Verdes Landfill, Spadra Landfill,
Mission Canyon Landfill/ Calabasas Landfill, Scholl Canyon
Landfill, and the Puente Hills Landfill. All these sites
are open to private citizens, rubbish haulers, gardeners,
municipalities, school districts, County agencies, and
industry. These groups are permitted to dispose of waste
material including: combustible and noncombustible rubbish,
garbage, paper, grass, tree and yard trimmings, lumber,
tires, cans, bottles etc., but excluding separate loads
of garbage. The charge for disposing of each of these
materials depends on the site involved; Figures 6.2 through
6.7 provide a detailed listing of these charges at each
of the County landfills.
There are now a total of twenty-seven County Sani-
tation Districts, the first established in 1923, They
are governed by a Board of Directors made up of the mayors
and county supervisors whose jurisdictional areas are
involved. The Districts stretch from the eastern border of
the City of Los Angeles to Pomona and from Long Beach to
Lancaster and include a total of seventy-one cities. By
combining under a single administration it is possible to
attack problems on a regional basis. Prior to 1957 the
343
-------
Districts were concerned strictly with sewage treatment,
but with the legislated demise of the backyard incinerator
they were thrust into the refuse problem. Preparations had
been underway since 1950 to provide for the transfer and
disposal of refuse, though the house to house collection
remained in the hands of local authorities [7],
At the present time the Board of Directors feels
that sanitary landfills are the most economical and nuisance
free method for the disposal of solid waste. Depleted
gravel pits and unused mountain canyons provide the
County with many acceptable sites for land reclamation under
this method of disposal. During the filling operation,
the refuse is compacted and covered with earth each day
in order to maintain a clean land mass. Properly managed
landfill sites have become aesthetically acceptable land
which has been used to create parks, playgrounds, golf
courses, and other community improvement projects. Other
refuse disposal methods, such as composting, incineration,
and ocean disposal have been considered and evaluated;
however, none have been found to be as practical as the
sanitary landfill method for use at this time. Sites
chosen by the City and County are selected with great care.
Among the important factors considered in such selection are
the feasibility of land acquisition at a reasonable cost,
344
-------
the suitability of the reclaimed land for subsequent bene-
ficial use, and the proximity of the site to the collection
area to be served. The reader may refer to Figure 6.1 to
see the distribution of all landfill sites throughout
Los Angeles County [73.
There are fifteen privately operated landfill sites
in Los Angeles County which handle approximately 38% of the
total solid waste volume. The majority of these sites are
abandoned gravel pits which, being nothing but huge excava-
tions, have little value in other uses. The private opera-
tors run the landfill until the pit is filled and then
attempt to sell the site to the State or County as park or
recreation land. As one would expect of abandoned gravel
pitsi these sites are mostly located in industrial areas,
near operating gravel pits or other types of heavy industry.
Although the private sites will accept business from any
source (private, municipal, and industrial) the latter
source provides the greatest volume due to its proximity
to these sites.
For the purposes of this study four landfills have
been chosen: Palos Verdes, Calabasas, Sheldon-Arleta and
Tujunga. Descriptions of these four sanitary fills are given
in Section D. These fills were chosen because of their
proximity to residential areas. Several sites were rejected
345
-------
as being unsuitable for the purposes of this study. For
instance, the Mission Canyon landfill is rimmed by mountains,
and there are no homes within two miles of the site. How-
ever, a new tract has just been started one-half mile away
and it may be of interest to study their property values a
few years from now.
The North Valley site is also not suitable because
it is over five miles away from residential areas. The
Bradley pit is surrounded by industry and freeway and the
closest home is over one-half mile away. None of the homes
have a direct view of the landfill. The Burbank dump was
also rejected because it is in the mountains, is completely
hidden from view and is over a mile from the nearest homes.
Both Owl Park and Alpha Investment are in the same highly
industrialized area. They are both gravel pits used by
neighboring firms to dispose of heavy materials. There are
some homes about a mile away but they have no view of the
fill. Furthermore, the gravel pits themselves are much
more offensive because they cause dust and noise and have
machinery which stand over 100 feet high.
346
-------
Table 6.1 OPERATING LANDFILL SITES IN LOS ANGELES COUNTY
Site (Acreage)
Fill Rate (Tons/Year)
Sanitation Municipal Private
Districts
1) Palos Verdes Landfill (220)* 1,600,000
2) Spadra Landfill (185) 180,000
3) Mission Canyon Landfill (485) 1,400,000
4) Scholl Canyon Landfill (345) 500,000
5) Calabasas Landfill (380) 260,000
6) Puente Hills Landfill (265) 940,000
7) Sheldon-Arleta Pit (60) 510,000
8) Toyon Canyon Landfill (150) 730,000
9) Burbank City Landfill (150) 120,000
10) City of Whittier Landfill (32) 36,000
11) North Valley Refuse Center (300) 360,000
12) Bradley Avenue 600,000
13) Hewitt Pit (64) 280,000
14) Tujunga Pit (66) 370,000
15) Azusa Rock and Sand (40) 250,000
16) Owl Park (60) 84,000
17) Canyon Park Development (15) 95,000
18) B.K. Co. (45) 560,000
19) Operating Industries (182) 900,000
*The location of a landfill is identified in Figure 6.1 by the
corresponding number in the table.
347
-------
Table 6.1 (continued). OPERATING LANDFILL SITES IN LOS ANGELES COUNTY
Site (Acreage)
Fill Rate (Tons/Year)
Sanitation Municipal Private
Districts
20) Norwalk Dump (18) 14,000
21) Kobra Landfill (15) 41,000
22) Ascon (60) 300,000
23) Lancaster Dump (24) 18,000
24) Antelope Valley Landfill (40) 20,000
25) Chiguita Canyon Landfill (97) 28,000
Subtotals 4,880,000 1,396,000 3,910,000
Total 10,186,000
Source: Existing Landfills in Metropolitan Los Angeles County,
County Sanitation Districts of Los Angeles County, 1970.
Note: The figures in parentheses represent the areas, in
acres, of the corresponding landfill.
348
-------
Fiqi.re 6.1 EXISTING LANDFILLS IN METROPOLITAN LOS ANGELES COUNTY.
Ul
EXISTING LANDFILLS
Metropolitan Los Angeles County
REDONOO \\Q O
BEACH I TORRANCE
A
PAI.OS VERGES
ESTATES //
a
PRIVATE
MUNICIPAL
C.S.D.
0 I 2 5 4 B
-------
Figure 6.2 PALPS VERDES, LANDFILL NO. 1
Qnvercd loads mean cleaner reads //
PALOS VERDES LANDFILL NO. 1
(Operated by County Sanitation District*)
WHERZ J6S01 So. Crenshaw Blvd.. Rolling Hillf. Calif. 772-2631. 377-3S14
MHO Open to all private citiiens. rubbish haulers, gardeners.
municipalities, school districts, county agencies, industries, etc.
WHAT Solid and liquid wastes, coebustible and noncombuttible rubbish,
wrapped household garbage, paper, grass, tree and yard trii-nir.qs.
lueber, tires, cans, bottles, etc., hut excluding icp»rjt» loads
of garbage.
BOW MUCH Refuse $2.00 per ton
Solid inert material (rock, dirt, etc.). . . . 1.25 per ton
Kard-to-handle. bulky material* 3.00 per ton
Liquid wastes 2.00 per ton
HlnlsHji charge 2.00 per load
Load pull-off 2.00 per pull
Tank washout *-00 per wash
WHEN 8:00 A.M. to 5:00 P.M. daily excepting Sundays and Holidays
CREDIT See weighwaster for application form, cr contact Sanitation District
office, Z020 Beverly Blvd., Los Angeles, Calif. 90057, 464-1J70.
Checks cannot b« accepted at the landfill.
SAMITATICeJ Cover loads In accordance with all city, county and state regulations.
LOCATION Sc« map shown below. Duap only at spots designated by landfill
personnel. Observe speed regulations on site.
Vi/n
350
-------
Figure 6.3 SPADPA, LANDFILL NO. 2
leads mean cleaner rr*ch //
SPA3RA LANDFILL NO. 2
(operated by County Sanitation Districts)
WHZRT 4125 West Valley Blvd., Walnut, California. (714) 595-2710
WHO Open to all private citizens, rubbish haulers, gardeners,
Bunicipallties, school districts, county agencies, industries, etc.
WHAT Waste Material, including combustible and noncombustible rubbish,
garbage, paper, grass, tree and yard trimings, lur^er, tirei,
Cans, bottles, etc., but « eluding separate loads of_ garbage.
BOM MUCK Refuse $1.60 per ton
Solid inert material (rick, dirt, etc.). . . . 1.10 per ton
Hard-to-handle. bulky materials 2.SO per ton
Liquid wastes 1.60 per ton
Hinimsi charge 2.00 per load
Load pull-off 2.00 per pull
Tank washout 2.00 per wash
WHEN BiOO A.M. to 5:00 P.M. Monday Through Saturday. 12:00 Noon to
4:00 P.M. Sundays. Closed Holidays.
CREDIT See weighnaster for application ferns, or contact Sanitation
District office, 2020 Beverly Blvd., Los Angeles, Calif. 90057,
484-1370. Checks cannot be accepted at the landfill.
SANITATION Cover loads in accordance with all city, county and stat*
regulations.
LOCATION See map shown below. Dump only at spots designated by landfill
personnel. Observe speed regulations on site.
7/1/72
351
-------
Figure 6.4 MISSION CANYON-SEPULVEDA, LANDFILL NO. 3
j^S^gsrrTzr — - (Covered loads mean cleaner roads //
MISSION CANYON-SERTLVEDA LANDFILL NO. 3
{Operated by County Sanitation Districts)
WHERE 2201 No. Sepulveda Blvd., West Los Angeles, Calif. 272-291*2
WHO Open to all private citizens, rubbish haulers, gardeners,
municipalities, school districts, county agencies, industries, etc.
WHAT Waste materials, including combustible and noncorabustible rubbish,
wrapped household garbage, paper, grass, tree and yard trlsmings,
lumber, tires, cans, bottles, etc., but excluding separate loads
of garbage.
HOW MUCH Refuse $1.60 p«l ton
Solid inert material (rock, dirt, etc.). ... 1.10 per ton
Hard-to-handle, bulky materials. 2.50 per ton
Minimum charge 2.0O per load
Load pull-off 2.00 per pull
WHEN 8:00 A.M. to 5:00 P.M. dally excepting Sundays
CREDIT See welghmaster for application forms, or contact Sanitation
District office, 2G2O Beverly Blvd., Los Angeles, Calif. 90057,
Wt-1370. Checks cannot be accepted at the landfill.
SANITATION Cover loads in accordance vith all city, county and state
regulations.
LOCATION See map on reverse side. Dump only at spots designated, by landfill
personnel. Observe speed regulations on site.
7/1/72
352
-------
Figure 6.5 MAP OF MISSION CANYON LANDFILL
MISSION CANYON LANDFILL
2200 NO. SEPULVEDA BLVD.
WEST LOS ANGELES.CALIF.
TO VENTURA FRWY
MULHOLLAND
DRIVE
A
CHALON
ROAD
OFF RAMP
TO SANTA MONICA
353
-------
Figure 6.6 SCROLL CANYON, LANDFILL NO. 4
Qjvered /ax/s mean cleaner roads //
SCHOLL CANYON LANDFILL NO. U
(Operated by County Sanitation Districts)
WHERE 75>»6 No. Flgueroa St., Los Angeles, Calif. 2^
yjjo Open to all private citizens, rubbish haulers, gardeners,
municipalities, school districts, county agencies, industries, etc.
WHAT Waste materials, including combustible and noncombustible rubbish,
wrapped household garbage, paper, grass, tree and yard trimmings,
lumber, tires, cans, bottles, etc., but excluding sepirate loads
of garbage.
HOW MUCH Refuse $1.60 per ton
Solid inert material (rock, dirt, etc.). . . . 1.10 per ton
Hard-to-handle, bulky materials 2.85 per ton
Minimum charge 2.CO per load
Load pull-off 2-00 per pull
WHEN 8:00 A.M. to 5:00 P.M. daily excepting Sundays
CREDIT See weighmaster for application forma, or contact Sanitation
District office, 2020 Beverly Blvd., Los Angeles, Calif. 90057,
ltfllt-1370. Checks cannot be accepted at the landfill.
SANITATION Cover all loads in accordance vith all city, county and state
regulations.
LOCATION See map on reverse side. Dump only at spots designated by landfill
personnel. Observe speed regulations on site.
7A/72
354
-------
Figure 6.7 MAP OF SCHOLL CANYON LANDFILL
SCHOLL CANYON LANDFILL
7546 NO.FIGUEROA ST.
LOS ANGELES,CALIF 90041
355
-------
Figure 6.8 .CALABASAS, LANDFILL NO. 5
•l^^r "£Sr^ - ___- . (jyvened fax/s mean cleaner mwb //
CALABASAS LANDFILL NO. S
by County Sanitation District*)
(DM MUCH
CREDIT
26919 West Ventura freeway. A-roura, Calif. Phon* - 689-1410
On« Hilt welt of Lai Virgenes Road.
Open too all private citizens, rubbish hauler*, gardener*.
Municipalities, school districts, county agencies, industries, etc.
Solid and liquid wastes, combustible and noncoBbustible rubbish,
wrapped household garbage, paper, grass, tree and yard trimings,
lunber, tires, cans, bottles, etc., but excluding separate loads
of garbage.
JUfuse J1.6C per ton
Solid inert eaterial (rock, dirt, etc.). . . . 1.10 per ton
Rard-to»handle, bulky materials 2. 50 p«r ton
Liquid wastes 1.60 per ton
KlnisuB charge 2.00 per load
Load pull-off 2.00 per pull
Tank washout 2.00 per wash
8:00 AJ1. to 5:00 P.M. dally excepting Sundays and Holidays.
See veighnaster for application form, or contact Sanitation District
office, 2020 Beverly Blvd., Los Angeles, Calif. 900S7, 484-1370.
Checks cannot be accepted at the landfill.
SANITATION Cover loads in accordance with all city, county and state regulations.
LOCATION
Take Lost Hills Road turnoff fron Ventura Freeway.
See stap shown below.
7A/72
356
-------
Figure 6.9 PUENTE HILLS, LANDFILL NO. 6
\J,
lilu^^_ — - Qjvered loads mean cleaner road* !l
FUEKTE HILLS LANDFILL NO. 6
(Operated by County Sanitation Districts)
WHERE 2800 So. Workman Mill Road, La Puente, Calif.
Local: 699-520U. From Los Angeles: 723-9261*.
VHO Open to all private citizens, rubbish haulers, gardeners,
municipalities, school districts, county agencies, industries,
etc., but excluding separate loads of garbage.
WHAT Waste material, including combustible and noncombustible rubbish,
garbage, paper, grass, tree and yard trimmings, lumber, tires,
cans, bottles, etc.
HOW MUCH Refuse &..&) per ton
Solid inert material (rock, dirt, etc.). ... 1.10 per ton
Hard-to-handle, bullny materials 2.50 per ton
Liquid wastes 1-60 per ton
MinUnura charGe 2-00 per load
Load pull-off 2-0° J*r f"*11
Tank vashout . 2.00 per vash
WHEN 6:00 A.M. to 5:00 P.M., Monday through Saturday.
8:00 A.M. to l*:00 P.M. Sunday.
CBEDIT See weigbmaster on site for application forms, or contact
Sanitation District office, 2020 Beverly Blvd., Los Angeles,
Calif. 90057, WU-1370. Checks cannot be accepted at the
landfill.
SANITATION Cover loads in accordance vith all city, county and state
regulations.
LOCATION See map on reverse side. Dump only at spots designated by
landfill personnel. Observe speed regulations on site.
7A/72
357
-------
Figure 6.10 MAP OF PUENTE HILLS LANDFILL
NOTE: TAKE PECK ROAD
OFF-RAMPS FROM
FREEWAYS.
PUENTE HILLS LANDFILL
2800 SO. WORKMAN MILL ROAD
LA PteNTE, CALIFORNIA
o
«IO HONDO
COLIIOI
358
-------
C. DETERMINANTS OF RESIDENTIAL PROPERTY VALUES
It is well recognized that the sale price or assessed
value of a given property depends on a number of attributes
of the property. Thus, the relationship between the price
and characteristics of a house can be expressed in the
form
Pi " f(Xil' Xi2' *•" Xij' '•" Xin)
where P. is the price and x.. is the jth attribute of the
ith residence. A similar relationship can be written down
between the assessed value of a house and the attributes.
The exact specification of the relationship will be dis-
cussed later. The number of individual characteristics
of any property is unmanageably large. In practice/ we are
thus limited to measuring the most important attributes in
each, of several broad categories and modeling the effects of
the unmeasured attributes as an error or disturbance term.
The characteristics of a property are basically of
two types, (1) characteristics of the house and (2) charac-
teristics of the neighborhood. The house characteristics
can be further subdivided into two types, basic attributes
and improvements. Neighborhood characteristics can take
the form of amenities as well as disamenities. In this
359
-------
section, we describe the various attributes of a property
for which information could be obtained. The nature
of the expected relationship between price or assessed
value and the characteristics of a residence are also
discussed. The complete list of the variables and their
definitions are given in Table 6.2.
As indicated above, property values are measured
by two variables: the price at which it was last sold
(PRICE) and the assessed value (ASSVAL). In this study/
models in which each of these measures is used as the de-
pendent variable are estimated. Some of the issues
involved in specifying how the house and neighborhood cha-
racteristics influence property values can be made evident
by considering how the amount of living space (AREA) and
land area (LOTSIZ) might influence property values. Both
PRICE and ASSVAL include the value of both land and house.
Consequently, we would expect that the larger the size
of the lot (measured in square feet) , the higher will be
the value of the residential property. We can also expect
that the living area (measured in square feet) would
exhibit a positive relationship to residential value.
Ceteris paribus, a house with a larger living area is
likely to be more valuable. However, to the extent that
in our sample larger houses are found on larger lots, it
360
-------
Table 6.2 LIST OF VARIABLES AND THEIR DEFINITIONS
AFLOOR
AGERDD
AIRCON
APPL
AREA
ASSVAL
AV/LOT
BDROOM
BTHRUM
CLFRNT
CLLFRN
CONDN
CONDO
FLOORS • AREA
1 - exp[R(AGE - DD)]
where AGE is the age of the house in
years, DD is the useful lifetime, and
R is the discount rate. Alternative
values for R = .02, .04, .06, for
DD = 50, 100, 150 (DD = 100 and 150
are denoted by 10 and 15). Note that
a newer house means a higher value for
AGERDD.
1 if no air conditioning
2 if wall or evaporative cooler
3 if central air conditioning
1 if no range/oven or dishwasher
2 with one of them
3 with both of them
Living area in square feet
Assessed value in dollars (including
that of land)
Assessed value per square foot of lot
size
Number of bedrooms
Number of bathrooms
CULDSC • FRONT
CULDSC • log (FRONT)
Condition of maintenance:
1 = poor
2 = fair
3 = average
4 = good
5 = excellent
1 for condominium, 0 otherwise
361
-------
Table 6.2 (continued). LIST OF VARIABLES AND THEIR DEFINITIONS
CONST
CORNLT
CRFRNT
CRLFRN
CULDSC
DFILL
DFRWY
DGOLF
DJUNK
DOWNP
DPARK
DRAMP
DSHOP
DTRUCK
DUMP!
FIREPL
FLOORS
FRONT
GARAGE
LARFLR
LEASE
LLOTPL
Constant term
1 for a corner lot, 0 otherwise
CORNLT • FRONT
CORNLT • log(FRONT)
1 for cul-de-sac, 0 otherwise
Distance to sanitary landfill (all
distances are in hundreds of feet)
Distance to nearest point of freeway
Distance to golf course, if any
Distance to any disamenities
Down payment as a fraction of sale
price
Distance to nearest park
Distance to freeway ramp
Distance to nearest shopping center
Distance to dump truck route
1 for dump # i (i = 1,..., 4)
1 4- number of fireplaces
1 for more than one floor
Front footage of lot measured in feet
1 + number of car spaces in garage
log(AREA) • FLOORS
1 for lease
log(LOTSIZ) » log(POOL)
362
-------
Table 6.2 (continued). LIST OF VARIABLES AND THEIR DEFINITIONS
LOTSIZ
LTSZPL
Size of lot in square feet
LOTSIZ • POOL
(Other variables beginning with L or LG denote the logarithm
of what follows; three examples are given below)
LAG415
LAV/LT
LGAPPL
MAINST
ODDLOT
OTHRUM
POOL
PRICE
PR/LOT
QUALTY
REMODL
SALDAT
VIEWFL
log(AGE415) (with R = .04 and DD = 150)
log(AV/LOT)
log(APPL)
1 if house is on main street
where P = estimated perimeter, S = area of lot
R = average PA/S for sample houses
near a fill
Number of other rooms excluding bedrooms,
bathrooms, and kitchen
1 if no pool
2 if non-heated pool
3 if heated pool
Sale price in hundreds of dollars
Price per square foot of lot size
Quality of construction:
1 = poor
2 = average
3 = good
4 = excellent or luxury
1 if no remodelling
2 if kitchen, bathroom, or both remodelled
3 if completely remodelled
Date of sale in fractional years (e.g.,
March, 1970 = 70.25)
1 if fill is visible from the residence
363
-------
Table 6.2 (continued) . LIST OF VARIABLES AND THEIR DEFINITIONS
VIEWLT 1 if lot has good view
WIND Number of degrees away from prevailing
downwind from the fill
YARDSZ LOTSIZ - AREA
364
-------
may be difficult to distinguish the effects of these two
variables. In the case of both area and lot size, it
is reasonable to expect that the additional units of each
variable would add to the property value at diminishing
rate. Thus, the extra value of an additional 100 square
feet of area or lot size is likely to be more for a small
house than for a large house. Furthermore, the value of
extra square feet of living space may depend on the lot
size and vice versa.
There are at least two alternative ways in which the
hypothesis implicit in the above statement may be tested.
First, consider the following specification:
PRICE = a + a,AREA + a., (AREA)2 + a,LOTSIZ + a (LOTSIZ)2
oX2 3 4
+ a AREA-LOTSIZ + Z (6.1)
where Z represents the effects of other characteristics of
a house which are held constant for the moment, and a. 's
are the unknown coefficients to be estimated from the data.
Partially differentiating this with respect to AREA, we
have
3PRICE
= a, + 2a AREA -I- a LOTSIZ (6.2)
3AREA •»• 2 5
365
-------
The above relationship specifies that the addition in
the price due to a unit addition in the living area de-
pends on the living area itself and the size of the lot.
A similar statement can be made with respect to the effect
of LOTSIZ on PRICE. The hypothesis that the owner of a
large house may not be willing to pay very much for an
additional 100 square feet of area as compared to an owner
of a small house, will hold provided a- has a negative
sign and is significantly different from zero. The com-
posite variable (AREA»LOTSIZ) is the interaction term
designed to measure the extent to which LOTSIZ affects the
marginal effect of area on price. If a* is statistically
insignificant, then we can conclude that the additional
price of an extra unit of area does not depend on the lot
size.
A second formulation which can also be used to test
the hypotheses described earlier is the following:
PRICE = 3 -AREA -LOTSIZ «Z (6.3)
o
This is the well-known Cobb-Douglas (or double-log) for-
mulation which can also be written as
log(PRICE) = logS + 0,log(AREA) + B log(LOTSIZ)
o 1 2
+ log Z . (6.4)
366
-------
It is readily noted that
3log(PRICE)
x Slog(AREA)
3 is thus the constant elasticity of the price with
respect to the living area. It is approximately the
percentage change in price for a unit percentage change
in the area. The marginal effect of area on price is
given by
3PRICE 6-1 B
«• 6 B,AREA * LOTSIZ Z • Z (6.5)
3AREA ° X
If 0
-------
garding the significance of individual coefficients more
difficult. This problem is less serious in the Cobb-
Douglas formulation, since each attribute appears only once.
With more attributes, there are more nonlinear and inter-
action effects, and this difference becomes even greater.
The double log formulation captures these effects while
at the same time reducing multicollinearity and economizing
on the number of parameters to be estimated, thus increasing
the power of tests (.in the statistical sense). For these
reasons the emphasis throughout this study will be on the
double log formulation even though it is somewhat less
flexible than C6.ll. We do, however, present the results
of other formulations as well.
In the following paragraphs we discuss briefly the
reasons for including the other explanatory variables in
tfije analysis and we suggest in some cases how they might
enter the statistical model.
Lot size less area; To separate the influences on
price of the living area from that of the size of the
yard we have obtained the yard size which is measured as
the lot size less the living area.
Age; The age of the house would also be important
in determining its price. Other things being equal, a
newer house would cost more, and hence, would also be
368
-------
assessed more. Would the market value decline at an
increasing rate with age or at a decreasing rate? Emerson
[4] has argued that the market value of the residence
declines at an increasing rate with age. Let P be the current
price of the property, and let B be the value of the annual
flow of benefits it yields. For simplicity, this flow is
assumed constant over the total useful life of the property,
which, we call D. Finally, let A be the current age of the
residence, and let r be the relevant real discount rate.
Then the relationship between the current price and the
annual flow of benefits is given by
P * Bdt - 11 - e. (6.6)
0 r
It is easily shown that 3P/3A < 0 and 32P/3A2 < 0. Thus,
the value of a property declines with age at an increasing
rate. In this study, age enters the relation in the form
1 - expJrtA-D)] or its logarithm, and we have alternatively
assumed that D = 50, 100, and 150 years, and that r = .02,
.04, .06. Thus, the relationship between price and age
is specified as follows:
log(PRICE) = a log {1 - expIr(A-D)]} + Z or
f\
PRICE = B {1 - exp[r(A-D)]> + Z.
C6.7)
369
-------
It is evident from (6.6) that the double log formulation
is more appropriate here also.
Sale date: The date of sale has also been incorporated
as one of the independent variables. A more recently sold
home is likely to have a higher market value. The sale
date is measured in fractional years. For example, March
1970 is denoted by 70.25. Thus, a larger value for this
variable represents a more recently sold house.
Quality; Besides age, the quality of construction is
also important in determining the value of a house. The
higher the quality, the greater the market value of the house.
The quality of the construction was rated as follows: Poor,
average, good, excellent and luxury. These were translated
into numerical scores with poor =1, average = 2, and
so on, with excellent and luxury getting the same score.
This is evidently not a satisfactory procedure, because a
unit score increase from poor to average may have a much
different increase in value than the same unit score in-
crease from good to excellent. To some extent, this can
be taken into account by entering quality nonlinearly
in the functional form. An alternative is to assign a dummy
variable to each of the types of quality. In general we
have avoided using too many dummy variables in this study
because the large number of qualitative variables would
370
-------
have meant a substantial loss of degrees of freedom, espe-
cially with interaction terms.
Condition; It is evident that a house maintained
in good condition is likely to be more valuable than one
which is not. This variable also is in qualitative form
with ratings poor/ fair, average, good and excellent. As
for the quality variable, this has been converted to a
numerical score with poor = 1.
Bedrooms, bathrooms and other rooms: The number of
bedrooms, bathrooms and other rooms are of obvious impor-
tance in affecting the residential property values. Other
rooms are defined as total number of rooms (excluding kitchen
and baths) less number of bedrooms. These will thus be
living room, den, playroom, family room, etc.
Garage space; This variable is measured as 1 + number
of cars which can be parked inside the garage. A house with
no separate garage will thus have a score of one.
Number of floors; It is clear that for a residence
with a given lot size and area, the value will depend on
the number of stories it has. A two-story house would mean
a stronger and hence a more expensive foundation, and would
also have more outside area. We may therefore expect a
positive relationship between price and number of floors.
This variable has been defined as a dummy with the value
371
-------
of one for a residence with more than one floor. Because
it is not possible to take the logarithm of a dummy variable,
it is specified additively in a linear formulation and
multiplicatively in the Cobb-Douglas form (or equivalently,
additively in the double-log form). Thus, for instance,
the relationship between price and FLOORS can be specified
as follows:
log(PRICE) = a FLOORS + Z
or
PRICE = BF FLOORS + Z . (6.8)
This procedure has been adopted for all the dummy variables
used in the study. The interaction terms, however, are
specified differently. For example, the interaction be-
tween FLOORS and AREA is specified in the linear form as
(FLOORS-AREA = AFLOOR) whereas in the double log form it
is specified as (FLOORS-log AREA = LARFLR). The latter
form simply states that the elasticity of PRICE with respect
to AREA depends on the number of floors.
Fireplace: A residence with one or more fireplaces
is likely to be more valuable than one without a fireplace.
This variable is measured as 1 + number of fireplaces in
the residence.
372
-------
Built-in appliances: This refers to the presence
or absence of a built-in dishwasher and/or range-oven, but
has been translated into a numerical score with one for a
house with1 no range-oven or dishwasher, two for a residence
with only one of the two built-ins and three for a house
with both. Houses vary substantially in the quality of
the appliances. For instance, a residence may be equipped
with a self-cleaning oven which certainly costs more. How-
ever, since such data are not available, only the above
scores are used.
Air-conditioning: In Los Angeles where smog is often
a serious problem, many home owners prefer to air-condition
their homes and hence, we would expect air-conditioned
houses to have a higher market price. A score of one was
given to a home with no air-conditioning, two for a residence
with either a wall air-conditioning or an evaporative cooler
system, and three for a home with central air-conditioning.
Pool; Homes which had swimming pools had either a
heated pool or ordinary pool. To distinguish between these,
a score of three was assigned to a heated pool, two for a
non-heated pool, and one for a home without a swimming pool.
Remodeling: Remodeling very often augments the value
of a property. A house which has not been remodeled was
rated one, a residence completely remodeled was rated
373
-------
three, and others in between (kitchen, bath or both
remodeled) were assigned a score of two. Again these
scores are inadequate variables because they do not take
into account the extent and quality of remodelling.
Front-footage; The front-footage of a lot,
measured in feet, has also been used in this study to
test whether this variable significantly affects the
price of the property. Because of a possible high
correlation between the size of the lot and the front-
footage, it may be difficult to separate the influence
of this variable.
Corner lots; Corner lots may exert both positive
and negative influences on house value. On the one hand
a corner lot has greater accessibility to the street
and increased privacy due to having only two neighbors
instead of three. However, a corner lot may also have
the disadvantage of greater exposure to traffic noise
and fumes, and having pedestrians cut across the
property. This variable is measured as a dummy variable
with a corner assuming a value of one.
Cul-de-sacs; A residence located in a cul-de-sac
may be more valuable because of reduced traffic and
greater privacy. This variable is also measured as a
dummy with one for cul-de-sac.
374
-------
View lot; A house with a good view will obviously
have a higher market value than one without. Although
a rating scale giving various types of view would be
appropriate, this variable is only available as a dummy
with one for a house with a good view.
Main street; Similar to a corner lot, a house
located on a main street may have both desirable and
undesirable elements, although the undesirable features
are likely to dominate. A house located on a main
street, takes the value of one for this dummy variable.
Odd lot; It is hypothesized that, other things
being equal, a residence situated in an odd-shaped lot
will have a different value as compared to "normal"
lots. To test this hypothesis, we have constructed a
variable called ODDLOT which is obtained as follows.
If we take the view that a normal lot is more square
shaped than an odd lot, then the ratio of the perimeter
of the lot to the square root of lot size is a rough
measure of the oddness of a lot. If all the lots in the
neighborhood are odd by this criterion, then there may
be no difference in property values due to this
particular variable. We therefore consider the absolute
value of the deviation of this ratio from its mean
for the entire sample. More formally, this variable
375
-------
is defined as follows:
ODDLOT = P/s - R | (6.9)
/*
where P is the estiminated perimeter of the lot, S is the
L ^ A
lot size and R is the mean of P/S for the sample residences
A
near a given fill. The perimeter, P, is estimated as
follows:
ss
(a) If the lot is a corner lot, then P is twice
the front footage (F).
(b) In the case of a lot not in a corner or a
A
cul-de-sac, P is estimated as 2 [P + S/P], on the
assumption that it is a rectangular lot.
(c) If a lot is situated in a cul-de-sac, we
assume that it has the shape indicated in the following
diagram:
40
y
376
-------
The perimeter is given by P = F + 2x + y. We
also assume that an average cul-de-sac has a 40 foot
radius to the lot. If 9 is the angle, in radians, made
by the sides of the lot as indicated above, then
409 = F and y = (40+x)9. To obtain x, we use the lot
size S which is the area ABCD. We have
S « OABCDO - OADO ~ 1(40 + x) 2 9 - - 402 9.
2 2
This gives the following quadratic equation and solution
for x.
x2 + 80x - 25/9 = 0
x « -40 +
Therefore,
ft
P « P-80 + (2+F/40)1600+80S/F (6.10)
In double log formulations, LODDLT = log (ODDLOT) is
used as an explanatory variable.
Downpayment; The downpayment made at the time of
purchase of the property may influence its price, al-
though it is not clear what its direction of effect is
likely to be, if any. Downpayment has been measured
as a fraction of sale price.
Lease hold; To the extent that owners who reside
in the home take better care of the residence, a house
rented out is likely to fetch a smaller price than one
in which the owner is living. It would therefore be
377
-------
desirable to included a dummy variable which takes the
value one if the property is leased. However, since
none of the sample houses were rented, this variable
is not used in the study.
Condominium: A house in a condominium can be
expected to have a lower value than a similar single
family house, because the latter has greater privacy.
This variable is also measured as a dummy which takes
the value one if the residence is in a condominium
and zero otherwise. This variable also is omitted
because none of the sample residences were in a con-
dominium.
Distance to shopping centers; The distance to
the nearest shopping center is likely to have a non-
linear effect on property values. Being too close to the
shopping center is generally not desirable because of
traffic congestion, noise and fumes. Similarly, being
too far away is also not desirable because of increasing
commuting costs. We have therefore included the
distance to the nearest shopping center as one of the
explanatory variables.
Distance to fche nearest park; This variable will
measure the premium, if any, that people are willing
to pay to- be close to a neighborhood park. Here also,
378
-------
there may be a non-linear effect.
Distance to the nearest freeway ramp; Being not
too far away from a freeway entrance is also desirable
because of greater accessibility to other parts of the
city. However, being closer to the freeway may also
mean greater noise associated with it.
Distance to the nearest point of the freeway; In
some instances, a residence may be located very near
the freeway itself, even though the access to it may
be some distance away. This is not a desirable
feature because of increased noise and possible
visual pollution.
Distance to the sanitary landfill: This variable
is of major interest to this study. If it is true that
homeowners are adverse to being close to a landfill,
however sanitary it may be, then we will expect that,
ceteris paribus, the further away a house is located
from a fill, the higher its value. In other words, we
would expect a positive coefficient for this variable.
It should, however, be pointed out that the effect of
the sanitary landfill would also depend on the future
use to which it will be put once the landfill has been
closed. For instance, if the landfill is expected to
be turned into a golf course or other recreation area
379
-------
after a few years, then the effect of the landfill
could be positive if the residents are willing to
put up with current disamenities arising from the
landfill in order to enjoy future benefits.
Distance to the dump truck route; Being close to
the route the trucks take to carry waste to the land-
fill is clearly undesirable because of the added noise
and possible damage to health and property from the
exhaust fumes emitted by the trucks.
View of the sanitary landfill; To the extent that
people are willing to pay a higher premium for a better
view from their residence, a direct view of the fill
may be quite undesirable. It may be, however, that
expected future benefits of a view of the landfill when
it would become a recreation area will outweigh
present disamenities. This variable is measured as
a dummy with a value of one if the residence has a
direct view of the landfill.
Degrees from prevailing downwind; An undesirable
feature of a landfill is the possibility of bad odor
emitting from it. Houses that are downwind of the
landfill may receive more odor and dust than other houses
which may be closer to the fill site but not down-
wind. For this reason, we have measured the number
380
-------
of degrees away from the prevailing downwind direction that
the residence in question is situated.
Distance to other disamenities; We also measured
the distance to other disamenities that may be present
in the nieghborhood, For instance, Tujunga has a gravel
pit nearby which is clearly a disamenity.
Interaction terms; The list of explanatory
variables also includes several interaction terms. For
instance f the effect of the front-footage of a house
on property value is likely to depend on whether the
house is on a corner lot or is situated in a cul-de-sac.
Typically, houses in a cul-de-sac have smaller front-
footage and much larger back footage. The fact that it
has a smaller front-footage may not decrease its value
because it is situated in a cul-de-sac. Another inter-
action term takes account of the joint effect of lot
size and pool. The pool variable does not measure the
size of a pool, which may be important because a larger
pool may increase the value of the house. A larger
pool is unlikely unless the lot size is big. Thus,
there is a possibility of an interaction between lot
size and pool. Another interaction used here is that
between the living area and the number of floors. The
elasticity of price with respect to the area may
381
-------
depend on the number of floors in the house. Table 6.2
includes all the interaction terms used in the analysis
and the corresponding names.
D. EMPIRICAL RESULTS
In this section we present the results of the study
of the effects of sanitary landfills on the four sanitary
landfills in Los Angeles county. The sample sizes for
each of the fills were as follows: Palos Verdes, 447;
Tujunga, 279; Sheldon-Arleta, 179; Calabasas, 61. Because
the characteristics of each neighborhood are quite different,
the estimation has been done for each fill separately. In
each case double log and linear relations using price and
assessed value as the dependent variables and a long set of
independent variables, were estimated for the different
useful lifetimes (D) and discount rates (r) specified in
Section C in defining the AGE variable. The values of D
and r which gave the "best" fit in terms of the highest
_2
R*-and F statistics were chosen. Although there were some
exceptions, it was found that in most cases D = 150 and
r = .02 (as defined by the age variable AGE215 or LAG215)
gave the best fit. For the sake of convenience and com-
parability these values have been used throughout the rest
of the estimation for all landfills and all other dependent
variables. A comparison of the average values for a number
382
-------
of variables for each of the four sites is given in Table
6.3. The results of the analysis for each fill are presented
in the following four sections.
Palos Verdes
The Falos Verdes Landfill is one of the largest sanitary
landfills in Los Angeles County. It handles the largest
yearly volume of refuse (approximately 1/600,000 tons) and
covers an area of approximately 220 acres (Table 6.1).
It is situated at the base of the eastern slope of the Palos
Verdes Peninsula, a mountainous region bordering the Pacific
Ocean in the southwestern corner of Los Angeles County.
Surrounding and overlooking the landfill site to the west
and south are the upper-income residential communities of
Rolling Hills and Rolling Hills Estates. The terrain in
this area is one of partially wooded, sharply rising foot-
s
hills, steep narrow canyons, and small winding roads. Con-
sequently, many of the homes in this area have an unrestricted
view of the landfill. The atmosphere in these communities is
quiet, spacious, well-groomed, and surburban. This ambiance
offers a sharp contrast to the commercial and industrial
sections of Torrance which borders the site to the north and
west.
The Palos Verdes Landfill began oj: orations in 1957
and is scheduled for completion in 1977, at which time a
383
-------
golf course is planned for the site. Another portion of this
site which was completed some time ago has already been
turned into a botanical garden.
The average sale price of the 447 sample residences
was $50,110 with an average assessed value of $43,545. The
mean size of the lot is 12,908 square feet and the living
area is 1,843 square feet. In terms of house values, Palos
Verdes is clearly the most expensive of the four neighborhoods
under study. In lot size, number of bedrooms and bathrooms,
however, Calabasas dominates the other three communities.
It is interesting to note that for the two low value neigh-
borhoods, Tujunga and Sheldon-Arleta, the ratio of assessed
value to sale price was much higher than that of the wealthy
neighborhoods of Palos Verdes and Calabasas.
Tables 6.4 and 6.5 present the estimated coefficients
and their t-statistics for PRICE, ASSVAL, log(PRICE) and
log(ASSVAL). They indicate that the explanatory variables
used in this study explain 90 percent of the variation in
price, 81.3 percent of the variation in assessed value,
90 percent of the variation in log(PRICE) and 80 percent
of the variation in log(ASSVAL). Considering that this
is a cross-section of residences with a fairly large sample,
the goodness of fit is quite good. It is not surprising,
however, that a number of independent variables are statis-
384
-------
tically insignificant (at a 95 percent confidence level).
With a large list of highly related variables we can expect
multicollinearity to seriously affect the results making
it difficult to make inferences about individual coefficients.
In order to examine the sensitivity of the coefficients
to alternative specifications of the relationships, several
different formulations were tried, although in each case
several key variables Csuch. as area, lotsize, bedrooms
and bathrooms} were retained regardless of whether they
*ere statistically significant. Table 6.6 presents the
coefficients for logCPRICEj. and logCASSVALl for one such
formulation. Most of the coefficients in Table 6.6 are
significant. CWe employ two-tailed tests at the 5% level.)
While. R2 does not change much between Tables 6.4 and 6.6,
the F statistics (which test the joint significance of all
the independent variables) more than double for both
dependent variables. The evident reason for this is the
substantial gain in degrees of freedom (D.F.) in Table 6.6.
The variables describing the sanitary landfill are
of maximum interest for this study. Among the fill variables
the wind direction (LGWIND) was the only one statistically
insignificant. The most interesting result is that both
distance to the sanitary landfill and view of fill showed
opposite signs from what was initially expected. For instance,
385
-------
houses with a direct view of the fill (16 percent of the
sample) had a significantly higher market price and inte-
restingly enough, were also assessed higher compared to
similar residences without a view of the fill. This
conclusion holds for both linear and double log formula-
tions and even if the full list of explanatory variables
are included (i.e., even with substantial multicollinearity/
which is likely to raise estimated standard errors and thus
lower t-statistics). Similarly, houses closer to the sani-
tary landfill (the average distance is a mile) have higher
prices, as indicated by the fact that the distance to the
fill has a negative and significant coefficient both in
the linear and double log formulations. The elasticity of
price with respect to distance to fill is -6.9 percent
(Table 6.6). This means that if the distance from the
fill is doubled keeping other characteristics of a residence
the same, then its price will decrease by 6.9 percent.
Similarly, a direct view of the fill tends to raise the
sale price an average of 2.8 percent.
The explanations for this counter intuitive result
are not difficult to find in the case of Palos Verdes.
As mentioned earlier, a portion of the landfill has already
been converted into a lovely botanical garden. The rest
of the sanitary fill is expected to become a golf course
386
-------
in 1977. Since Palos Verdes is an upper-income area, one
might expect actual and potential residents to place a high
value on proximity to a golf course or botanical garden.
The net value of proximity to the landfill must depend on the
present discounted values of (.1) long-term proximity to the
garden, C21 a few years of fill-related disamenities, and (3)
long-term proximity to the golf course after the fill is
converted. One explanation for our results is that the
relevant discount rate is low, so that the third value is
large enough relative to the second to make the sum of
all three positive.
Although home owners near the Palos Verdes landfill
are willing to pay higher prices even though there is a
direct view of the fill or the residences are closer to the
fill, they apparently do regard as costs the noise, visual
pollution, congestion and air pollution associated with the
trucks that carry solid waste to the fill. The distance
to the truck route showed a significant positive coeffi-
cient indicating that the farther away from the truck
route a property is situated, the higher its value. This
conclusion also holds for the assessed value and is inde-
pendent of the functional forms used to formulate the esti-
mation models (Tables 6.4 through 6.6). The elasticity
for distance to the truck route was 0.0281 indicating that
387
-------
a residence situated at a distance twice that of another
residence of similar characteristics has a 2.81 percent
higher price on average. However, the effects of being
near or having a view of the landfill are larger than the
effect of being near the truck route which suggests that,
for the Palos Verdes area, the presence of the sanitary
landfill has not been detrimental to property values but
in fact has been beneficial.
While the focus for this study is on the variables
associated with the presence of the sanitary landfill,
other characteristics of a property are also of considerable
interest. In interpreting the effects of other variables
on price and assessed value, we have used only the double
log formulation which was justified in Section C as being
more powerful. The results are presented in Table 6.6. The
variables which significantly affect the sale price of a
residence are discussed first.
It was hypothesized in Section C that lot size and
living area will both have positive but declining effects.
In other words, additions to either of the two variables
would increase property values but at a decreasing rate.
The results support this hypothesis. The price elasticity
with respect to area was 49.9 percent whereas that for lot
size was only 5.7 percent (Table 6.6). Since both are
388
-------
less than 100 percent, it follows that there is diminishing
marginal returns to the size of the lot and living area; that
is, the owner of a large house will be willing to pay less
for an extra 100 square feet area (or lot size) than the
owner of a smaller residence. When the size of the yard
was used instead of the lot size, the elasticity for area
became 55.3 percent and that for yard size was 4.1 percent.
Among bedrooms, bathrooms and other rooms (excluding
kitchen), only the number of bathrooms had a statistically
significant coefficient. A doubling of the number of bathrooms
would increase the price by 14.1 percent (Table 6.6). The
number of bedrooms did not have a significant coefficient,
which implies that an additional bedroom (keeping number
of bathrooms, living area, lot size and all the other
characteristics of the residence constant), does not affect
the price of a property.
As can be expected, a more recently sold house, as
indicated by a higher value for SALDAT, had a higher market
price. In fact, this variable had the highest t-statistic
among all the independent variables. The dummy variable for
a view lot is also significant implying that residences with
a good view had a higher market value than similar houses
without a view. On average, view adds 2.8 percent to the
price of a house. The condition of maintenance of a property
389
-------
was also important in explaining the variation in price.
A doubling of the CONDN rating increased the average price
by 12 percent. Another significant variable was the down
payment as a fraction of the sale price. A doubling of this
fraction would increase the sale price by 15.6 percent.
This suggests that buyers are willing to pay the higher
mortgage rates associated with higher loan to market value
ratios only if they get a break on the price of the house.
It was mentioned earlier that a house situated on a
main street will have both desirable and undesirable elements.
The undesirable features — the noise/ air pollution and
hazards associated with the increased traffic — are likely
to dominate any advantage due to nearness to shopping areas/
bus routes or schools. In the Palos Verdes area, a main street
location means reduced property values as shown by a signi-
ficant negative coefficient for this dummy variable. A
house located on a main street will only sell for approxi-
mately 93.5 percent of a similar house not on a main street
(Table 6.6). Also consistent with our expectations is the
finding that homes nearer a park are more valuable than
similar homes farther away. The distance from the park
shows a significantly negative coefficient, although the
magnitude of the elasticity itself is not large. A doubling
of the distance from the park would reduce market values by
390
-------
by an average of 2.8 percent.
Residences with a swimming pool do have a higher sale
price as indicated by the fact that the variable LGPOOL has
a significant positive coefficient. Having a heated pool
raises the market price even more. The quality of construc-
tion LGQUAL, however, does not have a significant coefficient
in Table 6.6, but if other variables with a low t are excluded,
LGQUAL became significant, which suggests that its statistical
insignificance resulted from the high correlation* between
quality of construction and other variables.
A number of variables in Table 6.4 which do not have
significant coefficients were tested for joint significance
with other variables with insignificant coefficients. None
of the joint F-tests involving CORNLT, CULDSC, LODDLT,
LGAPPL, LFRONT (and its interaction terms with corner lot
and cul-de-sac) were significant and these variables were
excluded from Table 6.6.
All the results interpreted above refer only to the
(determinants of the sale price of the residences in the Palos
Verdes area. It is of considerable interest to examine the
variables that are of significance in explaining the assessed
value of the property as well. To a large extent there will
be a high correlation between the sale price and the assessed
Value of a property because past sale price of the same or
391
-------
a similar house in the neighborhood is usually taken into
account by the assessor in setting the value of a property
for tax purposes. However, the criteria used by the assessor
need not coincide with those which determine the market
demand for residential property.
It is of interest to note that for the Palos Verdes
area as well as the other three neighborhoods, the list of
variables that significantly explain the variation in sale
price is not the same as that of variables explaining the
variation in the assessed value of the property. For instance,
the condition of maintenance of the property, down payment,
main street location and distance to the nearest park, all
had statistically significant coefficients for the dependent
variable LPRICE, but none of these variables significantly
affected LASSVL. On the other hand, fireplace, the oddness
of a lot and the number of degrees away from the downwind
from the fill, significantly affect the assessed value of
the property but not the sale price. As one might expect,
the more the number of fireplaces the higher the assessed
value. Also, the variable LODDLT which measures the oddness
of a lot has a significantly negative regression coefficient.
In other words, the more odd the shape of a lot (as measured
in this study) relative to other lots in the neighborhood,
the less the assessed value of the property. The variable
392
-------
LGWIND, which did not significantly affect the sale price
of a property, has a positive and significant coefficient
for the assessed value dependent variable LASSVL. This, of
course, does not imply that the assessor's office consciously
takes the degrees away from wind (or the oddness of the lot)
into account when setting the assessed value of a property.
It is also of interest to compare how a variable which
significantly affects both the assessed value and the sale
price differs in the respective elasticities. Some of the
differences are quite small. The direct view of the fill,
for example, increases prices by 2.8 percent whereas it
increases the assessed value an average of 3.1 percent. A
doubling of the distance from the fill would decrease the
sale price by 6.9 percent whereas it would decrease the as-
sessed value by only 4 percent. The elasticity with respect
to the distance to the dump truck route was .036 for assessed
value as compared to .028 for the price.
The most significant difference between the regressions
for the sale price and those for the assess value is in the
SALDAT variable. This variable had a significant but negative
coefficient for the assessed value but a positive and very
highly significant coefficient for price. The coefficient
from the price regression is easy to interpret: since the
property prices have been rising steadily for many years, the
393
-------
more recent the date of sale for which the price was recorded,
the higher is the price. The coefficient from the assessed
value regression does not have so obvious an explanation. One
possibility is that assessed values of properties that are
not sold in any given year rise more rapidly than market
values. But since assessments are supposed to reflect market
values, whenever a property is sold, the assessor must equate
assessed value with price. Consequently, the more recently
a property is sold the less will the assessed value have
risen relative to market value and the lower the assessed value
will be. If assessed values rose in proportion to market
values, then the coefficient on SALDAT would not be signifi-
cantly different from zero.
Another difference between the assessed value and
price regressions is in the age variable. A newer house (as
indicated by a larger value for AGE215 and LAG215) had a
significantly higher sale price than a similar, but older,
house. However, for the assessed value regressions, the age
variable did not have statistically significant coefficients.
The number of bathrooms also showed a substantial difference
in the effect on price and assessed value. For example, a
doubling of the number of bathrooms would increase the
assessed value by an average of 7.8 percent, whereas in the
case of the sale price, the increase would have been as
394
-------
much as 14.1 percent.
A number of independent variables had to be excluded
from the analysis because they could not be measured for the
Palos Verdes area for a variety of reasons. These were:
Distance to the nearest freeway and the nearest off ramp
(there is no freeway nearby), distance to the shopping center
and the dummy variables for condominium and leasehold.
395
-------
Table 6.3 MEANS AND STANDARD DEVIATIONS (IN PARENTHESES) OF SELECTED VARIABLES
Variable
PRICE ($)
ASSVAL ($)
LOTSIZ (sq. ft.)
AREA (sq. ft.)
Bedrooms
Bathrooms
Other rooms
Viewlots (%)
Leased (%)
Corner lot (%)
Main street (%)
Cul-de-sac (%)
View of fill (%)
DFILL (ft.)
DTRUK (ft.)
DPARK (ft.)
DJUNK (ft.)
DFRWY (ft.)
DRAM? (ft.)
Sample Size
Palos Verdes
50,110 (14,706)
43,545 (11,571)
12,908 (9/296)
1,.843 (512)
3.4 (.72)
2.1 (.51)
3.3 (3.4)
36.2
0.2
21.5
2.9
6.0
16.1
5,290 (2,000)
550 (1,351)
3,899 (1,664)
9,400 (3,806)
None
None
447
Tuj unga
22,962 (4,474)
20,793 (3,801)
7,285 (4,401)
1,149 (275)
2.7 (.57)
1.3 (.40)
2.5 (.67)
None
None
15.3
12.5
2.5
None
3,059 ^,160)
3,190 (1,081)
4,552 (1,338)
2,244 (753)
4,980 (1,231)
5,021 (979)
279
Sheldon-Arleta
24,097 (5,164)
'21,772 (4,869)
7,382 (1,909)
1,228 (336)
2.9 (.75)
1.4 (.49)
2.6 (.61)
None
None
)%6
. . 9
None
None
?,^33 (964)
j.,550 (682)
2,255 (1,348)
3,470 (821)
1,733 (1,156)
2,179 (943)
179
Calabasas
44,880 (6,360)
38,785 (6,045)
9,994 (1,890)
2,588 (433)
4.0 (.86)
2.7 (.50)
3.7 (.51)
None
None
3.3
None
None
None
3,007 (424)
621 (418)
None
None
1,774 (471)
1,907 (326)
61
U)
ID
CTV
-------
Table 6.4 NAME OF LANDFILL: PALOS VERDES
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
LOTHRM
SALDAT
LGPOOL
LODDLT
CORNLT
LFRONT
CULDSC
CRLFRN
CLLFRN
LGAPPL
LAIRCN
VIEWLT
LREMOD
LFIRPL
LGARAG
FLOORS
LARFLR
DOWNP
LGQUAL
LCONDN
LLOTPL
LDPARK
LDJUNK
MAINST
LDFILL
LGWIND
LDTRUK
VIEWFL
R2
F-
D.F.
Dependent Variable
LPRICE
-1.3343 (-2.79)*
4.8636 (4.44)*
0.0422 (2.36)*
0.4891 (12.66)*
-0.0123 (-0.43)
0.1503 (4.92)*
-0.0024 (-0.16)
0.0553 (12.47)*
-0.1958 (-0.83)
0.0122 (0.62)
0.1954 (0.92)
0.0153 (0.56)
0.0890 (0.68)
-0.0401 (-0.96)
-0.0174 (-0.52)
0.0132 (0.91)
-0.0756 (-1.32)
0.0251 (2.74)*
0.0272 (1.04)
0.0058 (0.27)
-0.0153 (-0.81)
-0.4892 (-0.51)
0.0623 (-0.51)
0.1495 (2.77)*
0.0465 (1.77)
0.1169 (5.25)*
0.0295 (1.18)
-0.0301 (-2.34)*
-0.0284 (-0.80)
-0.0611 (-2.46)*
-0.0581 (-2.85)*
-0.0253 (-0.68)
0.0260 (5.39)*
0.0280 (2.15)*
0.900
121.974
413
LASSVL
7.2940
2.1362
0.0919
0.5449
-0.0489
0.0999
0.0056
-0.0168
0.5089
-0.0791
-0.5147
-0.0791
-0.0258
0.1013
0.0038
0.0377
0.0611
0.0196
0.0115
0.0563
0.001
0.6254
-0.0868
0.1082
0.0102
0.0112
-0.0450
-0.0078
-0.564
-0.373
-0.0247
0.1244
0.0359
0.0383
0
54
(11.41)*
(1.56)
(4.08)*
(11.17)*
(-1.36)
(2.59)*
(0.30)
(-3.13)*
(1.70)
(-3.18)*
(-1.93)
(-2.28)*
(-0.16)
(1.91)
(0.09)
(2.05)*
(0.85)
(1.69)
(0.35)
(2.07)*
(0.00)
(0.52)
(-0.57)
(1.59)
(0.31)
(0.40)
(-1.42)
(-0.48)
(-1.26)
(-1.19)
(-0.96)
(2.64)*
(5.89)*
(2.33)*
.798
.427
413
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
397
-------
Table 6.5 NAME OP LANDFILL: PALOS VERDES
Independent
Variables
CONST
AGE215
LOTS I Z
AREA
BDROOM
BTHRUM
OTHRUM
SALDAT
POOL
ODDLOT
CORNLT
FRONT
CULDSC
CRFRNT
CLFRNT
APPL
AIRCON
VIEWLT
REMODL
FIREPL
GARAGE
FLOORS
AFLOOR
DOWNP
QUALTY
CONDN
LTSZPL
DPARK
DJUNK
MAINST
DFILL
WIND
DTRUCK
VIEWPL
R2
F
D.F.
Dependent Variable*
PRICE
-4119.08
2607.18
0.0014
0.1397
-7.0247
36.4577
0.1037
24.9603
13.3779
-0.0012
20.8133
-0.1778
3.7293
-0.0401
-0.2442
-0.9563
-21.5588
9.0710
9.5899
7.3473
-5.0673
-86.3110
0.0371
80.8308
9.9332
22.7120
0.0008
-0.2824
-0.0610
-22.6179
-0.5202
0.0128
1.6723
16.1407
0
120
(-6.09)*
(4.00)*
(1.77)
(12.91)*
(-1.55)
(4.71)*
(0.15)
(10.21)*
(1.74)
(12.14)*
(1.38)
(-1.66)
(0.20)
(-0.33)
(-0.78)
(-0.23)
(-1.24)
(1.80)
(1.04)
(1.16)
(-1.01)
(-1.24)
(1.38)
(2.70)*
(1.76)
(5.09)*
(1.67)
(-1.07)
(-0.91)
(-1.65)
(-3.19)*
(-.07)
(7.53)*
(2.18)*
.899
.794
413
ASSVAL
2470.91
65,976.3
0.2363
14.3407
-808.384
2078.66
29.7112
-834.302
3194.24
-0.2136
158.033
-30.7953
1391.79
19.9589
-45.6136
500.783
1272.81
663.728
-721.489
1290.04
326.642
1913.98
-2.0199
3971.53
524.130
383.273
-0.0803
-52.9937
0.3238
-607.202
-51,5351
26.3665
164.611
1951.22
0.
59.
(0.04)
(0.95)
(2.86)*
(12.40)*
(-1.67)
(2.51)*
(0.41)
(-3.71)*
(3.88)*
(-3.71)*
(0.10)
(-2.70)*
(0.71)
(1.52)
(-1.36)
(1.14)
(0.68)
(1.23)
(-0.74)
(1.91)
(0.61)
(0.26)
(-0.70)
(1.24)
(0.87)
(0.80)
(-1.61)
(-1.88)
(0.05)
(-0.42)
(-2.95)*
(1.34)
(7.03)*
(2.47)*
813
875
413
*The figures in-parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
398
-------
Table 6.6 NAME OF LANDFILL: PALOS VERDES
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
LOTHRM
SALDAT
VIEWLT
DOWNP
LCONDN
LDPARK
MAINST
LDFILL
LGWIND
LDTRUK
VIEWFL
LGQUAL
LGPOOL
LODDLT
LFIRPL
R2
F
D.F.
Dependent Variable*
LPRICE
-1.4182
4.7036
0.0568
0.4992
-0.0024
0.1405
-0.0002
0.0524
0.0283
0.1563
0.1200
-0.0284
-0.0677
-0.0688
-0.0080
0.0281
0.0277
0.0395
0.0810
(-4.24)*
(4.70)*
(5.06)*
(13.75)*
(-0.09)
(4.90)*
(-0.01)
(18.93)*
(3.19)*
(2.96)*
(5.46)*
(-2.46)*
(-2.81)*
(-5.30)*
(-0.25)
(6.53)*
(2.25)*
(1.56)
(6.73)*
0.900
225.161
428
LASSVAL
6.4364
1.4125
0.0612
0.5811
-0.0508
0.0784
-0.0171
0.0262
-0.402
0.1548
0.0361
0.0306
0.0835
-0.0209
0.0528
0.
123.
(12.77)*
(1.14)
(4.07)*
(13.51)*
(-1.51)
(2.15)*
(-4.02)*
(2.34)*
(-2.45)*
(4.71)*
(6.79)*
(1.95)
(5.42)*
(-3.13)*
(2.58)*
794
930
432
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
399
-------
Tujunga
This is a privately operated landfill using what was
at one time a sand and gravel pit. It is located on Tujunga
Avenue, only about two miles from the Sheldon-Arleta disposal
site. This site comprises 66 acres and handles approximately
370,000 tons a year of solid waste volume (Table 6.1).
Homes along two streets, Strathern and Irvine, have
a direct view of the landfill, but they also have a view of
other industry in the immediate vicinity. There are operating
gravel pits bordering on two sides of the landfill site.
Many of the homes in the vicinity of this site are old and
rundown. This site was started in 1962 and should be com-
pleted around 1978, at which time the owners will try to
sell the site to the State or County for use as a park.
The average sale price of the 279 sample residences
is $22,962 and the mean assessed value is $20,793 which are
the lowest of the four sanitary landfill areas studies
(Table 6.2). The average lot size, living area, number of
bedrooms and bathrooms are also the smallest as compared
to the other three neighborhoods. Although some homes along
two streets have a direct view of the landfill, none of the
sample residences had a direct view. In addition to this
variable, several others had to be excluded from the analysis
because of insufficient information. These were: Quality of
400
-------
construction, condition of maintenance/ number of floors,
down payment, distance to the nearest shopping center and
dummy variables for view lot, condominium and leasehold.
Tables 6.7 and 6.8 present the results for regressions
involving the dependent variables PRICE, ASSVAL, LPRICE and
LASSVL. The goodness of fit (R2) for PRICE and LPRICE was
much lower for Tujunga as compared to Palos Verdes. Only
about 66 to 68 percent of the variation in these dependent
variables could be explained by the independent variables
measured for residences in this neighborhood. The varia-
tion in the assessed value and its logarithm is better ac-
counted for with 81 to 82 percent explained by the model.
Table 6.9 presents the estimated coefficients and their
t-statistics after many of the insignificant variables have
been eliminated, with the exception of the variables repre-
senting the relationship between the property and the sanitary
landfill. As in the case of Palos Verdes, only the double
log formulation is used in interpretation for the reasons
given in Section C.
Among the fill related variables LDFILL, LDTRUK and
LGWIND, only the wind direction had a statistically signi-
ficant regression coefficient in the regression with LPRICE
as the dependent variable. In the case of assessed value,
none of the fill variables was significant. Even in the
401
-------
price equation, for which the wind direction had a significant
coefficient, the sign is counter intuitive. The sign was
negative implying that the farther away from the prevailing
wind direction from the fill the less the sale price of a
residence. A possible explanation for this unusual result
could be the fact that there is a gravel pit on the other
side which is currently in operation. Therefore, a property
away from the downwind from the landfill may be closer to
the wind from the gravel pit bringing dust, sand and noise.
In contrast, the landfill is sanitary and is hence less likely
to emit dust or odor, unlike the gravel pit. Home owners
may therefore be willing to pay a higher price to avoid the
wind from the gravel pit and hence be closer to the prevailing
wind from the landfill. Thus a house nearer the wind direction
from the sanitary landfill may carry a higher price tag.
Although the distances to the nearest freeway ramp and
the' nearest point of the freeway had positive coefficients
with low t-ratios, the joint F-test of these two variables
was quite significant. The F value of 8.34 was well above
even the 1% critical value of 4.61, indicating that these
two variables together have a significant effect on the
sale price of residential property. The evidence shows that
people are willing to pay a higher price to be away from a
freeway. A doubling of the distance to the nearest point
402
-------
of freeway would raise the sale price by an average of 16.8
percent (Table 6.9) whereas a doubling of the distance to
the nearest off ramp would raise the price by only 4.1 per-
cent. In contrast/ a doubling of the distance to the
sanitary landfill will raise the price by 3.9 percent which
is even lower.
As in the case of Palos Verdes we will also discuss
the effects on price and assessed value of the independent
variables other than those that refer to the sanitary landfill,
The effects on price are discussed first. A number of in-
dependent variables showed consistently insignificant co-
efficients no matter what the formulation was. Even joint
F-tests did not point out any significant effect of these
variables on price. They are: dummy variables for
corner lot, cul-de-sac and main street/ fireplace,
remodelling/ garage space/ distance to the neighborhood
park and any disamenities/ front footage and all the
interaction of terms. Among independent variables which
significantly explain the variation in prices, the age of
the residence is important even at the 1 percent level with,
as one might expect, newer houses having higher sale prices
as compared to older residences with the same characteristics.
The elasticities of the size of a lot and the living
area are also significant at the 1 percent level. As in
403
-------
Palos Verdes, home owners are willing to pay much more for
a doubling of the living area as compared to a doubling
of the lot size. A 100 percent increase in the living area
would increase the sale price by nearly 50 percent, whereas
a similar increase in the size of the lot would increase
the sale price by only 8.5 percent. When LYARD was used
instead of LLOTSZ, the elasticities became 45.2 percent
for area and 7.3 percent for size of the yard. Of the
variables LBDRM, LBTHRM and LOTHRM, only LBDRUM had a
significant t-statistic, and even then its sign was negative
indicating that the more the number of bedrooms/ keeping
all other variables constant/ the less the sale price.
This result is counter intuitive and hard to explain. A
possible reason could be that when the number of bedrooms
is increased keeping the living area the same, the size
of a bedroom becomes quite small, especially in a neighbor-
hood like Tujunga in which the average living area is only
1,150 square feet. This may make the addition of a bedroom
(with total area remaining the same) less desirable and
hence may decrease the sale price.
The date of sale is of high significance in explaining
the sale price of a residential property. As one would
expect, more recently sold homes carry higher price tags.
In the case of Palos Verdes, neither built-in appliances
404
-------
nor air conditioning significantly affected sale price. In
Tujunga, however, both these attributes seemed to matter.
Our estimates indicate that adding a range/oven or a dish-
washer to a residence in Tujunga with neither would raise
the expected sale price by 4.4 [=100(2-1}] percent
(Tables 6.2 and 6.9). Installing central air conditioning
in a home with no air conditioning would similarly produce
a 5.4 percent increase. Adding a non-heated pool to a
residence in Tujunga would raise the expected sale price
by 10.1 percent, versus only 5,8 percent in Palos Verdes.
The greater effect in Tujunga is not surprising, since the
average sale price there is less than half that in Palos
Verdes, and we would thus expect a pool to add a larger
percentage to the value of a typical home.
In the case of Tujunga also there were several variables
which affected the sale price significantly but had no
appreciable effect on the assessed value and vice versa.
As expected, the date of sale had a very high t-value (8.9)
in the LPRICE case but was insignificant for LASSVL. Proxi-
mity to the neighborhood park and a main street location
were important characteristics of a property influencing
tha assessed value, but not the sale price. The most
significant difference between the regressions for LPRICE
and LASSVL was in the effect of the POOL variable. Unlike
405
-------
the sale price, the elasticity of assessed value with
respect to POOL depends on the size of the lot as evidenced
by the significance of the interaction term LLOTPL. This
elasticity is given by -2.1742 + 0.2624 log(LOTSIZ). Thus
the bigger the size of the lot, the greater the percentage
increase in the assessed value attributable to the swimming
pool.
Another significant difference is in the fill variables
LDFILL, LDTRUK and LGWIND. None of these variables were
individually significant, nor did an F-test for their joint
significance show any significance. We can, therefore,
conclude that for the sample residences in the Tujunga area
the presence of the sanitary landfill had no significant
effect on their assessed values. Among the three variables
LBDRUM, LBTHRM and LOTHRM, only the number of bathrooms
had a significant regression coefficient. This is in
contrast to the result for the sale price for which the
number of bedrooms rather than the number of bathrooms
was an important determinant. The elasticity for the number
of bathrooms was 0.0989 which indicates that a 100 percent
increase in their number would raise the assessed value
•
by 9.9 percent (Table 6.9).
Unlike the sale price, the assessed value is signifi-
cantly influenced by distance to the neighborhood park
406
-------
and a main street location. The positive coefficient of
LDPARK, however, has the surprising implication that assessed
value rises with distance from the park. Being on a main
street reduces assessed value, by an average of 2.8 percent,
as we might expect. Adding a range/oven or dishwasher to a
home with neither would raise expected assessed value by
2,8 percent, even though expected sale price would increase
by 4.4 percent.
Finally, the distances to the nearest freeway exit and
the nearest point of the freeway were jointly (but not
individuallyI significant, with positive but rather different
coefficients. The elasticity of assessed value with respect
to distance to the off ramp was 9.Q percent, while for sale
price it was only 4.1 percent. In contrast, the elasticity
with, respect to the distance to the nearest point of the
freeway was 8.1 percent for assessed value and 16.8 percent
for sale price.
407
-------
Table 6.7 NAME OF LANDFILL: TUJUNGA
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
LOTHRM
SALDAT
LGPOOL
LODDLT
CORNLT
LFRONT
CULDSC
CRLFRN
CLLFRN
LGAPPL
LAIRCN
LREMOD
LFIRPL
LGARAG
LLOTPL
LDPARK
LDJUNK
MAINST
LDFILL
LGWIND
LDTRUK
LDFRWY
LDRAMP
R2
F
D.F.
Dependent Variable*
LPRICE
-0.7567
4.2851
0.0960
0.4000
-0.1110
0.0850
-0.0407
0.0389
-2.6004
-0.0170
0.6247
-0.0062
0.4945
-0.1228
-0.1324
0.0573
0.0496
-0.0107
0.0334
-0.0239
0.3087
-0.0138
-0.0317
-0.146
0.0464
-0.1309
-0.369
0.1348
0.1400
0
21
(-1.01)
(4.78)*
(2.11)*
(6.54)*
(-2.64)*
(2.07)*
(-1.17)
(5.91)*
(-1.63)
(-0.35)
(0.76)
(-0.10)
(0.74)
(-0.76)
(-0.72)
(2.65)*
(2.16)*
(-0.29)
(1.27)
(-1.00)
(1.72)
(-0.55)
( 1.38)
(-0.72)
(1.69)
(-2.04)*
(-0.96)
(1.13)
(1.18)
.675
.624
250
LASSVL
6.2719
3.7153
0.0435
0.3983
-0.0077
0.1100
-0.0288
0.0025
-2.2672
0.0187
-1.7309
0.0444
0.2513
9.3253
-0.0537
0.0375
0.0418
-0.0199
-0.0099
-0.0140
0.2731
0.0442
-0.0103
-0.0292
0.0197
-0.0856
-0.0545
0.0581
0.1337
0.
45.
(11.79)*
(6.09)*
(1.35)
(9.18)*
(-0.26)
(3.76)*
(-1.17)
(0.54)
(-2.01)*
(0.54)
(-2.96)*
(1.01)
(0.53)
(2.84)*
(-0.41)
(2.44)*
(2.57)*
(-0.76)
(-0.53)
(-0.83)
(2.14)*
(2.49)*
(-0.63)
{ 2.04)*
(1.01)
(-1.88)
(-2.00)*
(0.69)
(1.59)
818
689
250
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
408
-------
Table 6.8 NAME OF LANDFILL: TUJUNGA
Independent
Variables
CONST
AGE215
LOTSIZ
AREA
BDROOM
BTHRUM
OTHRUM
SALDAT
POOL
ODDLOT
CORNLT
FRONT
CULDSC
CRFRNT
CLFRNT
APPL
AIRCON
REMODL
FIREPL
GARAGE
LTSZPL
DPARK
DJUNK
MAINST
DFILL
WIND
DTRUCK
DFRWY
DRAMP
R2
F
D.F.
Dependent Variable*
PRICE
-1516.97
1156.26
-0.0067
0.0673
-7.7063
12.1749
-2.3774
8.1502
-28.7209
0.0013
19.4175
0.1495
34.6758
-0.2299
-0.7333
9.0320
9.3877
-0.6567
7.8643
-3.1240
0.0072
0.0845
-0.6136
-3.7207
0.2254
-0.1690
-0.2721
0.5158
0.7562
0
20
(-5.91)*
(4.94)*
(-2.09)*
(5.20)*
(-1.96)*
(1.72)
(-0.76)
(5.30)*
(-1.23)
(0.58)
(0.60)
(1.48)
(0.79)
(-1.12)
(-0.67)
(2.82)*
(2.83)*
(-0.12)
(1.78)
(-1.03)
(2.26)*
(0.35)
(-1.90)
(-0.76)
(0.80)
(-1.24)
(-0.70)
(0.77)
(1.18)
.661
.372
250
ASSVAL
-72,156.0 (-5.21)*
8'9,708.8 (6.49)*
-0.4552 (-2.30)*
6.6481 (8.38)*
-201.541 (-0.84)
1682.01 (3.88)*
-309.193 (-1.62)
-44.6159 (-0.48)
-1164.84 (--0.82)
0.0547 (0.38)
-7564.04 (-3.79)*
15.6751 (2.54)*
2920.82 (1.09)
30.5504 (2.44)*
-33.9963 (-0.51)
665.670 (3.39)*
708.609 (3.49)*
-235.915 (-0.69)
-1.0107 (-0.00)
-49.5687 (-0.27)
0.4739 (2.42)*
30.3960 (2.04)*
-40.9464 (-2.07)*
-513.560 (-1.72)
12.7101 (0.73)
-8.2178 (-0.99)
-43.8028 (-1.85)
-10.3803 (-0.25)
107.214 (2.78)*
0.824
47.456
250
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the. 5% level on a two-tailed test.
409
-------
Table 6.9 NAME OF LANDFILL: TUJUNGA
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
LOTHRM
SALDAT
LGAPPL
LAIRCN
LGPOOL
LDFILL
LGWIND
LDTRUK
LDFRWY
LDRAMP
LDPARK
MAINST
LLOTPL
R2
F
D.F.
Dependent Variable*
LPRICE
-0.7050
4.2180
0.0845
0.4297
-0.1015
0.0588
-0.0437
0.0362
0.0621
0.0467
0.1391
0.0392
-0.1231
-0.0308
0.1675
0.0406
0
40
(-1.24)
(5.00)*
(3.38)*
(7.54)*
(-2.50)*
(1.47)
(-1.29)
(8.88)*
(3.01)*
(2.09)*
(4.44)*
(1.56)
(-2.06)*
(-0.89)
(1.47)
(0.42)
.680
.309
263
LASSVL
6.4170
3.9695
0.0713
0.4096
-0.0163
0.0989
-0.0315
0.0398
0.0408
-2.1742
0.0070
-0.0793
-0.0428
0.0805
0.0898
0.0444
-0.0281
0.2624
0.
71.
(19.17)*
(6.70)*
(3.79)*
(10.02)*
(-0.56)
(3.56)*
(-1.29)
(2.63)*
(2.57)*
(-2.01)*
(0.38)
(-1.79)
(-1.63)
(1.00)
(1.25)
(2.54)*
(-1.97)*
(2.15)*
813
915
261
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test.
410
-------
Sheldon-Arleta
This municipal site is only 60 acres, but it handles
over 500,000 tons of solid waste per year (Table 6.1).
It is located at the intersection of the Hollywood and
Santa Ana Freeways, and thus there are sources of noise
unrelated to the landfill operation. The site is bordered
on two sides by freeways, one side by a school, and on the
fourth side by a residential area. There are a limited
number of homes along Sharp Street with a direct view of
the site, but because of the flat topography and the
elevated freeways on two sides of the site only those homes
directly adjacent to the site have any view. The residential
area is composed of older homes and is a middle class
neighborhood with an occasional rundown house on many of
the streets.
This site was started in 1961 and will be turned
into a park including Little League fields in about five
years. In fact, a small portion of the site has already
been turned into baseball fields.
Similar to Tujunga, this area also has low property
values. The average sale price of the 179 sample homes
was $24,097 and the average assessed value was $21,772.
The size of the lot is 7,382 on average and the mean value
of the living area is 1,228 square feet. The average
411
-------
number of bedrooms and bathrooms are respectively 2.9 and
1.4. The sample residences are not too far away from the
sanitary landfill, the average distance being only one-half
mile. However/ as mentioned earlier, there are freeways
nearby, the average distance being only 1,733 feet.
Tables 6.10, 6.11 and 6.12 present the results of
the regression analysis for the 179 sample residences.
Because of the lack of availability of data, a number of
independent variables had to be excluded from the analysis.
These were: view of the fill, distance to the shopping
center, down payment, dummy variable for a view lot, quality
of construction, and the dummy variables for leasehold
and condominium. The results indicate that the included
independent variables explain 80 percent of the variation
in price, 92 percent of the variation in assessed value,
85 percent of the variation in log (PRICE) and 92 percent
of the variation in log(ASSVAL).
Table 6.12 presents the regression results after
eliminating roost of the insignificant variables, but
retaining the variables associated with the sanitary land-
fill. None of the fill variables are statistically signi-
ficant in any of the formulations. This could be because
these variables are unimportant in explaining the variation
in price and assessed value, or because multicollinearity
412
-------
among these variables might be lowering the estimated
t-statistics. in order to exclude the latter possi-
bility, we carried out an F-test for the joint significance
of the fill variables LDFILL, LDTRUK and LGWIND. The value
of F was 0.9 for the dependent variable LPRICE and 2.2 for
the dependent variable LASSVL, neither of which are signi-
ficant because the critical value of F for the 5 percent
level of significance is 2.60. We therefore conclude
that in the case of Sheldon-Arleta neither the sale price
nor the assessed value of a property are significantly
affected by the presence of the sanitary landfill.
Although the variables associated with the fill were
not statistically significant, a number of other indepen-
dent variables do have substantial influences on the sale
price and assessed value. As in the other neighborhoods,
the age of a house is an important determinant of both
price and assessed value, with newer homes having higher
values. Similarly, the lot size had a substantial effect
on both price and assessed value. A 100 percent increase
in the size of the lot would have raised the average sale
price by 9.3 percent (Table 6.12). This is slightly
higher than that of Tujunga which, like Sheldon-Arleta, is
a low income neighborhood. The corresponding elasticity
for assessed value was 16.2 percent which is considerably
413
-------
higher than the elasticity for price. For the living area,
however/ the elasticities are nearly the same, with a
value of 38.3 percent for price and 36.5 percent for assessed
value. These figures are somewhat lower than those for
Tujunga. If the lot size is replaced by the size of the
yard, the price elasticity with respect to living area
becomes 39.0 percent and that for the yard size is 6.6
percent, well below that for lot size.
With LPRICE as the dependent variable, the log of
the number of bedrooms had a coefficient that was signi-
ficantly different from zero, while the log of the number
of bathrooms did not. The results for the assessed value
regression, however, were exactly the opposite. The number
of bathrooms had a significant regression coefficient,
but the number of bedrooms did not. The elasticities
were also quite different between price and assessed value.
For instance, a doubling of the number of bathrooms would
have increased the assessed value by 11.6 percent but price
by only 4.8 percent. In contrast, a doubling of the number
of bedrooms (keeping all other characteristics the same)
would have raised the average assessed value by only 3.0
percent whereas it would have increased the sale price by
as much as 9.1 percent.
Another interesting contrast between sale price and
414
-------
assessed value is the effect of SALDAT. While, as one
would expect, a recently sold house would carry a higher
price tag than a similar residence sold earlier, in the
case of assessed value the result was the opposite. Other
things being constant, a recently sold property has a
significantly smaller assessed value than a similar property
sold at an earlier date. The same result occurred in the
regressions for Palos Verdes.
As Table 6.12 shows, the distance to the park had a
negative coefficient. The corresponding t-statistic of -1.81,
though, indicates significance at only the 10% level on a
two^tailed test. However, if the insignificant landfill
variables are eliminated and the resultant equation is re-
estimated, this statistic rises above the 5% level in absolute
value. In other words, the Sheldon-Arleta residents seem
to be willing to pay a higher price for being closer to
the neighborhood park. In the case of assessed value this
independent variable has no significant effect. The number
of built-in appliances does affect both sale price and as-
sessed value substantially. The elasticity for price was
6.7 percent while that for assessed value was somewhat
less at 4.1 percent.
Air conditioning also has an appreciable influence on
price and assessed value, with nearly identical elasticities.
415
-------
The addition of central air conditioning to a home with no
air conditioning would raise expected price by 4..44 [=100
(2'0396_ijj percent and expected assessed value by 4.36
percent. While adding a two-car garage to a residence with
no garage would increase its expected sale price by 4.3
percent, the corresponding effect on expected assessed value
is not statistically significant.
As one would expect, better maintained houses tend to
Have higher prices and higher assessed values. The mag-
nitudes of the effects are quite different, however. For
example, raising the condition rating of a residence from
"average" to "excellent" would be expected to increase its
sale price by 6.5 percent, but to increase its assessed value
fcy only 3,2 percent. Similarly, having more than one floor,
all other things being equal, has substantial and significant
positive effects on both price and assessed value, but it
would raise the former by 15.5 percent and the latter by
11,5 percent, on average. On the other hand, the addition
of a non-cheated swimming pool would raise both expected sale
price and expected assessed value by about 10.1 percent.
Interestingly enough, while location on a main street
does not significantly affect the sale price of a typical
residential property, our estimates show that such a loca-
tion would serve to lower expected assessed value by nearly 5.5
416
-------
percent as compared to a similar property not located
on a main street. Similarly, a corner lot did not differ
much in the sale price but had a significantly smaller
assessed value. Although the absolute value of the
t-statistics given in Table 6.12 for CORNLT is not
above the 5 percent critical level, when the insigni-
ficant landfill variables were omitted, CORNLT had a
statistically significant negative coefficient. Location
in a cul-de-sac, however, had opposite results. It was
not a significant determinant of the assessed value
of a residence but has a significant positive effect
on sale price.
In both, the sale price and assessed value models,
both, distance to the nearest freeway off-ramp and distance
to the. nearest point on the freeway had positive coefficients
While the. coefficient of the former variable was not
significant at any reasonable level in either model, it
should be noted that these two quantities were very closely
correlated in this sample. Doubling the distance to the
nearest point on the freeway would raise the expected sale
price of a typical residence by 3.5 percent and its
expected assessed value by 2.9 percent. Our estimates
indicate clearly that residents and potential residents
of the Sheldon-Arleta area dislike living near the freeway.
417
-------
Table 6.10 NAME OP LANDFILL: SHELDON-ARLETA
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
LOTHRM
SALDAT
LGPOOL
LODDLT
CORNLT
LFRONT
CULDSC
CRLFRN
CLLFRN
LGAPPL
LAIRCN
LREMOD
LFIRPL
LGARAG
FLOORS
LARFLR
LCONDN
LLOTPL
LDPARK
LDJUNK
MAINST
LDFILL
LGWIND
LDTRUK
LDFRWY
LDRAMP
If*
F
D.F.
Dependent Variable*
LPRICE
0.7641
2.3404
0.1829
0.3523
0.1085
0.0581
0.0375
0.0072
1.8169
-0.1198
-0.3241
-0.0043
1.2170
0.0445
-0.3497
0.0517
0.0396
-0.251
-0.0384
0.0833
4.9893
-0.6625
0.1302
-0.1847
-0.0707
-0.0047
-0.399
0.0461
0.0406
-0.0160
0.0459
0.0562
0
32
(0.96)
(2.08)*
(1.20)
(4.73)*
(2.43)*
(1.27)
(0.90)
(2.21)*
(1.51)
(-0.59)
(-0.39)
(-0.02)
(2.68)*
(0.26)
(-2.77)*
(2.03)*
(2.15)*
(-0.63)
(-1.23)
(3.71)*
(1.78)
(-1.72)
(3.93)*
(-1.39)
(-2.09)*
(-0.08)
(-1.64)
(1.31)
(0.72)
(-0.46)
(1.82)
(1.10)
.847
.729
147
LASSVL
4.8975
3.0359
0.2032
0.3375
0.0412
0.1328
0.0120
-0.0052
3.8372
0.0808
1.1391
0.1040
0.7163
-0.2373
-0.2373
0.0348
0.0414
-0.0607
0.0131
-0.0152
2.0625
-0.2696
0.00667
-0.4097
-0.0265
0.0767
-0.0685
0.0146
0.0243
0.0147
0.0458
0.0159
0.
64.
(7.98)*
(3.72)*
(1.74)
(5.92)*
(1.20)
(3.78)*
(0.37)
(-1.93)
(4.17)*
(0.52)
(1.77)
(0.74)
(2.06)*
(-1.81)
(-1.81)
(1.78)
(2.93)*
(-2.00)*
(-.55)
(-0.90)
(0.96)
(-0.91)
(2.62)*
(-4.02)*
(-1.02)
(1.62)
(-3.66)*
(0.54)
(0.56)
(0.55)
(2.36)*
(0.41)
918
904
147
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test.
418
-------
Table 6.11 NAME OF LANDFILL: SHELDON-ARLETA
Independent
Variables
CONST
AGE215
LOTSIZ
AREA
BDROOM
BTHRUM
OTHRUM
SALDAT
POOL
ODDLOT
CORNLT
FRONT
CULDSC
CRFRNT
CLFRNT
APPL
AIRCON
REMODL
FIREPL
GARAGE
FLOORS
AFLOOR
CONDN
LTSZPL
DPARK
DJUNK
MAINST
DFILL
WIND
DTRUCK
DFRWY
DRAMP
R2
F
D.F.
Dependent Variable*
PRICE
-597.544 (-2.10)*
542.927 (1.92)
0.0108 (2.48)*
0.0718 (5.22)*
7.9098 (1.94)
7.4455 (0.95)
4.3414 (1.16)
1.7944 (2.38)*
57.4965 (2.98)*
-0.6711 (-1.30)
-16.0849 (-0.48)
-0.1714 (-0.46)
74.4602 (2.55)*
0.1094 (0.33)
-2.4731 (-2.60)*
6.6585 (1.85)
6.0026 (2.42)*
-4.3982 (-0.73)
-1.2352 (-0.24)
9.8802 (3.35)*
192.369 (2.19)*
-0.0967 (-1.65)
13.9752 (4.57)*
-0.0039 (-1.79)
-0.3362 (-0.54)
-1.3245 (-2.32)*
-10.8048 (-1.90)
0.0006 (0.00)
-0.0312 (-0.27)
-1.8221 (-2.66)*
1.3826 (2.39)*
0.1750 (0.25)
0.861
36.671
147
ASSVAL
-62,323.5 (-3.47)*
69,666.0 (3.50)*
1.4345 (4.49)*
5.9634 (5.883*
304.697 (1.01)
1939.73 (3.35)*
270.167 (0.98)
-96.5404 (-1.60)
7435.87 (5.23)*
6.5504 (0.17)
2374.43 (0.97)
2.8518 (0.11)
2223.87 (1.03)
-17.5588 (-0.72)
-86.3865 (-1.23)
434.785 (1.64)
511.450 (2.80)*
-976.798 (-2.20)*
360.091 (0.95)
-186.434 (-0.87)
6061.28 (0.94)
-2.3383 (-0.54)
752.211 (3.33)*
-0.6617 (-4.11)*
3.6574 (0.08)
-40.6818 (-0.97)
-1601.47 (-3,82)*
-6.6928 (-0.12)
-4.8122 (-0.57)
-84.8047 (-1.68)
125.574 (2.94)*
-20.5574 (-0.40)
0.915
63.087
147
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
-------
Table 6.12 NAME OF LANDFILL: SHELDON-ARLETA
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
CULDSC
SALDAT
LDPARK
CLLFRN
LGAPPL
LAIRCN
LGARAG
LCONDN
LGPOOL
FLOORS
LDFRWY
LDRAMP
LDFILL
LGWIND
LDTRUK
MAINST
CORNLT
-2
R
F
D.F.
Dependent Variable*
LPRICE
0.8515
2.3297
0.0931
0.3826
0.0909
0 0479
0.9307
0.0080
-0.0568
-0.2711
0.0672
0.0396
0.0609
0.1204
0.1382
0.1441
0.0500
0.0536
0.0300
0.0387
-0.105
49
(1.57)
(2.13)*
(1.99)*
(6.05)*
(2.26)*
(1.09)
(2.35)*
(2.51)*
(-1.81)
(-2.32)*
(2.76)*
(2.24)*
(3.37)*
(3.69)*
(4.13)*
(2.63)*
(2.17)*
(1.23)
(0.92)
(0.832)
(-0.481)
.844
.088
158
LASSVL
6.3022
3.3353
0.1616
0.3646
0.0299
0.1163
-0.0077
0.0414
0.0390
0.0620
0.1399
0.1091
0.0413
0.0312
-0.0164
0.0420
-0.0254
-0.0563
-0.0301
•
102.
(15.15)*
(4.35)*
(4.43)*
(7.34)*
(0.95)
(3.39)*
(-3.83)*
(2.20)*
(2.83)*
(2.45)*
(5.08)*
(2.60)*
(2.47)*
(1.16)
(-0.69)
(1.32)
(-1.55)
(-3.03)*
(-1.85)
911
494
160
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
420
-------
Calabasas
Although the Calabasas site is large in area,
approximately 380 acres/ it is currently handling
only about 260,000 tons of solid waste per year
(Table 6.1). The surrounding area is still rather
sparsely populated, thus the site will probably
gain volume as the local population density increases.
This site is located in the city of Calabasas, about
one mile north of the Ventura Freeway (Figure 6.1).
The topography of the area is predominantly flat with
some gentle hills surrounding the site itself. Thus
there are no homes with a direct view of the landfill.
There is, however, only one access road to the site,
Lost Hills Road, which all the solid waste disposal
vehicles must use. This road runs along the boundary
of a new housing tract which has been built between
the landfill site and the freeway. The tract includes
homes in the $40,000 to $60,000 price range,well
maintained and landscaped by the residents, making
this a very appealing residential setting. This
County operated landfill was started in 1957 and is
scheduled for completion sometime around 1985; it
will then be made into a park or golf course.
The average sale price of the 61 sample resi-
421
-------
dences was $44,800 with an assessessed value of
$38,785. Thus this area is second in house values
among the four sanitary landfills studies. However,
in living area and the average number of bedrooms,
bath rooms and other rooms, this tract was the largest
of the four areas. The average lot size was 9,994
sg. feet, well below the 12,908 value for Palos
Verdes but on the average these have 4.0 bedrooms,
2.7 bath rooms and 3.7 other rooms. In contrast,
homes in the Palos Verdes area have 3.4 bedrooms,
2.1 bath rooms and 3.3 other rooms.
Tables 6.13, 6.14 and 6.15 present the results
of the regression analysis using LPRICE, LASSVL,
PRICE, and ASSVAL as dependent variables. The models
used here explain 90 to 92 percent of the variation
in log(PRICE) and log (ASSVAL), and 86 to 90 percent
of the variation in PRICE and ASSVAL. Table 6.15
presents the coefficients and t-statistics for the
double log formulations, after most of the insig-
nificant variables, with the exception of the inde-
pendent variables associated with the sanitary landfill,
have been eliminated. The results of this formulation
form the basis for the discussion that follows.
None of the independent variables associated
422
-------
with the sanitary landfill have significant coefficients
when LPRICE is the dependent variable. As this may be
attributable to multicollinearity involving the fill-
related variables, a joint F test of these quantities
(LDFILL, LDTRUK, and LGWIND) was performed. The
F-statistic was l.l/ well below the critical value at the
5 percent level. We therefore conclude that as
far as the sale price is concerned, the presence of
the Calabasas sanitary landfill has no statistically
significant Cpositive or negative) effect. In the case
of assessed value, however, there is a significant
effect. For instance, the distance to the fill
variable LDFILL has a significant (at 1 percent
confidence level) negative coefficient, implying that
the farther away from the fill, the less the assessed
value of a residence. A doubling of this distance
would lower the assessed value fay nearly 22 percent.
The corresponding coefficient for price also was
negative but not significant at the 5 percent level
Csignifleant, however, at the 10 percent level). The
distance to the truck route and the number of degrees
away from the prevailing downwind from the fill are
not significant Ceven at the 10 percent level) either
for price or assessed value.
423
-------
Unlike any of the other neighborhoods, in
Calabasas, the lot size variable has no significant
influence on either the price or the assessed values/
the reason for which is not clear. The living area,
however, is very highly significant. A 100 percent
increase in the living area would raise the sale price
by 48.7 percent (Table 6.15) and the assessed value
by 64.4 percent. Although the number of bedrooms
and bath rooms did not show significant effects on the
sale price, a joint F-test gives an F-value of 3.86
which is significant at the 5 percent level. The
elasticities for sale price with respect to the
number of bedrooms and bath rooms are, respectively,
9.5 percent and 6.9 percent. These variables, how-
ever, do not have a significant effect on the assessed
value, either individually or jointly (the F-value
was only 0.4). The number of other rooms was
statistically significant for both price and assessed
value with positive coefficients. A doubling of the
number of other rooms would increase the sale price
25.7 percent whereas they would raise the assessed
value by only 11.5 percent.
The presence of a swimming pool also has
substantial effects on price and assessed value.
424
-------
Residences with an unheated pool would have a higher
price as well as assessed value than homes without a
swimming pool, and homes with a heated pool will
have even higher values. The respective elasticities
are 15.7 percent for price and 9.7 percent for
assessed value. A corner lot location also had a
significant positive effect on the sale price, but not
on the assessed value. A residence in a corner lot
would have, on the average, a sale price about 10
to 10.5 percent higher than a similar one not located
on a corner lot. The quality of construction and
the condition of maintenance are two other independent
variables which had significant regression coefficients
in the case of the sale price but not in the case
of the assessed value. The elasticity of price with
respect to quality of construction was 11.5 percent
while that with respect to condition of maintenance
was 20.5 percent.
A number of independent variables were excluded
from the analysis for Calabasas because the relevant
information was not available. These were: view of
landfill, distances to the neighborhood park, shopping
center and any disamenities, number of appliances,
remodeling, and dummy variables for view lot, main
425
-------
street location, cul-de-sac/ condominium and lease-
hold.
426
-------
Table 6.13 NAME OF LANDFILL: CALABASAS
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
LOTHRM
SALDAT
LGPOOL
LODDLT
CORNLT
LFRONT
CRLFRN
LGAPPL
LAIRCN
LFIRPL
LGARAG
FLOORS
LGQUAL
LCONDN
LLOTPL
DOWNP
LDFILL
LGWIND
LDTRUK
LDFRWY
LDRAMP
R2
F
D.F.
Dependent Variable*
LPRICE
3.2965
40.4265
-0.9168
0.4381
0.0473
0.2356
0.2414
0.385
22.3124
1.2771
10.8827
0.8838
-2.0967
0.1322
-0.0399
-0.1281
-0.0352
-0.0333
0.0356
0.0714
-2.3781
0.1764
-0.3129
-0.0991
-0.0023
-0.1295
-0.0772
0
26
(1.30)
(2.58)*
(-2.10)*
(3.85)*
(0.47)
(2.89)*
(2.90)*
(2.20)*
(1.76)
(2.11)*
(1.88)
(1.75)
(-1.81)
(0.66)
(-0.73)
(-1.87)
(-0.89)
(-0.80)
(0.70)
(0.61)
(-1.74)
(1.48)
(-0.96)
(-0.60)
(-0.09)
(-1.08)
(-0.64)
.918
.700
34
LASSVL
10.2051
40.2258
-0.7921
0.6456
-0.0950
0.1432
0.1200
-0.0161
20.0601
1.1065
8.4235
0.667
-1.606
-0.0642
-0.0038
-0.0707
0.0230
0.0146
0.0502
-0.0695
-2.1467
0.2385
-0.2730
-0.0244
-0.0047
-0.1555
0.0950
0
20
(2.40)*
(2.08)*
(-1.45)
(4.54)*
(-0.75)
(1.40)
(1.15)
(-1.59)
(1.28)
(1.46)
(1.16)
(1.06)
(-1.11)
(-0.25)
(-0.06)
(-0.82)
(0.46)
(0.28)
(0.79)
(-0.47)
(-1.28)
(1.60)
(-0.67)
(-0.12)
(-0.15)
(-1.04)
(0.63)
.894
.512
34
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test
427
-------
Table 6.14 NAME OF LANDFILL: CALABASAS
Independent
Variables
CONST
AGE215
LOTSIZ
AREA
BDROOM
BTHRUM
OTHRUM
SALDAT
POOL
ODDLOT
CORNLT
FRONT
CRFRNT
APPL
AIRCON
FIREPL
GARAGE
FLOORS
QUALTY
CONDN
LTSZPL
DOWNP
DFILL
WIND
DTRUCK
DFRWY
DRAMP
R2
F
D.F.
Dependent Variable*
PRICE
-11840.2
11499.9
-0.0061
0.0771
16.7044
23.3779
25.7683
8.3192
146.594
5.3481
489.269
1.9524
-3.2130
-0.2298
-8.8553
-8.6902
-14.7552
-3.8574
15.7738
17.1163
-0.0084
90.8317
-0.4387
0.4848
-0.8700
2.2921
-3.9030
0.
21.
34
(-1.99)
(2.02)
(-0.22)
(3.32)*
(1.20)
(1.55)
(1.96)
(1.24)
(0.50)
(2.07)*
(1.73)
(1.53)
(-1.49)
(-0.01)
(-0.65)
(-0.53)
(-1.57)
(-0.20)
(1.53)
(1.12)
(-0.31)
(1.53)
(-0.06)
(0.30)
(-0.18)
(0.25)
(-0.77)
900
710
ASSVAL
-1,197,980
1,305,940
1.7918
11.0596
-1008.37
1201.33
1755.08
-618.240
31,298.3
297.505
11,398.0
27.5307
-27.6338
-2269.19
189.335
-717.650
230.316
939.046
724.162
-17.5375
-2.7202
9287.13
-659.647
44.5255
-263.024
-958.192
642.157
(-2.01)
(2.12)*
(0.59)
(4.36)*
(-0.66)
(0.73)
(1.22)
(-1.40)
(0.99)
(1.06)
(0.37)
(0.20)
(-0.12)
(-0.59)
(0.13)
(-0.40)
(0.23)
(0.44)
(0.64)
(-0.01)
(-0.92)
(1.43)
(-0.85)
(0.25)
(-0.49)
(-0.96)
(1.17)
0.868
16.116
34
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
428
-------
Table 6.15 NAME OF LANDFILL: CALABASAS
Independent
Variables
CONST
LAG215
LLOTSZ
LGAREA
LBDRUM
LBTHRM
LOTHRM
LGPOOL
LBQUAL
LCONDN
CORNLOT
LDFILL
LDTRUK
LGWIND
R2
F
D.F.
Dependent Variable*
LPRICE
2.1593
11.8274
-0.0132
0.4873
0.0947
0.0686
0.2565
0.1571
0.1151
0.2053
0.0987
-0.1353
-0.0228
0.1262
0
42
(2.25)*
(2.70)*
(-0.30)
(5.76)*
(1.43)
(1.21)
(5.36)*
(3.51)*
(2.74)*
(2.50)*
(1.95)
(-1.73)
(-0.95)
(0.98)
.901
.939
47
LASSVL
6.3815
16.1180
-0.0101
0.6444
0.0583
-0.0029
0.1147
0.0967
-0.2199
-0.0173
0.1592
0
49
(5.72)*
(3.04)*
(-0.21)
(6.82)*
(0.85)
(-0.05)
(2.08)*
(2.11)*
(-2.89)*
(-0.62)
(1.10)
.890
.755
50
*The figures in parentheses are t-statistics. Asterisks
indicate significance at the 5% level on a two-tailed test,
429
-------
E. SUMMARY OF RESULTS
In this section we summarize the results ob-
tained from, the analysis of data for residential
areas in the proximity of sanitary landfills.
Regression models designed to explain variation in
property values were estimated separately for each
fill area. Four explanatory variables were included
in the model in order to account for the effects of
each fill on the prices and assessed values of nearby
residential properties. The estimated coefficients
for these fill variables from the regressions in
double log form are summarized in Tables 6.16 and
6.17.
Table 6.16 ELASTICITY OF PRICE WITH RESPECT
TO FILL VARIABLES
PalosSheldon-
Variable Verde Tijunga Arleta Calabasas
Distance from fill -.07* .04 .03 -.14
Direct view of fill .03* NA NA NA
Distance from truck route .03* -.03 -.01 -.02
Degrees away from downwind -.01 -.12* .03 .13
* Significantly different from zero at 95 percent
confidence level.
430
-------
Table 6.17 ELASTICITY OF ASSESSED VALUE
WITH RESPECT TO FILL VARIABLES
Palos Sheldon-
Variable Verde Tijunga Arleta Calabasas
Distance from fill -.04* .01 -.02 -.22*
Direct view of fill .03 NA NA NA
Distance from truck route .04* -.04 -.03 -.02
Degrees away from downwind .15* -.08 .04 .16
* Significantly different from zero at 95 percent
confidence level.
The general conclusion suggested by these results
is that these four sanitary landfills have not had
significant detrimental effects on surrounding property
values. In the cases of Tijunga, Sheldon-Arleta and
Calabasas there were no significant coefficients
which were consistent with the hypothesis that sanitary
landfills have harmful environmental impacts. In
the case of Palos Verde, proximity to the truck route
was associated with lower property values, but this
negative impact seems to be dominated by the positive
impacts of being near to and having a view of the
fill which results/ presumably, from the anticipation
of the future use of the fill site for recreational
purposes.
Further insight into these findings may be gained by
431
-------
recalling some of the discussions in earlier chapters.
As Chapters II and V stressed, the basic hypothesis under-
lying studies of this sort is that if external effects are
properly perceived and understood, they will be reflected
in the rental values of impacted properties. Since, as
Chapter IV has stressed, the matter of ground water pol-
lution is quite complex, it would not be reasonable to
expect a property value study to capture related external
damages, even if they are in fact important.
Further, available data provide information on property
values, not on explicit or implicit rents. These values
are logically viewed as having been obtained by capitalizing
the relevant streams of benefits or (explicit or implicit)
rents at some discount rate. These facts have two main
implications. First, if one wants to derive a dollar
measure of damages, it is necessary to consider the rela-
tionship between this discount rate and the socially appro-
priate discount rate. Since the present study is mainly
interested in whether damages in fact exist, this issue
does not arise here. Second, since we can only observe,
at best, the present value of the changes in benefit
streams brought about by the proximity of the landfill,
it is difficult or impossible to infer anything about the
432
-------
change in benefits at any single point in time. This issue
arises in connection with Palos Verdes, where we argued
that proximity to the disposal site presumably lowers bene-
fits for a few future years and raises them thereafter.
Our results suggest that the resultant change in present
value is positive, but they can say nothing about the
absolute magnitude of current costs or future benefits.
Further, as Chapter II stressed, even if these problems
did not exist, estimates such as ours could not be used by
themselves to compute dollar values of landfill-related
externalities. In theory, any change in the relative
attractiveness of various parcels of land will bring about
shifts in the pattern of land use throughout the relevant
jnarket area. In order to derive a defensible measure of
external costs, these shifts must be predicted. While
approximations and shortcuts do exist, they must be used
with, extreme care.
We are aware of only two other attempts to evaluate
the external effects of solid waste disposal sites; both
were discussed in Chapter III. The study by Goldberg, et
al J6J takes a very different approach from ours. Their
use of control neighborhoods is questionable on empirical
grounds, and their focus on changes- in property values
433
-------
over time has no theoretical justification.
The paper by Havlicek, Richardson, and Davies [5],
hereafter referred to as HRD, is much closer in spirit to
our study. Both relate various attributes of residential
properties, including proximity to solid waste disposal sites,
to property values. Because of data limitations, neither
is able to investigate the Anderson-Crocker [1] contention
that owner characteristics, in particular income, belong
in such regressions. There are a number of important
differences between these studies, however.
First, HRD used a linear specification throughout.
We have discussed at length above the shortcomings of this
specification, and we regard our log-linear model as more
sensible. The weakness of the linear specification is
especially clear with respect to the "distance from the
site" variable. According to the HRD estimates, each foot
one moves away from the disposal site adds the same dollar
amount to property value, regardless of the original distance
or of any characteristics of the property. This is surely
unreasonable. Another weakness in their specification is
their treatment of the age variable. Even though they
cite Emerson I4J , HRD make no use of his analysis of this
quantity, and age enters linearly into their regressions.
Second, our analysis had the advantage of a much richer
434
-------
data base. The HRD sample consisted of 182 residences,
while ours included a total of 966 homes. While HRD consi-
dered only 12 attributes of individual properties, we analyzed
over 40. The fact that we had more information suggests
that our results may be more reliable. It should be under-
stood, however, that, strictly speaking, our estimates
pertain only to the sites we actually studied. Solid
waste disposal sites differ enormously, and our findings
clearly cannot be asserted to apply to all sites everywhere
without considerable additional research.
Finally, while neither study employed data on neighbor-
hood characteristics, the problem of differences among dis-
posal sites and their surrounding neighborhoods was handled
very differently. HRD simply pooled all their observations,
even though they related to five different sites, and, thus
five different neighborhoods. They allowed intercepts but
not slopes to differ among the five groups. In particular,
the external effects of all five disposal sites were assumed
identical on the margin. Our analysis, on the other hand,
treated each landfill separately. All the homes considered
at any one time belonged to the same general neighborhood
and were affected, if at all, by the same disposal site.
There was thus no need to control for characteristics of
individual neighborhoods or landfills. Our regression
435
-------
results confirmed our initial suspicions: individual
variables did indeed have different impacts on property
values in different neighborhoods. This suggests that/ at
least in our sample, pooling of the HRD sort would have
been quite inappropriate. Whether this is true for their
sample and whether their results are accordingly biased,
we have no way of knowing.
The foregoing comparisons should not be interpreted
as suggesting that the HRD effort was worthless. Theirs
was truly a pioneering study, and we learned a great deal
from careful consideration of its strengths and weaknesses.
Still, we feel safe in asserting that the research reported
here represents an advance over the HRD study.
The regression models we employed were specified after
considerable theoretical analysis, and they explained high
proportions of the variance in observed property values.
For the most part the explanatory variables had coefficients
that were consistent with the predictions of economic
theory, and some of the landfill-related variables had
significant and explicable coefficients. These results
support the contention that the external effects that can
be perceived by actual and potential property owners will
be reflected in market values and can be evaluated by fit-
ting a properly specified econometric model to data on the
436
-------
values and characteristics of affected properties.
Obviously, as a general matter, more work needs to
be done on modelling the way in which property character-
istics determine property values. In particular, a good
case can be made for the explicit use of nonlinear estimation
techniques in such models. Further, additional research
is needed to gather and evaluate data on solid waste dis-
posal site characteristics that are likely to reflect
external effects. This study indicates that such efforts,
if done with care, are likely to yield meaningful results.
437
-------
REFERENCES
1. Anderson R.J., and R.D. Crocker. Air Pollution and
Residential Property Values. Urban Studies. Vol. 8
(October 1971): p. 171-180.
2. City of Los Angeles Bureau of Sanitation. Liquid
and Solid Waste: Collection/ Treatment, Disposal.
3. County Sanitation Districts of Los Angeles County.
Existing Landfills in Metropolitan Los Angeles
County.
4. Emerson, F.C. The Determinants of Residential
Value with Special Reference to the Effects of
Aircraft Nuisance and Other Environmental Features.
University of Minnesota, Ph.D. Dissertation, 1970.
5. Havlicek, J., R. Richardson, and L. Davies. Measuring
the Impacts of Solid Waste Disposal Site Location on
Property Values. (Unpublished, c. 1972.)
6. Goldberg, L., et al. The Effects of Solid Waste Disposal
Sites on Property Values. Environmental Protection
Agency, Washington, D.C. 1972.
7, Parkhurst, J.D. An Introduction to the Sanitatioji
Districts of Los Angeles County.
438
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/5-75-010
2.
3. RECIPIENT'S ACCESSION NO.
». TITLE AND SUBTITLE
Measuring External Effects of Solid Waste
Management
5. REPORT DATE
March 1975
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Richard Schmalensee, Ramachandra Ramathan, Wolfhard
Ramm, Dennis Smallwood
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Institute for Policy Analysis
8961 Nottingham Place
La Jolla, California 92037
10. PROGRAM ELEMENT NO.
1DB314 R/T 02 AAE 06
11. CONTRACT/GRANT NO.
Grant No. R-801673
12. SPONSORING AGENCY NAME AND ADDRESS
Resource Analysis Staff
Washington Environmental Research Center
U.S. Environmental Protection Agency
Washington, D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This study analyzes the environmental impact of using sanitary landfills for
disposing of solid waste. The relevant economic theory concerned with the measure-
ment and valuation of external effects is developed and previous empirical impacts
of various activities are reviewed. Both the theoretical and empirical evidence
suggests that the costs and benefits of external effects are extremely difficult to
measure directly but under certain circumstances property value studies can be used
to obtain indirect estimates.
A survey of the technology of sanitary landfills suggests that a properly
designed fill will cause very little air and water pollution, but may impose visual
and noise pollution on nearby residents. These hypotheses are tested with data on
property surrounding four sanitary landfills in Los Angeles County. A model of the
determinants of residential property values is formulated and estimated. The model
includes variables which describe the characteristics of the property and variables
which describe the relationship of the property to the fill. Statistical estimates
of the parameters of the model indicate that proximity to and view of a sanitary
landfill do not significantly reduce the market or assessed value of surrounding
property.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
external effects, environmental damages,
property values, sanitary landfills,
solid waste, economics.
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
448
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |