3EPA
United States
Environmental Protection
Agency
Municipal Environmental Research EPA-600/9-80-001
Laboratory May 1980
Cincinnati OH 45268
Research and Development
Proceedings of a
Symposium on
Economic
Approaches to Solid
Waste Management
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology, Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3, Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8 "Special" Reports
9. Miscellaneous Reports
This document is available to the public through the National Technical Informa*
tion Service, Springfield, Virginia 22161.
-------
X)
EPA-600/9-80-001
May 1980
PROCEEDINGS OF A SYMPOSIUM
ON ECONOMIC APPROACHES
TO SOLID WASTE MANAGEMENT
by
Robert J. Anderson, Jr.
MATHTECH, Inc.
Princeton, New Jersey 08540
Contract No. 68-03-2673
Project Officer
Oscar W. Albrecht
Solid and Hazardous Waste Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U, S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem solu-
tion and it involves defining the problem, measuring its impact, and search-
ing for solutions. The Municipal Environmental Research Laboratory develops
new and improved technology and systems for the prevention, treatment, and
management of wastewater and solid and hazardous waste pollutant discharges
from municipal and community sources, for the preservation and treatment of
public drinking water supplies, and to minimize the adverse economic, social,
health, and aesthetic effects of pollution. This publication is one of the
products of that research; a most vital communications link between the re-
searcher and the user community.
On September 19 and 20, 1978, a group of economists interested in
finding solutions to the solid waste problem met at Philadelphia, Pennsylvania.
The express purpose of the meeting was to exchange ideas and views on the
potential role of pricing in municipal solid waste management. This report
contains the formal papers given at that Symposium.
Francis T. Mayo
Director
Municipal Environmental
Research Laboratory
iii
-------
ABSTRACT
This report contains the papers presented at a Symposium on Economic
Approaches to Solid Waste Management at Philadelphia, Pennsylvania, on
September 19 and 20, 1978. The symposium was sponsored by the Municipal
Environmental Research Laboratory of the U.S. Environmental Protection
Agency. Its purpose was to review the state of the art in user charges and
product charges for non-hazardous municipal solid waste, and to provide a
forum for researchers to exchange their views on methods and results in the
field.
To aid in its assessment of the state of the art, the symposium
participants considered four key questions:
1. Do incentive systems designed to alter waste
generation behavior actually improve the economic
efficiency of solid waste management and/or the
equity of its financing?
2. What is the optimal combination of incentive
mechanisms? For example, would the best system
include some mix of a user charge and a product
charge?
3. How do incentives relate to the choice of resource
recovery technology?
4. What are the best ways to pursue the testing of the
effects of incentive mechanisms? Should we use
historical data? Should we try to conduct
experiments?
The report was submitted to MATHTECH, Inc., as a subcontractor to
JACA Corp., JACA Contract 68-03-2673 under the sponsorship of the U.S.
Environmental Protection Agency.
IV
-------
CONTENTS
Foreword iii
Abs tract iv
Section 1. Introduction and Summary 1
"Introduction and Summary" (Robert J. Anderson, Jr.) .... 2
"Can Economic Incentives Help Resolve Solid Waste
Management Problems? (Haynes C. Goddard) 14
Section 2. Pricing and the Efficiency of Solid Waste Management ... 17
"The Economic Efficiency of User Fees: Some
Preliminary Empirical Results" (William N. Lanen) 18
"Evaluating the Efficiency of the Solid Waste
Charge" (Steve Buchanan) 47
"Economic Approaches to Solid Waste Management:
European Experience" (W. David Conn) 66
"The Administrative Costs of User Charges"
(Barbara J. Stevens) 75
Section 3. Pricing and the Equity of Solid Waste Financing 92
"Equity Considerations in User Charges for Residential
Refuse Collection and Disposal" (Roger Bolton) 93
"Pricing Municipal Refuse Service: Potential Effects
on Municipal Budgets" (Kenneth L. Wertz) 114
Section 4. Comparison of User Charges and Product Charges 120
"The Optimal Policy Mix for Solid Waste Pricing"
(James F. Hudson) 121
"Fundamental Comparisons of Alternative Solid Waste
Management Policies" (Allen K. Miedema) 130
-------
Section 5. Pricing and Resource Recovery Technology 155
"Conditionally Predictive Estimates of
Secondary Materials Supply" (Tayler H. Bingham,
Curtis E. Youngblood, Philip C. Cooley) 156
"Secondary Material Demand and Supply Responses"
(Robert C. Anderson) 189
Section 6. Pricing and Solid Waste Management 205
"The Seattle Solid Waste Management Experiment: User
Charges and Separate Collection of Recyclables"
(Bernard H. Booms) 206
Section 7. Research Directions 231
"Research Activities in EPA" (Oscar W. Albrecht) 232
"Solid Waste Collection/Disposal Economic Research
Strategies" (G.S. Tolley, V.S. Hastings) 240
Appendix A. Agenda « 248
Appendix B. List of Participants 249
vi
-------
SECTION 1
INTRODUCTION AND SUMMARY
"Introduction and Summary" (Robert J. Anderson, Jr.)«
"Can Economic Incentives Help Resolve Solid Waste Management Problems?
(Haynes C. Goddard).
-------
INTRODUCTION AND SUMMARY
by
ROBERT J. ANDERSON, JR.
INTRODUCTION
In late spring of 1978, Oscar Al brecht and Haynes Goddard of the
Municipal Environental Research Laboratory, U. S. Environmental Protection
Agency began to make tentative plans for a workshop on the role of pricing in
municipal solid waste management. Al brecht and Goddard, both of whom are
resource economists actively studying pricing policies for residential solid
waste management, felt that the pace of research on the subject had so
quickened over the last few years that the time was ripe for researchers
working in the area to come together. The objectives of so doing were to
define the current state of knowledge concerning the role of pricing policies
in solid waste management and to identify research needs for the coming
years.
To host the workshop, EPA contracted with the JACA Corporation which
subcontracted with MATHTECH to organize the workshop sessions and to prepare
the workshop report. The workshop was organized into five sessions at which
prepared papers were delivered and discussed. The agenda for the workshop is
contained in an appendix to this volume.
The workshop tasks, as noted above, were i) to identify the current
state of knowledge concerning the role of pricing in solid waste management,
and ii) to identify research needs in this area. In an effort to accomplish
this, these tasks were narrowed in two important ways. First, the workshop
considered only non-hazardous wastes. Economists generally are in agreement
that there is little or no role for pricing policies in management of wastes
which, if mismanaged, can result in catastrophy.
Second, the workshop participants organized their papers and
discussions around two of the most widely-discussed and intensely-researched
classes of pricing policies, "user charges" and "product charges". The
difference between these two approaches lies in the point at which the price
is levied. Under the product charge, the price is levied directly upon
-------
producers of goods which, when consumed, generate solid wastes. Under the
user charge, the price is levied directly upon consumers as they dispose
their wastes.
It was clear from the outset of the workshop that Albrecht and
Goddard were right about the need for researchers to meet. The workshop
discussions were pointed at some times heated and researchers clearly
needed the opportunity afforded by this workshop to exchange ideas and argue
out points of difference. Indeed, several of the discussions begun at the
workshop have continued afterward.
This report contains the papers and a summary of the discussions that
took place at the workshop. Even a casual examination of its contents will
show that a great deal of progress has been made in understanding the effects
of pricing policies on the cost and effectiveness of solid waste management.
In particular, the papers and discussions in the pages to follow contain
evidence that pricing policies can improve the cost and effectiveness of
solid waste management, although the gains over present policies that do not
incorporate pricing approaches seem to be relatively small given today's
economics. Tomorrow, however, the story may be quite different. If landfill
costs and the prices of virgin materials exhibit the marked relative
increases which many observers forecast, pricing policies could result in a
substantial improvement in the efficiency of solid waste management.
The following pages will also show that our knowledge is relatively
weak on what might be called the "political economy" of pricing approaches to
solid waste management. Workshop participants were unanimous in their call
for expanded effort in the study of administrative, political, and legal
factors bearing upon the workability of pricing approaches.
SUMMARY OF PAPERS
Charge to the Workshop
The charge to the workshop was delivered in an opening address by
Haynes Goddard. Pointing out that economic approaches to solid waste
management have failed for the most part to grasp the attention of those
responsible for making and implementing policy, Goddard laid much of the
blame squarely on economists. Our faults, Goddard noted, are an excessive
preoccupation with theory as opposed to practice, and a failure to deal
effectively and convincingly with difficult (but solvable) empirical problems
in our applied work.
As a point of departure for correcting these faults, Goddard
recommended that the economics profession take stock of current and needed
knowledge. To this end, Goddard asked workshop participants to consider four
key questions:
-------
1. Do incentive systems designed to alter waste generation
behavior actually improve the economic efficiency of
solid waste management and/or the equity of its
financing?
2. What is the optimal combination of incentive mechanisms?
For example, would the best system include some mix of a
user charge and a product charge?
3. How do incentives relate to the choice of resource
recovery technol ogy?
A. What are the best ways to pursue the testing of the
effects of incentive mechanisms? Should we use
historical data? Should we try to conduct experiments?
As will become clear below, the papers presented at the workshop and
the discussions that surrounded them all related to one or more of these
questions. Accordingly, I have chosen to organize my summary of the papers
presented and workshop discussions by the questions to which they pertain.
Pricing and the Efficiency of Solid Waste Management
Four of the papers presented pertained directly to the question
raised by Goddard concerning the effects of incentives of the efficiency of
solid waste management. These papers included those by William Lanen, Steve
Buchanan, W. David Conn, and Barbara Stevens.
In a paper on the effect of user charges on the efficiency of solid
waste management, Lanen presented the results of case studies of user charges
in five municipalities. In general, Lanen's results showed no statistically
significant effect of user charges on the quantity of waste presented for
collection, the usage of special services, or the degree of littering. He
did not examine the effect of user charges on costs of collection or
administration of solid solid waste systems.
Lanen cautioned that his failure to find significant relationships
between user charges and household solid waste behavior does not mean that
there is no relationship. Rather, he concluded that, due to data and
specificaton problems, the historical case study approach is not particularly
well-suited to measuring any relationships which may exist. His results thus
leave open the question concerning the effect of user charges on solid waste
management efficiency.
Buchanan's paper focused on efficiency aspects of product charge
systems. Analyzing data compiled in studies carried out for the Resource
Conservation Committee, Buchanan determined that the effect of a product
charge on national solid waste flows is likely to be rather small, with
estimated resource savings of between $25 million and $60 million per year.
Buchanan points out, however, that the magnitude of net savings actually
realized may be very sensitive to picking the correct product charge rate.
At rates other than the correct rate, resource savings may fall off very
-------
rapidly, and even become negative. When the additional costs of
administering a product charge are considered, Buchanan concludes that we
don't know whether or not a product charge would contribute to more efficient
solid waste management. He is quick to add, however, that the relevant
question is not how the product charge approach compares with the theoretical
ideal charge, but rather how it compares with what we are doing now, and with
other practical alternative approaches to solid waste management.
Conn's paper, like Lanen's, is based upon case studies. Conn
examined the experience of a number of European countries and municipalities
with various forms of user and/or product charge. He found that it is
extremely difficult to get the kind of hard data that is needed to make a
fair appraisal of the workings of the various incentive systems that have
been tr.ied. As a result, Conn's data consist almost entirely of anecdotes
concerning the transition between systems that do not include incentives and
systems that do.
In general, Conn found that user-charge systems, once introduced,
have not taken hold. Citing increases in littering and/or difficulty of
administering such systems, communities that have had them under
consideration have not adopted them, and communities that had instituted them
have disconitinued them.
European experience with the product charge, according to Conn, has
been somewhat more hopeful. Conn notes that a container charge in Norway has
been effective in preventing any growth in the use of non-refill abl e
containers for beer and soft drinks. In Sweden there has been a transition
from non-refill abl es to refill abl es; somewhat mysteriously, use of cans has
been on the increase.
Virtually all of the papers summarized above raised questions
concerning the effect of incentive approaches on the costs of administering
solid waste management systems. Barbara Stevens' paper presents two case
studies and the results of a statistical analysis of the administrative costs
associated with user-charge systems. In her case studies, Stevens shows that
administrative costs depend heavily upon the "skill" with which the system is
administered. Indeed, cost differences due to differences in skill probably
are far more significant than are differences due to use or non-use of a user
charge approach.
Stevens finds that on the average administrative costs for user
charge systems probably range between 4 percent and 8 percent of collection
costs. Because she has no firm estimates of the responsiveness to
user-charges of quantities of waste disposed, she is unable to make an exact
estimate of the net savings that result from switch from tax-based financng
to a flat-fee or quantity-based fee system. Ranges of estimates presented by
her suggest however that quantity reductions and/or savings of costs through
change in the collections system would have to be substantial in order to
offset increased administrative costs of a fee system.
-------
Pricing and the Equity of Solid Waste Financing
Efficiency effects are not the sole grounds upon which economists
have rested their case for economic incentives. Arguing that polluters
should pay the costs that their wastes impose upon society, economists have
also stressed that pricing approaches to financing waste management are more
equitable than are other approaches. The workshop heard two papers on the
question of equity and public finance aspects of pricing alternatives for
solid waste management.
The first paper, by Roger Bolton, discussed considerations relevant
to study of the distributional consequences of financing solid waste
management via user charges in contrast to finance via property taxes.
Bolton argues that a complete analysis requires consideration both of effects
of charges on household behavior and on the costs of solid waste management
-- the same evidence required to understand the effects of pricing
alternatives on solid waste management efficiency.
Noting that a complete analysis is beyond reach given the current
state of empirical knowledge, Bolton nonetheless was able to make some
approximate calculations of the relationship between income level and burden
of solid waste finance under a user charges policy, and to compare these with
the burden that would result if solid waste management were financed through
the property tax. His finding is that the distribution of the burden by
income level is about the same under both alternatives. He concludes that
there is "no reason to fear any noticeable increase in regressivity from
substituting user charges for the property tax, particularly if it is the
residential property tax which is being replaced."
Kenneth Wertz, in the second paper on equity and government finance
heard by the workshop, compared the effects of user charges with those of
product charges. Wertz argues that local governments and their citizenry are
likely to prefer the product charge approach to approaches based on
locally-levied user charges. While there are, Wertz notes, several
advantages that accrue uniquely to locally-levied user charges (e.g., placing
the burden of finance directly on the waste generator, and development of
data that may improve the management of solid waste collection and disposal),
he also notes that product charges avoid certain "liabilities" of a local
pricing system. "Under federal pricing, administrative costs exist but are
essentially borne elsewhere. Responsibility for the enactment of a new
revenue program can also be put elsewhere. The risk of increased littering
that goes with local pricing is removed by federal pricing, and the risk of
having promoted a bad idea if in fact pricing fails to induce much
reduction in the output of refuse is shifted to other parties." He
concludes that, because of these budgetary impacts and risks, product charges
are likely to gain more widespread political backing than are local user
charges.
Comparison of User Charges and Product Charges
Two papers heard by the workshop compared user charges and product
charges as alternative incentive mechanisms. One paper, by James Hudson,
-------
argued that product charges and user charges are complementary, and could be
combined to provide a more complete set of incentives than either taken alone
would provide. Hudson shows how a properly designed combination of the two
approaches can avoid double-pricing of wastes and can encompass a wider range
of wastes than either alone could cover.
The other paper, by Allen Miedema, compared the economic efficiency
of a product charge, user charge, and recycling subsidy in a simple general
equilibrium framework. Miedema finds that, in the illustrative framework he
has constructed, a product charges policy yields the greatest gain in
efficiency of the three pricing policies examined. He notes, however, that
the treatment of user charges in his model precludes recycling activity, and
thereby biases his results against user charges.
Pricing and Resource Recovery Technology
Two of the papers heard by the workshop addressed aspects of the
relationship between incentive mechanisms and resource recovery technology.
The first of these papers, by Tayler Bingham, presents the results of a study
in which a process model was used to predict the quantity of secondary
materials recovered from the solid waste stream under alternative prices for
these materials. Applying techniques used elsewhere by James Griffin,
Bingham estimates supply elasticities that confirm earlier findings that the
recovery of secondary material s form the municipal waste stream is fairly
insensitive to price changes. This finding has important implications for
pricing approaches. In particular, it implies that charges levied on
disposal of waste (either product charges or user charges) may not be very
effective in encouraging resource recovery.
The second paper on the relationship between incentive systems and
resource recovery, by Robert C. Anderson, presents some new estimates of
demand and supply elasticities in the markets for ferrous scrap and waste
paper. Anderson, in contrast to Bingham's approach, develops his elasticity
estimates from econometric models of the wastepaper and ferrous scrap
markets. While neither Anderson's or Bingham's paper contains all of the
information required to make a precise comparison, it is interesting that
Anderson's estimate of ferrous scrap supply elasticity is almost two times
Bingham's estimate. Anderson notes, moreover, that his estimate should be
adjusted upward by a factor of approximately two to obtain an estimate of the
supply elasticity for ferrous scrap other than prompt. Adjusted in this
manner, Anderson's elasticity is about 3.5 times higher than Bingham's.
Other things being equal, Anderson's results thus imply a substantially
greater shift to ferrous scrap in the face of disincentives to use virgin
materials.
A near infinitude of explanations could be offered for this
difference. Unfortunately, little has yet been done in reconciling process
modeling approaches with econometric approaches, as the discussion of
Bingham's paper makes clear. A tentative explanation may be that Bingham's
process model embodies more constrained production possibilities than in fact
obtain, with the result that supply is estimated to be less responsive to
price than it really is. Or, of course, it may be that supply really is as
-------
obtain, with the result that supply is estimated to be less responsive to
price than it really is. Or, of course, it may be that supply really is as
unresponsive as Bingham estimates, the econometric results being spurious.
Pricing and Sol id Waste Management
The papers discussed above for the most part dealt with the
relationship between pricing policies and the efficiency and equity of solid
waste management. Roughly speaking, each of the authors of these papers took
as given aspects of solid waste management other than pricing policy. The
fourth question raised by Goddard asks us to consider what happens in
real-world policy-making and service delivery situations where all aspects of
solid waste management may be subject to change when pricing is instituted.
Further, it asks us to consider what criteria are relevant to evaluating
alternatives in such an environment.
The workshop heard one paper on this subject by Bernard Booms. Booms
described experience to date in Seattle, Washington, as that city has moved
to experiment with user charges in combination with various other solid waste
management alternatives. He noted that the Seattle experience provides a
nearly unique opportunity for studying the implementation process. In
Seattle, the entire solid waste management system has been changed to
accommodate pricing. It is possible, therefore, to observe both the
political and administrative processes that come into play as system
management is changed. Importantly, Booms notes that the Seattle experiment
does provide for monitoring and evaluation of effects on a wide variety of
dimensions of solid waste management.
Research Directions
Two papers were delivered on research. One, by Oscar ALbrecht,
outlined the research activities of the U. S. Environmental Protection
Agency. Albrecht noted that although EPA has funded a great deal of research
in the economics of solid waste, much more remains to be done. Citing the
immense regulatory responsibilities of the Agency, he expressed the concern
tha important research needs may be overlooked. He called upon workshop
participants to articulate clearly research needs.
The second paper heard on research offered George Tolley's and Steven
Basting's response to the request put by Goddard and Albrecht. Tolley and
Hastings divide their discussion into two parts, one of which discussed
needed research on the cost of solid waste management services and the other
of which discusses research needs relating to the demand for solid waste
management services. On the cost side, the most pressing need identified by
Tolley and Hastings is for more refined studies of the relationship between
services provided and cost. Noting that solid waste management services have
a multitude of cost-causing attributes (including, for example, frequency of
service, location of collection, source separation requirements, quantity,
volume, container type), Tolley and Hastings called for detailed cost studies
that explore the relationships between each of the attributes and service
cost.
-------
Tolley and Hastings note that even less is known about the demand
side than the cost side. Here they call for a two-pronged approach. First,
they recommend that crude demand curves be constructed from information on
the cost of alternatives to collection and disposal. Presumably what they
have in mind is something similar to the analysis of resource recovery
presented in Bingham's paper, with construction of household process models
describing alternatives to municipal collection and disposal of solid wastes.
Second, they recommend that pricing experiments be conducted, controlled and
monitored so as to produce the data needed to infer households' demand for
solid waste collection services. To guide such experiments, they present a
comprehensive list of questions that should be provided for in the design of
each.
SUMMARY OF DISCUSSION
The papers summarized above provided the starting point for the
workshop's deliberations on the various questions posed by Haynes Goddard at
the beginning of our sessions. The full transcript of these deliberations
runs to almost 600 pages. It is impossible for me to do more here than to
try to draw out of the transcript some of the key points of agreement and
disagreement among workshop participants. Complete copies of the transcripts
are on file with the Municipal Environmental Research Laboratory in
Cincinnati.
Pricing and the Efficiency of Solid Waste Management
Discussion of this topic touched on several aspects of the
relationship between pricing and efficiency. The first aspect of this
relationship to be discussed was that between pricing and households'
generation of waste and selection of waste disposal methods. While several
conference participants hypothesized that changes of price and/or pricing
policies of the magnitudes currently observed around the nation have only a
small effect on household behavior, the consensus of the workshop seemed to
be that this hypothesis remains to be tested conclusively. The results of
quantitative studies that have been done all are acknowledged by their
authors to be subject to serious data and specification problems. While
these studies provide no basis for thinking that the effect of prices in
ranges currently observed are large, conference participants concluded that
until more careful research has been conducted, judgement should be reserved.
A second aspect of the efficiency question addressed by the workshop
was the cost of administering a pricing policy. Here again the workshop
participants felt that available data are inadequate to reach any firm
conclusions. However, most participants seemed to conclude tentatively that
any extra, costs associated with administration of a pricing policy are small.
A third aspect of the efficiency question concerns the balancing of
efficiency gains induced by changes in behavior associated with pricing
policies with the costs of administering them. Because of the uncertainties
noted above, the workshop was unable to reach even a tentative conclusion
concerning the net effect.
-------
Pricing and the Equity of Solid Waste Financing
Workshop discussions centered about two aspects of the relationship
between pricing policies for solid waste and the equity of solid waste
financing. The first aspect is the public perception of equity issues. In
this regard, workshop participants noted that there may be a growing public
perception that finance alternatives which distribute the burden of financing
solid waste management according to service usage as do user charges
and/or product charges are more fair than are alternatives which
distribute the burden in a manner unrelated to service usage. If so, we may
expect to observe more and more interest in solid waste service pricing
policies. Many workshop participants also noted that if measures to limit
governmental tax revenues continue to grow in popularity, public perception
of the unfairness of using general tax funds to provide free services which
otherwise could be priced will be heightened.
The second aspect of the relationship between pricing and equity
considered by the workshop is the effect of moving from a property tax
financed system. While several participants cautioned that the evidence is
far from complete, most seemed to agree that the distributional affects of
switching from tax based financing to financing via revenue raised from
pricing of solid waste service would not be great.
Optimal Combination of Incentive Mechanisms
This question was the subject of a great deal of discussion since
natiowide product charges are being considered actively at the federal level.
Many workshop participants expressed doubts that a product charge approach to
pricing solid waste collection and disposal services makes much sense. Among
the basic difficulties with product charges cited by those expressing this
view are the following:
1. Product charges could not be levied on many household
wastes. For example, product charges would not (and,
more importantly, could not reasonably) apply to yard
wastes.
2. Product charges set at a nationwide level would reflect
national average costs of assumed methods of collection
and disposal rather than local marginal costs of actual
methods of collection and disposal.
Most workshop participants who expressed reservations about product
charges argued that local user charges, which would apply to all household
wastes presented for collection and would be set according to local cost
conditions, may be a preferable pricing approach. However, the workshop,
including even those most skeptical about product charges, stopped well short
of concluding that user charges are a superior pricing alternative. It was
generally agreed that user charges also are subject to certain difficulties,
including (for example) the following:
10
-------
1. User charges, given current technology, cannot
differentiate among different types of waste. Since
collection and disposal costs probably vary widely by
type, user charges also do not reflect costs accurately.
2. Institutional considerations of administrative cost,
political feasibility, market structure and so forth may
weigh in favor of a product charge approach as against a
user charge approach.
Workshop participants concluded that the only way to resolve the
question of what sort of incentive mechanism is best is to obtain more
empirical information on the political economy of pricing approaches. More
information is needed on household responsiveness to incentives, the
relationshp of incentives to other aspects of solid waste management, the
costs of administration, and the process of implementing incentive
approaches.
Pricing and Solid Waste Management
As the above paragraphs make clear, progress has been made with
regard to study of particular aspects of the relationship between incentives
and particular aspect of solid waste management. While workshop participants
were reluctant to reach firm conclusions about the responsiveness of
households' waste generation to user charge/product charge levels, evidence
to date does support some weak positive conclusions.
This situation contrasts sharply with the state of our knowledge
concerning the relationship between pricing policy and the entire process of
solid waste management. We do not know, for example, what the relationship
between pricing policy and efficiency of collection and disposal may be.
Some observers have hypothesized that the use of prices makes costs of solid
waste management more "visible", and hence promotes efficient management of
solid waste collection and disposal. We do not know whether or not the
effect of prices on household generation behavior depends upon the nature of
collection and disposal option (e.g., self-hauling, recycling centers, source
separation programs), although we hypothesize that it does. Still another
interesting and important aspect about which next to nothing is known is the
relationship between information about waste content of goods and services,
pricing, and solid waste management. Workshop participants felt that the
scope of research on pricing and solid waste management must be expanded to
cover these as yet unexplored areas.
Research Directions
Implicit in the papers we heard and the discussion we conducted is a
long list of specific research topics. These topics certainly would include
the foil owing:
11
-------
1. Effects on household behavior of changing from tax-based
finance to pricing.
2. Effects on household behavior of availability and/or
prices of methods for solid waste handling such as home
compaction, garbage disposals, self-hauling, separate
collecting of yard wastes, and source separation and
recycling programs.
3. Effects of alternative combinations of system design,
operating procedures, and waste loads on the cost of
solid waste management.
4. Effects of pricing policy on cost of solid waste
management.
5. Effects of information on waste content and
collection/disposal cost on household behavior.
6. Institutional considerations in the design and operation
of solid waste management systems.
Workshop participants also spent a substantial amount of time
discussing research methods. Noting that the data that have been used in
studies conducted to date are deficient in imporant respects, workshop
participants seemed to be unanimous in recommending improved research design.
Among the specifics raised by workshop participants are the following:
1. An effort should be made to improve and expand the
collection of data on current operations and costs of
solid waste management systems.
2. Preparation should be made to capitalize on the
experimental opportunity presented when a community
decides to change its solid waste management system, or
the method or level of pricing of solid waste management
services. Case studies of these natural experiments
should be conducted.
3. Case studies of changes in solid waste management in
municipalities abroad should be conducted.
4. Serious consideration should be given to conducting fully
designed and controlled experiments.
CONCLUSIONS
In general, workshop participants felt that the time for an expanded
research effort on pricing in solid waste management has arrived. While
participants did differ in their views about the seriousness of
inefficiencies caused by failure to price solid waste management services
correctly, there was a consensus that pricing policy will become vitally
12
-------
important in the future. This will come about for several reasons, including
increasing cost of landfill, and growing pressures to reduce tax burdens. If
we begin now to design and execute a plan of research designed to answer
conclusively the practical questions raised elsewhere in this volume, we will
be ready to help communities manage their solid waste services more
efficiently.
13
-------
CAN ECONOMIC INCENTIVES HELP RESOLVE SOLID WASTE MANAGEMENT PROBLEMS?
by
HAYNES C. GODDARD
The purpose of this symposium is to bring together the principal and
most active researchers in the country on the question of the proper role of
economic incentives in a comprehensive approach to resolving municipal solid
waste management (SWM) problems. Outside of the investigations of the
comparative handful of researchers assembled here today, this area has
received only a small fraction of the attention that has been devoted to
economic incentives for the management of air and water residuals.
Furthermore, except for the product charge concept (a manufacturer's excise
tax on packaging materials), the question of economic incentives for SWM has
tended to be accorded the same degree of hospitality by governmental research
and program administrators as for other economic incentives: a large yawn.
None the less, with the gradually increasing attention paid to the
question by members of this group, our understanding of the potential for
incentives in this area is growing, and has led the organizers of this
symoposium to conclude that that the time is propitious for a concentrated
review of the major issues. By bringing together for the time those with
research interests in the area, we hope to enlarge our collective and
individual perspectives and understanding of the area.
We are assembled here with the common belief that it is worth testing
the hypothesis that economic incentive systems in SWM will raise economic,
engineering, and administrative efficiency and perhaps also equity by making
the costs of the various characteristics of SWM explicit to the waste
generators. Such characteristics are weight, volume, packaging content,
location of pick-up, frequency, and any other locally unique characteristics.
We hope that major outcomes of this symposium will be a specification of the
relevant research and policy questions that need to be investigated, and a
set of relatively specific recommendations indicating how this research
should be conducted.
Most of you are aware that, in general, the official reception of
economic approaches to residuals management has never been a very hospitable
14
-------
one and at times has been downright hostile. Our colleagues in engineering
and law have been much more successful in convincing policy makers of the
applicability and validity of their proposed solutions to the residuals
management problems and while, of course, physical and regulatory measures
are necessary conditions for efficient and effective pollution control, they
are not sufficient. Undoubtedly part of the problem we confront can be
explained by the low level of economic literacy that characterizes the
population in general, but this notwithstanding, the major share of the blame
for the lack of success which economists have suffered in obtaining
receptivity for their ideas must lie on the economists themselves.
Two of the internal impediments that some have observed to diminish
the receptivity of proposed economic solutions are:
a rather exclusive interest on the part of academic
researchers in the theoretical aspects of pollution
control, with a corresponding relative neglect of
the empirical side of the problem (of which I myself
am guilty);
a tendency to build models with substantial
specification error, either because of mathematical
convenience and/or because data on the correct
dimension to be tested is unavailable.
A major external impediment on the research side is that our
colleagues in engineering typically are given a limited exposure to economic
analysis, and thus conclude that they can perform what economic analysis is
needed to resolve pollution problems.
If we are to increase our effectiveness in making contributions to
the resolution of pollution problems, we must find ways to remove these and
other impediments, and we must become more effective spokesmen for economic
understanding, but, of course, we must have an understanding to conveythat
is why we are here.
Research on economic approaches to SWM does not have a long history,
dating back only 4-5 years, but nonetheless, this research has been pursued
in several directions. These directions have included user charges for SWM,
manufacturing excise taxes (product charges), litter control incentives,
relative efficiencies of alternative institutional arrangements, the impacts
of federal tax policies on waste generation and materials recovery, and the
incremental costs of collection and disposal among others. Our
understanding in each of the areas is growing, but is still insufficient to
make reasonably complete evaluations of the impacts of alternative incentive
mechanisms.
The speakers today and tomorrow will be touching on a large number
of current research and policy questions, and I will not attempt to
enumerate them here. I do want to mention, however, some of the major
questions that have not been answered, and hope that we will remember to
reference them frequently throughout our proceedings. These questions are:
15
-------
Do incentive systems that are designed to alter
waste generation behavior in fact raise the economic
efficiency of SWM, that is, produce large net
benefits?
What is the optimal combination of alternative
incentive mechanisms? How should the product charge
and the user charge be related to one another "in the
most efficient configuration of complementary and
substitute incentive systems?
How do these incentive mechanisms relate to the
choice of resource recovery technologies?
What is the optimal combination of all SWM
alternatives, and what criteria should be employed
to make the choice?
What is the best way to pursue the testing of the
effects of the incentive mechanisms, with historical
or experimental data?
I hope that this symposium will help us much in our search for
answers to these and other related questions. Hopefully, we can also find
some ways to remove impediments to the employment of economic knowledge in
the quest for efficient and effective environmental management.
16
-------
SECTION 2
PRICING AND THE EFFICIENCY OF SOLID WASTE MANAGEMENT
"The Economic Efficiency of User Fees: Some Preliminary Empirical Results"
(William N. Lanen).
"Evaluating the Efficiency of the Solid Waste Charge" (Steve Buchanan).
"Economic Approaches to Solid Waste Management: European Experience"
(W. David Conn).
"The Administrative Costs of User Charges" (Barbara J. Stevens).
17
-------
THE ECONOMIC EFFICIENCY OF USER FEES:
SOME PRELIMINARY EMPIRICAL RESULTS
by
William N. Lanen
One of the most fundamental propositions of economic theory is that
the demand for any good or service declines as the price of that good or
service Increases. This is, undoubtedly, one of the major reasons economists
traditionally look to pricing solutions rather than regulatory solutions for
the control of environmental pollutants. Today, economic solutions involving
charges are being employed to deal with air, water, noise, and land pollution
problems.*
Individuals charged with the management of systems (for example solid
waste management systems) have, on the other hand, tended to distrust pricing
approaches. The reasons for this distrust include notions of equity (for
example, that charges for solid waste collection are regressive) or
practicality (e.g., that user charges are administratively burdensome).2 It
then becomes important to determine whether user charge systems, especially
incremental charge systems, lead to improvements in economic efficiency.
The approach an economist would take to make that determination has
been given the name "cost-benefit analysis." Cost-benefit analysis is
nothing more than a set of procedures for performing an analysis of the
merits of alternative approaches to problem solution. As such, it is done,
perhaps only implicitly, by decision-making units at all levels. The
difference between the results of the analysis (i.e., the reason why
different groups come up with different "optimal" solutions) is that each
group accrues different benefits and incurs different costs. As we discuss
below, the conflict between system managers and economists about the optimal
solution to the financing of solid waste management may often be a result of
this difference in the organizational unit considered.
1. See, for example, the discussion in Anderson, et. al.
2. American Public Works Association, Solid Waste Collection Practice, 4th
Edition, Chicago, 1975.
18
-------
The purpose of this paper Is to present some preliminary results of a
study conducted by MATHTECH, Inc. for the Municipal Environmental Research
Laboratories of the U.S. Environmental Protection Agency. A case study
approach was employed using five cities, each with a different structure for
a user fee. (The term "structure" will be defined below.) The study was
primarily empirical in nature and its purpose was to evaluate the impact of
user fees residential solid waste behavior.
In the first section we discuss the concept of economic efficiency
and its role in evaluating user fee systems. We then describe the different
types of fee structures. Finally, we evaluate the qualitative effects which
we would expect each of the fee structures to possess.
In the second section a selective review of the literature is
presented. It is selective in that it discusses only work which has
attempted to estimate price or income elasticities. As we will show, while
there is some statistical evidence of positive income elasticities of waste
generation, no study to date has provided convincing evidence of a
statistically significant non-zero price elasticity. Non-zero price
elasticities, while they cannot be used to measure economic efficiency, are a
necessary condition for the employment of user fees to lead to an improvement
in efficiency.
In the third section the results of our study are presented.^ Both
the models and the statistical results are discussed. We will see that there
is some additional evidence of positive income elasticities. While some
evidence points to the existence of statistically significant price
elasticities, it is not strong.
Finally, in the last section we present our conclusions with respect
to the effects of a user charge system on residential solid waste behavior
and the consequent implications for their role in increasing economic
efficiency.
USER FEE SYSTEMS AND ECONOMIC EFFICIENCY
To an economist, a system that is economically efficient is one in
which net social benefits are maximized. Net social benefits are, of course,
nothing more than the difference between social benefits and social costs.
Under traditional cost-benefit analysis, the measure of social benefits is
"consumers' surplus." In other words, it is the difference between what
consumers would be willing-to-pay for a particular good or service (e.g.,
carryout service) and what they are required to pay (if they take the
service). Social costs are defined in terms of opportunity costs. Making
3. At this time, all of the results of the study must be considered
tentative.
4. The terms "benefits" and "costs" can, in practice, be misleading since,
for example, depending upon the baseline taken, benefits can be, and
often are, reductions in costs.
19
-------
use of these two concepts, we next show how they might be used to evaluate
the economic efficiency of a user fee system.
An Illustration
Consider the question of carryout service. If refuse collection
services are provided by the city, it has three basic alternatives: provide
the service to all residents; provide the service to no residents; or, charge
a fee for the service and let the individual resident choose. The
comparative benefits can be assessed using the information presented in
Figure 1.
Price
AC
Figure 1. Evaluating economic efficiency
In Figure 1, the demand for carryout service is drawn as a linear
function of price only for simplicity. It is assumed that at zero price, all
residents will demand the service. While this may not be true (if, for
example, some residents feel there are additional costs to having collectors
enter their yards) nothing in the analysis is altered by this assumption.
Suppose that, if the service is offered for a fee, the fee (p) will be set
to equal the average cost (AC) of providing the service. (Average costs are
drawn as constant In Figure 1, implying that average costs equal marginal
costs. We have excluded from the figure any administrative costs which might
be associated with the use of a fee.)
Consider now the net social benefits associated with the three
alternatives outlined above. The first alternative, providing the service to
everyone without an explicit change, is equivalent to charging a zero price.
20
-------
Therefore, the demand Is 100%. The total willingness-to-pay (i.e., the area
under the demand curve) is equal to the area OAD. The total cost of
providing the service is N x AC where N is the number of residents.
(There are no administrative costs since the service is provided to everyone
and no monitoring is required.) Therefore, the net benefits to society are
equal to the total benefits less net costs and depend on the relative size of
the two shaded areas in Figure 1..
Now consider the net social benefits of not providing the service at
all. In this case the net benefits are zero -- no costs and no benefits.5
Finally consider charging a price p (assumed to be equal to average
cost). The percentage choosing the service will be W. The total benefits
(i.e., w.illingness-to-pay) are OABW. The total costs are AC x (W)(N). The
net social benefits are the vertically shaded area pAB. Note that the
benefits associated with this arrangement are greater than those associated
with providing the service to everyone by the amount WBD (the horizontally
shaded area) neglecting the administrative costs.
Similarly the net benefits associated with the charge system are
greater than providing no service (again neglecting administrative costs) by
the amount pAB. The question of the best choice depends then on the level
of administrative costs. If they are less than pAB, a charge system is, on
net, beneficial when compared to providing no service. If they are less than
the area BCD, then the charge system has greater net benefits than the free
provision. Of course, the level of the administrative cost and the relative
magnitude of the shaded area are empirical questions.
Suppose, however, that in the process of trying to evaluate the net
social benefits associated with a particular fee structure, that the analyst
is missing some of the crucial information; e.g., the actual demand curve for
the service. Is there some necessary condition that must hold for a user fee
to improve economic efficiency (not equity)? The answer is yes. Basically,
it must be true that the demand curve has a negative (i.e., non-zero) slope.
Consider what happens if the slope (or elasticity) is zero in the relevant
range.5 If the elasticity is zero, the waste presented for collection
remains the same and, therefore, total costs remain, the same. All that
happens is that there is a transfer payment from residents to the solid waste
system, and under traditional cost-benefit analysis such a transfer has no
effect on benefits.
5. Again, this result is dependent on the definition of the baseline.
6. Obviously, at some price, demand for the service will fall. In
discussions, such as this, with policy implications, it is of interest
to determine whether there are significant price effects within the
range of user fees normally encountered in practice.
21
-------
A Classification of Fee Structures
As discussed above, the approach followed in this study has been a
case study approach of five cities, each of which has a different type of fee
structure. By "different" we mean merely that the resident in each of these
cities is faced with a different set of choices concerning the disposal of
the solid waste generated within the household. Before discussing the
findings in each of the five cities, however, it is useful first to discuss
the classification systems used to develop the five different fee structures.
Once this has been done, we can analyze what the qualitative effect of a fee
increase would be.
The nature of the fee systems used by the roughly 1400 communities
identified in the study that employ user fees for the collection of
residential solid waste vary greatly. However, within that variation, there
is one characteristic which is (a) easy to identify, (b) useful in the
development of hypotheses concerning behavior, and (c) restricted in
relevant possibilities. This characteristic is, therefore, an obvious one to
use for the classification of user fee systems for the purposes of selecting
the case study cities. This characteristic is the set of decision choices
the resident has for the disposal of solid waste. Although there may be
others, we have selected five different possibilities as including
(hopefully) most of the fee structures currently used. Below we identify and
define each type.
The first, simplest, and by far the most widely used fee system is
the uniform or flat fee structure. In this system, the resident (assuming
service is mandatory) has no choice over his disposal options.7 The goal of
this type of fee structure is basically one of revenue raising. In other
words, the flat fee service is an alternative to general tax revenues for
supporting the solid waste system.
A more complex system is one in which the resident has the option to
choose the number of containers (i.e., the capacity) for waste for pickup.
While it is not necessary to assume that the resident must remain within that
constraint once and for all, it is assumed that there are some transaction
costs associated with frequent changes from one level of capacity to another
that are sufficiently large to discourage such changes. We will refer to
this as a capacity-based structure.
A third type of fee structure is one in which the resident can choose
alternative levels of service. Service levels can be defined in terms of the
point of collection (curbside or backyard) or frequency of collection (e.g.,
weekly or semi-weekly). Most cities that have a service-based system provide
a choice in terms of location. In this system, the number of containers that
can be presented is specified (although it may be unlimited). This is
referred to as a service-based structure.
7. Certainly, the household can choose illegal modes of disposal.
22
-------
Metered-bag systems are those systems where the resident presents all
waste to be collected in a specially-marked bag. This system is quite
similar to the capacity-based system discussed earlier. The difference, and
this is an important one, is that the resident can alter, at no cost (other
than the cost of the bag) the number of containers presented each time.
While this system is not a true quantity-based system, where the resident
pays for collection according to the weight or volume of the waste presented,
it is a step closer to that ideal because of the freedom offered at each
collection period.
Finally, some municipalities employ fee structures that are
combinations of two or more of the other systems. For example, in Tacoma,
Washington, the resident chooses both the number of containers and the level
of service (in terms of pickup location). This type of system may be
expected to be more expensive ye.t more flexible than a system offering only
one choice. This will be referred to as a combination-based structure.
Analyzing the Effect of Different Fee Structures
In theory, one could qualitatively analyze the impact upon resident
behavior from the change in the level of the fee for any type of fee
structure by first developing a utility maximization model with each of the
fee features. Then, comparative statics could be used to determine the
direction of change in the amount of waste generated. Such an approach was
used, for example, by Kenneth Wertz (1976) when analyzing the effect of price
and service level upon waste generation behavior. The problem with such an
approach is that, of necessity, most of the fee systems that exist include in
their rate structures such large discontinuities (e.g., the number of
containers) that applying such methods may not be particularly fruitful when
examining actual systems. The results of the Wertz analysis are helpful,
however, in that they provide a check on a more simple approach to the
qualitative determination of the effects of a fee change.
Consider, for example, an "ideal" fee structure where the resident
pays by the pound (or cubic foot, or both) for the amount of waste disposed.
Further, suppose that 'he is allowed to choose the point of collection (for a
fee). A system of demand equations that might be expected to model such a
decision process adequately would be like the two equations:
3 - 8Cps, y) (2)
where q represents the quantity of waste disposed, pw the price per unit
of waste, y is disposable income (after payment of any flat fee for
disposal), s is a measure of service level (e.g., the number of feet from
the curb), and ps is the unit price of the service level. Noted above each
23
-------
of the independent variables in both equations are the effects we would
expect changes to have. For example, following elementary theory we would
expect that an increase in the price of either quantity or service would
decrease the demand for the respective good, and similarly for income. In
equation (1), we show that the effect of an Increase in service level leads
to an increase in the amount of waste generated. This is consistent with the
findings of Wertz. Intuitively, its justification is quite straightforward.
The disposal of residential solid waste requires the labor of two different
units: the collector and the household. Use of household labor is costly
just as is the use of the collectors' labor. Increasing the service level
while holding all other variables constant, means that the cost to the
household of presenting an additional unit of waste has fallen. We would
therefore expect to observe an increase in the amount of waste presented.
We can now use equations (1) and (2) to determine the effect of a
change in the price of either quantity or service level. The effect of a
change in the price of quantity is known directly from the demand curve. The
effect of a change in the price of service level upon the amount of waste
generated depends on the indirect effect of the service price on the level of
service. Thus,
9q/9p = 9q/9s . 9s/9p < 0 (3)
8 S
The effect of an increase in either price therefore is expected to be a
decrease in the amount of waste presented for collection. These results were
for the ideal system. We now evaluate the five types of fee structures
described above.
In the case of the uniform fee, the only equation is:
q = f(y) <4>
Note that the only variable influencing the amount of waste presented for
collection is the disposable income. Suppose we define,
y = m - p (5)
where m is household income before the fee and p is the flat fee. Then,
9q/3p - dq/dy 9y/9p - - 9q/9m (6)
That is, the effect of an increase in the flat fee is equivalent to the
effect of an equal decrease in income. In terms of elasticities, however,
24
-------
"n = 3q/3p . p/q = - 3q/3m . p/q » - 1m P/m (7)
This means that if the income elasticity of waste generation is, say,
on the order of .30, then the elasticity with respect to a change in the
level of the flat fee is insignificant.
The capacity-based system differs from the ideal because the resident
does not face an explicit charge each collection. Rather, he faces what is
essentially a capacity constraint. Therefore, his choice can be modeled as:
.,+ +.
q - f(c, y) (8)
+
c - 8Cpc, y) (9)
Again, the signs above "the independent variables represent our assumptions
about the effects of changes in these variables on the amount of waste
generated. In these equations, C represents the number of cans selected.
Ignoring the discontinuity of the number of cans, we can analyze the effect
of an increase in the price per can on the amount of waste generated. It is:
3q/3pc - 3q/3c . 3c/3pc < o (10)
The product of the two terms in (10) is negative by virtue of the effect of
price on the number of containers selected. However, if the first factor is
less than one, the resident can moderate any inhibiting effect of the price
change. This factor 3q/3c measures the density change resulting from a
change In the number of containers. That is, If it is less than bne, part of
the effect of the change in price is to have residents packing each container
slightly more full. This would be a perfectly rational thing to do since, as
the price of an additional can increases, the relative cost of the resident's
own labor in the preparation of his waste for collection falls.
This suggests that one of the "weaknesses" of the capacity-based
system is that it provides a means for the resident to avoid the full effect
of the price increase. Of course, the quantitative importance of this effect
is an empirical question.
The model for a service-based fee would be
., .
f(s, y) (11)
g(? , y) (12)
25
-------
Here we are Interested not so much in the effect of changes in the level of
the fee on the choice of service level by the resident, which we know must be
negative, but the effect on the quantities of waste generated. But this is
just,
dq/dp - dq/da . ds/dpa < 0 (13)
S B
Therefore, we see that the effect is dependent upon two quantities; the
effect of price on service and the effect of service on quantity.
The bag system differs from the capacity-based system in two
important respects. First, the resident chooses the number of containers to
use each collection period. The second is the fact that each bag used
includes the cost of the bag. With a container, the container was not
"consumed" when filled. With the bag it is. Therefore, the system of demand
equations for this system can be specified as:
q = fCp , B, y) (14)
B
B - gCp , y) (15)
B
where pB is the price of a bag. Note that the price of bags affects the
demand for two separate goods, the collection of waste and the number of
bags. Therefore, we would expect the effect of a price change to be somewhat
more complicated. In fact, it is,
dq/dp = dq/dp + dq/d'B 9B/3p < 0 (16)
B B B
The combination systems combine the effects of two or more of the
types of fee structures. Assuming that the combination is for the number of
containers and the location for pickup, the model for analyzing the effect of
fee changes would be:
q - f(c, s, y) (17)
c " 8(? , y, s) (18)
C
s - M? , y, c) (19)
S
26
-------
Measuring the effect of a change in the price for containers is now
complicated by the fact that the container fee will affect both the number of
containers selected and the level of service chosen. The effect of a change
in the fee for a container is
= dq/dc dc/dp + dq/ds ds/dc dc/dp < 0 ; (20)
c c c
the effect of the change in the fee for level of service is
dq/dp =» dq/da ds/dpo + dq/dc dd/ds da/dp < 0 . (21)
S S 8
We see that in both cases the impact of an increase in either fee is to
reduce the amount of waste presented for collection,
REVIEW OF THE LITERATURE
While there have been innumerable studies of the cost of providing
solid waste collection services (along with many prescriptions for decreasing
agency, not necessarily social, costs with particular technologies), there
have been relatively few studies that empirically address the effects of
pricing on residential solid waste behavior. Without such Information,
however, the analyses of the "optimal" type of fee structure would be
impossible to conduct in terms of cost-benefit analysis. The purpose of the
following is to provide a brief, and admittedly selective, review of the
empirical literature that has developed In this area. The reason for focusing
on the following studies is that they form the basis for much of the current
discussion regarding solid waste pricing and, because they provide numerical
estimates of elasticities and other economic data, have come to be relied
upon in policy-making discussions. Therefore, it is important that they be
assessed critically in order that the faith placed in the results be
commensurate with that due.
Wertz
The Wertz article is primarily concerned with the development of
qualitative implications for the effects of changes In service fees or
quantity fees. However, throughout the paper are several brief examples
which provide some support for the theoretical propositions developed. The
ones which we are primarily interested in are those relating to income
elasticities, price effects, and service level effects.
With respect to income effects, Wertz used data from 10 suburbs of
Detroit which had similar financing and collection policies. Using a linear
functional form for the demand curve, Wertz found implied income elasticities
27
-------
of .279 and .272 (depending on the actual sample used). In a conclusion that
does not seem to be entirely warranted, Wertz states "The foregoing mixture
of theory and observation suggests the expected: residential refuse
quantities should decline as t (price) increases."
The "foregoing mixture" of theory appears to be the derivation of the
usual classification of effects into substitution and income effects. This
is perfectly straightforward following usual demand theory. What is not
clear is how the observations about income elasticities from cities employing
no incremental user charge can support expectations about price elasticities.
Following this conclusion, Wertz examines some additional evidence
bearing on the price effect. He cites the fact that per capita generation in
San Francisco, which employs an incremental user fee, was substantially less
than "for all urban areas where general financing prevails." As Wertz notes,
there are too many variables to place too much belief in the numerical
accuracy of the implied price elasticity of .15. In addition to the lack of
data, however, is the fact that the comparison was between two different
years made comparable by applying "an average growth rate." In addition, the
growth rate was applied to a figure which was composed of measured tons and
estimated (from volume) tons.
The estimates Wertz provides on income elasticities appear to be
consistent with the findings of others and are based on statistically sound
methods. The empirical evidence of a non-zero price elasticity is not,
however.
Tolley, Hastings, and Rudzitis
In an updated version of an earlier study, Tolley, et al. , provide
estimates of income elasticities based on cross-sectional data from several
wards in Chicago. Because waste collection services are financed out of
general revenues in Chicago, no estimates of price elasticities were
possible.
The findings of Tolley, et al., were consistent with the earlier
work. Namely, the estimated income elasticity in Chicago appears to be .3
and .7, depending on the season.
Again, these findings are also consistent with those of others that
there is an income elasticity which is positive but less than one.
McFarland
One of the most oft-cited studies concerning the existence of a
significant price elasticity is the study by the University of California on
solid waste practices in that state. Chapter IV, which formed the economic
basis of the report, was authored by McFarland and has come to be known by
that name. We will therefore continue to use it.
28
-------
The approach McFarland used to estimate the price elasticity was to
apply ordinary least squares to the following equation:
al &2 a3
Qd = a0Xl X2 X3
where Qd is the annual per capita quantity of waste generated, Xi was
average revenue, X2 was per capita income, and X. was population density.
The results of the regression were:
InQ - 6.9 - .455 InX. + .178 lnX? - .212 InX,
(3.2) (.36) * (1.5)
where t-statistics are noted in parentheses. McFarland used this result to
state "This indicates that people will definitely respond to price incentives
or disincentives in their use of the service." There are, however, several
problems with these results -- problems which essentially vitiate the
results.
First, the price proxy used was average revenue. Unfortunately,
McFarland did not specify the actual 13 cities used in the analysis.
However, since only two out of the 58 in the entire study appear to employ an
incremental user fee, the majority had to be flat-fee cities. In that case,
there is no price elasticity. A second problem is that many flat-fee cities
impose quantity limits. This means that service levels are not held
constant.
Finally, there is undoubtedly a good deal of simultaneity. Most
municipal systems are designed so that the flat fee is, at least somewhat,
related to system costs. But system costs are related to the level of waste
collection. Therefore, by not including a second equation in the model and
using simultaneous methods, the results do not have the usual desirable
properties.
These problems may also have led to the finding of an income
elasticity not significantly different from zero, a finding which conflicts
with other studies.
McFarland goes on in the analysis to discuss the effect of incremental
fees on litter the primary externality to be expected from the imposition
of a user fee. To do this, she classifies cities as internalizing or
externalizing, depending upon their mode of financing and quantity
limitations. Externalizing cities are those cities not "providing unlimited
or generous service at zero marginal costs." She then found a significant
difference in the costs associated with the solid waste management system
between the two types of cities. She attributed this difference to the extra
costs of litter cleanup in the externalizing cities.
29
-------
The results of this analysis have been criticized elsewhere (see;
e.g., Goddard). However, an important point that we have not seen mentioned
is that flat-fee cities imposing quantity limits (while perhaps by definition
incremental fee cities; i.e., with an infinite incremental fee) are not what
is generally meant by incremental fee cities.
The empirical results of McFarland, therefore, do not appear to be
sufficient evidence of the existence of non-zero price elasticities either
directly from demand equations or indirectly through litter effects.
Stevens
In an unpublished research paper, Barbara Stevens has presented some
results, both theoretical and empirical, concerning service level pricing.
In the theoretical analysis she extends the results of Wertz by explicitly
including frequency of collection and location of pickup simultaneously. She
concludes the theoretical section with:
All mandatory collection systems, whether the fee is
explicit or implicit, cause the consumer to generate less
wastes when prices are increased (provided that either the
price elasticity of demand for service or that the relation
between refuse generation and income is small) and cause
the consumer to generate more wastes in response to a
costless increase in service level. Only the service level
fee pricing scheme has the joint advantage of encouraging
consumers to value goods implicitly according to the
disposal cost of their refuse component and of
implementability.
It is important to note her qualifier "mandatory." She goes on to
assert:
When a pricing scheme does not require mandatory
participation of households, none of the above conclusions,
with respect to any of the pricing models, can be stated
with any confidence. In any such non mandatory
arrangement, increases in the price of refuse collection
services can lead to a decrease1 in the proportion of
households selecting organized collection services. The
households opting out of the system may choose to self haul
to legal disposal site (sic), to dump in illegal locations,
to burn refuse, etc. Any or all of these alternative
disposal technologies may result in a perceived increase in
disposable income and a consequent increase in refuse
generation. In addition, some of these alternate disposal
technologies may result in increased total costs of
collection and disposal of refuse to be borne by the
society as a whole.
The reason for such a counterproductive generation effect, at least in terms
of economic theory, is unclear. For if the resident could make use of the
30
-------
alternative facilities at lower cost than at the new user fees, the same
alternatives were available at the lower cost. That is, let p(Q) be the
price of disposing of an amount Qc through "conventional" means. Let L(Q)
be the "price" of some alternative form of disposal (e.g., littering). Then
the resident will choose quantities Q , Q. to be disposed of
conventionally and littered so that c
TC - p(Q)Q + L(Q)QL
C
is the total cost of waste disposal. Total waste generated is just Q +
QL. Now suppose there is a price increase in conventional disposal.
Regardless of the functional forms of the price functions p(Q) and L(Q) , the
total cost cannot fall. (It is true, however, that there will be more waste
disposed through illegal means and, therefore, social costs may increase.)
The difference between mandatory and non-mandatory systems is that,
under the latter, the resident has more substitution possibilities'
available. Thus, in such systems, we would expect the inhibiting effect of
any price increase to be more moderate than with a mandatory system.
With respect to the empirical portion of the paper, Stevens provides
a useful test of the efficiency of service-based fee systems. Stevens
estimates three demand curves: one for waste presented for collection;
demand for service frequency; and demand for service location. Using the
latter two, we can estimate consumers' surplus and, hopefully, say something
about the effect on economic efficiency of service-based plans.
For example, the demand curve for location of pickup is
BY = 271 + . 003Y + 1. 4QH + . 04FRE - 3. 3P - . 028DEN (22)
where BY = percent .of households selecting backyard service
Y = mean annual family income
QH - annual tons of refuse per household
FRE = percent selecting higher frequency of service
P - price (per month) for carryout service
DEN = persons per square mile
Significant explanatory variables were income and price.
Substituting the mean values for each of the variables except orice
into (22), we get
BY = 268 - 3.3p
BY
31
-------
Equation (23) can be used to estimate consumers' surplus. First note that
(23) is linear and therefore consumers' surplus is just the triangle bounded
by the demand curve, the price, and the price such that BY is just zero.
That is,
CS = BY x (p - E^ ) x .5
max T3Y
(24)
where Pmax is fc^e Price such that BY is just zero. From (23) p =
$81. Using mean values for BY and p , we find
max
CS = .28 x (81 - 2) x .5 = $11.06/year
The $11 is per household (since the quantity term is percentage of
households). To determine whether the offering of such a service leads to
an increase in efficiency, we need to compare this benefit to the
administrative costs of providing the optional service (we assume that the
price includes the additional collection costs). While there is no data, we
can infer likely effects. Stevens' sample generated 1.71 tons per household
annually. Assuming average collection and disposal costs of $30, total
annual costs per household would be $51. The benefits of $11 annually per
household represent 22 percent of this. Edwards and Stevens found
administrative costs to be between 3% and 18% of collection and disposal
costs. Thus, it appears that providing the optional service increases
economic welfare.
Because of the assumptions necessary to estimate consumers' surplus
in this case, we would not want to say that the Stevens' results can be used
to infer that service-based plans increase economic efficiency. We would
argue, however, that they are highly suggestive.
Based on the theory and previous studies relatively little can be
said about the magnitudes of the price and income elasticities. There
exists no statistically valid evidence of a quantity effect. The Stevens'
result suggests that price is important to the resident in selecting service
levels.
In the next section, we present our analyses of the effects of five
different types of fee structures. As with the studies reviewed here, the
results (or lack of results) depends on the amount of data available.
RESULTS
In this section, we present some preliminary empirical results of our
study. The emphasis in this section will be on results about elasticities,
both income and price, that can be derived from each of the case studies.
32
-------
Burbank, California
The City of Burbank, California employs a flat-fee system for
residential solid waste collection. Service from the City is mandatory for
all residents in single family dwellings and is available to residents in
multifamily housing. Service is offered once-weekly and there is no limit to
the amount of waste that may be presented for collection provided only that
it is properly packaged. These two factors mean that City residents have no
incentive for either reducing the amount of waste generated or to engage in
illegal disposal methods in response to a fee increase.
As we will show below, the income elasticity estimated for Burbank,
given the assumptions and model specification discussed below, is consistent
with previous findings. Recall that with a flat fee there is no non-zero
price which the resident faces but rather a one-time charge that operates
through an income effect. Therefore, there is no price elasticity to be
estimated.
The basic equation estimated for the Burbank case study is:
b c
Q = a(RS) s
where RS represents retail sales, which was taken to be a surrogate for
income. Taking logs on both sides,
ln(Q) = ln(a) + b ln(RS) + c ln(s) (25)
In this section, we present the results of the econometric investigation of
equation (25).
The first step is to select seasonal variables. For the estimation
of the equation, we have used dummy variables to represent Winter (December,
January, and February), Spring (March, April, and May), and Summer (June,
July, and August).
Using monthly totals on waste collected and including the 30% of
"commercial" waste that is estimated by city personnel to be residential, the
results of applying ordinary least squares to equation (25) gives (standard
errors are in parentheses):
InQ - 6.18 + .1711nRS - .06 WINTER + .08 SPR + .07 SUMMER
(.11) (.025) (.025) (.024)
R2 - .38
N - 83
33
-------
As shown In the results of the regression, the coefficient in the
income surrogate is not significant at the 95% confidence level (the
t-statistic is 1.62). Before failing to reject the null hypothesis of no
income effect, some additional investigation was performed. The most
plausible reason for the finding of no effect is that the effects of income
are felt on waste generation only after some lag. Therefore, we use lagged
(by one month) retail sales in the next regression. The results are:
InQ = 5.73 + .221 In(RS-l) - .058 WINTER + .081 SPR
(.106) (.024) (.024)
+ .068 SUMMER (26)
(.023)
R2 = .36
N = 81
The coefficient on the income variable implies an income elasticity
of .22 and is statistically significant at the 95% level. This figure is
consistent with previous findings as discussed in Section B above.
To check the sensitivity of the results to the particular quantity.
variable, the results of several other regressions are summarized in Table 1.
As shown there, none of the Implied income elasticities are significantly
different from one another. (All regressions Include the same set of
independent variables as in equation (26)).
TABLE 1. ALTERNATIVE INCOME ELASTICITIES
Excluding 30% of
commercial waste
Daily totals including
30% commercial
Daily totals excluding
30% commercial
Per capita monthly
Estimated
income
elasticity
.221
.272
.273
.233
t-statistic
2.05
2.83
2.81
2.13
R2
.36
.28
.27
.35
including 30%
Note: For all regressions, the number of observations was 81. The
period was February, 1972 to December, 1977 excluding
December, 1973 and January, 1974.
34
-------
These results lend support to previous findings In the range of .2-.4
for the Income elasticity. Because this analysis Is a time-series analysis
we would expect the elasticities estimated to be short-run elasticities and,
therefore, somewhat lower than those estimated by cross-sectional analyses.
Sacramento, California
The City of Sacramento, California allows residents to choose, for a
fee, the number of cans to be picked up during regular collections. We have
referred to such a system as a capacity-based fee. Service from the city is
mandatory for all residents. The basic service is offered once per week with
special collections, for an additional fee, available upon request. As we
have discussed above, such a fee system may induce residents to dispose of
less waste by imposing an additional fee. In addition, it may induce
additional use of illegal disposal methods.
Unfortunately, the data in Sacramento are not detailed enough for us
to make a definitive statement about these hypotheses. While good data on
the amounts of waste disposed over time are available, other important data
items, such as the number of containers chosen, are not.
An important difference in the Sacramento system is the provision for
the pickup of lawn and garden refuse. Lawn and garden refuse in Sacramento
is not collected by refuse crews. Residents are not required to place their
garden waste in containers (except those in areas without curbs) but are
required to pile the waste in the street no sooner than one day prior to the
scheduled pickup. Thus all references to the solid waste system in this
section exclude lawn and garden trash.
The basic equation estimated was:
bed
Q ap y s
Taking logs on both sides gives:
ln(Q) - ln(a) + bln(p) + cln(y) + dln(s) (27)
Again, as with Burbank, we will use dummy variables to represent three of the
four seasons. Further, a surrogate is required again for the income
variable. We again will use retail sales in the City of Sacramento. The
basic one-can fee is not subtracted since it is small relative to income and
no true income variable is available.
Using the data described above, the following results were obtained
after applying OLS to equation (27) (again standard errors are in
parentheses):
35
-------
4.7 + .09 ln(RS) + .23 ln(p) + .01 WINTER
(.07) (.09) (.01)
- .01 SPRING - .02 SUMMER
(.02) (.01) (28)
R2 = .40
N = 26
There are three important results to note. First, the income elasticity,
while positive, is insignificantly different from zero. Second, the price
elasticity is positive -- and significant. Third, the seasonal dummies are
not significant.
Because the existence of a positive price elasticity is theoretically
implausible, the results in equation (28) were analyzed further. The first
step is to recognize that, as with Burbank, the effect of income and price
changes may occur only after some lag. The effect of lagging retail sales
and incremental price by one period is:
Q = 3.6 + .20 ln(RS) + .22 ln(p) + .01 WINTER
(.08) (.10) (.01)
+ .02 SPRING - .01 SUMMER (29)
(.02) (.01)
R2 = .50
N - 22
Performing this regression results in a positive and significant income
elasticity of .20. However, the problem of a positive and significant price
elasticity remains.
Because the price variable used is deflated by the CPI and, for
several months, the nominal fee is unchanged, the effect is that the
incremental fee may be acting as a trend variable. To check this
possibility, we substitute a trend variable for the seasonal dummies. When
this is done, the results are:
ln(Q) = 1.95 + .35 ln(RS) + .09 ln(p) - .002T
(.12) (.09) (.001)
R - .48 (30)
N - 22
The effect of this substitution is to retain the positive, and significant
income elasticity (although the magnitude of the elasticity has increased to
36
-------
.35). While the price elasticity is again positive, it is insignificantly
different from zero. Further, while the R is about the same as before,
there are variables making the regression significant at a higher level.
Provo, Utah
Among cities with user fees for household solid waste collection,
relatively few employ a variable fee structure based exclusively on pickup
location, charging one rate for backyard service and another rate for
curbside service. Provo, Utah, is one such city. Residents of Provo may
choose between paying $2.50 per month for curbside collection and paying
$5.00 per month for backyard collection.
Above, a demand model for a service-based fee was presented. In that
model, the amount of waste generated was seen to be affected by the price of
the service through the effect of service levels on waste generation.
Unfortunately, the lack of data in Provo makes it impossible to estimate the
demand specifications presented.
Other data that is available provides an opportunity to indirectly
test for an income effect in the demand for increased service levels. This
data is the breakdown by billing tracts of the percentage of residents
choosing backyard service. While these tracts do not correspond to Census
Tracts, they can be identified in terms of "housing value" which might be
taken as an (albeit imperfect) surrogate for income.
We specify, therefore, the following equation:
al
BY = a.P + a-TRACT. + a-TRACT9 + a^TRACTo (31)
0 6 1 J " " J
al
where BY is the fraction of residents choosing backyard service, P is
the price of backyard service, and the TRACT^ are dummy variables
representing the different areas of the city. There are 57 observations for
each of the years 1975 and 1977. These represent years before and after the
fee.
Applying ordinary least squares to (31), we obtain:
BY - -5.14 + .13P + 2.7TRACT. + 1.7TRACT, + .82TRACT-
(.13) (.33) l (.32) * (.33)
R2 = .43
N - 114
We see that backyard service in each of the tracts, which are in declining
order of housing value, is significantly and positively related to P. Note,
also, that both the magnitude and significance of the estimated coefficients
decline with TRACT number.
37
-------
These findings in Provo lend some support to the existing evidence of
positive income elasticities for solid waste services. However, because of
the nature of the data, the support is not strong.
Grand Rapids, Michigan
Grand Rapids, Michigan operates a non-mandatory solid waste
collection service which participating households pay for by purchasing
plastic bags or tag cards whose price includes payment for municipal
collection and disposal. Specially marked city garbage bags are sold for
$2.30 per dozen; tags are sold for $1.25 for 10. The city also sells special
plastic trash cans whose contents are collected upon payment of an annual fee
of $4.35 or $6.50 depending on the container size. In addition, several
privately owned collection firms operate in Grand Rapids, and residential
households may arrange service with these rather than pay for municipal
service.
A metered-bag system might be expected to affect residential solid
waste behavior in two ways that are different from capacity based systems.
First, the capacity of the bag is generally less than that of a waste can.
Therefore, a metered-bag system faces the resident with a smaller increment
of waste allowed between addition of a new bag. Second, because the bag is
consumed when filled, the resident faces a direct charge for each additional
bag used. With a capacity based system, once the number of containers is
selected, there is no direct charge to the customer for using the number of
cans chosen. Thus, the metered bag appears to fall in between the "ideal"
system outlined above and a capacity-based system such as that in Sacramento.
This hybrid nature of the metered bag system makes the modeling of
consumer demand for waste collection services somewhat more tricky than that
of the capacity-based system. Fortunately, it turns out that it is possible
with a single equation model to estimate the effect of price on waste
generation whether the bag system is considered to be more like the "ideal"
system with a charge varying with weight or more like the capacity system
with the unit of capacity being the bag rather than the can.
Recall that with an ideal system, the resident is charged for the
weight presented for collection each collection period. If the metered-bag
system were a reasonably good approximation of the ideal system, the resident
would not be able to "hide" additional waste by filling each bag with more
waste. Therefore, the ratio of bags to waste would be constant and the
"price per pound" could be calculated as:
P = PBAG/(q/BAGS) (32)
where p is the price per pound, P-RAQ is the price per bag, q is the
weight of the waste presented for collectTon, and BAGS is the number of bags
used.
38
-------
If a double log specification is assumed for the waste generation
function, it would look like:
al *2 *3
q - a P y s (33)
where s represents seasonal factors, y is income, and q and p are as
before. Substituting (32) into (33) gives:
al ~al ~a2 ~&3
q - a p (q/BAGS) y s (34)
If the assumption concerning the constancy of the waste to bags ratio
is correct, we can rewrite (34) as:
a p y s (35)
0 BAG
where:
a - a (q/BAGS)
Thus, we needn't make use of the data we have on the number of bags sold in
Grand Rapids. This is fortunate since the data available are on wholesale
sales and, therefore, incorporate inventory phenomena not on the part of the
household but on the part of the retail outlets.
Suppose that instead of being viewed as an ideal system, the
metered-bag system were merely another capacity based system with a different
representation of capacity. Then, from the discussion in above, the
theoretical representation of the demand model would be (assuming again a
double-log specification) :
q = b BAGS y s (36)
0
c c c
BAGS = c p y q (37)
0 BACT
where the variables are defined as above. Substituting (37) into (36) and
simplifying gives:
bi b? b^
« ' VBAG y ' (38>
39
-------
Note that the specification in (38) is the same as that in (35). Therefore,
estimation of either equation provides information about both possibilities.
The difference is that if the raetered bag system is actually a capacity based
system, then the coefficients estimated for (38) no longer represent
elasticities of price and income. Given some reasonable restrictions on the
elasticities in (36) and (37), however, we can determine the sign on the
price and income terms if the actual elasticities are to be negative and
positive, respectively.
For example, if it is assumed that:
o > b > -1, 1 + aj(l + b3) > o, ja^l < &2
then the signs on bl and b£ would be as expected from the economic
theory. Note, however, that the numerical estimates of bj and b^ in (38)
cannot be construed as elasticities since they are complicated functions of
other parameters and are not identifiable.
In order to estimate Equation (35) [or equivalently, Equation (38)],
we must first transform it into the linear form:
ln(q) = In(b') + b'ln(p ) + b'ln(y) + b's (39)
0 1 BAG <£ j
The next step is to identify the actual data series used in the estimation.
For total waste disposed, the series presented above was employed. Because
some of the data is expressed in cubic yards and some in tons, it was
necessary to make the two comparable. To do this, we used the formula:
TONS - .29 YARDS
The factor of .29 is based on experience from Grand Rapids. For the "price
of bags," we use the price of a bag. For the period during which bags were
not available and tags were used, we use the implicit price of a bag based on
the price of a tag. Therefore, when bags were not available but tags were
and they cost $.10, we use .167 as the price of a bag.
For income, a better substitute was available in Grand Rapids than in
any of the other cities. Grand Rapids has a city income tax whose rate has
not changed over the period estimated. Therefore, we use quarterly
observations on income tax collections as the surrogate for personal income
with the quarterly observation being used in all three months.
For the seasonal variable, we use a dummy variable for Spring. This
dummy serves two purposes. First, there is a high generation point in the
months of May and June. Second, personal income tax collections peak in
40
-------
April with the filing of returns. Both of these factors are held constant by
the use of the Spring dummy.
Applying ordinary least squares to (39) gives the following result:
ln(q) - 7.1 - .331n(p ) + .021n(y) + .14SPR (40)
(.46) BAG (.25) (.10)
R2 - .08
N =50
As we can see, the coefficients on price and income have the expected signs
but are insignificantly different from zero.
It is unlikely, however, that the effect of price or income changes
are felt immediately. If a distributed lag model is specified and estimated
the results are (standard errors in parentheses):
ln(q) = 5.4 - 1.21n(p ) + .501n(p ) + .291n(p )
(.89) BAG (1.12)BAG-1 BAG-2
- .0091n(y) - .431n(y) + ,701n(y) + .23s
(.37) (.42) -1 (.31) "Z (.11)
R2 = .20 (41)
N =46
None of the coefficients on price are significant and only the second lagged
income term is significant (and of the proper sign).
The positive price elasticities are somewhat troubling even though
they are insignificant. We, therefore, next specified and estimated the
model:
ln(q) =4.4 - .501n(p ) + .401n(y) + .17SPR
(.49) BAG (.21) "2 (.09)
R2 - .15 (42)
N =46
The signs of the coefficients are all as expected. The coefficients on both
the price and income terms are insignificant at the 952 level.
41
-------
Tacoma, Washington
Tacoma, Washington has some fifty years' experience in financing
solid waste collection and disposal through user fees. Over the years the
city has considered or experimented with most of the various fee systems in
current use around the country. The present fee structure is a combination
of a capacity-based system and a location- or service-based system. Hence,
it provides an example of the effects of combining several fee structures.
Residential households living in single unit or duplex structures
receive mandatory weekly collection at a rate determined by the number of
containers presented for collection and the level of carryout service
provided. Households presenting refuse within 25 feet of a legal collection
point (usually curbside, but alleys in some districts) are charged $2.45 per
month for the first can and $1.15 per month for each additional can. Those
presenting refuse between 25 and 75 feet from collection points are charged
$3.90 for the first can and $2.65 for each subsequent can. Between 75 and
200 feet, and over 200 feet, the base prices are $5.30 and $6.65 and marginal
prices are $4.05 and $5.45, respectively. A charge of $1.60 times the number
of cans is levied for each flight of up to six stairs between points of
collection and presentation. These are monthly fees paid along with
electricity, water, and sewage bills on a monthly or bi-monthly basis. In
addition, occasional extra bags of refuse left alongside regular cans are
collected for a charge of $.75 each. Residents are required to pay for the
minimum service of one can at less than 25 feet; about 30% of residential
customers demand some level of optional service, a proportion .that has
steadily increased historically. Records from 1947 and 1958 indicate that in
those years about 92% of all households received the minimum level of
service; in recent years the figure has been about 75%. Households living in
structures of three or more units are charged jointly for service at
commercial rates. Finally, Tacoma residents may haul extra refuse to the
city landfill and dispose of it at no charge; about 40% of all household
refuse by weight is disposed in this way.
The first step in estimating the model for a combination structure
presented above is to express each of the equations in a linear form. This
results in:
ln(q) = Ina + a InC 4- a InS + a Iny -f a.lnZ
U A &* J 4
ln(C) - Inb + b Inp + b InS 4- b Iny
01 2 3
ln(S) = Ind + d Inp + d InC + d Iny
01 2 3
For income, sales tax receipts were used. The seasonal variable used is
precipitation (primarily because rainfall is relatively great in Western
Washington).
42
-------
r*rma «7 -J Varfables 1» c» and S, we have used the detail available in
terms of the numbers of households selecting a particular service
Therefore, the monthly quantity variable is in terms of ?ons per household'
* "
the* c-f h 8vCe eve seeete we
we use the d f V0r? tha° the baSlC leVel °ffered' That ls. *«r cans,
we use the odds of choosing more than one can of service which is the ratio
of the number of households choosing two or more cans of service to the
Using two-stage least squares, the resulting estimates
are:
ln(q) - -5.5 + .85 InC - .34 InS + .23 Iny - .0002 PRECIP
(.93) (.77) (.22) (.003)
ln(C) - 3.2 - .12 Inp - .06 Iny + .41 ln(q) + .61 InS
Ml) (-13) (.42) (.12)
ln(S) = -5.9 + .09 Inp + 1.76 InC + .13 Iny - .87 Iny
(15> (.41) (.30) (.97)
the only significant explanatory variables are between the
chosen.
variableoresentin 8i*«ific* ««» because the dummy
m°nth8 aPPears to be important in many of
ln(q) - -2.88 + 1.0 InC - .42 ln(S) - .02 Iny - .08 WINTER
(-86) (.76) (.21) (.04)
-6 .9
+ 2.7 x 10 , EXT + 9.9 x 10 _ SELF
(5.3 x 10-°) (1.3) xlO-8 R2 . .36 (43)
ln(C) - 2.3 - 18 Inp - .004 ln(y) + .08 ln(q) + .72 ln(S)
(-05) (.06) (.08> (.05)
R2 = .92 (44)
43
-------
ln(S) = -3.1 + .15 ln(p ) - .0002 ln(y) - .14 ln(q) + 1.5 ln(C)
(.06) (.10) (.13) (.13)
R2 = .88 (45)
The results of these regressions are Interesting. Looking at equation (44)
we see that there is a negative and significant price elasticity of .18.
That is, for every 10% increase in incremental can price, the percentage of
households choosing more than one can of service falls by 1.8%. However,
from Equation (43) it appears that waste generation per household is not
significantly related to the percentage of households choosing more than one
can of service. If true, this would suggest that the way people respond to
increases in the incremental price of containers is by demanding fewer
containers and using those demanded more intensively. The price elasticity
for location is, however positive and significant.
A second approach to testing for the existence of negative and
statistically significant price elasticities is to derive the reduced form
equation for waste collected and to use OLS. The reduced form (excluding
SELF and EXT) can be shown to be:
a' a' a' a'
q = a* p p y WINTER (46)
0 L c
When OLS is applied to the logs of (46), the result is:
ln(q) = -2.3 - .47 Inp + .59 Inp + .01 Iny - .10 WINTER
(.26) L (.49) c (.18) (.03)
R2 = .43
N =41
Neither of the coefficients on the price of an incremental can nor on
location is significant at the .95 level.
Because the prices on containers and location are set by the same
authority and change at the same time, it might be suspected that the
coefficient estimates are inefficient (but not biased) because of
muliticollinearity. However, the two terms are not perfectly correlated. To
regress quantity disposed on one of the prices (either container or location)
would result in specification error. Further, because the exact form of
specification error would be one of omitted variables and the omitted
variable would be positively related to the included variable(s). Therefore,
the coefficient estimated for the included variable would be biased away from
zero.
44
-------
CONCLUSIONS
The empirical results from the study are somewhat mixed. Income
elasticities consistent with those found in previous studies have been
estimated for Burbank and Sacramento and additional, but indirect, evidence
of an income effect on service level was found in Provo, Utah. Insignificant
income elasticities were estimated in Grand Rapids and Tacoma.
With respect to price elasticities, the only indication of
significant coefficients is from Tacoma. This finding is, however,
complicated by the fact that there are two price elasticities of interest,
one on incremental containers and one on location. The finding of a positive
price elasticity for container price suggests that some type of specification
error might be the cause.
Additional, and indirect, evidence concerning price elasticities
comes from attempts to examine the effect of user charge on illicit disposal
methods. With data from the waste deposited in city parks (or other
surrogates) we regressed quantity of "litter" on user fee. For no city was a
significant relationship found. This finding is consistent with a finding of
no significant price elasticity.
45
-------
REFERENCES
1. Anderson, F., et. al., Environmental Improvement Through Economic
Incentives, Resources for the Future.
2. Edwards, F. and B. Stevens, "Testing the Efficiency of Alternative
Organizational Schemes," in Evaluating the Organization of Service
Delivery; Solid Waste Collection and Disposal, E.S. Savas and B.
Stevens, Principal Investigators, Report to the National Science
Foundation (New York: Center for Government Studies, Columbia
University, 1977).
3. McFarland, J.M., et. al., "Economics of Solid Waste Management," from
Comprehensive Studies of Solid Waste Management, Final Report, Sanitary
Engineering Research Laboratory, College of Engineering and School of
Public Health, Report No. 72-3, University of California, Berkeley, May
1972.
4. Stevens, B., Pricing Schemes for Refuse Collection Services; The Impact
on Refuse Generation, Research Paper No. 154, Graduate School of
Business, Columbia University, New York, January 1977.
5. Tolley, G.S., et. al., Economics of Municipal Solid Waste Collection and
Disposal Services.
6. Wertz, Kenneth L., "Economic Factors Influencing Households' Production
of Refuse," Journal of Environmental Economics and Management, Vol. 2,
pp. 263-272, April 1976.
46
-------
EVALUATING THE EFFICIENCY OF THE SOLID WASTE CHARGE
by
STEVE BUCHANAN
INTRODUCTION
The national Solid Waste Charge (SWC) policy has received
considerable attention from the EPA over the past few years. The proposal is
now being evaluated by the Resource Conservation Committee, and one specific
legislative version has been introduced by Senator Hart. There is good
reason for interest on the part of policy makers because the potential
advantages of the SWC are quite striking, particularly in comparison to some
of the alternative policies.
It is simple enough to state the logic or rationale of the SWC. The
charge would act as an excise tax on that component of products entering the
municipal waste stream. The amount of the tax would reflect the associated
costs of collection and disposal, causing producers and consumers to reckon
with those costs as they are incurred. So, the charge could, in theory,
correct resource allocation relative to the present situation in which
incremental waste contributions are non-priced and waste flows and materials
consumption are correspondingly excessive. In addition, the charge would
further reduce waste flows and virgin materials consumption by incorporating
a recycling incentive'in the form of exemption (in whole or part) of the
recycled content of the taxed products. Since the charge would, in most
cases, be levied on producers who purchase secondary and virgin inputs, the
exemption would encourage increased secondary materials use. In most
versions of the policy, the charge revenues would be distributed back to
municipalities to fund solid waste management. In that case they would
effectively reduce local property taxes. However, they could be channeled
into other public goods or used to reduce personal income taxation, as well.
Thus, potentially, a solid waste charge can confer benefits both by:
1) improving the efficiency of public finance or pricing of waste
management, and 2) by reducing rates of environmental spoilage and natural
resource consumption. The first objective (which we call "PF") can be
quantified in terms of the resource misallocation or "excess burden" of
various taxes which could be used to finance the needed services. The second
objective, conservation of environment and productive resource stocks ("CE")
47
-------
is much more difficult to evaluate, as everyone is aware. PF and CE are not
entirely disparate since because of existing market failures we are creating
too much solid wasteto the detriment of CE. An efficient level of solid
waste flow, even measured in purely market values, would improve CE.
Now we have reached the point of departure for this papernamely the
"efficiency" of the SWC. The first thing we can say is that if the charge
policy aims at two different objectives there can be no unambiguous measure
or index of efficiency, particularly when one of the objective functions
("CE") has mostly subjective variables for its domain. Second, we must
recognize that the SWC proposal includes a range of design options and that
the problem of optimizing the design of the charge is itself endogenous to
assessing its effciency with reference to either of the stated objectives.
The major variables in the charge design are: 1) the level(s) of the
charge(s); 2) the relative size of exemptions for secondary materials; 3)
the points at which charges and exemptions are assessed; and 4) the use of
the resulting revenues.
Two approaches to evaluating the efficiency of the solid waste charge
will be described here. The first approach is based on a narow view of the
range of policy concerns: that the charge is intended only to internalize
the attributable costs of collection and disposal in the unit price of
products. "Attributable costs" here means some estimate of short or long-run
marginal cost, leaving aside the social cost concerns reflected in CE.
Using consumer and producer surplus changes resulting from the direct
burden of the charge and adding the cost savings brought about by a smaller
total waste flow a lower bound for potential net benefit of the swc can be
found. Then the approximate sensitivity of net benefits to certain
unavoidable sources of error in the level of the charge is analyzed. On the
basis of current (but still relatively crude) empirical estimates of the
variables involved, we conclude that the SWC is not at all likely to be an
efficient means of internalizing (a narrow measure of) solid waste management
costs.
The second approach to efficiency described here acknowledges the
second best nature of all the policy proposals, including the status quo.
Thus the disadvantages in terms of resource misallocation, equity and the
other environmental and conservation dimensions inherent in the status quo of
public finance of solid'waste management and of materials production
incentives would be accommodated. Enlargening the basis of comparability in
this way would clearly allow a more comprehensive statement of the social
opportunity costs of the various policy alternatives. However, this approach
is limited by our ability to determine the social costs of the status quo and
by the prevalence of subjective variables. In addition, a number of
difficult-to-access empirical quantities come into this larger picturesuch
as long-run elasticities of supply of secondary materials, responsiveness of
industry to charges in relative prices or to subsidies, the impact of
environmental standards on long-run disposal costs and many others.
Here we do not attempt a comprehensive analysis of the various
policies, their possible interactions and their probable effectiveness in
48
-------
forwarding the two objectives we have mentioned. Rather we identify some
considerations bearing on such an analysis which would tend to improve our
initial assessment of the solid waste charge. In light of those
considerationsparticularly those relating to the infirmities of the status
quo policyit seems that the SWC could possibly be part of a superior policy
prescription. Some more study would be necessary to confirm or deny that
proposition.
PERFORMANCE OF THE SWC AS A PRICING INSTRUMENT
Here we develop a lower bound measure for the potential net benefits
conferred by an ideal SWC without, a recycling exception. This amounts to a
worst case assumption in that producers do not respond to the recycling
incentive nor take steps to reduce the overall waste content of their
products, but simply pass the charge through. The purpose of the following
analysis is to establish an acceptable measure of the sensitivity of net
benefits to error in the charge level; then we can make some rough estimates
of how well the actual charge is likely to perform given that there are
significant sources of error in the charge level. The main sources are: 1)
substantial variation among municipalities in solid waste management costs;
2) ambiguity in allocating solid waste management costs among waste
components; and 3) error in estimating a national average marginal cost,
thus creating bias in the charge level.
Net Benefits of a Solid Waste Charge
To begin with, the net welfare benefit of the solid waste disposal
charge can be analyzed as the algebraic sum of several components of welfare
change. Initially we will assume a homogeneous waste stream, having
geographically uniform and exactly measurable per unit marginal costs of
collection and disposal. Thus, the level of the national solid waste charge
corresponds precisely to the real resource savings associated with disposing
of one less unit of post-consumer waste. For the time being, long and short
run marginal solid waste management costs are assumed to be constant. The
implications of large incremental costs incurred in switching to a new
disposal site will be discussed later. We are considering, then, an ideal
charge, which perfectly internalizes the relevant costs.
The solid waste charge will raise prices of the waste component of
final products in accordance with the associated solid waste management
costs. Unless demand for the product is perfectly inelastic or if the supply
curve is a horizontal line (reflecting constant costs) the new quantity of
product traded will be less than the original quantity and the new price will
be greater than the original by an amount less than the per-unit solid waste
charge. The increased price and reduced quantity imply a reduction in
welfare for consumers and, in general, for producers of goods bearing the
chargethat is, a loss in producers' and consumers' surplus. Part of the
loss in producers' and consumers' surplus is regained by society in the form
of the charge revenues. Additionally, solid waste management resources are
saved by the reduced waste load corresponding to the lower quantities of
waste-producing goods consumed. That resource savings represents a social
gain. Finally, the amount of solid waste management services demanded may be
49
-------
further reduced by increased recycling, if the charge scheme includes an
exemption for recycled materials. Recycling would lessen municipal waste
management costs to the extent that secondary materials are diverted from the
collection and/or disposal stages.
dearly, the resultant net welfare changes should be positive after
correctly internalizing waste management costs. However, the costs of
administering the charge and of disbursing the revenues must be subtracted
from the total net benefit. For the time being we will assume that these
costs are a positive function of the charge revenues, denoted a(R).
Now we can state the elements of net welfare change in the following
general expression:
AW = - (CS + PS) + WR + SR + R - a(R) (1)
where
CS, PS are the losses in consumers' and producers' sur-
plus, respectively,
SR, SR are the solid waste management cost savings attri-
butable to reduced quantities of post-consumer
product waste and increased usage of secondary
materials, respectively, and
R, a(R) are the charge revenues and administrative costs,
respectively.
It should be noted that some actual use must be made of the charge
revenues, if they are to be reckoned as a benefit. The funds could not
produce benefit if they were, for example, placed in escrow indefinitely,
except for the possibly salutory effects of reducing the velocity of money.
In fact, they should be applied to the margins of social or private
expenditure prevailing at the new relative prices induced by the charge.
Lower Bound of Potential Net Benefits
It would be useful to determine what level of potential benefit an
ideal solid waste charge- would yield. The analysis is restricted to
"potential" benefits because the administrative costs will be left aside.
Other work has been, and will be done to assess the size of administrative
costs. Also, we will initially ignore the savings in solid waste management
costs arising from increased secondary materials use. In other words, firms
pass the entire amount of the charge forward as increased product prices.
These assumptions characterize a lower bound for potential net benefit of an
ideal charge policy, since the recycling benefits are not included.
The expression for lower bound of potential net welfare charge, as
just defined, is the following:
50
-------
AW
- (CS + PS) + WR + R
(2)
It should be noted that the values of each of the right-hand-side
terms in equation (2) would in general be different from the corresponding
terms of equation (1). This is true because the price quantity equilibrium
in the case of (2) is predicated on all of the solid waste charge being
shifted forward, whereas equation (1) includes recycling effects, which would
reduce foward shifting.
The potential lower bound of net welfare benefit can now be found.
Denote the per unit charge level by t. Then the charge causes the final
product supply curve to shift vertically, by the distance t in Figure 1,
from s to s . The resulting surplus loss is represented by the rectangle
p ade plus the "deadweight loss" triangle acd. The consumers' surplus loss
is the area above the old price line, pcb, and the producers' loss is that
below the old price line. The area of rectangle p ade equals t .q , which
is also equal to the total revenues yielded by the charge. Therefore, those
components of total welfare change "net out" as a transfer payment. The
remaining terms, then, are the waste reduction effect and the deadweight
triangle, acd.
price
Figure 1. Effects of an excise tax on the
component of products.
Quantity
solid waste
51
-------
Since the per unit charge, t, accurately measures the per unit
savings in waste management costs, those savings amount to t. q. That
benefit is offset by the deadweight loss which is equal to (t. q)/2. To see
this, note that the area of the deadweight loss triangle acd can be found
as
acd = acb + bed
= ApAq/2 + (t - Ap)Aq/2
= tAq/2
Here it is assumed that the demand and supply curve segments, ac and dc , are
linear. Given the small relative price change brought about by the charge,
this assumption should give rise to virtually no error.
Summarizing these results in the form of equation (2), the net
welfare change, under the assumptions detailed above, can be expressed as
AW = - (tqj + tAq/2) + tAq + tq j^ (3)
= tAq/2
This holds true irrespective of the elasticities of demand and supply, or if
the good is supplied under conditions of constant costs. The elasticities
do, of course, enter in determining the size of Aq, however.
Therefore, a lower bound of potential net benefit of the solid waste
charge may be calculated as the per unit charge times one half of the
reduction in quantity consumed. For example, if a solid waste charge of 3
cents per container were passed through entirely by the container
manufacturer, and the resulting decrease in annual quantity traded were
300,000 units, then the potential annual net benefit of charging direct solid
waste costs for those containers would be $4,500. It should be pointed out
that the benefits of the solid waste charge rise in direct proportion to the
size of the per-unit costs internalized. To the extent that the relative
price of solid waste management continues to rise, the potential discounted
future net benefits of the policy will also increase.
Sensitivity of Net Benefits, to Error in the Charge Level
As has been mentioned, one prominent issue in the design of the solid
waste charge concerns the preciseness of the charge level in reflecting
actual solid waste management costs. Under somewhat limiting assumptions we
will now examine the welfare losses that occur when the charge level and
costs of waste management differ.
We will again assume a homogeneous solid waste stream, having per
unit collection and disposal costs denoted by c. Also, we assume for the
time being that the taxed product is supplied under constant cost conditions
and that competition prevails. Thus the supply curve is a horizontal line.
The per unit tax shifts the price up by the full amount, t, of the tax
52
-------
because the supply curve is horizontal. Because of error in the
specification of the charge level, t and c are not equal. The expression for
net welfare change becomes
AW = - (tq][ + l/2tAq(t)) + cAq (t) + t^ (4)
= Aq(t)(2c - t)
2
The change in quantity traded depends on the charge level and so must
be written asAq(t). Whent t = c, the welfare change is the maximal value,
W* , equal to Aqc/2 as before. Again assuming a constant elasticity of
demand, e , Aq( t) can be computed as
Aq(t) = qE (5)
Let "a" denote a constant of proportionality between t and c, viz., t = ac .
Then we may examine the sensitivity of AW to deviation between the charge
level and waste management costs using the function AW(a).
AW(a) = i|ai (2c - ac) (6)
= (2c.2a - a2c2) fl (6a)
This function behaves as it should, having its maximum of AW*
(=c'iqE/2p) at a = 1, since
;and (7)
A second important and obvious characteristic of W(a) is that it is zero for
a = 2 and negative for larger values. Similarly, for a - 0 W is zero, as we
would expect. Thus, if the solid waste charge overstates the actual relevant
costs by 100 percent then it results in no net benefit and actually reduces
welfare if the overstatement exceeds 100 percent.
However, under the present assumptions, we find the net benefit
calculus is quite forgiving of relatively small values of error. The
relevant part of the parabolic function W(a) is graphed in Figure 2 (for
= 1) and pairwise values of W and charge error are set forth in Table 1. A
10 percent error in the charge level gives rise to only a one percent loss of
lower bound potential welfare, and a 20 percent error corresponds to a loss
of four percent. The rate of benefit dropoff increases more rapidly with
greater deviation between the charge level and the relevant solid waste
management costs. While a 25 percent error in the charge results in only a
6.2 percent loss of welfare, an additional 25 percent error leads to a loss
53
-------
in realized benefit of 25 percent, or a fourfold loss resulting from a
twofold greater error. A third increment of 25 percent erroror 75 percent
errorreduces benefits by about 56 perdent.
TABLE 1. SENSITIVITY OF LOWER BOUND POTENTIAL NET
BENEFIT TO ERROR IN THE CHARGE LEVEL FOR
e = 1
Value of
a (=t/c)
1.0
.9 or 1.1
.8 or 1.2
.75 or 1.25
.67 or 1.33
.5 or 1.5
.33 or 1.67
.25 or 1.75
.2 or 1.8
.1 or 1.9
0 or .2
Percent. error in
charge level
(|t - c|/c)100
0
10.0
20.0
25.0
33.0
50.0
67.0
75.0
80.0
90.0
100.0
Percent of
potential welfare lost
((AW* - W(a))/AW*) 100
0
1.0
4.0
6.2
11.6
25.0
44.9
56.2
64.0
81.0
100.0
For the present purpose we retain the assumption of unitary
elasticity of demand, as a convenient approximation to the range of
elasticities for products affected by the solid waste charge. It would not
be difficult, in a more detailed analysis, to utilize the elasticity and
consumption shares data for major product categories used by Research
Triangle Institute in their economic model of the solid waste charge. The
effect of higher elasticity is to increase the sensitivity of net benefits to
error. In othe words, for products having price elasticities of demand in
excess of one, the parabola in Figure 2 would be more steeply shouldered.
54
-------
AW (a)
AW*
5AW*
a = t/c
Figure 2. Lower bound of potential net benefit as a
function of error in the charge level.
Applicability of the Analysis
Now it is appropriate to inquire whether the results just described
have much relevance to estimating the net benefits of the charge under
Lnn hi T U/?nS* Preliminaryily. we conclude that the method is
applicable and useful as it stands, and can be readily refined further. The
fTctio'ns ° th8 °fhC°nSta"t elasticity and linearity of demand and supply
n^M % ^8 TradiCt°ry' CSnn0t contr^ute significant error except
possibly in those few cases where the charge induces a very noticeable jump
aive riT T ' ^ ^ " iS unlikel? that these assumptions would
give rise to error greater than that which would result in an attempt to
55
-------
estimate entire supply and demand schedules. Clearly there is no payoff in
the kind of refinement.
The major assumption is that of the constant cost supply function.
It seems reasonable that a great percentage of consumer products are in fact
supplied under conditions approximating constant costs, particularly since we
are considering aggregated national markets. Yet it would be useful to
determine what happens to our analysis if that is not the case, and for
certain segments of the market it will not be.
With a rising supply curve some of the tax burden will be borne by
producers. That case, which is of course the general one, is depicted in
Figure 1. There the tax wedge raises the consumer's price by p and reduces
the suppliers' price by ep. Ultimately the producers' share of the tax
would be shifted back to owners of those inputs having the lowest
elasticities of supply employed in the industry, in the form of reduced
rents.
However, producers also have the option of substituting away from the
taxed dimension of the product, which in this case is essentially the
packaging component of the product, or the virgin materials content of the
packaging. There are two basic kinds of substitution possibilities which
allow producers to reduce their tax burden:
reducing the proportion of virgin materials at the basic
materials level
reducing the packaging component of the product, or at
least the taxed dimensions of the packaging component.
For the most products these two possibilities occur at different
stages of the production process, usually in different firms. But the
observaton still holds, regardless of which stage of production is taxed,
that producers having relatively strongly upward sloping supply functions
have greater incentive to reduce the tax. This effect makes the lower bound
measure of net benefit analyzed above more applicable to the upward sloping
supply case, insofar as producers substitute away from the waste component as
a whole. It makes no difference which side of the market brings about waste
reduction effects in measuring the net benefit of the charge.
To the extent the proportion of recycled materials is increased in
order to reduce the tax, our formulation understates the net benefit by
leaving out the recycling benefits, viz., the term "SR" in equation (1). Yet
the value of such benefits, incorporated in the level of the secondary
materials exemption, is the expenditure of solid waste management
expenditures avoided by recycling. This mode of substitution amounts to what
is also basically a waste reduction effect.
Thus, on the whole, the lower bound method just described above
should be a useful and acceptable approximation for policy analysis.
56
-------
To the extent this analysis is actually representative of the net
benefit behavior of the solid waste charge, the message is clear: the
efficiency benefits of the policy disappear entirely if the degree of error
in measuring the relevant costs and in implementing the charge is comparable
to the size of the charge itself. It seems that this in fact is the case.
Assessing Net Benefit Losses
Source of Error in the Charge/Exemption Levels
Several sources of error or bias in the level of a national solid
waste charge can cause misstatement of the relevant costs which the charge is
intended to internalize. They are:
Differences among municipalities in the relevant costs
of waste collection and disposal.
The existence of local user fees which partly or wholly
recover waste management costs in about 10 percent of
local jurisdictions.
Ambiguity in allocating solid waste costs among materials
types and waste components.
Determining the amount of the recycled materials exemption,
based on estimates of the waste management costs avoided
by increased secondary materials use.
Determining the level of the charge based on estimates of
the national average waste management costs imposed by
increased disposal per household at the margin.
The first point above is probably the most significant, because
interlocal solid waste management cost variability appears to be very
substantial. So, the charge overstates or understates the relevant costs in
accordance with the variance of observed costs, in effect creating a
component of random error in the charge from the point of view of individual
communities. We will concentrate on assessing the impact of the error
component here, not only because it probably dominates other kinds of error
or bias, but also because there is at least some of the needed empirical
information available.
The existence of local user fees could create some significant bias
in the charge/exemption levels. Local user fees combined with the SWC would
create a situation of double-charging for waste collection in certain areas.
It seems reasonable to expect that adjusting or discontinuing local user fees
would be part of the start-up costs of the SVJC, if, in fact, the charge
revenues were to be redistributed to cover local solid waste costs.
Otherwise, inhabitants of those areas would be systematically overcharged,
relative to the rest of the nation. This would happen to the extent that
local fees were more highly correlated with waste management costs per
household than the property tax or other sources of funds relied upon
57
-------
elsewhere. Evidence suggests that that correlation is quite poor in the case
of the property tax and noticeably imperfect in the case of user fees. We
will not attempt to assess benefit losses from imposing charges in addition
to local fees. The main points here are that evaluating those losses depends
on the use of the charge revenues and on the second best nature of the
alternatives, which we discuss in more detail later on.
The problem of allocating costs among waste materials components
seems to be a rather severe one, even at first glance. First, those costs
clearly depend on the weight and volume characteristics of waste components.
However, even in a given jurisdiction, the weight/volume cost relationships
can change if packer trucks have different compaction ratios, or if differing
pickup (e.g., for apartments vs. houses) or disposal methods are used. Also,
the weather can significantly alter the relative cost shares of waste
components.
The volume/weight ratios of cardboard and newspaper change
drastically when they become wet, thus creating daily and seasonal changes in
cost shares. Finally, the compaction characteristics of any given item
depend on the mix of waste surrounding it. For example, cans will compress
relatively more when surrounded only by other cans than by corrugated boxes
and plastic containers.
EPA has initiated a project with the National Bureau of Standards to
measure compaction ratios of waste components in a "typical" mixed waste
environment and to devise a cost allocation procedure, using representative
assumptions on pickup, transport and disposal in a landfill. No results are
available as yet, but obviously there will be some error in those cost
estimates. Here we will simply make some hopefully reasonable assumptions
about the size of that error in order to assess the resulting net benefit
losses.
The final two points mentioned above, determination of the charge and
exemption levels, entail both conceptual and empirical questions. The
conceptual issues are the economic bases for the levels of the charge and
exempion. The charge should reflect all costs attributable to the additional
unit of household post-consumer waste. Those costs could be reckoned
relatively narrowly or broadly as we discuss in Section III, below. At
minimum, a more inclusive measure would also reflect costs of capacity
expansion and environmental upgrading of landfills. The more inclusive cost
bias notion would be more applicable in some areas than in others. Likewise,
the exemption should reflect costs avoided by recycling. Those avoided costs
may be collection costs, and/or disposal costs, depending on locality and
type of material.
In addition to bias and errors arising from Improper economic basis
of the charge and exemption, some further inaccuracy will occur in estimating
national average values of the cost components used in setting the
charge/exempion levels. Some of the deficiencies of present best estimates
of marginal collection and disposal costs are mentioned later. The estimates
of exemption levels rely on 'the original cost estimates as well as on other
empirial information which has not yet been adequately sampled or analyzed.
58
-------
The lack of precision in determining national average charge and exemption
values will lead to some degree of bias in the impact of the charge on
economic incentives. Again, we must rely on guesswork in looking at the
economic effects of that bias. Presumably, such effects would tend to abate
if the charge were properly implemented, with some attention placed on
improving data bases and empirical estimates.
Probable Magnitude of Benefit Loss
As should be clear by now, the best we can do at present in assessing
the efficiency of the SWC as a pricing instrument is to approximate a likely
range of potential benefit loss resulting from the various sources of error.
Ideally, we should treat each source of error as a stochastic variable with a
known or estimated distribution, and the total effect should be computed as a
sum of random variables. Also, the welfare loss estimates should be based on
estimated elasticities of demand and supply for the particular markets
affected. For this purpose however, the assumptions of competitive supply
and unitary price elasticity of demand do not appear to be unreasonable. The
consumption data used in the RTI model indicate a demand elasticity of .99
for the non-food, non-beverage segment of the marketwhich bears by far the
greatest burden of the charge on paper and flexible packaging. Also, that
market segment represents 75 percent of personal consumption expenditures.
The major deviations from unitary demand elasticities occur in beverage,
beer, wine and spirits categories comprising 2.7 percent of personal
consumption expenditures. Demand elasticities range from .4 (beer) to 1.7
soft drinks to 2.2 (wine) in those markets. Canned foods, which could also
be relatively strongly affected by the charge, exhibit low price elasticities
(.17) and comprise 1.3 percent of household expenditures. The consumer
surplus loss in inelastic demand markets would be relatively less sensitive
to error in the charge level.
Now it remains to combine our scanty knowledge of the various sources
of error and bias with the welfare loss function of Figure 2 and Table 1.
The variance of local waste management costs can be assesed only rather
indirectly. Data on average cost per ton (collection only) are available for
177 cities of greater than 2,500 population from the Columbia Survey. Table
2 shows the distribution of those costs in 1974 dollars. The national
average collection cost in 1974 was $24.91/ton. Disposal costs averaged
$4/ton bringing the 1974 average total cost per ton to $29. The distribution
of disposal costs (as measured by fees charged to private sector collectors)
also showed considerable dispersion according to EPA staff.
Assuming that the charge was set at $25/ton, the efficiency losses
resulting from variation of local costs are shown in Table 2.
The overall efficiency of the SWC computed on this basis is 65
percent. In other words, only 65 percent of the welfare gains from properly
internalizing waste management costs are realized.- That figure neglects the
sizable net losses that would occur in about 10 percent of the communities
59
-------
TABLE 2. DISTRIBUTION OF AVERAGE PER TON COLLECTION COSTS
AMONG MUNICIPALITIES AND CORRESPONDING SWC
EFFICIENCY LOSSES
Cost class
$/ton
0-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
45-50
50-70
70
Number
in
class
19
29
32
34
15
14
12
8
5
3
6
Percent
in
class
10.7
16.4
18.0
19.2
8.5
7.9
6.8
4.5
2.8
1.7
3.4
Deviation
from charge SWC
level % of efficiency
class median by class
233 (median = 7.5)
100
43
11
9
23
33
41
47
58
66 (medium = 75)
0
.84
.98
.98
.93
.88
.84
.76
.65
.55
Weighted
efficiency
(%)
15.12
18.81
8.82
7.35
5.98
3.78
2.13
1.11
1.87
65%
Based on national average collection costs of $25/ton, 1974 dollars.
60
-------
where the cost are less than half the charge rate. More importantly, the
estimate also is based on average costs, not marginal costs. Only if the
distribution of average costs is similar to that of marginal costs does our
65 percent efficiency value have much significance. Engineering analysis of
municipal waste production functions seems to be the only feasible way to
determine whether the relevant marginal cost would have a broader
distribution than average costs. I am inclined to think that is not the
case, but if it is, then it appears that a big share of the potential benefit
of the charge stands in jeopardy. It should also be noted that if the
relevant cost distribution has a mode in the neighborhood of one-half the
charge rate or less, then net losses occur for inhabitants of those
communities. That could undermine the efficacy of the SWC policy
considerably. The empirical data sample used here does not adequately
represent major population centers, where costs are higher.
Assuming that local cost variation does result in approximately a 35
percent loss in net SWC benefits, it seems reasonable to expect that the
other Implementation problemsestimating precise central values for the
charge/exempion level and allocating costs among waste componentscould
easily reduce the overall efficiency of the SWC to below 50 percent.
Some empirical work is necessary to guage the degree of error in
assigning cost shares among various waste components. In one sense, it seems
to be misleading to think that those shares are fundamentally well-defined.
There is a problem akin to an index number problem in economics, in assigning
those cost shares. Beyond that, there will always be some geographic
variation in cost shares, owing differences in relative importance of
collection, transport and disposal costs, differences in climate and
differences in mix of waste, among other things. If error in assigning cost
shares averaged 33 percent across all goods and communities, then the net
benefit of the charge could be expected to decline an additional 11 percent.
(It may be possible to minimize those losses if the cost shares of relatively
elastically demanded goods, e.g., wine, soft drinks, prepared beverages could
be determined relatively more precisely.)
Finally, if bias in the charge and exemption estimates yielded a
further loss of 5 percent of the benefits, we would find that the SWC yielded
only about half of its potential benefits. (Potential benefits here excludes
waste reduction effects associated with increased recycling. Those effects
do not figure importantly in the RTI estimates of total net benefit.)
According to RTI's recent estimates of total net benefit, a reduction by
one-half would place the net benefit of the charge in the range of $13-$30
million annually, depending on whether elasticities of supply of secondary
materials are assumed to be relatively low or high, respectively.
Subtracting $10 million for estimated annual administrative costs, we find
that the SWC appears likely to be a rather marginal policy alternative. It
should be noted, though, that the RTI estimates are preliminary. Also, they
are based on a $30/ton charge rate, which probably comes much nearer to
average cost pricing rather than marginal cost pricing.
Obviously, further work is necessary before we can assess either the
efficiency or the net benefit of the charge as a pricing instrument with
61
-------
sufficient reliability to make a policy determination. However, if we
consider the SWC as one of several second-best public finance instruments,
then much further refinement of data and estimating procedures appears to be
somewhat lower in priority.
THE SOLID WASTE CHARGE IN A SECOND-BEST SETTING
The previous discussion of the efficiency of the SWC purely as a
pricing or public finance instrument assumes that there are no pre-existing
distortions elsewhere in the economy. Otherwise the demand and supply curves
central to the analysis do not fairly represent the opportunity costs
associated with price and quantity movements. Because of the smallness of
the SWC relative to prices of most of the affected goods, we are dealing with
truly marginal changes, so our estimates of the efficiency performance of the
charge are fairly robust. However, what of the distortions inherent in the
status quo or in the other contemplated policy alternatives? We would not
have formulated our two objectives at all if there were not pre-existing
market failure in the waste sector. So, although the charges may in
actuality yield little of its theoretical potential benefit or may even yield
negative net benefits, the net benefits of the status quo and/or other
alternatives may be worse yet. So, the less-than-ideal expected performance
of the SWC must be matched against that of the other policies. Below we
discuss several considerations arising out of the second-best approach which
could conceivably place the SWC proposal in ascendancy, all things taken
together.
We also discuss the level of charge. The potential net benefits
increase with the size of the cost internalized in the charge but the usual
marginal cost pricing logic combined with current econometric estimates of
marginal cost may yield a charge level which falls substantially short of the
optimal one. We take up this issue first.
Cost Basis of the SWC
Determining what the level of the SWC should be raises two obvious
problems. First is the problem of deciding which elements of cost should be
included in the charge. Even if we leave aside extramarket social cost
(reflecting environmental and conservation objectives) the answer is not
obvious. Having decided on a cost basis there is, then, the issue of
reliably estimating the costs.
The studies now being conducted by the EPA rely on estimates of
long-run marginal cost to set the charge level. There are several margins in
municipal waste management cost functions. In collection the margins occur
in providing service to an additional household, providing a higher level of
service, and in picking up a greater quantity per household. Additionally,
there are costs of transport, transfer and disposal which depend largely on
the location of the disposal site and somewhat on route patterns. Finally,
there are long-run costs of providing new disposal facilities. In the
short-run the solid waste charge would impinge directly only on the quantity
per household margin, and very slightly on transport and disposal.
62
-------
Estimates of long-run marginal costs using the best available
(Columbia University) data indicates a very low elasticity of cost with
respect to quantity per householdcertainly no greater than 15 percent of
average per ton costs. These estimates are unreliable for several reasons:
The sample of cities, while quite large enough, excludes
cities of greater than 700,000 population, where the per
capita costs would be highest. Also, towns of less than
2,500 were excluded.
Costs of transfer stations and incineration are excluded.
Disposal costs are measured by the disposal fees muni-
cipalities charge.
EPA is working to overcome these data deficiencies and to improve estimation
procedures, but even then one would have to question the efficacy of
econometric techniques in producing reliable estimates of the needed kind.
The regressions produced so far show inconsistencies in signs of
coefficients, spotty statistical significance and gross differences in
results on a ton vs cubic yards measure of output. Given the sensitivity of
net benefits of the SWC to the relevant cost level, it seems inadvisable to
rely on econometric estimates of cost elasticities. Engineering-based
estimates would surely be preferable in this instance.
Returning to the question of selecting the cost basis of the charge,
it would seem the empirical estimates of long-run marginal costs could not
capture some of the relevant costs associated with higher overall rates of
disposal. The costs of acquiring new landfills and of operation with new
disposal sites are reported to be increasing rapidly. Those costs must be
incurred sooner without a SWC than otherwise (unless implementing user fees
would reduce wasteloads more than the national charge). Perhaps more
important, the costs of upgrading disposal sites to environmentally
acceptable standards, which may be on the order of $3 ton per year, are not
adequately reflected in the 1974 data. Some portion of both of these
categories of incremental costs should be included in the solid waste charge
level.
The next question concerns the other elements of cost mentioned
above. Should some of those be included in the SWC, so that it approximates
more nearly an average rather than a marginal cost change? This would amount
to using the SWC as an additional taxation instrument to raise social
overhead revenues associated with the fixed cost elements of waste
management.
Optimal Taxation Considerations
Using the SWC as a means of raising social overhead implies the
substitution of an excise tax for a property tax (and to a minor extent local
income taxes). In that case, the charge could be used to recover all the
cost components except those related to level of service which would probably
63
-------
be more efficiently and equitably recovered through a flat rate related to
service level. An average cost recovery scheme would clearly result in a
much higher charge level, (by factor of 3 to 6) than the presently
contemplated marginal cost-based charge. It would create some excess burden
effects, which must be compared to those of the existing property tax-based
financing arrangements.
Research done to date indicates that the static, and especially the
dynamic, efficiency losses associated with the SWC would be noticably less
than those for the property tax. The horizontal equity of the SWC would
clearly be greater, assuming that redistribution of revenues could be kept
reasonably in line with local costs. The incidence of the property tax is
undoubtedly more poorly correlated with household waste generation rates than
the SWC. In fact, the solid waste charge may resemble the formulation for
optimal commodity taxation which prescribe relative small, pervasive excises
on commodities with relatively low price and income elasticity. Some
additional research is needed to evaluate that resemblance further.
Offsetting Pre-Existing Distortions
Repeal of tax laws which lead to excessive rates of mining and
harvest of productive resource stocks does not seem to be in the offing. The
SWC probably could be used to offset those tax subsidies. Whether it would
be an effective instrument for that purpose remains to be seen. Other policy
alternatives which would have similar effects are subsidies to recycling
firms and standards for recycled content. Those policies have other serious
disadvantages, however.
Conservation of Environment and Productive Resources
The SWC may have some significant conservation advantages relative to
other material policies and charging schemes. Since it is a direct tax on
producers, it constitutes a direct incentive for them to substitute away from
waste components of products and particularly virgin material components.
Volume-related user fees may do this indirectly, but may also raise littering
and unauthorized dumping, both of which are very costly not only in
environmental but in fiscal terms. The SWC also directly induces consumers
to substitute away from waste intensive products, which no other policy does.
Finally, it encourages recycling in accordance with market demand through
price incentives. One must be skeptical of lump sum handouts to budding
industries in order to get them started. Commercial capital and
entrepreneur ship should become available when the market is created.
The advantages just mentioned may not be particularly great. Some
imponderable notional values and quite difficult-to-estimate empirical
factors are involved. However, whatever positive conservation and
environmental effect could be ascribed to the SWC will certainly increase as
the level of the charge increases. Here is another argument, then, for
financing fixed-cost related municipal waste management overhead out of SWC
revenues. That approach not only yields excess extra-market benefits, but
also reduces the waste load and overall economic burden on households.
64
-------
The intention here has not been to campaign for the SWC policy, but
rather to describe two hopefully diagnostic points of view in analyzing the
probable benefits of such a policy. Those benefits pertain to the public
finance and conservation objectives implicit in the SWC proposal. The
evaluation method favored by the RCC and followed in Section II above
confronts only the PF objective. And it does that in an incomplete manner,
by leaving aside the second-best nature of the other PF instruments. It
seems that such an approach is definitely too narrow, given the stated
concerns of the Resource Conservation Committee and of the EPA mandate, for
that matter. Hopefully some of the lines of enquiry mentioned here can be
followed further in order to produce a truer assessment of the various
initiatives, relative to both the PF and CE objectives.
This argument gains force when the SWC and other policies, for that
matter, are looked at in the longer view, rather than merely in terms of
discounted net policy benefits for the next few years. Are incentives in the
solid waste generation and management sector genuinely askew? Clearly the
attendant environmental and productive resource allocation costs are rising,
if not accellerating. At this point, I think debate should focus on how to
restructure incentive patterns to create long-term corrective influences.
Quite possibly no single proposal or mix of proposals could outweigh its
distortional and administrative costs, but such a judgment cannot be
supported without examining more thoroughly the disadvantages of the status
quo» in terms of both the PF and CE perspectives, than we have as yet done.
65
-------
ECONOMIC APPROACHES TO SOLID WASTE MANAGEMENT:
EUROPEAN EXPERIENCE
by
W. DAVID CONN
ACKNOWLEDGEMENTS
Information contained in this paper was collected by the author under
contract to the U. S. Environmental Protection Agency. However, any views
expressed herein are those of the author alone. Thanks are due to the many
individuals and organizations in Europe and the United States who assisted
with the research.
INTRODUCTION
This paper discusses the application in Western Europe of (i)
continuously varying user charges for household solid waste services (that
is, charges that vary for each collection according to the quantity of waste
collected), and (ii) product charges (that is, solid waste-related charges
levied on products at the time of manufacture or distribution). The
discussion of user charges is based on research conducted between May -
September, 1978, while the information on product charges was collected
largely between November, 1976 - January, 1977 (with some subsequent
up-dating).
USER CHARGES
Earlier Studies
A 1973 study by the French consulting firm CERU[1] compared the
method of financing solid waste services then generally employed in France
with those used in the U.K., Germany (Federal Republic), and Sweden. The
French at that time typically paid for solid waste services via a tax
collected at the same time as property taxes. The study identified a
continuum of alternative approaches ranging from:
(i) the use of local general revenues (collected via property taxes,
etc-.), with no direct connection between the amount paid by an individual
household and the cost actually incurred (as in the U.K.); through
66
-------
(ii) the establishment of a charge that varies according to the size
and number of containers regularly collected from each household (as in many
parts of Germany); to
(iii) the establishment of a charge that discriminates more finely
between households according not only to the size and number of containers
but also to the distance from the storage area to the street, the distance
between pick-ups, the frequency of collection, etc. (as in some metropolitan
areas in Sweden).
The study pointed out that the use of general revenues may be
justified on the grounds of providing a public service advantageous to all
members of the community, as well as on the grounds of simplicity (there
being no need to collect an extra tax or charge). On the other hand, the
study also mentioned three major problems, resulting from the absence of a
direct link between the payments made and the costs incurred, the likelihood
of inadequate financing (due to competition for scarce funds in the municipal
budget), and the lack of any incentive that can be used to encourage "good"
solid waste practices on the part of households.
Solid waste charges were said to be based on two principles, namely
that the total sum collected should cover the total cost of the service
provided and that the contribution of each user should equal the cost of the
service received by that user. The study argued that the use of charges
gives flexibility in maximizing the efficiency of the service, and that it
provides an incentive for households to cooperate (or be charged
accordingly) . A charging system can be made more or less elaborate, with
varying costs of administration, accounting, etc.
The possibility of establishing charges that vary for each collection
according to the quantity of waste collected was one variation of the
charging approach considered in the report. The system could employ trash
cans (or some other kind of reusable receptacles) or disposable bags.
Trash Can System
In its most complicated form, this would require the collection crew
to make detailed estimates of the quantities of waste placed out for
collection, which would then provide the basis for billing. The study
rejected this method owing to the difficulties of estimating the quantities,
of keeping a daily record, and of resolving queries. A simplification would
be to base the charges just on the number of containers (irrespective of the
quantities contained); another possibility, which would avoid the need for
recordkeeping or billing, would be to levy the charge at the time of
collection, e.g., using tokens attached to the containers. The study was
unenthusiastic about these methods; instead, it presented the alternative of
having households pay a regular subscription for the right to place out a
pre-specified number of containers at each collection (the charges no longer
varying on a day-to-day basis with the actual amount of waste collected).
67
-------
Bag System
The study discussed the possibility of selling bags at a price that
would include the costs of collection and disposal. The advantage would be
the avoidance of record-keeping or billing, but several disadvantages were
also identified; these included the fact that the method ignores differences
in the costs of collecting and disposing of different bags (depending on
their location, the number picked up at any one location, etc.), the problems
of requiring users to go to buy the bags, and the incentives that users would
have to dump waste improperly or to misuse the bags (e.g., by overfilling),
creating difficulties for collection. The study concluded that this method
has to be approached with caution.
A later (1976) study by the same firm examined solid waste financing
options for the community of Beaune[2], a small but growing town (and center
for the production of burgundy wines) in Cote d'Or, near Dijon. At the time
of the study, the municipal service was collecting waste from a population of
about 19,000. The existing system involved the use of bags that were
distributed regularly (usually annually) by the collection service to all
houses, apartment buildings, etc., the number of bags depending on the number
of people resident. The system was financed by a tax which in 1974 covered
only part of the cost incurred but which was increased the following year to
cover the full cost. It was felt that the system was inequitable since the
tax was the same for each household, while calculations showed that the cost
varied significantly, depending on the number of people in the household.
The study proposed either that a charge should be levied depending on
the number of people in the household, or that bags should be sold at a price
covering the cost of the service. In connection with this second
alternative, the study identified as problems the fact that arrangements
would have to be made for selling the bags (and commercial firms might not
wish to act, in effect, as revenue-collectors for the municipality) and the
fact that the system might encourage improper dumping or the placing out of
wastes without bags. On the other hand, the system would be equitable and
would not require a register of households to be maintained. The study made
no conclusive recommendation as to which system should be implemented.
In reality, the community decided against using the sale of bags as
the means of charging for solid waste services, for fear that people would
evade the charge by improper dumping[3]. Instead, the community is examining
how a register of households might be established, so that charges can be
based on the number of people living in a residence.
Actual Experience; Baden-Wurttemberg, Germany
A system of charging based on containers was introduced many years
ago in a part of the Lande (State) of Baden-Wurttemberg, near the city of
Stuttgart, on the grounds that it was an equitable means of pricing[4].
Users indicated that they had paid for the solid waste service by attaching
tokens to each (standardized, 35 or 50 litre) container; they had the choice
of purchasing either an annual "stamp" or individual "strips" for each
collection (the strips being broken when the containers were emptied). The
68
-------
tokens were issued by the local authorities and sold by local banks, without
any surcharge.
One problem was that the strips were sometimes mutilated or detached
prior to the collection, e.g., by children. As a result, an attachment to
the strips was introduced which could be removed by the householder and kept
as proof that the charge had been paid. In practice, containers without
strips were generally emptied anyway, especially when it was apparent that
the strips had been improperly detached.
The prices charged for the tokens were re-calculated annually based
on an estimate of likely expenditures on solid waste services for the coming
year (including separate collection of bulky wastes, provided at no direct
cost to.householders) and an estimate of the number of stamps/strips expected
to be sold. The stamp was sold at a slight discount compared to 52 strips.
The system was used in the cities of Leonberg, Fellbach, Waiblingen,
Reutlingen, and Tuebingen but it was discontinued in all but a part (about
one-quarter) of Waiblingen because of two problems, namely:*
(i) difficulties in estimating the weekly amount of waste collected,
necessary for efficient planning of routes, vehicle capacities, etc.; and
(ii) considerable increases in the quantities of bulky waste and
litter (although the quantities of household waste placed in the standardized
containers decreased). Specifically, the frequency of removal of bulky waste
had to be increased 2 times per year to 8-10 times per year, signifying an
estimated increase in bulky waste volume of 400%-500%. Although the ratio of
household to bulky waste generation is known to depend on the size of
containers used for the household waste (the use of smaller containers
typically causes an increase in bulky waste), it is reported that communities
using the token system generated much more bulky waste than other comparable
communities (using the same types and sizes of containers).
Overall, costs were found to increase. The proceeds from the sale of
tokens were insufficient to cover the costs, and the difference had to be
made up from other sources, e.g., general revenues. However, users generally
liked the system and some applied (unsuccessfully) to the courts for its
continuance.
It may be noted that a 1973/74 study[5] conducted in the district of
Kurten in the county of Kreis (a rural/resort area) revealed that where
residents were charged for the use of bags for household waste, the amount of
waste collected from each household was approximately half by weight of that
collected from an urban district with a normal container system; however, the
amount received at the disposal site (including household, bulky, and
commercial wastes) was about the same (roughly 350 kg per person per year).
1. It has also been reported that the municipal finance officers did not
like receiving revenue in 52 installments.
69
-------
Netherlands
A new Wastes Act was passed in the Netherlands in 1977. The
supporting text[6] mentioned that some communities paid for solid waste
services via charges for bags, and stated (without supplying evidence) that
"such a system encourages illicit dumping of normal and bulky domestic
refuse. Moreover, the high price of the bags is an inducement to stuff too
much into the bags so that they often tear". It is now reported[7] that due
to the expense and problems caused by illegal dumping (as experienced by the
solid waste departments), less and less communities are using this system.
Where it is still in operation, the price of the bags does not typically
cover the full costs of collection and disposal but, rather, an arbitrary
fraction thereof.
Discussion
It appears that the use of systems in which charges vary for each
collection according to the amount of waste collected is not extensive, and
is currently on the decrease. While the potential advantages of increased
equity and efficiency (compared to a flat-rate system) and administrative
simplicity (where billing can be avoided) are widely recognized, they are
thought to be outweighted by the disadvantages, of which the likelihood of an
increase in improper dumping and litter seems to be viewed as the most
serious. Such an increase has been reported in communities where the system
has been tried, although quantitative data unfortunately seem to be lacking.
Interestingly, there seems to have been little discussion in Europe
of the potential use of the system specifically to encourage waste reduction
or resource recovery; rather, the Europeans have been more concerned with
establishing an equitable and efficient method of financing than with
reducing the generation of waste.
Furthermore, the concern for providing a satisfactory public service
(with adequate protection of the public from the hazards caused by improper
dumping and littering) is evidently stronger than any concern for waste
reduction. For example, in many parts of Germany where the municipality
decides the size of container to be used (and often supplies it), there is a
trend toward requiring larger containers to remove any temptation to dump
improperly[A] : in addition, the public can take small quantities of waste
directly to the disposal site without charge, and there is always the
possibility of placing wastes in the litter bins on street corners. In
Denmark, also, solid waste collection and disposal are viewed primarily as a
public service[8]; if it were found that the container provided to each
household were not large enough, its size would be increased.^ Households
can also take wastes free of charge to container sites throughout the cities.
2. To accommodate temporary surpluses of waste, households can purchase
bags at nominal cost. It is pointed out that the marginal cost of
handling these bags is probably very low.
70
-------
PRODUCT CHARGES[9]
At the present time, Sweden, Norway, and Finland each have taxes or
charges on certain beverage containers; the French government has legislative
authorization to introduce charges and is proposing to do so initially on
beverage containers; and legislation enacted in 1977 in the Netherlands
authorizes the imposition of charges on certain goods.
Nature and Purpose of the Taxes/Charges
Norway and Finland introduced taxes for the specific purpose of
discouraging the use of non-refillable containers for beer and soft drinks.
The Norwegian tax on non-refillables is about $.15 (plus 20% value-added tax)
per container and represents a significant portion of the price of the
beverage; when the tax was first introduced in 1974 this proportion was about
30%, but beverage prices have since risen, and it has now dropped to about
15%. The Finnish tax on non-ref illables, introduced in 1976, represents a
smaller, though still significant proportion of the beverage price; the tax
is about $.08 on beer containers (some 6% of the price of the beer) and $.16
on soft drink containers (some 13% of the price of the drink).
The original purpose of the Swedish tax on containers for
ready-to-serve drinks between 20 cl and 3 litres, first introduced in 1973,
was to finance a freeze on certain food prices rather than to discourage the
use of any particular kinds of containers. The tax is small (slightly more
than $.02) and it is added to the cost of non-refillables, whereas it becomes
part of the refundable deposit on refillables (i.e., it is paid on the latter
only when they are not returned).
In France a proposal is being developed to impose a charge (under the
existing legislation) on all containers ranging from 10 ml to 3 litres for
wine, beer, soft drinks, fruit juices, mineral water, and possibly milk
(although milk may be excluded for political reasons since it is currently
subsidized). The charge would be 5-10 centimes (about $.01-$.02) per unit
and would be intended to reflect the approximate cost of handling and
disposing of the container once discarded. The tax would probably be levied
when the containers are manufactured or assembled; the revenues would be
collected by the Ministry of Finance and then disbursed to local authorities
for solid waste projects including litter cleanup, demonstration trials,
etc., as well as to industry (possibly) to finance the development of low
waste technologies. The Ministry preparing the proposal is taking a
pragmatic approach to the development of the charging scheme, keeping it as
simple as possible; for example, no attempt would be made to "fine tune" the
charge according to the disposability or other characteristics of a given
container (except possibly that a distinction might be made between standard
and non-standard containers). No reduction or exemption would be granted for
the use of secondary materials in a container; for while this would almost
certainly increase the level of resource recovery from the waste stream, it
would tend to lessen the effectiveness of the charges in achieving their
primary objective of reducing the rate of generation of wastes.
71
-------
In the Wastes Act enacted during 1977 by the Netherlands parliament,
there is provision for the introduction of "levies" both to limit the
generation of waste and to provide revenue to carry out the provisions of the
Act. The levies may be imposed on those who manufacture, import, hold in
stock for sale, offer for sale, sell, deliver or use certain goods that are
difficult to reuse, difficult to handle, cause an excessive increase in
waste, or are likely to be littered.
However, it is now reported[7] that implementation of this provision
by the present government is most unlikely, especially while problems of
unemployment persist. The provision is considered too expensive to
implement, and potentially too unpopular to be politically acceptable.
Effects of the Taxes/Charges
The Norwegian tax has been effective in preventing any growth in the
use of non-refillable containers for beer and soft drinks. Although domestic
bottlers have traditionally used refillables, beer imported from Denmark and
Sweden was previously packaged in cans; now most imported beer is brought to
Norway for packaging in refillable bottles. In Finland, also, there is very
little canned beer.
In Sweden, on the other hand, while there has been a transition in
the bottle sector from non-refillables to refillables, there was at first no
change and subsequently an increase in the use of cans for beer (despite the
fact that consumers must pay about twice as much for the same amount of beer
packed in a can rather than a bottle). Other effects of the tax in Sweden,
as reported by a government-appointed committee, include a reduction in the
production of all bottles (due mainly to an increase in the number of trips
made by each bottle, as well as to a shift to the use of larger bottle sizes)
and various attempts to avoid the tax (e.g., by introducing new packages
containing volumes outside the dutiable limits and by substituting undiluted
drinks for ready-to-serve drinks). However, the committee also concluded
that the tax has had no significant effects on costs, employment, or the
environment.
Discussion
The taxes/charges have been applied differently according to the
objective(s) intended. For example, it seems that the primary objective of
the current French proposal for product charges on beverage containers is to
improve market efficiency by internalizing in the producer's decisions the
subsequent costs of handling/disposal (so that he has an incentive to take
these costs into account). The monetary amounts involved are small (relative
to the beverage costs) and so the charges are not expected to have major
impact on the physical characteristics of the waste stream. The charges
will, however, raise revenue that can be used to finance solid waste
services. On the other hand, in Norway and Finland the objective has been
specifically to discourage the use of non-refillable beverage containers, and
therefore the latter have been relatively highly taxed.
72
-------
In Sweden, the present duty on beverage containers was imposed not
for waste reduction purposes but instead for an entirely unrelated reason (to
subsidize a price freeze on certain foods); the amount is relatively small,
and while there has been some shift from non-refillable bottles, the overall
impact on the economy and on the environment has been reported to be
insignificant. An increase in the duty is currently being proposed, with the
intention of raising revenue to finance various anti-litter activities.
It is important to establish with what objective taxes or other
fiscal mechanisms might be introduced in the United States. The product
charging scheme that is being studied by the Resource Conservation Committee
(as specified in the Resource Conservation and Recovery Act of 1976) appears
to have the primary objective of internalizing solid waste management costs,
and therefore seems most closely related to the French proposal. Careful
attention should therefore be paid to developments if/when the latter is
implemented. It should be noted that the charges currently proposed in
France would apply only to beverage containers (whereas a U.S. scheme might
apply to all consumer products), and that the French would grant no exemption
for the use of secondary materials (on the grounds that this might encourage
resource recovery at the expense of efforts to reduce the rate at which
resources first enter the waste stream). One of the most striking features
of the French approach is that every effort is being made to keep the scheme
as simple as possible; rather than placing too much reliance on studies that
seek to determine the theoretically "correct" level of charges, etc., the
authorities are prepared to make reasonable assumptions and in general are
adopting a pragmatic line.
A problem encountered in several European countries, and apparently
common to the U.S. also, is that the use of taxes or other fiscal mechanisms
for waste reduction purposes tends to be viewed with great suspicion by
finance ministries and policy-makers in general. It seems that they
frequently have difficulty in understanding that fiscal measures may be used
not only for raising revenue but also for incentive purposes. Furthermore,
finance ministries on the whole prefer to keep revenues and expenditures
separate; they generally dislike funds to be "ear-marked" for particular
activities (such as solid waste projects) over which they may have little
control. An educational effort is required to ensure that those making
decisions at least have a reasonable understanding of the advantages and
disadvantages of fiscal approaches.
73
-------
REFERENCES
1. CERU, Tarification du Service de Collecte et de Traitement des Ordures
Menageres, Prepared for Ministere de la Protection de la Nature et de
1'Environnement, Paris, 1973.
2. CERU, Elimination des Ordures Menageres, Prepared for Ministere de la
Qualite de la Vie - Environnement, Paris, 1976.
3. Montangerand, C., Le Secretaire General, Mairie de Beaune, Personal
Communication, 1978.
4. Schmitt-Tegge, J., Goosman, G., Orlich, J., and Wicke, Umweldbundesaint,
Berlin, Personal Communication, 1978.
5. I.G.T., Investigation of Waste Arisings in Rheinisch-Bergischer Kreis,
1974/75 (quoted by Goosman, G., Umweldbundesamt, Berlin).
6. Netherlands, Regulations Concerning Domestic Refuse, Derelict Motor
Vehicles, and Other Categories of Wastes (Wastes Act), Explanatory
Memorandum, Session 1974-1975.
7. Erasmus, J. H. , Waste Management Policy Division, Ministerie van
Volkgezondheid en Milieuhygiene, Leidschendam, Personal Communication,
1978.
8. Welinder, A. S., Miljstyrelsen, Copenhagen, Personal Communication,
1978.
9. Conn, W. D., Waste Reduction Activities in Europe, Prepared for Office
of Solid Waste, U.S. Environmental Protection Agency, Washington, 1977.
74
-------
THE ADMINISTRATIVE COSTS OF USER CHARGES
by
BARBARA J. STEVENS
INTRODUCTION
Since the passage of the resource Conservation and Recovery Act
increasing attention has been focused on the problem of solid waste
management.! More and more officials and citizens are becoming aware that
solid waste disposal is not a simple matter, and that not only is the
availability of sites for safe disposal of wastes rapidly decreasing, but
that also the costs of such disposal are rapidly increasing. Responses to
the problems of finite resources and finite numbers of suitable locations for
the disposal of the residuals from these resources vary. For example, some
groups attempt to encourage the recycling of used goods by instituting source
separation programs for paper. Others attempt to develop procedures to
create useful products from municipal waste, such as fuel or energy. A third
group agrees that a market failure is at least in part responsible for the
solid waste problem. Correction of this market failure, this group would
argue, would do much to help solve the problem of increasing waste generation
coupled with rising costs of disposal. It is the argument of this latter
group which is addressed here.
Economists concerned about the market failure in solid waste
collection generally cite the fact that generators of solid waste
especially households do not face a price vector for the collection of
such wastes which truly reflects the costs of solid waste handling. In the
absence of such a price vector consumers' decisions are, in general,
suboptimal. The lack of a true price for solid waste collection and disposal
signals that in some sense society as a whole may not be as well off as it
could be.
Whether or not this loss in social welfare can be remedied through
the introduction of a properly specified price vector for solid waste
collection and disposal services is only partly addressed here. The entire
answer depends upon the relative magnitude of the gain in. social welfare to
1. Resource Conservation and Recovery Act of 1976; (IL 94-58)
75
-------
be expected from institution of a price vector reflecting accurately the
costs of solid waste collection and disposal services versus the opportunity
cost of the resources used to implement and administer such a program. This
paper attempts to quantify the resources which would be required to
administer a user charge system for solid waste disposal service. The next
section describes the economic framework in which society as a whole would
benefit from initiation of a user charge system. While social welfare is
most likely to increase when the type of user charge implemented is a
marginal cost based system, practical considerations make implementation of
such a system most difficult. Thus, Section III presents empirical evidence
about the costs of administering various types of user charge systems. The
final section of this paper presents some guidelines about when and how user
charges can and should be initiated and summarizes the empirical information
presented earlier.
ECONOMIC FRAMEWORK
Goods prices do not include waste handling prices, and waste handling
prices do not, in general, vary with marginal changes in waste generation.
One of the "problems" of solid waste management stems from the fact
that the costs of waste disposal are not included in product prices when
consumers make their purchasing decisions. When households are faced with
prices which do not vary with decisions made at the margin, the optimizing
process is intercepted.
Most households in the United States receive refuse collection
services from a public agency, and the most common means of financing those
services is from general revenues. Thus, the household is typically unaware
that it pays about $35 per year for refuse collection and disposal. Even if
the household is aware of the costs of refuse collection, it is not able to
impact its own charges through individual behavior; there is thus no
pecuniary reason for a household to cut back on refuse generation.
Pricing schemes for users of refuse collection services fall into two
broad categories: those where the marginal cost is zero and those where it
is positive. All methods of financing refuse collection can have an impact
on refuse generation; however, those user charge systems with a marginal
price of zero will affect production of waste only through the income effect.
Such systems pother charge the household a flat fee or finance refuse
collection from general revenues. Clearly, the higher the fee (or cost
financed from general revenues) the smaller the disposable income of the
2. Savas and Stevens, "The Cost of Residential Refuse Collection and the
Effect of Service Arrangement," Municipal Year Book 1977.
3. The Cost per ton for refuse collection average about $20; using a
household generation rate of 1.7 tons, the household pays an implicit or
explicit fee of about $35 per year. See Savas and Stevens, ibid.
76
-------
household and (assuming that refuse generation increases with income) the
smaller is the quantity of refuse generated.4 Systems where the marginal
price is positive should impact the volume of refuse generated through both a
substitution and an income effect. Such pricing mechanisms have the very
desirable attribute of causing households to consider the implicit cost of
disposing of the refuse component of a good, allowing households to trade off
goods with respect to their total costpurchase price plus implicit disposal
cost.
While the theoretical superiority of user charge systems with a
positive marginal price are quite clear, they present problems in that they
are difficult and costly to implement. In order to affect the marginal
consumption and disposal generating decision of the consumer it is necessary
for the. user charge plan to contain a marginal price for refuse generated.
In fact, a true marginal pricing scheme does not exist. The closest
approximation to such a marginal price system is the one in which households
are charged by the can or container of refuse presented for collection.
Since the contents of a container of refuse can vary widely, even this system
does not truly present the consumer with a valid marginal price vector to
impact on his marginal consumption decision.
In short, economic theory would indicate that institution of a
pricing scheme or user charge system for refuse collection which truly
charged consumers for the disposal of each marginal item of refuse generated
would (a) cause a reduction in waste generation due to both the income and
the substitution effects and (b) result in an increase in social welfare to
the extent that foregone use of resources for waste disposal and the value of
goods not disposed of exceeds the value of resources used in administering
the program. Administrtion costs are defined to include the costs of
billing users of the solid waste service, collecting from delinquent
accounts, enforcing the system, and handling of complaints arising solely
from the user charge system. To be complete, one should state that
administrative costs should be net of any changes in productivity arising
from the imposition of the system, as, for example, might occur were the
system that of metered bags.5 In the following discussion, however, no
4. The value of resources used in administering the program can be defined
as equal to not only straightforward administrative cost but also costs
associated with other activities which may be reinforced via the
imposition of the user charge system. For example, littering may
increase when user charges are unpaid. As the cost of litter collection
exceeds the cost of regular solid waste collection, this may represent
(if the increase in the propensity to litter is great enough) a real
drain on resources. See McFarland, et.al. Comprehensive Studies on
Solid Waste Management, Final Report, Sanitary Engineering Research
Laboratory Report 72-3, University of California, Berkeley, 1972.
5. Hudson, Issues in Pricing Solid Waste Management Urban Systems Research
and Engineering, Inc., Cambridge, Massachusetts 02138, #68-01-4362 for
USEPA, June, 1978 finds significant savings in costs due to productivity
increases associated with initiation of a metered bag or other user
charge system.
77
-------
attempt will be made to account for anything other than the straightforward
administrative costs of a user charge system.
From a pure, maximize-social-welfare-point-of-view , then
quantification of several factors is an essential first step in evaluating
whether a user charge system is desirable. Assuming that one institutes a
marginal based system in a market where prices for refuse collection are
already present, one first needs to know the price elasticity of refuse
generation. From this information, together with known data on current
refuse generation patterns," it is possible to calculate the quantity of
resources which would be freed from entry into the disposal stream and thus
saved for other, presumably superior, uses by society. Second, the quantity
of refuse going into recycling patterns should be determined. These
quantities would already be included in the first category as refuse not
going to the final disposal site; they should be considered as resources
newly being directed -to an optimal use, as a direct result of the institution
of the new pricing scheme. These two factors represent the benefits to
society of the user charge system. No attempt is made here to value this
latter benefit stream. In order to estimate net benefits to society, it is
of course necessary to know the true marginal social cost of refuse
collection and disposal. Net costs are the costs of administering the system
less the net costs foregone for refuse not generated and not collected.
While this discussion is straightforward and presents no new
theoretical insight, it does highlight the basis on which the user charge
system can best be evaluated. First, how large are the benefits to society
from institution of this system? Where the system is one in which a flat or
lumpy user charge system is instituted instead of a true marginal cost
pricing scheme, it is possible that there might be no reduction in refuse
generation and, thus, no social benefit. Without any attempt being made to
evaluate the benefits from reduction of the consumption stream, we can say
that a user charge system is desirable where the value of resources freed
from refuse collection and disposal activities is at least as great as the
cost of administering the program. An'attempt at this final calculation is
presented in the next sections.
EMPIRICAL EVIDENCE
Two approaches will be used to evaluate the magnitude of
administrative costs for user fees relative to the cost savings resulting
from initiation of such fees. First, two cities' experience with such fees
Several studies have indicated that refuse generation increases with
income. See Center for Urban Studies (A Condensation of Battelle
Columbus Laboratories) Some Economic Factors Affecting Demand for
Municipal Collection of Household Refuse, University of Chicago, August,
1973, 47 pp., J. McFarland, et. al. Comprehensive Studies of Solid Waste
Management, Final Report Sanitary Engineering Research Laboratory,
College of Engineering and School of Public Health SERL Report No. 72-3.
Berkeley, California, University of California, (May, 1972).
78
-------
will be presented. Each has a user charge system which is variable by
quantity of refuse collected. The cities differ dramatically in how the
charges are administered and consequently, in costs. The low cost city
illustrates a "best case" for the administrative cost of a user charge system
while the contrast between the two cities highlights procedures to be
followed in instituting an efficiently administered system. Second, overall
data is presented for thirty-eight cities located throughout the United
States that have municipal refuse collection service and bill households for
this service. These average administrative cost data are compared with
refuse generation and the costs of service delivery to obtain average cost
figures for administering a user charge system.
Having estimated the administrative costs of user charges on a best
case and an average case basis, these data are related to best case and
average case estimates of the cost savings to be expected from imposition of
the user charge system. These data are drawn from a 1974' survey as well as
other engineering and economic estimates. Together, the probable impact of
user charges can then be evaluated.
Best Case-Worst Case
The two cities selected to illustrate the characteristics of an
efficiently designed user charge system are both located in the Western part
of the United States.8 Each was paired with another city as part of a series
of case studies comparing cities whose costs differed but which were alike in
refuse collection service delivery, scale and location. The matched cities'
respective residuals from an estimated cost equation were at least two
standard deviations apart, with one city's residual being positive and the
other's negative. One city selected for discussion here was the low cost
outlier of the matched pair located in the Mountatin Region while the other
was the high cost outlier from the matched pair located in Northern
California. The high cost outlier, Norcahigh, is, not surprisingly, here the
city with the high costs for administration of the user charge system. The
low cost outlier, Mountainlow, is associated here with an extremely efficient
low cost system for administering a user charge system. The costs of refuse
collection and administration are greater in Norcahigh than in Mountainlow
for several reasons, one of which is different ways responsibility for
billing and collection of fees is allocated in the two cities.
Norcahigh has a very complex way of administering their variable fee
user charge system. Norcahigh has a contract with WCF for the collection of
residential refuse. WCF is a national firm that has been in business in this
city since the mid-1970's. Residential service in Norcahigh is once a week
curbside at the base rate of $3.11 per month per household for two cans.
Additional cans are collected at $1.00 per month per can. A resident may
also pay a higher fee for backyard service.
7. See Savas, ed., The Organization and Efficiency of Solid Waste
Collection, Lexington Books, Massachusetts, 1977.
8. Stevens, Norcahigh-Norcalow available from Ecodata, 1700 Broadway, and
Mountainhigh-Mountainlow.
79
-------
The Office of Property Code Enforcement in the City Government is
specifically empowered to enforce contract provisions. It deals with
delinquent accounts and unanswered complaints, and is the liaison with the
health department for matters such as vehicle inspection. This office is run
by a deputy health officer; twenty-one of its employees spend part of their
time monitoring the refuse collection contract. In the last quarter of 1976,
1,000 liens against property were issued due to delinquency in payment of
refuse bills. The large numer is no doubt the result of a notably cumbersome
billing procedure.
WCF must act as a bill collection agency as well as a refuse
collection firm. More than ten full-time employees work at WCF to update the
list of accounts to be billed for refuse collection service. Owners of
property are supposed to institute services themselves, but, in fact, the
firm generally must itself identify new residents in order for billing to
begin. Once a resident has been billed, the firm's collection problems have
only begun. If a resident does not pay, very specific procedures are
followed. First, WCF sends a delinquency note. If there is no response to
the delinquency note, after sixty days, WCF sends a second delinquency notice
and simultaneously files a copy with the health officer indicating who has
not paid and how much is due. If the bill remains delinquent for seventy
five days, WCF gives the bill to the city for lien and special assessment
proceedings. After twenty days, the health officer can send the owner
notificaion that he owes the bill plus ten percent charge and also runs the
risk of incurring a lien on his property for the delinquent amount. At that
time, lien proceedings are begun and the health officer notifies the city
council which in turn posts a time for a hearing on the proposed lien. After
the hearing, if the city council decides the lien is justified, it is
established as the greater of $25 or ten percent of the refuse collection
service charge. The lien is recorded, the owner is notified officially, and
proceedings to have the sum of service, delinquency charges, administrative
and assessment charges specially assessed against the owner's property begun.
This occurs if the lien is still not paid. If the special assessment is
approved upon hearing by the city council, it becomes part of taxes payable.
Two private firms, Oaks & Sons and MoLo, service residential accounts
in Mountainlow. Each has an exclusive area and is franchised to service all
households in the area. MoLo collects mostly from recently annexed areas of
MoLo county, while Oakes & Sons picks up from the majority of households in
Mountainlow. Oakes & Sons has been in business for thirty-one years. For a
fee, households have a choice of location (curb or carry out) and quantity of
refuse to be collected. In Mountainlow, the base fees of $2.25 per month for
one can, $2.75 for two cans, $3.10 for three cans, and $3.45 for four cans
also include costs of billing and disposal. Unlike in Norcahigh, however, in
Mountainlow these two functions are performed by the city government, which
takes 21% of gross fees to pay for billing costs, disposal expenses and
franchise fees. In both cities, yard trash is collected along with regular
commercial customers on residential routes, although the fees for commercial
establishments generally differ from those for households.
Mountainlow's City Attorney and the Council take responsibility for
writing franchise agreements, for selecting the refuse collection contractor,
80
-------
and for determining annual charges. In addition to these functions, this
city has a separate complaint and billing department, employing an equivalent
of four full time persons to provide the complaint handling and billing
services for refuse collection.
Mountainlow's City Attorney and the Council take responsibility for
writing franchise agreements, for selecting the refuse collection contractor,
and for determining annual charges. In addition to these functions, this
city has a separate complaint and billing department, employing an equivalent
of four full time persons to provide the complaint handling and billing
services for refuse collection.
Mountainlow operates a computerized billing system for Oakes & Sons,
employing two full time equivalent persons in that function. Residents are
divided into groups, each of which receives quarterly bills on different
months. The city receives both weekly and monthly status reports on all
accounts. Customers wishing additional service on an occasional basis attach
a coupon (obtained from Oakes & Sons) to each extra can or bag of refuse; the
loader remits this coupon to the city, and the resident is billed for the
additional service on the next regular mailing. The city provides Oakes &
Sons, at no charge, with a monthly printout of the status of its accounts.
Oakes & Sons must pay the city for additional computer services, such as
mailing lists.
The city remits a monthly check to Oakes & Sons. The amount the firm
receives is gross receipts less bad checks, a 3% franchise fee, a 6% billing
fee, and a 12% landfill fee. The city's concern with low cost service is
evidenced by a current controversy about the amount charged the city by the
county for use of its landfill. Since 1973, Mountainlow has claimed that the
county is charging too high a landfill fee (i.e., 12% of receipts), since its
charge to individual residents is much lower. The city has withheld half of
the 12% it owes, holding the 6% in a trust account. Despite the
disagreement, Oakes & Sons in effect pays 12% of gross receipts for disposal,
and, netting out all charges, receives (in 1976) $1.78 per household per
month for refuse collection services.
Mountainlow has a city operated complaint handling facility. Two
persons work full time receiving complaints and monitoring their handling.
The city opened its complaint and information center in 1974. It is staffed
by a former public relations consultant, and his assistant. Residents
usually call Oakes & Sons with a complaint. If the problem is not rectified,
they then call the city. When the complaint office receives a call, a staff
member writes it up, trying to record the complaint in the caller's own
words. The office personnel will try by tactful, questioning to discover
whether the resident has observed all the "hard card" rules printed up and
distributed by Oakes & Sons. Complaints are followed up by the city
sanitarian, or the police department if the complaint involves a tip-over by
dogs.
There is a regular daily procedure for handling complaints. The
sanitarian comes to the complaint office each day at 1:30 P.M. He collects
all the complaint slips and investigates by contacting either the resident or
81
-------
the private contractor. Attempts are made to resolve complaints the same day
they are received. This is of great concern to the city.
The main complaint is missed collections or litter around the can.
Complaints increase during the summer months, as corroborated by the
complaint officer. Complaints are categorized by number and type. Indeed,
there is even a list of "chronic complainers". As is often the case, missed
collections are the most common problems, with lids off the can the second
most frequent complaint.
Mountainlow is involved in the service on several levels. On a
day-to-day basis the city monitors the quality of the service through its
complaint system. On a monthly basis it monitors the financial activity of
the firm through its billing system. Lastly, the city has very specifically
defined in its ordinances the parameters of the service it wants and the fair
price that it will pay. Approximately four full time equivalent municipal
employees are devoted to the refuse collection service. These employees are
paid from the franchise and billing fees charged to the contractor.
Table 1 presents a summary of the best-worst case history of
administrative costs for these variable fee user charge systems. Both
Norcahigh and Mountainlow have quite similar collection costs per
household$25.29 and $23.40 respectively. Due to differences in generation
rates in the two cities, the cost per ton collected in Norcahigh exceeds the
$13.47 amount achieved in Mountainlow by 35%. The different administrative
procedures followed by the two cities for billing results in an even more
striking cost difference. Mountainlow's cost per household per year for
billing is but $1.64, while that for Norcahigh is $4.48over four times more
than its comparison city.
While administration costs of a user charge system are defined here
to refer mostly to the expenses associated with billing, some extra
complaints may arise due to the charge system, and these expenses should also
be included. For example, households coulfl complain about inaccurate billing,
or about service so unsatisfactory as to justify non-payment of invoices.
While there is no evidence to indicate that the type of complaint or
frequency of complaint received in these two cities is any different from
that received in cities without user charges, the costs of complaint handling
are presented in Table 1. Including these costs as part of administrative
costs of user charges would result in an overestimate of the magnitude of
that cost. As shown on Table 1, the complaint handling costs in Norcahigh
averaged 5.1% of the annual per household collection costs while the same
function in Mountainlow averaged at 8.6% of the annual collection costs per
household. Together, then, the administrative costs of billing and complaint
handling amount to 22.9% of collection costs in Norcahigh$5.97 per
household per yearand 16.5% of collection costs in Mountainlow$3.86 per
household per year. Excluding complaint handling reduces the annual per
household costs for administration to $4.48 and $1.64 respectively.
It is clear that the cumbersome procedures adopted by Norcahigh for
billing result directly in very high costs. The smoother-working procedures
in Mountainlow contrast sharply with those in Norcahigh. Billing by the same
82
-------
TABLE 1. REFUSE COLLECTION COSTS: COMPARISON OF NORCAHIGH
AND MOUNTAINLOW
Cost measure Norcahigh Mountainlow
1976 costs
Total cost $3,160,978 $468,000
Number of households served 125,000 20,000
Cost/ton $18.25 $13.47
Cost/cubic yard $ 4.44 $ 3.36
Cost/household $25.29 $23.40
Billing data
Billing personnel (full time
equivalent)
City 10 2
Firm 10 2
Billing costs as % of cost/household 17.8% 7.9%
Cost/bill $ 1.12 $.41
Complaint data
Complaint personnel (full time
equivalent)
City 4 2
Firm 2 .5
Complaint cost as % of cost/household 5.1% 8.6%
*It should be stressed that the fees charged by both these firms are at
least 25% less than the expected costs for a municipal agency delivering the
same services to its residents as do these firms to the residents of these
cities. All costs are exclusive of disposal fees and percentages are as
percent of collections costs only.
83
-------
agency which has responsibility for termination of service for non payment
seem highly desirable, as does joint billing for several services. Billing
for utilities and refuse collection at the same time makes record keeping of
new arrivals in the city much easier than when refuse collection is
separately billed. New arrivals must have utilities officially connected in
order to start receiving service whereas this is not necessarily required for
waste handlingcollectors often service refuse containers set out at the
proper time and place. In short, joint billing with utilities facilities
accurate billings of all service recipients, decreases the deadbeat problem,
and results in lower administrative costs.
Survey Results
In a 1974 survey of 101 cities providing municipal refuse collection
services to residents, cost data was carefully collected on site using
consistent cost categorizations.° One of the categories for which data was
obtained was billing costs in cases where cities had user charge systems.
Thirty-eight of the 101 cities surveyed had user charge systems of one type
or another: five based their charge on quantity collected while the
remainder charged a flat fee or variable fee based on the frequency or
location of pickup service selected by the recipient households. Figure 1
and Table 2 display the data collected in the survey.
The data show little relationship between billing costs and city
size, though the variance in billing cost as a percentage of the total costs
of collection does decrease with increases in city size. Several other
factors were calculated from these survey data. The average number of
households served in the cities with user charges studied here was 14,513;
the market ranged from under a thousand to almost ninety thousand. The
average cost per city for billing was $12,000. This amounted to about
eighty-six cents per household per year for billing expendituresthe range
was from a disconcertingly low three cents per household per year to $2.75
per household per year. Of the thirty-eight cities in the sample, 43% used
joint billing. Contrary to what one might believe, those cities which used
joint billing had higher costs attributable to residental refuse collection
bills than did those cities which billed refuse collection separately. This
result is so counterintuitive that one must conclude only that cities where
more than one department pays for billing are more likely to keep accurate
records of the true costs of billing and to enable accurate separation of
these costs than do cities where billing occurs within a single department.
Very often, where billing occurs within a single department, in collecting
the cost data, interviewers would be willing to lump together some of the
true costs of billing (for example, the cost of a clerk's time) in other
categories rather than separating them out, so long as the total cost figures
for refuse collection service delivery were accurate. This readily explains
why the cost per household for joint billing might exceed those for single
billing.
9. See Stevens, "The Cost of Residential Refuse Collection" in Savas, ed.,
op. cit. for a description of the survey.
84
-------
Billing Cost as % of
Total Co ection Cost
6
3 X
O
Key:
x = joint billing
o = single bill
10 15 20 25 30
Number of Households (Thousands)
35 40 45 80
Figure 1. Billing costs by city size.
85
-------
TABLE 2. USER CHARGES AND ADMINISTRATIVE COSTS: 1974 DATA
Households served
Service level
% of households with Ix/week service
% of households with curb or alley
service
Monthly charge per household
Billing' costs:
Average cost/city
% of cities with joint billing
Average billing cost per household
Average billing cost per household
for cities with joint billing (?8)
Billing cost/Total Collection Cost
Billing cost/Total Collection Cost
for city with joint billing
Mean
14,513
30%
67%
$2.70
$12,506
43%
$.86
$1.28
3.18%
3.80%
Minimum
762
0
0
$1.00
$50.00
-
.03
.11
1%
1%
Maximum
88,143
100
100
$6.00
$88,312
$2.75
$3.75
6.83%
8.33%
86
-------
Overall, the average billing cost as a percentage of collection cost
was 3.18%. In cities with joint billing an average o'f 3.8% of collection
costs were attributable solely to the administrative expenses of billing for
refuse collection services. As this number seems more reliable than that for
single billing of refuse collection services, we can take the 3.8% figure as
a low estimate of the administrative cost of user charges.
Although the 3.8% figure may indicate a reasonable: level for the
expenses of instituting a user charge system, it should be remembered that
the administrative cost of instituting a quantity based user charge system
might be much greater than that figure. This would be so because charging on
the basis of quantity requires more interaction between the billing system
and the refuse collector himself. When the refuse collector himself must be
involved with determining the eligibility of a household's presented refuse
for collection at each stop, work productivity is probably decreased, which
may in effect increase the administrative cost of a system more than with a
non-quantity based user charge system.
While not as detailed as one might like, these data do represent some
of the best information on the actual costs of administering a user charge
system. Further, detailed study on these matters would of course tend to
increase the assurance with which policy decisions could be made.
CONCLUSIONS AND POLICY IMPLICATIONS
The empirical evidence has shown that the administrative cost of user
charge system are on average likely to be at least 3.8% of the total cost of
residential refuse collection. Indeed, for two cities where billing costs
were collected with great care and detail, it was found that the
administrative cost of billing ranged from a low of 7.9% to a high of 17.8%
of collection and disposal cost per household.
The 3.8% figure is representative of costs where billing is mostly
done on a non-quantity basis and by public agencies billing services jointly.
As such, it represents the low end of the administrative cost scale. The
7.9% figure is consistent with a quantity based user charge system where the
billing is performed by a public agency billing jointly for several services
at one time. The 17.8% figure, the high end of the administrative cost
scale, is obtained when a private firm bills individually for a quantity
based service.
These three data points can be used to derive estimates of the likely
impact of imposition of user charge systems on social welfare. Two scenarios
for imposition of the quantity based charge will be considered:
1) imposition of a quantity based charge over an existing flat fee
user charge system, and
2) imposition of the quantity based user charge system over a
service financed through general tax revenues.
87
-------
Both of these systems are hypothesized to impact household refuse generation
and the costs of refuse collection in similar fashion. Financing the service
through property taxes as compared to a flat fee user charge differently
affects aggregate refuse generation only to the extent that the incidence of
the property tax differs from equality across individuals. If the property
tax incidence is sufficiently regressive to approximate the impact of a flat
fee user charge in a heterogeneous community, then little difference in
aggregate refuse generation would be expected as a result of the financing
scheme. Many smaller communities are quite homogeneous in terms of income,
and there the probability is high that the incidence of the flat fee user
charge and the property tax would not be too different. It is assumed, then,
that household refuse generation is the same under the flat fee or the
property tax financing scheme. The imposition of a quantity based user
charge system on these financing schemes will, however, result in different
increases in the administrative costs of the two systems.
Figure 2 portrays graphically the marginal savings and costs to be
expected from imposition of a quantity based user charge system on the two
financing schemes. In both cases, households generate Q refuse prior to
imposition and Q refuse post imposition. The flat fee system is shown as
saving some of its pre imposition billing costs as refuse to be billed for
decreases after imposition of the user charge system.
Both systems also reap whatever cost savings are attributable to
collecting and disposing of a reduced quantity of wastes. The quantity based
user charge system is showed as causing the equilibrim price to equal 1.079
times the property-based cost of refuse collection. There is, obviously a
greater increase in billing costs for the property tax financing scheme than
for the flat fee user charge system. Clearly, then imposition of the
quantity based system is desirable in either case if the extra administrative
costs do not exceed the cost savings from reduced generation of refuse.
Estimating the size of the cost savings from reduced refuse
generation is difficult. There is no clear cut evidence about households'
reactions to imposition of such a charge. Studies of existing cities with
quantity based user charge systems indicates that the price elasticity of
refuse generation may be very low, but this evidence is by no means
conclusive. Rather than estimate these refuse generation responses directly,
then, they will be calculated as a residual, indicating the minimum response
necessary to offset increased administrative costs. These estimates are
presented in Table 3.
Table 3's data is based on the following general assumptions: that
the additional administrative costs imposed in moving from a property tax
financing scheme are $2.96 per household per year; that the comparable figure
for moving from the flat fee user charge system is $1.53 per household per
88
-------
, quantity based
fee schedule
1.038x=flat fee cost
x-cost of collection
Q* Q° Q
Total Tons Collected
Property tax system:
Collection
Cost Savings
Administrative
Cost Increases
Flat fee system:
Figure 2. Costs and savings from imposition of a
quantity based user charge system.
89
-------
TABLE 3. REFUSE GENERATION RESPONSES REQUIRED
FOR ACCEPTABILITY OF QUANTITY BASED
USER CHARGE SYSTEM
General assumptions:
Refuse per household, pre imposition 1.5 tons/year
Cost per household, pre imposition $37. 50/year
Extra administrative costs household
Property tax to quantity based user fee $2.96
Flat fee to quantity based user fee $1.53
COST SAVINGS FROM REDUCED REFUSE GENERATION
#
Alternatives: Required refuse generation reduction
Property tax to Flat fee user charge
quantity based to quantity based
user charge user charge
Cost savings = 90% of average 8.8% 4.5%
collection costs
Cost savings = 60% of average 13. 1% 6.8%
collection costs
Cost savings = 25% of average 31.6% 16.3%
collection costs
* Reduction required such that cost savings will equal increase in
administrative costs.
-------
year.10 Three alternative assumptions are made regarding the collection cost
savings which would be experienced as a result of collecting less refuse at
each stop: savings will be 30%, 60%, or 25% of what the average cost of
collecting the refuse no longer being generated would have been. The 60%
figure results from engineering models, while the 25% figure is a point
elasticity calculated from econometric work. *
If cities are efficient and can reroute and reschedule in response to
refuse generation reductions, then cost savings can be high, and the required
refuse reduction to justify the quantity based user charge system is quite
low4.5 to 8.8%. The converse also holds true. These simple calculations
highlight the fact that for successful Imposition of a quantity based user
charge system, a streamlined administrative system and a somewhat price
sensitive refuse generation behavior are only necessary conditions.
Sufficient conditions require also that the refuse collection organization
itself responds efficiently to the changed levels of demand for its services.
Development of procedures to aid managers in effecting these efficient
responses should be a cornerstone of any quantity based user charge program.
10. These costs are calculated 40 a percentage of baseline collection costs.
No downward adjustment it made for reduced generation affecting thit
product, ae it is assumed that sdmnistrative coit will be more sensitive
to number of sccounte billed than to tha quantity of refuse generated*
11, See Stevens, and Hudson, Work undar progress for the Resource
Conservation Committee, B«P.A,, 1978 and op. cit. 1978.
91
-------
SECTION 3
PRICING AND THE EQUITY OF SOLID WASTE FINANCING
"Equity Considerations in User Charges for Residential Refuse Collection and
Disposal" (Roger Bolton).
"Pricing Municipal Refuse Service: Potential Effects on Municipal Budgets"
(Kenneth L. Wertz).
92
-------
EQUITY CONSIDERATIONS IN USER CHARGES FOR
RESIDENTIAL REFUSE COLLECTION AND DISPOSAL
by
ROGER BOLTON
THE GENERAL CONTEXT
I first want to set the context for considering equity
considerations. We are concerned with the distribution of effects of a
change from one system of public sector solid waste management (henceforth,
SWM), call it System A, to another system, call it B. We are motivated
chiefly by the possibility that System B has user charges, System A does not.
Indeed, in many discussions of this problem writers implicitly or explicitly
assume that System A and System B differ only in that B has charges and A has
some other financing method. More than one author discusses the equity
effects of user charges as if the conditions of refuse management are
unchanged when changes are imposed. By conditions I mean to include the
aggregate quantity of refuse and the costs of handling it, as well as the
distribution of those aggregates over households. Costs here must include
the cost incurred by households as well as the public sector.
I think that is too restricted and too hypothetical a comparison. In
the real world, if System B includes user charges it will probably also show
other changes from System A. Those other changes will affect the total costs
in the public sector. There are the billing costs, for one thingevery one
mentions them. But in addition there are changes in household behavior,
induced by the introduction of charges. Those changes in household behavior
are precisely the changes which are emphasized in discussions of the
efficiency aspects of charges. If the price system "works" when user charges
are introduced, then the aggregate quantities of refuse and the aggregate
public sector costs of handling it will change. And so will the public
sector costs of handling any one household's refuse; one cannot assume the
costs imposed on the public sector by a household's refuse under System B
will be exactly the same as they were under System A. Nor can one assume
that the costs imposed by an income class's refuse will be the same.
Finally, the changes in household behavior will'change the aggregate costs
incurred by households, and they are relevant for an analysis of equity.
Again, one cannot assume the distribution of household-incurred costs is the
same under System B as it was under System A.
93
-------
The upshot of this is that the analytical and empirical problem is
one of "balanced budget incidence," rather than one of "differential tax
incidence" (11, ch. 10). It is balanced budget incidence because both the
income and expenditure sides of the public budget will change: the
expenditure side changes because quantities of refuse and costs of handling
change; the income side changes if income exactly covers costs. The
literature seems to have sidestepped this problem, and dealt with the problem
of differential tax incidence, instead. Writers have used differential tax
incidence, because they have assumed total expenditure on SWM is constant,
and also the expenditure required to handle each household's, or each income
class's, waste.
No doubt, with the uncertainties and controversies over tax
incidence, especially over property tax, analysts have felt they had enough
on their plates if they tackled the differential incidence problem. They
have also had the problem that most of the data on public sector costs are
for cities where System A is in place, and System B has never been tried.
Then, too, some writers feel the differences in costs between System A and
System B are not large enough to worry about not large enough to require
a balanced budget incidence model. They may argue that my initial point is
rather academic. That may be. However, I do find it strange that there is
so little discussion of the point, in analyses of a policy change which is
recommended in the first place on efficiency grounds. Economists recommend
charges because they will induce changes in the quantities of refuse offered
by households for collection and disposal, and, therefore, changes in public
sector costs.
CHANGES IN THE CONDITIONS OF SWM
Changes Initiated by the Public Sector
First, we may have significant changes in the range of choice offered
by the public SWM system. The move to user charges may be accompanied by
changes in the frequency of collection, the regulations on types of
containers, the locations at which containers will be picked up (curb side.
back yard, etc.), regulations on separation of garbage frm other refuse or
separation of newspapers from other refuse, the availability of bulky item
pickups, etc. Whether the range of choice will be expanded or narrowed will
depend on specific circumstances in the municipality. The nature of the
previous system, the impetus to user charges in the first place, and the
socioeconomic characteristics of the residents are all relevant. It's hard
to generalise, and one should probably do no more than point out the need to
keep these things in mind when analyzing the equity effects of a change in
policy.
Are such changes important in practice? Can't one assume them away
when the financing method is changed? I think it would be wrong to do so,
for some very definite reasons. For one thing, SWM is a
politically-sensitive function, and changes in it will produce "flak" for the
managers. That is especially true if changes include a move to an untested
and basically unpopularin the short run--pricing system. If solid waste
managers even thought about introducing any other changes in their system,
94
-------
they may decide to bite the bullet and introduce them along with charges.
They may prefer to endure one period of intensive criticism and then be done
with it, rather than endure several recurring periods of criticism.
For another thing, some changes in the conditions of collection and
disposal may be called for by the efficient and cost-saving administration of
the user charges scheme itself.
But the most important reasons for expecting multiple changes by the
solid waste managers are different. One of the reasons is that a change to
user charges is likely to be stimulated by a crisis or quasi-crisis situation
in the municipality. It may be a budgetary crisis, or a landfill-space
crisis. In either case, user charges are likely to be part of a package tc
cope with the problem. The municipality may eliminate some costly service
options, such as frequent collection, or backyard pickup. It may eliminate
these altogether rather than go to the trouble of pricing them correctly. It
may also try to increase resource recovery revenue to the municipality or to
reduce landfill space requirements; for example, it may require that
newspapers be separated out from other items.
Solid waste managers may find it very difficultvery costly, that
isto maintain the same environmental standards when they substitute System
B for System A. We often worry, for example, that some households will
respond to the user charges by increasing what is euphemistically called
"alternative disposal". This may include illicit disposal by dumping the
rubbish in vacant lots and parks or open burning. It may also include
increased use of garbage disposals in the kitchen, which transfer the
enironmental burden from the SWM system to the sewage disposal system, and
increase the public sector sewage costs and/or the pollution damages
accordingly. Some analysts discuss user charges in solid waste under the
explicit assumption that environmental standards are unchanged. This seems
wrong. Either there will be increased public sector costs to insure the
maintenance of the standards, or else the standards will not be met. In
either case, the usual limited approach of differential tax incidence is
incomplete, and misleadingly so. In the first case, public expenditures will
be affected. Also, in most jurisdictions sewage disposal is already financed
by user charges, the equity aspects of which may be quite different from
those of the general property tax which one usually associates with the SWM
function. In the second case, the distribution of increased environmental
costs is itself an issue; many of these costs are likely to bear heavily on
the poor.
Changes Initiated by Households
There are at least three kinds of changes here. At the risk of dull
repetition, I point out again that they are the very things given the biggest
play in the literature on efficiency, sometimes by authors who go on to
ignore them in their analysis of equity considerations. Furthermore, the
basic theory of household behavior inescapably concludes that the income
levels of households may be a prime determinant of the changes; that leaves
us with an equally inescapable conclusion that the equity or income
distribution Impacts of added household costs needs some careful analysis.
95
-------
Households may initiate these changes:
1. Change in consumption habits in general. Different market goods
and services have different amounts of solid waste attached to them; user
charges for solid waste cause effective prices of goods and services to
change, theoretically leading to changes in consumption. Thus, solid waste
charges might affect the consumption of newspapers, of frozen food, of
beverages, neat lawns, and other things. The net impact on the welfare of
the household after it has made its optimizing adjustments will vary with
income and with some important family characteristics, such as the
opportunity cost and psychic cost of time spent in waste handling by various
members of the family.
2. Change in household time spent in handling waste. Time may be
spent in alternative disposal, separating kinds of waste in order to meet new
regulations or to earn money from recycling, carrying containers longer or
shorter distances for pickup, storing waste more carefully in order to reduce
the frequency of pickup service, etc. Unlike in the previous paragraph, here
we are concerned with the impact on the household of the changes in its use
of time per se, rather than the impact of the changes in its consumption of
market goods and services. However, the same determinants mentioned in the
previous paragraph are operating hereincome and characteristics affecting
the opportunity and psychic cost of using time in various ways.
3. Change in household money expenditures on some particular market
goods and services, which are used in solid waste management in the
household. Chief examples are items used in alternative disposal, such as
automobile and kitchen garbage disposal costs, and items used in waste
storage, such as compactors and containers. These are merely a subset of the
wider set of changes discussed under Section 1, consumption habits in
general, but it is useful to distinguish them because user charges set up
definite incentives for alternative disposal and storage methods. Again,
income and value of time are important determinants of the changes.
My emphasis on household time and on expenditures for market goods
and services is inspired by a Becker-Lancaster household production model.
That model is fruitful in analyzing the household responses which need to be
taken account of; in the household production approach, some broad good like
"household environmental quality" is an argument in the utility function.
The household chooses between its own production activities and public
management services as substitutes in providing environmental quality; the
presence or absence of user charges for the public management services
obviously affects that choice. But user charges also affect the price of
environmental quality per se, no matter how the household achieves it, so
they affect the consumption of household environmental quality relative to
other things which give utility. Thus, they affect consumption habits in
general.
There are also production functions for the household activities in
question, the inputs in which are time of family members and market goods and
services. User charges affect the optimum level of these household
96
-------
production activities, so they affect the optimum allocation of family time
and also the purchase on the market of inputs into production. Thus, when
all the household's reactions to charges (and accompanying public sector
changes) have been worked out and a new optimum allocation of time and money
income achieved, it is possible that all sorts of things have changed. The
mix of environmental quality and other things in the utility function must be
achieved simultaneously with the mix of household production activity in
waste management and the use of public SWM services, and both must be
achieved simultaneously with the optimum utilization of time and
market-purchased inputs in household production activities. After all the
dust has settled, the household may be better or worse off, depending on the
balance between its user charges and the amount of taxes it saves, but also
on the nature of the accompanying changes in the public system, on its own
household production functions, and, of course, on its tastes.
Clearly, the task of equity analysis theoretically should include a
prediction of how the net result of these household decisions varies across
households at the same income level. That is essential for the question of
whether "equals" are affected "equally," or the "horizontal equity" question.
It also must include a method for expressing the net result of the decisions
in some kind of monetary equivalent form, such as a "consumer surplus". And,
finally, it must include a prediction on how the results of household
decisions are correlated with income levelsthat is essential for the
question of "acceptable distribution of income" or "vertical equity".
One way to make the point I am trying to make is to refer to Figure
1. Here we have three different households' demand functions for public
collection and disposal. Under System A, in which there is a zero charge for
the service, the three households all offer the same quantity of waste, 4,000
units per year. However, their demand curves have very different
elasticities, so that if a user charge of $.03 per unit is imposed they
respond in very different ways. At the $.03 price, for example, C's
elasticity is unity, D's is 1/3, and E's is zero.
If for simplicity we accept consumer's surplus analysis, we see that
the households' loss of surplus generally is not correctly estimated by an
amount equal to the new price, $.03, times the quantities offered at the
original position. The impact as measured in that way is $120 for each
household, but that is the true loss of surplus only for Household E. The
$120 figure overestimates D's loss of surplus, which is only $105, and C's
loss which is only $90. This is an obvious and simple point, and needs no
belaboring here. It is an example of the familiar principle that one must not
be content with an analysis based on "marginal" changes when the changes are
actually quite non-marginal. This point is especially relevant in the
present context, for when user charges are introduced for the first time they
will be decidedly non-marginal!
In order to use this analytical point in equity analysis, one must of
course make some conjectures on how the elasticities vary with income levels,
and how they vary across households who are at the same income level, or who
are in the "same circumstances" from the point of view of horizontal equity
analysis. There are hardly any good studies on the price elasticity of
97
-------
0 1,000 3,000 3,000 4,000
Figure 1
98
-------
demand at all, let alone studies on how it varies in those ways. One would
expect, however, that it does vary considerably across households at the same
income level. How the elasticity varies with income is a difficult question.
For example, higher income households probably have stronger preferences for
some market goods and services which generate a lot of wastepackaged foods,
newspapers and magazines, and large well-kept lawns, for example. And they
probably have stronger preferences for a tidy household environment as well.
And they are more likely to have higher opportunity costs of time, at least
if we consider the adult members of the family. These all would seem to make
their demand less elastic. On the other hand, they are more likely to buy
compactors and kitchen garbage disposals, which would work in the opposite
direction.
THE MEANING OF EQUITY
It is customary to speak of horizontal equity and vertical equity.
The first is said to require equal treatment of persons in equal positions;
the second to require equal taxes for equal abilities to pay, which most
politicians accept as requiring some progressivity relative to income.
If we have a genuine public good, or at least a good with such
pervasive positive externalities that the benefit theory of taxation can be
disregarded, there is perhaps a simpler way of expressing the two principles.
Then it is simpler to say there is only one principle: equity requires the
equal treatment of persons in equal positions, but "equal positions" must in
fact be determined with reference to equal ability to pay and "equal
treatment" implies something like equal sacrifice. From the one basic
principle stems both the horizontal equity and the vertical equity principle.
Horizontal equity means requiring persons with the same index of ability to
pay and in the same general circumstances to pay the same tax. Vertical
equity means requiring persona, with different indexes of ability to pay, to
pay different taxes, and a generally accepted implication of this is that as
income increases, taxes should increase as a proportion of income.
This is overly simplified, but seems to me to capture the essence of
generally accepted notions of equity in taxation for public goods. But in
passing it must be admitted that the notion of "equal circumstances," when
circumstances include aspects other than income or other index of ability to
pay, is a slippery one. I have suggested earlier, for example, that certain
family characteristics will affect the responses of households to user
charges. Family characteristics affect the opportunity cost and psychic cost
of the time required for aternative diposal, and also the ability to finance
compactors and garbage disposals. While the incomes of households are
important determinants of these impacts, the impacts are not perfectly
correlated with income. Thus, different families may feel widely different
effects of user charges, even though they have identical incomes. For
example, a family with several young children or a nonworking spouse, who can
handle the rubbish, may feel the adverse effects-of user charges much less
than another family with equal income but different circumstances. How is
that to be evaluated in equity analysis?
99
-------
This discussion has been for the case of public goods. When the
public service is a business-like one, and the benefits and public sector
costs vary considerably from household to another, we might abandon the
previous notion of equity and substitute the principle of benefit taxation.
We are especially likely to do so for a function like solid waste management,
where the service is so business-like that it is left to business to provide
in many places! And where the price needed to cover costs is such a
smallalmost trivialproportion of total income.
The benefit theory means that the recipients of SWM services, should
pay in aggregate a sum equal to the total costs of the services (this may
require a multi-part tariff pricing scheme in order to preserve marginal-cost
pricing). In addition to this aggregate requirement, it requires that the
distribution of payments across households be the same as the distribution
either of benefits received or of the public sector costs incurred.
Now, efficiency requires that the payments required of a household be
based on the public sector costs incurred to serve it. However, here as
elsewhere in economic policy, efficiency is not necessarily the only goal.
The distribution of benefits from the public activity is a relevant equity
consideration; if the SWM benefits were larger, relative to total economic
welfare, we might pay a lot more attention to their distribution than we do.
But the benefits are not very largenot nearly on the same scale as
education or medical care, for exampleso we are much less interested in
sacrificing efficiency in order to get a more equitable distribution of them.
The benefit principle leaves no room at all for vertical equity in
the case of something as small in scale as solid waste management. Or so it
seems to me. If we didn't have to worry about the externalities from
alternative disposal, we wouldn't need to talk about benefit principles of
taxation at all. We could simply talk about business pricing.
However, not every observer will accept the benefit principle. The
principle tends to be most appealing to those who believe firmly in the
desirability of the so-called Musgrave-Tiebout "layer cake" model of
government (8, referring to 11,17). .In that model the functions of
governments are sharply divided into the allocative, distributive, and
stabilizing functions, and they are divided up among the levels of
government, in a federal or other multi-level political system, so that only
the lowest level handles an allocation function like solid waste management.
In the model, SWM services are properly financed by the lowest levels because
their externalities are not widespread geographically. Furthermore, those
levels finance such services without regard for distributional consequences.
They leave distribution to the higher levels of government, which
redistribute income in general way, rather than through price subsidies
which distort the efficient allocation of resources.
Not everyone accepts the layer-cake theory. Some critics would argue
that the distributional consequences of every government policy, even one as
mundane as the management of the garbage, must be considered explicitly by
the level of government to which history and political realities have left
the function. The critics would argue that no local government can assume
100
-------
that higher levels will in fact redistribute income in the desirable
direction, let alone in the desirable, non-distorting way. They may argue
that Thurow's notion of "individual-societal preferences" (16), or Tobin's
notion of "specific egalitarianism" (18), both of which are originally put
forward for goods looming as large as medical care and education, can be
legitimately extended to solid waste. They may claim that the externalities
of SWM are in fact pervasive enough that it is close to the public good end
of the continuum. A related argument is this: in our modern technologically
advanced society, the environmental media are so interconnected and the
production of packaging materials is so far beyond the control of the
individual consumer, that the management of waste is in fact an essential
function for government and it is no longer realistic to expect individuals
to feel any responsibility at all it. These arguments, while they are not
likely-to be accepted by many of us, and while the last one in particular is
quite pernicious, are not totally ridiculous in the politicial arena.
REVIEW OF EMPIRICAL STUDIES
The sensitive and broad-minded solid waste manager will want to know
something about all the changes in the distribution of benefits when he moves
from management system A to management system B, even if in the end he risks
a judgment that some effects are not awfully important. And from what I've
said, it should be clear that he will want to know about all the changes in
conditions of waste mangement, which accompany or are induced by the move to
user charges.
The ideal empirical study for him would be a "before-and-after" study
in some community similar to his own. Alas, he is unlikely to get such a
quasi-controlled experimental result* At the moment, the best he can do is
to look at the few studies which relate quantities of refuse offered for
public collection and disposal to income and other household characteristics,
under one single existing management policy. In other words, he will have to
be content for now with analyses of the variations in so-called "demand" for
refuse services under only one system, which corresponds to System A in our
hypothetical comparison. He will find a few such analyses, but unfortunately
they tend to be silent on some of the things which might actually happen when
System B is substituted for System A. From them he will have to make his own
conjectures on households' elasticities of demand for public service and how
the elasticities vary with income levels and other characteristics.
I shall review the salient features of some of the major empirical
studies here. In this, I am especially concerned, of course, with the
estimates of the income elasticity of refuse generation, as that is very
important in inferring vertical equity effects. There is a strong consistent
pattern to the estimates of income elasticity. In addition, for horizontal
equity, we must note how well any statistical demand function fits the
behavior of households. The more poorly it fits, the more variation in
refuse generation there is, at given levels of household income and other
variables measuring "equal circumstances." That implies something for
horizontal equity. However, rather than review all the details of the
101
-------
studies in that respect, let me say here that the statistical functions fit
well enough for us to be confident on the Income elasticity of demand, but
also poorly enough to indicate that horizontal equity would be seriously
violated under a financing scheme which collected the same amount from all
households at the same income level, regardless of the refuse they generate.
Downing
Perhaps the most useful study which confronts equity aspects
empirically is the one described in two papers by Paul Downing (5). Downing
analyzed variation in refuse quantities, collection costs, and estimated
property tax burdens across 64 collection routes (henceforth CR's) in
Riverside, California in 1971. First, he estimated average volume of refuse
per household in a CR, as a function of average family income, average family
size, and land area per household in the CR. Volume was based on one single
pickup in each CR during February or March 1971. We know that generation of
refuse is highly seasonal, and the income elasticity varies considerably by
season, at least in other parts of the country (2, 14, 15, 19), and Downing
in fact recognizes that his results cannot easily be extrapolated to other
seasons and/or localities. The income elasticity in the equation was .39.
That income elasticity is a "net" or "partial" elasticity, in that
the equation holds constant other household and CR characteristics. As 1
have pointed out elsewhere (2, p. 3), in an analysis of income distribution
effects, the "total" or "gross" elasticity, not holding other variables
constant, is more important. If income and refuse quantities/costs are
correlated, that is relevant even if other variables, correlated with both
income and refuse, improve the fit of a statistical demand equation.
Although Downing does not explain this point very well, presumably the gross
income elasticity in Riverside was .55, for that is the value he uses later
in discussing income distribution effects.
Downing's average income measure for each CR had to be constructed
from Census estimates for Census Tracts for 1970. CR's are not coterminous
with Census Tracts, so he had to construct a weighted average of the Tracts
in each CR, weighting each Tract by the number of households living in the
part of the Tract within the CR. This is apparently the same method used in
the Kemper-Quigley and Tolley studies reported on later. Such an estimate is
valid, of course, only if Tracts are geographically homogeneous. However,
there is ample evidence that Tracts are not very homogeneous as far as income
is concerned.
Downing then analyzed the variation in collection cost across CR's.
He estimated cost per household per year on the basis of careful studies of
labor requirements and user cost of equipment. In a regression, he found
cost to be a function of refuse per household (based, again, on the one
observation per CR in February-March 1971), truck travel times per household
(to capture effects of density), and percentage of housing units which are
single family. All these variables had positive and significant effects.
102
-------
Downing added to collection costs estimates for costs of
transportation and disposal at landfill. He then regressed total management
cost incurred to handle a CR's refuse on family income in the CR; the
estimated income elasticity was .44. Thus, if a user charge were imposed,
and fell on households in proportion to the costs of handling their waste, it
would have an income elasticity of .44.
An analysis of the impact of the alternative financing scheme, the
property tax, requires estimates of the property tax payments by households
in each CR. Downing first computed property values in each CR from 1970
Census block data. He assumed that the incremental property tax for refuse
management would fall on each CR in proportion to the property values. Fully
recognizing the shortcomings of many of his procedures, Downing then
regressed estimated property values against family income, and found an
elasticity coefficient of 1.16; that indicates that the property tax is
progressive, and that the elasticity is well above what the elasticity would
be for user charges based on management costs. Thus, Downing concluded that
user charges would be much more regressive than the property tax.
Remember that Downing found the income elasticity of volume of refuse
collected was .55. Accepting this as a measure of benefits, he concluded
that the user charge would distribute costs in about the same proportion as
benefits, because .44 is very near .55. The property tax method, however,
has a net effect which is progressive, in the sense that tax costs to
households are more progressively distributed than benefits (1.16 is higher
than .55), and also in the sense that tax costs to households are more
progressively distributed than would be user charges (1.16 is higher than
.44).
Kemper-Quigley
One of the most prominent of studies is that by Kemper and Quigley
(7). They analyzed New Haven data on weight of annual household refuse and
family income estimates for 33 residential collection routes in New Haven,
Connecticut, in 1972. They tried a variety of econometric specifications,
using income and other variables to explain the variation in refuse
generation per household across collection routes. The estimates of gross
income elasticity of quantity varied with the specification, but were
generally between .65 and .75 for the better fitting specifications.
For example, in a linear equation, the gross elasticity was .45
(income is the only independent variable), and the net .35 (household size
and density also in the equation). However, log-linear equations fitted the
New Haven data slightly better, and in those the elasticity rose to .82 when
income alone was used, and to .67 when household size (significant) and
density (not significant) were added.
In alternative models for the New Haven data, they used refuse per
capita as the dependent variable, rather than per household, and personal per
capita income as the independent income, rather than household income. In
103
-------
the log-linear versions of these, gross income elasticity was .39 and net
.32. The estimates in linear forms were almost exactly the same. Household
size was not significant in these regressions; density was significant in the
linear form but not the log-linear form.
Kemper and Quigley also analyzed a pooled sample, consisting of the
33 New Haven routes plus municipality-wide data for 35 other municipalities
in Connecticut. There was considerable measurement error for the 35
municipal observations. The degree of fit was very much smaller. The
estimates of income elasticity were .65 (gross) and .51 (net) in a linear
equation in which household size and density were not significant, but had t
values above one. In the log-linear form, income was more significant than
in the linear form, and gross elasticity was .67 and net .45. Household size
was significant.
On the equity issue, Quigley and Kemper note that the total amounts
are small in any event, perhaps only .3 percent of median family income for
their sample of Connecticut towns. Their analysis, however, does provide
"strong evidence that lower income households generate less refuse than
higher income households" (7, p. 89). That pattern can be compared to
existing estimates of the incidence of the property tax nationwide in order
to conclude something about equity. As I explain in the next section, the
authors accept some traditional estimates of property tax burden, which show
it as very regressive; Kemper and Quigley note the older estimates give
elasticities of property tax burden which are even lower than their estimates
of refuse quantities. They conclude, therefore, that the property tax
financing of residential refuse collection is much more regressive than a
system of user charges which is based on quantity collected. They conclude
that it would be "highly progressive" if cities went from a property tax to
user charges based on tonnage. Thus, they disagree sharply with Downing.
It should be noted that Kemper and Quigley did not attempt to
estimate the income elasticity of the cost of collecting refuse. This is in
spite of their having accumulated vast amounts of data on collection cost by
collection route, and analyzed it in other connections. It would have been
most interesting if they had estimated a relationship between cost and income
in CR's.
Chicago Study
I note two other empirical studies here. They are "demand" studies,
which go into considerable detail on the variation in refuse weight or volume
by income and demographic characteristics of collection routes. But they do
not match this information with estimates of the income distribution effects
of various financing schemes in the sample cities. Therefore, they are
relevant here only because they give solid additional support to the
conclusion that the income elasticity of refuse generation is well below
unity.
The study of Chicago's 50 CR's by Tolley, Hastings, and Rudzitis (19;
this is an updated version of a previous study by Sheaffer and Tolley, 14 and
104
-------
15) Is especially important, because it analyzes generation at different
seasons. It also includes some variables which other authors have not, such
as race and variance of household income in a CR. In the specifications
which the authors feel are best, net income elasticities varied between .38
and .66, depending on the season and the exact specification, and gross
income elasticities varied between .50 and .64. The elasticities tend to be
highest in the fall and late fall/early winter, lowest in mid-winter
(February), and lower in the spring and summer than in the spring and the
fall. From this seasonal pattern, the authors conclude that there is a
"basic, year-round volume of wastes, represented by the mid-winter volume,
that is relatively less sensitive to income than 'excess' volumes in other
parts of the year, except possibly mid-summer" (19, pp. 88-89). In this,
they reinforce the earlier conclusion of Sheaffer and Tolley (14, 15).
This is an unusually rich and sophisticated demand study. I cannot
go into more details here, however, because it does not confront the equity
issues empirically.
Springfield Study
Finally, I may note my own unpublished study of refuse generation
demand in 1972 in Springfield, Massachusetts (2; see also 3). My analysis of
the city's 20 collection routes attempted to explain variation in refuse by
variation in several social and demographic variables: property value per
household; percentages non-white, younger than 18 and older than 62; and
percentage of housing units occupied by their owners. These functions fitted
rather well, except in the winter season. I made no effort to construct
income estimates for CR's, having less faith in the homogeneity of Census
Tracts than Downing, Kemper and Quigley, and Tolley. However, I constructed
indirect estimates of gross income elasticity in the following way: (a) I
estimated refuse demand equations using CR's as units of observation, as
noted above; (b) I then assumed that the same demand function would apply if
Census Tracts were the units of observation, and used the function (based on
CR's) and the observations on independent variables (by Tract) to estimate
the refuse generated in each Tract; and (c) I then regressed those estimates
of refuse quantities on income in Tracts to get a gross income elasticity.
The results varied by season: .17 in winter; .44 in spring and also in
summer; and .72 in fall (the fall elasticity is high obviously because of all
the New England leaves falling on the large lots of higher income
neighborhoods).
I had no direct data on costs, and did not engage in the same useful
exercise of estimating them carefully, as did Downing and Kemper and Quigley.
I resorted to very rough estimates of relative collection costs in each area.
To do this, I assumed that in a CR, the collection cost per household
relative to the city average, would be the same as the relative number of
standard collection crews required to service the CR in each season. Each
crew consisted of the same number of men and truckst so I felt able to assume
the collection cost in CR's varied in proportion to the number of crews* By
assuming this, I estimated an elasticity of collection cost per household
with respect to refuse per household: .79 in winter, .75 in spring, .81 in
105
-------
summer, and .63 in fall. When I multiplied those by the elasticities of
refuse quantity with respect to income, previously estimated, I got these
estimates of elasticity of cost per household with respect to family income:
winter, .13; spring, .33; summer, .36; and fall, .46. While my cost data
fell far short of the quality of Downing's, they are consistent with his
results, and extend them to other seasons: the cost of serving households on
a route rises much less than proportionately to family income on the route.
I did not attempt to estimate the pattern of property tax incidence
in Springfield. But I concluded that the property tax elasticities are at
least a bit higher than the refuse elasticities just given. Therefore, I
concluded that if Springfield refuse and costs data were typical, one could
say that "the use of property taxes to finance refuse management may have a
very slight equalizing effect on the income distribution" (2, p. 2).
IMPORTANCE OF ASSUMPTION ON INCIDENCE OF PROPERTY TAX
The most likely alternative to user charges for financing SWM is the
local property tax. Thus, what one assumes is the incidence of that tax is
crucial for his analysis of both horizontal and vertical equity effects of
substituting user charges for it. That was shown by the conclusions of the
Downing and Kemper-Quigley studies referred to in my previous section. The
authors reached different conclusions on the distributional impact, not
because they assumed very different income elasticities of refuse collected,
but because they assumed radically different income elasticities of the
property tax burden. Therefore, I think it useful to delve into the theory
and literature on property tax incidence as part of a discussion of the
equity considerations in SWM.
I have dipped into the voluminous literature on property tax
incidence several times, and I tried to refresh my knowledge of it in writing
this paper. We all know there is a "New View" on this, which argues that
even the residential property part of the tax is much less regressive, and
may even be slightly progressive, than the conventional wisdom of 10-15 years
ago. I think it is generally agreed by now that the New View is sound enough
to force a major revision of thinking, although not all are so bold as to
insist that the tax is regressive or is^ progressive. Note that by "New View"
here I am not referring to the abstract theoretical literature which argues
that the property tax is highly progressive under certain artificial
assumptions. There is such literature, but by New View I mean rather the
Aaron-Netzer-Mieszowski theoretical and empirical studies which amend the
abstract theory in applying it to the real world (1, 13, 9, 10).
Such a champion of the older conventional wisdom as Dick Netzer has
partially recanted his earlier position, and now agrees that the residential
property tax is not as regressive as once thought, and that the total
property tax is probably slightly progressive in a typical jurisdiction (13).
I may add that Kemper and Quigley appear to have relied much more on Netzer's
earlier position, and on other writers using the same assumptions, than on
his later change-of-mind and other representatives of the New View.
(Netzer's earlier study was (12)).
106
-------
Netzer's partial recantation is such a useful review of all the
arguments that I have relied heavily on it in structuring my owa discussion
here. What he is most convincing on is that the whole subject is incredibly
complicated, and that it is rather misleading to speak of "the" property tax
in a "typical" jurisdiction. However, 1 feel the following remarks represent
fairly the opinion of the majority of public finance economists at this time.
1. First, under either the older view or the newer one, the
residential property tax is more regressive or less progressive than the
nonresidential property tax. Netzer takes pains to argue that the
residential part is still probably regressive in a typical jurisdiction, with
an elasticity of property tax burdens with respect to income of about .7,
even though the other parts of the tax and the tax as a whole are
proportional or slightly progressive. Thus, we have at the start a problem
of context: is it the residential tax or the general tax which is the
alternative to user charges? Clearly, it depends on the jurisdiction we are
studying. Most writers assume it is the residential tax which is the
alternative, because SWM services are provided to persons living in housing
units. However, I am sure that many jurisdictions cannot legally or
politically change the residential tax rate without also changing the rate on
other property. Over a period of time, of course, different assessment
ratios for different kinds of property appear, whether legal or not. But
ratios would not change merely because user charges are subsituted in one
public function; namely, SWM. Thus, at the margin, it appears that in some
jurisdictions the decrease in tax rate would have to apply to all property.
2. Neither the residential part nor the total tax is as
regressive as the conventional view of 10-15 years ago suggested it was.
This point needs elaboration in depth, and I return to it later.
3. Strictly speaking, there is no such thing as "the" property
tax. The tax is levied at widely different rates, and the differences per se
are relevant for incidence theory. Even within a jurisdiction, some property
is exempt, and in addition the rates differ greatly across jurisdictions.
The jurisdictions differ in the possibilities property owners have for
shifting the tax. This means the best analysis is one which is tailor-made
to the jurisdiction. It must allow for characteristics such as assessment
practices, possibilities for tax exporting (shifting to other regions) by the
industries whose capital is taxed, the elasticity of substitution between
capital and other factors in those industries, the ratios of market value to
annual rent in each housing market, the competitiveness of the rental housing
market and the kinds of individuals who own rental housing units, and many
more. J
4. The conclusion on incidence depends greatly on whether one
adopts annual income or a longer-run, "permanent" income as the income
measure. The tax is less progressive or more regressive, the shorter the
period for which income is measured. The New View contains as one crucial
element a conviction that a longer-run measure is more relevant.
107
-------
However, if one adopts so permanent a concept as lifetime income, the
question becomes complicated by the fact that many families move from
jurisdiction to jurisdiction several times in a lifetime, and some of them
move each time from a lower tax base base (per capita) community to a higher
tax base (per capita) community. Thus, even if the tax is progressive for
all the families living in a jurisdiction at any one time, it is possible
that over a whole lifetime the tax price of local services rises much less
slowly than income. Some who push this argument point out that many of the
poorest families are confined for their entire lifetimes to central cities
where tax bases are low, tax rates are high, and public services are poor.
However, even if multiple jurisdictions in a lifetime are important in
analyzing the property tax as an institution, I don't think they are very
important in analyzing the changes in tax rates in any one jurisdiction.
Along the same line as the multiple jurisdiction point, Hamilton has
suggested that the property tax ultimately is far more regressive than any
cross-section evidence on incomes and estimates of tax burdens can suggest
(6). That is because the property tax is part and parcel of a comprehensive
system of local finance in which restrictive zoning is used to limit high tax
base communities to high income individuals. He hypothesizes that the
restrictive zoning, which is employed by affluent communities precisely
because they must rely heavily on local property taxes to provide services,
raises the price of housing in poor communities. This adds significantly to
the impact of the property tax payments per se. This argument probably
should be ignored in our present context, which is the financing of a single
function which does not absorb much of the local budget. In other words,
changing the financing of SWM will not in itself have much impact on the
system Hamilton is describing.
With these summary conclusions as background, allow me to sketch the
basic theoretical arguments. Under the old view, the tax decreases the
returns to capital and land in the short run. In the long run, the supply of
land is inelastic, so the rate cannot be restored by any shifting of land out
of taxed uses, if the tax is general. Therefore, the land tax is borne by
landowners, which means it is quite progressive. But the supply of
reproducible capital has some elasticity in the long run, and the owners of
taxed capital will shift the tax at least partly on to consumers, and perhaps
other factors, by reducing the quantity supplied until the gross return rises
enough to restore the net return to equality with other investments. In the
end there is some fall in the.net return to capital, but the lion's share
falls on consumers of goods, and services, and disproportionately on housing,
a very capital-intensive good. As the elasticity of general consumption with
respect to current income is below one, and the elasticity of housing
expenditures well below one, the tax on reproducible capital is seen as
regressive and the residential property tax especially so.
The New View comprises two parts. First, the theoretical derivation
of the incidence of a general tax at a uniform rate on all capital and land;
second, the adjustment of the theoretical results to allow for the fact that
the tax is not actually levied on all capital, and not at a uniform rate on
the capital which is taxed.
108
-------
A key assumption In the New View is that the supply of reproducible
capital to the economy is inelastic, even in the long run, just as is the
supply of land. In that case, a general and uniform tax on capital must fall
on the owners of capital. The inelasticity of supply eliminates the
possibility of shifting the burden to consumers or to other factors of
production. As the earnings from capital are concentrated in the higher
income groups, this means the tax is quite progressive.
In the real world, the tax is not uniform. Some capital is taxed at
much higher rates than other capital, partly because of intra-jurisdietional
variation and partly because of inter-jurisdictional variation. The owners
of capital which is taxed at more than average rates can and will shift
some of the burden by moving their capital out of the high tax sectors and
into the low or zero tax sectors; that will raise the prices of goods and
services produced using the heavily taxed capital. The more uses of capital
are taxed, and the less the variation in tax rates, the less of the burden
can owners shift off themselves. Of course, the process will result in lower
prices for goods using less heavily taxed capital. To the extent there are
such shifts forward on to consumers, the incidence of those excise-type
effects will depend on how the consumption of various goods and services
varies with income. Some backward shifting is also possible, as the
restriction of capital in some industries and some localities may reduce the
demand for that labor which is relatively immobile and must accept reduced
wages to gain employment. However, these backward shifting effects are
probably quantitatively minor compared to the forward shifting or excise
effect.
Housing is very capital intensive, and the property tax rates on
capital in housing appear to be higher than average. That means the excise
effect will raise the price of housing to a significant degree. From a
practical standpoint, housing is probably the most important good the price
of which is raised or lowered by the excise effects of differential tax
rates. The fact that the price of housing does rise, therefore, is a central
fact under the New View, just as it was under the older view; however, the
New View stresses that the rise will be less than formerly thought, because
the fairly widespread taxation of capital and the presumed inelasticity of
supply reduces shifting. Furthermore, the same process of forward shifting
which increases the price of housing will reduce the prices of lightly taxed
goods, the consumption of which may be roughly proportional to permanent
Income. For consumers as a whole, this may offset the housing effect.
Break has concluded:
"Property tax burdens on consumers ... are probably less
important than commonly supposed. They are a function only
of interregional and intraurban tax differentials and not
of the total property tax rate; the excise tax effects
generated by those tax differentials contain labor and
landlord burdens which reduce the pressures making for
consumer burdens; and such national consumer burdens as do
109
-------
exist fall not on consumers In general but only on those
with relatively strong tastes for the output of the
industries that are more heavily taxed than others." (4,
p. 165.)
The New View not only has a different conclusion on the extent of
forward shifting, it also has a different concept of income. It favors a
"permanent" notion of income, rather than a single year's income. The
empirical studies of housing demand show that the elasticity with respect to
permanent income is considerably higher than the elasticity with respect to
annual income it is perhaps unity or even slightly higher. Thus, even if
housing consumers paid all the residential property tax, it might not be
regressive!
Somewhat strengthening this conclusion is other empirical evidence
that the housing market produces ratios of market value to annual rent which
rise with income. If the annual rental is proportional to permanent income,
the value then rises more than proportionately. If the property tax is
proportional to value, then the tax rises more than proportionately as well.
Thus, the New View has several strings to its bow. The widespread
taxation of capital limits the shifting by capital owners; the income
elasticity of housing expenditure with respect to permanent income is around
unity; and the value of property, which is the tax base, rises with income
even faster than housing rental expenditures. And even if there is a
regressive housing effect, there are some offsetting effects for the average
consumer from lower prices for other goods and services.
Against these there are some counterarguments, a few of which have
already been mentioned. The housing occupied by lower income families may be
systematically overassessed, which offsets the value-rent ratio phenomenon.
Many feel the permanent income is not a solid enough empirical concept on
which to rest a policy decision. Others Accept the permanent income concept,
but feel that if properly extended to lifetime income it points to
regressivity; this is the multiple-jurisdictions-in-a-lifetime phenomenon
mentioned earlier. Another consideration is the possibility that much rental
residential property is owned by low and middle income families, who rent out
parts of their own homes or who rent out at most one or two houses. If
rental property owners are lower income people, and their tenants are also
lower income people, it doesn't matter how much of the property tax on the
property is shifted forward and how much reduces the net returns to owners.
Finally, the deducibility of property tax for Federal income tax purposes
means that the net impact on higher income owner-occupiers is reduced. One's
feeling on how important this is depends on whether he thinks the
deductibility feature should be analyzed as a feature of the property tax or
as a feature of the Federal income tax. It is true that a change from the
property tax toward user charges will have a blunted favorable effect on high
income families because part of the reduction in property tax will be offset
by higher Federal income tax. In an empirical analysis, one must remember
the very recent trend toward raising the standard deduction and reducing the
number of Federal taxpayers who itemize deductions.
VI0
-------
By now, if you have never gotten into this literature before, your
head may be swimming. If you have gotten into it, your head may be swimming
anyway, for you may be nagged by still additional considerations which I have
not had time to mention. The safest conclusion, I think, is Netzer's:
"There seems no substitute for empirical work that recognizes local
differentials and perhaps, no case for the kind of nationwide estimates of
incidence by income class that I and others have made in the past." (13, p.
519.)
However, if one does insist on generalizing from national studies to
a particular jurisdiction, say a "typical" jurisdiction, one could conclude
with Netzer that the property tax on the whole has an elasticity with respect
to income of about 1.1 and the part on residential housing and consumer-owned
durables an elasticity of about .7. Netzer makes these ball-park estimates
on the assumption that income is measured as annual income, but with some
modification at the lower end of the scale to raise incomes there above the
ones reported as annual incomes. This modification is to allow for the fact
that many poor people are elderly homeowners who have a high asset-income
ratio, for the fact that some other poor people are poor only temporarily and
also have comfortable asset positions or at least other reasons for higher
income prospects, and for the underreporting by some poor people receiving
public assistance.
CONCLUSION
As noted earlier, the empirical studies of the income elasticity of
solid waste volume and collection cost produce estimates which are about
equal or a bit lower than the Netzer guess on residential property tax
burdens. It is unlikely that the income elasticity of solid waste is above
unity. This suggests that the vertical equity of user charges and the
residential property tax are not much different from each other. Indeed, it
hardly matters anyway, because the dollar amounts are very small to start
with. Therefore, there is no reason to fear any noticeable increase in
regressivity from substituting user charges for the property tax, especially
if it is the residential property tax which is being replaced.
Nor are there valid arguments, in my opinion, against user charges on
grounds of "merit goods" or on grounds that society should want to
redistribute income in a way dependent on poor persons' use of solid waste
management services. As I have said earlier, I do not think we can apply to
solid waste the notions which justify specific price subsidies on medical
care, housing, and education. In fact, I would be prepared to argue the
opposite, that persons should be encouraged, even at some cost in terms of
efficiency, to take care of their own waste. Society as a whole should want
to encourage habit of resource conservation, for one thing, and the price
system in solid waste may be useful in this respect.
Ill
-------
REFERENCES
1. Aaron, Henry, Who Pays the Property Tax. Brookings, Washington: 1975.
2. Bolton, Roger, "Household Income and Refuse Generation," unpublished.
3. Bolton, Roger, with P. M. Meier and J. Kuhner, An Assessment of Wet
Systems for Residential Refuse Collection: A Case Study for
Springfield, Massachusetts, for U.S. Environmental Protection Agency,
1974 (NTIS).
4. Break, George, "The Incidence and Economic Effects of Taxation," in Alan
Blinder and Robert Solow, £t al, The Economics of Public Finance.
Brookings, Washington: 1974, pp. 119-240.~~
5. Downing, Paul, "Intra-Municipal Variations in the Cost of Residential
Refuse Removal," and "The Distributional Impact of User Charges for
Refuse Collection," unpublished, mimeo.
6. Hamilton, Bruce, "Local Government, the Property Tax, and the Quality of
Life: Some Findings on Progressivity," in Lowden Wingo and Alan Evans,
eds., Public Economics and the Quality of Life, Johns Hopkins Press for
Resources for the Future, Baltimore: 1978, pp. 111-122.
7. Kemper, Peter and John Quigley, The Economics of Refuse Collection.
Ballinger, Cambridge, Massachusetts!1976.
8. McClure, Charles, "Revenue Sharing: Alternative to Rational Fiscal
Federalism?" Public Policy, Summer 1971, pp. 457-478.
9. Mieszkowski, Peter, "Tax Incidence Theory: The Effects of Taxes on the
Distribution of Income," Journal of Economic Literature. December 1969,
pp. 1103-1124. ~~
10. Mieszkowski, Peter, "The Property Tax: An Excise or a Profits Tax?"
Journal of Public Economics, April 1972, pp. 73-96.
11. Musgrave, The Theory of Public Finance. McGraw-Hill, New York: 1959.
12. Netzer, Dick, Economics of the Property Tax. Brookings, Washington:
1966*
13. Netzer, Dick, "The Incidence of the Property Tax Revisited," National
Tax Journal, December 1973, pp. 515-535.
112
-------
14. Sheaffer, John R. and George S. Tolley, Decision Making and Solid Waste
Disposal, Center for Urban Studies, University of Chicago, 1971.
15. Sheaffer, John R. and George Tolley, Socio-Economic Factors Affecting
Demand for Municipal Collection of Household Refuse (condensation), U.S.
Environmental Protection Agency, Cincinnati, Ohio: 1973.
16. Thurow, Lester, "Cash Versus In-Kind Transfers," American Economic
Review, May 1974, pp. 190-195.
17. Tiebout, Charles, "An Economic Theory of Fiscal Decentralization," in
Public Finances; Needs, Sources, and Utilization, Princeton University
Press for National Bureau of Economic Research, 1961, pp. 79-122.
18. Tobin, James, "On Limiting the Domain of Inequality," Journal of Law and
Economics, October 1970, pp. 263-78.
19. Tolley, George, V.S. Hastings, and G. Rudzitis, Economics of Municipal
Solid Waste Management; The Chicago Case, U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1978.
113
-------
PRICING MUNICIPAL REFUSE SERVICE: POTENTIAL EFFECTS ON MUNICIPAL BUDGETS
by
KENNETH L. WERTZ
I wish to examine how two approaches to pricing household's
generation of refuse could affect municipal budgets. The local pricing
approach contemplates a municipal charge system for residential service under
which a property owner would pay more as he presented more standard
containers of refuse for collection and disposal. The federal pricing
approach would create specific excises against certain products used in the
packaging of consumer goods and distribute the revenue to local governments
to assist their financing of refuse service.
In the opening sentence I emphasized "could" because it connotes
potential and uncertainty of result. The budgetary impact of a pricing
program depends not only on the responses of households but also on decisions
made by municipal officers. So long as the dollar amount is not trivial, the
introduction of a new component of revenue creates the leeway for managerial
discretion and may quicken the propensity to use it. I cannot predict the
specific responses (and their budgetary images) of municipal officers who
operate under heterogeneous conditions in- the nation's 18,517 municipalities,
but I do endeavor to describe what the opportunities might be. I utilize
many approximations to suggest the order of magnitude of certain effects, and
indicate the method of calculation in the event others care to employ
alternative estimates.
The first section looks at the revenue side of the budget, and the
second at the expenditure side. A final section concludes with a bit of
speculation.
REVENUE SIDE
The revenue yield of a pricing system depends on the charge
structure, the height of the charge, the base to which it applies, and the
reactions of consumers and producers. With a federal program there is
additionally the matter of how revenue is to be returned to municipalities.
For concreteness assume that a municipality's total revenue from a pricing
system, whether locally or federally administered, just equals what the
municipality was previously spending for residential refuse service. The
114
-------
basic question for municipal officers is whether these receipts will be
allowed to change the total amount of local governmental revenue from own
sources that had previously been planned for in preliminary budgets. The
polar possibilities are that charge receipts are treated as an addition to
own revenue or are used to fund reductions (or to defer planned increases) in
local tax rates.
If own tax reduction is chosen with the intent of holding constant
total local governmental revenue from all sources, then municipal officers
must plan on reducing own tax revenue by an amount that is less than
anticipated charge receipts. This is so because some intergovernmental
transfers, of which local governments are the net beneficiaries, are allotted
according to formulae that reward a local government for collecting tax
revenue but not charge revenue. A universal case in point is the federal
general revenue-sharing program. To a first approximation, the grant to a
unit of local government varies directly and proportionately with the
magnitude of its "adjusted" tax collections. Thus if a municipality were to
replace X% of its own tax collections with charge revenue, its federal
revenue-sharing grant would decline by X%, other things being equal and
assuming that the municipality's allotment is not determined by the
provisions governing maximal and minimal allotments. A rough estimate is
possible for the aggregate of municipalities. In 1975 all municipalities
combined raised $21.1 billions in own taxes and $3.6 billions in
miscellaneous own general revenue, of which $3.6 billions were spent for
education (municipal expenditures for education were actually $7.2 billions
in 1975, but only 49% of local governments' expenditure for education came
from own sources). Thus "adjusted" taxes the adjustment excludes revenue
raised for education expenditures were about $21.1 billions. If we take
municipal outlays for residential refuse service and financed from local
general revenue to have been $1.2 billions in 1975 (actually, $1.8 billions
was spent for "sanitation other than sewerage," but some of this was not for
residential service and some was not financed by own taxation), then its
replacement by an equal amount of charge revenue would have implied a
reduction in revenue-sharing grants to all municipalities of about 6%. The
dollar loss would have been $125 millions, since all municipalities received
$2.2 billions in federal revenue-sharing payments in that year. The
calculation indicates that on average a municipality could enact local tax
relief of $.94 and lose $.06 in intergovernmental transfers for each dollar
collected in charges, if total revenue from all sources is to be unchanged.
These figures will differ from place to place, depending primarily on the
budget share of public residential refuse service in a given municipality and
on higher order interactions in the workings of the revenue-sharing formula.
For municipalities in some states even less local tax relief could be-
afforded because of state-local intergovernmental flows. For example, in
1973 the Michigan state government distributed $54 millions to local
governments according to formulae that increased the share of any particular
1. Outlined at Advisory Commission on Intergovernmental Relations,
Federal-State-Local Finances, 1974, pp.79-81.
115
-------
local government as the latter's own tax collections increased. Similarly,
the North Carolina state government distributed $26 millions according to
local property tax levies. Such programs* were also to be found in
California, Ohio, and Wisconsin in 1973, and it is quite possible that their
adoption has since spread in imitation of the federal general revenue-sharing
program.
The discussion has so far regarded charge revenue as a source of
local tax reduction, but it applies as well if charge revenue is meant to
enlarge local revenue from all sources. When an extra dollar of revenue is
raised through charges rather than taxes, the local government foregoes the
bonus that higher level governments will pay on the latter.
EXPENDITURE SIDE
One obvious difference between a local and a federal pricing system
is the incidence of the costs of running the system. S. 1281 (95th Cong.,
1st sess., 1977) to my reading would erect no condition that should occasion
signifcant municipal expenditure solely for the purpose of making the
municipality eligible for the return of charge revenue from the federal
government. A local pricing system, however, is sure to cause additional
local expenditure for administration. It is necessary to have an account
with each property owner; to publicize rates and information about standard
containers; to bill, collect, and perhaps issue receipts; and to adjust
disagreements about the correct charge. Although municipalities have already
opened accounts for purposes of property taxation and although they could
economize by combining billing with tax or public utility billing, extra work
and extra expenditure must still be expected. Furthermore, it is necessary
to observe and record the number of containers emptied at each property.
This task involves additional expenditure too, for such data are not compiled
where general revenue financing is employed.
There is nothing incorrect about this fatiguing recital of
conventional wisdom. But it takes too narrow a view of "administrative
costs" by ignoring what they could buy. One return on the costs of running a
local pricing system is the regular collection of data which might be used
for the more efficient supply of service. The quantitative importance of
this potential is unknown, but the state of municipal recordkeeping that
researchers have encountered is at the very least suggestive that that
potential remains largely unrealized. Microdata on refuse generation would
make it possible to search thoroughly for cost reducing route designs and
factor assignments. Aggregated and kept over time the data could be used to
sharpen planning for capital expenditures, manpower adjustments, disposal
site acquisition, and adjustments to seasonal variations in refuse
generation. Efficiency savings from these sources, assuming that they are
realized through managerial action, are properly considered as offsets
against the direct cost of operating a local charge system. The federal
product charge system generates no comparable information, and no comparable
efficiency savings should be expected.
A second type of return on administative costs is an Intangible that
appears in no governmental budget, even though much public spending and
116
-------
revenue raising are undertaken in its name. The intangible is equity. The
content of the term changes from one application to the next, and unanimity
about its proper content in any one application is rare. These realities
understood, we may note that some of the enthusiasm for pricing proposals is
based on the view that a person ought to pay for service in proportion to the
costs which his refuse generation visits upon the public treasury. The
regular satisfaction of that norm under local tax financing, which is
essentially property tax financing, is just not in the cards: The
connections between determinants of refuse generation and property value and
the connection between assessed value and property value are too irregular.
How the burden of local taxation for refuse service might be distributed is
scrambled another turn by the deductability of local property taxes (but not
charges and then only for itemizers) from the federal (and many state)
individual income tax base(s). A local charge system, on the other hand, is
predicated on a clear and direct connection between payments made and costs
imposed, from one person to the next, provided that the logic of the system
is carried out. Administrative costs for running a local charge system are
then costs of producing horizontal equity. "Local" is emphasized because the
same results are unlikely to follow from a federal product charge system that
returns revenue to local governments on a capitation basis.2 No systematic
relief from current inequities can be expected if municipal officers utilize
their federal grants for improved refuse service (as contemplated by S.
1281), because neither the magnitude nor the distribution of local taxes
would be affected, and they are the putative sources of the inequities. Some
systematic relief would occur if municipal officers were instead to utilize
federal grants to fund reductions in local taxes. But in either case federal
pricing would create new variances between payments made and costs imposed,
because product charge collections are based on refuse generation while the
return of revenue is based on population, and considerable variance appears
to exist in per capita rates of generation from one place to the next.
In summary, the potential for efficiency savings from data
availability and an improvement in horizontal equity are direct offsets
against administrative costs of managing a local pricing system. It is not
possible to arrive at a monetary figure for net administrative costs, because
one of the products is an intangible and will be valued differently even by
persons in the same locality.
The collection and disposal of refuse is the second major area where
pricing is likely to have a noticeable effect on the expenditure side of
municipal budgets. Federal pricing would create a private marginal cost
against one's generating extra refuse. Local pricing would create a private
marginal cost against one's presenting extra refuse for public collection.
Local tax financing of service does neither. Hence pricing, relative to the
status quo, would reduce the amount of refuse which the municipality would
have to treat, and service with constant features (frequency, place of
2. This position is elaborated at K. L. Wertz, "Equity Impacts of Proposals
for Federal Product Charges," Journal of Environmental Economics and
Management, Dec. 1977.
117
-------
collection, etc.) could be supplied for a smaller municipal outlay. 3 Since
the charge-induced reduction in operating costs is to be realized year after
year, it could function as an endowment for programs whose budgetary impact
would last over many periods. The cost savings could be returned to citizens
through reduced local tax rates, distributed to municipal employees as higher
effective wages, applied to providing refuse service of a higher quality, or
applied to augment other municipal programs.
An order of magnitude estimate of the charge-induced reduction in
operating costs is possible through a succession of approximations. I have
elsewhere instanced the fact that in 1970 per capita collections of
residential refuse in San Francisco, where pricing is practiced, were 25%
less by weight than the per capita figure for all urban areas combined, where
pricing is very seldom practiced. This is a gross comparison; it does not
adjust, up or down, for 'other possible influences on refuse generation. Let
us stipulate a 20% reduction to follow from the introduction of pricing, the
downward adjustment made because the height of the charge in San Francisco
probably exceeds the average level that would be contemplated in other
municipalities. The second figure that is needed is the elasticity of total
cost of refuse service with respect to refuse tonnage. Barbara Stevens has
produced estimates for collection costs that range from 0.8 to 1.0, depending
on city size. Let us use an intermediate value, 0.9, and assume that it
applies to disposal costs as well. The conclusion, then, is that pricing
could result in an 18% decrease in annual operating costs.
The significance of this figure in the aggregate can be better seen
by carrying the approximation through a few more steps. Assuming that
municipalities as a group are now spending about $1.8 billions per annum for
residential service, an 18% decrease equals annual savings of $0.32 billions
in 1978 dollars. Assume finally that the current municipal borrowing rate is
6.67%; that savings in operating costs are realized in perpetuity; and that
both the nominal savings and nominal borrowing rate grow at the same rate as
inflation generally. Then the present value of the charge-induced reduction
in operating costs to all municipalities combined would be $4.8 billions in
1978 dollars. This is a large amount, but not an imposing one. It is, for
3. The metering method which is employed under local pricing could
contribute to savings in operating costs, independently of any
charge-induced effects on the output of refuse. For example, if the use
of standard disposable bags were required in lieu of conventional
containers so that quantity for billing purposes is determined by the
number of bags set out, then the costly return of empty containers from
street to dwelling would be eliminated. Such possibilities are not
considered in the estimates in the text.
4. K. L. Wertz, "Economic Factors Influencing Households' Production of
Refuse," Journal of Environmental Economics and Management." Apr. 1976.
5- B« ! Stevens, "Scale, Market Structure, and the Cost of Refuse
Collection," Review of Economics and Statistics. Aug. 1978.
118
-------
example, only double the amount that all municipalities are now receiving in
federal general revenue-sharing grants in one year.
The use of the foregoing analytical structure will probably lead to
underestimates of the charge-induced savings in operating costs experienced
by any one municipality in the event that a majority of municipalities adopts
local pricing or that federal pricing is installed. It would then become
reasonable to think of secondhand markets becoming more active and of
producers taking the existence of charges into account when making decisions
about packaging, two changes that would reduce the amount of refuse which a
municipality collects and disposes of.
CONCLUSION: A RACE OF ONE?
A municipal officer who might have earlier considered the adoption of
a local charge system could have framed the matter this way. "I get charge
revenue, a modest reduction in operating costs, the likelihood of cost
savings due to data generation, and much better satisfaction of a particular
distributional norm. In return I incur a certain increase in administrative
costs and the equally certain unpopularity of having proposed a new % tax'
(even if existing taxes were adjusted so that revenue collections from all
sources were held constant)." However attractive this package might once
have been, its attractiveness probably evaporated with the recent legislative
advances of the federal proposals for product charges. From a municipal
officer's perspective, or so I should think, the federal plan would have none
of the liabilities of a local pricing system. Under federal pricing
administrative costs exist but are essentially borne elsewhere.
responsibility for the enactment of a new revenue program can also be put
elsewhere. The risk of increased littering that goes with local pricing is
removed by federal pricing, and the risk of having promoted a bad ideaif in
fact pricing fails to induce much reduction in the output of refuseis
shifted to other parties. Yet the federal plan would provide revenue and
stimulate whatever cost-reducing decreases in waste generation there are to
be realized. Only the potential efficiencies that arise from data generation
and the increased equity in finance that are unique to local pricing need be
given up. As witnessed in other intergovernmental aid programs, the prospect
that the federal government may in the future bear the costs, broadly
conceived, of providing many of the benefits that could be had by local
action is probably enough to retard local action now. I suspect that the
contest of pricing propoals is now a race of one.
119
-------
SECTION 4
COMPARISON OF USER CHARGES AND PRODUCT CHARGES
"The Optimal Policy Mix for Solid Waste Pricing" (James F. Hudson).
"Fundamental Comparisons of Alternative Solid Waste Management Policies"
(Allen K. Miedema) .
120
-------
THE OPTIMAL POLICY MIX FOR SOLID WASTE PRICING
by
JAMES F. HUDSON
INTRODUCTION
It would be difficult to claim a connection between the concept of
"optimality" and the field of solid waste management, despite the title of
this paper. The handout from our recent work for the Resource Conservation
Committee, and other papers in this seminar, summarize the state of the
theoretical and practical knowledge about solid waste management and
potential pricing policies. According to these papers, there is little
likelihood that we will approach a mix of policies which could be called
"optimal", as defined by economists.
Still, it may be possible to locally optimize our policy choices,
given the set of options which appear politically feasible. It is this more
limited sort of optimization which this paper addresses. It describes steps
which can be taken towards optimization, and it defines reasonable goals in
this regard but stops short of determining some form of globally optimal
policy mix.
This paper will ignore issues such as the assumptions behind the
optimality of marginal cost pricing, the requirement for constant returns to
scale, and the declining marginal benefit curve for pollution. Further, the
paper will not repeat the descriptions of research contained in the handout,
but rather will focus on the conclusions arising from the research, and the
potential for policy improvements in the field of solid waste management.
Although the conclusions reached in this paper are somewhat speculative, they
are based on the author's experience in the field of solid waste pricing and
management.
APPROACHES
There are two basic economic approaches which have been suggested for
reducing the quantity of solid waste reachingfinal disposal; these are
product charges and local user charges. Both act through reducing the amount
of product-related wastes, and through increasing the amounts recycled. This
paper will argue that, under a reasonableat least to the authorset of
121
-------
assumptions, the interactions can be generally beneficial and conducive to
achieving efficiency and equity goals. It will make arguments related
particularly to the incentive structures of the various actors in the system,
and the directions of their probable responses.
The effects of pricing policies are described schematically in Figure
1, along with the impact points of the charges. While this schema represents
a simplification, it leads us to suggest a number of actors in the process:
extraction firms;
production firms;
fabrication firms;
wholesaling/retailing;
households and other consumers;
solid waste management firms and agencies;
secondary materials retailers/wholesalers;
state and federal governments;
equipment suppliers for waste management.
Product Charges
Product charges act relatively early in the cycle, as described in
the handout. They will reduce the level of output and revenues for
extraction firms. This is intended to correct the inefficient or excessive
output levels prevalent prior to the imposition of charges. There will be a
direct incentive to increase the production and fabrication of recycled
materials and/or to decrease the use of materials for packaging. Price
increases will be passed on to consumers, leading to a loss in consumer
surplus, which could be counterbalanced somewhat by reduced solid waste
management costs. In addition, the product charge revenues will be
distributed to local governments, to offset solid waste management costs, or
increase the spending for other social services, leading to (it is hoped) a
net benefit consumer surplus. The local solid waste mainagernent agency will
be able to lower cost, and may receive revenue transfers from the system.
Recycling firms at the wholesale and retail levels will see increased demand,
and probably higher prices; current users of the wastes, however, will also
have to pay higher prices for their goods*
In essence, the product charge provides revenues to local
governments, increases the demand for post-consumer wastes, and provides an
incentive for firms to modify their use of packaging, with bonuses for forma
which minimize waste-handling costs.
122
-------
Extraction
Production
Key:
Original material flows
~ Revenue flows
Fabrication
CNJ
OJ
Recycling
Use
Collection/
Disposal
Litter
Other
Self-Disposal
Figure 1. Impact points of pricing options.
-------
User Charges
User charges related to the quantity of waste may provide some
incentive for consumers or firms to change their consumption patterns and to
select products with relatively small contributions to the wasteload. The
basic incentive is to reduce the quantity of waste placed for
collection/disposal: there are a number of methods to achieve this. One
method is to increase product lifetimes: this is likely to result in reduced
consumption, with potentially significant effects on waste quantities. Other
techniques include at-home recycling (such as composting yard waste); product
recycling; littering (including disposal at workplace, etc.) and other forms
of disposal such as direct hauling to the disposal site, or outside
contracting. Inventory increases, for example of newspapers, are not going
to affect total waste quantities for the long-term, though they may lead to
change in the amount of collected waste.1 The handout suggests in Section
4.3 that it is relatively easy to achieve 30-50% reductions in wasteloads if
yard wastes can be used on-site, and if recycling programs are available for
beverage containers and newspaper. With user charges, recycling firms should
see an increased supply of waste. Collection/disposal costs may go down;
litter collection costs, on the other hand, might rise, depending on
household behavior. The direct incentive of the charge is to reduce the
quantity of wastes placed for collection/disposal.
MARKETS
There are two submarkets of interest in analyzing solid waste: the
traditional production/consumption submarket, and the disposal/recycling
submarket. We are primarily concerned with the operation of the latter, and
the interactions between the markets, which occur through:
transfer of revenues (e.g. product charges);
changes in consumption caused by disposal practices;
supply and demand for post-consumer wastes for re-
cycling.
The disposal submarket consists of the household engaged in
post-consumption waste disposal and recycling, the firms or agencies charged
with solid waste collection and disposal, litter management, and the
retailing of recycled materials.
ANALYSIS OF SYSTEM COMBINATIONS
To identify a desirable mix of policies, the effects of combinations
of product charges and incremental user charges should be considered. These
include effects on consumption, household income, levels of recycling (and
prices for recycled products), operations of the extraction/production/
fabrication firms, operations of the solid waste collection and disposal
agencies or firms, and requirements for other forms of disposal.
124
-------
Effects on Consumption
As mentioned above, the product charge is intended to have an effect
on consumption, shifting the purchases of cost-conscious households away from
heavily packaged products. The result should be a shift in consumption
towards efficiency of at least the materials affected by product charges.
Because of the averaging that takes place in creating a =an or bag of
solid waste, and the low cost per item, I do not expect any change in
consumption to result from imposition of incremental user charges. Product
lifetimes may be increased somewhat, but I would expect this effect to be
relatively small.
With respect to consumption, then, there may be some adverse
interaction of the product charge and the deposit system, but this will be
slight.
Effects on Household Income
With the product charge households will experience increases in
prices with an effective loss in income somewhat less than the product charge
payments (because of shifts in consumption). The distribution of this burden
across households depends upon their consumption patterns.
Product charge revenues will be used to defray the administrative
costs of the program, and to provide funds to state and local governments
which will either be used for reducing solid waste disposal charges, property
tax relief, or other social programs. If we assume that consumption of
product-related waste rises with income, i.e. the price increase burden is
progressive, and that the property tax and waste charges are relatively
regressive, while the social programs are progressive, then the net result
should be a progressive shift in the distribution of income. In addition, if
the administrative costs of the product charge are minor, then adjustments in
consumption would probably lead to a net welfare gain, resulting from the
improvement in allocative efficiency.
Levels of Recycling
The product charge will increase the demand for product-related
waste, because of the lack of charges for reused materials. The incremental
user charge can be expected to increase the supply of such wastes provided by
households, by giving them an incentive to dispose of their wastes by methods
other than collection. If the product charge is on the order of 1/2
cent/pound, then the price increase should initially be on the order of
$107 ton. This will have little effect on recycling of metals such as
aluminum, but should affect other materials, such as glass, steel cans, or
newspapers. The price support may also lead to use of these materials as
base-load sources, rather than marginal ones,, and smooth out the price
variations described in the literature.
On the other side of the system, recycling cans, bottles, and
newspapers will result in greatly reduced wasteloads, and reduced solid waste
125
-------
collection costs. The easy availability of mechanisms for recycling, such as
separate collection, will reduce the chance of diversion to other less
desirable modes of disposal, such as littering, and lead to benefits from the
incremental user charge scheme. Therefore, the product charge will aid in
the diversion of waste from collection to recycling. The imposition of
incremental user charges should reinforce this effect, particularly in areas
where recycling is available. Reductions in waste collection costs, and
increased availability of revenues for municipal recycling programs, may also
result.
If the incremental user charge is set at $.50/bag, or about $30/ton,
then the total incentive for recycling will be $40/ton for the household if
both systems are implemented. However, unless the recycling firms can charge
households for the materials collected, they will see a price increase of
$10/ton, and, probably, an overabundance of materials available for
recycling. The interaction of the two systems will mean that sufficient
supply will be available, with demand likely to remain the constraint. A
related effect is that, without the product charge, it will be difficult to
recycle all the materials removed from the collection/disposal stream. This
is a major problem with the incremental user charge system, implemented
without the product charge.
Effects on Production Firms
The incremental user charge is not expected to affect these firms,
except slightly through the purchase of disposable containers. Therefore,
there is no major interaction here.
Effects on Solid Waste Agencies and Firms
This is the area where most of our research has been directed. At
present, we believe that most systems operate fairly inefficiently, and that
the elasticity of cost with respect to waste generation would be:
0.2 for inefficient firms;
0.5 to 0.8 for efficient firms;
0.7 to 0.9 for firms unconstrained by the real world;
The latter are a special case, similar to the frictionless plane of.physics:
firms and agencies are not constrained by actual vehicle sizes, the slack
related to the need for fixed routes, or the variation in load from week to
week; instead, they can operate efficiently under any condition, and
therefore can make the best possible adjustments to reductions in the level
of waste to be collected. The "efficient" firms go as far in this direction
as is practicable, while the inefficient firms make no major adjustments in
the face of changes in wasteload, except reductions in overtime and equipment
wear.
For major reductions in wasteload, as might occur with introduction
of incremental user charges, an elasticity of 0.5 would appear to us to be
126
-------
reasonable. Changes such as this are usually simultaneous with other major
system revisions, hopefully in the direction of efficiency. Therefore, the
result should he an initial capturing of the potential savings in many cases.
If the wasteload reductions are large, for example 40%, then the cost savings
will be on the order of 20%. System revenues, though, would drop 40% if the
charge was set to cover the before-introduction average cost, and there would
be a net deficit.
In order to eliminate this deficit, an increase in charges of 33% is
required, given the elasticity of 0.5.
If the product charge revenues were distributed according to
population (or the number of households), then communities reducing their
waste loads through incremental user charges would benefit. Not only would
their waste management costs drop, but their product charge revenues would
remain constant, reducing the necessary local contribution.
Let us consider a simple example, with initial collection costs of
$30/ton, and 1 ton generated per household annually. The elasticity is 0.5,
and the product charge receipts amount to half the waste disposal costs at
$30 per ton. The community decides to use the product charge to defray waste
collection and disposal costs, and there are no administrative costs.
Therefore, in the initial case, the product charge provides
$15/household annually, and another $15 has to be found to cover the costs of
collection. Now- introduce an incremental user charge set at marginal cost,
which is $15/ton or $.25/bag. Assume that use of metered bags breaks even in
this case, and that wasteload drops 40%.
We now have the household generating 0.6 tons, and paying $9 annually
for collection and disposal. But the product charge adds $15/household,
resulting in a total revenue of $24/household, just equal to the collection
cost which is reduced by 20%. The product charge revenues cover the fixed
costs, as long as they are disbursed on the basis of population.
The product charge itself will have some effect on collection and
disposal costs through changing the character of the wastes, and reducing
total quantity. This may be significant in some cases, but appears to us to
be small in comparison to the effects expected through incremental user
charges.
One major issue for solid waste management systems, however, is the
ability to find sufficient funds to operate effectively. Many suggestions
have been made that these systems are under-funded; this claim, though, has
not been sufficiently documented. Both incremental user charges and product
charges have been proposed as potential remedies for these problems.
However, product charge revenues will-not necessarily be spent on
solid waste management only. Because of the number of firms engaged in solid
waste collection and disposal, and because of the fragmentation of the
industry, it would be difficult to distribute the revenues to all such
entities. Assuming that local governments attempt to maximize welfare
127
-------
through their budgeting procedures, there is no need to do so. Tax relief of
any form will have similar effects to tax relief for solid waste. Additional
services will, it is hoped, have a net social benefit. Thus, there is no
strong need to limit the use of the product charge revenues to waste
management, given the administrative problems Which may be expected.
User charges are likely to represent an attractive revenue source.
It is very possible that local governments will view charge levies as an easy
method for increasing their budgets for "under-funded" services, leading to
gold-plated operations, particularly since they are often monopolies. The
imposition of incremental user charges may accentuate the problem of
inefficiency by making the system appear equitable to observers and by
reducing waste volumes. Such charges, then, may have more chance of adverse
taxation effects than flat-rate charges, though this is speculative.
Other Modes of Disposal
The product charges are unlikely to have an effect on the use of
littering, garbage grinders, or other modes of disposal; therefore, there is
no interaction of concern.
RESULTS AND CONCLUSIONS
Figure 2 shows the relative effects of product charges and
incremental user charges based on the above discussion. In general, most of
the interactions are beneficial, and the two systems would appear to be able
to operate together effectively. The major interactions, in system
financing, consumption, and recycling, all appear to be positive, based on
certain design assumptions.
However, these design assumptions are fairly critical and should be
restated. We have assumed an elasticity of cost with waste generation of 0.5
throughout, which is effectively an assumption about system efficiency. We
have assumed that the product charge revenues are disbursed on the basis of
population, and that local governments use them to maximize social welfare,
either for tax relief or for the provision of services, which may include
solid waste management. We have assumed that waste disposal charges have no
effect on consumption behavior, or at least that such effects are minimal,
and we have assumed that recycling centers will not charge for the waste they
receive.
In addition, we have assumed that the incremental user charge will
use the marginal cost of collection as the marginal price, with a flat rate
for service or another revenue source used for the first block on the rate.
This would apparently result in an overly large amount of recycling. Given
assumptions such as these, however, it should be possible to design a
combination of the incremental user charge and the product charge which does
not double-count any costs, and does provide a more complete set of
incentives than either form individually.
128
-------
Effect on
Product charge
Incremental user charge
Combination
Consumption.
Income
Shift towards less waste
Small effect, probably
net increase
Minimal
Small effect, probably
net increase
Like product charge
Small effect, probably
net increase
Recycling
Increased demand for
base load of re-
cycled product waste
Incentive for households
to supply recycled
products
Increased recycling
with a probable
imbalance towards
exce a s supply
NO
Production firms
Shift towards reduced waste
in packaging and in-
creased use of recycled
materials
No effect
Like Product Charge
Waste management
firms
Slightly reduced costs
Reduced costs
Increased availability
of revenues
Reduced costs
Increased availability of
revenues
Other modes
of disposal
None
Incentive for increased
use
Incentive for Increased
use
Figure 2. Relative effects of user and product charges.
-------
FUNDAMENTAL COMPARISONS OF ALTERNATIVE SOLID
WASTE MANAGEMENT POLICIES
by
ALLEN K. MIEDEMA
ACKNOWLEDGEMENTS
The motivation for this paper was the absence of a fairly rigorous
case for the conceptual superiority of the product charge. The need arose
because our empirical analyses of the product charge and recycling subsidies
could not be easily and intuitively related to a paradigm structure that
could disentangle otherwise puzzling quantitative results. The crude
outlines of the analysis in this paper were suggested by Fred L. Smith, Jr.
in a briefing session on a related project during May 1977. Other thoughts
in this paper, especially the consideration of practical complications, are
indebted to my discussions with both Tayler H. Bingham and Curtis A.
Youngblood at RTI.
INTRODUCTION
The "solid waste problem" has received considerable attention for
many years and particularly during the last decade. It has been argued that,
like air and water pollution, solid wastes "need" to be "controlled." Yet
the reasoned case for the control of effluents to common property air and
water resources hardly extends in any direct manner to solid waste flows.
The consumption and production residuals that are disposed to the air
and water inflict costs on society in the form of health deterioration or, at
the very least, an unsightly, uncomfortable, or unenjoyable physical
environment. Given the institutions that have emerged in our society, the
residuals, aside from litter, that are disposed to the land as solid waste
inflict only a financial cost, a cost that is borne ultimately by all
individuals who directly or indirectly pay those costs. In the U.S. a large
number of state and local ordinances and the efforts of the U.S. Public
Health Service have virtually eliminated the egregious damages to human
health and physical environments caused by "unabated" solid wastes. (This is
certainly not true in less developed countries where fetid open dumps and
sewers constantly threaten the lives and well-being of their citizens.) The
"abatement" of solid waste has, by and large, taken the form either of
130
-------
government provision of the resources necessary to collect and dispose of all
wastes in a sanitary manner or of the imposition of ordinances that require
the self-provision of those services by private agents. Where then is the
externality?
Superficially it might appear that the current methods of handling
solid waste in the U.S. could be cited as an exemplary method for curing the
air and water pollution externalities: simply have the government pay for
all the abatement costs or have the government pass ordinances that require
those expenditures by private agents. Indeed that has apparently been
considered a sterling example. With rare exception the chosen "cures" for
air and water pollution have been both regulations and government subsidies.
The latter include a full complement of government assisted abatement schemes
like pollution equipment loan guarantees, extension of preferential municipal
bond rates to abating industries, rapid abatement cost write-offs, federal
development of abatement technology, government supported technical
assistance for abatement methods, etc. The main point is that, regardless of
the institutional mechanisms used, the achievement of the "same level" of air
and water pollution abatement as is already realized for solid waste might
well be viewed by many individuals as considerable, if not fully adequate,
progress in the battle against all three externalities.
Thus if the analyst assumes that current solid waste "control" is
provided efficiently, it is not surprising that he would defend the
materials-, packaging-, and disposable goods-intensive society. He can
argue, e.g., Johnson (I960), Becker (1965), and Arthur D. Little, Inc.
(1966), that materialism in the form of relatively high consumption of waste
intensive goods in both industrial and household production functions
enhances social welfare because it conserves other apparently scarcer
resources like labor time and energy.
What solid waste imposed externality do we cite in rebutting this
argument? The critical first step in the debate is to recognize that "the
externality" in solid waste flows is not unlike "the externality" that would
still prevail if all air pollution were "controlled" through the use of
regulations. Again aside from littering, an externality from solid waste no
longer exists in the sense that individual A is directly physically harmed or
interfered with by individual B's production or consumption activity. Rather
most of the externality in current solid waste management practices is
attributable to the maze of market distortions imposed by inefficient,
regulation-like pricing, institutional, and legal structures that combine to
produce solid waste "control."
A number of those inefficient structures have been enumerated in
other sources; e.g., Anderson (1977), Page (1976), and Fiekowsky (1976).
Each is mentioned here again for completeness:
Virgin Materials-Biased Tax-Policies
Current tax laws provide for virgin material depletion allowances,
capital gains tax advantages on standing timber, and favorable treatment of
expenses for exploration and development. All these policies allegedly
131
-------
reduce the relative price of virgin materials and hence cause both the
underutilization of recyclables and the underpricing of waste intensive
goods.
Virgin Materials-Biased Regulations
A number of administrative policies also distort relative prices of
virgin and recycled materials. First, many regulated freight rates are
systematically lower for virgin than for recyclable materials. Second, an
archaic federal law grants free mineral rights to those who make discoveries
on open federal land. Third, under the pretext of quality-assurance,
governments enforce virgin materials-biased labelling and procurement
requirements.
Free Rider Problem
The predominant method of payment for post-consumer solid waste
collection and disposal costs is flat assessments; i.e., the
"all-you-can-eat-for-$X-buffet" pricing scheme. The most common method of
administering flat assessments is through the application of property taxes,
local income taxes, or, given their fungibility, federal revenue sharing
funds in payment for municipally provided services. The second most
prevalent method is the direct administration of flat assessments by waste
collection and disposal contractors. It is argued (Wertz, 1976, p. 269) that
this financing method fails to make individuals aware of the waste collection
and disposal component of the marginal social costs of the goods they
purchase -- like eating another slice of apple pie at the
"all-you-can-eat-buffet," discarding another unit of waste costs the
individual nothing on the margin. Thus, only the fairly weak, negative
income effect from increased flat assessmnt rates, will reduce the purchase
of waste intensive goods. Society forfeits the potentially more powerful
stimulus for waste reduction in the form of the substitution effect that
might be achieved by full marginal cost goods prices that include waste
handling costs. This is the classic free rider problem.
Indirect Subsidization of Virgin Materials
Another less commonly mentioned distortion (Summers, 1973, Chap. 3;
Smith, 1974, p. 3; Train, 1976, p.l) is attributed to the market's failure to
incorporate the eventual collection and disposal costs of some virgin
materials in their prices. By comparison, the use of recyclables avoids
disposal costs while collection costs are included in their market prices.
Therefore, even apart from the distortions in the relative price of virgin
and recyclable materials mentioned in 1 and 2 above, this indirect subsidy
alone is allegedly accountable for underpriced virgin materials; supposedly
this simultaneously induces both excessive use of underpriced materials, and
hence excessive generation of wastes, and uneconomically inhibits the use of
recyclables.
Besides these four core arguments for the existence of "an
externality" in materials markets two other arguments have been propounded.
The first is that solid waste flows do entail a genuine externality in the
132
-------
form of littering through which individual A's production and consumption
activity adversely affects individual B.^ The second is that, compared to
recyling, virgin material extraction and processing causes a significant
amount of other externalities such as air and water pollution and spoilage of
scenic natural environments. However, as Anderson (1977, p. 355) points out,
the potency of this argument is diminished by the partial removal of these
externalities through the implementation of the National Environmental Policy
Act, the Federal Water Pollution Control Act, and the Clean Air Act.
A number of federal policies have been advocated or are under
consideration to correct these market distortions; e.g., see especially Sec.
8002 (j) of U.S. Congress (1976). Three major generic alternatives have been
suggested: (1) product disposal charges, (2) recycling subsidies, and (3)
user fees. Product charges would provide for a tax on virgin materials that
are eventually to be disposed as post consumer solid waste; the tax would
equal the marginal social cost of solid waste handling. Recycling subsidies
would provide for direct payments to users of recycled materials on a unit
basis. User fees would be direct, volume sensitive charges for waste
handling services in proportion to the extent that households and businesses
use them.
To date little analytical rigor has been applied to define clearly
the comparative strengths and weaknesses of each of these policies.
Since it may cost as much as 19 times more to collect a littered
container than to collect that container from a curbside (Goddard, 1975,
p. 39), it is sometimes argued that a corrective "litter tax" is
warranted. Although this argument initially seems appropriate and
although a significant amount of littering still does occur, the
magnitude of the "litter tax" that appears to follow logically from this
argument is infinitesimal. The logic would be to set the litter tax
equal to the cost of collecting a unit of litter times the proportion of
all units that are actually 1 ittered. Under such a scheme the proceeds
from the litter tax would just cover litter collection costs. For
example, an official of the National Soft Drink Association, obviously a
vested interest group in view of recent legislative activities on behalf
of mandatory deposits, recently proposed a litter tax of .1 cent per
beverage container (see "NSDA's Reed Proposes Litter Law; Bottlers Would
be Taxed for Program Litter Control Funding," Beverage Industry. 1977).
The flaw in this logic derives from waste disposal patterns that result
from flat assessment methods for funding solid waste handling services
(mentioned above in the discussion of the free rider problem). Since
the marginal cost of disposal under this payment method is zero, it
undoubtedly accounts for the sanitary processing of what would be a
significant volume of litter if point-of-dtsposal charges were levied.
Therefore the collective decisionmaking that leads to "free" waste
handling services also undoubtedly accounts for litter volumes far below
those that would be observed if those services were prices at their full
marginal social costs.
133
-------
Generally, comparisons of alternative solid waste management policies; e.g.,
Conn (1975), consist of fairly casual inferences about the general
directional effect each policy would have on the waste stream. Evidently
previous work fails to distill the essential elements of "the externality"
from solid waste flows into a simple model that both precisely defines the
market distortions that characterize materials markets and provides a
rigorous theoretical basis for inferences about the conceptual superiority of
one solid waste management policy over another.
The purpose of this paper is to examine the real income and market
effects of each of the three major alternative solid waste management
policies and of the status quo and to determine whether these policies can be
ranked in terms of their ability to increase social welfare.
The analysis begins with the development of a paradigm model of a
simple economy in the next section. The model assumes the existence of only
the free rider problem and the indirect subsidization of virgin materials;
i.e., not only all three alleged sources of downward biases in relative
virgin materials prices, but also the littering problem are assumed to be
nonexistent. Also, zero transactions costs are assumed. In a later section
this paradigm is used to analyze each of the four policy settings. The
policy simulations presented there are used to draw tentative conclusions
about the relative real income, net waste, waste generation, resource
recovery and recycling rate effects of the alternative policies.
THE MODEL
Since the objective of this analysis is to make statements about the
welfare characteristics of alternative solid waste management policies, it is
necessary to define a very rudimentary general equilibrium model of a
paradigm economy. The model development parallels that of neoclassical price
theory, with appropriate departures to account for the materials flows that
cause the solid waste problem.
The macroeconomic setting is assumed to be identical to that of the
standard (timeless) general equilibrium model, except that a constitution and
government are assumed. The government represents all individuals through
its constitutional authority to protect resource ownership and to expropriate
(tax) and expend those resources. As usual the economy is assumed to be
autarkic and both exchange and (in this model) government functions are
undertaken costlessly.
All production activities occur within profit maximizing competitive
firms that are fully owned by individuals. The firms themselves are not
endowed with resources and do not consume. Rather all resources are owned
and provided by individuals who are paid the value of the output (including
rents) attributable to the resources they provide. Thus the macro accounting
identity is met: the total value of output (GNP) is identical to total
national income.
A major departure of this model from the standard neoclassical model
is that all resources other than a single natural resource are subsumed into
134
-------
a single factor of production. This resource endowment is a composite of
human and physical capital from which a fixed total, k units, of services
are available. The resource endowment is owned in aliquot shares by the
individuals and is completely (and costlessly) mobile among all production
activities in the economy. The society is also endowed with a "mine" from
which the natural resource, "clay," is extracted. Once extracted, this
material flows through the economy and can either be recycled or disposed to
a sanitary landfill.
On the demand side all individuals have identical but independent
preference functions. Therefore, envy and benevolence are absent and a
single indifference curve can be used to represent society's real income.
Also all resource expropriation (taxation) by the government, whether by
excise taxation or direct property (income) taxation, is borne in aliquot
shares by the individuals in the economy. Similarly, the benefits of
government expenditures are enjoyed equally because individuals' preference
functions and resource endowments are identical.
With this general setting it is now possible to define more fully the
specific characteristics of production and consumption activities within the
paradigm economy. Only two final goods are produced: a "beverage" and a
residual (numeraire) good. The beverage is packaged in a "clay" container
which, when disposed, constitutes the only solid waste generated within the
economy. A single unit of "contents," f, is assumed to be required along
with a single container in producing a unit of beverage. This fixed
coefficients beverage production activity simply represents an operation that
(costlessly) blends the beverage contents and containers; i.e.,
q = min(j.f) (1)
where q » beverage output,
j » containers, and
f - contents.
Because of the specialized form of(l), q - j - f.
The two intermediate goods used in (1), containers and contents, are
produced by applying the capital output. The production process for contents
is assumed to be a simple constant-returns process,
f - Pk (2)
where P parameter of the beverage contents production process.
A "used" container is like modern non-refillable beverage
containers unsuitable for reuse; it must either be disposed or recycled
135
-------
into a new container production process. The latter is described by a simple
Cobb-Douglas production function,
r = «0kal (3)
where r = number of containers produced from recycled containers,
and
«0, «. « parameters of the recycling and container reconstitution
process.
The activities represented by the production process of (3) include the
collection, transfer, and reconstitution of "used" containers into new
containers.
Alternatively a new container can be produced from virgin "clay"
extracted from the "mine." The entire process of mining, materials
processing, and container manufacturing from virgin material is also
described as a simple Cobb-Douglas production process.
v = |9lA (4)
where v = number of containers produced from virgin material, and
/? /? parameters of the mining and container manufacturing
process.
The total number of new containers (and of beverage units produced)
is identically the sum of those produced from recycled containers and those
produced from virgin materials; i.e.,
q = j = v + r. (5)
The quality and appearance of containers produced from recycled material is
assumed identical to those produced from virgin materials.
Both the basic production processes of (3) and (4) and the (costless)
blending operations of (1) and (5) are assumed to be completely free of
materials losses. Also it is assumed that none of the used containers are
lost. Thus the total flow, h, of potential solid waste is exactly equal to
the number, j, of containers used or, equivalently, the number of units, q,
of beverages consumed.
The numeraire good, which represents all other goods besides
beverages, is also assumed to be produced by a simple constant returns
production operation,
136
-------
X - Trk (6)
where x - numeraire good, and
if = parameter of the numeraire good production process.
The utility function shared by all individuals is a simple
exponential function
A. 1-A.
u = q x (7)
where u - utils enjoyed by each individual, and
A - parameter of the utility function.
It is also assumed that if even a single container is littered; i.e., neither
recycled nor disposed, all individuals' utility becomes negative.2
Therefore, all individuals will unanimously vote to have the government
collect and dispose all used containers if the government is as efficient as
individuals in completing that activity. Finally individuals' utility is
assumed to be completely unaffected by whether a used container is disposed
as solid waste or recycled; i.e., utility is independent of the recycling
rate.
If prices in the paradigm economy are denominated in units of capital
services, and if disposable incomes are regarded by individuals as completely
exogenous, then individual utility maximization requires that for the society
as a whole the following expression is maximized
A 1-A
0 - q x + A(E-p*q-wx) (8)
where E - society's aggregate disposable income,
p* - price of beverages including excise taxes, if any, and
A - Lagrangian operator.
Since E is actually the sum of individuals' disposable incomes and since the
latter are assumed to be perceived as exogenously given, the partial
differentials of E with respect to q and x are zero irrespective of the solid
2. If L is the number of units littered, a more precise definition of the
utility function of (7) might be u - (l-^L)qAcl"A where l> 1.
137
-------
waste management policy that is applied. This is assumed even though E may
be substantially affected at the macro level by choices of q and x, since
individuals always require their government to tax them for the cost.of
collecting and disposing the used containers that are not recycled.
Under these assumptions the solutions to the first order conditions
yield the aggregate demand functions for q and x in terms of aggregate
disposable income, the price of beverages, and the parameter of the utility
function. Specificaly, the aggregate demand functions are
q - AE/p* and (9)
x = (l-A)E/7r
The production process for the collection and disposal activity is
also assumed to be a constant returns process and is assumed to be identical
whether the service is provided by the government or by individuals,
d - /xk (10)
where d = number of used containers collected and disposed to a
sanitary landfill, and
M - parameter of the collection and disposal process.
Figure 1 summarizes resource and product flows in the paradigm
economy. Only beverages and the numeraire good are consumed directly as
final products. All other products in the economy are intermediate goods
used in or associated with the waste intensive good. Three of the five basic
industries in the economy, i.e., those industries that directly use the
capital input, are intermediate goods production processes for f, r, and v;
the other two are the numeraire good production activity, x, and waste
collection and disposal, d. Economic forces together with solid waste
management policies determine the extent to which wastes are either recycled
or disposed to a sanitary landfill, the extent to which virgin
materials-based and recycled materials-based containers are substituted in
container production, and the extent to which beverages and the numeraire
good are substituted in consumption.
ANALYSIS
The paradigm of the previous section can now be used to examine the
materials and resource flows as well as the price and real income effects
associated with four alternative states of the world implied by four policy
settings. The four settings to be analyzed are: (1) non-intervention of the
government in either final product or input markets; i.e., the status quo,
(2) government payment of a subsidy for recycling, (3) government imposition
of a collection and disposal or user fee, or equivalently in this model, of
an excise tax on beverages, and (4) government imposition of a product
138
-------
virgin
material-
based
container
'composite1
capital
input, IT
beverage
container
beverage
contents
production
numeraire
good
beverage
production
on-
sumption/
ide prod-
uct gen-
eration
container
recycling
and
reconstitution
used container
collection
and
disposal
Figure 1. Paradigm economy resource and product flows.
-------
disposal charge on the virgin material content of eventual side products.
These four policies will be referred to as the status quo, user fee,
recycling subsidy and product charge policy options.
The analysis of each policy develops the equations whose solutions
characterize market equilibria and real income levels in the paradigm economy
under alternative policy choices. Specifically, a total of six equations and
six endogenous variables define market equilibria. These variables are: (1)
capital services used to produce virgin materials-based containers, k ; (2)
capital services used to produce recycled materials-based containers, k ; (3)
aggregate consumption of beverages, q; (A) the f.o.b. price of beverages, p;
(5) rents from container production, R; and (6) aggregate disposable income
level, E. From the six endogenous variables others such as the level of
numeraire goods consumption, x, and the real income level, u, can be computed
using (9) and (7), respectively. As would be expected, some of the six
equations differ depending upon the particular policy choice that is assumed.
Therefore, the next section first defines the particular general equilibrium
system that obtains under each of the four policy alternatives. Then, since
all systems are fully nonlinear, we present some policy simulations for
assumed values of the various parameters in each general equilibrium system.
The results of these simulations are then used to draw conclusions about the
comparative desirability of the four alternative solid waste management
policies. Table 1 defines the parameters and unknowns that appear in the
alternative six equation systems.
TABLE 1. GENERAL EQUILIBRIUM SYSTEM VARIABLES
Classification Variables
Function parameters a, ft, y
System parameters w, k
Policy parameters t, a
System unknowns k , k , q, p, R, E
140
-------
Policy-Specific General Equilibrium Equations Systems
Each of the four policies considered here implies a different method
of government Intervention to manage solid wastes. Each method in turn may
imply a unique profit function for beverage producers. Therefore the profit
equation is the first specified below in the general equilibrium systems
associated with each policy. Then, using the profit maximization assumption,
the second equation expresses the optimum quantity of capital to be used in
recycled materials-based container production (determined from the first
order conditions) as a function of that to be used in virgin materials-based
container production. The third equation, also derived from the first order
conditions tor profit maximization, expresses the price of beverages as a
function of the optimum quantity of capital services in recycled
materials-based container production. The fourth equation defines disposable
national income; i.e., national income net of government expenditures for
waste handling services. The fifth equation simply specializes the aggregate
demand function of (9) for each policy option. Finally the identity of (5)
completely identifies the system for each policy.
To simplify the algebra the production functions of (2), (3), (4),
and (6) are further specialized by assuming that P-«o=/3o!a7r=l and by
suppressing the subscript on «j and /Jj. While they are notationally
convenient, these assumptions do not substantively affect the conclusions
that derive from the policy simulations of the next section.
Status Quo
The status quo policy is intended to reflect the basics of current
waste handling practices. Specifically, government (or a group of
competitive private firms) is assumed to provide waste collection and
disposal services under a flat assessment payment scheme. The property tax,
for example, is one method by which this is accomplished in practice. A
similar pricing technique used by private waste collection and disposal firms
is to charge a flat rate per period for waste removal services regardless of
the volume generated.
Under the status quo beverage industry profits, R, will simply be the
difference between beverage sales and total input costs. Since one unit of
contents costing one unit of capital services (P - 1) is required per unit of
beverage output, the cost of contents will be q. The cost of containers will
just be the sum of all units of capital services used in container
production. Thus industry profits are
R
where R - beverage industry profits.
141
-------
When the specialized forms of the container production functions, (3) and
(4), are substituted into (5) and when the latter replaces q in the above
expression, beverage industry profits can be re-expressed as
(ID
When (11) is differentiated with respect to k and k and the
resulting expressions are set equal to zero, these two first order conditions
can be solved to express the profit maximizing quantity of k as a function
of the production function parameters and k ,; i.e.,
The first order condition in which the expression of (11) is
differentiated with respect to kr can also be solved to express the f .o.b.
price of beverages as a function of k ,
p = (l~a/«) + 1. (13)
Disposable national income under the_status quo is the value of its
fixed endowment of capital services, k, plus profits from beverage
production, R, less income taxes required to pay for waste handling; i.e.,
less the product of the unit cost of waste handling, w, and the flow of
virgin materials used (and discarded) or
E = k + R - wk ^ (14)
v
Finally since no excise taxes are imposed under the status quo, the
f .o.b. price is the retail price, p = p*, and the aggregate demand function
of (9) becomes
q = XE/p. (15)
Table 2 summarizes the six equations of the general equilibrium
system that is defined by a status quo solid waste management policy. This
system can be solved tor k , k , q, p, R, and E, although not in closed form.
142
-------
TABLE 2. GENERAL EQUILIBRIUM EQUATIONS SYSTEMS
Policy choice
Status quo
Recycling subsidy
User fee
Product charge
Equation System
(H),
(16),
(11),
(22),
(12),
(17),
(12),
(23),
(13),
(18),
(13),
(13),
(14),
(19),
(20),
(24),
(15),
(15),
(21),
(15),
(5)
(5)
(5)
(5)
Recycling Subsidies
Among the .proposed legislative remedies for the solid waste problem a
recycling subsidy policy is one of the most frequently mentioned. Anderson
(1977) reviews in some detail two major federal bills, H.R. 148 and H.R.
10612, that would grant alternative torms of recycling subsidies. Those
documents and others referenced by the National Commission on Supplies and
Shortages (1976, pp. 155-72) provide details on the nuances of specific
subsidy plans.
For this analysis the simplest possible form of a recycling subsidy
is assumed. Specifically it is assumed that a flat subsidy rate, a, is paid
directly to beverage producers for each recycled container they use.
Therefore the beverage industry profit function under a recycling subsidy
differs from that under the status quo policy in that profits are augmented
by subsidy payments; i.e,
R = (p-lHkr* + k^f) - kr - kv + a kra. (16)
Differentiating (16) and solving the first order conditions yields an
alternative expression for the optimum usage of capital services in recycled
materials-based container production,
I-/? !/(«-!)
k - [/3/(a(k +]8a))] . (17)
r v
143
-------
Also, one of the first order conditions for a maximum of (16) can be
solved to develop a price equation analogous to (13),
p - (klr~a/«) - a + 1. (18)
Disposable national Income under the recycling policy differs from
that under the status quo in that incomes must be further taxed to pay the
recycling subsidies. Therefore, disposable income will be
E - k + R - wk /8- ak ". (19)
Of course the absence of an excise tax on beverages implies that the
aggregate demand function under a recycling subsidy is the same as under the
status quo; i.e., equation (15). The six equation general equilibrium system
under the recycling subsidy is also summarized in Table 3.
User Fee
The user fee policy in the paradigm economy represents either of two
situations. First, the policy can represent the imposition of a
volume-sensitive waste tax at the point of disposal. For example, a tax
equal to the cost of used container collection and disposal could be levied
on each container that is thrown away. Alternatively, given the assumption
that transactions costs are zero, the policy is Identical, in the paradigm
economy, to a litter tax, since government can costlessly impose a charge
either at the point of disposal (user fee) or point of purchase (litter tax).
Under a user fee policy the profit function and the associated
solutions for k and p from the first order conditions are identical to those
under the status quo equations (11), (12), and (13) because the policy
does not directly affect beverage producers.
However, when a user fee is applied at the rate of t per unit of
beverages consumed (and per unit of container discarded by consumers),
disposable national income is increased over that under the status quo by the
amount of the user fee proceeds. Therefore disposable national income will
be
E-k + R+tq- wkv^ (20)
Of course, the specification of the aggregate beverage demand
function will also be affected since the effective price that consumers face
will be the sum of the f.o.b. price of beverages and the user fee; i.e., p* -
p + t so
144
-------
q - XE/Cp+t). (21)
The general equilibrium equation system under the user fee' policy will,
therefore, consist of equations (11), (12), (13), (20), (21), and (5).
Product Charge
A product charge policy would, as usually conceived (see Smith, 1974;
Page, 1976; and U.S. Senate, 1976), impose a unit tax on all virgin materials
that are eventually disposed. In the paradigm economy this implies that
every virgin materials-based container will be taxed at, say, the rate t.
Therefore, beverage producer profits will be reduced by the amount of these
taxes so the modified profit function is
r - Ky - t kv. (22)
The corresponding solution of the first order conditions for k is
(a"l) . (23)
The solution for p is identical to that under the status quo and user
fee policies; i.e., to (13).
Aggregate disposable income will be increased over that under the
status quo by the amount of the product charge proceeds so
E - k + R - (w-t)ky . (24)
Finally the specification of the aggregate demand function for
beverages will be the same under the product charge policy as under the
status quo and recycling subsidies. The resulting six equation system is
summarized in Table 2.
Solid Waste Management Policy Simulations
Since closed-form solutions of the four alternative general
equilibrium systems are not attainable, fixed values were chosen for the
parameters identified in Table 2 and the resulting nonlinear systems were
then solved numerically using the simple half interval method (Southworth and
Deleeuw, 1966). For all policy simulations the following parameters are held
constant at the indicated values: A - .2, w - 10,. k- 10,400. Then in three
different sets of simulations the production function parameters a and /? were
varied to analyze the effects of differences In the degree of diminishing
returns (diseconomies of scale) in the virgin materials-based as opposed to
the recycled materials-based container manufacturing industries. The assumed
145
-------
alternative parameter values are given in Table 3 and, as indicated,
correspond to three possible situations: (I) both the virgin and recycled
materials-based container manufacturing industries have equal diseconomies of
scale, (II) the recycled materials-based industry has greater diseconomies,
and (III) the virgin materials-based industry has greater diseconomies.
Diseconomies of scale obviously imply that increasing amounts of inputs must
be used to produce additional containers. In the recycling materials-based
industry this could be caused by diminishing quality of recyclable materials
or by increasing unit collection costs as the recycling rate increases. In
the virgin materials-based industry it could be caused by diminishing ore
quality and capital fixities in mining industries.
TABLE 3. CONTAINER PRODUCTION FUNCTION PARAMETERS FOR
THREE ALTERNATIVE SETS OF POLICY SIMULATIONS
Policy
simulation
I
II
III
Production function
parameter values
ft
.5 .5
.4 .6
.6 .4
Containe r manufacturing
industry with greater
diseconomies of scale
Neither
Recycled materials-based
Virgin materials-based
For each of the three sets of policy simulations the single solution
for the status quo was obtained. Since the status quo is always a special
case of all three alternative policies i.e., all four equation systems of
Table 3 are identical when t = a = 0 that solution is also a solution for
the zero value of the policy parameter associated with each. Then the policy
parameter was incremented by a value of 2 up to a value of 20 (twice the
marginal cost of collection and disposal). The resulting solution values of
interest are displayed in the three panels (labeled I, II, and III to
correspond with policy simulations I, II, and III) of Figures 1-4. The
figures display, respectively, the real income, net waste, waste generation
and resource recovery, and recycling rate effects of the alternative
policies. Each of those effects is discussed in turn in the following four
sections which compare the four alternative policies.
146
-------
Real Income Effects
Figure 2 displays the most important results because it shows the
fate of individuals' real income under alternative solid waste management
policies. Several critical conclusions are suggested by these results:
The product charge maximizes social welfare when the charge rate is set equal
to the marginal cost of waste collection and disposal. No other policy at
any setting of their policy parameters can generate as high a level of real
income.
Unless the product charge rate is well in excess of the marginal cost of
waste handling, w, society will be better off with a product charge set at
any level than under the status quo.
At any identical charge, subsidy, or fee rate the product charge is superior
to both the user fee and recycling subsidy policies and the user fee is in
turn superior to the recycling subsidy.
When the diseconomies of scale are substantially greater in virgin
materials-based container manufacturing than in recycled materials-based
manufacturing none of the innovative solid waste management policies are
markedly superior to the status quo.
Only the product charge is always superior to the status quo when the policy
rate is set equal to the marginal cost of waste collection and disposal.
All three innovative policies are superior to the status quo for values of
the appropriate policy rate that are close to zero.
Net Waste Effects
Figure 3 shows the associated net waste that is generated. All three
alternative policies lead to a lower net flow of waste that needs to be
collected and disposed by the government. At identical charge, subsidy, and
fee rates the product charge causes the lowest net waste flows. If
diseconomies of scale are greater in the virgin material-based container
manufacturing industry user fees are associated with the highest net waste
flows. Otherwise recycling subsidies are.
Waste Generation and Resource Recovery Effects
The reasons for the preceding results are displayed in Figure 4 which
shows both the gross waste generation and recycling levels. The upper set of
curves in each panel represents beverage consumption (waste generation)
levels and the lower set represents recycling levels. The vertical distance
between the two curves for each policy are the net waste effects shown in
Figure 3. The superiority of the product charge is also suggested by the
fact that it causes both a positive waste reduction effect (the upper dotted
line has a negative slope) and a positive resource recovery effect (the lower
dotted line has a positive slope). These reinforcing effects contrast with
the recycling subsidy which causes a perverse (negative) waste reduction
effect with a positive resource recovery effect and with the user fee which
causes a positive waste reduction effect with a comparatively perverse
resource recovery effect.
147
-------
(I)
(II)
(III)
8 12 16
Policy Rate
©
Product Charge
User Fee
> Recycling Subsidy
Status Quo
4 8 12 16
Policy Rate
o>
at
3.42
3.38
4 8 12 16
Policy Rate
Figure 2. Real income effects.
148
-------
(I)
(II)
(III)
4 8 12 16
Policy Race
©
Product Charge
User Fee
Recycling Subsidy
Status Ouo
8 12 16
Policy Race
4 8 12 16
Policy Rate
Figure 3. Net waste effects.
149
-------
50
1 (
*H / C
£ 45
nJ
u
V
C ? ,3 jU
CO Pi
o 25
0 /
V
01
00
dl
Waste Generation .
A A A""\ ' A
*--*--*! Recycling *' .'
* - *- -*J ,'
,.'' y '
' f'
'' *
s /
f *' j'
t' y'
,'' ,'
lf" *'
.' '
N' '_ »'
». - ' ^ . *
^7 v/
^*-~ '">»,
""*-^ "^-.
*., """-i
S* >»
*"^ ^"~~.^
*""-
8 12 16
Policy Rate
(I)
48 12 16
Policy Rate
(ID
8 12 16
Policy Rate
(III)
Figure 4. Waste generation and resource recovery effects.
-------
Recycling Rate Effects--
Figure 5 shows the recycling rates (recycling levels divided by gross
waste generation levels) associated with each policy. The recycling rates
are greater under the product charge at any particular charge/sub sidy rate
than under the recycling policy, except under policy simulation III when the
comparatively greater returns to scale in recycled materials-based container
manufacture lead to higher recycling rates under a recycling subsidy. Also,
for all positive charge and subsidy rates, the recycling rates under either
of those two policies exceeds that associated with the status quo. That is
not the case for the user fee under which the recycling rate may actually
decline (simulation III),
This figure combined with Figure 2 makes a point that should be, but
apparently isn't always, obvious: policy-induced increases in the recycling
rate are very poor indicators of associated changes in social welfare. For
example, even though the user fee leads to much lower recycling rates than
the recycling subsidy (as shown by Figure A), it is nonetheless uniformly
superior to the recycling subsidy in a social welfare sense (as shown by
Figure 1). Also, a very high product charge rate would actually make society
worse off than under the status quo, but it would at the same time further
increase the recycling rate.
151
-------
(I)
(II)
(III)
c
0)
-------
REFERENCES
1. Anderson, Robert C. "Public Policies toward the Use of Scrap
Materials." American Economic Review. 56, No. 1 (Feb., 1977):
355-358.
2. Arthur D. Little, Inc. The Role of Packaging in the U.S. Economy. New
York: American Foundation for Management Research, Inc., 1966.
3. Becker, G. S. "A Theory of the Allocation of Time." The Economic
Journal. 75, No. 299 (Sept., 1965): 493-517.
4. Conn, W. D. "Analysis of Source Reduction Policies." Unpublished draft
for the California Solid Waste Management Board. (December, 1975).
5. Fiekowsky, S. "Tax Subsidies for Recycling Solid Waste Materials: An
Overview." Treasury Department, Unpublished Staff Memorandum BSM-76-15.
(1975).
6. Johnson, H. G. "The Political Economy of Opulence." Canadian Journal
of Economics and Political Science. 26, No. 4 (Nov., 1960): 552-64.
7. National Commission on Supplies and Shortages. Government and the
Nations Resurces. Washington: National Commission on Supplies and
Shortages, 1976.
8. "NSDA's Reed Proposes Litter Law: Bottlers, Brewers, Would be Taxed for
Program Litter Control Funding," Beverage Industry, Vol. 62, No.
(February 18, 1977), p. 17 ff.
9, Page, R. T. Conservation and Economic Efficiency; An Approach to
Materials Policy. Baltimore: Johns Hopkins Press for Resources for the
Future, 1977.
10. Smith, F.L. "The Disposal Charge Concept." Environmental Protection
Agency, Unpublished Staff Paper. (Sept., 1974).
11. Summers, Wall en M. "Externalities of Paper and Paperboard."
Unpublished dissertation, Harvard University, 1973.
12. Train, R. E. Official letter to Senator Russell B. Long, Chairman,
Committee on Finance on a revised Senate Recycling Tax Credit proposal.
February 27, 1976.
153
-------
13. U.S. Congress. Resource Conservation and Recovery Act of 1976. Public
Law 94-580, 94th Cong., 2nd Sess., 1976.
14. U.S. Senate, Committee on Public Works. Hearings, to Consider the
Effects of Product Disposal Charges on Municipal Waste Recovery and
Reuse. 94th Cong., 2nd Sess., May 20, 1976.
15. Wertz, K. L. "Economic Factors Influencing Households' Production of
Re fuse." Journal of Environmental Economics and Management 2, No. 4
(April 1976): 263-72.
154
-------
SECTION 5
PRICING AND RESOURCE RECOVERY TECHNOLOGY
"Conditionally Predictive Estimates of Secondary Materials Supply" (Tayler H.
Bingham, Curtis E. Youngblood, Philip C. Cooley).
"Secondary Material Demand and Supply Responses" (Robert C. Anderson).
155
-------
CONDITIONALLY PREDICTIVE ESTIMATES OF SECONDARY MATERIALS SUPPLY
by
TAYLER H. BINGHAM
CURTIS E. YOUNGBLOOD
PHILIP C. COOLEY
INTRODUCTION
The effectiveness of a number of solid waste policies, most notably
the disposal charge, is influenced by the potential supply of secondary
materials. The initial efforts in this area were, necessarily, descriptive
studies that presented materials flows, described solid waste management
practices and discussed the institutional and economic impediments to
additional resource recovery. Quantitative studies of secondary materials
supply from the post-consumer (municipal) waste stream have been hampered by
a number of obstaclesmost notably the lack of an adequate data base, and
for many materials, the lack of any significant level of recycling.
There are basically two approaches to supply estimation: econometric
and process modeling. The econometric method utilizes statistical techniques
and empirical observations to develop positive estimates of behavioral
responses to exogenous events. These estimates relate to what has happened
in the past and can provide the basis for good predictions if there has been
no change in the economic or institutional relationships. The process model
approach to supply estimation is normativeit describes what should happen,
given various underlying factors and the objective function of the
institution. The quality of the predictions developed from the use of a
process model is only as good as the set of choices that are modeled and the
relevance of the assumed objective function to that of decision makers.
A small number of econometric studies of secondary materials supply
have been undertaken. Most of the estimates indicate an inelastic supply
function for the secondary mater)als studied. Frequently, however, the
materials studied have not been post-consumer wastes but rather obsolete
scrap from industrial establishments.
This paper presents the results of utilizing a process model to
predict the quantity of secondary materials recovered from the solid waste
stream under alternative prices for these materials. The quantity-price data
156
-------
are then used to estimate a conditionally predictive set of own- and
cross-price supply elasticities for secondary materials.
The general approach has been employed by Griffin (1977) in another
setting where the joint product nature of production was importantpetroleum
refining. The joint product problem, which arises when the production of one
product (e.g., beef) results in the concomitant production of another (e.g.,
hides) is not handled well by current econometric methods. Many of the
existing resource recovery systems have important joint product features.
For example, aluminum recovery requires the recovery of ferrous materials.
Material flows are identified in the following section. This is
followed by a discussion of the solid waste management options, their costs,
and the process model. Finally, statistical estimates of the supply of
secondary ferrous, glass and aluminum are presented and some conclusions
drawn.
MATERIALS FLOWS
The process model presented later in this paper uses a materials
balance framework to account for all materials flows. In this section the
materials and their major paths are identified.
In 1976, individuals, commercial establishments, and institutions
generated almost 145 million tons of solid waste (about 3.7 pounds per person
per day) of which nine million tons, or six percent, were recycled. Thus,
approximately 136 million tons were disposed of by dumping, littering
landfilling, and incinerating. Table 1 shows the breakdown of municipal
solid wastes used in this study. The category "all other waste" is the
residual which is included for completeness. It does not include
agricultural and mining wastes, construction and demolition debris, abandoned
vehicles, or industrial wastes.
Current levels of recycling are presented in table 1 for each of the
five components of the municipal waste stream. Paper and aluminum have the
highest recycling rates. The relative availability of aluminum recycling
centers and the price paid for aluminum cans ($340 per ton in 1976) make
aluminum an attractive material to recover. Newspapers, corrugated
containers, and office paper constituted over three-fourths of all paper
recycled in 1976. When separated from other wastes at the point of
generation (commercial and business establishments), old corrugated
containers and sorted white ledger (office paper) have the characteristics
cited by Page (1976) that make a material more amenable to recycling: mass,
concentration of mass, known contaminants, and homogeneity, or consistency.
Newspapers, while generated in sufficient quantities, are dispersed over a
much wider area than are corrugated container and office paper. A paper
drive accumulates these dispersed quantities into a mass sufficient for
recycling.
Glass and ferrous metals from postconsumer wastes are not recovered
in large quantities. The raw materials used to make glass are so cheap and
abundant that there is little incentive to recover glass from postconsumer
157
-------
TABLE 1. COMPOSITION AND DISPOSITION OF POSTCONSUMER SOLID
WASTES, 1976
Disposition
Composition
Paper
Glass
Ferrous
Aluminum
All other waste
Total (Average)
Amount
(106 tons)
50.4
13.9
11.2
1.2
68.0
144.7
Proportion
.348
.096
.077
.008
.470
1.000
Recycled
Share
(%)
16.1
2.7
0.9
8.8
0.0
6.0
Amount
(106 tons)
3.1
.4
.1
.1
0.0
8.7
Disposed
(106 tons)
42.3
13.5
11.1
1.1
68.0
136.0
Source: Adapted from estimates by Franklin Associates, Ltd. for the Resource
Recovery Division, Office of Solid Waste, U.S. Environmental
Protection Agency, January 1978.
158
-------
wastes. Gullet of known quality needed in the glassmaking process is easily
generated by the producer. Gullet from municipal wastes, on the other hand,
must be cleaned of contaminants like metal and ceramics and color sorted!
Its weight-to-value ratio is so large that transporting it more than a few
miles is uneconomical. Ferrous metals, too, suffer from contaminants,
chiefly tin, that must be removed prior to reuse in steelmaking. Ferrous
scrap is also much less valuable than aluminum.
As shown in figure 1, the generated solid waste may either go
directly to the municipal or private collector, or some of the more useful
materials may be source separated and recycled. Source separation can be
either voluntary or mandatory. The majority of the recycling activity
presented in table 1 represents the voluntary action of the waste generators.
Voluntary source separation alters not only the composition of the
waste stream but also its energy content. This affects the economics of the
various waste management options facing municipalities, as will be seen
below. Counterbalancing this effect, however, voluntary source separation
decreases the volume of municipal wastes that must be collected and disposed.
One difference between mandatory and voluntary source separation is that the
revenues from the sale of the secondary materials accrue directly to the
municipality under mandatory separation, thus offsetting some or all of the
incremental costs of separate collection.
The actual waste collected by the private firm or municipality is
usually unseparated. However, in some areas, separation of certain fractions
is required as a condition of servicemandatory source separation. The city
(or the city's contracted hauler) collects these source-separated wastes for
delivery to a processor. In voluntary activities, this service is provided
by the private sector.
A survey of resource recovery and waste reduction programs throughout
the U.S. (McEwen, 1977) showed that as of August 1974, 174 cities had
implemented separate collection programs. Of these, 156 collected newsprint
only; the remaining 18 collected various combinations of mixed paper, glass,
and cans. Whether all of these programs are mandatory in the sense that we
use the word is doubtful.
Unlike some forms of voluntary source separation, there are not
direct financial incentives for generators to comply; since enforcement of
the municipal ordinance requiring source separation is practically
impossible, compliance has a strong voluntary element. If the benefits are
to outweigh the costs, it requires the cooperation of waste generators.
Mandatory source separation reduces the volume of waste requiring
disposal, just as voluntary recycling does. This is increasingly important
as landfill space becomes more scarce. Thus, avoided disposal costs are one
of the benefits of mandatory source separation. It also alters the volume,
composition, and energy content of the solid waste stream, which naturally
affects the economics of processing and disposing of the remaining wastes.
159
-------
Haste
Generation
Voluntary
Source
' Separation
Waste
Collection
Mandatory
Source
"Separation
T
Pre-processing
Materials/
Energy
^Recovery
Sanitary
Landfill
Paper
Glass
Ferrous
Muminum
Energy
Figure 1. Solid waste material flows.
-------
The unseparated wastes next follow one of three paths leading to
final disposition: 1) disposal, 2) preprocessing and disposal, or 3)
preprocessing, resource recovery, and disposal.
Neglecting open dumping, the sanitary landfill is the most widespread
method of waste disposal currently used in the U.S. Although this method
does require expenditures on heavy equipment, such as compactors, scrapers,
and graders, the major cost for most communities is the cost of land.
Because of recent restrictions on and requirements for landfills, the
availability of suitable sites has shrunk. The reluctance of property owners
to sell acreage for use as a landfill and their resistance to having a
landfill adjacent to their properties further constricts supply. Additional
processing to achieve more size reduction and more resource recovery will
extend the life of sanitary landfills.
Processing and disposal, the second major alternative, involves
shredding the waste stream prior to disposal. The size reduction achieved by
shredding is substantial. For instance, a sanitary landfill with an expected
life of 20 years designed to handle 1,000 tons per day requires about 31
percent more land than does a landfill of the same capacity at which the
wastes are shredded prior to disposal. Additional size reduction can be
achieved by incineration, after the waste has passed through a one-stage
shredding operation. The residue remaining is landfilled; it represents
about 10 percent by volume and about 25 to 35 percent by weight of the
incoming waste stream (Levy and Rigo, 1976, p. 26).
The third major alternative available to municipalities is to process
the waste stream, recover certain of the resources, and dispose of the
remainder. The basic processing step is to shred the wastes. Resource
recovery connotes both retrieval of the glass, ferrous, and aluminum fraction
for recycling and the recovery of the energy content of the combustible
fraction. The two methods used to recover energy from the waste stream are:
1) air classification which separates the "light" (combustible) fraction
from the "heavy" (inorganic) fraction to produce a refuse-derived fuel (fluff
RDF), and 2) incineration to generate steam. The steam is a more valuable
product than fluff RDF, but requires more processing to make it so.
Materials recovery takes place at several points following the
shredding operation. In plants that produce a fluff RDF, the ferrous portion
is magnetically separated prior to and just after air classification while
the glass and aluminum fraction is separated and recovered after the waste
stream has been air classified. For the incineration processes, we assume
that the ferrous, glass, and aluminum fractions are separated and recovered
prior to incineration. Disposal of the wastes remaining following resource
recovery operations is usually by sanitary landiflling.
Table 2 shows the resource recovery facilities in operation in 1976
across the nation. Quite obviously, most solid wastes do not enter resource
recovery facilities but are directly disposed of.
161
-------
TABLE 2* MIXED-WASTE RESOURCE RECOVERY FACILITIES IN OPERATION (SUMMER 1976)
Location
AHoona, Pennsylvania
Ames, Iowa
Blythevflle, Arkansas
Braintree, Massachusetts
Chicago, Illinois (Southwest)
E. Brldgewater, Massachusetts
Franklin, Ohio
Grove ton, New Hampshire
Harrlsburg, Pennsylvania
Merrlck, Mew York
Miami, Florida
Nashville, Tennessee
Norfolk, Virginia
Oceans Ide, New York
Palos Verdes, California
St. Louis, Missouri
Saugus, Massachusetts
SI loam Springs, Arkansas
S. Charleston, West Virginia
Type
Compost
RDF
Incineration
Incineration
Incineration
RDF
Wet pulp
Incineration
Incineration
Incineration
Incineration
Incineration
Incineration
Incineration
Methane recovery
RDF
Incineration
Incineration
Pyrolysis
Capacity
(tons per day)
200
400
50
240
1,200
160
150
30
720
600
900
720
360
750
300
1,200
20
200
Products /markets
Humus
RDF, Fe, A1
Steam/process
Steam/process
Steam/Industrial park
RDF/utility
Fiber, Fe, glass, Al
Steam/process
Steam/ sludge drying
Electricity
Steam
Steam heating & cooling
Steam/navy base
Steam
Gas utility & Fe
RDF coal-fired utility
Steam/process
Steam/canning Industry
Gas, Fe
Startup.
date
1963
9/75
11/75
1971
1963
1974
1971
1975
1972
1952
1956
7/74
1967
1965/74
6/75
1972
4/76
9/75
1974
Source: L. B. McEwen, Jr., Waste Reduction and Resource Recovery Activities; A Nationwide Survey.
Environmental Protection Publication SW-142, Washington, 1977, p. 2.
-------
SOLID WASTE MANAGEMENT OPTIONS
The solid waste management options represented in the process model
include the major alternatives available to municipalities. These
alternatives can, as shown in figure 1, be divided into six broad solid waste
management strategies:
1) sanitary landfill,
2) mandatory source separation-sanitary landfill,
3) mandatory source separation-preprocessing-sanitary landfill,
4) mandatory source separation-preprocessing-resource recovery-
sanitary landfill,
5) preprocessing-sanitary landfill, and
6) preprocessing-resource recovery-sanitary landfill.
The six strategies represent combinations of four options (santiary
landfill* mandatory source separation, preprocessing, and resource recovery.
Table 3 identifies the choices that the decision maker has within each
option. These options represent the majority of those currently available to
municipalities although, as shown in table 2, the experience with most is
limited.
TABLE 3. SOLID WASTE MANAGE-
MENT OPTIONS
Sanitary landfill
Mandatory source separation
Paper recovery
Glass recovery
Ferrous recovery
Aluminum recovery
Preprocessing
Shredding
Incineration
Resource recovery
Glass recovery
Ferrous recovery
Aluminum recovery
Refuse-derived fuel
Steam
163
-------
Sanitary Landfill
The sanitary landfill is currently the most common method of waste
management. The location of the landfill is criticalan attempt is usually
made to locate it so that the externalities imposed on residents of the
municipality are minimized: the heavy truck traffic, the noise of heavy
machinery, the dust, and the noxious odors all contribute to a desire on the
part of residents to have the landfill located as far away from themselves as
possible. The geology of the landfill site must also be selected or altered
so that the leachate (runoff water) does not contaminate water supplies.
The landfill in our model has a lifetime of 20 years. The site has
twice as many acres as are actually needed for disposal; this helps minimize
the externalities mentioned above and provides room for access roads for
garbage trucks and other vehicles. Each acre used for disposal is dug out to
a depth of 25 feet and can hold 20,000 tons of waste, when properly processed
by compactors, scrapers, graders, and truck-type tractors that constitute the
landfill's capital equipment. In the usual procedure, a load of waste is
deposited in the pit, compacted, leveled and then covered with two or three
feet of earth. The layers of dirt help prevent odors, rodent infestations,
and unsightlinessmajor complaints leveled against the open dumps still in
use in some areas of the country.
Properly managed, the site can be reclaimed for other usesusually
recreational at the end of its useful life as a landfill. Once a site is
closed, it sits idle for a period of time to allow decomposition and settling
to take place. Since the site does have alternative uses, the original
investment in the land can be recouped. Thus, the true cost of the land to
the municipality is the cost of the money used to purchase the land, i.e.,
the interest rate.
Mandatory Source Separation
Once the most valuable wastes have been skimmed, the remaining wastes
are collected from their points of generation: households, businesses, and
other commercial establishments. It is at this point that mandatory source
separation can take place. There are two major types of separate collection
programs, one in which newspapers only are picked up, the other in which
newspapers and mixed cans and glass are collected.
Newspapers are the easiest component of residential solid waste to
separate and to maintain in a relatively clean condition, i.e., free from
contaminants such as moist food wastes. The newspapers/mixed cans and glass
program requires each participant to separate his wastes into three
categories: newspapers, cans and glass, and all other wastes. The time and
storage requirements are greater than those for the newspaper program, but
the possibility of diverting the large can and glass fraction of municipal
solid wastes could prove economically attractive enough for the city to
mandate such a program. The attractiveness must be tempered, however, with
the realization that the more time and effort the program imposes on
participants, the greater the likelihood that people will not cooperate.
164
-------
The success of mandatory source separation programs depends on four
factors (EPA, 1977): 1) the availability of long-term markets for the
recovered materials, 2) an ongoing publicity campaign to maintain the
public s awareness of the program, 3) careful planning, and A) an
antiscavenger ordinance that prohibits pickup of the separated materials by
any party other than the municipality or the municipality's contract hauler.
Success can be partially measured by the participation ratethe percentage
of residents in the collection area who place separated materials out for
collection. Because many of the incremental collection costs aresfixed,
increasing participation in the program means lower costs and usually greater
revenues to offset those costs.
This is plainly shown when comparing the Marblehead and Somerville
separate collection programs. During the first nine months of 1976, the
Marblehead program with a participation rate of 75 to 80 percent resulted in
net savings* of around $2,857 per month. Approximately 26 percent of the
total waste stream (by weight) was diverted from disposal. In the Somerville
program, on the other hand, participation was much lower during the same
period of time so that net savings averaged about $333 per month. Only 8
percent of Somerville's waste stream was diverted. Eliciting the community's
support and cooperation, then, is probably the key factor in determining the
success or failure of a separate collection program. Details of these
programs can be found in several publications (EPA, 1977; McEwen 1977-
Resource Planning Associates, Inc., 1976). ' '
The future of such source separation programs appears promising.
They involve smaller capital expenditures than do other methods of processing
municipal wastes. There are fewer uncertainties associated with this
approach than there are with other more capital-intensive methods. Also
with disposal sites becoming increasingly scarce and more expensive'
mandatory source separation programs offer a relatively inexpensive way to
extend the life of sanitary landfills.
Refuse-Derived Fuel Plants
While shredding can reduce landfill costs to a certain extent, the
municipality is still disposing of a large proportion of its solid wastes
(betwen 80 and 90 percent, even if the glass, ferrous, and aluminum fractions
are recovered). In cases where landfill costs are prohibitive, or where
energy prices are favorable, the municipality may find that some form of
energy recovery is necessary to keep disposal costs at a manageable level.
There are two routes a municipality can take to recover energy from its solid
wastes: 1) produce a refuse-derived fuel (RDF), or 2) incinerate the
wastes to generate steam. This section considers the former alternative-
incineration is discussed in a later section.
incremental collection costs.
165
-------
The RDF option's produce a fluff RDF consisting of combustible
particles with an average size of one inch and with most of the dense
organics and inorganics removed. It is primarily used as a supplementary
fuel in coal-burning facilities. Typical processing operations are depicted
in the flowchart of figure 2.
The waste stream is first shredded to an average particle size of
four to eight inches. A portion of the ferrous metals is recovered by
magnetic separation and the remaining wastes are air classified. The light
fraction is passed through a trommel to separate glass fines and any
remaining heavies. The wastes that do not fall through the trommel are then
shredded again to produce particles ranging from I/A inch to 2 inches in size
and with a Btu content of about 5,000 Btu per pound. The fluff is then piped
pneumatically to a storage location and then to a coal-fired boiler, where it
constitutes between 10 and 20 percent of the charge.
If no materials recovery is being undertaken, the heavy fraction,
including the materials that fell through the trommel, is landfilled.
However, if any materials recovery is to take place, the heavies from the air
classifier and the trommel undergo a second stage of magnetic separation.
Aluminum is recovered (if desired) by eddy current separation and glass (if
desired) by froth flotation. Any remaining residue is landfilled.
This procedure can recover approximately 90 percent of the light
fraction in the form of fluff RDF. This would, of course, mean a
considerable savings in disposal costs, since the lights constitute about
three-fourths of municipal solid wastes. Adding materials recovery modules
to the basic RDF system will also reduce disposal costs.
Shredding and Materials Recovery
Under certain circumstances, such as high land prices or favorable
secondary materials prices, it may be more economical for a municipality to
process its waste stream prior to disposal or to undertake the recovery of
certain materials, rather than simply landfill all of it wastes. The options
shown in figure 1 give the municipality this sort of flexibility in
determining its least-cost waste management program.
The flowchart of figure 3 illustrates the shredder options. The
first stage of this set of alternatives is to pass all wastes through a
shredder. The size reduction achieved by shredding is considerable; the
wastes leaving the shredder have an increased density, which means that
larger quantities of waste can be landfilled in the same amount of land
compared with unprocessed wastes. After shredding, the density of the waste
166
-------
Mixed
Hastes
Primary
Shredder
Key:
mixed solid wastes
r._ ,'-\ ___
Drum I^^J Air Classifier
Magnet ^^^^
i
L_ _ _ _ _^
. organlcs, paper
Trammel
A
*
Drum
Magnet
'~A '
i
i
i
i
i
..J
Secondar
Shredder
y
Eddy
Current
i
*
0 «
Froth
Flotation
1
1
1
1
1
._ *__
Fluff
RDF
* Ferrous
J Aluminum
J Glass
1
Sanitary
Landfill
Figure 2. Solid waste processing in refuse-derived fuel plant.
-------
oo
Drum
Magnet
Air Classifier
L f
Key:
mixed solid wastes
"light" fraction, e.g.. paper. Inorganics
"heavy" fraction, e.g.. Inorganics
i J
Drum
Magnet
Trommel I 1->
Eddy Current
Aluminum
Figure 3. Solid waste processing by shredding.
-------
stream is about 30 percent greater than that of unprocessed waste, i.e.,
about 26,000 tons of shredded waste can now be landfilled on each acre of
land. The result is to extend the life of an existing landfill or to reduce
the number of acres required to dispose of a fixed amount of waste. The
shred-dispose option, one of two process-dispose options noted in the
overview above, is represented in the flowchart by the heavy arrow connecting
the boxes labeled "shredder" and "sanitary landfill."
If any type of materials recovery is to be undertaken, the shredded
wastes first move on a conveyor belt under a drum magnet, the first stage of
ferrous metals separation. If no other materials are being recovered, the
shredded wastes are passed under a second drum magnet; the wastes are
landfilled and the recovered ferrous scrap undergoes cleaning and some
additional processing prior to sale. Magnetic separation can recover
approximately 90 percent of the ferrous metals in the waste stream. It is a
proven tehnology. The recovered materials can be sold to detitmers or
directly to some steel producers.
The processing sequence is slightly altered if glass or aluminum is
to be recovered in addition to the ferrous metals. As shown in the
flowchart, after passing under the first drum magnet the shredded waste
stream is air classified by dumping it into an upward-moving column of air.
Basically, the air classifier divides the waste stream into a "light"
fraction consisting of organics, such as paper, food, and yard wastes, and a
"heavy" fraction composed of inorganics, such as metals, glass, dirt, and
rocks, and dense organics, such as wet food and yard wastes. The light
fraction is landfilled. The "heavies" pass under a second drum magnet for
additional recovery of ferrous metals, then into a trommel, a cylindrical
revolving screen that allows "heavy" particles smaller than four inches to
drop through for further processing. Particles remaining in the trommel are
landfilled.
The feed to the glass or aluminum separation modules consists of
glass, aluminum, rocks, dirt, bones, some ferrous metals, heavy organics and
some other inorganics (Levy and Rigo, 1976). These modules are
complementaryseparation of aluminum leaves a fraction that is rich in
glass, and vice versa. However, as shown by the flowchart, we allow for the
recovery of one, the other, or both. If glass and aluminum are to be
recovered, the remaining heavies are passed "on a conveyor belt over a linear
motor. Electric current in the motor creates a magnetic field that induces
an eddy current in any conductor (metallics) passing through the field. The
induced current in the conductors opposes that being generated by the linear
motor strongly enough to knock the conductors off the belt. Eddy current
separation, also known as the "aluminum magnet," can recover about 70 percent
of the aluminum fraction.
169
-------
Glass cullet is separated from the remaining stream of heavies using
froth flotation. In this process the stream is treated with a reagent that
gives glass "hydrophobia surface characteristics," i.e., that tends to make
it float. This stream is then fed into the flotation chamber, which contains
a liquid undergoing constant agitation by rotors. The air bubbles produced
by the agitation attach themselves to the glass particles, which then float
to the surface. The resulting froth is skimmed and cleaned. The product is
a clean mixed cullet, or "sand," that can be used in small amounts in
glassmaking. Compared to color-sorted cullet, however, the product of this
model is a low quality aggregate mainly suitable for use in making brpcks or
"glassphalt."
This system can recover about 66 percent of the glass in the solid
waste stream. The residue remaining in the flotation chamber (and which does
not float to the surface) is removed and landfilled.
Incineration
Incineration is the second major form of energy recovery available to
municipalities and included in the model. Two types of incinerators are
available to municipalities, depending on their capacity requirements. We
assume that a municipality processing less than 500 tons per day will use the
small modular, or package, incinerators. These are available in capacities
of 25 and 50 tons per day and can be connected in series to achieve the
desired capacity. For those cities requiring processing capacity between 500
and 3,000 tons per day, the larger waterwall incinerators are available. The
combustion chambers of these incinerators are enclosed by closely-spaced
water-filled tubes that recover heat from the burning waste. Steam is
generated by built-in recovery boilers, which also reduce the temperature and
volume of the exhaust gases.
Solid waste processing by incineration is depicted in the flowchart
of figure 4. One feature of this flowchart that bears noting is that any
materials recovery that occurs takes place on the "front end," i.e., prior to
incineration. While materials recovery from the incinerated residue is
possible, the recovered materials, especially the ferrous scrap, are
fundamentally different from the product recovered prior to incineration.
Incineration of tin-plated steel causes the tin to fuse with the steel, thus
making detinning impossible. Incinerator bundles, as they are called, are
only suitable for use as scrap inputs to foundries. To make the materials
recovered by a municipal waste mangement program homogeneous, regardless of
the option selected, we employed "front end" processing of wastes destined
for incineration.
In the processing sequence depicted in the flowchart, bulky wastes,
e.g., appliances, are manually separated from the waste stream prior to
shredding. This applies only to municipalities using the modular
incinerators, since the shredders feeding these incinerators are too small to
process large, bulky wastes. After shredding, the waste stream is air
classified and divided into light and heavy fractions. The light fraction is
fed into the furnace by spreader stokers that move the wastes across the
combustion chamber to a traveling grate. This is a semi-suspension
170
-------
Mixed
Wastes
Key:
Bulky Shredder . Air Classifier '
Waste } * 1
1 i
« ^
' i
i
i
I £
solid wastes
> Drum .j, Trommel _ -A, Sanitary
Magnet ^ 7k ^ Landfill
_ i
' i
^, j
.... _ |~ I
Currpnt 1 1
i
I
i
"light" fraction, e.g., organics, paper
"heavy" fraction, e.g.. Inorganics
Froth
Flotation
->I Glass
Figure 4. Solid waste processing by incineration.
-------
incineration process in which the waste is partially burned while falling
through the chamber and finishes burning while lying on the grate. Air is
introduced below (underfire air) and above (overfire air) the grate to
promote the mixing of gases and to ensure complete combustion. The ash and
any incinerated inorganics are landfilled. The ferrous, aluminum, and glass
in the heavy fraction are removed (if at all) by magnetic separation, eddy
current separation, and froth flotation, respectively. All unrecovered
residue is landfilled.
Incineration reduces raw waste to approximately 25 to 35 percent of
its original weight, including the glass, ferrous, and aluminum fractions.
The steam generated is about three times more valuable than fluff RDF; this
occurs because the incineration option includes an additional processing step
(combustion), which increases the cost. Approximately 59 percent of the
energy content of the original waste stream is recovered in the form of
marketable steam (Levy and Rigo, 1976).
SOLID WASTE MANAGEMENT COSTS
The solid waste management costs for the options identified in the
previous section are discussed below. The costs were primarily derived from
data developed for the Kansas City Metropolitan region in a 1977 study by
Franklin Associates and Black & Veatch, Consulting Engineers. In that study
costs were determined from contacts with equipment suppliers and from
operating RDF plants for 500 and 1,000 tons per day with materials recovery,
and shredding plants with materials recovery. Extrapolations were made for
other plant capacities. Because of the limited data, mandatory source
separation is not covered.
Although it is conceivable that in a few areas of the country, plants
larger than 3,000 tons per day may be the best option, those plants were not
considered for this analysis. Plants larger than 3,000 tons/day have special
siting problems primarily due to the heavy truck traffic in and out of the
plants, and high transportation costs. They cannot generally be treated as
simple extensions of the lower capacity plants.
Three basic types of mixed solid waste (MSW) plants are considered in
this analysis. One is a plant which genrates a "fluff" RDF to be burned as a
supplementary fuel in a coal-burning facility and which also recovers some
combination of steel, aluminum, and glass for recycling. Another type of
plant simply shreds the solid waste for landfilling and has the capacity for
adding material recovery facilities. The third type is an incineration
facility that generates steam.
The RDF plants are generally similar to the facilities in Ames, Iowa,
Southwest Supplementary Fuel Processi g facility in Chicago, and the
Americology plant in Milwaukee. The shredding plant is similar to the
Recovery I facility in New Orleans.
The RDF and shredding plants were evaluated at capacities of 500,
1,000, 2,000, and 3,000 tons of solid waste per day. The incineration
facilities were evaluated at capacities of 50, 200, 500, 1,000, and 2,000
172
-------
tons per day. For each of the capacities, costs were developed for plants
with no material recovery, steel recovery only, steel and aluminum recovery,
steel and glass recovery, and steel, aluminum, and glass recovery. Plants
with aluminum and/or glass recovery without steel recovery were not
considered since it is essential to remove the steel from the waste stream to
successfully remove the aluminum and glass.
Sanitary Landfill
The initial per acre cost of the land depends on several things:
alternative uses of surrounding land, distance of the site from population
centers, population density of the region, etc. They are also undoubtedly
influenced by a host of region-specific factors, such as the growth rates of
industry and population in the region, climate, employment, etc. A
preliminary estimate of landfill costs was developed using data on landfill
practices in 17 communities (table 4). The 1975 populations of these cities
ranged from 21,000 to 368,000 and represented five of the 10 size categories
used in the process model (table 5). Data on the cost and size of the
landfill sites were available for 15 of the 17 cities (two municipalities
were leasing the sites). From these data, the per acre price of the land was
calculated; it ranged from $163 to $10,000 per acre. A simple functional
form relating price to the population and to the square of the population was
fitted to these data using ordinary least squares:
LPRICE = 3.1461(POP) + 0.0492(POP) where
(0.47) (2.01)
LPRICE = price per acre
POP = population (in thousands).
The t-values for the coefficient estimates are in parentheses. The R for
this function was .76. The predicted land values for each representative
city size in the model are given in table 6. The estimated land prices for
the larger and smaller city sizes tend to be extremely large and small,
respectively. This occurs because we extrapolated beyond the range of our
dataa risky procedure at best, but one that is superior to the arbitrary
assignment of "reasonable" values to those out-of-range cities.
Refuse-Derived Fuel Plants
The RDF plants for which costs were developed are complete systems.
The items, not all itemized in the analysis, but which are included are
listed below.
Truck scales
Tipping floor conveyors
Primary and secondary shredders with necessary input and output
conveyors. (Only 50 tons per hour or larger shredders
were considered.)
Trommel screenused only when there is metal or glass recovery
Air classifier subsystem
173
-------
TABLE 4. DATA USED TO ESTIMATE PRICE OF LAND FUNCTION
Landfill data
City-
Cape Glrardeau, Ho.*
Fairfax County, Va
Fresno, Ca.
High Point, N.C.
Independence, Mo.
JopUn, Mo.
KenosJra, Wfs.
Lancaster, Va.
Madison, Wis.
Madison, Wfs.
Nashua, N.H.
Norfolk, Va.
Odessa, Tex.*
Raleigh, N.C.
Richardson, Tex.*
Toledo, Oh.
Topeka, Kans.
Total land
expenditures
(10*$)
___
85
190
60
94
165
40
108
56
50
295
83
245
233
680
65
Size
(acres)
60
81
200
60
44.7
130
33
215-
H
24
295
83
1500
137
227
63
160
Price
($/acre)
_*_
1049
950
1000
2103
917
1212
502
4000
2083
1000
1000
163
2066
10000
406
Population, 1975
(10-3 persons)
31
21
177
67
111
39
80
ISO
168
168
62
287
78
111-
57
368
119
* Landfill site was leased.
Source: EPA, Office of Solid Waste Management Programs.
174
-------
TABLE 5. DISTRIBUTION OF CITIES, BY SIZE CATEGORY, 1976
Size categories
U
500 - 1.000
250 - 500
100 - 250
50 - 100
25 - 50
10 - 25
5 - 10
2.5 - 5
Less than 2.5
Total (Average)
Number
of
places*
s
20
30
100
240
520
1385
1839
2295
627
7062
Total
(103)
19,748.6
13,738.2
10,947.6
15,026.1
17,602.0
13,390.0
22,539.2
13,738.2
8,586.4
853.6
141,674.0
Population
Percent of
U.S. population*
9.2
6.4
5.1
7.0
3.2
8.3
10.5
6.4
4.0
0.4
66.0
Average city
size (103)
3,291.4
686.9
364.9
150.3
73.3
36.3
16.3
7.5
3.7
1.4
20.1
* 1970 valued. The 1976 population was used In calculating the total
population and average city size for each category.
Source: U.S. Bureau of the Census, Statistical Abstract of the United
States; 1977. (98th edition) Washington, D. C., 1977, Table 20.
175
-------
TABLE 6. ESTIMATED PRICE OF LAND USED FOR SANITARY LANDFILL
FOR REPRESENTATIVE CITIES
Size
Category
1,000 +
500
250
100
50
25
10
5
,2.5
- 1,000
- 500*
- 250*
- 100*
- 50*
- 25*
- 10
- 5
< 2.5
Average
population
(103 person)
3291.4
686.9
364.9
150.3
73.3
36.3
15.3
7.5
3.7
1.4
Land
price
(S/acre)
543354. 1
25375.2
7699.1
1111.4
495.0
179.0
64.4
26.4
12.3
4.5
Estimates for these city sizes fall within the range of the data from which the
coefficients were estimated.
176
-------
Pneumatic RDF delivery system to storage and to boiler
Dust collection system
RDF storage system
Control panels
Residue conveyors and bins
Coal burner modifications
Truck loading facilities
Ferrous metal recovery subsystem consisting of two stages of
magnetic separation for each processing line and the
associated conveyors and bins
Aluminum recovery subsystem consisting of one aluminum magnet
per processing line and the associated conveyors and
bins
Glass recovery subsystem (froth flotation)
Delivery and electrical and mechanical installation of all
equipment
Mobile equipment
Supplies and furniture
Some of the assumptions made for this cost analysis are:
1. The fuel generated is "fluff" RDF.
2. Costs for plants as a function of capacity were
determined by the
function:
Ca /Sa\n
Cb " Vsb/
where C is plant costs, S is plant capacity, and n is 0.8
for all costs except labor and land. For labor and land
the exponent is 0.6. These exponents were chosen on the
basis of costs for the 500 and 1,000 TPD plants.
3. Land costs for RDF plants are calculated at
$50,000/acre.
4. The buildings (reinforced concrete) are large enough to
permit the addition of material recovery subsystems.
5. Interest during construction is equal to 6 months of
interest on the total capital investment. (It is assumed
that construction time is 2 years, during which time part
of the capital may be reinvested.)
6. Startup costs are equal to the operating and
maintenance costs for one year. (This is to operate the
plant during shakedown, when there is little or no revenue
expected.)
177
-------
7. Labor costs are based on Kansas City, Missouri rates
and include 40 percent for fringe benefits, overtime, etc.
8. ML RDF plants are assumed to process solid waste 2
shifts per day, 300 days per year with an 85 percent
availability. The availability factor is to allow for
routine maintenance and breakdowns. System costs were not
developed on the premise of providing standby process lines
for use in the event of breakdown. However, diverters and
RDF trailers are included to be used in the event of power
plant breakdown.
Shredding and Materials Recovery
The analysis for the shredding and shredding with material recovery
plants is similar to the RDF plants. However, only one stage of shredding is
needed and facilities and equipment for energy recovery are not included. In
the plant with no material recovery or with only steel recovery the air
classifier system is not needed.
Incinerators
Cost for the modular incinerator are given for 50, 200, and 500 tons
of solid waste per day. The data for the 50 and 200 TPD plants without
materials recovery came from the Franklin Associates Ltd. report; the 500 TPD
data were obtained by extrapolation. The costs for recovering materials were
estimated, based on the previously submitted RDF data. There are no known
municipal modular incinerators in operation where materials are mechanically
separated prior to incineration.
The mechanical processing consists of single stage shredding, air
classification, magnetic separation, aluminum magnet separation, and glass
separation by screening and froth flotation. Because of the smaller size
requirement for the shredders, large bulky items would need to be handpicked
from the waste. (In order to shred bulky wastes, 50 ton per hour or larger
shredders would be required.) Included are cost estimates for truck scales,
mobile equipment for feeding the solid waste into the incinerators, and
delivery and installation of all equipment. The systems for which cost
estimates are made include single lines of mechanical processing and two
parallel incinerator units for the 50 TPD plants, four for the 200 TPD plant,
and eight for the 500 TPD plant. Each incinerator is equipped with a steam
boiler which generates about 3 pounds of low pressure steam per pound of
solid waste. Steam lines are provided to a user approximately 500 feet away
from the boiler.
It is assumed that incinerators would operate continuously, with the
mechanical processing being done eight hours per day, five days per week.
Amortization costs are calculated at 7 percent and 20 years.
178
-------
It is assumed that Incinerators would operate continuously, with the
mechanical processing being done eight hours per day, five days per week.
Amortization costs are calculated at 7 percent and 20 years.
The semi-suspension cost data were derived from actual contract costs
of the 1,000 TPD recycle energy system under construction in Akron, Ohio/
The process In Akron consists of coarse single stage shredding of the solid
waste; combustion of the light fraction in semi-suspension fired waterwall
boilers; recovery of ferrous metals; and future recovery of glass and
aluminum. The aluminum magnet and glass screening and froth flotation system
costs were added from the Franklin Associates Ltd. data.
In the semi-suspension process, the shredded light fraction is
approximately 50 percent burned while falling through the flame, thus
improving the boiler efficiency over mass burning. The remainder is burned
on a grate. The energy product is steam which is distributed to steam
customers which must be located within a mile or two of the plant. In the
Akron plant, steam generation is estimated at 4 pounds per pound of solid
waste. After the steam for in-house use and distribution system losses is
deducted, it is expected that 3.2 pounds steam per pound of solid waste will
be available for sale.
Amortization costs for the equipment and buildings are calculated at
7 percent and 20 years.
Average availability of the modular plant was assumed to be the same
as for the semi-suspension plant, that is, 85 percent, 350 days per year. It
is assumed that the incinerator will produce saturated steam at 560 psig
continuously and that the mechanical processing equipment will operate 5 days
per week, 8 hours per day.
PROCESS MODEL
The sol id waste management options and their costs are included in a
solid waste management process model (SWMPtt) along with relevant waste
generation rates and materials prices. The decision maker, confronted with
an exogenously established flow of solid wastes is assumed to select the
solid waste management option that minimizes the cost of waste management.
The analysis is done for the average city sizes shown in table 5. The flow
of operations is SWMPM is presented in figure 5 and discussed below.
Enter Secondary Materials Prices
The first step in the operation of the model is to determine the set
of secondary materials prices that will confront each representative
2. Hileman, G. P., and F, B. Pyle, "Recycle Energy for Central Heating and
Process Steam," presented to International Heating Association, 1977.
179
-------
C START )
0~
1
i
Select secondary
material price vector
/Average, high,
/ each material
and /
Select cost minimizing
solid waste management
option
Weight results
Population, by
municipal ity
Per capita waste
generation rate
Masta component
Estimate voluntary
recycling, by
material
< 1
*
Estimate recycl ing
through mandatory
source separation
programs
*,
Adjust volume and
composition of
waste stream
4-
/ proper
\
Compute vo
stream, by
/Cost of m<
tions /
r
lume and
n of waste
municipality
/ . Supply elasticities, /
Y hy mflf«"l»l /
/ . Baseline recycling /
/ rates, by material/
mdatory source /
on programs /
Cost of sol td waste
management options :
Energy prices
Interest rates
Outputs:
. Price and quantity supplied,
by material
. Energy price and quantity
recovered as energy, by
material
Summary of waste management
options selected, by
municipality
Figure 5. Flow of operations in the supply model,
180
-------
ow' v- Elem*ntS °l thlS Set are drawn from the universe of "high" and
low deviations from the average price of each material. A total of 16
unique price vectors^ is selected by the model; each drives the model and
the resulting price-quantity pairs constitute the data set.
Selection of a price vector has two major effects: 1) for each
material the percentage change of the high or low price from the average
price elicits some percentage change in the level of voluntary source
separation through an elasticity calculation, 2) secondary material prices
obviously impact the conomics of municipal solid waste management by altering
the potential revenues available from recovering and selling the materials?
The rationale for selecting the range of price variation and the exact method
used to vary the prices are discussed in the section on Price Variation.
Calculate Amount and Composition of Solid Waste
For each representative municipality, the total waste generated each
year is the product of the population and the per capita waste generation
rate. We apply a vector of fixed material proportions to decompose this
total into the five material components used in the model: paper, glass
ferrous materials, aluminum, and all other wastes.
Estimate Extent of Voluntary Source Separation
At this point we estimate the amount of each secondary material that
is voluntary separated at its source of generation, because the level of this
activity directly affects the amount of waste collected and disposed by the
municipality. Voluntary source separation activities are influenced by the
prices of secondary materials. Thus, each material is represented by a
baseline recycling rate that is a function of the average material price; any
deviations from this price alter the baseline recycling rate through a supply
elasticity. The resulting percentage change in quantity supplied is applied
to the amount of that waste material to determine the absolute level of
voluntary source separation taking place.
Estimate Effects of Mandatory Source Separation
As explained above, two programs are available to municipalities, one
involving separate collection of newspapers, the other for collecting
newspapers and mixed cans and bottles. The costs of both programs depend on
their efficiency in recovering the relevant materials. In this and the next
step Uhe cost minimization routine) the model investigates three
alternatives: 1) all municipalities opt for separate collection of
The number of possible unique combinations of high and low prices for
four materials is 2 . The price of the fifth material is held constant
(at zero), since no materials are recovered from "all other" waste.
Energy is recovered, and we account for it in the model, but we did not
need to know the own- and cross-price elasticities of the supply of
energy from solid wastes, so its price was also held constant.
181
-------
newspapers, 2) all municipalities opt for separate collection of newspapers
and cans and bottles, 3) no municipality engages in separate collections.
waste management; the least-cost alternative is then chosen, ^he
municipality in effect has 48 waste management strategies to choose from..
Revise Waste Stream Composition
The amount of each material in the waste stream is adjusted to
reflect the quantities diverted by voluntary and mandatory (if any) source
separation activities.
Minimize Processing and Disposal Costs
The municipality must process and dispose of all remaining solid
wastes by using the least-cost mix of options available to it. The cost of
each option is a nonlinear function of the amount of waste to be processed.
Revenues are derived from the sale of secondary materials and energy.
Unrecovered wastes are disposed of at a cost that also varies with the
amount. This is the general form of the net cost function that is minimized
by the nonlinear optimization algorithm:
min rnin
k L i
2i
a [X(l-Rk) + a3.[X(l-RkMl-E.)
I4i
-X(l-Rk)E.PJ
for k = 1,2,3
i = 1, .. . , 16
subject to
X(l-Rk)Ei
- Rk)(l -E.)
= X*-V
where
X
R
E.
1
total waste to be processed by the municipality
proportion of waste stream diverted by the k
tory source separation program
th
manda-
proportion of waste stream recovered by the i
basic option
price of recovered materials
fch
4. The 48 alternatives come from the 16 basic options, each applied in
conjunction with the three separate collection options.
182
-------
^k = cost of the k mandatory source separation program
X* » total waste generated in the municipality
V «* amount of waste stream diverted by voluntary
source separation
ai » estimated cost coefficients.
The first term is the cost of processing all wastes (net of those
diverted by a mandatory source separation program) by the i basic option;
the second term is the cost of disposing of the wastes remaining after
processing. These costs have three components: land, operating and
maintenance, including labor, and capital. The third term is the revenues
generated by processing wastes using the i basic option. The fourth and
fifth terms represent the costs and revenues, respectively, of the k
separate collection program. The constraint is that the amount recovered
(first term) plus the amount disposed (second term) plus the amount diverted
by mandatory source separation (third term) must equal the total amount of
waste generated in a municipality less the amount that is voluntary recycled.
Outputs
Once the above steps have been performed for each representative
city, the results are weighted by the number of cities in the U. S. in each
size category. The outputs of primary interest are the price-quantity pairs
constituting the data set. Other outputs include the amount of materials
recovered as energy and a summary of the options selected by the
municipalities for handling their solid wastes.
SUPPLY FUNCTIONS
Supply functions were estimated for three secondary materialsglass,
ferrous, and aluminum, using the data generated by the process model
presented above. Since paper is only recovered as a fiber under voluntary
and mandatory source separation, its supply was estimated outside the model
but incorporated in the model, as its recovery affects the energy content of
the waste stream and the economics of resource recovery systems.
Price Variation
The model requires three prices for each material: the average, low,
and high prices. Deviations above and below the average price are used to
determine the amount of "front end" recylcing (i.e., source separation) that
occurs. The municipality is confronted with combinations of the high and low
secondary materials prices16 in all.
This system of price variation is a 24 factorial design In which two
prices for paper, two for glass, two for steel, and two for aluminum are
tested in all combinations. The 16 resulting price vectors represent points
at the 16 corners of the experimental region, which is a four-dimensional
hypercube. Thus, the range of price variation determines the size of the
183
-------
experimental region over which the mathematical model of secondary materials
supply is assumed to be valid. However, as the size of this region
increases, the likelihood that the model is valid over the entire region
decreases. Hence, the range over which the prices are varied is a critical
factor in the estimation procedure. In determining the range of price
variation, we started with average 1976 prices and considered the historical
variation in secondary material prices along with the probable effects of the
disposal charge. The values are shown in table 7.
TABLE 7. SECONDARY MATERIALS PRICES AND SUPPLY
ELASTICITIES
Price
(dollars per ton)
Paper
Glass §
Ferrous metals §
Aluminum
Low
5.0
0.1
9.0
200.0
Average
25.4
10.0
39.0
300.0
High
60.0
40.0
69.0
380.0
Supply
elasticity*
Low
.4
.1
.1
1.1
High
1.7
.71
3.9
4.3
Energy ($/thousand Btu)
RDF 0.5
Steam 1. 2
1.5
3.6
* Estimates supplied by the Resource Conservation Committee.
+ Prices developed from Official Board Markets. Since OMB quotes
are delivered prices, they were multiplied by . 5 to approximate
prices paid to generators.
§ Franklin Associates Ltd. , Memoranda to EPA and ICF, February
17, 1978.
184
-------
Supply Estimates
The data set generated by the process model consisted of 16 sets of
secondary material quantities and prices. A function that was linear in the
logarithms of its variables was fitted to these data using ordinary least
squares. The functional form is given by
InO. =
where Q.
a.
i
> V P3>P4
= the quantity of the ith material recovered, i=2-4,
= the regression coefficients,
= prices the municipality receives for the seondary
materials, and
P5 ~ price the municipality receives for recovered energy.
The resulting elasticity of supply estimates are shown below. They
incorporate voluntary source separation elasticities (Table 7) along with the
estimates of the response of municipalities to the price changes.
;-
Paper
Glass
Steel
Aluminum
Energy
Paper
1.7
0.022*
(0.85)
0.068
(2.56)
0.050
(1.97)
-0.123
(-28.23)
Coefficients
(t-statistics)
Glass
0.0
0.190
(17.63)
0.049
(4.43)
0.044
(4.16)
0.003
(1.76)
Steel
0.0
0.136
(4.27)
0.396
(12.13)
0.246
(7.96)
0.016
(2.93)
Aluminum
0.0
0.120*
(1.19)
0.207
(2.01)
1.102
(11.27)
0.010*
(0.59)
Energy
0.0
0.014*
(0.70)
0.065
(3.05)
0.051
(2.56)
0.236
(68.01)
R2
.928
.882
.896
.995
*Not significant at a = .10 level.
185
-------
The estimated supply elasticities confirm the belief that the supply
of secondary materials from the municipal solid waste stream is fairly
unresponsive to price changes. The data confirm the interdependence of the
supply of secondary materials from the municipal waste stream.
CONCLUSIONS
The combined use of process and econometric models provide a means
for estimating the supply elasticities for products for which there is no
adequate historical record or data base for use in a more positive approach.
It can be used to make predictions of quantities supplied for sets of prices
not used in the estimation (the new price vectors must fall within the
experimental region, however). Because we are using a process model, the
resulting data depict ponts on a well-defined production surface.
The data are subject to two types of errors (Griffin, 1977): 1)
errors in measurement that result from differences between the observed
technological matrix and the technically correct matrix, and 2) behavioral
errors that result because the frictionless long run cost minimization
subsumed by the process analysis/statistical reduction approach does not
obtain in the short run. Another problem involves determining the size of
the experimental region. If it is too small, then the process model tends to
generate an "all or nothing" response which can render statistical estimates
useless. Over too large a region, the mathematical model may not accurately
represent the responses.
186
-------
REFERENCES
1. Baseline Forecasts of Resource Recovery, 1972 to 1990. Midwest Research
Institute. Final Report for Office of Solid Waste Management. U.S. EPA,
March 19757 ~
2. Darnay, Arsen and William E. Franklin, The Role of Packaging in Solid
Waste Management 1966 to 1976. Prepared for Public Health Service
(U.S.H.E.W.) by Midwest Research Institute, 1969.
3. Griffin, James M. "The Econometrics of Joint Production: Another
Approach," The Review of Economics and Statistics. LIX, November 1977,
pp. 389-397T ~~~
4. "Long-Run Production Modeling with Pseudo Data: Electric Power
Generation," The Bell Journal of Economics 8(1), Spring 1977, pp.
112-127. ~~
5. Hunt, Robert G., and William L. Bider, Analysis of Environmental and
Economic Impacts of Waste Reduction Procedures and Policies. Final
Report prepared for Office of Solid Waste Management, U.S. EPA, December
1977.
6. McEwen, L. B., Jr., Waste Reduction and Resource Recovery Activities; A
Nationwide Survey. Environmental Protection Publication SW-142.
Washington, D.C., U.S. Government Printing Office, 1977, 78 p.
7. Quiraby, Thomas H. E., Recycling; The Alternative to Disposal. The Johns
Hopkins University Press, Baltimore, 1975.
8. Savas, E. S., The Organization and Efficiency of Solid Waste Collection.
D. C. Heath and Company, Lexington, Mass., 1977.
9. Sawyer, James W., Automotive Scrap Recycling; Processes, Prices, and
Prospects. Resources for the Future, the Johns Hopkins University Press:
Baltimore, 1974.
10. Shumway, C. Richard and Anne A. Chang, "Linear Programming versus
Positvely Estimated Supply Functions: An Empirical and Methodological
Critique," American Journal of Agricultural Economics, Vol. 59 (2), May
1977, pp. 344-357.
11. Stevens, Barbara J., "Pricing of Solid Waste Collection and Disposal,"
Memorandum, date unknown.
187
-------
12. "Cost of Refuse Collection and Disposal," Memorandum to Fred Smith,
March 1, 1978.
13. U. S. Department of Commerce, Bureau of the Census, Statistical Abstract
of the United States; 1977, (98th edition), Washington, D. C., 1977.
14. U. S. Environmental Protection Agency, Office of Solid Waste Management
Programs, Resource Recovery and Waste Reduction, Fourth report to
Congress, Environmental Protection Publication SW-600, Washington, U. S.
Government Printing Office, 1977, 142 p.
188
-------
SECONDARY MATERIAL DEMAND AND SUPPLY RESPONSES
by
ROBERT C. ANDERSON
INTRODUCTION
The desirability of either the solid waste user fee or the consumer
goods product charge can be evaluated according to several criteria:
administrative feasibility, effectiveness, efficiency, equity, legal and
political feasibility. Other participants in this conference have addressed
the issues of efficiency, equity and administrative feasibility. In this
paper I will address the issue of effectiveness with special reference to the
consumer goods product charge. The principal analytical tools are a set of
econometric models of the ferrous scrap and wastepaper markets, developed
over the past few years under grants from the Environmental Protection
Agency.
As currently envisioned, the consumer goods product charge would be
structured as an excise tax on inputs to the production of consumer goods
which ultimately enter the solid waste stream (e.g., beverage containers,
newspapers, packaging materials, etc.). Of two principal production inputs'
virgin materials and secondary materials, only the virgin inputs would be
subject to the product charge. This would create a change in the relative
prices of virgin and secondary materials and, theoretically, produce shifts
in resource use patterns and in resource recovery as the effects filter
through the economic system.
In order to project accurately the probable impacts of regulatory
actions such as the product charge, one must have a detailed knowledge of the
market parameters of interest, the demand and supply elasticities for both
virgin and secondary materials. This paper is principally devoted to a
discussion of our most recent empirical- work in this area. Quantitative
assessments of the likely impacts of the product charge based upon estimates
presented here are left to others.
189
-------
As noted the econometric models presented here are extensions of the
ferrous scrap and wastepaper demand and supply models developed for EPA in a
previous study. * Three major modifications have been made to the earlier
published models. (1) The original ferrous scrap model was based on the
years 1962-1972; this model has been updated to include data through October
1977. The wastepaper model was previously estimated over the period January
1962-December 1974 and has been extended to August 1977. This additional
period is especially interesting because of the large price changes in both
materials which occurred between 1973 and 1975. (2) The earlier ELI models
for wastepaper and ferrous scrap were two equation demand and supply systems.
Additional equations are added here to represent market behavior more
accurately. The ferrous scrap model now contains supply, demand, export
demand, and mill inventory equations. (3) In the wastepaper model, the
supply of ferrous scrap was initially hypothesized to be a function of price,
activity in the prompt producing sectors (automobile production), scrap
exports, lagged prices and a trend variable. Some alterations were made to
this equation to improve the fit (exclusion of the trend variable) and to
accommodate the export equation (scrap exports were added to the left-hand
side of the equation). A few other minor changes in the demand equation for
ferrous scrap and the wastepaper supply and demand equations are discussed
below.
THE WASTEPAPER MODEL
The wastepaper model, as it now stands, is composed of four
equations.
Demand : ( 1 )
QtD = a0 + alPt + a2X(t) + a3X3(t) + ^4(t) + Ut
Supply: (2)
Inventory: (3)
lt = cO + clI(t-l)+c2Q°+c3AQf+c4p(t-l) + Vt
Environmental Law Institute, Impact of the Federal Tax Code on Resource
Recovery. Prepared for the Environmental Protection Agency.
Washington, D.G.: PB-264 886, 1976. Also, Anderson, R. C. and
Spiegelman, R. D., "Tax Policy and Secondary Material Use." Journal of
Environmental Economics and Management (JEEM), A (1977): 68-82.
190
-------
Where:
Identity: (4)
QSt=QtDH-dt-I(t-1))
Q. is the consumption of wastapaper in thousands of
tons
o
Q, is the purchases of wastepaper in thousands of tons
P, is the wholesale price index of wastepaper deflated
by the wholesale price index for all commodities
(1967 = 199) and multiplied by 1000
X. is the Federal Reserve Board Index of paperboard
container production (1967 = 100) multiplied by 10
X- is the Federal Reserve Board Index of converted
paper product production (1967 = 100) multiplied
by 10
X, is the output of construction paper and board in
thousands of tons
X^ is the output of combination paperboard in
thousands of tons
W is the wholesale price index for market wood pulp
(all grades) deflated by the wholesale price index
for all commodities and multiplied by 1000
P,. .. is an Almon lag over wastepaper prices lagged one
* "^ to four months
I, is inventories of wastepaper held by major
consuming mills at the beginning of the month
in 1000's of tons
Estimates of the structural coefficients (the a's, b's and c's) of
equations (1) through (3) were obtained by two stage least squares. All
equations were corrected for first order serial correlation. The final
estimated equations are reported below. T-ratios appear in parenthesis under
the respective variable.^
2. The properties of T-ratios and other test statistics in simultaneous
equation systems have not been fully developed. The T-ratios should be
interpreted as guidelines to the significance of the variables.
191
-------
Demand: (5)
Q? = 41.9 - 0.13P. - 0.03W. + 0.37X,. + 1.51X4.
* (-4.3) * (-1.0) * (4.8) (15.9)
Supply: (6)
Q? = 51.9 + 0. 36P, + 0.08X. ,. ., + O.Z5X , ?. - 0. 004P .
t (11.4) * (1.5) I(t"1} (4.8) *(t-*} (0.09) lt>J)
P = .45
Inventory: (7)
I, = -39.6 + 0.951,, .. + 0.04QJ3 - 0.08 AQ?5 + 0.03P, ..
t (21.8) (t"1} (0.7) * (-1.2) t (1.7) ^-J'
P = -0.05
All the variables but one have the proper sign as suggested by
economic theory. Demand responds negatively to price. The estimated
elasticity of demand, evaluated at the means, is 0.19. Consumption of
wastepaper is a positive function of activity in construction paper and
board, two sectors of the paper industry that are large users of wastepaper.
While one would expect the price of woodpulp to have a positive impact on
wastepaper demand, the estimates presented here show a negative sign,
although the variable is not significant.
The supply of wastepaper is positively related to price, with an
estimated short-run elasticity of 0.5. The negative sign for lagged prices
is also expected since the level of prices in past time periods provides a
good index to the extent of depletion to the reservoir of available scrap.
No medium term elasticity is reported because the lagged price variable is
insignificant. The generation of prompt industrial scrap, represented by the
lagged values for paper board container and converted paper production, acts
to increase the supply of scrap as indicated by the positive sign for these
variables.
Comparison of the current wastepaper estimates with estimates from
the earlier ELI models, reproduced in Table 1, must be done with some care.
In order to take advantage of the publication of a more consistent series,
the price variable for wastepaper was changed in the updated model. The
previous model used a wholesale price index for wastepaper that was based on
two different base years, 1957-1959 and 1967 = 100, the new series has been
weighted to a single base year, 1967 = 100. It does not seem likely, though,
that this modification can account for the over twofold increase in the
demand elasticity, 0.08 in the first study to .19 in the current work. The
192
-------
same sort of change was made to the woodpulp price index, but cannot be taken
as the cause of the negative sign for that variable. The estimates of the
supply elasticity are similar in both studies.
TABLE 1. COMPARISON OF ESTIMATED ELASTICITIES
FROM PAST AND CURRENT ELI WASTEPAPER MODELS
Elasticity
Demand
Gross price
Supply (short -run)
Supply (medium- run
EPA
0.16
0.13
0.40
0.15
*
Estimate
JEEM
0.08
0.17
0.53
0.40
Current
0.19
>!<*
0. 12
0.50
0.50
**
Reported elasticities refer to the EPA study cited earlier based
on the years 1962-1972, then the results published in JEEM
(based on the years 1962-1974).
Estimates based on the years 1962-1972.
One cause of the difference in the price coefficients of the demand
equations may be found in the market convulsions of 1973-75. The beginning
of the period was marked by a world wide increase in demand for paper
products, as well as most other commodities. This was shortly followed by a
more than proportional increase in the demand for wastepaper as
non-traditional users entered the market. Mills that normally relied upon
woodpulp inputs were buying wastepaper to compensate for a developing
shortage of pulp for paper production.
In response to increased demand, the price of wastepaper more than
doubled from early 1973 to the first quarter of 1974. The price of woodpulp
also rose during this time period. During the period from October 1973 to
November 1974 consuming mills increased their inventories of wastepaper,
possibly in anticipation of even further price increases. (Time series data
for these variables are plotted in Figure 1).
By mid-1974, the commodity boom had begun to slip into the recession
period of 1975. As demand for paper products declined, demand for wastepaper
inputs fell even more as mills used up accumulated inventories instead of
buying in the market. The wholesale price index for wastepaper fell from a
193
-------
II LIONS OF SHORT TONS
O
i-
wastepaper p ice
index A.
purchases or
wastepaper /
stocks or wastepaper
- A/^
TIME
Figure 1. Time series plot of wastepaper data, from January 1962 through August 1977
-------
high of 222.4 in March 1974 to 56.1 in March 1975. The price of woodpulp did
not experience such a severe drop. Actually, the higher prices of woodpulp,
established during the 1973-74 period, held fairly constant throughout the
recession period of 1975. As shown in Figure 2, these price patterns
signalled a change in the historical relationship between woodpulp price and
the consumption of wastepaper.
More formally, the inability of the wastepaper demand model to
capture the substitution effect of woodpulp for wastepaper can be formulated
as an error-in-variables problem. It is likely that while posted woodpulp
prices remained high in the aftermath of the shortage period, actual prices
were somewhat lower due to discounts being given buyers. The wholesale price
index for woodpulp would not account for such discounting. The use of this
index would generate biased estimates of the woodpulp price coefficient and
unreliable elasticity estimates.
To demonstrate the effect of the different price structure for
woodpulp on the demand for wastepaper, the demand equation was re-estimated
using two woodpulp price variables, one representing prices before 1973 and
another representing prices after 1973. It was anticipated that the price of
woodpulp prior to 1973 would have a positive sign and the post-1973 price
would capture the change in the woodpulp price - wastepaper consumption
relationship. The pre-1973 woodpulp price had a marginally significant
positive sign and an elasticity of 0.12. The post-1973 price was not
significantly different from zero.
The higher price elasticity of demand, although still quite
inelastic, can be explained, at least in part, by the large increases and
ensuing decreases in the consumption and price of wastepaper. It seems
likely that industry decisionmakers behave differently at different points in
the business cycle. Mills may be more responsive to prices that are rapidly
increasing than to prices that are falling. The fact that the elasticity
estimates presented in Table 1 vary depending on the time periods on which
the models were estimated lends some credence to this hypothesis.
Although there were major price and therefore demand changes from
late 1972 to early 1975, an analysis of the residuals from the demand
equation suggests that the estimated equation was able to account for most of
the shifts in demand that occurred during this period.
All of the variables in the inventory equation have the proper sign
as suggested by the theory outlined earlier, although only the once lagged
price ad inventory variables were significant.3 The extremely small
coefficient of adjustment (8), .05, suggests that inventories are adjusted
only a fraction each month towards the desired level. This is understandable
3. The specification of the inventory equation is discussed in detail in
"Analysis of Scrap Material Supply and Demand." Draft Interior Report
No. 2 to the EPA. Prepared by E.L.I., September 1978.
195
-------
MILLIONS OF SHORT TONS
C
~
TIME
Figure 2. Time series plot of woodpulp price and wastepaper consumption data.
-------
given the variable nature of scrap prices and the high cost of holding
wastepaper in storage. Past rates of change in consumption appear to be more
important in determining inventory adjustment plans than current consumption.
This would be the case if many of the mills that consume wastepaper do not
produce strictly to order.
Several different forms of lagged prices were used in the inventory
equation and all generated similar results. The once lagged price variable
was chosen based on maximum explanatory power of the equation. The fact that
stocks of wastepaper are adjusted in response to expected price is not
surprising in light of inventory behavior during the 1973-1974 period
described above.
THE FERROUS SCRAP MODEL
The four structural equations making up the ferrous scrap model are:
(8)
Qs
where:
t = aO + alPt + a2 X3t + a3P(t-j)
= C0 +ClPt + C2X4t
dO + Vft-l) + d2X5t + d3AX5t
Q is net receipts of ferrous scrap (all grades) in thousands
of tons
s
Qt is total receipts of ferrous scrap (net receipts plus
exports) in thousands of tons
Pt is the wholesale price index of ferrous scrap (all grades)
deflated by the wholesale price index for all
commodities (1967 = 100) and multiplied by 1000
X It is total u*s> steel production in thousands of tons
X2t iS an ^mon la§ of the wholesale price index of market pig
iron deflated by the wholesale price index for all
commodities (1967 - 100) multiplied by 1000, lagged
five to eight months
X 3t is the Federal Reserve Board Index of automobile production
(1967=100) multiplied by 10
197
-------
X is an index of total foreign steel production
p>
Q is net U.S. exports (exports - imports) of ferrous scrap
* all grades) in thousands of tons
I is beginning of month stocks of ferrous scrap (all grades)
in thousands of tons
X_ is the U.S. consumption of ferrous scrap in thousands of
tons
(t-j)
is an Almon lag on the ferrous scrap wholesale price index
lagged one to four months
The structural coefficients of equations 8-11 were estimated by
two-stage least squares and corrected for first order serial correlation.
The estimated equations were:
QD = 353.58 - 0.90P.+ 0.26X7. + 0.65X,/f -x (12)
* (-1.8) * (9.5) " (1.2) *{t~3)
.53
Q^ = 350.85 + 2.16P. + 0.67X.. - 0.04P (13)
1 (18.9) T (11.9) (.24) lt>J)
P = .77
QE = 402.49 - 0.05P.+ 0.076X4- (14)
t r-n.^^ r-^.Ti
(-0.6) (3.3)
.65
= 270.12 + 0.961,. .. - 0.02XC,. - 0.02AXC. + 0.13P,. .x (15)
(46.4) (t-1) (-0.9) bt (-0.6) 5t (2.2) (t-j)
0.1
The demand for ferrous scrap is negatively related to price although
the coefficient is only marginally significant. The elasticity of demand,
estimated at the means, is 0.34. The elasticity of scrap demand with respect
198
-------
to total production, 0.92, indicates that demand for ferrous scrap rises
almost proportionately with increases in steel production. A one ton
increase in the production of steel increases the demand for scrap by .26
tons. The lagged pig iron price has the expected positive sign but is not
significantly different from zero.
All of the variables entering the supply equation have the
appropriate sign, although lagged scrap prices are not significant. The
short-run supply elasticity at the means is 0.71. The production of
automobiles, which in the process generates prompt scrap, augments the supply
of scrap forthcoming at any given price.
Although the U.S. scrap prices are shown, in the export demand
equation, to have an insignificant effect on the foreign demand for U.S.
scrap, activity in foreign steel industry is an important determinant of
demand. An increase in total foreign steel production has a positive impact
on the quantity of scrap exported. There are several possible reasons for
the lack of an estimated price effect. (1) It may be that exports are
exogenously determined from the rest of the model. If this is the case,
scrap exports should appear as an explanatory variable in the supply
equation. (2) While foreign consumers may actually respond to the price of
U.S. scrap, the price variable used in the equation may not reflect the true
U.S. export price for scrap. The price index used in the analysis is highly
aggregated and may give more weight to prices in non-exporting cities or
grades of scrap that are not normally exported in large quantities. (3) The
equation may omit other variables that are important in a foreign scrap
consumer's decision to import U.S. scrap. The most obvious example of this
is the lack of a variable defining the costs of transporting scrap to the
importing country. It seems reasonable that the level of transportation
costs would be a significant factor in explaining demand for U.S. scrap. (4)
While current prices may not affect current export demand, an expected or
lagged price variable may play a more prominent role in determining the level
of scrap exports at any given point in time. This same argument applies to
the activity variable. More attention must be paid to the specification of
the export demand equation before firm conclusions concerning the existence
of a price effect can i»e made.
Initially it was felt that the imposition of export controls during
part of the data period may have been one reason for the insignificant price
effect in the export demand equation. Export controls on ferrous scrap were
imposed by the U.S. Department of Commerce in July 1973 in an attempt to
assure domestic consumers adequate supplies of ferrous scrap during a period
of unusually strong demand. The first form of export control required all
export shipments of over 500 tons to be -licensed, beginning with orders
accepted after July 1. Regulations for orders of 500 tons or more received
after July 1 were not issued until later. All orders of less than 500 tons
would be granted licenses. The effect of these control regulations was to
reduce the size of individual orders but appears to have little immediate
effect on the total amount of exports. In August 1973, scrap exports were
actually higher than they had been in July.
199
-------
In September, the Commerce Department acted on the problem with
unlimited small orders and stopped consideration of any small orders received
after September 11, although orders over 500 tons which had been received
prior to July 1 could still be filled. In addition, Canada and Mexico were
allowed to import up to 75,000 tons of U.S. scrap per month.
The licensing system was replaced by a quota system in the first
quarter of 1974. The quotas were based on historical export patterns and
licenses were granted based on the country to which the scrap was to be
exported. The quota was set at 2.1 million tons for each quarter until
export controls expired on December 31, 1974.
An attempt was made to estimate the impact of export controls on
export demand. A dummy variable was included in the export demand equation
that had a value of 0 for the period before and after controls, and a value
of 1 for the period August 1973 - December 1974. The coefficient of the
dummy variable was insignificant indicating that export controls may have had
little impact, perhaps because the demand for exports was already declining
by the time controls were imposed.
The inventory equation suggests that the level of stocks held by
scrap consuming mills is predominantly a positive function of stocks and a
negative function of scrap prices in the previous month. The extremely small
coefficient of adjustment ( ), 0.04, indicates that like wastepaper
inventories, actual stocks of ferrous scrap adjust slowly to desired levels.
During a one-month period stocks are adjusted only 4% of the amount of
adjustment required to reach the desired level of inventories. The absence
of a strong short-run buffer stock motivation for holding stocks, evidenced
by the insignificant consumption and change in consumption variables, may
indicate the consuming mills react to longer-run considerations. The
positive sign for the lagged price variable seems consistent with a visual
analysis of the price and stocks data plotted in Figure 3. For the most
part, the level of stocks rises with increases in the ferrous scrap price.
Two periods are exceptions to this rule. From early 1972 to late 1973,
stocks of scrap at consuming mills fell quite regularly in the face of
sharply increasing prices. Mills were depleting inventory levels to hold
down the cost of raw materials. This trend was reversed in late 1974. As
scrap prices began to fall, mills replenished their inventories. When prices
returned to more normal levels, stocks on hand continued to increase above
the historical average. Further research is required before the inventory
behavior in these two periods can be adequately specified in the model.
Table 2 reports summary statistics of elasticities for this model and
our earlier model. The fact that the current model shows the demand and
supply of ferrous scrap to be less responsive to price signals (more
inelastic) than in the earlier model may be the result of several factors.
These will be discussed in turn.
200
-------
:-
O
MILLIONS OF SHORT TONS
M.» -I
- user mills stocks of scrap A. J( .' \ ls
^v/^t^
^--
.
/ '-'
by user mills
'^
TIME
Figure 3. Time series plot of ferrous scrap data, January 1962 through October 1977,
-------
TABLE 2. COMPARISON OF ESTIMATED DEMAND
AND SUPPLY ELASTICITIES FOR FERROUS SCRAP
Estimates
Elasticity EPA* Current
Demand 0.64' 0.34
Supply:
Short- run
Medium- run
1.43
1.12
0.71
0.71**
* Environmental Law Institute.
** No significant medium term effect.
(1) Like wastepaper demand, the demand for ferrous scrap inputs
increased during the commodity boom of 1973 more than proportionately to the
demand for final steel outputs. This was due to a relative shortage of coke
for pig iron production. To compensate for the lack of pig iron, mill demand
for scrap increased. In response, the,price for scrap rose to all-time
highs, as shown in Figure 3, between the beginning of 1973 and early 1974.
Unlike wastepaper, though, the supply of scrap did not increase significantly
in response to the new higher prices. Several structural changes in the
market acted in concert to limit scrap supply and create a shortage of
ferrous scrap. 4 (a) There was a decrease in the availability of home and
prompt scrap. Prompt scrap generators returned scrap back to the originating
mills instead of selling scrap to brokers or dealers. In addition, higher
yield steel products were being produced, cutting down on home scrap
production. (b) A. scarcity of gondola cars constrained the movement of
scrap to areas with the greatest demand. (c) In anticipation of higher
prices and further shortages, mills began to build up inventories above
levels required by current consumption needs. The lower estimated elasticity
of supply from the current model can, in large part, be ascribed to the
shortage conditions described above.
4. The Commodity Shortages of 1973-1974; Case Studies. Prepared for the
National Commission on Supplies and Shortages. Washington, B.C.: U.S.
Government Printing Office, 1976.
202
-------
(2) The price controls, that were in effect from 1972-1974, may have
had some effect on the responsiveness of demand to price. Analysis of the
theoretical impact is clouded by the different price control regulations that
were enacted during this period. Generally, controls were placed on all
unprocessed industrial scrap. Small scrap dealers were exempt from price
controls while dealers with over 60 employees were not. In December 1973 a
one-time cost pass through was allowed on a dollar-for-dollar basis for both
buyers and sellers. Prices were set at the average price during the last
quarter of 1973. By early 1974, all ferrous scrap was exempted from price
controls. Our attempts to estimate the impacts of price controls on ferrous
scrap supply and demand were unsuccessful.
(3) The price indexes used here are highly aggregated, fixed weight
averages of the published market prices. These prices may not be accurate
representations of the true prices to which scrap consuming mills react. To
the extent that they do not reflect the actual market prices, both the demand
and the supply elasticities may be understated and biased.
Furthermore, the nature of the prices making up the pig iron price
index may provide some insight into the insignificance of the variable in
determining demand for scrap. Since the market price for pig iron is a
posted price, and then only accounts for a small portion of total pig iron
consumed, it may not accurately reflect the relative scarcity or abundance of
the input. The impact of the availability of pig iron on the demand for
scrap might be better captured by a variable depicting the production of pig
iron and not its price. A pig iron production index was included in the
scrap demand equation. Although it did not generate entirely satisfactory
results, pig iron production was a significant and positive determinant of
the demand for scrap.
(4) Finally, the definition of the dependent variables for the
supply and demand equations may not be specific enough to display all of the
subtle changes that occur in the ferrous scrap markets during periods of
rapidly increasing (or decreasing) prices. For example, the average grade of
purchased scrap is known to change with changes in the price of scrap.
CONCLUSIONS
This paper demonstrates that econometric models of secondary
materials markets yield results which are consistent with known institutional
relationships and with economic theory. It is important to recognize the
limitations of the econometric approach, however, before directly applying
the parameter estimated developed here to policy problems. The estimated
demand elasticities are believed to understate seriously the actual long-run
parameters of the scrap consumers. For policy purposes it may be reasonable
to employ much higher direct price elasticities of demand and infinite cross
elasticities.
5. This is a highly simplified view of the complex system of price control
policies and regulations enacted during 1972-1974. A more detailed
outline of the price regulatory events can be found in The Commodity
Shortages of 1973-1974; Case Studies.
203
-------
On the supply side, one important adjustment must be made. If one is
interested in the impact of product charges or other similar forms of market
intervention on the recovery of post-consumer secondary materials, one must
adjust the elasticity estimates. This adjustment is necessitated by the fact
On the supply side, one important adjustment must be made. If one is
interested in the impact of product charges or other similar forms of market
intervention on the recovery of post-consumer secondary materials, one must
adjust the elasticity estimates. This adjustment is necessitated by the fact
that the market models are estimated for all scrap (prompt industrial scrap
plus post-consumer scrap) as dictated by the manner in which the data are
reported. If one assumes that prompt scrap recovery is unaffected by price,
the fact that prompt ferrous scrap comprises approximately 50% of total
supply would dictate that the reported supply elasticity be doubled. For
wastepaper, approximately 40% is prompt, indicating the wastepaper supply
elasticity should be increased to 1.67 times its reported value.
204
-------
SECTION 6
PRICING AND SOLID WASTE MANAGEMENT
"The Seattle Solid Waste Management Experiment: User Charges and Separate
Collection of Recyclables" (Bernard H. Booms).
Z05
-------
THE SEATTLE SOLID WASTE MANAGEMENT EXPERIMENT:
USER CHARGES AND SEPARATE COLLECTION OF RECYCLABLES
by
BERNARD H. BOOMS
ACKNOWLEDGEMENT
The author gratefully acknowledges the comments and suggestions of
Dr. Haynes Goddard, EPA, Cincinnati, Ohio; and Mr. Ed Steyh, Manager of the
Seattle Solid Waste Experiment, City of Seattle.
THE SERVICE AREA - THE SETTING
This section provides a general profile of the Seattle area and
describes the current residential solid waste management system operated by
the City of Seattle. The purpose of this descriptive material is to
familiarize the reader with the setting for the Seattle experiment and to
highlight situational factors which could conceivably influence the
development and outcomes of the experiment.
Description of the Area
Population and Geography
The City of Seattle is located in Washington State on the shores of
Puget Sound. The current 1978 population is estimated to be 490,000. The
boundaries of the city encompass about 88.5 square land miles (91.5 sq. miles
of area including lakes and waterways). The area is bounded on two sides by
water, with Puget Sound to the west and Lake Washington to the east. An
aerial view of the land mass reveals an hourglass shape. (Perhaps because of
her "figure" Seattle is called the Queen City!) The city is thus
geographically divided into north and south sections with the downtown area
at the narrowest part of the hourglass being the usually accepted dividing
line between the north and south sections (See Map 1).
Topography
Seattle was once a city of seven hills, but the marvels of
engineering reduced the number to six hills and one regarded bump!
Nevertheless, the general terrain is still far from flat with many areas of
rolling hills and numerous areas with steep hills. Obviously, the hill
206
-------
N
--
-------
nature of the area makes the operation of collection vehicles and the
collection of solid waste more difficult than in flat areas. This is
especially true during the winter months when the ground is occasionally snow
or ice covered.
Climate
The climate is generally mild. Summer daytime temperatures usually
reach the upper 60's to middle 70's. Summer days with temperatures over 80
are unusual. Winter daytime temperatures range in the high 30's to 40's.
Below freezing readings occur only about 15 days in an average year.
Seattle has a reputation for being rainy. This reputation is based
less on absolute amounts (average 36 inches/year) and more on consistency.
The Seattle area is often cloud-covered, and rain, when it falls, occurs at a
light rate for long periods.
Winds are seldom strong. Normal wind speeds are 3 to 7 miles per
hour. Winds are rarely strong enough to blow debris about during the
collection process and cause collection difficulties.
Population and Housing Characteristics
Table 1 depicts population and housing statistics for Seattle and
offers comparative statistics in some instances with selected cities over
250,000 population surveyed by R. L. Polk and Company. The comparative
statistics were calculated by Polk based on 1975 data.
PopulationSeattle has a population mix quite different in several
important dimensions from most other cities of similar size. The age
distribution of Seattle's population tends to be bimodal to the younger and
older ends of the scale. Seattle has become predominately a city of young,
single, often highly educated, high income persons, and of retired, older,
lower income persons. The degree of Seattle's singleness obviously
influences the proportion of households with children, with a higher
proportion being childless in comparison to other cities. It appears that
national trends of urban population composition are exaggerated in Seattle.
The high proportion of young, single or young married persons with high
education and income levels in Seattle suggests a great potential interest in
and willingness to participate in social experiments. The attitudes found in
Seattle can be a strong determinant in the observed outcomes of the
experiment, and any attempt to transfer the findings from Seattle to other
areas should be done with caution.
HousingApproximately 65% of the population live in single family
housing units. There is a trend for more single people to occupy single
family housing. Solid waste collection for the 35% of the population living
in multiple family units is most often carried on using large collector
containers with resultant economy of collection. But approximately 5% of the
structures in the city have 5 or more housing units, so that large scale
collector systems are not extensively used. Despite a high proportion of
single persons in the population, Seattle continues to retain a relatively
low population density.
208
-------
TABLE I. SUMMARY OF POPULATION AND HOUSING DATA IN SEATTLE AND
OTHER CITIES OVER 250, 000
Other cities over
Seattle 250, 000, 1975 data
Population
% White
% Non-White
% Black
% Other
Households
Average per sons /HH
% One person HH
490,000
85%
15%
9%
6%
203,237
2.4
35% Average 26%
% HH with children
% Retired heads of HH
Housing units
% Single family units
% Multiple family units
24%
27%
216.058
65%
35%
Range 20% -- 31%
Average 38%
Range 30% -- 40%
Average 20%
Range 16% 26%
Source: 1977 Polk Data and City of Seattle.
209
-------
Collection and Disposal
Solid waste operations for the City of Seattle are run as a utility.
Utility status means that the solid waste operation must "stand on its own
bottom," i.e., operate without subsidy from the City's general fund. Fees
are charged for the service and these fees must cover all costs of operation.
For solid waste management purposes the city is divided into two
service areas. These areas coincide with the north and south geographic
sections of the city described earlier.
The city operates a transfer station in each service area. These
transfer stations are the points where the two contracted commercial haulers
deliver their collections. The city then transports the solid waste from the
transfer station to the city-operated disposal sites. The transfer stations
are also open for residents (and for non-residents at a fee) for individual
dropoff of solid waste.
The Solid Waste Utility controls and manages only residential solid
waste. Disposal of commercial solid waste is left to the private sector,
without city involvement. One major commercial firm dominates the commercial
collection and disposal business. This firm operates a marine transfer
station in the downtown area. This station is used to transfer commercial
solid waste to a barge. The waste is then hauled out of the county area and
disposed of on a private landfill site north of Seattle.
Map 2 shows the boundaries of the service areas and the location of
the transfer stations.
The city has recently passed an ordinance that could require that all
material be disposed of at city-operated landfills and run through the
city-operated transfer stations.
Charges for the collection and disposal of residential solid waste
are collected on a flat fee basis by the city. Charges per single family
household are currently $5.20 a month for an allowable limit of 4
conventional garbage cans plus two ordinance units (boxes or bags of
non-putrescible substances) per can (total allowable ordinance units, 8).
Collection
The collection service is operated on a contract basis by the city
with the two service areas being separate bidding areas. Award rules for
service contracts prohibit any single contractor from being awarded the
contract for both areas.
Collection frequency is once a week. Backyard pick-up is provided.
Though curbside placement is purely voluntary, it is estimated that from 15%
to 20% of the households do set their cans at curbside on collection day.
Setting cans at curbside is one way residents can obtain special favors from
collection crews, like accepting occasional oversized or overweight items.
210
-------
Map 2
NORTH
TRANSFER
STATION
SOUTH
TRANSFER
STATION
SERVICE AREA BOUNDARIES:
'/////////.-HQi\t\\ South
O 3 -Marine (Some Solid
Waste from East ol City)
Source: City of Seattle, Departments of Engineering and
Lighting, Seattle's Solid Waste ... An Untapped
Resource, May 1974,
211
-------
Approximately 30% of households are on alleys; in these areas,
collection is via the alley. Alley collection usually involves shorter
distances for the collection crews to walk in picking up the solid waste. A
recent survey showed that 52% of the housing units locate their refuse
containers only in the backyard, while the remaining 48% utilized a
combination of backyard, curb or alley locations for their containers.
The current distribution of can usage per household within the City
is estimated to be:
Zero cans 1%
1 can 20%
2 cans 60%
3 cans 15%
4 cans 4%
A recent survey of container usage and volume of refuse indicated that the
average number of units per stop was 1.9 cans and 0.1 ordinance units. The
average weight of material per stop was 53 Ibs., ranging from 40 to 69 Ibs.
Disposal
Two disposal sites are currently leased and operated by the Solid
Waste Utility. The two landfill sites are both outside of the city limits.
The close-in Midway site is nearly up to grade and is limited to
non-putrescible wastes. The Kent-Highlands site is estimated to have only
three years of capacity remaining at current rates of solid waste generation.
The city pays a disposal royalty of 65 cents/ton for the use of the
landfills.
Waste Flow
Approximately 550,000 tons of solid waste are collected each year in
Seattle. This flow is composed of approximately 218,200 tons/year of
residential waste and 331,800 tons/year of commercial waste.
A study of percent composition by weight of solid waste in the area
that includes Seattle estimated the following percentage breakdown by waste
type.
212
-------
TABLE 2. PERCENTAGE COMPOSITION BY WEIGHT OF COLLECTED SOLID WASTE
IN COUNTY AREA THAT INCLUDES SEATTLE
Waste type
Paper
Newspaper
Cardboard
Mixed Paper
1971 data
residential
41%
15%
6%
20%
1973 data
industrial/commercial
22%
0
7%
15%
Plastic 5% 1%
Ferrous Metals 6% 3%
Aluminum 1% 1%
Glass 7% 15%
Garbage (putrescible) 13% 5%
Other combustibles 13% 34%
Wood 1% 19%
Garden 10% 0
Other 2% 15%
Inerts 14% 19%
Source: Market Analyses of Recovered Materials and Energy from Solid Waste,
CH2M Hill Study, 1977.
A follow-up study using these percentage estimates attempted to
determine the proportions of paper, ferrous metal, aluminum and glass
currently not entering the collected solid waste stream and being recycled
instead. Table 3 illustrates the estimated level of recycling activity prior
to the beginning of the Seattle experiment and compares these levels with
"maximum possible" recovery rates published by EPA.1
1. Maximum possible effort estimates are based on use of all presently
accepted technologies and favorable market prices for recyclables.
213
-------
TABLE 3. ESTIMATES OF RECYCLED MATERIAL IN SEATTLE AND EPA MAXIMUM EFFORT ESTIMATES
ro
Waste type (tons)
Paper
Newsprint
Cardboard
Mixed paper
Plastic
Ferrous metals
Aluminum
Glass
Garbage
Other combustibles
Wood
Garden
Other
Inerts
Totals
(1) total recycled/
collected material
225,791
53,922
60,268
111,601
14,228
30,973
9,820
88,929
44,956
141,178
65,224
21,820
54,134
93,590
649, 465
(2) recycled (3) collected
material material
63,333 162,458
21,192 32,730
23,950 36,318
18,191 93,410
14,228
7,927 23,046
4,320 5,500
23,885 65,044
44,956
141,178
65,224
21,820
54, 1 34
93,590
99,465 550,000
(4) percentage (5) EPA/s percent
re cycled /total "maximum effort"
28 46
39.3 55
39.7 55
16.3 35
25.5 63
43.9 46
26.8 50
15.3
Source: City of Seattle.
-------
It is clear that Seattle is an active recycling area. Approximately
15 percent of all material is being recycled. This compares with a recycling
rate of 7 percent on the national level.
Twenty-eight percent of the paper, 44 percent of the aluminum, 27
percent of the glass and 26 percent of the ferrous metal are being recycled.
This high recycling rate may be explained by the attitudes and
characteristics of the population which were described earlier. A large
number of groups and individuals are involved in recycling operations
throughout the city. One estimate places the number of recycling operations
(church and school paper drives, etc.) around 120. At least 18 recycling
locations pay cash for one or more recyclable materials.
Figures 1 and 2 give an indication of the monthly and daily averages
of solid waste material flow from the Seattle area.
BACKGROUND
In this section the factors leading up to the experiment will be
outlined, and the nature of the citizen and governmental initiative
described.
Factors Leading to the Experiment
Citizen Concerns-
During 1976 the City of Seattle renegotiated the refuse collection
contracts. As a part of this process, the City Council held public hearings.
Citizen input at these hearings revealed that some individuals were not
satisfied with a flat fee charge. These vocal individuals argued that the
flat fee was unequitable because the fee was fixed regardless of the amount
of waste generated (up to four containers). Thus the driving factor behind
the variable rate portion of the experiment seems to be a concern with equity.
Landfill Capacity
For a number of years the City has realized that at current rates of
solid waste generation it will soon exhaust its present landfill capacity.
This concern has led to the study of a number of alternative disposal
methods, including using the waste for fuel or for making ammonia. Source
separation and recycling have also been discussed as ways of extending the
life of the landfill. Landfill capacity and imminent decision on locating
another landfill or using other technology were also factors leading to the
experiment.
Thrust of the Experiment
Several broad areas of concern could be said to be factors behind the
experiment.
215
-------
January
February
March
April
May
June
July
August
September
October
November
December
Total (TPY)
Monthly Avg.
T/Mo.
41,818
41, 150
45,770
48,086
49,908
49,217
48,621
47,549
46,036
45,256
43,057
42,568
549,036
45,753
Average
Av. T/Day
1,349
1,470
1,476
1,603
1,610
1,641
1,568
1, 534
1,535
1,460
1,435
1, 373
NOTE: Data include the estimated Marine Transfer
Facility quantities of 220, 800 T//Yr. , or
18,400 T./Mo. , assumed constant year-to-
year and month-to-month.
50
o 40
o
o
£ 30
c
o
"? 20
10
0
H
50
40
30
20
10
0
JFMAMJJASOND
Source: City of Seattle
Figure 10 Five-year annual average of solid waste.
216
-------
Out
1
2
3
4
s
6
7
3
9
10
11
12
13
14
IS
16
17
18
19
JO
21
22
23
24
25
26
27
28
29
30
31
D.y
Su
Mo
Tu
W«
Th
Fr
Si
Sv
Mo
Tu
W«
Th
fr
Sa
Su
Mo
Tu
W«
Th
Fr
Si
Su
Mo
Tu
We
Th
Fr
Si
Su
Mo
Tu
NTS STS
171
1.116
1.094
99O
(.021
951
157
122
1.099
1.120
1.075
1.074
964
279
116
1.122
1.072
1.040
1.003
956
229
212
1.123
1.105
1.04fi
993
947
220
178
I.10S
1.149
STS
Cm
110
132
140
96
102
89
118
133
1O1
106
109
108
83
110
62
89
101
106
77
97
179
112
114
104
92
86
82
123
S3
107
115
MTF
0
767
767
767
767
767
383
0
767
767
767
767
767
383
0
767
767
767
767
767
383
0
767
767
767
767
767
383
0
767
765
Tot.l
281
2.015
2,00 1
1.853
1.89O
1.807
658
2S4
1567
1.993
1.951
1349
1.8)4
722
178
1.978
1.940
1513
1.847
1,815
791
324
2.004
1.976
1.904
1.846
1.796
726
266
1.979
2.031
X
19
134
133
124
126
121 ,
44
17
131
133
130
130
121
48
12
132
129
128
123
121
53
22
134
132
127
123
120
48
18
132
135
Oiily Avaraga 1.499 tpd
* Based on an estimated rate of receipt of 18,400
tons per month, received at an assumed equal
daily rate 5-1/2 days per week.
Source: City of Seattle.
Panel A. Solid waste received at North (NTS) and
South Transfer Stations (STS), and the
Marine Transfer Facility (MTF), in tons
per day.
SMTWTFSSMTWTFSSMTWTFSSMTWTFSSMT
AUGUST 1976
Source: City of Seattle.
Panel B. Seattle solid waste handled daily, August 1976
Figure 2
217
-------
Equity
It has already been suggested that equity among households seems to
be the most often stated factor motivating the experiment. This concern is
readily understandable by the average citizen and has broad appeal.
Ironically, if the service level of the collection system, i.e., the
frequency and location of pick-up, account for a large proportion of costs
and these costs are not appreciably influenced by the amount of material
collected, then a flat fee charge may be more equitable than a variable
charge based on the quantity of refuse! (It should be noted that the fee
structure can vary according to service level. Some cities are currently
using such a rate structure.)
Efficiency
Concerns relating to the allocation of resources focus on social
costs and benefits being equal to individual costs and benefits and the true
costs and benefits being borne by the individual making the resource
allocation decisions. In this way the consumer can correctly compare the
consequences of alternative resource allocation decisions. The variable rate
is one way of trying to ensure that the costs of solid waste generation are
borne by the generator, thus facilitating informed decision making by the
consumer in trading off between more or less waste generation activities.
However, little mention is made about efficiency in discussions of the
experiment.
Implementation--
Another area of concern that could relate to the experiment is
implementation questions. The experiment is viewed, in part, as a method for
identifying implementation problems and for trying alternative implementation
schemes. The experiment is an opportunity to learn on a trial basis
how to operate alternative solid waste systems. City government is
especially interested in discovering problems without having to risk the
possibility of a big failure (see below). So information potentially useful
to implementation concerns is also a partial basis for the experiment.
Initiative in .the Experiment
It is important to realize that the initiative for the experiment
started with the citizens and was picked up by the City Council. The
Council, not the City Utility (city department) initiated the call for an
experiment. The city department has been reactive to the idea of an
experiment, rather than proactive.
The environment in local government may help to explain why the
initiative lies where it does. As a result, error is terror for city
bureaucrats. The constituents are close at hand to complain if things go
wrong and there are often individuals willing to turn mistakes into political
advantage. Most often the facts of life for bureaucrats in this setting are
such that if you do the same old thing and something goes wrong, it is an act
of God, but, if you try something new and It goes wrong, it is your fault!
Therefore, organizational resistance.to experiment and change is increased.
218
-------
It is important to understand the environment in which the experiment
is being conducted when considering the experiment and thinking about
suggestions for modification to the experiment.
THE EXPERIMENT
In this section the objectives and design of the Seattle Solid Waste
Experiment are discussed. The experiment was designed under contract with a
consulting engineering firm. The program is currently being managed by the
City's Solid Waste Utility. The actual running of the experiment and the
evaluation have been contracted out by the City to private firms. These
firms are responsible for the regular collection of solid waste under a
variable rate structure, a recycling program based on home collection, and
evaluation as well as design of the experiment.
The experiment consists essentially of three test treatments in the
mangement of solid waste. The treatments are:
Variable Can Rate
In selected areas of the City, residents are charged for refuse
collection and disposal on the basis of the number of refuse containers they
use. Household participation in the test areas is on a voluntary basis.
Source Separation with Separate Truck Collection - Recycling
In selected areas, residents have the opportunity to voluntarily
separate recyclables from their refuse for separate collection on a schedule
different from the regular solid waste pick-up schedule. The resident is
asked to separate waste into three recyclable categories: glass, metals, and
newspaper.
Variable Rate Plus Source Separation
In selected areas, both of the above described treatments are being
tried simultaneously. The variable rate applies only to containers used for
refuse, and not to containers used for the recyclable materials. Again,
participation in both treatments is voluntary, and the resident can volunteer
for one or both of the elements of the treatment.
An evaluation of these various treatments is being conducted as a
part of the Seattle Solid Waste Experiment.
The Objectives of the Experiment
There are several written statements outlining the objectives or
goals of the experiment. The main objectives are as follows:
To test the feasibility of a variable can rate structure and
determine its impact on refuse generation and collection.
To test the viability and economic feasibility of separate
collection of household sorted recyclable materials.
219
-------
To test the impact of separate collection of household sorted
recyclables on refuse generation and collection.
To test the impact of a variable can rate structure on household
participation rates for separate collection of sorted recyclables.
The research design developed for the City under contract was
prepared to meet these stated objectives. The research design contains
little discussion or analytical framework relating to possible hypothesized
impacts (both wanted and unwanted) of the various experimental treatments.
Without the benefit of such a model or analysis it is difficult to determine
a methodology for evaluating the experiment and, in fact, to design the
experiment and identify data collection needs.
The Research Design
The research design was constrained by several factors. First, was
a concern and a requirement that the experiment be as self-supporting as
possible, and that in no case should the experiment result in more than
$250,000 of utility funds being dedicated to the project, or result in more
than $55,000 of revenue loss to the utility. In addition, the experiment
had to meet all tests of legality.
The test of legality raised the most difficult complication in
research design. The City Attorney rendered an opinion that variable rates
exceeding existing standard rates would be illegal under State law requiring
that all utility customers, of the same class be charged at the same rate.
Thus, the possibilities for designing a rate structure that would reward
residents who generate little waste and penalize those who generate large
amounts of waste, while at the same time generating the revenue needed to
cover the total costs of the service, were greatly limited. As a result, a
proposed variable rate that had considerable differentiation between numbers
of containers had to be abandoned. The variable rate is discussed in detail
below.
Three Treatments
Three treatments are involved in the experiment: variable rates,
source separation and .separate collection of recyclables, and a combination
of these two treatments. Participation in all cases is voluntary. A part of
the experiment is to determine the response of the residents to the various
treatments, given various promotional and educational efforts.
Treatment Areas
Thirty areas in the City were identified for inclusion in the
experiment. In addition, five control areas similar to the test areas were
designated for inclusion in the study so as to provide baseline data which
allow for measurement of seasonal and weather factors that influence solid
waste collection.
The thirty treatment areas represent regular collection routes or
combinations of regular collection routes.
220
-------
Each of the three treatments is tried in 10 of the 30 test areas.
Each set of 10 test areas was picked so as to be representative of the City
as a whole. The areas were selected on the basis of income, housing density,
topography, and location within the City with income levels being the
dominant selection factor. For each treatment, there are two low income
areas, two high income areas, and six middle income areas.
The treatment or test areas are not contiguous, and are spread
throughout the City.
Test areas 1 through 10 are the variable rate areas; areas 11 through
20 receive a combination treatment of variables rates and recycling, and
areas 21 through 30 are the recycling only areas (Maps detailing these areas
are available from the City of Seattle.).
Variable Rate
The collection and disposal of solid waste is being conducted by the
regular contractors. The level of service is identified by special stickers
on the containers. Cans without stickers are not collected. This system was
chosen over route books because it was believed to be more efficient, and
also because the high proportion of alley pick-up in the City limits the
suitability of the address-based route book system.
The variable rate treatment areas are being monitored by five City
employees hired throught the CETA program. (The CETA contribution to the
experiment amounts to approximately $75,000.) The monitors follow the
collection crews, recording numbers of containers collected and the time
required for the various components of the collection process. The monitors
also make general observations about the collection process and the
experiment. The monitors are responsible for attaching the stickers to the
containers and for making sure the sticker system is kept operational.
The billing for the variable rate areas is conducted through the
regular City billing system.
Early in the design phase of the experiment there was much discussion
concerning how to determine the "appropriate" variable rate structure.
Ideally, the rates should reflect the marginal costs of collection and
disposal. In the absence of any data on marginal costs (on a container
basis) the discussion focused on the requirement to generate revenues needed
to cover the cost of the experiment and on guesses as to what rate spread
would provide the desired incentives to residents. The following rate
structure was proposed.
221
-------
Number of Containers Rate
Zero cans $1.00
1 can 4.60
2 cans 5. 60
3 cans 7. 00
4 cans 8. 50
When the City Attorney was asked to review this structure he
suggested that it would be illegal on the basis that several of the proposed
rates exceeded existing rates for residents outside of the test area
receiving similar service.
Two alternate rate structures were considered (see below), but these
were judged inappropriate either because they generated insufficient revenue
or because they did not appear to provide a large enough incentive for
residents to modify their behavior.
Alternative Rate Structures Considered
No. of Cans Rate No. of Cans Rate
Zero $1.00 Zero $1.00
1 2.80 I 4.00
2 4.00 2 4.95
3 or 4 5.20 3 or 4 5.20
Rate Structure Finally Adopted
No. of Cans Rate
Zero $1.00
1 4.00
2, 3 or 4 5.20
This rate was judged to minimize revenue loss and to provide a
"reasonable" incentive to switch from two can to one can service. (It waa
222
-------
felt few residents would be willing to move to a zero can service level.) It
was noted in the discussion that thir rate structure has the drawback of not
providing a rate break between the 3 and 4 can users. It remains to be seen
if the experimental rate structure provides incentives for the big waste
generators to reduce waste output.
It is important to make clear the fact that the one can service does
not include collection of ordinance units. What doesn't fit in the one can
is not collected, and the resident must dispose of any extra himself or store
it for collection at the time of the next pick-up. No provision is made for
occasional increases in the amount of solid waste generated, e.g., from yard
cuttings or special events such as the holiday seasons. This fact may have
cut down on the number of residents willing to adopt the one can service
level. Note also that in treatment areas where both the variable rate and
recycling tests are in operation, containers for the recyclable material are
not counted in the variable rate.
Generally, the discussion about what rate structure to adopt was
based on guesses, hunches, common sense, with the greatest emphasis being on
revenue needs. Words like "ideal," "appropriate," and "reasonable" were used
to describe various alternative rates, without the words themselves being
given precise meaning. There is a need for cost information on which to base
such rate making decisions.
Source Separation and Separate Collection of Recyclables
The sorting and recycling part of the experiment is being conducted
under contract by a private recycler Seattle Recycling, Inc. The research
design specified certain requirements that the recycling contractor had to
satisfy in his operation of the experiment. These requirements seem to be
aimed at ensuring that the technology and procedures used in the recycling
process would represent efficient equipment and processes for large scale
operations. In this way experimental operating experience would be
indicative of operating results that might be obtained in a citywide program.
The aim of the experiment is to simulate a citywide program. To further this
end publicity requirements, processing requirements and requirements for
the marketing of collected recyclables are specified in the research design.
To assist the household in the storing of source separated material,
each participating household is provided with two cloth mesh bags (potato
bags) for holding metal and glass and one plastic bag for newspapers at
the time of each collection. The bags are cleaned and recycled back to the
households.
Combination of Variable Rate and Recycling
The combination of the two treatments is carried on in 10 test areas.
The contractors each conduct their separate portions of the experiment. No
modifications to the two treatments or special arrangements are used in the
combination test areas.
223
-------
Current Evaluation Design
Considerable mention is made in the research design of the data to be
collected. The data monitoring is summarized in Figure 3, taken directly
from the Final Research Design.
The current evaluation effort appears to be heavily focused on the
"time and motion" aspects of the collection and recycling process.
Experiment Modification
The managers of the experiment have indicated from the conception of
the experiment that the program is and will continue to be open to
modification. Thus, the possibility exists to make additions or changes in
both the research design and the evaluation procedures. As early results
become available, modifications will begin to be considered.
SOME EARLY RESULTS AND OBSERVATIONS
Some results are available based on the first month of operation of
the Seattle experiment.
Response and Participation Rates
Figure 4 and Tables 4 through 7 summarize the response rates and
actual participation rates for the various treatments and test areas.
Variable Rate
The participation in the variable (one can rate) rate test averaged
12%. There was no difference in the participation rate between those areas
wit and without separate collection of recyclables. This finding is contrary
to the expected outcome. One might assume that collection of recyclables
would make movement to the one can level of service feasible for most
households.
The 11% participation rate was obtained using only a mail
solicitation. In addition, media coverage described the variable rate
program.
Approximately 30 households chose the zero rate. The households have
been given composting information and are being monitored by the city health
department.
Income levels did not seem to systematically influence the observed
participation rates among the various test areas.
Recycling1
A total of 58% of the households responded favorably to being
involved in the recycling experiment. A higher proportion of households in
the combination treatment areas (61%) expressed a willingness to be involved
than in the recycling-only treatment areas (55%).
224
-------
Variable Rate Program
Monitoring In the variable rate area! (and control areai a< well) will Include recording of the following Informa-
tion Lolh prior to implementation and once every Iwo months thereafter:
Frequency distribution! for the number of refule container! and ordinance units per atop:
Average quantity of refute per atop;
e Frequency Dlmrlbutlon for the collection time per ttop;
e Average Travel time between stops;
Total stops serviced per day;
Total ton> collected per crtw per day;
Total lime to service pilot and control areas;
. MUcellanemn travel limes and distancea (yard and route, route to landfill, landfill to yard, etc.):
Number of trip* lo landfill;
Unavoidable delay! and breakdown!;
Scheduled break*.
Source Separation and Separate Truck Collection
Monitoring In the pilot area! performing source ieparation of reeyclablea la more exteniive than in the variable
rale only areas.
Prior to implementation, monitoring'tasks are identical to those summarised In the laat section on the variable
can rate teat. Once lource aeparatlon beglne. It will be necetaary to monitor not only ongoing reCuie collection activities but
alao the source separate collection ae well.
Below U a Hit of the data required for (valuation of the tource separate collection of glass, metal, and newa-
pnper.
Addresses of homea participating by aeltlng out recyclable! (percent participation);
e Identification of Tecyclalilea let our (newspaper, glass, metal);
e Average quantity of reeyclablea by type per atop;
e Frequency distribution of time per atop for recyclable* collection;
Average travel time between atopt:
a Total time for separate truck to service pilot area;
Miscellaneous; time* and distances (yard to route, rout* to yeard, etc.);
Unavoidable delays and breakdowns;
Scheduled breaka.
With this data, the source separate collection method can be evaluated for performance and cost. A sample form
for collection of this data follows. Naturally, the control areaa will not be involved nor monitored far source separation. ,
At the yard area, monitoring will again be performed to Identify the manpower required and time associated with
unloading, processing and storage of the (lass, metals, and newspaper collected. This data will allow cost trade-offs to be
nude, and other sorting/storage techniques to be evaluated.
Monitoring will be relatively simple consisting basically of recording times and labor requirements for the various
processing and sorting operations previously described. Typical activities lo b* timed Include]
, e Unhook trailer;
a Paper unload from pick-up to roll-off bin;
a Position and dump metal bin, overaee convey, magnet, crushing system;
Position, unload glass bin.
« Sort glass.
Variable Can Bats Plus Source Separation
Tlie monitoring requirements for the combined variable can and source separation pilot program are virtually
Identical lo those for source separation above. Afstn, pilot areas and control areas will be selected.
Source: Final Research Desljm Seattle Variable Rale Source Separation Pilot Programs, City of Seattle. 1978.
Figure 3. Monitoring Design
225
-------
rtKltHI 1UU
90
80
KEY:
oorbell 70
hone
rn
ar. Rate 50
Irst
ol lect ion nn
30
20
10
TEST AREAS
TREATMENT
DATA
Doorbell
Mai 1
Variable rate
Total HHH
SE£
1st collection
.
1
2
H
§
3
M
4
_
5
M
6
H
7
L
_
o
0
9
L
10
VARIABLE RATE ONLY
.
83
786
M
.
62
H
.
48
328
M
.
55
469
H
71
539
M
.
45
451
H
.
30
361
L
.
17
203
M
_
73
522
L
65
457
M
%
m
K
WM
Wk
,;-'-'::
IP
"'.:*.
mfc^?
p^^^p^
P^
m
:::::
^*-
m
w*
f -^jiiiiii ;;: f*"*1 '"'- Uj
! iW-x^tiXWfcxWvlxffiwliiiii.J '
11
II
12
VAR
101
160
66
448
H
1S4
118
l?2
61
643
M
184
x.WS
13
^-i-:'
M
15
H
16
17
L
^
iff
wn
' :
^
//"//
::-:::-:
18
19
'"""> :;:$: SxvfcSJSr" '' 1 K^*S?%
20
L
ABLE RATE AND RECYCLING
81
ns
1 71
64
480
M
170
175
121
1 22
71
713
H
179
95
S4
445
H
142
77
62
46
54
291
M
63
92
85
72
68
475
L
67
'!'
76
23
327
M
98
138
112
82
70
H
133
146
2(|
47
409
L
90
21
22
H
23
H
25
26
L
RECYCLING
122
132
4R
^36
M
146
81
64
399
H
SI
94
40
346
H
122
109
20
423
M
137 il34 124J123
109
168
92
658
H
220
iSI
100
14
537
L
27
28
mm
29
mst
30
L
ni.'LY
119
127
19
595
H
75 125
135
136
31
546
M
L56
125
122
1\
564
75
114
85
17
522
L
63
Source: Seattle Recycling, Inc. and City of Seattle.
Figure 4. Percent response Seattle experiment.
-------
TABLE 4. PARTICIPATION SUMMARY VARIABLE RATE TREATMENT
Area*
1 - 10
11 -20
Totml
Treatment
Variable only
Combination
543
572
1115
#HH total on map
4605
4827
9432
%
12%
12%
12%
Source: City of Seattle
TABLE 5. RESPONSE SUMMARY RECYCLING TREATMENT
Area*
11 - 20
21 - 30
Total
Treatment
Combo
Recycle only
Mail
768
368
1136
Phone
1030
1164
2194
Doorbell
1138
1254
2392
Total f HH
responding
2936
2786
5722
Total # HH
on mapt
4827
5076
9903
%
response
61%
. 55%
58%
Source: Seattle Recycling, Inc.
TABLE 6.. FOLLOW-THROUGH SUMMARY (1ST COLLECTION) RECYCLING TREATMENT
Area*
11 - 10
21 - iO
Total
Treatment
Combo
Recycle only
Total HH
collected
1280
1272
2552
Total # HH
responding
2936
2786
5722
%
follow-through
44%
46%
45%
Source: Seattle Recycling, Inc.
TABLE 7. ACTUAL PARTICIPATION SUMMARY (1ST COLLECTION) RECYCLING TREATMENT
Areas
11 - 20
21 - 30
Total
Treatment
Combo
Recycle only
Total HH
collected
1280
1272
2552
Total I HH
on mapa
4827
5076
9903
%
participation
27%
Z5%
26%
Source: Seattle Recycling, Inc.
227
-------
The follow-through percentage, i.e., the proportion o± those
households who said they would participate and actually did, overall was
45%. In other words, over half of the households who said they would
participate did not put out recyclable materials during the first collection
period. Time will tell if this "drop out" rate reflects actual rejection of
the program or just forgetfulness in putting out materials for collection.
The recycle-only treatment areas had a higher follow-through rate
(46%) than the combination treatment areas (44%), thus offsetting the greater
initial response rates for the combination treatment areas.
Actual participation rates for all recycling areas averaged out to
26%. The combination treatment areas averaged a 27% participation rate,
whereas the recycle-only treatment areas had a 25% response rate. How much
of the 2% point difference can be attributed to incentives created by the
existence of a one can rate is difficult to judge. In any case, 2% points do
not seem like a very large difference between the two treatments.
In the case of recycling treatments, the income levels in the test
areas did seem to have a systematic influence on participation. High income
areas tended to have higher participation rates and lower income areas lower
rates. (This finding parallels results from previous studies.) The highest
participation rate for an area was 46% (test area 22, a high income area);
the lowest rate was 12% (test area 30, a low income area).
Material Collected and Revenues
Trash
Information on the amount of non-recycled material is being recorded,
but is not yet analyzed.
Recycled Material
The data on the amount of recycled material collected are presented
in Table 8.
Revenues--
Information on materials collected, market prices for the materials
and the resultant revenues are given in Table 9.
Other Data
Time and motion type information and costs data are being recorded,
but these statistics are not yet available.
Some Observations
Several implementation difficulties have been encountered during the
early operations of the experiment.
228
-------
TABLE 8. AMOUNT OF RECYCLABLE MATERIAL COLLECTED (1ST COLLECTION) IN POUNDS
(June 19. 1978 -- July 14, 1978)
toute
'll
12
13
14
15
16
17
18
19
20
'21
22
23
24
25
26
27
28
29
,30
Metal
677
885
861
929
600
353
255
250
640
374
623
805
458
567
1,521
310
620
585
260
280
Giles
3,494
4,065
3,042
3,409
3.359
1,200
1,345
1,650
2,480
1,889
2,310
3,816
2,880
1,826
5,367
1,250
1,760
3,060
1,422
1,325
Newspaper
4,503
3,864
4, 139
3,926
3,332
1,068
581
2,380
2,010
2.877
3, 100
6,799
4,180
3,448
6, 289
2,036
4,391
3.610
1,950
1,006
Total
8,674
8,804
8,042
8,264
6,291
2,621
2. 181
4,280
5,130
5,140
6,033
11,420
7,518
5,841
13,177
3,596
6.771
7,255
3,632
2,611
Percent
participation Average
34%
29%
35%
25%
32%
22%
14%
30%
22%
22%
28%
46%
36%
29%
33%
14%
21%
28%
13%
10%
1
N
£
Source: Seattle Recycling, Inc.
TABLE 9. AMOUNT OF MATERIAL COLLECTED, MARKET PRICES AND REVENUES
(1ST COLLECTION) (June 19, 1978 -- July 14, 1978)
Material
Paper
Glass (cullet)
Tin/steel
Aluminum
Beer bottles
Amount Collected in Tons Unless Indicated
Market Amount collected
$36. 00 /ton
20. 00/ton
60. 00/long ton
.22/lb.
, 60/caae
33
21
4 long tons
2331 Ibs.
496 cases
Total
Revenues
$ 1,188
420
240
513
298
$ 2,651
Source: Seattle Recycling, Inc.
229
-------
Variable Rate
The main problem encountered with implementing the variable rate has
been the use of stickers for differentiating service levels. Trouble first
arose when residents objected to the City monitors placing the stickers on
the cans.
Several phone calls were received from residents reporting strangers
walking around their yards and garbage cans. Some householders have removed
the stickers.
In addition, the different colored stickers have faded so that it is
difficult to tell the colors apart. The problem is further complicated by
the fact that the stickers do not adhere well to some cans. Some effort will
have to be made to improve sticker design and technology.
Recycling
The recycled materials are collected at curbside. A significant
number of scavenging incidents have been reported. Efforts are being made to
record the extent of this theft problem.
CONCLUSION
The Seattle solid waste management experiment involves the test .of a
variable user charge for residential refuse service and separate collection
of household sorted recyclables. The Seattle program is probably the most
extensive residential refuse experiment conducted to date in the United
States. It is hoped that the experiment will shed light on the economic
feasibility of, and the impact of these proposed innovations, and at the same
time provide insights into important implementation questions related to the
possible adoption of these suggested innovations.
The Seattle experiment deserves close watching by all those concerned
with the study or management of urban refuse systems. The results from the
experiment presented here are only preliminary. As of this writing the
experiment is still on-going.
230
-------
SECTION 7
RESEARCH DIRECTIONS
"Research Activities in EPA" (Oscar W. Albrecht).
"Solid Waste Collection/Disposal Economic Research Strategies" (G.S. Tolley,
V.S. Hastings).
231
-------
RESEARCH ACTIVITIES IN EPA
by
OSCAR W. ALBRECHT
INTRODUCTION
I have the privilege of discussing an area that is probably less
controversial than some of the areas we've been discussing up to now. The
papers have all been interesting and have generated some lively discussion.
My assignment is to discuss the research programs of the U.S. Environmental
Protection Agency (EPA). I will briefly mention the research and activities
of the major offices, and then discuss in a little more detail the research
conducted by the Office of Research and Development, and particularly at the
Cincinnati Laboratory which is sponsoring this workshop.
RESEARCH BY OFFICES, EXCLUDING THE OFFICE OF RESEARCH AND DEVELOPMENT
The U.S. Environmental Protection Agency (EPA) was established as an
independent agency in the executive branch of the federal government,
effective December 2, 1970. Many of EPA's activities are
enforcement-oriented, but the Agency is involved in numerous research
activities throughout the various offices, divisions, and branches of the
organization.
The EPA is directed by an Administrator, presently Douglas M. Costle,
and reporting to him are various offices headed by assistant administrators.
Within these offices are various divisions and branches which conduct
research; for example, the Office of Planning and Management has an economic
analysis division which evaluates impacts of Agency programs and policies.
These impacts include potential plant closings, effects on employment,
prices, and Gross National Product (GNP). Estimates of abatement costs and
cost/benefit analyses are made for selected industries, principally those
pertaining to air and water regulations. The Office of Planning and
Management (0PM) frequently performs the Impact analyses when program offices
do not have sufficient capability or resources to conduct such analyses on
their own. The 0PM also performs the impact analyses on selected industries
where more than one regulation is involved. The investigations by the Office
of Planning and Management do not preclude similar research by the various
program offices within their own areas of responsibility.
232
-------
The Office of Water and Hazardous Material performs economic analyses
and prepares inflationary impact statements to support regulations developed
by the Office of Water Planning and Standards. The Office also performs
cost/benefit analyses, develops and analyzes alternative control options and
prepares issue papers describing the costs, risks, benefits, legal, and other
aspects of alternative control options.
The Office of Solid Waste (OSW) within the Office of Water and
Hazardous Materials has three divisions, each doing research related to their
own particular areas of concern. The three divisions in the OSW are: the
Hazardous Waste Management Division (HWMD), the Systems Management Division
(SMD), and the Resource Recovery Division (RRD). The HWMD conducts economic
and policy analyses related to the handling and disposal of industrial and
hazardous wastes. It is at present involved in developing criteria and
procedures for implementing the regulations mandated under Subtitle C of the
Resource Conservation and Recovery Act (RCRA) of 1976. The HWMD is also
conducting economic impact analyses of the proposed hazardous waste
management regulations for selected industries.
The SMD in OSW has responsibility for a national program to improve
the collection, processing, storing, and disposal of municipal and other
nonhazardous solid waste. Its major areas of responsibility are providing
technical assistance to cities in improving the efficiency and safety of
solid waste management, developing a better understanding of the solid waste
disposal to land practices, and demonstrating system management plans.
The RRD within OSW performs studies related to the recovery of energy
and materials from solid waste. It is also responsible for implementation of
a national program to reduce the generation of solid waste. A major focus of
RRD has been on large scale resource recovery demonstration projects. The
Division has also studied the use of incentives to stimulate demand for
seconday materials and waste reduction. Other economic studies by RRD
include estimates of consumer demand elasticities for selected products,
product charges for solid waste, study of the demand and supply for secondary
fiber in the U.S. paper board industry, and the economics of waste oil
recovery.
Other offices which conduct economic research include the Office of
Air and Water Management which has an Environmental Analysis Division; and
the Office of Toxic Substances which has a Benefits and Field Studies
Division. Without giving further detail on these, in the remaining time I
would like to discuss research by the Office of Research and Development.
RESEARCH CONDUCTED BY OFFICE OF RESEARCH AND DEVELOPMENT
The Office of Research and Development (ORD) is responsible for the
national research program for control of all forms of pollution. The ORD's
funtions are Agency-wide in satisfying the research and development needs of
the Agency's operating programs, and it is responsible for conduct of
integrated research within the Agency. Within ORD are four main offices;
233
-------
these are: the Office of Monitoring and Technical Support; the Office of
Energy, Minerals, and Industry; the Office of Air, Land, and Water Use; and
the Office of Health and Ecological Effects.
The Office of Monitoring and Technical Support, as the name suggests,
is engaged primarily in activities related to monitoring and enforcement.
The Office of Energy, Minerals, and Industry (EMI) has responsibility for
assessment of environmental and socio-economic Impacts resulting from energy
and mineral resource extraction, processing, and utilization.
The Office of Air, Land, and Water Use (ALWU) is responsible for
assessment of the environmental and socio-economic impacts of land, water,
and air pollution control and management activities. This office is also the
focal point within ORD for providing liaison with the Office of Water and
Hazardous Materials and the Office of Air and Waste Management (which are
outside the ORD organization). Within the Office of ALWU is a Waste
Management Division which has responsibility for assessment of the
environmental and socio-economic impacts of methods to control and manage the
discharge of wastes from municipal, recreational and other domestic sources.
There is also the Agriculture and Non-Point Sources Management Division
within the ALWU, which assesses the environmental and socio-economic impacts
of alternative control methods for pollution from agriculture and other
non-point sources.
The ORD organization also includes the research laboratories at seven
locations over the country. These are at Cincinnati, Ohio; Research Triangle
Park, North Carolina; Athens, Georgia; Corvallis, Oregon; Duluth, Minnesota;
Naragansett, Rhode Island; and Gulf Breeze, Florida.
Before discussing the research at Cincinnati laboratory, let me
briefly mention the other two offices within the ORD. The Office of Health
and Ecological Effects (OHEE) has a criteria development and special studies
division which is developing economic models and methodologies for conducting
socio-economic assessments relating to the benefits of pollution control on
an individual and.broad program basis. The Office of Toxic Substances has a
benefits and field studies division which conducts benefit-cost analyses on
the impacts of cancellations or suspensions and the emergency use of
pesticides.
RESEARCH CONDUCTED BY MERL, CINCINNATI
Research activities at Cincinnati, by the Municipal Environmental
Research Laboratory (MERL), are directed at three principal areas of concern:
solid and hazardous waste, wastewater treatment, and water supply.
The Solid and Hazardous Waste Research Division (SHWRD) within the
MERL, which is sponsoring this workshop, has to an increasing degree been
directing its research to meet the program needs of the OSW which is
responding to the RCRA legislation. In our disposal branch, considerable
effort is being expended on waste characterization in order to learn more
about pollutant transport, including movement through soils by leachates and
gases, in order to develop more effective treatment and control methods. The
234
-------
problems associated with co-disposal of municipal, industrial and hazardous
waste are also areas Investigated and economic and environmental effects of
the disposal alternatives are being analyzed.
Waste characterization is also an area of concern for the processing
branch because of the implications for resource recovery systems. Our
current projects in the area of resource recovery include research on
technologies associated with waste-as-f uels, preprocessing techniques for
recovery of various fractions from the solid waste stream, and the design of
resource recovery systems. Some of our research characterized as economic is
more appropriately described as operations research, but it is mentioned here
mainly because it does not involve technological hardware.
The first few years after formation of the EPA, several research
projects involved studies on the collection, storing, and disposal activities
of municipal solid waste. The feasibility of wet systems for collection and
transport of residential refuse was investigated, as well as the potential
for wage incentives to increase the efficiency of solid waste collection and
disposal. Institutional arrangements for regional and solid waste management
were studied, and sample and test procedures developed for materials
recovered from municipal solid waste. Management systems for hospital and
institutional solid waste were examined in detail. A survey was made of
housewives' knowledge and attitudes towards solid waste and resource
recovery; there was also a study on citizens' attitudes toward disposal of
hazardous waste. The manner in which federal government procurement policies
and practices affected solid waste recycling was also studied.
On the issue of externalities, we funded a study to measure the
external effects of solid waste management. This study, by the Institute of
Policy Analysis (IPA), used regression analysis to determine the effects on
property values. Using several landfill operations in Los Angeles County
(California), it was found that proximity to and view of a sanitary landfill
did not significantly affect surrounding property values. This study pointed
out, however, that additional work should be done on the specification and
exercising of the model as there was inadequate information on the
relationship between property values and the non-environmental variables.
In the recent past, there was considerable concern over the abandoned
automobile problem. We suported a study to determine the extent of the
problem and to evaluate strategies for dealing with it. One suggested
strategy was a mandatory deposit fee, refundable upon certification of proper
disposal. A case study approach to analyzing the salvage industry was also
conducted. This study showed that many.salvage firms are family owned, and
that salvage firms tended to vary widely in the extent of their operations
and capital investments. We also supported a study of the beverage container
problem by Research Triangle Institute. The results suggested that
government intervention could not be justified solely on the basis of a
misallocation of resource inputs. And that to justify government
intervention solely on the basis of energy savings would ignore the utility
that consumers derive from the convenience of nonrefillable containers.
235
-------
An important component of the solid waste stream is the combustible
portion. This component has been of increasing interest recently because of
the concern over diminishing energy sources and high consumption. A study by
the International Research and Technology for EPA estimated the magnitude of
combustible solid waste; in a subsequent study they developed an input-output
model for predicting the quantity and composition of the municipal solid
waste. Additional efforts are currently underway to update the estimates
using EPA's SEAS model and forecasts of economic expenditure patterns.
Wastepaper is the most important component in municipal solid waste
in that it comprises about 35 percent of the total residential and commercial
solid waste. Several EPA sponsored studies have investigated this component.
A study by Resource Planning Institute, using a pro forma income analysis
approach, investigated the potential role of incentives for increasing
recycling. It suggested that high subsidies would be needed to obtain a
significant increase in recyclng of wastepaper.
Another area of concern has been the issue or allegation of
descriminaton in regulated transportation rates, particularly rail freight,
and its affect on the recycling of secondary materials. Two studies funded
by the EPA arrived at somewhat different conclusions. The Resource Planning
Institute study, funded by our office, concluded that on a
chemical-equivalent basis ferrous scrap moved at a lower rate than the virgin
material while the Moshman Associates study sponsored by the OSW arrived at
somewhat opposite results. These differences resulted from variations in
the manner in which the data were used, and varying opinions on the
appropriate chemical formulae and points of substitution between secondary
and competing virgin materials. A shortcoming of these studies, particularly
the Moshman study, was the short-run time perspective in which the issues
were addressed. At present, we are supporting work at the University of
Cincinnati on demand modeling which we hope will ultimately shed further
light on this issue.
Another area of concern is the discarded passenger tire disposal
problem. Although some discarded tires are recycled as door mats, highway
crash barriers, artificial reefs, and the like, these uses constitute only a
small solution to the overall scrap tire disposal problem. One recent study,
initiated by our office, investigated the net benefits derived from using
ground scrap rubber from tires as an additive to asphalt roadways. These
potential benefits are currently undergoing further study with actual field
tests on several highways. Another study just being implemented will
investigate the state-of-the-art, trends, and impediments to various
utilizations of the discarded tires.
Resource conservation and recycling have been and continue to be
proclaimed as desirable goals which this Nation ought to pursue. A study by
the Environmental Law Institute addressed the question as to whether market
prices determine resource conservation and recycling in a socially optimal
fashion. While the study provided no quantitative estimates of the benefits
of conservation, it pointed out that numerous external forces act to create
divergencies between the socially optimal prices and the prevailing prices in
the market for natural resources. These forces, which act in varying
236
-------
directions and with varying intensities over time, are difficult to measure.
Moreover, from an intergenerational viewpoint, a method for determining the
socially optimal prices has yet to be devised.
One of the external forces affecting recycling is taxes. The
Institute found that elimination of subsidies to extractors through the
federal tax code could cause an upward shift of at least 1 percent in the
virgin material supply curve for paper; for steel, 2 percent; and copper 5
percent. If taxes were equalized between secondary and virgin materials,
considering the supply and demand elasticities, the percentage increases in
recycling were estimated as 0.63 percent for wastepaper; 0.42 percent for
scrap steel; 0.35 percent for copper; and 1.00 percent for aluminum.
.The Environmental Law Institute also conducted a study on the
feasibility of using scrap futures markets for encouraging the use of
secondary materials. They concluded that futures trading for ferrous scrap
and wastepaper was technically feasible, and could result in net gains to
society by reducing uncertainty and stabilizing prices and incomes of
suppliers and users of secondary materials. Successful operation of such
markets however, would depend upon adequate participation by the private
sector. A participating role by the government was argued because of the
public good aspect of the information generated by the market.
CURRENT RESEARCH AT MERL (SHWRD), CINCINNATI, OHIO
In addition to the projects just mentioned, several new projects have
been initiated recently in response to provisions of the Resource
Conservation and Recovery Act. The National Academy of Science is
establishing for us priorities for government supported research in the area
of resource recovery. This study will establish priorities for recovery of
both materials and energy from solid waste, taking into consideration the
lack of incentives in the private sector to do the needed research, and the
net benefits expected to be gained by society.
Another study will investigate the impediments to successful
operation of resource recovery facilities. Substantial research efforts in
the past have been directed at the technological "know-how" of facilities and
techniques for recovering materials from municipal solid waste, but yet the
overall recycling rate achieved to date is disappointedly low. This study
will investigate in particular how initial plant design and expectations
compared with the actual market experiences once the facilities became
operational.
In response to the recent legislation (RCRA), we have initiated
several new studies dealing with hazardous waste. One study will determine
the cost-effectiveness of various treatment and disposal alternatives and the
relevance of the various options to management of hazardous waste in the
municipal waste streams. Another study will evaluate existing federal
programs for risk management, with particular emphasis on- the methodologies
used by different government agencies in assessing risk, and their
applicability to risk assessment for management of hazardous waste.
237
-------
We are continuing to have an interest in issues related to efficient
management of solid waste. Several early studies investigated the
theoretical basis for pricing solid waste management and you have just heard
the report on the five case studies of user charges conducted for us by
Mathtech (the Contractor for this study). Several other studies related to
pricing have been mentioned. A list of the past research on pricing and
related research was included in the bibliography mailed to each of the
participants. The substance of these reports generally emphasize the
difficulties encountered in acquiring adequate and meaningful data and
indicate the need for further research before any conclusions can be reached.
We are, of course, interested in your recommendations for further research in
this area.
CONCLUSION
I hope this discussion has given you a general overview of the
research activities by the EPA. You have, no doubt, noted that my discussion
was mainly on economic research the list of research projects in the
technological area would be much longer. As stated earlier, the EPA's total
research effort is not confined to the ORD, but is spread out among various
offices of the Agency. Since the ORD is separate from the
enforcement-oriented program offices, perceptions of research needs
frequently differ as the research laboratories are inclined towards a
longer-run view of research programs. Maintaining an effective degree of
continuity in these research projects is sometimes difficult.
Another problem is the coordination of research within the Agency.
Up to now, the research efforts in the area of solid waste have been minimal
when compared with those for air and water pollution. I think we all agree
that the issues on solid waste we've been discussing are important and
research support and funding are needed if they are to be resolved. You can
help in this effort by articulating these needs whenever you have the
opportunity to do so.
238
-------
REFERENCES
1. Richard Schmalensee, et al. Institute for Policy Analysis. Measuring
External Effects of Solid Waste Management. EPA-600/5-75-011, March
1975.
2. Booz-Allen Applied Research. An Analysis of the Abandoned Automobile
Problem. NTIS PB-221879, 1973.
3. Taylor H. Bingham and Paul E. Mulligan. Research Triangle Institute.
The Beverage Container Problem: Analysis and Recommendations. NTIS
PB-213341, 1972.
4. Frederick A. Cardin. Resource Planning Institute. Secondary Fiber
Recovery Incentive Analysis. NTIS PB-241-082, October, 1974.
5. John F. Foran, et al. Raw Materials Transportation Costs and Their
Influence on the Use of Wastepaper and Scrap Iron and Steel.
NTISPB-229816, April 1974: David G. Abraham, et al. Transportation
Rates and Costs for Selected Virgin and Secondary Materials. NTIS
PB-233871, 1974.
6. Haynes C. Goddard. An Economic Evaluation of Technical Systems for
Scrap Tire Recycling. NTIS PB-249197/5WP.
7. Robert C. Anderson. Evaluation of Economic Benefits of Resource
Conservation. (In Print).
8. Barbara J. Stevens. Impact of the Federal Tax Code on Resource
Recovery: A Condensation. EPA-600/8-77-012, August 1977.
9. Robert C. Anderson and Roger C. Dower. Environmental Law Institute. An
Analysis of Scrap Futures Markets for Stimulating Resource Recovery (In
Print).
239
-------
SOLID WASTE COLLECTION/DISPOSAL ECONOMIC RESEARCH STRATEGIES
by
G. S. TOLLEY
V. S. HASTINGS
A question of major concern in municipal solid waste management deals
with the rising costs of solid waste collection/disposal. Can there be a
payoff in holding down these costs in ways that take into account economic
considerations of more than a strictly engineering or technical nature in
providing solid waste management services? The rising costs are largely
associated with rising input costs, particularly labor, the real cost of
which continues to rise, and rising waste volumes. Concerns associated with
rising volumes include environmental degradation and increased use of scarce
land resources. These concerns are all worth economic investigation if costs
can be significantly reduced by economic approaches that would increase net
social benefits from solid waste collection/disposal. This paper addresses
the subject of research strategies related to the use of incremental user
charges as an economic approach for this purpose. Cost-side and then
demand-side strategies are discussed, and finally a list of research
questions that waste managers might ask is offered.
COST-SIDE STRATEGIES
It is a basic economic proposition that net social benefits are
maximized when services are charged for at the marginal cost of providing the
services. Under existing methods of financing solid waste services, usually
flat user charges or general fund financing, there are no charges related to
the amount or to the marginal cost of collection/disposal. This tends to
lead to overuse of services in terms of quantity, to the point where costs
exceed benefits. However, this does not mean that all charges for solid
waste co11ection/disposal services should be based on amounts
collected/disposed. Some services provided are not waste quantity related.
Charges for these services should be based on their marginal costs.
This point is worth emphasizing not only because it can make a
considerable difference, but because there has been confusion on this point.
For example, Ernst (1975) assumed that a choice must be made between either a
flat charge or a quantity charge. He related all costs to quantity. This
led him to attribute certain empirical findings to scale factors related to
240
-------
population size, whereas in fact more plausible and reasonable explanations
were possible based on true cost relationships. Ernst's cost relationships,
for example, would result in no costs for traveling a route or going to the
back door to look when no wastes were collected. Costs are associated with
these actions regardless of quantities collected. Ernst's conclusions
concerning the efficacy of marginal cost pricing as a general proposition are
thus cast in doubt, even though his conclusions concerning strictly
quantity-based incremental charges (for example, his conclusion that
incremental charges would not cover costs) are probably correct.
These considerations lead to the research strategy proposition that,
for properly setting marginal cost prices, different services associated with
solid waste collection/disposal must be identified and their costs
determined. In addition to amounts collected, these services include
customer services, frequency of collection services, location of collection
services, and kind of waste services.
Customer Services
Certain costs are associated strictly with adding a customer. These
include billing and adding a stop to the route. This means that part of the
total charge should be a flat charge. As an example in another field,
electric utilities now often estimate and sometimes include a specific
customer charge in their billing, based on billing costs and amortization of
customer hookup costs. This avoids the erroneous inference from declining
block charge structures that there is a scale factor involved in the cost of
electricity to a customer. Declining block charges can be replaced with two
marginal cost pricing components that actually apply marginal customer cost
and marginal electricity cost components. A summer cottage customer would
then be given a true picture of his hookup service costs, and could decide
whether to take service based on these costs. He would not impose costs on
the system greater than the value of the service to him. In the solid waste
field, a customer two miles off the normal route could decide whether he was
willing to pay the additional customer service cost to him. Even more to the
point, no customer's use of the service in terms of quantities offered for
collection would be distorted by an erroneous inclusion of a customer cost in
the incremental quantity charge, nor would the solid waste collection agency
have the problem of marginal cost prices not covering costs.
Other Services
Other important services include frequency of collection, location of
collection such as curb or backdoor, and kinds of waste collected. There may
be more, such as service on Mondays as opposed to other days of the week, and
quieter service. Each of these other services would have a cost function
separate from the quantity of service cost function. The frequency of
collection cost function would depend on the cost of traveling the route an
additional time per base unit of time, such as per week, the back-door
service on distances to back door and time costs for labor and route travel
time extension, and the kind of waste on the salvage or special disposal
costs of particular kinds of waste. Service on Mondays, if Monday service
were in heavy demand for disposing of weekend collected yard wastes, for
241
-------
example, would involve costs for peaking on a Monday (equipment for that
day's collection alone). A demand for particularly quiet service on a
particular route would involve the added cost of providing equipment and
training personnel for quieter service.
The literature has indicated that greater frequency of service (Quon
1965) and backdoor service (Wertz 1976) lead to greater volumes for
collection/disposal. However, while these interrelationships are important,
they do not affect the basic cost functions for the different services.
Thus, the cost functions can be estimated independently of what these
relationships might be.
Interrelationships Affecting Costs
Switching to quantity based charges would involve additional customer
costs due to the need .for measurement of quantitites of various services and
of quantity billing. There may also be added littering costs associated with
quantity-based charges. Furthermore, switching from general fund financing
would involve a cost due to loss of income tax advantages. (Such financing
is federal income tax free, whereas individually borne user -charges are not.)
These considerations lead to the need for a benefit-cost assessment to
determine if there would in fact be net advantages in switching to marginal
cost pricing for the services.
Cost Variation
Basic cost functions will vary from place to place. In some areas,
least cost disposal may be land disposal at close-in low price per acre
sites, in other areas the opposite. In some areas, back doors may be close
to the street, in others far removed from the street. In some areas, housing
may be high density, others low density, leading to different costs for route
coverage. Some areas may be more susceptible to littering with incremental
quantity charges imposed than others, possibly depending on socioeconomic
characteristics. Because of these various differences, any general study of
the efficacy of marginal cost user charges must consider the possible ranges
in variations of conditions and associated least costs for providing various
services.
DEMAND-SIDE STRATEGIES
To estimate whether the potential gains from marginal cost pricing
are substantial and if such pricing is worthwhile, various demand curves for
services must be estimated. However, estimation has been limited by a lack
of sufficient data. Application of marginal cost pricing or any form of
incremental pricing that would generate the needed data has been limited.
Given this situation, a possible approach for at least partial estimation of
demand curves is one based on estimating the cost of alternatives to
community collection/disposal services.
242
-------
Cost of Production Approach
Volumes of wastes collected and associated waste management services
can only be substantially decreased if there are lower cost alternatives to
waste generation, to community collection, and to the associated services
such as frequency of service and location of service. The shape of demand
functions for the various collection/disposal services depends on these
alternatives and their costs to users. These alternatives include sink
disposals, carrying wastes to curb, storing wastes over a longer period,
finding alternative collectors for certain wastes and possibly storing those
particular wastes longer (e.g., paper to Boy Scouts), and changing
consumption patterns to generate less wastes (e.g., use of returnable
bottles). Lower cost alternatives to individuals, but not necessarily to
society, may also include littering and on-site burning. Some progress
toward estimation of demand curves can be made through research into these
costs.
The cost of production approach would cover total costs to society
rather than to the individual alone for determining optimum demand for waste
services functions. The costs added to individual costs would include those
of littering and on-site burning. The point can be made that even if greater
experience in the application of marginal cost pricing, and data from these
applications, were available, it would still be necessary to modify demand
curves so generated by social cost considerations for determining net social
gains from marginal cost pricing.
Creation of More Experiments
Actual pricing experiments would still be necessary. Some costs,
such as the bother costs of returnable bottles, and of carrying wastes to the
curb, would be difficult to estimate in the absence of such experiments.
Indeed, the research strategem most needed for investigation of economic
approaches to solid waste collection/disposal services is the creation of
more pricing experiments and the generation of demand side information from
more experiments. The depth of the concerns over the costs of increasing
waste volumes and evidence to date on potential gains from marginal cost
pricing appear to justify this strategem.
Two Basic Approaches to the Creation of More Experiments
Two basic approaches are possible for the creation of more
experiments and generation of more experimental data. One is the creation of
a few well designed controlled experiments, with certain control areas
remaining on flat charges for cross section comparison. Time series analysis
would also be possible if historical information were available on quantities
collected by subareas. The advantages of this approach lie in the care that
can be applied for obtaining needed and reliable comparative results. A
second approach is the promotion of more widespread adoption of pricing of
solid waste collection/disposal service. In this case, cross section
comparison of results with communities not adopting pricing would be used.
Multiple regression techniques would be used to control for other variables.
Again, time series analysis would also be possible given appropriate
243
-------
historical information. The advantages of this approach lie in the fact that
greater variation in other variables can be considered with a greater output
of data to cover various conditions for the same budget and time constraints.
A LIST OF RESEARCH QUESTIONS TO ASK IN SOLID WASTE SERVICES PRICING
EXPERIMENTS
The focus of solid waste services pricing experiments should be on
ultimate applicability, providing answers to solid waste systems managers'
questions. Their questions would include what form financing should take,
what specific charges should be set and how the level should be determined,
and how to measure the services provided. They would also include what
services to provide, and what degree of environmental amenities, such as
litter control and clean up, to provide. As a research approach, it is
useful to consider the various questions that might be asked. The following
outline of possible question is offered:
1. What collection services and level of collection services should be
provided?
a. What frequency of service should be provided? Should this be
decided on a systemwide or individual collection area basis?
b. Should curb or back door service be provided? If back door,
should it be optional, and if optional, should it be optional at each
collection?
c. Should unlimited-quantity service be provided, or should
consideration be given to approaches that might limit or reduce quantities to
be .collected?
d. Should combined, ,same frequency, , and location of collection be
offered for all residential solid wastes, or should separate collection be
offered?
2. What degree of community cleanliness, in terms of freedom from litter and
other pollution from solid wastes, is demanded, and how is this best
achieved? ,
a. What are the marginal benefits (what is the demand function for)
different degrees of cleanliness?
b. What are the various means for achieving cleanliness and what are
the marginal cost functions (cost per unit of cleanliness achieved) of the
various means?
c. What is the optimum level of cleanliness and what are the ,optimum
levels of input of the various means, given these demand and cost functions?
3. What disposal services and level of disposal services should be provided?
244
-------
a. Should there be processing of wastes prior to ultimate disposal
(e.g., incineration)?
(1) What are the benefits of processing in terms of reducing
transportation, ultimate disposal, or other costs?
(2j) What are the costs of processing, including economic
costs (net of any energy values obtained) and environmental costs (air
pollution)?
b. What are the options for ultimate disposal? If the least-cost
alternative is by land disposal, what sanitary land-disposal standards should
be set for this disposal?
(1) What are the marginal environmental benefits of added
efforts to protect against environmental damages from the disposal operation?
These might include benefits from:
(a) More frequent covering or burying of wastes.
(b) More careful site selection, for example, to
find less permeable soils.
(c) More remote sites, for example, to remove the
site further from population centers, or to remove further from groundwater
sources.
(d) More care to separate off and neutralize the
possible health and safety effects of potentially hazardous wastes that might
be introduced into residential solid wastes.
(2) What are the marginal costs associated with achieving
the marginal benefits?
4. How should collection/disposal services be financed?
a. What are the alternative ways of financing collection/disposal
services? These may include:
(1) Flat charges, based on average costs of services
actually rendered over a period of time.
(2) Incremental user charges, based on the incremental costs
of the various collection/disposal services rendered.
(3) Strictly marginal cost pricing, which may be some
combination of flat and incremental charges.
b. What are the advantages/disadvantages (benefits-costs) of
alternative financing methods compared to existing financing methods, for
example incremental compared to flat charge methods?
245
-------
(1) What are the estimated changes in the amounts of the
various services in shifting to incremental user charges?
(2) What are the direct benefits (estimated direct cost
savings) from reducing these amounts?
(3) What are the direct costs to households (estimated loss
in benefits) from reducing these amounts?
(4) What are the indirect costs from shifting to incremental
charges?
(a) What are the added (or reduced) costs of
administering such charges?
(b) What are the added environmental cleanliness
costs?
(c) What are the tax/revenue sharing costs?
5. How can service requirements be best projected for planning and decision
purposes?
a. What are the various projection methodologies and how would they
be implemented?
b. Which projection methodologies are the more reliable in terms of
accuracy? Which in terms of greater flexibility for different conditions or
different modes of operation?
c. Which projection methodology, or methodologies, should be used
for the particular system? If more than one, could there be irreconcilable
differences in results, and how would this be handled?
6. What specific-system research is needed to make all of the above
determinations?
a. What are the specific questions to be asked and answered?
b. What are the research procedures?
c. What are the data needs?
246
-------
REFERENCES
1. Ernst, Ulrich. 1975. Evaluation of the Feasibility and Economic
Implications of Pricing Mechanisms in Solid Waste Management. Abt
Associates, Inc., Cambridge, Massachusetts. US Environmental Protection
Agency, January 1975.
2. Quon, Jimmie E., Masaru Tanaka, and Abraham Charnes. 1968. "Refuse
Quantities and Frequency of Service," Journal of the Sanitary
Engineering Division, Proceedings of the American Society of Civil
Engineers, 94, 403-420, April 1968.
3. Wertz, Kenneth L. 1976. "Economic Factors influencing Households'
duction Production of Refuse," Journal of Environmental Economics and
Management, 2,263-272 April 1976.
Z47
-------
APPENDIX A
AGENDA
TUESDAY, SEPTEMBER 19, 1978
WEDNESDAY, SEPTEMBER 20, 1978
8:30-
9)00
WELCOME: Robert J . Andereon
KEYNOTE ADDRESS! Elaynee Coddard
ADMINISTRATIVE COSTS OF USER FEES
Barbara Stevena
9:00-
9.JO
EFFICIENCY OF USER CHARGES
William Lanen
COMPARISONS OF ALTERNATIVE SOLID
WASTE MANAGEMENT POLICIES
Allen Mledatnft
9:30-
lOiOO
EVALUATING THE EFFICIENCY OF THE
SOLID WASTE CHARGE
Steve Buchanan
TAX POLICY AND SECONDARY MATERIAL USE
Robert C. Anderaon
10:00-
10:20
BREAK
BREAK
10:20-
10:50
EQUITY CONSIDERATIONS IN USER CHARGES
Roger Bolton
DISCUSSION
Frank Smith Seymour Flefcowaky
10:50-
11:20
OPTIONAL, POLICY MIX
Jamea Hudaon
11:20-
12:00
DISCUSSION
Robert C, Anderson Peter Kemper Kernel Goddard
SUMMARY GENERAL DISCUSSIONi
PRICING AND SOLID WASTE MANAGEMENT:
FACT. FAITH, AND FICTION
12:00-
1:30
LUNCH
LUNCH
Ii30-
2:00
THE SEATTLE EXPERIMENT
Bernard Boome
RESEARCH PROGRAMS AT EPA
Oacar Albrecht
2:00-
2t30
EXPERIENCE IN EUROPE
W. David Conn
RESEARCH STRATEGIES
Steve Haattaga
2:30-
3.00
ESTIMATES OF SECONDARY MATERIALS SUPPLY
Taylor Blngham
WHAT THE SOLID WASTE MANAGER NEEDS
Jack Dunn
3:00-
3:20
JiZO-
J.50
3:50-
4:30
7:20-
8|30
BREAK
PRICINQ MUNICIPAL REFUSE SERVICE
Kenneth WertB
SUMMARY GENERAL DISCUSSION:
RESEARCH PRIORITIES
DISCUSSION
John Tbompaon M0rrlle« Bonney
Jochen Kuhner Seymour Flekowaky
ADDRESS BY:
David Tundarman, CEQ
248
-------
APPENDIX B
LIST OF PARTICIPANTS
Oscar W. Albrecht
U. S. Environmental Protection
Agency
Cincinnati, OH
Robert C. Anderson
Environmental Law Institute
Washington, D. C.
Robert J. Anderson, Jr.
MATHTECH, Inc.
Princeton, NJ
Tayler Bingham
Research Triangle Institute
Research Triangle Park, NC
Roger Bolton
Williams College
Williamstown, PA
Bernard H. Booms
University of Washington
Seattle, WA
Steve Buchanan
Public Interest Economics Center
Washington, D. C.
W. David Conn
University of California at
Los Angeles
Los Angeles, CA
Philip C. Cooley
Research Triangle Institute
Research Triangle Park, NC
Haynes C. Goddard
University of Cincinnati
Cincinnati, OH
V. Stevens Hastings
WAPORA, Inc.
Chicago, IL
James F. Hudson
Urban Systems Research
& Engineering, Inc.
Cambridge, MA
William N. Lanen
MATHTECH, Inc.
Princeton, NJ
Allen K. Miedema
Research Triangle Institute
Research Triangle Park, NC
Barbara J. Stevens
ECODATA, Inc.
New York, NY
Curtis E. Youngblood
Research Triangle Institute
Research Triangle Park, NC
Kenneth L. Wertz
University of North Carolina
Chapel Hill, NC
George S. Tolley
University of Chicago
Chicago, IL
249
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/9-80-001
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Proceedings of a Symposium on Economic Approaches
to Solid Waste Management
5. REPORT DATE
May 1980 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Robert J. Anderson, Jr.
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Mathtech, Inc.
Box 2392
Princeton, New Jersey 08540
10. PROGRAM ELEMENT NO.
1DC818 - SOS #5, Task 06
11. CONTRACT/GRANT NO.
68-03-2673
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory--Gin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA 600/14
15. SUPPLEMENTARY NOTES
Project Officer - Oscar W. Albrecht
513/684-7881
16. ABSTRACT
Incremental user charges or prices as a means of assuring equity and efficiency in
solid waste management have interested economists for some time. The ultimate
effects of pricing are not well-known, however, and most communities continue to
rely on revenues from property taxes and flat charges to finance the collection and
disposal of residential solid waste. The Symposium sponsored by the Municipal
Environmental Research Laboratory (Cincinnati, Ohio) provided an opportunity for
economists interested in solid waste problems to exchange ideas and views on the
potential role of pricing in municipal solid waste management. This report
contains the formal papers presented at the Symposium held at Philadelphia,
Pennsylvania on September 19 and 20, 1978.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
Demand (economics)
Mathematical models
Econometrics
User charges
User fees
Solid waste
5c
68c
91a
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
256
20. SECURITY CLASS (THitpage)
22. PRICE
EPA Form 2220-1 19-73)
250
ftUSQPO: 1980 657-146/5695
------- |