1 r3 .
IV“: J

$21 .p

1.
, $ 1 F
{25 «SR

 

{‘1

V I A!
Efiro :
a: mug».
‘nIVmg... ..
iflfiul; «haunt
w? 11:9 3..

as

 

.A 5w; 1.?

in»; 3.319.»: x
(i.
“23:...

v. J9...“ P4.

 

 

.D
\‘K‘Lfl'v I
4.3. .. ltl .\v x

r}.F!~|fia..Z.~3u~.\I.§al. wi.‘wuv11e..=dlflni_251fimdl 1. W.‘H‘ums..fl‘wum.v ,. , . u oifimnf 1m".- .. . .. g ‘ VWRWrn. ”xvi-Q . 1.6.5.: 3»? .Wvuuilmw.
‘ ‘ ‘ 7 l. . .. . .J.Hw.y4utwvn nu...

 

 

'
1m
l
O'th‘l

.’ _/ \ '. cl
,, at, i) 3 Q43

This is to certify that the
thesis entitled

AN ECONOMIC ANALYSIS OF THE EFFECTS OF
RECYCLING FOCUSED ON BOTTLE COLLECTION
PROGRAMS IN THE PACKAGING SECTOR

‘ presented by
Yongseong Ha

has been accepted towards fulfillment
of the requirements for the

MS. degree in Packaging

Major Professor’s Signature
J% .22 #03
, ,

Date

 

 

 

MSU is an Affirmative Action/Equal Opportunity Institution

~v--.--—-

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

MAR 1 3 2005
34 f

 

A!"
U0

”I: D

Q

nhr“

I

 

1v . M:
g‘ ‘4 ' A
, W ' t- I v

\.

m

j

 

APR 0 8 2013

 

 

 

 

 

 

 

 

 

 

 

6/01 cJClFlC/DateDuopGS—p. 1 5

 

AN ECONOMIC ANALYSIS OF THE EFFECTS OF RECYCLING FOCUSED ON
BOTTLE COLLECTION PROGRAMS IN THE PACKAGING SECTOR

By

Yongseong Ha

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

School of Packaging

ABSTRACT

AN ECONOMIC ANALYSIS OF THE EFFECTS OF RECYCLING FOCUSED ON
BOTTLE COLLECTION PROGRAMS IN THE PACKAGING SECTOR

By

Yongseong Ha

With a variety of efforts, the infrastructure for recovering beverage containers has
been well developed. This study investigated the major factors that affected recycling
activities during the span of 1989 to 2002.

Bottle collection methods and bottle deposit rates are the main areas of interest.
Mandatory bottle deposit laws and curbside collection programs are two major ways to
collect for recycling. Currently nine states have bottle deposit laws and California has a
redemption system. Deposit amounts vary by state: 2.5 cents, 5 cents and 10 cents.
Alternatively, curbside recycling programs are found in most US. states.

This study focused on the two main aspects: recycling and bottle deposit systems.
First, the relationship between the recycling rate and various socioeconomic variables
was examined. Second, the effect of different deposit rates was investigated.

Using multiple regression analysis, this study found that: Waste generation per
person shows the highest correlation with recycling activity. The 5 cent system is more
associated with high recycling rates among the three deposit rates. And bottle deposit

systems contribute to recycling efforts more than curbside programs.

ACKNOWLEDGMENTS

Without the guidance and assistance from many people, this study would never
have been completed. First of all, I would like to express my appreciation to my advisor,
Dr. Susan E. Selke for not only her invaluable comments and criticism, but her
encouragement and willingness to persevere. She freely gave countless hours working
with me and her ceaseless assistance made all the difference. I also would like to thank
my guidance committee, Dr Diana Twede and Dr. David Johnson for their interest and
participation.

I want to thank all my friends in School of Packaging, who have my time at MSU
enjoyable. Many of you are gone, but you are not forgotten.

I give special thanks to Mr. Sangchoon Icon for his assists in dealing with data
analysis. He was ready anytime with willingness to help when I needed him most.

I would like to thank my parents for their everlasting encouragement and support.
The person that deserves the most recognition is my lovely wife, Kyungnam. Her
sacrifices, patience, advice, encouragement and understanding are greatly appreciated.

My appreciation is extended to my lovely sons, Daewan and Chang-Woo.

iii

TABLE OF CONTENTS

LIST OF TABLES

 

LIST OF FIGURES

 

CHAPTER 1: INTRODUCTION

 

INTRODUCTION

 

TRENDS IN MUNICIPAL SOLID WASTE DISPOSAL

 

WASTE MANAGEMENT STRATEGIES

 

Economic Instruments

vi

vii

 

 

Characteristics of Packaging Waste
MAJOR BOTTLE COLLECTION METHODS

 

Mandatory Deposit Systems

 

 

Curbside Programs

 

RESEARCH PURPOSES
ORGANIZATION OF THE RESEARCH

 

CHAPTER 2: LITERATURE REVIEW

 

INTRODUCTION

 

WASTE MANAGEMENT STRATEGIES FOR RECYCLING

 

 

Determinants of Residential Recycling Decisions
Environmental Attributes and Behaviors for Recycling

 

THE EFFECTS OF BOTTLE COLLECTION PROGRAMS

 

 

CONCLUSION OF LITERATURE REVIEW

CHAPTER 3: METHOD

 

 

INTRODUCTION
RESEARCH DATA

 

THE MODELING METHOD OF ESTIMATION

 

 

ESTIMATION FOR FACTORS AFFECTING RECYCLING
Hypotheses

 

Regression Model

 

ESTIMATION FOR FACTORS AFFECTING RECYCLING RATE

 

IN THE BOTTLE REFUND STATES
Hypothesis

 

Regression Model

 

iv

i—nr—tr—I
OOQUJWNOQb-fiht—‘W‘W

19
19
19
19
23
26
29

31
31
31
35
35
35
39

40
4o
41

CHAPTER 4: RESULTS

 

INTRODUCTION

 

 

RESULTS OF ANALYSIS
Estimation for Factors Affecting Recycling

 

Estimation for Factors Affecting Recycling Rate in the
Bottle Refund States

 

SUMMARY OF THE RESEARCH RESULTS

 

 

CHAPTER 5: CONCLUSION
INTRODUCTION

 

CONCLUSIONS

 

DISCUSSIONS AND LIMITATIONS OF THE STUDY

 

 

RECOMMENDATIONS FOR FUTURE STUDY

APPENDICES

 

BIBLIOGRAPHY

 

43
43
43
43

51
56

58
58
58
61
63

64

80

LIST OF TABLES

Table 1 — Landfill Tipping Fees by States in 2000

 

Table 2 — MSW Generation Rate from 1960 to 2000

 

 

Table 3 —— Packaging Waste by Products in 2000

Table 4 — The Summary Statistics of Variables

 

Table 5 — Test of the Factors of Recycling Rate

 

Table 6 — Estimated Correlation Matrix of Variables

 

 

Table 7 — Breusch-Pagan Test for Overall Recycling Rate
Table 8 — The Result of GLS Regression for Overall Recycling Rate

Table 9 — The Result of Regression for the Deposit States

 

 

Table 10 — The Result of Stepwise Regression for the Deposit States

 

 

Table 11 — VIF Test for Bottle Refund Regression

 

Table 12 — Breusch—Pagan Test for Bottle Refund System

Table 13 — Summary of the Research Results

 

vi

34

44

45

47

49

52

53

54

56

57

LIST OF FIGURES

 

Figure 1 - Trend of Recycling Rates in the US.

Figure 2 — Number of Curbside Programs

 

Figure 3 — Population Served in Curbside Programs

 

Figure 4 — Residual vs. Predicted Value for Overall States

Figure 5 — Residual vs. Predicted Value for Deposit States

vii

 

 

12

15

16

46

55

CHAPTER 1: INTRODUCTION

INTRODUCTION

The human activities of production and consumption give rise to solid waste
because the level of economic activity relates to the level of waste emitted. According to
the laws of thermodynamics, it is not possible to convert waste residuals completely into
beneficial or even harmless products. Although some amount of solid waste is an
inevitable consequence of human activity, it may be possible to minimize the amount of
solid waste generated in our environmental system. This introductory chapter reviews the
current waste management options and discusses the research problems including the

research objectives.

TRENDS IN MUNICIPAL SOLID WASTE DISPOSAL

The handling and disposal of solid waste is one of the major concerns for
communities throughout the United States. The current solid waste problem is related to
both a waste disposal problem and a waste generation problem. The United States is
generating more waste now than ever before. From 1960 to 2000, total US. municipal
solid waste (MSW) generation increased over 260 percent from 88.1 million tons to
231.9 million tons per year, while per capita generation increased over 170 percent from
2.68 to 4.51 pounds per person per day (EPA, 2002).

To reduce or manage MSW, various methods have been applied. The main
components of the waste management policy are source reduction, reuse of products,

recycling of materials, and waste combustion and landfilling. Although source reduction

and recycling are EPA’s preferred practices, landfilling is the current dominant method of
MSW management. According to EPA, 55 percent of MSW generation was landfilled
and the rest was handled by recycling (30%) and combustion (15%) in 2000.

In using the landfilling and combustion techniques, however, some restrictions
have been made. EPA (2002) reported that there are fewer years of landfill capacity
available and most states have less than a decade of capacity left. Besides, landfill tipping
fees vary by states. For example, Table 1 shows the landfill tipping fees of eight states.
Vermont has the highest fee at $75 per ton while Colorado has the lowest fee at $11 per
ton. Besides these eight states, Goldstein and Madtes (2001) also reported that fourteen
states are in the $20 to $30 per ton range and eight states are in the $30 to $40 per ton

range.

Table]. Landfill Tipping Fees by State in 2000

 

 

State Fee ($/ton)
Vermont 75.0
Massachusetts 67.0
New Hampshire 66.0
Maine 65.0
Delaware 58.5
New Jersey 55.0
Nevada 20.0
Colorado 1 1.0

 

Source: Goldstein and Madtes, 2001

Many states also depend on combustion systems for managing MSW. According
to Goldstein and Madtes (2001), the District of Columbia had the highest combustion rate
at 79 percent. In Connecticut, 65 percent of MSW was incinerated, followed by Maine
(39 percent), Massachusetts (36 percent), and Hawaii (32 percent). In using combustion
systems, however, air pollution is caused because some residue reaches the air. To deal
with this problem, various emission control systems are used such as settling chambers,
cyclones, filtration, electronic precipitators and wet scrubbers. There are also solid
residues that generally must be landfilled. EPA (2001) estimated that MSW combustor
ash amounts to about 25 percent (dry weight) of unprocessed MSW input.

Under these circumstances, recycling is becoming the more preferred practice to
deal with MSW. Recyclable materials must be collected before they are processed and
recycled into valuable products. In general, collection of recyclables from residential
sources is separate from commercial recyclables collection programs. Most residential
recycling is recovered from curbside collection, drop-off centers, buy-back programs, and
container deposit systems. On the other hand, the largest quantity of recovered materials
in the commercial sector is provided by old corrugated containers (OCC) and office paper.
The recovered materials are picked up by a paper dealer or self-delivered to the recycler.
In 2002, EPA estimated that residential waste constituted between 55 and 65 percent of
total MSW generation and commercial waste was between 35 and 45 percent of MSW.

In 2001, Goldstein and Madtes reported that recycling is considered as a preferred
method for managing MSW in the United States. Ten states have mandatory recycling

goals for waste reduction, and four states have already attained their recycling goals.

Massachusetts and Rhode Island have the largest recycling goals at 70 percent. In

addition, sixteen states aim to reach at least half of their wastes recycled.

WASTE MANAGEMENT STRATEGIES
Economic Instruments

Economic analysis has been used in decision-making in MSW management due
mainly to the advantage of economic instruments, which help modify human behavior
through the price mechanism. In managing packaging waste generation, it should be
defined where in the solid waste generation and disposal system is the optimal point to
apply incentive-based policies. In general, incentives applied at the point of disposal
(downstream) are considered as more efficient than those applied at the point of purchase
or manufacture (upstream).

A variety of economic instruments have been used or considered in waste-related
policy. All are attempts to achieve an efficient solution by influencing the wide discretion
of producers and consumers. Although it is difficult to decide which option should be
used, an incentive must start from the baseline goal that the generation of packaging
waste is optimally prevented or that waste generated is treated as efficiently as possible.
The following economic instruments can be applied in waste management policy for
packaging materials:

1) Virgin materials tax; this tax is put on raw materials used for the production of

packaging products. As source reduction, it is applied to reduce the use of raw

materials.

2) Product charges; they are added to the price of packaging products as taxes. When
products are fully made from recycled materials, they can be exempted, while a
lower charge can be imposed if products are partly made from recycled materials.
These charges can facilitate reuse and recycling.

3) Waste charge; payments can be charged when waste is collected or disposed. As a
user charge, this system increases the final costs of waste disposal. The charge
can be directly related to disposal of packaging, when packaging is a major
component of the waste.

4) Recycling credits; Credits can be paid to those who reuse containers or recycle
materials. This method can be used to encourage the division of waste fi'om
landfill to recycle.

5) Deposit-refund systems; basically this instrument is a type of product charge since
the surcharge is applied to the packaging products. However, it is a refimdable
surcharge because the surcharge will be returned if the charged products are
prevented from final disposal as waste. This system can be applied as either
voluntary or mandatory.

Using a legal framework, various economic instruments provide a choice to the
consumers or producers to take an action, either “abate and save” or “pollute and pay for
it.” The instruments described can be applied in a mixed system, depending on waste
management policies. However, charges should be carefully reviewed because they can
be double-counted, possibly causing damage to the production and consumption of
products. In addition, whatever economic instruments are used for recycling, they should

be connected to the whole MSW stream.

Characteristics of Packaging Waste

Table 2 shows that containers and packaging waste is the largest proportion of
total MSW generation. In 2000, containers and packaging waste consisted of over 32
percent (74.7 million tons) followed by nondurable goods at 27.5 percent (about 63.7
million tons). As a percentage of MSW, the long term trend of packaging waste
generation was higher in 1980 and decreased somewhat in 2000 (EPA 2002). However,
the amount of waste generated has increased dramatically for several decades. Despite
various efforts to reduce waste generation, the amount of packaging waste is still
increasing. To understand the increase in waste, two main motivating factors should be
considered. First, manufacturers and sellers of products and packaging usually have no
responsibility for handling materials after they are discarded. Second, recycling competes
with raw materials in uncertain economic conditions, in which the market price of

recycled materials is very changeable.

Table 2. MSW Generation Rate from 1960 to 2000

 

 

 

Products 1960 1980 2000

1000 tons (%) 1000 tons (%) 1000 tons (%)
Durable Goods 9920 11.3 21300 14.4 36330 15.7
Nondurable Goods 17330 19.7 34420 22.7 63660 27.5
Containers and PKG 27370 31.1 52670 34.7 74730 32.2
Food Waste 12200 13.8 13000 8.6 25900 11.2
Yard Trimming 20000 22.7 27500 18.1 27730 12.0
Others 1300 1.5 2250 1.5 3500 1.5

 

Source: EPA, 2002

Table 3 shows generation and recovery of packaging wastes in each product
category in 2000. Paper and paperboard comprised the largest fraction of packaging
waste generation at over 39 million tons, followed by plastics and glass packaging which
were equal at 11.2 million tons. On the recovery side, over 58 percent of steel packaging
waste was recovered from the MSW stream and over 56 percent of paper and paperboard
packaging waste was also recovered. The recovery rate of plastic packaging wastes was
relatively lower than other packaging products, at 9 percent. On the other hand, bottle
recovery in each packaging type contributed to increasing the recovery rate of packaging
waste. More than half of steel and aluminum cans are recovered from the waste stream
and over 30 percent of plastic soft drink bottles are also collected for recycling (EPA
2002)

This trend results from the fact that recyclable beverage containers have relatively
higher market value. In particular, aluminum cans have a high recycling value. However,
EPA (2002) reports that the recycling amount for plastic beverage bottles stayed constant

and recently decreased for glass bottles and aluminum cans.

Table 3. Packaging Waste by Products in 2000

 

 

 

 

 

 

 

 

Generation Recovery Recovery Rate
(1000 tons) (1000 tons) (°/o)
Glass
Beer and Soft Drink Bottles 5860 1560 26.6
Wine and Liquor Bottles 1970 440 22.3
Food and Other Bottles & Jars 3360 940 28
Total Glass Packaging 11190 2940 26.3
Steel Packaging
Food and Other Cans 2640 1510 57.2
Other Steel Packaging 240 160 66.7
Total Steel Packaging 2880 1670 58
Aluminum
Beer and Soft Drink Cans 1520 830 54.6
Other Cans 50 Neg. Neg.
Foil and Closures 380 40 10.5
Total Aluminum Packaging 1950 870 44.6
Paper and Paperboard
Corrugated Boxes 30210 21360 70.7
Milk Cartons 490 Neg. Neg.
Folding Cartons 5580 430 7.7
Other Paperboard Packaging 200 Neg. Neg.
Bags and Sacks 1550 300 19.4
Other Paper Packaging 1370 Neg. Neg.
Total Paper & Paperboard
Packaging 39400 22090 5 6. 1
Plastics

Soft Drink Bottles 830 290 34.9
Milk Bottles 690 210 30.4
Other Containers 2630 260 9.9
Bags and Sacks 1650 10 0.6
Wraps 2550 170 6.7
Other Plastic Packaging 2840 90 3.2
Total Plastic Packaging 11190 1030 9.2
Others Misc. Packaging 8120 480 6.1

 

Neg. is less than 5000 tons or 0.05 percent
Source: EPA, 2002

MAJOR BOTTLE COLLECTION METHODS

To control the generation of packaging waste, various collection methods are
operating in the market. Especially, beverage containers such as cans and bottles are the
aim at the state level. Typical methods for container collection include deposit systems,
curbside programs, drop-off centers and buy-back centers. However, only two collection
methods, mandatory deposit systems and curbside recycling programs, are the focus of

this discussion because these two programs are widely used.

Mandatory Deposit Systems

Mandatory deposit-refund systems called “bottle bills” have been implemented
for a reduction of packaging waste from beverage containers in the United States. In
dealing with packaging waste, it may be considered that this system increases return of
packaging and diverts recyclable materials from the MSW stream. Current deposits apply
to glass, metal, and plastic containers for carbonated soft drinks, beer, wine, and some
bottled water and other beverages. Nine states have enacted container deposit systems:
Connecticut, Delaware, Iowa, Maine, Massachusetts, Michigan, New York, Oregon, and
Vermont. Except for Michigan, which has a 10 cent deposit, the deposit in the remaining
8 states is the same at 5 cents. In these nine states, consumers have to pay a deposit on
beverage containers when they buy the products. The deposit will be redeemed if the
empty containers are returned to retailers. This law in most states requires retailers to
accept for refunds containers of the same brand and type that they offer for sale. In
addition, retailers in most cases are paid a fee to cover the costs of handling returned

containers (OECD, 1993). California applies a rather different system in which the

consumer pays no separate deposit, but containers can be redeemed when they are
returned to designated redemption centers. In California and Maine, more containers are
covered than in other deposit states (Appendix A).

Although bottle deposit laws only aim at beverage containers, the actual amount
of these containers accounts for less than 4 percent of total MSW generation. Of this
portion, about 35 percent of all recovery of beverage containers comes from the nine
deposit states, and another 20 percent comes from California (EPA, 2002).

The rationale for why beverage container wastes have been targeted includes:

1) The typical lifetimes of beverage containers are less than one month (Stilwell et
al., 1991). This fact indicates that a large volume of packaging waste is disposed
of by consumers who just want the product, not the packaging. With little
understanding of the benefits of packaging, packaging is treated as a by-product
of the purchase, and discarded right after the contents are removed.

2) Manufacturers and sellers usually have no responsibility for the environmental
impact of their products such as collecting, recycling, or landfilling discarded
materials (GRRN, 2000). Therefore, taxpayers are forced to pay for disposal costs
in general. This situation may create a free rider problem in handling the
packaging wastes because someone can produce more wastes without paying
more control costs.

These characteristics imply that in most cases the generation of waste beverage

containers is caused by individual people. Consumers have a direct influence on the
generation of waste beverage containers. To deal with this characteristic, container

deposit systems are introduced because they are focused on modifying consumer

10

behavior using the polluter pays principle. OECD (1974) explained the principle as the

following:
“The polluter should bear the expenses of preventing and controlling
pollution ‘to ensure that the environment is in an acceptable state.’ The
notion of an ‘acceptable state’ decided by public authorities, implies
that through a collective choice and with respect to the limited
information available, the advantage of a further reduction in the
residual social damage involved is considered as being smaller than
the social cost of further prevention and control. In fact the Polluter-
Pays Principle is no more than an efficiency principle for allocating
costs and does not involve bringing pollution down to an optimum
level of any type, although it does not exclude the possibility of doing
so.”

The effectiveness of container deposit systems is positively proved in the market. For
instance, OECD (1993) assessed the impacts of the container deposit systems and found
that return rates of containers were between 72 to 98 percent in the nine deposit states. In
addition, reductions of litter and solid waste volumes have been found as a result of the
deposit return systems. Additionally, EPA (2002) estimated that about 55 percent of
recovered beverage containers come from the nine deposit states and California. Figure 1
shows the effect of recycling efforts between the deposit states and the non-deposit states
in the nation. The trend reveals that the deposit states have a higher recycling rate than
the non-deposit states in recent decades.

Successful recycling basically depends on collecting high-quality post-consumer
materials. BEAR (2002) noted that deposit systems provide the highest quality materials
with the highest market values, as mandatory deposit systems avoid cross-contamination
of materials at the point of collection and sorting. Therefore, it is considered that

mandatory deposit programs have efficiently influenced the recovery of materials of

higher quality from the waste stream.

11

 

 

Figure 1. Trend of Recycling Rates in the us.

     

 

40 .
35 ,
30 .
ES 25 , ......
a
g 20 .
2. .
5 15 . +Deposfl
10 j +Non-deposit
5 ‘ ...... Average
0 I r j

 

198919901991 1992 1993 19941995 1996199719981999 2000

Year

 

 

Source: Glenn (1990, 1992, 1998, 1999), Glenn and Riggle (1991), Goldstein (1997,
2000), Goldstein and Madtes (2001), Steuteville (1994, 1995, 1996),
Steuteville and Goldstein (1993)

12

Curbside Programs

Curbside recycling programs are offered to residents and may be operated by
local governments or private businesses. These collection programs typically cover most
single-family homes and some multi-family residences. In general, curbside participants
are asked to place their recyclables in curbside collection bins. When the citizen separates
recyclable materials into discrete containers at curbside, collection crews load materials
on a specially designed vehicle. If all materials are commingled into one bin, the
collection crew sorts the recyclables, or the commingled recyclables are transported to a
special facility for separation.

In most cases, recyclables are easily commingled in the process of curbside
collection because typically the programs allow households to put recyclables in bins or
carts without any sorting. Due to the commingled nature of the collected materials, such
programs require materials recovery facilities (MRFs), where commingled materials are
sorted into single-stream materials. This recovery system is designed to limit cross-
contamination of materials, which results in lower-valued materials or even unmarketable
ones. Along with the MRFs, truck-side sorting is an option that is sometimes used.

From the economic point of view, constructing MRFs requires a significant
financial expense and also results in substantial costs for maintaining and operating the
facility. For this reason, MRFs are typically located in large population areas that can
manage the significant costs. For instance, Mullins et al. (1997) reported that cities with
populations of 250,000 and over typically have two MRFs and the average cost to
maintain a MRF is $643,271. Despite the high cost requirements for MRFs, curbside

programs have been rapidly growing for the last decade in the United States. In 2000,

13

more than 9700 curbside programs were operated (Figure 1) and about 140 million
people received curbside recycling services (Figure 2).

Goldstein and Madtes (2001) reported that New York had 1500 curbside
recycling programs in 2000, leading the nation. Pennsylvania was the second with 892
curbside programs, followed by Minnesota (765) and Wisconsin (631). California led the
nation with 31.1million people served by this service and New York was the next with
17.2 million people. This report also found that 100 percent of the population had access
to a curbside program in Connecticut. New York and New Jersey were at 90 percent or
higher, and California, Nevada, Rhode Island and Washington were at over 80 percent.
Nationally, about 50 percent of the U. S. population was served by curbside programs in
2000.

Curbside programs are relatively convenient, compared to mandatory deposit
systems, because they provide easy access to the collection point for households. In
addition, only low capital and operating costs for processing will be necessary if
materials are highly sorted (NREL, 1992).

Collection schedules vary from once a week to once a month. The collection
frequency can influence the participation rate and the total quantity recycled. The special
type of curbside program will be selected as a function of the community’s demographics,
the availability and reliability of processing facilities, community value, and the type of
collection vehicles used (McGrath, 1990). Therefore, all curbside programs are not
identical. One program may be poorly run, covering few materials, while another may be

well publicized, very convenient, covering many more materials.

14

 

 

Figure 2. Number of Curbside Programs

12000-

1CD00-

Curbside Programs

 

 

I I I I I

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
Year

 

Source: Glenn (1992, 1998, 1999), Glenn and Riggle (1991), Goldstein (1997,
2000), Goldstein and Madtes (2001), Steuteville (1994, 1995, 1996),
Steuteville and Goldstein (1993)

15

 

 

 

Figure 3. Population Served in Curbside Programs

§

#1

E

120.

Population (million)

oases?

 

 

19901991199219931994199519$199719981999 2000
Year

 

Source: Glenn (1992, 1998, 1999), Glenn and Riggle (1991), Goldstein (1997,
2000), Goldstein and Madtes (2001), Steuteville (1994, 1995, 1996),
Steuteville and Goldstein (1993)

16

 

RESEARCH PURPOSES

As discussed, beverage containers are recovered through various collection
systems. Beverage container materials maintain a higher market value than other
packaging materials. With a variety of efforts, therefore, the infrastructure for recovering
beverage containers has been well developed. However, the efforts to reduce beverage
container materials in the waste stream vary state by state. As discussed, some states have
enacted mandatory deposit systems to recover beverage containers for recycling and
others have several collection programs in which residents participate voluntarily. This
gives rise to interest in how the different collection systems and different deposit levels
are related to the recycling activity in the nation.

This research will investigate and test some major factors that affected consumer
attitudes toward recycling practices during the span of 1989 to 2000. Using an economic
analysis, this study will present empirical evidence. The findings can be applied in the
development of effective and efficient options for waste management policy. The
research will focus on tWo main aspects: recycling and deposit-retum systems. First, the
research examines the relationship between the recycling rate and various socioeconomic
variables. Emphasis will be placed on the examination of the premise that the level of
waste generation has a positive impact on recycling efforts. With a set of explanatory
variables, this test will show how the recycling activity is affected by the amount of
individual’s behavior.

Second, this study examines the effect of container deposit legislation. Although
this legislation is solely focused on the packaging waste to divert recyclable containers

from the MSW stream, bottle deposit laws will be applied to identify how recycling

l7

efforts in the packaging sector affect the overall recycling activity. Currently the nine
deposit states and California have enacted bottle bills, but they have different levels of
deposit such as 2% cents, 5 cents, and 10 cents. The effects of this difference will be
examined to investigate what level of deposit is most effective. The differences between
the bottle-bill states and non bottle-bill states will also be identified. In addition, the

effects of curbside collection programs will be evaluated.

ORGANIZATION OF THE RESEARCH

The following chapters of this research are organized as follows: Chapter 2
presents a review of the relevant literature, mainly focusing on recycling programs,
mandatory deposit systems and curbside programs, and recycling behaviors. Chapter 3
describes the methods of analysis, data collection methods, research hypotheses and
parameter selection. In Chapter 4, the statistical analyses and findings are discussed.
Finally, Chapter 5 presents the conclusions of the research. It presents the summary of the

empirical findings, policy implications and directions for future research.

18

CHAPTER 2: LITERATURE REVIEW

INTRODUCTION

This chapter reviews the research related to the proposed study. Current issues in
the solid waste management area will be identified, along with analyzing a variety of
relevant studies and previous contributions. It consists of two broad parts: 1) waste
management strategies for recycling activities and 2) effects of bottle collection programs.
Because collection of commercial solid waste is usually implemented by contract with

private firms, the commercial sector will not be emphasized.

WASTE MANAGEMENT STRATEGIES FOR RECYCLING

Most waste management strategies start with designing some type of incentive for
recycling, and then it is presented to citizens as either a set package or a choice from a
number of incentives. In developing solid waste policy, a critical factor is to understand
citizen’s opinions and whether there are options that make the disposal alternative more
attractive to them. As proper and well developed methods are implemented, a better
physical and social infrastructure will be provided. Therefore, a variety of recycling
programs have been provided to create an incentive for residents to recycle as much as

possible and dispose as little as possible.

Determinants of Residential Recycling Decisions

In the past, researchers have examined the factors that influence participation in

recycling efforts. Along with these efforts, local officials also desire to initiate or to refine

l9

recycling programs. The main reason for this phenomenon is that if they can manage the
factors that are important for achieving a high rate of recycling, the recycling activity can
be more easily applied in our society (De Young, 1986; Sundeen, 1988).

To identify significant determinants, the socioeconomic attributes of the residents
have been explicitly related to recycling activities. These attributes typically include
personal factors such as income, age and education. Many investigations have provided
empirical evidence showing that these variables strongly affect environmental behavior,
especially recycling.

Using OLS (ordinary least squares) estimation techniques, Richardson and
Havlicek (1978) analyzed economic and social factors affecting the generation of
household solid wastes. In the empirical model, the household is treated as a production
unit generating solid wastes. This study found that the household income is positively
related to the waste generation of glass containers, aluminum and other metal products,
but not plastics in general. In addition, household size and age (18 to 61 years of age)
have a positive correlation in the waste generation. In terms of the effects of income and
age variables, this result is consistent with Mohai and Twight (1987), who especially
pointed out that the young indicate the greatest level of environmental concerns, while
middle-aged people are the most likely to voluntarily engage in recycling programs. As a
similar point of view, De Young (1986) found that recyclers are older (mean 42 years of
age) than non-recyclers (mean 35 years of age), and recyclers have higher income levels
($40,000-$50,000 range) than non-recyclers. In contrast, Sundeen (1988) found that

income, homeownership, and age are not related positively to recycling activities.

20

According to Vining and Ebreo (1990), residents with higher incomes and
education are more likely to recycle. They explained this result as due to either better
access to information or because they have relatively more materials to recycle than those
with lower incomes.

Focusing on a pricing system for waste collection services, Hong et a1. (1993)
showed that the disposal fee significantly affects households’ curbside recycling
participation. For example, the participation rate in curbside recycling increases as the
disposal fee increases. Due to the same effect, the quantity of non-recyclables decreases
as the fee increases. On the other hand, changes in income are not likely to convert
household’s generation of total solid wastes into curbside recycling. Therefore, they
insisted that income is considered as an insignificant factor for the participation decision
on curbside recycling through the empirical analysis.

From a somewhat different point of view, Lansana (1992) compared several
differences between recyclers and non-recyclers using socioeconomic factors. This study
found that age and education are significant factors for recycling. For instance, in a group
with the highest recycling average, the age group of recyclers was 40 to 64 while non-
recyclers were less than 40 years old. Recyclers also had at least 7 to 12 years of
education, which indicates more education than non-recyclers. However, there were no
significant differences between the groups on the basis of their income. The age factor
was also observed by Lake et al. (1996) who found that households with a middle aged
(45 to 64 years of age) head of the household have higher recycling rates. They explained

that in this age group, people spend more time engaging in recycling activities.

21

Lake et a1. (1996) also calculated the diversion rate, defined as the quantity of
household waste which is diverted from disposal to recycling. According to the analysis,
the average diversion rate per individual household (25.1%) exceeds the overall diversion
rate (22.9%). This result is explained as houses producing a large amount of waste are
recycling a lower percentage of their waste than houses with small total weights. When
all those households who recycle less than 10 percent by weight are excluded from the
analysis, the overall diversion rate rises to 29 percent and the average individual rate
increases to 32 percent as well. This result suggests that the participation of those
households who do not now take part in recycling is an important factor in developing
waste management policies.

Steuteville (1996) pointed out that recycling has apparently become part of
American life and a responsibility of local government. To verify this trend, Mullins et a1.
(1997) examined changes in local government practices. They surveyed 1071 cities with
populations of 5000 or above in the US. and found that population size is highly related
to waste services. For instance, larger cities (with more than 250,000 people) more
frequently offer solid waste pickups for residential waste collection. This study noted that
about 67 percent of the 975 local governments that provided information require
recycling by state law or local ordinance. In detail, 83 percent of all cities recycle
aluminum, 81 percent glass, and 76 percent plastic beverage bottles. For collecting
recyclables, 80 percent of surveyed cities offer curbside programs and 69 percent of
communities provide drop-off centers. This study also found that localities are not likely

to use a single method, otherwise they want to use not only multiple providers for

22

collecting materials but also multiple methods for disposing of solid waste. However,
they only used curbside pickup and drop-off centers as recycling methods.

To induce optimal recycling activity, Dinan (1993) examined several alternative
policies for the recovery of old newspapers (ONP). The policies discussed included a
virgin material tax, a combination of disposal tax and reuse subsidy, and a combination of
disposal tax and reuse subsidy with unit-based pricing, which is a direct method to charge
households by the amount of waste that they generate. This article shows that a combined
disposal tax and reuse subsidy policy could be an efficient method of reducing waste and
is suited to deal with recyclable waste items, while a virgin material tax is not. A
particular concern results from the unit-based pricing. Because households pay the same
amount for each unit of trash, illegal disposal increased. This study concluded that
implementing a combined policy provides more efficient recycling results than using just

one policy.

Environmental Attributes and Behaviors for Recycling

A number of recycling studies have focused on two behavioral approaches to
understand and explain people’s motives for recycling. Most groups emphasized intrinsic
motives for recycling behavior, while a few groups investigated the extrinsic motives of
recycling incentives. In an extrinsic study, Luyben and Cummings (1981-1982) explained
individual recycling behavior and audited the behavior of subjects. They found that the
combination of information fliers, lotteries and contest prizes is likely to increase
recycling in a dormitory system. In addition, Goldsby (1998) explained the extrinsic

motive as the most consistent, strongest influence on behavioral reaction. He considered

23

convenience as one of the important extrinsic motives and found that 43.5 percent of
respondents would like to recycle more if a more convenient system was provided. This
study shows the direct effect of extrinsic motives on recycling behavior and also
demonstrated that such external factors help to improve recycling.

On the other hand, intrinsic motives are defined as the characteristics inherent
within the individual that lead his behavior (Guagnano et al., 1995). Research on intrinsic
motivation has suggested that a good deal of human behavior is best explained in terms
of goals and rewards that lead participants to maintain recycling activity (Deci, 1975;
Eckblad, 1981).

De Young (1985-1986) found that intrinsic motivation and personal satisfaction
are the most dominant factors in people’s decision to recycle, while monetary reward is
not an important reason. Likewise, Resource Recycling (1982) reported that economic
rewards have only the slightest effect on people’s willingness to recycle. In addition,
studies of behavior-change strategies have suggested the importance of attitude change to
achieve people’s substantial behavior change (Burn and Oskamp, 1986; Katzev and
Pardini, 1987-1988). Hopper and Nielsen (1991) insisted that recycling is a form of
altruistic behavior and any type of social motivation will substantially increase recycling
efforts.

In a comparison study of recyclers and non-recyclers, Vining and Ebreo (1990)
presented knowledge, motives, and demographic factors. They found that non-recyclers
are more concerned with financial incentives, rewards, and the convenience of recycling,
while concerns for the environment are equal between the two groups. De Young (1986)

indicated that changes in recycling behavior motivated by monetary incentives do not last

24

when the reward is removed, suggesting that on-going forms of economic incentives are
necessary to make long-term changes in recycling behavior. This study suggested that a
waste management policy should be selected to encourage non-recyclers to begin
recycling. This study provides the relative importance of different motivational factors;
mandatory deposit laws such as beverage container deposit legislation are not considered
in the analysis.

To determine characteristics affecting the successful management of different
recycling programs, F 012 and Hazlett (1991) conducted a national survey. This survey
indicated that mandatory recycling programs attain participation and diversion rates twice
as high as voluntary programs. Higher participation also occurred in cities that have more
expensive waste collection fees, which are used for financing the operating costs of the
recycling program. This result suggests that more residents are motivated to recycle due
to the higher fees. Unlike previous studies, this study observes that recycling success is
not related to socioeconomic parameters but depends more on the policies chosen.
Therefore, they insist that policy selection and its implementation are the most important
factors for the success of local recycling programs, regardless of the per capita income,
education, or any other characteristics. This finding supports the study by Thomas (1990)
and the position that citizens are more likely to participate effectively in collection efforts
when they have been party to the policy decision.

Lansana (1992) indicated that the attitudes of residents toward the recycling
program are significant predictors of their behavior. In particular, the recyclers are
willing to recycle additional materials. For instance, overall, 92 percent of recyclers are

willing to add new items to be recycled, while 97 percent of the non-recyclers are not

25

willing to participate in the recycling program. This finding suggests the need for careful
planning and implementation of recycling programs when a new strategy is developed. In
behavior patterns, recyclers prefer mandatory programs because they assume that the
policy will ensure that people in the community recycle. In addition, many households
also prefer the use of curbside programs rather than the use of drop—off centers because of

easy access to the collection point.

THE EFFECTS OF BOTTLE COLLECTION PROGRAMS

With concern from states about recovering beverage bottles from the waste stream,
a number of studies have estimated the effects of bottle deposit laws. As an early study of
the impact of mandatory deposits, Porter (1978) estimated the social benefits and costs of
mandatory deposits on beer and soft drink containers in Michigan. The social benefits of
a change to mandatory deposits were increased control of litter and solid waste and the
cost savings. In a similar point of view, Porter (1983) also evaluated the economic
efficiency of mandatory deposits in Michigan. This study confirmed the previous study in
that solid waste pickup and disposal costs were saved and amenity benefits from litter
reduction were positive because of the mandatory deposit legislation. However, these
results cannot be generalized to other states since these costs depend mainly on the return
rate of deposit containers in each state.

Due to this limited application, Naughton et al. (1990) applied a methodology for
analyzing the impacts of bottle bills in California. They found that return value is more
important than convenience, while Porter (1983) found that the convenience of return is

as important as redemption values. With the same five cent deposit level, they also

26

indicated that the level of urbanization is one of the key factors. For instance, rural states
such as Vermont, Maine, and Oregon have higher return rates than more urban states
such as Massachusetts and New York. To estimate the net cost impacts of bottle bills on
consumers, three types of redemption values, such as 1 cent, 3 cents and 5 cents, were
used. Among the three values, the 5 cent redemption value was recommended to achieve
80 percent recycling. In addition, overall beverage container returns increased from 40 to
57 percent in California. This increase was mainly caused by the refund system, and a
decrease in both beverage container litter (by 29 percent) and beverage container solid
waste (by 17 percent) also resulted.

In previous studies, general equilibrium models were built to explain household
waste generation and disposal (Dinan, 1993; Jenkins, 1991; Morris and Holthausen,
1994). Fullerton and Kinnaman (1995) included illicit burning and dumping in their
model, with an assumption of no tax on the garbage. When applying the various
parameters, they found that a negative impact caused by garbage can be corrected by a
tax on garbage, while a tax on virgin materials is not useful for encouraging recycling. If
illicit burning is added as a third option, the tax on garbage turns out to be insignificant.
Furthermore, the model shows that the downstream tax system (at the point of disposal)
turns out to be the most efficient policy, while the upstream tax (at the point of
production) is easier to implement. Finally, this study found that the optimal regulatory
structure is a deposit-retum system because it is a tax on all output plus a rebate on
proper disposal. In this study, a variety of consumption goods are included in the analysis

rather than considering beverage bottles as the only item.

27

As one cost-effective way to encourage household recycling, Reschovsky and
Stone (1994) applied market incentives which replaced the common flat-fee pricing for
waste disposal with quantity-based pricing. They found that the most effective waste
management policy, among eight policy combinations, is the combination of mandatory
recycling and curbside pickup. The study suggests that households are more sensitive to
the private costs of waste reduction and less sensitive to the costs of waste disposal. This
finding is consistent with other studies (Vining and Ebreo, 1990; Markowitz, 1991). In
addition, they demonstrated that efforts to encourage recycling by imposing high
quantity-based fees or by enforcing mandatory recycling without providing a convenient
way to recycle cause households to regard recycling as a meaningless activity. It was also
found that curbside programs show the greatest effect on reported recycling behavior
when they are coupled with the other programs, such as the combination of mandatory
recycling and curbside pickup.

In a recent study, BEAR (Business and Environmentalists Allied for Recycling,
2002) presented a report on beverage container recovery programs, using data from 1999.
When various recovery programs are rated, deposit systems show the highest level of
recovery. For example, the average redemption rate of deposit states is 78 percent,
varying from a high of 95 percent in Michigan to a low of 72 percent in Massachusetts. In
a regional comparison of all types of recovery programs, the ten deposit states achieved
an overall recovery rate of 71.6 percent, while the recovery rate in non-deposit states was
27.9 percent. This result shows that there is a big difference between those two systems.
The second highest level of recovery is curbside programs, and the last is drop-off centers.

In deposit states, curbside programs collect 9.5 percent of containers generated in

28

residences and drop-off centers collect 1.6 percent. These rates increase to 18.5 percent
and 4.5 percent, respectively, in non-deposit states because more containers are available
than in deposit states.

In terms of the cost issue, Okamoto (1991) pointed out that deposit laws are costly
especially in that consumers bear economic costs related to the time required to return
containers and collect deposits. In addition to the comparison of recovery programs,
BEAR (2002) estimated the costs and revenues associated with recovery programs using
the value chain assessment, which includes all entities involved in the production, sale
and consumption of beverages and in recovering, processing and recycling beverage
containers. As the same trend with the recovery activities, deposit systems have the
highest gross cost. Curbside programs are the second highest and the lowest cost is drop-
off programs. Gross costs for deposit systems are 3.16 cents per container, and net costs
including revenue from material sales are 2.21 cents. Despite the high cost, it is pointed
out that deposit systems yield the highest quality containers, providing the highest market
values. The redemption system in California has the lowest net operating costs because

distributors and retailers do not handle containers and redeemed containers are not sorted

by brand.

CONCLUSION OF LITERATURE REVIEW

The review of previous studies shows that there are some options for various
economic incentives for recycling activities. Given the range and diversity of people’s
opinions toward recycling and other waste disposal methods, policy makers should seek

proper strategies that will enhance the management of MSW issues in their state or

29

community. The switch in policies and the creation of programs to develop alternative
waste management strategies will fundamentally change the way waste is discarded (EPA,
2001). From this point of view, recycling programs should be efficiently designed to
collect the majority of available materials generated in the waste stream. Otherwise,
recycling goals will not be achieved and large amounts of recyclables will be introduced
to landfills or combustion again.

Optimal recycling programs are likely to vary considerably from community to
community; thus results from previous studies cannot be generalized to other
communities. Therefore, in most cases, the results are regarded as suggestive and
compared to the experience of other communities. Despite various limitations, it is
necessary that the level of information is improved to assess the effectiveness of
recycling programs. This is mainly because recycling is a feasible part of the waste
disposal structure for some materials, especially for beverage containers. Furthermore,
this effort can provide substantial influence on recycling activity since as the markets for
recyclables mature, other materials will be economically feasible to recycle as well. This
research will seek to improve the level of information available by an economic analysis

method.

30

CHAPTER 3: METHOD

INTRODUCTION

This chapter reviews the framework of the data set used in the study and then
examines the estimation method, hypotheses and regression models developed and used
in the study. First, the estimation method of the study is explained. Second, the
hypotheses and associated variables are defined. Finally, the regression models are
discussed. Several regression models were constructed to determine the significance and
importance of variances of interest. In the first part of the regression models, the factors
that influence recycling activities are identified. In the second part, the effects of different

deposit incentives imposed on packaging are estimated.

RESEARCH DATA

This study focuses on the macro aspects of recent recycling efforts, from 1989 to
2000. According to EPA (2002), recycling activity was not significant before the 19803
and a rapid increase in the infrastructure for recycling started in the late 19805. In
addition, BioCycle started the annual national survey to deal with solid waste
management practices at the state level in 1989. Therefore, the given period of time was
used for the study. The analysis was carried out at the state level which includes 50 states
and the District of Columbia. Each state has a 12 yearly time periods for the independent
variables. Therefore, total observations for a variable are 612 since the matrix of data for
one variable is 12 by 51. For the analysis of the study, a statistical program called “Stata

version 7” was used. There are some missing values in the variables. For the treatment of

31

the missing values, cells are left empty in the data. Then Stata will automatically deal
with the missing values (Hamilton, 2003).

The main data source for waste-related variables such as recycling rate, MSW
generation and curbside programs was BioCycle reports titled “The State of Garbage in
America.” The annual report compiles waste stream estimates in the United States. The
data for MSW per person is calculated using both MSW generation and population data
since no direct data on personal waste generation by state is provided. For the curbside
programs, the reported number of programs within the state was selected as the variable.

The BioCycle data are useful to this analysis in two ways: First, as the EPA
(1995) pointed out, BioCycle’s data covers more waste streams than EPA’s annual
update. Second, data are reported by state, including all states in the United States. In the
report, the recycling rate includes all materials generated from the municipal solid waste
stream. Finally, there are two reasons for using BioCycle’s recycling rate for this study:
First, little other recycling data at the state level is available. Second, the effect of various
bottle deposit systems can be easily calculated in terms of how they affect the overall
recycling activity.

To measure the population and metropolitan area population, data are used from
the annual report of the US. Department of Commerce (2001) titled “Statistical Abstract
of the United States.” As a dummy variable, metropolitan area population is used as a
measure of the extent of urbanization. According to the US. Department of Commerce,
the average value of metropolitan and consolidated metropolitan area population is about
80 percent for the last decade in the United States. Because precise measures of

urbanization of each state are not available, the average value of the metropolitan

32

population is used as a proxy for urbanization. Therefore, if the percentage of the
metropolitan area population in a state surpasses the national average, it is noted as 1,
otherwise it is zero. Among the 50 states and the District of Columbia, 19 states are
higher than the national average. In addition, data for personal income are from the US.
Department of Commerce (2001) report titled “Survey of Current Business.”

Several categorical variables are used to measure the differences between the
deposit states and non-deposit states and between the five cent system and other deposit
systems. The dummy variables describe a shift in the y-intercept or constant. In the
regression for overall recycling trend, the dummy is named DEP. DEP equals 1 for
deposit states and zero in non-deposit states.

In the comparison among three deposit systems, two systems are treated as
dummy variables since one of these will be redundant. If all three dummies are included,
Stata will automatically drop one because multicollinearity otherwise makes the
caiculati’orrimpossible (Hamilton, 2003). The dummies are named FIVE and REDEEM.
FIVE equals 1 in the 5 cent system and zero in all other deposit systems and REDEEM
equals 1 for the redemption system and zero otherwise.

All 5 cent deposit systems are not the same in terms of included containers and
deposit amounts for liquor. Despite the difference, all 8 states are classified as 5 cent
deposit states in most studies. One of the main reasons for this classification might be that
separate data for each container type are not available. In the study, it is assumed that all
5 cent deposit states have the same bottle deposit system.

In California, a variety of deposit items were added after 1998 (Appendix A).

Although the alteration may cause a structural change in the recycling activity, it cannot

33

be included in the analysis because of the lack of data. Therefore, it is assumed that the

alteration does not influence the regression analysis of the study.

Except for the categorical variables, the data summary for all variables used in the

analysis is presented in Table 4.

Table 4. The Summary Statistics of Variables

 

 

Variables Mean Std. Dev. Min Max
RECYCLE' 19.18 12.07 0 59
Mswp2 1.126 0.349 0.42 2.66
POP’ 5131.78 5680.69 454 33872
CURB" 136.15 224.99 0 1500
INCOME’ 22183.85 4938.26 11915 40870

 

Note: 1: percentage
2: tons per person
3: thousands
4: numbers
5: dollars
a: sample mean

34

THE MODELING METHOD OF ESTIMATION

To estimate the set of equations proposed in the study, panel data are used. Data
are formed by the pooling of observations for the 51 states, including the District of
Columbia, over 12 time periods. Since panel data consist of both cross-section and time-
series data, there are a variety of benefits for evaluation of the effects of economic
programs. For instance, panel data allow us to control for each state’s heterogeneity and
to construct and test more complicated behavioral models than pure cross-section or pure
time-series data (Hsiao, 1985; Solon, 1989). In addition, variables can be more accurately
measured because data are usually collected on micro levels (Klevmarken, 1989).

To estimate the values of unknown parameters on the basis of observation, the
statistical technique of regression analysis will be applied. In the analysis, it is expected
that the data follow the statistical properties of ordinary least squares (OLS) estimates
under the normality assumption. Before the regression results are accepted as decision
criteria, a diagnostic approach will be conducted to validate the regression analysis.
When any potential problems, which cause any reason to distrust the regression results,
are found in the initial regression analysis, a specialized regression method will be

provided to make the regression results valid.

ESTIMATION FOR FACTORS AFFECTING RECYCLING
Hypotheses

Using various socio-economic variables, previous studies have tried to identify
the most important factors for recycling activity. Most of them have attempted to find

relationships between waste generation and other variables which are assumed to

35

influence MSW generation. These studies result in determination of a positive or a
negative correlation between factors. In most studies, data were from a random survey of
either households or officials who are working in the related area. Only a few studies
have focused on the relationship between recycling rates and socio-economic variables.

In the study, an economic analysis is conducted to evaluate the correlation
between recycling rate and a set of independent variables which were most often used in
other studies. For this analysis, data were collected through a variety of economic and
statistical reports. Data used in this study present people’s reactions to the country’s
recycling efforts.

The expected causal relationship between recycling rate and a set of explanatory
variables is hypothesized as follows. First, the generation of MSW per person has been
considered as one of the most important factors for recycling activity in the country
because recycling activity is related to personal behavior, especially in packaging bottle
recycling. In this study, it is assumed that the recycling rate increases as personal waste
generation goes up. Thus, it can be hypothesized that:

H1: The amount of MSW generation per person will have a positive correlation

with the recycling rate.

There are generally four kinds of possible relationships between population and
waste generation. First, waste generation increases as population goes up. This effect is
common and can be seen in most areas. A related relationship is that waste generation
decreases as population goes down. Third, as population increases, waste generation
decreases. This indicates that some consumer behavior has changed that affects waste

generation. This is a favorable situation since one of the main goals of waste-related

36

policies is to modify consumer behavior. The last case is that as population decreases,
waste generation increases. However, these relationships described above can differ when
recycling is considered instead of waste generation. In this study, it is assumed that
increased consumption by population growth has not increased recycling rates. Therefore,
it is assumed that the recycling rate is not related to the population change although waste
generation is positively related to population in general. Thus, the second hypothesis
states that:

H2: There is no correlation between population and recycling rate.

Income is related to people’s purchasing power. In a review of the interaction
between income and demand for products, Varian (1993) explained that:

“In general, when income goes up, the demand for a good could
increase more or less rapidly than income increases.”
In addition, according to economic theory, population growth will result in consumption
of more products since the entire demand curve shifts, causing the equilibrium price and
quantity to rise.

This explanation confirms that income can be positively associated with waste
generation. Besides, higher-income people are on average more educated than lower-
income people and so they can be expected to be more concerned about environmental
issues. With these factors considered, it is expected that higher-income people are more
likely to recycle than lower-income ones. Thus the third hypothesis states that:

H3: There will be a positive correlation between income and recycling rate.

The design and management of recycling programs is affected by the total number

of residents. Although it cannot be said simply that more populous areas produce more

37

waste than do less populated areas, an increase in density of the resident population will
produce rather different recycling outcomes. To indicate the degree of urbanization of
each state, population density is used as an independent variable. In this study, the focus
is only on highly populated areas.

Urban areas generally rely on standard methods of MSW pickup and disposal.
Even though people who live in populated areas know about the recycling programs in
their community, they may be less sensitive about recycling because it is easy to neglect
in such crowded areas. In addition, Jenkins (1991) explained that greater urbanization
leads to a lot of packaging waste because small quantities were purchased frequently due
to easy access to retail stores and little storage space. Thus, the fourth hypothesis states
that:

H4: Recycling activity will have a negative association with metropolitan areas

where population density is generally higher.

As a waste policy, bottle bills target only packaging products, especially beverage
bottles. As explained, this mandatory legislation has a very strong effect on the
composition of the MSW stream. Including California, currently 10 states have bottle
bills and others have different recycling programs. All bottle deposit systems, including
the redemption system, are expected to have a positive impact although they have some
different characteristics. Because bottle bills are mandatory legislation, they are expected
to directly affect the recycling rate. Thus, the fifth hypothesis states that:

H5: There will be a positive association between bottle deposit laws and the

recycling rate.

38

In general, curbside programs are assumed to be less effective than bottle deposit
laws since they are not based on a pecuniary motive and most of them are voluntarily
operated in many communities. However, as a substitute for bottle deposit laws, curbside
programs are found in most states throughout the country, even in bottle bill states. In
addition, it is found that this recycling system has also influenced consumer behavior.
Thus, the sixth hypothesis states that:

H6: There will be a positive correlation between the recycling rate and availability

of curbside programs.

Regression Model
A simple regression model was constructed to determine the significance and
importance of variables of interest for recycling activity. The testable regression model
can be presented as follows:
RECYCLE = a + [SIMSWp + BzPOP + B3INCOME + [34METRO + BsDEP
+ I36CURB + e (1)
where: MSWp = MSW generation per person (tons per year)
POP = population of each state (thousands)
INCOME = per capita income (dollar per year)
METRO = dummy variable for the metropolitan population areas
DEP = dummy variable for deposit states
CURB = the number of curbside programs

e = residual

39

The first independent variable tests the impact of each person’s MSW generations.
The second variable examines the effect of population differences. The third variable
tests the impact of personal income. The fourth variable evaluates the effect of
urbanization; it is 1 if a state has more than 80 percent urbanization on average and zero
otherwise. The fifth variable examines the effect of implementing bottle deposit laws; it
equals 1 if a state has bottle bill laws and zero otherwise. The last variable tests the

impact of curbside programs.

ESTIMATION FOR FACTORS AFFECTING RECYCLING RATE IN THE
BOTTLE REFUND STATES
Hypothesis

A variety of economic instruments, including variable fees and taxes, have been
applied in waste-related policies. To achieve an efficient solution by allowing consumer
discretion, various incentives have been introduced. The typical instruments are bottle
deposit laws and curbside programs in the packaging sector. In terms of economic
measures, the relationships between these two variables are defined in the previous part
of the study. In this analysis, in contrast, bottle refund states will be the focus to measure
the effect of different refund rates for recycling. Therefore, only data for the 10 bottle
refund states will be used for the study.

As described before, currently three deposit rates are used in 10 states: 5 cents, 10
cents and redemption (2.5 cents). Due to the different refund rates, the effect of the
refund systems for recycling can be different even though all bottle refund systems are

mandatory. From an economic point of view, it is expected that a high deposit rate is

40

more effective than a low rate because it can induce more consumers to follow the bottle
legislation. Therefore, it is assumed that among the three refund systems, the 10 cent
refund system is more effective than other refund systems.

To verify the impact, a regression analysis will be applied. In the model, the 5
cent system will be set as a target variable. Since currently many states adopt the 5 cent
refund rate, it will be tested as a common method for recycling. Other refund systems are

used for the control variable to explain the regression model. Thus, it can be hypothesized

that:
H7: The correlation rate with the recycling rate in the 5 cent deposit system will
be less than the recycling rate in the 10 cent refund system.
Regression Model

A multiple regression model will be introduced to test hypotheses concerning
recycling variation in the 5 cent refirndl states. Most of the independent variables have
been examined and established as important determinants of recycling activity in
previous studies. Even though the 5 cent deposit rate is widely accepted as one of the
primary deposit systems, the recycling effect of this level has not been extensively
examined. By introducing dummy variables, the regression model is expressed as
follows:

RECYCLE = a + [31MSWp + BZCURB + B3FIVE + B4POP+ leNCOME
+ BGMETRO + B7REDEEM + e (2)
where MSWp = MSW generation per person (tons)

CURB = the number of curbside programs

41

FIVE = dummy variable for 5 cent refund states

POP = population (thousands)

INCOME = per capita income (dollar per year)

METRO = dummy variable for the metropolitan population areas
REDEEM = dummy variable for the redemption state

e = residual

The third variable is 1 if the states have the given deposit rate and zero otherwise.
The null hypothesis (B3 = 0) implies that the 5 cent deposit system is not significantly
different from the 10 cent refund system. Other independent variables are used as control
variables and they have the same conditions applied in the first regression (Equation 1). If
some of the independent variables become invalid determinants, another regression

method will be employed to provide a statistically significant regression analysis.

42

CHAPTER 4: RESULTS

INTRODUCTION

This chapter presents the results of the research. Each section discusses the
empirical analyses that assess the variable characteristics. In each regression analysis,
diagnostics will be conducted for model validation. This discussion is followed by a

summary of the research findings.

RESULTS OF ANALYSIS
Estimation for Factors Affecting Recycling

A multiple regression model was used in order to test factors that are related to
the recycling rate. Empirical results of the regression analysis are reported in Table 5.
The proposed regression model explains about 50 percent of the variation in the recycling
rate of 51 states. The variables POP and METRO have a negative relationship to the
recycling rate while others are positive. In addition, the variables POP and INCOME
have an extremely low effect on the dependent variable. Diagnostics will be conducted to
verify the result in terms of statistical properties before this result is considered as the
final regression model.

Table 6 is the matrix of correlations between variables. This matrix is used to
check for the evidence of collinearity between the independent variables. In general, high
correlations between pairs of variables indicate possible collinearity problems. The

variables CURB and POP are the only pair in which correlation is over 50 percent.

43

Table 5. Test of the Factors of Recycling Rate

 

 

Independent [3 Coef. t P> | t |
Variables
MSWp 8.8152 7.65 0.000
(1.1518)
POP -0.0002 -2.49 0.013
(0.0001)
INCOME 0.0010 1 1.45 0.000
(0.0001)
METRO -2.93 90 -3.12 0.002
(0.9431)
DEP 3.6400 3 .91 0.000
(0.9298)
CURB 0.0183 8.91 0.000
(0.0020)
CONSTANT -14.6054 -7.50 0.000
(1.9467)
R2 0.50
F
93.87

 

Note: parentheses are standard errors

significant at a = 0.05
Prob > F = 0.0000
obs. = 579

44

Table 6. Estimated Correlation Matrix of Variables

 

 

MSWp POP INCOME METRO DEP CURB
MSWp 1.0000
POP 0.3367 1.0000
INCOME 0.2789 0.1950 1.0000
METRO 0.2166 0.4934 0.4555 1.0000
DEP 0.1326 0.2191 0.2234 0.2239 1.0000
CURB 0.1292 0.5382 0.3957 0.2913 0.2052 1.0000

 

 

High correlation of over 40 percent is presented in the pair of variables METRO and POP
and the pair of variables METRO and INCOME.

The correlations imply that the variable POP or METRO might not be a suitable
parameter for the regression model. Because in most cases the independent variables are
correlated with each other, removing a particular variable from the model will alter the
estimated regression coefficients of the remaining variables. Therefore, a stepwise
regression, which has been widely used for providing ways to automate the process of
model selection, is applied to choose any variable to eliminate. The result of the stepwise
regression indicates that no variable should be dropped from the model (Appendix B),
thus it is decided that all proposed independent variables will be used for the analysis.

In using cross-sectional data, homoscedasticity or equal variance is one of the
most important assumptions since it implies that all independent variables are equally

important. To detect heteroscedasticity or unequal variance, the estimated residuals are

45

plotted against the estimated recycling rate from the regression line (Figure 4). The
scattergram shows that the variance of the error term has a systematic pattern. In
particular, the estimated recycling rate has a linear association on the lower part of the

graph, suggesting a symptom of unequal variance in the data.

 

 

 

 

 

 

Figure 4. Residual vs. Predicted Value for Overall
States
40
3O _
o
o, .
20 . 99:.
9 e
.3. . ‘;.I‘ 9 0’
g 10 - O Q ' ‘ 0‘.
9 I A, O o
E g .- o 3
3 o,” Isrx.‘,4 ’.
m ‘Q.‘§ ‘9’ . ~ r v r r
. . . . ’ . ..
~10 I ffi . 0’.
o
.. o 0% : 9 ° .9
-20 ”o I 0’ 0 °
-30
0 10 20 30 40 50 60 70
Predicted Value

 

 

 

46

Due to the systematic pattern between the variables, the Breusch-Pagan test was
also used to confirm whether the presence of the unequal variance was significant (Table
7). The Observed chi-square value of 20.85 is significant at the 5% level of significance
because it exceeds the critical chi-square value of 3.84 from the chi-square table. Because
this result confirms that there is unequal variance in the regression, the data should be

transformed in such a manner that in the regression the variance of the residual is equally

spread.

Table 7. Breusch-Pagan Test for Overall Recycling Rate

 

Ho: Constant variance

Chi2(1) 20.85

Prob > chi2

0.0000

 

The method Of generalized least squares (GLS) is applied in the transforming
method. GLS is a special case of OLS on the transformed variables that satisfy the
general assumptions of the classical model. The difference between GLS and OLS is that
equal weight or importance to each observation is assigned in GLS by making the error

term a constant (Gujarati, 1995). To use GLS, the method of theoretical change will be

introduced from OLS to GLS.

47

In general, the two-variable model is described as follows:

Y1: 0X01+I3X11+ Ci (3)
where Y, is the dependent variable, Xm is the explanatory variable for each i, or is the
intercept, B is the slope coefficient and e, is the error term. When the original variables
are divided by the variance, Equation 3 can be written as

Yi‘ = (I. X01. + I3t X11. + 61‘ (4)
where transformed variables are marked by asterisks. Equation 4 indicates that the
variance of the error term is homoscedastic or equal variance. Therefore, the estimated [3.
is now the unbiased estimator, providing true values.

Because the data used in the study combine all cross-section and time-series units,
finally the two-variable model would be:

Yit=u+ BXit+eit fori= 1,2, 3, ------- ,N (5)
t=1’2’3’ ....... ,T

where N is the number of cross-section units and T is the number of time periods. With
one large pooled data set, the regression is run with NT observations.

The GLS result is presented in Table 8. All values are somewhat changed,
compared to the OLS result in Table 5. The p value of the variable METRO is significant
at the 0.021 percent level while others are significant at the 0.0001 percent level. The
result indicates that the values of the t-statistic lie in the critical region, thus the test
statistic is statistically significant.

The first independent variable MSWp tests the effect of the MSW generation per
person on recycling rate. The hypothesis, that a positive association exists between the
two variables, can be accepted. The estimated parameter of 8.77 indicates that this

variable significantly affects the recycling rate among the other independent variables.

48

Table 8. The Result of GLS Regression for Overall Recycling Rate

 

 

Independent B Coef. z P> | z |

Variables

MSWp 8.7745 7.48 0.000
(1.1725)

POP -0.0003 -3.53 0.000
(0.0001)

INCOME 0.0010 10.91 0.000
(0.0001)

METRO -2.0891 -2.31 0.021
(0.9025)

DEP 3.8304 4.31 0.000
(0.8884)

CURB 0.0207 9.78 0.000
(0.0021)

CONSTANT -14.4344 -7.69 0.000
(1.8764)

 

Note: parentheses are standard errors

significant at a = 0.05

z is a transformed t value
P> | z|isapvalue forz
Obs. = 579

49

The coefficient value 8.77 implies that by holding other variables constant the recycling
rate on average increased by 8.77 percent for every one percent point increase in the
MSW generation of each person over the period 1989-2000.

The second independent variable POP tests the effect of the population change.
The hypothesis presuming no relationship between population and recycling rate can be
accepted. The association of the two variables suggests that population growth will have
no effect for recycling over the country but its impact will not be practical significance.

The third independent variable INCOME tests the effects of per capital income
on recycling rate. The hypothesis stating a positive association between the two variables
can be accepted. However, a 1 percent increase in income causes the recycling rate to
increase on average by only 0.001 percent. This result indicates that income is not a
significant variable for recycling.

As a dummy variable, the fourth independent variable METRO tests the effect of
population density on overall recycling rate. The hypothesis proposing a negative effect
can be accepted. This result shows that recycling is less effective in highly urbanized
areas.

The fifth and sixth variables DEP and CURB test the effect of different recycling
systems. The hypotheses of positive recycling rate can be accepted for both variables.
The coefficient of bottle deposit laws is 3.34 and the curbside collection is 0.02. This
result suggests that the bottle deposit system is more significant than curbside programs

for recycling.

50

Estimation for Factors Affecting Recycling Rate in the Bottle Refund States

To compare the difference among the three deposit systems, a multiple regression
was conducted (Table 9). The regression result shows that the proposed model explains
about 61 percent of the variation in the recycling rate. Most variables have a positive
association with the recycling rate except for the variables POP and METRO. The
variables CURB, POP and REDEEM have relatively higher p values. In addition, the
correlation between the independent variables shows that the variables POP and
REDEEM are associated over 50 percent with some other variables (Appendix).

The results indicate that it should be decided whether any variable is not
necessary for the regression model. The stepwise regression, therefore, is applied to
detect unnecessary variables in the model (Table 10). The result reveals that the
independent variable POP should be removed from the model. The stepwise regression
shows that the p value of the target variable FIVE is reduced to become statistically more
significant. However, the variables CURB and REDEEM still have somewhat higher p
values. These values are control variables, thus they will be kept in the model unless
regression diagnostics show they are not necessary.

Except for the variable METRO, the signs of the independent variables are
positive. The variable MSWp is the major factor with the parameter coefficient of 18.3
while the variables CURB and INCOME have extremely small effects on the dependent

variable. This trend is the same as shown in the first regression for overall recycling rate.

51

Table 9. The Result of Regression for the Deposit States

 

 

Independent B Coef. t P> | t |

Variables

MSWp 18.313] 7.91 0.000
(2.3158)

CURB 0.0033 0.99 0.324
(0.0033)

FIVE 6.4261 2.46 0.016
(2.6139)

POP -0.0000 -O.13 0.894
(0.0002)

INCOME 0.0011 6.36 0.000
(0.0002)

METRO -l3.8337 -6.38 0.000
(2.1680)

REDEEM 2.5002 0.48 0.633
(5.2269)

CONSTANT -21.3088 -4.70 0.000
(4.5363)

R2 0.61
F 24.94

 

Note: parentheses are standard errors
significant at a = 0.05
Prob > F = 0.0000
Obs. = 119

52

Table 10. The Result of Stepwise Regression for the Deposit States

 

 

Independent B Coef. t P> | t |
Variables
MSWp 18.2993 7.94 0.000
(2.3033)
CURB 0.0030 1.27 0.207
(0.0023)
FIVE 6.5071 2.57 0.011
(2.5318)
INCOME 0.0011 6.59 0.000
(0.0002)
METRO -13.9782 -7.47 0.000
(1.8724)
REDEEM 1.9236 0.65 0.516
(2.9517)
CONSTANT -21.4957 -5.00 0.000
(4.2975)
R2 0.61
F 29.35

 

Note: parentheses are standard errors
significant at 0r = 0.05
Prob > F = 0.0000
Obs. = 119

53

Before the stepwise regression is accepted as a final model, diagnostics are done
for the stepwise regression to verify the statistical validity. First, the variance inflation
factor (VIF) for each of the independent variables is calculated (Table 11). According to
the “rule of thumb” given by Chatterjee and Price (1991), there is no strong evidence of
multicollinearity since VIF values are smallerthan 10 and the mean VIF is not
considerably larger than 1.

To detect unequal variance, the estimated residuals are drawn against the
estimated dependent variable in Figure 5. The graph shows that the pattern of the
residuals is random. No systematic pattern between the two variables suggests that
unequal variance is not present in the data. In addition, the Breusch-Pagan Test (Table12)
confirms that the Observed chi-square value of 2.49 does not exceed the critical value of
3.84 at the 5 percent level. The tests indicate that there is no evidence of significant

unequal variance.

Table 11. VIF Test for Bottle Refund Regression

 

 

 

Variable VIF 1N IF
FIVE 2.44 0.4094
METRO 2.00 0.5007
REDEEM 1.87 0.5348
INCOME 1.84 0.5436
MSWp 1.58 0.6334
CURB 1.24 0.8034
Mean VIF 1.83

 

54

 

 

Residual

15

10

Figure 5. Residual vs. Predicted Value for Deposit

 

 

 

 

 

 

States
0
. O. O
Q 0 O.
0 . ‘ O. 9.
.’O .090. O
1 o ’0’ 0 ’0. ,
I . V W. .. l . I
O 0 O. .0: . Q.
0 0. ‘..
.09 ’ . 8
O
O O . . .
O
O
10 20 30 40 50 60 70

Predicted Value

 

55

 

Table 12. Breusch-Pagan Test for Bottle Refund System

 

Ho: Constant variance

(3102(1)

2.49

Prob > chi2 0.1148

 

In the comparison of the three deposit systems, the 10 cent refund rate is treated
as zero. The coefficient of the variable FIVE is 6.5 and it isl .9 for the variable REDEEM.
Since dummy variables are used to account for qualitative factors, this result says that
there is a very high correlation between the 5 cent refund system and recycling activity
which is about 6 times more association than with the 10 cent deposit system. The next
correlation is found in redemption which shows about 2 times more association than the
10 cent system. Thus, the hypothesis of more effectiveness of the 10 cent refund system
cannot be verified. This is presumably because the proposed independent variables may

be insufficient to prove the effectiveness of the 10 cent deposit system.

SUMMARY OF THE RESEARCH RESULTS

The results Of various analysis presented in this chapter generally supported most
research hypotheses, in spite of minor exceptions. Among the independent variables, it
was confirmed that the variable MSWp was the primary factor for recycling while the
variable METRO affected most negatively. Although expected hypotheses were mostly
confirmed in the variables POP and INCOME, the values of these parameter estimates
were turned out to be insignificant. In particular, the variable POP was deleted in the

regression model for deposit states. Only the assumption of difference between the bottle

56

deposit systems was not confirmed by the regression model. Table 13 shows the major

research hypotheses and results.

Table 13. Summary of the Research Results

 

 

Subject Hypothesis Result

MSW Generation Per Person Positive Confirmed
Population No Effect Mostly Confirmed
Income Positive Mostly Confirmed
Urbanization Negative Confirmed
Bottle Deposit Laws Positive Confirmed
Curbside Programs Positive Confirmed

5 Cent System Less Correlation Not Confirmed

 

57

CHAPTER 5: CONCLUSION

INTRODUCTION
This chapter presents the contribution of the research. First, implications of the
study’s findings are discussed. The discussions and limitations of the study will be

presented and the recommendations for future research will then be discussed.

CONCLUSIONS

Recycling has been developed as a major technology for taking care of the
nation’s waste. Every state of the United States has maintained various recycling
programs to deal with the MSW generation. For instance, anti-littering and recycling laws,
beverage industry recycling programs, Keep America Beautiful systems and other
recycling programs are used in the 41 non-deposit states (F DL, 1989). In this study, the
recycling rate was explained by several socio-economic factors. As major recycling
programs, curbside collections and bottle refund laws were compared.

Multiple regression analysis identified a significant association between the
recycling rate and the MSW generation per person. The relationship indicates that the
waste generation of an individual is the major factor for recycling. In particular, the
coefficient of the variable MSWp is much larger in the regression for bottle refund states
than in the regression for the overall states. From the finding, it can be assumed that
peOple living with a bottle refund system participate in recycle activities more than those

who live with other recycling systems.

58

In determining the recycling rate, demographic variables such as population and
metropolitan area population were negatively associated with recycling efforts. In
particular, the negative correlation between the recycling rate and urbanization was
higher in the 10 bottle refund states. As previous studies have shown, the finding suggests
that an increase in population or the total number of residents may cause more waste but
it cannot guarantee the participation of people in recycling programs. Therefore, it is
concluded that recycling effectiveness may not be increased by these demographic
variables.

As an economic variable, the effect of income turned out to be extremely small in
both regression models. The finding indicates that people of high income tend to recycle
somewhat more than those with less income, but there is no practical significance. The
result concerning income is consistent with some of the previous studies in the literature.

Recycling in the packaging sector has become a major component of MSW
management plans because packaging is directly implicated in two outcomes: litter and
solid waste. Due to this nature, packaging materials are frequently criticized as being a
main source of solid waste. In addition, many products are designed to be easily thrown
way for convenience of customers in these days. To minimize the environmental impact
of packaging and reduce waste generation, reducing MSW often means reducing
packaging materials. From this perspective, mandatory deposit legislation (or beverage
container deposit laws) has been adopted in nine states and California has used a
redemption system.

In the comparison of different deposit rates, it was confirmed that the three bottle

refunds have different effects on recycling. To distinguish the possible difference

59

between the three refund systems, they were separated by two categories, such as 5 cent
and redemption system. The coefficient of the 5 cent system is significantly higher than
other refund systems. The finding shows that the 5 cent refirnd system is a more effective
method for recycling because of the higher correlation with recycling rate, compared to
other deposit systems.

An interesting result, however, is the finding that redemption is the second most
effective method although it is the lowest refund rate at 2.5 cent, staying in between the
10 cent and 5 cent refund systems. Recall that California covers more items for deposit
than any other state. Thus, one of the possible reasons for the finding is that redemption
covers a wider range of containers. However, more descriptive variables that could Offer
an explanation are needed to verify this finding.

The examination of recycling methods revealed that bottle deposit laws are one of
the significant determinants for recycling. In addition, as mentioned, containers collected
through deposit systems have high quality, yielding the highest market value, compared
to other recycling systems. Therefore, the relative effectiveness of deposit systems will
also increase recycling rates for the covered containers. On the contrary, curbside systems
prove to be less significant as correlates of recycling efforts. This finding suggests that
current bottle deposit systems contribute significantly to recycling activities more than
any other recycling methods.

As a result, several factors should be noted: First, an individual’s behavior is the
primary variable in determining recycling success. Second, bottle refimd systems
positively affect the individual’s recycling activity which can lead people to recycle in

the long run. Three, a wider range of deposit items can lead people to recycle more than

60

an increased monetary valuation on the deposit containers. Four, beverage container
deposit laws imposed on the packaging sector turned out to be significantly effective for
recycling. Last, as a leading recycling method in the country, the bottle deposit system is
more effective than curbside programs.

This study examined the recycling activities in the country focused on the bottle
collection methods. The result provides how the recycling rate is affected by various
variables and what the difference is between the two bottle deposit methods, mandatory
bottle deposit systems and curbside programs. The results can be used for setting a bottle
recycling method and they are also used as basic information for waste management

approaches.

DISCUSSIONS AND LIMITATIONS OF THE STUDY

According to BEAR (2002), the redemption rate of recyclable containers in 1998
was 94 percent in Michigan (in 1999), 80 percent on average in 5 cent states and 69
percent in California. This report supports the assumption proposed by the study, which
the 10 cent system is highly associated with recycling rate compared to the 5 cent system.
However, the difference between the 5 cent and 10 cent system was not confirmed by the
proposed regression model. Instead, this study found the opposite result as the 5 cent had
a higher correlation with recycling rate.

This result can be explained as follows: First, the different state numbers in bottle
deposit systems cause an unexpected influence to the model. As described, only
Michigan has the 10 cent rate and California has the redemption system, while the other 8

states have 5 cent systems. If the shortcoming is resolved, the result can be more

61

acceptable with common knowledge. Second, people in Michigan recycle more of the
recyclable bottles but they recycle the total materials at a lower rate than is expected for a
deposit state. Therefore, the finding is that additional factors have to be considered in the
recycling study. Some factors influencing the lower overall recycling rate might be
relatively low landfill costs, lack of government support for recycling programs, a higher
reliance on voluntary rather than mandatory curbside programs, and less education and
publicity about recycling.

In an empirical study, the analysis is directly subject to the limitations of the data.
In particular, the accuracy and precision of data are the most important sources to
measure parameter estimation. Therefore, a number of sources can enter into the
regression model with uncertainty and error. In the study, one source of a possible error is
an inaccuracy in the data for curbside programs and recycling rate. There are some
inconsistent measures, and some data are under-reported.

The next possible source of error is that the included variables may be inadequate.
For instance, the finding that curbside programs were not as significant as deposit does
not necessarily mean that the variable is not a contributor to overall recycling activity.
This error can be minimized if the proper measure is used the model because it precisely
reflects the contribution of a variable.

A better predictive model would require adding other variables that are significant
contributors to recycling. As data for recycling becomes available in future years, the

model can be re-examined to reflect the recent recycling activities.

62

RECOMMENDATIONS FOR FUTURE STUDY

As an empirical study, this research is an initial attempt to examine some of the

causal factors that determine recycling efforts. It focused on several relevant parameters

that impact waste generation and collection. Future research can continue evaluations of

the recycling-based paradigm and develop the recycling model.

To deal with variability in recycling activities, follow-up research is

recommended in these areas:

1.

Conduct supplemental analysis for the three deposit systems to assess the
recycling effect of these systems to verify which system is most effective.
Examine the effect of mandatory deposit systems by comparison between deposit
states and non-deposit states.

Separate recycling rates by container type to include aluminum, glass, PET,

HDPE and other.

Assess factors which affect costs and revenue in operating recycling programs.

. Develop models for assessing the effect of energy savings and litter reduction as

substantial benefits from implementation of various recycling systems.
Expand the recycling models to include drop-off systems, non-residential

programs and buy-back centers.

63

APPENDICES

64

 

APPENDIX A

Table A-1 Summary of Deposit Containers

 

 

 

 

 

 

 

 

 

 

 

State Containers Deposit Rate
California soft drinks, beer, carbonated water, 2.5 cents < 24 oz
non-carbonated water and soda“, wine 5.0 cents Z 24 oz
and distilled spirit coolers“, fruit drinks*,
coffee and tea drinks"
Connecticut soft drinks, beer, malt, mineral water 5.0 cents
Delaware soft drinks, beer, ale, malt, mineral and 5.0 cents
soda water, Aluminum cans exempt from
deposit.
Iowa soft drinks, beer, soda and mineral water, 5.0 cents
wine, liquor, wine coolers
Maine soft drinks, beer, water, mineral water, 5.0 cents
wine, liquor, non-alcoholic carbonated 15 cents wine and
and non-carbonated drinks liquor
Massachusetts soft drinks, beer, carbonated water 5.0 cents
Michigan soft drinks, beer, carbonated water, 10 cents
mineral water, wine coolers, canned
cocktails
New York soft drinks, beer, malt, mineral water, 5.0 cents
soda wine, wine coolers
Oregon soft drinks, beer, malt, mineral and soda 5.0 cents
water
Vermont soft drinks, beer, malt, soda water, mixed 5.0 cents

 

wine drinks, liquor

 

15 cents liquor

 

Note: * added after 1998

Source: BEAR, 2002

65

 

Table B-1 Result of Stepwise Regression for Overall Recycling

APPENDIX B

 

 

Independent B Coef. t P> | t |
Variables
MSWp 8.8152 7.65 0.000
(1.1517)
POP -0.0002 -2.49 0.013
(0.0001)
INCOME 0.0010 1 1.45 0.000
(0.0001)
METRO -2.9390 -3. 12 0.002
(0.9431)
DEP 3.6400 3.91 0.000
(0.9298)
CURB 0.0183 8.91 0.000
(0.0020)
CONSTANT -14.6054 -7.50 0.000
(1.9467)
R2 0.50
F 93.87

 

Note: parentheses are standard errors

significant at a = 0.05
Prob > F = 0.0000
Obs = 579

66

Table B-2 Estimated Correlation Matrix of Variables for Deposit Systems

 

MSWp CURB FIVE POP INCOME METRO REDEEM

 

MSWp 1.0000

CURB 0.2944 1.0000

FIVE -04441 0.1245 1.0000

POP 0.4610 0.5203 -0.6893 1.0000

INCOME 0.3354 0.3512 0.0525 0.1679 1.0000

METRO 0.4914 0.1969 -0.4l33 0.4962 0.5289 1.0000

REDEEM 0.3207 0.2244 -0.6663 0.8480 0.0180 0.2754 1.0000

 

67

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88 88 88 88 88 88.8 88 ~88 88 88 88 88 898.22
88 88 88 :8 888 888 N88 N88 888 888 88 888 8886882
888 888 888 88 888 88 38 888 88 88 888 88 88.8.2
88 88 888 88 38 88 88 888 888 2.8 88 3.8 88.2
88 888 38 88 88 888 R8 888 88 88 88 88 8888..
88 88 88 888 88 88 88 88 88 38 88 88 82:3.
88 888 888 88 38 88 88 88 88 88 88 88 68:3.
88 88 88 88 38 888 88 888 38 88 88 88 838.
88 8.88 88 88 88 88 88 R8 88 88 88 888 88....
88 :8 888 N88 N88 88 88 88 88 R8 888 888 85...
38 88 88 88 88 88 88 888 88 88 38 88 9.8.
88 88 88 88 88 88 88 88 888 .688 88 888 838:
88 $8 $8 88 88 888 88 888 88 88 88 88 «@668
88 88 88 888 88 888 88 88.8 3.8 88.8 $8 88 88.“.
88 88 88 88 3.8 .88 88 88 $8 88 38 888 8528 86 .85
8.8 88 888 88 88 88 88 88 8.88 888 888 888 288.88
88 88 88 88 88 88 88 88 88 88 88 88 888888
88 88 88 88 £8 88 R8 38 888 {.8 888 38 88.6.8
88 888 :8 88.8 88 $8 38 88 3.8 88 88 888 958.8
2.8 88 888 888 88 88 88 88 888 38 88 88 888.2
N88 88 888 88 888 38 888 888 88 £8 88 88 868.2.
88 88 888 88 88 88 88 38 88 88 88 88 8.82
888 888 88 88 88 88 88 88 88 888 88 88 65882.
8888 8888 8888 8888 8888 8888 8888 8888 8888 8888 8888 8888
8888 8968 .68 205$sz 26: whim

 

 

 

U 5975;:

8382 6388.8 68 s8 88 To 6388

68

.88 Ba 82 8 88 68 .6 888 28 88 285 88> U.82.

 

 

0:. 08.8 08.8 08.8 0:. 3.? 00... 00.0 00.0 05.0 00.0 0Né wEEo>>>
00.0 05.0 50. —. 05.0 00. F 00. 8. 50. v 00.0 50.0 00.0 0v... V5.0 Ewcoomg
00.0 «5.0 0_.._. 0_.€ 0.? or; 08.? #00 0.0 00.0 00.0 00.8 was ~m0>>
0N._. 0:. 08.8 0...? 0N... 00.? mm... 09... :4 N0... 0:. 08.8 8.90550;
0.? 0_.._. 5»; 00.5 00.? 50.8 NNé 08.5 08.8 018 0v; 50.8 0.5%
00.0 No.0 00.0 N0.—. N0. F 00. v _.N. F N; 00.0 00.0 «0.0 00.0 Eco—:0)
00. _. Exp 00. _. «0.8 00.8 5‘? 00... #0.? «0.0 00.0 V0.0 «0.0 :33
0rd 05.9 P5.w N8... 0...? 08.? 00.8 00.8 N00 v0... 008. 00.8 «axe...
80.0 00. w 05. _. 55.8 N5é 0P... 0wé 00... 08.8 80.8 5.8 00.0 @3855...
00.0 05.0 00.0 00.0 3.8 3.9 08;. 088. «8.9 0:. «5.0 00. _. 80me 5300
N 5.8 min EN FNN Sam 00. F 00.8 50. P 00.« NF... N50 mp; 05:050 5300
00.8 N10 00.0 0v.0 00.0 3.5 50. w 0~€ 0N... 0a.? 00. P 00.? 0:03. 0005.
00.0 N00 55.0 05.0 00.0 05.0 05.0 05.0 05.0 00.0 55.0 05.0 £00252..ch
00.8 00.8 0w... 0:. 0:. 00... 00.? No.8 nré 0:. 55.0 «0.0 cambo
0 fir 00. F 00. v 05.0 05.0 05.0 55.0 00.0 v0.0 00.0 3.8 00.0 0:5:ch
0Né 50. w 08... 08.8 N0... 0rd 00d 00. w 0v... vvé 0N8. 0N? 030
00.0 05.0 05.0 00.0 05.0 05.0 05.0 05.0 05.0 00.0 05.0 05.0 90me 552
00.8 05. _. 50. w «0% 00... 006 08.? 00.0 36 00.0 00.0 80.0 05.0.60 552
#0... 00.8 00. 8. 00.8 0»; I: 01—. 00.8 0a.? NNé Nu; 3.8 fo> 302
00.8 05. 8 N0... 80.0 00.0 5 F... 3.8 00.0 «0.0 50.0 05.0 00.0 80822 302
00. F 00.0 00.0 N08 08; 50.8 00.0 00.0 00.0 0.0 80.8 0N6 5082. 302
00.0 50.w V5.0 N06 50.8 «0.0 —.0.0 00.0 No.8 00.0 00.0 00.0 9:23me 362
00. _. V54. 00. _. 00d 3d 00... 00.8 00.8 05.8 05.0 00.0 00.0 00902
00. 8. 00. _. 0N8 8N... rm... #0._. No.8 _.0.0 50.0 N00 05.0 05.0 9.3562
V0.0 0~.—. v8.8 0 fr 00.8 00.0 N00 00.0 _.0.0 845.0 05.0 05.0 29:22
#0... 05. w 9.8 8.8 0V8. 3.8 00. F 018 3.8 0.: 5 8.8 00. F 5.03:2
00. F «0.0 N..._. 00.0 00.0 00.0 00.0 00.0 V0.0 0.0 05.0 05.0 59.66632
0 F... 3.8 00. _. No.8 00.0 v0.9 80.? 50.0 00.0 00.0 00.0 N00 $39.55.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

86888 .8382 638.5 65 88 83 To 638

69

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88. 88 88 m8. .8. 88 88 8. 8. 8. 88 8.... 803.2
83 88 88 88 88 88 88 88. 8... 83. 88. 88. 87.285
88 9.8 88 8... 88 88 88 83 88 83 83 88 3885:
...m. 88. 8... 88. 88. 88. 8.8. 88. 88. 8... 8... 8... 895.2
38 we... 8. 8.. 8.. 88 8.8 8.8 88 88 88. 88 88:588.).
88 88 8.... 88 88 88 88 88 88 8.8 88. 8.. 298.2
88. m8. 88. a8. 88. m8. m8. 88. 8a. 8m 8m 8 gas.
.88 88 8... 38 8... 8.8 88 88 88 88 88 88 8328..
88 ..o.. 88 8.3 o8.” 88 8.8 .88 89. 88 88 88 8.03:8.
88 88 88 88 88 88 88 .38 88 88 8. 8. ”8:8.
88 88 28 88 .88 88 3.~ 88 88 88 o8~ 88 «so.
..8. 8.8 8.8 .... 88 88 88 83 88 8.... 88 83 «885
88. 8.8. 88. 88. 88. 88. 88. 8..... 8.... 88. 8.8. 88. 88...
8o. 3. .8 88 88 88 o8 8m 88 O8 o8 8. 9.8.
38. 38. 08. 88 8.... 88 o8~ o8~ 8. 8m. 8m. 8. 82...:
88. 83.. 8.9 88. «.8 88 88 88 88 83 83 83 mama
888 888 8.8 .88 ~88 ~88 .88 88. 83. 8.8. 88. 88. 88:
8 8m 8~ 98 .8 8m 8.. a... a... m... 8. 8. «5528 .o .85
88 o8~ 88 8.. .8 8w. 8.. 8o. 8. 8. 8.8 8 2858
88 8.8 .38 88 88 38 88 .88 88 88 88 88 .8888
88 88 88 88 :8 8.. 88 83 88 88 o8~ 88 08.28
8.8 888 .88 .88 83 83 88.. 883 A83 883 888 83 952:8
88 88 8.8 .83 88 88 88 88 88 88 o8~ 88. 888.2
8.... .88 «.8 8.8 88 83 8% 88 .33 88 8.8 8.8 88.2
88 m... 8.8 8 8 8 .8 8m. 8... Sn .8 83 9.82
88 2.8 88 88 :8 88 28 88 88 88 83 83 9882
88 a8. 8. .8. 88. m8. .8. 88. N8. .8. 8. 88.
.82 8.. 20.2%.sz Ems. whim

 

 

 

8:288 .82 s. 33 No 2%.

70

 

 

cme
comm
mmcm
mecv
mmm
mmvm
rmmv¢
comm
vwm
mmv¢
Pmmw
cmmwp
vvmw
mmmm
mmmwp
mmm
commP
ccwwm
mwvm
comm
mmcw
mmmm
mtmr
mmm

 

cmm
ccce
comm
mmmm
mm—m
mmm
mmmm
mmcwm
mwmm
vwm
movm
Fmv
comm
mwvc
mcmm
mrcmw
mmv
cocmm
cmmmm
mmmm
comm
vmmw
mmwm
cmmm
mmcF

 

cocm

coco?
cmm
cmvm
cmmmm
mwmm
owm
cwccw
omv
comm
ocwv
mvmm
mmmm—
Pom
mmmmw
ccmcm
cemm
comm
cmm
comm
cocm
wch

 

mmmm
cocm
mmmm

cmmm
mmmwm
mmvm
owm
mem
mmv

mmmm
comm
mmmmw
cwm
mtmm
commm
covp
comm
com—
mmmm
cocm
mmcr

 

cmm
vmvm
cocm
mvmm
mvmm

mmmm
mmmwm
mrem
mvm
mmmm

mmmm
vmmm
comm
mvvww
mom
mmmm
commm
mmmw
cmmm
mvmw
mmmm
cocm

 

mmm
vmvm
cocm
mmcm
mvmm

mmmm
mmcmm

otm
mmmm
comm

mmtm
comm
vcmvm

cocc—
commm
mmme
comm
omcw
mmmm
comm
mmm

 

vmvm
cocm
mrmm
cocm
com
cocm
mmcmm

ccwm
mmc—
comm
mmmm
comm
mvmmm

vmmm
cocmm
cmmp

mmcw
cmvm
cmmm
cmm

 

cmm
cocm
comm
mmcm
comm
com
mvmv
vcmvm
comm
mvm
comm
comm
comm
mcwm
comm
FmeF
com
wmmm
commm
mvmw
comm
more
comm
comm
v¢m

 

cmm
mmmm
comp
mcmm
comm
cmm
comm
mmwvr
comm
com
cocm
oom—
vmmm
cmmm
cocm
cctmv
mmv
mmmm
commm
mmtr
mrmm
mmww
comm
ccVF
vvm

 

cmm
cocm
comp
ccwm
cocm
cmm
com?
cocmw
cocm
com
cccv
ccmw
comm
comm
cocm
commm
cc?
cocm
cocmm
ccmw
ccwm
coww
cocm
comm
com

 

cocm
ccmw
comm

cow?
cocmm
cocm

comm
cocm
comm
comm
comm
occvw
cmv
cocm
cocmm
ccmF
cocvw
coop
cocm
cow?
com

 

cmm
comm
comm
comm
cocm
cmm
coww
commm
comm
cmm
comm
coop
comm
cmmm
comm
comm?
cm?
cocm
ccccm
cocw
comm
cccw
cocF
corp
com

mc_Eo>>>

 

£983.?

 

9535335

 

c2933;

 

£595

 

aco§o>

 

cm=4

 

v.98...

 

0335.8.

 

90me 530m

 

2:98 5:8.

 

25.2 2.95

 

«Emzxmccon.

 

comw5

 

2.2.25

 

ozo

 

80an 5.52

 

m5.0.60 ztoz

 

vto> 262

 

02me 302

 

>342. 162

 

238.5: 262

 

$952

 

9.8502

 

 

28:22

 

 

3.8330 £23850 BmE 8% San m-U 03mm

71

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NN NN 8 8 NN N. N. N. N. o. N N 38%
N. v. v. N. N. N. .. o. N N N N _&_NN_NN_s_
N.. ... Ne Nv Ne .4 I. .v NN .N N N. 2885.2
N. NE NN NN NN NN oN NN NN NN o o 59:22
NN NN 3 N... N... .N NN NN oN NN N. N Nummafimmmmz
NN NN oN NN NN NN NN NN N. o. o. N 95.5.2
8 Nv N.. ... NN NN NN on 8 N. N. N 8a:
N. N. N. v. N. N N N. o. o. N N 29.0.53
8 NN NN NN N. N. N. N. N. N. N.N o :02ch
N N. N. .. .. N N N N N N N 323.
NN NN ..N NN NN NN N. N. NN o. N.N N.N N32
NN NN NN NN NN N. N. N N N N N 2%...
NN NN NN NN NN NN N. NN .. N. N N was...
a? N2 NE NE NE E: o. N. o. N N N 2%.
«N NN «N NN NN NN N. .. v w v v .95:
N2 NE 8 NN 8 .3: N. N. N. N N.N o .238
NN NN NN ow 9. 8 NN NN NN .N N. v «23:
N. N. N N N N. NN 8 8 N N N 59:28 .o .35
NN NN NN .N .N NN NN NN N. N NN v $5.8
NN N N NN NN NN NN .N N. N. o N 508280
N N). N. N. N. N. N. NN NN N. N ... 8928
Nv NN NN NN NN NN NN .. .. N. .. N. 352:8
N.. N.. NN NN NN NN NN N. o. N N N N852
N. NN N. N. v. o. N o. N N N o 885
N N N N N N N N N N N N 9.32
NN NN NN NN NN N. N. N. N. N N N 9.582
NNNN NNN. NNN. NNN. NNN. NNN. «Na. NNN. NNN. .NN. NNN. NNN.
g 30>on HEN

 

 

m40>0mm NENtN> 2: New 8ND m0 NENF

72

 

 

 

a?
0?.
mm
mm
mm

cm
mm
3:
mt
em
mm
mm

m?
cm
cm
mm
m¢
c F
cw
vm
_. F
mm
P P

 

885%‘988918353880

m w
m _.
mm
mm
mv
c _.
me
mm
3
mm

 

9833880

0?

mv

mm
mm
mm
mp
m w
m
mm
mm
m F
mv
mm
m P
mm

 

 

NN
N.
NN
NN
NN
N.
N2
N2
NN
N.
«N
N.
NN
N.
.N

cm
mm
m r
mv

m_.
m_.

 

m
mm
mm
mm
mm
mm
mw
3.
mp
cm

m
Vm
cm
cm
m.
mm
m.

m
mm

m
3
m.
m?
m.

m

v
.m
mw
mm
mm
mm
mr
mv

m
2‘
cm
Fm
mw
mm
or
cm
mp

m
mm

m
mm
or
m.

m

m

 

 

v
vm
op
mm
cm
mm
mw
E.
or
c_.
c.
mw
5
mm
ow
m.
m.
V
.m
m
vm
c.
or
c.

m

 

m
m.
9.
cm
9
cm
or
or

m
or

m
m«
c_.
m
or

m
or
me
3

m
cm

m
c.
o_.

m

 

m.m
mm
o—.

m.m.
or

mw
mi
m.mm

3
m

 

cmmmSEu-Scu-mmfimflmwcmcfiegcom

mchm>>

 

swcoomcs

 

95975 .33

 

c2923;

 

£595

 

.co§m>

 

gas

 

meom

 

mmwmwccmm

 

29.8 snow

 

2:050 538

 

95.2 2.0:”.

 

m_:m>_>N:con_

 

cameo

 

mEocmEO

 

2.5

 

20me 5.52

 

m5.050 ztoz

 

{8. 3N2

 

02.82 3.02

 

3&2. :62

 

EENQENI 302

 

mum>mz

 

9.22an

 

 

28:22

 

 

8:528 .5055. 2N€N> 2.. 5N 95 3 2.3

73

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NNNN NNNN NN.N NN.N NNNN NNNN .NNN NNNN NN.N NN.N N..N NNNN £53....
N.NN NNNN NNNN NNNN N.NN SNN NNNN NNNN N.NN .NNN NNNN NNNN 25.3.3...
83 NNNv NNNv NNN.. NNN.. NNN.. NNNN «NN.. NNN.. NNS NNN.. NNN.. 9825.2
NNNN NNN N .8 83 ..NNN NNNN NNNN NNNN NN.N 88 NNNN NNNN 595.5.
N.NN NN.N NN.N ....N 88 NNNN NS 88 NNNN NNNN N.NN N.NN 9.9.2083...
NNNN NN.N NN.N NNoN NNNN NNoN NNN.. N3..- NNN.. NNN.. .NNN NNN.. 25.3.2
NNN. NNN. 3N. N..N. NNN. NNN. NNN. NNN. NNN. NNN. NNN. NNN. 2.22
NNN.. NNNN NNN.. .NNN NNN.. NNN.. NNNN NNNv NNNv .va NNNv NNNN NNN.N...S
N.N.. .NNN NNNN N.NN NNNN NNNN ..NNN «NNN NNNN N.NN NNNN NNNN .62ch
NNNN «NNN NNNN .NNN NNNN NNNN NNNN NNNN N.NN .NNN NN.N NNvN NNNch.
NNNN NNNN NNNN ..NNN N..NN .NNN NNNN .NNN 88 .NNN NNNN .NNN 25.
.NNN NNNN NNNN NNNN NNNN NNNN ...NN .NNN N.NN .NNN N..NN ..NNN NNN.N...
N.NN. NN.N. NNNN. NNN.. NNN.. NNN.. NNN.. N.N.. NNN.. NNN.. .N¢.. 2.... was...
3N. NNN. NNN. 8N. NN.. .N.. .N.. 8.. NNN. NNN. N8. N8 N..N...
N.N. NN.. NN.. NN.. NN.. NN.. NN.. .N.. .N.. NN.. NN.. NNN. ..Namz
NN.N NNNN N..NN NN.N ..NNN NN.N N..NN NNNN NNNN NNNN NN.N ....N N.N.ooo
NNNN. ...N. N.NN. NNN... NN.N: NN.N. «NNN. N.NN. NNNN. NNNN. NNNN. NNNN. 3.3.”.
NNN .N.N NNN NNN N..N NNN NNN NNN NNN 3N NNN ..NN £228 .0 .N.N
..NN NNN 3N NNN NNN N.N 2: 2: SN NNN NNN NNN 2258
NN.N NNNN ..NNN NNNN .NNN NNNN NNNN NNNN NNNN NNNN NNNN NNNN 52.828
.NNN NNN.. .NNN NNNN N.NN NNNN ..NNN NNNN .N.N NNNN ..NNN NNN.N 8828
NNNNN NN.NN NNNNN NN.NN NNN.N NNN.N NNN.N NN...” NNNNN NNNNN NNNNN N.NNN £59.30
NNNN .NNN NNNN NNNN NNNN NNNN .NNN ..NVN ..NNN NNNN .NNN NNNN NNNcNVE
.N.N NNN.. $3 $3 NNS NNN.. Ne... NNN.N NNNN NNNN NNNN NNNN 23.2
NNN NNN ...N N.N NNN NNN .NN NNN NNN NNN NNN N..N 9.32
N3... NNNN NNN.. NNN.. .NNN NNN.. NNN.. NN.N NN.N 89. $9. NNN.. mega...
NooN NNN. NNN. NNN. NNN. NNN. NNN. NNN. NNN. .8. 8m. NNN.
$288... .5 20.5339. PENN

 

 

 

NON N.NNE> 2.. 5N N.N.. .5 N.NN.

74

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.. NNN .N.. 8.. NN.. NNN NN.. NNN NNN NN.. ..N.. NNN 85.8.2.
..NNN NNNN ..NNN .NNN «N.N NN.N N8N NNNN N8N NNN.. NNN.. NNN.. c.2822.
NNN. NNN. ..N. N.N. NNN. NNN. N.N. N.N. NNN. NNN. NNN. NNN. «.595 .82.
88 NNNN NNNN N.NN N.NN NN.N NNNN N.NN NN.N N.NN NNN.. NNN.. 88:28.5
NNNN NNNN .NNN NNNN NNNN NNNN NNNN NN.N ..NNN ..NNN NN.N NN.N «.53..
N8 NNN .NN 8N NNN NNN NNN ..NN NNN NNN NNN NNN 285$
NNNN NN.N NN.N NNNN NNNN .8. N8. NNN. N.N. .NN. NNN. NNN. ..N.:
NNNNN $.8N NNNN. NNNN. NNNN. .NNN. NNNN. 88. NNNN. N.NN. NNNN. NNNN. 86.
NNNN NN.N .NNN NNNN NNNN NNNN NN.N NNoN N.NN N.N.. NNN.. .NNN 88:5.
NNN NNN NNN NNN NNN NNN NNN NNN N.N NNN NNN NNN 59.8 .58
N.NN NNNN NNNN NNNN NNNN 8NN NNNN NNNN 8NN NNNN NN.N NN.N 2.28 5:5
N3. .8 N8 NNN NNN N8 N8 N8 NS. «8. N8. .8. NNN.N. 82m
.NNN. 8.. .NNN. ..NN. «NNN. 88. N.NN. NNNN. ...N.. NNN.. NNN.. NNN.. £52.25“.
.NNN N.NN NNNN N..NN NN.N ....N NNNN NNNN «NNN N.NN N.NN .NNN c820
.N.N NNNN N..NN NNNN NNNN .NNN NNNN NNNN ..NNN NN.N NN.N NN.N «52.3.0
NNN.. NNN.. NNN.. NN... NN... NN... 8... NNN.. NNN.. .NNN. N..No. NNN... 95
NNN ..NN NNN ...N NNN ...N N..N NNN NNN ..NN NNN N..N 99.8 582
N.NN .NNN N..NN .NNN 82 NN.N .NNN 88 NNNN N..NN NNNN NNNN 2.28 532
NNNN. NN.N. NN.N. NN.N. NN.N. NN.N. NN.N. NN.N. NNNN. NNNN. .NNN. NNNN. €8.32
N.N. N..N. NNN. NNN. NNN. NNN. NNN. N.N. .NN. N..N. N.N. .8. 8.3.... .52
...NN NN.N N. .N NNNN NNNN NNNN N.NN NNNN NNNN ..NNN N.NN NNNN .028 3N2
NNN. ...N. NN.. NN.. NN.. N... NN.. NN.. v... NN.. 8.. NN.. 2.282in2
N8. N8. NNN. NNN. 98. NNN. NN... NNN. NNN. NNN. NNN. NN.. 89oz
..N. NNN. NNN. NNN. N..N. NNN. NNN. N.N. NNN. .NN. NNN. NNN. 9.582
N8 NNN NNN NNN NNN NNN NNN N..N NNN NNN 8N NNN 298....

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82.38 59. 28...... 2.. 5N N.NN. .5 N.NN.

75

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NN.NN N.NNN NN.NN NNNNN N..NNN NNN.N N.NNN ..NNN. N..NN. 88. 8:. NNNN. .598...
NNNNN NN.NN ..NNN. NNN... 85 88. .NNN. NNN... NN.N. N.NN. .NNN. N.N.. 8.8.8....
N.N.N NN.NN NN.NN N..NN NNNNN NNN.N NNNNN .NN.N NNN.N .NNN. NNNN. N..NN. 388......
.88 NNNNN NNNNN NoNNN NNN.N .NNNN NN.NN NNN.N NN8. N..NN. NNNN. NNNN. 58.2....
N.NNN NN.....N N.NNN 8N8 NN.NN ..NNN .NNNN ..NNNN N.NNN NNNNN NNNNN NNN.N 88:688....
.NNNN NNN.N N88 NNNNN NNNNN NNNNN NN.NN NNSN ..NNNN .NNNN 8NNN NN..N Esra...
8NNN NNN.N NNNNN .8NN N...N NN.NN ..8. N..NN. NNNN. «NNN. NNNN. NNNN. NN.N...
.3NN NNNNN «NN.N NNNNN .N8. .NNN. NN.N. N.NN. NN.N. ..NNN. .NNN. «NNN. N533.
NNN.N N.NNN NN.NN NN8N 88. .NNN. NNNN. 8NN. .NNN. NN.N. N..NN. NNNN. .99..ch
NN.NN N.NNN NNNNN NN.NN NNNNN NNN.N ..NNNN N.NNN ...NN. .NNN. NNNN. 88. 23.5.
NNNNN N..NNN .NN.N NNNNN NN.NN .8oN NNSN NNNN. .SN. .NNN. ..NNN. N..NN. 26.
NNNNN NNNNN NNNNN NN.NN ..NNN NNN.N .NNNN NNNN. NNNN. NNNN. N.NN. NNNN. NNN.N...
N..N.N NNNNN .NNNN N.NNN NNNNN NNNNN NNN.N NN.NN NNNNN .NNNN NN.NN .NNN. Nos...
N.NNN NNNNN NNN.N NNNNN .88 N.NN. ..NNN. NNNN. NNNN. NNNN. .NNN. .NN... 2.3.
N.NNN NNNNN NN.NN N.NNN N.NNN ..NNN NNNvN N..NNN .NNNN NNN.N No8N NN.N. ..Nzaz
8NNN NNNNN .NNNN NNNNN 8NNN NNN.N NNNNN NNNN. NN.NN. N..NN. NN.N. 88. 9888
NNNNN NNNNN NN.NN .NNNN NNNNN ..NNNN NNN.N NNN.N NNSN NN.N. NNNN. NNNN. 8.3.“.
..NNNN ..NNNN NNNNN ..NNN NNNNN NNN.N NNN.N NN.NN NNNNN NNNNN N.NNN ..NNNN on
.NN.N NNNNN NNNNN N.NNN NN.NN NNN.N NNNVN NNNNN NN.NN NNNNN N.NN. NNNN. N583
NNN? NoNNN NN.NN NNN.N ...NNN NNN.N N88 NN.NN .NNNN NNNNN NN.NN NNN.N 52.858
.38 NNNNN NNNNN NNNNN NNNNN N.NNN NN.NN NNNNN NNN.N N..NN. N.NN. NNNN. 890.8
NNNNN N.NNN NNNNN NNNNN NN.NN «NN.NN ..NNNN NNNNN NNNNN NNNNN ..NNoN NNNN. 352......
NNN.N .N..N .NNNN ..NNN. NNNN. NN.N. NN.N. NN.N. NNNN. 33.. ER. 88. N853
.8..N NNNNN .N.NN NNN.N 8N8 NNSN ..NNN. N.NN. ..NNN. NNNN. NNNN. NNNN. 88.2
NNNNN NNNNN N.NNN NNNNN NNNNN NNNNN .NNNN N.NNN .N.NN NNN.N .NNNN .NNN. N..NN...
NN.NN NNNNN N.N.N .NNNN NN.NN ..NNN. .NNN. NN.N: NNNN. N.NN. N8... NNNN. NEBNZ
NNNN NNN. NNN. NNN. NNN. NNN. NNN. NNN. NNN. .NN. NNN. NNN..
3. N58... .2022. 5.80 .N.. NENN

 

 

 

8.4007: oENE> 05 Now ENG 9.0 038:

76

 

 

mmvmm
mmcmm
mmmwm
mmpwm
mmcmm
vommm
vmmmm
mmmmm
mmmmm
mmmmm
mmmmm
mmwmm
mmmmm
mwmmm
mmmmm
wrmmm
ommvm
mvmmm
mcmvm
mmmrm
mrwmm
mccmm
memm
mmmmm
.vmmm

 

cmmmm
mmmmm
cmmcm
mmmmm
mommm
Vwmmm
mmmmm
mmmmm
mmmvm
wmv¢m
mommm
m.mmm
vmmmm
mmvmm
mmmmm
mmmmm
mmcmm
vwmmm
cmmmm
cmmcm

cmmcm
mmmmm
mmmmm
rwmrm

 

mmm¢m
mwcmm
mmmcm
mmmmm
mmcmm
mmmcm
cmmpm
mmcmm
mcpvm
vmwmm
mmwmm
cmmmm
mccmm
mccmm
mmmwm
mwmmm
mmmmm
mmmcm
mmmwm
cmmcm
mvmmm
mmmmm
cmwmm
mmmmm
mmmFm

 

mvmmm
vm¢¢m
mvmmw
mmvmm
mwvmm
mmcmm
mwmcm
mmmmm
«mmmm
mmmvm
mccwm
mmmmm
cmmmm
mmmvm
Pmmcm
cmmcm
mcmcm
mmcmm
vmmmm
cvmmw
mmmwm
cmmmm
mmmmm
mvwvm
mcmm.

 

vmmwm
mommm
mmmmw
mccmm
mvmmm
mmcmm
mwmmp
mmvmm
mmcmm
mowwm
mmccm
mvmwm
mmmcm
mmmmm
mmmmw
mmvmm
mcmcm
Pmmmm
vmmmm
mmmmr
mmmcm
.mmmm
mccmm
mmmmm
mmwmw

 

cwmwm
mmmmm
mmmmw
mmmmm
cmmwm
mvwrm
vwmm.
mmmwm
mmvpm
mmmmw
mmmmv
mmvmm
.wemm
mmmmm
vmwmw
«mmmm
cmmm.
mmvwm
cm—mm
mmvmw
wmmmm
mmmcm
mcmvm
mcmwm
mmmm.

 

mmccm
mmmwm
mmcmw
mommm
vmwmm
mmmcm
comm.
cmmcm
mmpcm
memv
mmcmw
mmvmm
mmvmm
cmmcm
vmcm.
cmmwm
chmw
mmccm
vmmmm
memw
mmmmm
mmmvm
vcmmm
wmvcm
momm.

 

mmmm.
mmmcm
mcvm.
mmpmm
mmmmm
memp
mmvmw
«mmmw
mmmm.
mcmm.
mvmm.
mmmvm
mwmwm
Pmmmw
cwmm.
mmmcm
mmwm.
mmmm.
cmmmm
mmmm.
mmmmm
mwmmm
cmmmm
mwmm.
mcmm.

 

mcmmw
mmcmw
mmmmw
mrmwm
cmmwm
mmmmw
mmmmw
mmmmw
mcmmw
mrwmw
mm¢mw
mmmcm
mmcwm
mvmmw
mmmmw
mmmm.
m.mmw
mmmm.
mcmwm
mommw
mmmmm
mvmmm
mwmvm
cmmm.
m.mm?

 

cmmm.
mmmmw
mmmww
mmvcm
mmccm
mmmmw
mmmvw
cmvmr
wcmm.
mmmmv
mmvmw
mmtmw
cvmmw
mmmmw
mmmmw
mwomw
mmmmw
mcmmw
mmmmm
mwmvw
mvmvm
mmcrm
mecm
mmcmm
mmmmv

 

mcmmw
mmmmp
mmmmp
mmmm.
mmmmw
mvvmp
cmcvr
mvmmr
mommm
cmmm.
mcwmr
mmcmv
mmmmv
mmwmw
mwpmw
mcmmw
memw
mmmm.
memm
mrmvw
mmPVm
mmmcm
vmmcm
mmmmw
vavw

 

cmmmw
mmcmm
mmmmw
mmcmv
mmmm.
.mmmv
chmw
mmmmv
vmcmp
mmwv.
vmmm.
Fv¢mw
vcmmw
mmmmw
mmwvw
wtmmw
mmmm.
mmmmw
mmmcm
mmmmw
wwwmm
mmmmw
cmmm.
cmcmw
mmwvw

#520;

 

c.2823

 

9:545 .82.

 

c2923;

 

0.595

 

.co::0>

 

:05

 

N0x0m

 

008053

 

99...... snow

 

8.98 snow

 

NNN.N. 82E

 

2:02.658

 

cameo

 

308.5

 

Orzo

 

99.8 532

 

05.9.00 552

 

...o> 2.02

 

8:85. 2.02

 

>022, 2.02

 

0.2380... 2.02

 

00902

 

8.3502

 

 

0:02.22

 

 

82.88 .8802. N.NNE> 2.. 8. N.NN. N-N M..NNN

77

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NN. 8. NN. NN. NN. NN. NN. NN. N.. NN N. N .582...
3 N. N. N N. N. NN N. N. NN N. N 398.8...
NNN NNN .NN .NN N.N NNN ..NN .NN .NN NN.. .NN NN. 282......
NNN .NN NNN NNN NNN NNN NN. NN. NNN NN. N.. N 596....
NN. NN. NN. NN. NN. NN. .N. N.. NN. NN NN N 0.336882
8 NN. NN. NN. 8. NN NN NN 8 . NN N NNN.N.N...
..N NN. x. 3 NN ..N ..N N.. N. N. N N NN.N...
NN NN NN NN Ne NN NN NN .N N. N. N 232.5.
N.. NN N.. N.. NN NN NN NN N.. N. N. N 2.8.ch
NN. 8 .N. 8 NN N. N. N. N. N. N N N85.
NNN NNN NNN NNN N..N NN.. NNN .N.. N..N NN N. . N..N.
NN. NN. NN. ..N. ..N NN NN ..N ..N 8 NN N NNN.N...
..N.. «N.. NN.. NN.. NN.. NN.. NN.. NN.. NNN NNN NN. NN N.NN...
NN N... N N N N N N N... N . N N..N...
N N N N N N N . N N N N .NN.N:
NN.. ..N. NN. N... NN. NN. NN. NN. NN NN NN N... 880
8N NNN N.N NNN NNN NNN NNN .NN N.N NNN NN. N.. 3.5.“.
. . . N . . . . . . . N N.Neaoo .0 .N.N
N N N N v N . . N N... N N 055.8
NN. NN. NN. NN. NN. NN. NN. NN. NN. NN. NN N.. 52.858
N... N... NN NN NN NN NN NN NN NN N N 8928
N..N .NN ..N NN.. ..N NNN NN.. ..N.. N3 NNN NNN NN. N_Eo...No
N... NN ..V 3. ... NN NN N. N. N N N N853
NN NN NN NN NN NN NN NN N. N. N . 23.2
N N . . . . N N N N N N N..NN...
NN NN NN NN NN NN NN NN NN N N N 958...

NNNN NNN. NNN. NNN. NNN. NNN. NNN. NNN. NNN. .NN. NNN. NNN.

.9352. 2:98.“. 8.3.8 NEPN

 

 

 

550 2.5... 2.. .NN N.N.. N-N N.NNN

78

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 2. N . N N N N . . N . N 322.2.
.NN NNN NNN NNN NNN NNN NNN NNN NNN NN. N2 NN 5.282...
.N .N NN NN NN NN NN NN NN NN N. N 2.32. .82.
NNN NN. NN. NN. NN. NN. NN. NN. 8 .N NN N. 3.822....
NN NN NN NN NN .N NN NN NN .N NN N. 25%
NN NN NN NN NN NN NN NN N. N. N . .85.N>
N 2.. N. N. N. N. N. N. N. N . N ..N.:
NN. NN. NN. NN. NN. NN. NN. NN. 8 N. N. N 2me
N2 N2 NN NN NN NN NN N. NN N N N 8825.
2.. N N N N . . . . . . N 29.8 N.NNN
NN. N2 NN. 2.. .N NN NN NN NN N N N 2.28 N.NNN
NN NN NN NN NN NN NN .N NN N. NN N 2%. 2.2.2
NNN NNN NNN NNN NNN NNN .NN NNN NNN NNN N.N NNN 252.25..
NN. NN. NN. NN. NN. NN. N.. N.. N.. N.. NN. NN. 5820
N N N N N N N N N. .. N . «.5225
NNN NNN NNN NNN NNN NNN NNN NNN NN. NN. NN NN 0.20
NN NN NN NN NN NN NN N. N. N N N 29.8 552
NNN NNN .NN NNN NNN NNN NNN NN. NN NN NN N 2.28 ..th
NNN. NNN. NNN. NNN. NNN. NNN NNN NNN NNN NNN NN. 2. ...N.. .52
N N N N N .N N N N N N N 8222 252
N.N N.N N.N N.N NNN NNN NNN NNN N .N NNN NNN NNN .222 .52
NN NN NN NN NN NN NN NN NN NN N. N 2225.... 52
N N N N N N N N N N . N 2252
.. .. N. N. N. .. .. N N N . N 2.2.22
2.. N N N N N N N N N N N 2252

 

 

 

 

 

 

 

 

 

 

 

82200 .550 222.. 2. 8. .20 N0 N.NNN

79

BIBLIOGRAPHY

8O

BIBLIOGRAPHY

BEAR (2002), Understanding Beverage Container Recycling: A Value Chain Assessment
Prepared for the Multi-Stakeholder Recovery Project, Stage One, Business and
Environmentalists Allied for Recycling, R.W. Beck, Inc.

Burn, SM. and S. Oskamp (1986), Increasing Community Recycling with Persuasive
Communication and Public Commitment, Journal of Applied Social Psychology
16: 29-41.

Chatterjee, S. and Price B. (1991), Regression Analysis by Example, 2nd Ed., New York,
Wiley.

De Young, Raymond (1985-1986), Encouraging Environmentally Appropriate Behavior:
The Role of Intrinsic Motivation, Journal of Environmental Systems 15(4): 281-
292.

De Young, R (1986), Some Psychological Aspects of Recycling: The Structure of
Conservation Satisfaction, Environment and Behavior 18: 435-449.

Deci, EL. (1975), Intrinsic Motivation, Plenum, New York.

Dinan, Terry M. (1993), Economic Efficiency Effects of Alternative Policies for
Reducing Waste Disposal, Journal of Environmental Economics and Management
25: 242-256.

Eckblad, G. (1981), Scheme Theory: A Conceptual Framework for Cognitive-
Motivational Processes, Academic Press, New York.

FDL (1989), The Staff of NSDA, There Are No Winners, Forced Deposit Laws, National
Soft Drink Association, USA.

81

Folz, David H. and Joseph M. Hazlett (1991), Public Participation and recycling

Performance; Explaining Program Success, Public Administration Review 51(6):
526-532.

Fullerton, Don and Thomas C. Kinnaman (1995), Garbage, Recycling, and Illicit Burning
or Dumping, Journal of Environmental Economics and Management 29: 78-91.

Glenn, Jim (1990), The State of Garbage in America; Part I, BioCycle 31(3): 48-53.

Glenn, Jim (1990), The State of Garbage in America; Part II, BioCycle 31(4): 34-41.

Glenn, Jim (1991), Improving Collection Efficiency for Curbside Recyclables, BioCycle
32(8): 30-32.

Glenn, Jim (1992), The State of Garbage in America, BioCycle 33(4): 46-55.

Glenn, Jim (1998), The State of Garbage in America; Part I, BioCycle 39(4): 32-43.

Glenn, Jim (1999), The State of Garbage in America, BioCycle 40(4): 60-71.

Glenn, Jim and David Riggle (1991), The State of Garbage in America; Part I, BioCycle
32(4): 34-38.

Goldsby, Thomas J. (1998), Consumer Recycling Programs: The Marketing and
Logistics Implications, Dissertation for the Michigan State University Department
of Marketing and Supply Chain Management.

Goldstein, Nora (1997), The State of Garbage in America; Part I, BioCycle 38(4) 60-67.

Goldstein, Nora (2000), The State of Garbage in America; Part I, BioCycle 41(4) 32-39.

Goldstein, Nora (2000), The State of Garbage in America; Part II, BioCycle 41(11) 40-79.

82

Goldstein, Nora and Celeste Madtes (2001), The State of Garbage in America, BioCycle,
42(12): 42-54.

GRRN (2000), Wasting and Recycling in the United States 2000, GrassRoots Recycling
Network, Athens, GA.

Guagnano, Gregory A., Paul C. Stern, and Thomas Dietz (1995), Influences on Attitude-
Behavior Relationships: A Natural Experiment with Curbside Recycling,
Environment and Behavior, 27(5): 699-718.

Gujarati, Damodar N. (1995), Basic Econometrics, 3rd Ed., McGraw-Hill, Inc.

Hamilton, Lawrence C. (2003), Statistics with STATA: Updated for Version 7, Duxbury,
Thomson Learning, Inc.

Hong, Seonghoon, Richard M. Adams, and H. Alan Love (1993), An Economic Analysis
of Household Recycling of Solid Wastes: The Case of Portland, Oregon, Journal
of Environmental Economics and Management 25: 136-146.

Hopper, Joseph R. and Joyce M. Nielsen (1991), Recycling as Altruistic Behavior:
Normative and Behavioral Strategies to Expend Participation in a Community
Recycling Program, Environment and Behavior 23(2): 195-220.

Hsiao, C. (1985), Benefits and Limitations of Panel Data, Econometric Reviews, 4: 121-
174.

Jenkins, Robin R. (1991), Municipal Demand for Solid Waste Disposal Services: The
Impact of User Fees, Dissertation for the University of Maryland Department of
Economics College Park, MD.

Katzev, RD. and A.U. Pardini (1987-1988), The Comparative Effectiveness of Reward
and Commitment Approaches in Motivating Community Recycling, Journal of
Environmental Systems 17: 93-113.

Klevmarken, N. A. (1989), Panel Studies: What can We Learn from Them? Introduction,
European Economic Review, 33: 523-529.

83

Lake, Iain R., Ian J. Bateman and Julian P. Parfitt (1996), Assessing a Kerbside
Recycling Scheme: A Quantitative and Willingness to Pay Case Study, Journal of
Environmental Management 46: 239-254.

Lansana, Florence M. (1992), Distinguishing Potential Recyclers from Nonrecyclers: A
Basis for Developing Recycling Strategies, Journal of Environmental Education,
23(2): 16-23.

Luyben, Paul D. and Susan Cummings (1981-82), Motivating Beverage Container

Recycling on a College Campus, Journal of Environmental Systems, 11(3): 23 5-
245.

Markowitz, Paul (1991), Winning Strategies for Curbside and Dropoff, BioCycle, 32(4):
39-41.

McGrath, S. (1990), Commingled vs Separation Program Design Criteria, Proceedings of
Sixth International Conference on Solid Waste Management and Secondary
Materials, Philadelphia, PA.

Mohai, Paul and Ben Twight (1987), Age and Environmental Concern: An Elaboration of
the Buttel Model Using National Survey Evidence, Social Science Quarterly 68:
798-815.

Morris, GE. and D.M. Holthausen (1994), The Economics of Household Solid Waste
Generation and Disposal, Journal of Environmental Economics and Management
26: 21 5-234.

Mullins, Daniel R., Alfred Tat-kei Ho, and Chia-Yin Chou (1997), The Solid Waste
Crisis: Is Recycling a Response?, The Municipal Year Book, International City
Manager’s Association.

Naughton, Michael, Frederic Sebold, and Thomas Mayer (1990), The Impacts of the
California Beverage Recycling and Litter Reduction Act on Consumers, The
Journal of Consumer Affairs, 24(1): 190-220.

84

NREL (1992), Data Summary of Municipal Solid Waste Management Alternatives,
National Renewable Energy Laboratory, NREL/TP-431-4988G, Golden,
Colorado.

OECD (1974), The Polluter-Pays Principle: Note on the Implementation of the Polluter-
Pays Principle, OECD, Paris.

OECD (1993), Appling Economic Instruments to Packaging Waste: Practical Issues for
Product Charges and Deposit-refund Systems, OECD Working Papers Vol.1 (14),
Paris.

Okamoto, Tetsuya (1991), Essential Factors to Enhance Growth of Integrated Recycling
Systems for Post-Consumer Aluminum Cans and Plastic Bottles in Japan, Thesis
for the Michigan State University Department of Agricultural and Natural
Resources.

Porter, Richard C. (1978), A Social Benefit-Cost Analysis of Mandatory Deposit on
Beverage Containers, Journal of Environmental Economics and Management, 5:
351-3 75.

Porter, Richard C. (1983), Michigan’s Experience with Mandatory Deposit on Beverage
Containers, Land Economics, 59(2): 177-194.

Resource Recycling (1982), The Public’s Opinion of Recycling, 1: 6-7.

Reschovsky, James D. and Sarah E. Stone (1994), Market Incentives to Encourage
Household Waste Recycling: Paying for What You Through Away, Journal of
Policy Analysis and Management 13(1): 120-139.

Richardson Robert A. and Joseph Havlicek Jr. (1978), Economic Analysis of the
Composition of Household Solid Wastes, Journal of Environmental Economics
and Management 5: 103-111.

Solon, G. S. (1989), The Value of Panel Data in Economic Research, in D. Kasprzyk, G.
J. Duncan, G. Kalton and M. P. Singh, eds., Panel Survey, John Wiley, New York,
486-496.

85

Steuteville, Robert (1994), The State of Garbage in America; Part I, BioCycle 35(4): 46-
52.

Steuteville, Robert (1995), The State of Garbage in America; Part I, BioCycle 36(4): 54-
63.

Steuteville, Robert (1996), The State of Garbage in America; Part I, BioCycle 37(4): 54-
61.

Steuteville, Robert and Nora Goldstein (1993), The State of Garbage in America; Part I,
BioCycle 34(5): 42-50.

Stilwell, E. Joseph, R. Claire Canty, Peter W. Kopf, and Anthony M. Montrone (1991),
Packaging for the Environment: A Partnership for Progress, Amacom, Arthur D.
Little, Inc.

Sundeen, Richard (1988), Explaining Participation in Coproduction: A Study of
Volunteers, Social Science Quarterly 69: 547-568.

Thomas, John Clayton (1990), Public Involvement in Public Management: Adapting and
Testing a Borrowed Theory, Public Administration Review, 50(July/August):
435-445.

US. Department of Commerce (2001), Statistical Abstract of the United States: 2001,
Economics and Statistics Administration, Bureau of the Census, Washington DC.

US. Department of Commerce (2001), Survey of Current Business, Bureau of Economic
Analysis, Washington DC.

US. Environmental Protection Agency (1995), Report to Congress on Flow Control and
Municipal Sold Waste, EPA530-R-95-008, March 1995.

US. Environmental Protection Agency (2001), Municipal Solid Waste in the United
States: 1999 Facts and Figures, EPA530-R-01-014, July 2001.

86

US. Environmental Protection Agency (2002), Municipal Solid Waste in the United
States: 2000 Facts and Figures, EPA530-R-02-001, June 2002

Varian, Hal R. (1993), Intermediate Microeconomics: A Modern Approach, 3rd Ed.,
Norton & Company, Inc.

Vining, Joanne and Angela Ebreo (1990), What Makes a Recycler? A Comparison of
Recyclers and Nonrecyclers, Environment and Behavior 22(1): 55-73.

87

        

11111111111111 *

  

111111111111111111111111111111111111111

3 12