IS

M A GIBSON

 

 

 

. .
2:: s
I»???
.r:

 

 

 

 

 

 

 

 

n is,

a. 03;}.

... 141;; my!
@aofihung.

Sis—r

 

s...»

. . I . .35...
id.? J. 2. .. L
3...: ‘133:

km“

.. . .
.téx...’ 34:

a

.1133. . .l‘

. .4 .I‘MJ. i1.

.37 I h;
Inna...) :. ..

 

.Ixtw
“than. 7..
whit? v .
‘ I .
{Havana

.. K: 2.1:.

. 5!.
n w: ‘

utmtfhék :13): :l.. 1
. .. ‘ .

 

.l.h....1._... awning. I? . 9mg Gmfirqflzwmmml 1.. A: . . ‘ ‘ ,. uhmwwuw gmnfiggnafip

 

F LIBRARY
Michigan State
University

___—-.

J62?!

 

This is to certify that the
thesis entitled

Stated Preference of Michigan Residents for Great Lakes
Coastal Wetland Program Characteristics

presented by

Melissa Ann Gibson

has been accepted towards fulfillment
of the requirements for the

 

 

 

 

__ MS. degree in Agricultural Economics
Major Professof's Signature
8‘ /zé/zoar
Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05' ' cICl—RCIDdthr-II—hdd-p. 1 s ‘

Stated Preference of Michigan Residents for Great Lakes Coastal Wetland
Program Characteristics ‘
By

Melissa Ann Gibson

A THESIS

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

MASTERS OF SCIENCE
Department of Agricultural Economics

2005

ABSTRACT

STATED PREFERENCE OF MICHIGAN RESIDNETS FOR GREAT LAKES
COASTAL WETLAND PROGRAM CHARACTERISTICS

By
Melissa Ann Gibson

Wetlands are unique and important ecosystems, and policy makers want to
develop appropriate conservation programs. Trade-offs must always be made when
developing environmental programs. This thesis presents the results of a state-wide,
stated choice questionnaire concerning programs for protecting and restoring Great Lake
coastal wetlands. The survey questionnaire was designed and implemented as part of a
wetland ecosystem economics research program at Michigan State University. The
survey questionnaire was mailed to potential respondents in spring 2004 and had a 44%
response rate. This thesis describes wetlands, their functions, research methods used in
this study, and results of the stated choice section of the survey. Stated choice methods
estimate preferences by developing statistical relationships between people’s choices and
the characteristics of the alternatives they choose among. Preferences were estimated for
characteristics of Great Lakes coastal wetland programs such as the ecological services
the program focuses on, the program’s mix of preservation and restoration, and the land
acquisition method. Since the literature review did not uncover previous research on
public preferences for coastal wetland programs, this research provides a first step toward
filling this gap. Knowledge gained fiom this study can aid policy makers in evaluating
current policies and programs and in developing new coastal wetland programs that are

in-line with public preferences.

To my family, my husband Daniel, my parents Michael and Marilyn Savard, my

sisters, and all the rest. Thank you for everything.

iii

Acknowledgements

There are numerous individuals and organizations that are responsible for,
and contributed, to my research. First, I would like to thank my committee
members, Dr. Michael Kaplowitz, Dr. John P. Hoehn, Dr. Patricia Norris, and
especially my thesis advisor Dr. Frank Lupi. Without their guidance and
assistance this research would not have been possible.

I would like to thank Michigan Sea Grant, Michigan State University Land
Policy Program and the Michigan State University Department of Agricultural
Economics for their support of my graduate studies. Their contributions made
this work possible. I would also like to thank the Lake Michigan Federation for
providing me with an Environmental Economics Fellowship, as well as the
Wisconsin Coastal Management Program for providing funding to allow me to
present this work to others. In addition, I would like to thank those that attended
my presentations, at the Great Lakes Costal Wetlands Conference and the
Society of Wetland Scientists Annual Conference, this past year. Thank you for
the opportunity to share my research.

Finally, I would like to thank my family for all their love and support. I
could not have made it this far on my own. Thank you Daniel for your constant
love and support, I never would have made it through these few years without

you.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
CHAPTER 1 ........................................................................................................................ 1
INTRODUCTION .......................................................................................................... l
COASTAL WETLANDS ............................................................................................... 3
Michigan Coastal Wetlands ........................................................................................ 3
Functions of Michigan Coastal Wetlands ................................................................... 5
CHAPTER 2 ........................................................................................................................ 8
RULES, REGULATIONS AND PROGRAMS ............................................................. 8
Wetland Policies and Regulations in the United States .............................................. 8
Programs and their F oci Protecting Wetlands in Michigan .............. . ....................... 12
PROGRAM ELEMENTS ............................................................................................. 17
Preservation and Restoration .................................................................................... 18
Land Acquisition ....................................................................................................... 19
Purchase of Property ............................................................................................. 20
Conservation Easements ....................................................................................... 20

Land Contracts ...................................................................................................... 22
RANKING AND VALUING ENVIRONMENTAL SERVICES ................................ 23
Michigan Coastal Wetlands ...................................................................................... 23
Environmental Valuation ......................... 24
Stated Preference Modeling ...................................................................................... 25
Stated Choice Method ............................................................................................... 26
Examples of the Stated Choice Method .................................................................... 28
WETLAND VALUATION .......................................................................................... 30
Great Lakes Wetland Valuation ................................................................................ 32
CHAPTER 3 ...................................................................................................................... 34
SURVEY METHODS .................................................................................................. 34
Survey Questionnaire Design ................................................................................... 34
SURVEY QUESTIONNAIRE PRE-TESTING ........................................................... 35
IMPLEMENTATION ................................................................................................... 38
DATA ENTRY AND RESPONSE RATE ................................................................... 41
Demographic Characteristics of the Sample ............................................................. 44
CHAPTER 4 ...................................................................................................................... 47
STATED CHOICE MODEL ........................................................................................ 47
Stated Choice ............................................................................................................ 47
Experimental Design of the Stated Choice Question ................................................ 52

THE MODEL ............................................................................................................... 53

ESTIMATION OF THE MODEL ................................................................................ 56
ESTIMATION RESULTS ............................................................................................ 59
Interpretation of Estimated Parameters ..................................................................... 62
CHOICE EXAMPLES ................................................................................................. 69
CHAPTER 5 ...................................................................................................................... 75
SUMMARY AND CONCLUSIONS .................................. ' ......................................... 75
APPENDIX ........................................................................................................................ 8O
BIBLIOGRAPHY ............................................................................................................ 103

vi

LIST OF TABLES

Table 1. List of Wetland Protection Programs Covering Michigan Coastal Wetlands... 13

Table 1. (cont’d) ............................................................................................................... 14
Table 2. Interview Discussion Guide. .............................................................................. 36
Table 3. Stratified Random Sample, Depicting the Seven Most Populous Counties ...... 39
Table 4. Response Rate Coding. ....................................................................................... 42

Table 5. Comparison between Survey Respondents and Census Results for Michigan... 45

Table 6. Example Survey Questionnaire Page Showing One Version of the Stated Choice

Question (Scenario). ................................................................................................. 49
Table 7. Priority: Programs Primary Focus. .................................................................... 50
Table 8. Mix: Effort Devoted to Preservation and Restoration. ...................................... 51
Table 9. Tool: Land Acquisition Method. ....................................................................... 51
Table 10. Variables Used in the Estimation of the Utility Function ................................ 57
Table 11. Estimated Parameters from the Stated Choice Model. .................................... 59
Table 12. Frequencies of Actual and Predicted Outcomes. ............................................. 60
Table 13. Percent Predicted Correctly. ............................................................................ 61
Table 14. Choice Model Parameters Tests of Equality: P-values for Priority Variables

and Levels of Significance. ....................................................................................... 65
Table 15. Marginal Effects on the Probability of A ......................................................... 67
Table 16. Predicted Choice Example 1. ........................................................................... 69
Table 17. Predicted Choice Example 2. ........................................................................... 71
Table 18. Predicted Choice Example 3. ........................................................................... 72

vii

LIST OF FIGURES

Figure 1. The Nonlinear Effect that Percent Preservation has on Estimated Utility ........ 63

viii

CHAPTER 1

INTRODUCTION

Wetlands are unique and important ecosystems. They support a diverse group of
plants and animals and perform important ecological functions such as water filtration
and water supply storage (Keating, 1995). Great Lakes coastal wetlands provide a
variety of environmental services including the support of non-game species; provision of
open space; protection of water quality; flood control; provision of fish and waterfowl
habitat; as well as maintenance of biodiversity.

In the United States, the development of wetlands is federally regulated by the
Clean Water Act and The Rivers and Harbors Act (Denison and Schmid, 1997; Gaddie
and Regens, 2000). The US Army Corps of Engineers (USACE) and the Environmental
Protection Agency (EPA) were given the authority to regulate the pollution of waters in
the United States under the 1972 amendments to the Federal Water Pollution Control Act
(National Research Council, 2001). While the amendments included regulation of US.
wetlands, their narrow construction only covers 15% of the total wetland acreage.

Based on demographic information, it is predicted that by the early 2 1 st century
over 70% of the United States population will live within the coastal regions of the
oceans and the Great Lakes (Vernberg, et a1., 2001). Although various Federal and State
laws regulate the use of wetlands in Michigan, they do not necessarily protect particular
wetland services, and coastal wetland fimctions in Michigan are still being degraded and
lost. Currently, there is newly proposed legislation in Congress, the Great Lakes Water

Quality Indicators and Monitoring Act, which would assess the health of the Great Lakes

(United States Senate, 2005). Private organizations and state agencies have programs
that attempt to overcome some of the shortfalls of government regulations, and protect or
restore additional wetland acreage and services. There are around thirty programs
protecting or restoring wetlands in Michigan. They vary from programs that focus on a
providing a single service such as protecting waterfowl habitat to programs that are

broadly aimed at acquiring high quality coastal land.

Problem Statement:

Current wetland programs vary in their scope and method for protection and
restoration. At the present, little is known about the public’s preference for program
characteristics regarding Great Lakes coastal wetlands. With limited funding available,
programs need to focus their efforts. Research into preferences for wetland programs can
aid in determining which trade-offs to make when developing coastal wetland programs,

and can provide information on how well current programs fit the needs of the public.

Objectives:
1. To learn about programs, and the characteristics of programs, protecting
coastal wetlands in the state of Michigan.
2. To gain information, through the use of a questionnaire, on individuals’
preferences for Great Lakes coastal wetland programs.
3. To estimate, through statistical analysis, the relationship between particular
coastal wetland program characteristics and Michigan residents’ preferences

for Great Lakes coastal wetland programs.

The focus of this study was to elicit individual’s preferences for particular Great
Lakes coastal wetland programs through the use of a stated choice questionnaire. The
questionnaire was designed to gain information on three different dimensions of possible
coastal wetland programs: primary focus of the program (i.e., which environmental
service is the top priority for the program), the mix of preservation and restoration used
by the program, and the method of land acquisition for the program.

The thesis continues with defining Great Lakes coastal wetlands, and is followed
by a description of their functions. Next, Chapter 2 will go over current laws and
regulations protecting wetlands in Michigan, and will review various wetland protection
program characteristics and how they can be utilized. The end of Chapter 2 provides a
review of literature on preference studies, environmental valuation, as well as the stated
choice method. Chapter 3, Methods, discusses the survey design and implementation.
The stated choice question is presented in Chapter 4, along with estimation of the model
and interpretation of the results. Conclusions and a summary make up Chapter 5. The

appendix contains copies of each letter sent to respondents and the survey questionnaire.

COASTAL WETLANDS

Michigan Coastal Wetlands

Wetlands are biologically diverse and ecologically important. Due to their
changing nature, wetlands are difficult to identify so it is best to begin with a legal

definition.

“Michigan's wetland statute, Part 303, Wetlands Protection, of the
Natural Resources and Environmental Protection Act, 1994 PA 451, as
amended, defines a wetland as ‘land characterized by the presence of
water at a frequency and duration sufficient to support, and that under
normal circumstances does support, wetland vegetation or aquatic life,
and is commonly referred to as a bog, swamp, or marsh.’ The definition
applies to public and private lands regardless of zoning or ownership”
(Michigan Department of Environmental Quality, 2005 (a)).

All wetlands have three main features. First, they have a pronounced presence of
water at the surface or root level, and may have periods when they are dry. Second,
wetlands have unique soil conditions, including accumulation of organic plant material,
which is substantially different from surrounding lands. Third, the vegetation in a
wetland is specifically adapted to wet conditions (hydrophytes) (Denison and Schmid,
1997; Hey, et al., 1999). If there are plants present that require dry conditions, the area is
probably not a wetland, just an area inundated with water because of some external
factors affecting it temporarily (Mitsch, 2000).

Great Lakes coastal wetlands are wetlands directly connected to the Great Lakes,
and have water levels that rise and fall with the level of the Great Lakes (Mitsch, 2000).
Wetlands a distance in from the shore, half a mile for example, may be considered a
coastal wetland if its water level is connected to the water level of the Great Lakes. On
the other hand, a wetland very close to the shore is not necessarily a coastal wetland
because it is possible that its water level is not affected by the level of the Great Lakes.

There are six main types of Michigan coastal wetlands: lagoon! barrier, embayed
and shoreline, riverine, deltas, dune and swale, and manmade dikes (Albert, 2003;
Michigan Sea Grant, 2002). Lagoon/ barrier wetlands are the most common, but

embayed and shoreline wetlands are the most visible (Canadian Wildlife Service, 2002).

Shoreline wetlands in Michigan are currently more visible now due to the low water
levels of the Great Lakes, creating more beaches. This recently exposed beach is covered
with wetland vegetation and a constant presence of water.

The coastline of the Great Lakes is not made of an insignificant amount of land.
In fact, including the islands present in the Great Lakes, there are 9,941 miles of coastline
between the United States and Canada on the five lakes. This is only 2,442 miles fewer
than the oceanic coastline for the entire United States, including Hawaii and Alaska, of
12,383 miles (World Atlas, 2004). Michigan alone has 2,138 miles of mainland shoreline
(Jaworski and Raphal, 1978). In 1972, coastal wetlands comprised only 3.5% of
Michigan’s total wetland acreage, however, they made up a larger percentage of total
habitat, 11-21% for different species (Jaworski and Raphal, 1978). In the 70’s, coastal
wetlands comprised 16% of the Michigan shoreline, with the highest percentage of
Michigan’s coastal wetlands around Lake Erie and Lake Huron, with a majority around
Lake Huron (J aworski and Raphal, 1978). Four counties around Saginaw Bay (on Lake

Huron) contain 33% of Michigan’s total coastal wetlands (Jaworski and Raphal, 1978).

Functions of Michigan Coastal Wetlands

Michigan coastal wetlands have many functions that are important to our
environment and us. They help improve water quality by absorbing nutrients and
chemicals that would otherwise be released into the environment and lead to poor water
quality. Wetlands filter sediment, reducing turbidity and making the water clearer (Hey,

et al., 1999). Wetlands provide an important function by recharging groundwater

supplies. They also provide an area for storage of flood waters, as well as supplies of
water during drought (Hey, etal., 1999).

Great Lakes coastal wetlands provide habitat for fish, birds, and other wildlife.
The spawning and nursery grounds that coastal wetlands supply are extremely important
for sustaining our Great Lakes fish populations. They support small fish used as food by
recreational and commercial fish. At least 75% of all Great Lakes fish use coastal
wetlands during their life span, for one reason or another (Canadian Wildlife Service,
2002). Coastal wetlands also provide necessary breeding and feeding grounds for
waterfowl. Michigan’s coastal wetlands are very important for fall and spring bird
migrations. Some species of waterfowl are only found in specific areas of Michigan’s
coastal wetlands.

Biological diversity provided by wetlands is another important function. Coastal
wetlands may contain rare or uncommon fauna and flora that are not found in other
environments. Wetlands hold concentrations of many species, including threatened
and/or endangered species that could become extinct if their habitat continues to
diminish. More wildlife and plants are supported by wetlands than any other habitat type
in Michigan (Michigan Sea Grant, 2002).

These few examples of wetland functions are not meant to be exhaustive, but
simply highlight some of the functions and services provided by wetlands. Other wetland
functions include timber production, educational and research values, and open space and
aesthetic values (Dennison and Schmid, 1997). Understanding the ecological functions

and services provided by Great Lakes coastal wetlands is an important step in

determining public preferences for wetland programs and insures that informed decisions

can be made.

CHAPTER 2

RULES, REGULATIONS AND PROGRAMS

Wetland Policies and Regulations in the United States

Between the late 1700’s and the mid-1980’s, 53% of the wetlands in the
continental United States were lost (Dennison and Schmid, 1997). The Swamp Land
Acts of 1849, 1850, and 1860 were the beginning of relevant federal regulations
regarding agricultural drainage (Hey, et al., 1999). Drainage of “swamps” was believed
to be a public benefit. In the 1890’s the Rivers and Harbors Act was enacted which gave
the Corps regulatory authority over work done on navigable waters, although
environmental impacts were typically not considered in decision making (National
Research Council, 2001). Through the Reclamation Act of 1902, land drainage had
active government involvement and in 1936 the Flood Control Act was passed and
land/waterways were either cleared, dredged or channels created in the name of flood
protection (Hey, et al., 1999).

While previous regulations were aimed at clearing and draining land for other
uses, in 1962, PL87-732 began to change the constant support of drainage for agricultural
purposes. Wetland drainage in Minnesota and the Dakotas could no longer be financially
supported by the USDA if the Secretary of the Interior determined that wildlife
preservation would be harmed by drainage of the land (Hey, et al., 1999). In 1977, all

cost-sharing for wetland drainage between farmers and the USDA was halted.

The Water Pollution Control Act of 1948 encouraged states to control water
pollution although no funding was provided and enforcement was left up to the individual
states; however, in 1956 states were authorized to set up water quality standards (Gaddie
and Regens, 2000). The Wetlands Loan Act of 1961, which provided interest free loans
for wetland acquisition and easements, was strengthened in 1965 by congress. The Fish
and Wildlife Coordination Act, as amended in 1958, required that wildlife be given equal
importance with other aspects of development when deciding on development of water
resources (Hey, et al., 1999).

The Water Bank Act of 1970 was in essence a small scale wetlands protection
policy with the goal of protect migratory waterfowl habitat. In exchange for not draining
wetlands along the Mississippi, landowners received financial compensation (Hey, et al.,
1999). This act, however, extended no further than the Mississippi river wetlands.

The National Environmental Policy Act (NEPA) of 1970 established, among
other things, the Council on Environmental Quality. One of the requirements of the
NEPA is that the effect on the environment must be considered in any major Federal
action (Hey, etal., 1999).

The first international treaty regarding natural resource conservation and use was
the Ramsar Convention in 1971, including the wise use of wetlands (Barbier, 1997). In
1972, amendments to the Federal Water Pollution Control Act (FWPCA) gave the US.
Army Corps of Engineers (USACE) and the Environmental Protection Agency (EPA) the
authority to regulate the pollution of waters in the United States (Gaddie and Regens,
2000; National Research Council, 2001). F WPCA was touted as the single most

important federal act protecting and restoring wetlands in the US. (Hey, et al., 1999).

The goal was to clean up the surface waters of the US, making them “fishable and
swimmable” and to eliminate pollution discharge by 1985. The act gave the majority of
the authority over development of wetland areas and wetland protection to the US.
Environmental Protection Agency (EPA), except for the issuance of development permits
(section 404), which is controlled by the Secretary of the Army, through the Corps of
Engineers. Section 404(b) gives the EPA the ability to set the standards, and section
404(c) gives the EPA veto power over Corps permits (Hey, et al., 1999).

An amendment to the FWPCA, section 404 (PL95-217), known as the Clean
Water Act, placed no limits to the jurisdiction of section 404 (Hey, et al., 1999). The
Clean Water Act, specifically sections 404 and 401 , and the Rivers and Harbors Act,
Section 10, are the primary federal laws regulating development of wetlands (Denison
and Schmid, 1997). The Endangered Species Act of 1973 (National Research Council,
2001), the National Environmental Policy Act (NEPA), the Coastal Zone Management
Act (CZMA), and the Swampbuster provisions of the Food Security Act of 1985, as
amended in 1990 and 1996, among others, are additional federal laws that may apply to
development of wetlands, depending on the circumstances (Denison and Schmid, 1997).

From the 1970’s through the 1980’s, 54% of the wetland loss was due to
agricultural conversion. The Swampbuster provisions of the Food Security Act in 1985
was designed to slow the conversion of wetlands to agricultural uses, and there is
evidence that wetland loss due to agriculture has slowed (Dennison and Schmid, 1997).
About 90,000 wetland acres were added nation wide from 1987 to 1990, due to programs

designed to restore wetlands, under the Food Security Act (Dennison and Schmid, 1997).

10

The North American Wetlands Conservation Act of 1986 was written to protect,
enhance, restore and create six million acres of wetland for waterfowl in North America
(Hey, et al., 1999). A “no net loss” goal was announced by President George H. W. Bush
in 1989, who then established an interagency task force to determine what would be
necessary to achieve that goal (Hey, et al., 1999).

Michigan began regulating dredging, filling and discharging activities into the
Great Lakes in 1966 (Gaddie and Regens, 2000). In the state of Michigan, the Geomare-
Anderson Wetlands Protection Act was passed in 1979, and subsequently became Part
303, Wetlands Protection, of the Natural Resources and Environmental Protection Act,
1994 PA 451, as amended (Michigan Department of Environmental Quality, 2005 (b)).
Coastal wetlands may also be protected under Part 323 (of the same Act), Shorelands
Protection and Management (Michigan Department of Environmental Quality, 2005 (b)).

In 1984, Michigan was given the authority to manage the Federal Clean Water
Act through out much of the state (Michigan Department of Environmental Quality, 2005
(b)). The Michigan Department of Environmental Quality (MDEQ) is allowed to issue
permits pertaining to wetland alteration, instead of requiring applicants to file a federal
application (Gaddie and Regens, 2000). While MDEQ has jurisdiction over in-land
wetlands, overlapping regulation in coastal zones requires a joint permit with the federal
government, and the US Army Corps of Engineers retains jurisdiction over the Great
Lakes and its coastal wetlands in conjunction with MDEQ (Michigan Department of
Environmental Quality, 2005 (b)).

While state and federal regulations require permits, local governments may also

regulate small wetlands not covered by the federal government or the State (Dempsey

11

and Kaplowitz, 2002). Almost forty townships and cities in Michigan have local wetland
regulations that require permit applications to be made with the local community

(Department of Environmental Quality, 2005 (b)).

Protecting Wetlands in Michigan: Programs and their Foci

Aside from laws and regulations, coastal wetlands are protected through a variety
of programs with various methods for conserving wetlands. There are around thirty
Federal, State, and private-organization programs protecting wetlands and wetland
services in the state of Michigan, all of which are voluntary. Table 1 is a list of programs
protecting or restoring wetlands in Michigan (Bailey, 2003).1 The programs have been
classified into six categories based on their primary focus: wetlands, waterfowl, natural
resources, water/ watersheds, wildlife habitat, and coastal land. These categories are not
mutually exclusive, and many programs have interests in more than one.

There are also other programs in the State of Michigan which aid the conservation
of Great Lakes coastal wetlands through the dissemination of information. These types
of programs, such as US. Geological Survey Great Lakes Science Center (which
provides information about biological functions in the Great Lakes Basin, including
information on wetlands), the Great Lakes Information Network and the Great Lakes Sea

Grant Network are not presented here.

 

' To the best of my knowledge, the list is as inclusive as possible, and much of the information is based on
work conducted by Deb Baily, 2003.

12

Table 1. List of Wetland Protection Programs Covering Michigan Coastal Wetlands.

 

 

Federal Programs Classification
United State Army Corps of Engineers
Continuing Authority Programs (CAP) Water/ watersheds
US. Department of Agriculture (USDA)
Natural Resource Conservation Service (NRCS)
Wetland Reserve Program Wildlife habitat
Wildlife Habitat Incentive Program Wildlife habitat
Farm Service Agency (CRP)
Conservation Reserve Program Wetlands
gzcgfrinégggvafim Reserve Enhancement Water/ watersheds
United States Fish and Wildlife
Coastal Program
Great Lakes Wildlife habitat
Coastal Zone Management Program (CZM) Coastal land
Division of Bird Habitat Conservation (NAWCA) Waterfowl
United States Environmental Protection Agency’s Great Wetlands
Lakes National Program Office
State Programs
Michigan Department of Environmental Quality (MDEQ)
Coastal Management Program, Section 306 and 306A Coastal land
Restoration / Enhancement Programs Water/ watersheds
Coastal Restoration Grants Coastal land
NOAA Coastal Management Fellowship Project Coastal land

Michigan Department of Natural Resources

Trust Fund (MNRTF)

Michigan Natural Features Inventory

Natural resources

Wildlife habitat

 

13

Table 1. (cont’d)

 

Non-Profit Programs Classification

 

The Conservation Fund

Great Lakes Revolving Loan Fund Coastal land

Michigan Dune Alliance . Coastal land
Ducks Unlimited

Partnership with FSA/MDA CREP Waterfowl

North American Wetland Conservation Act Waterfowl

Great Lakes Commission

Coastal Wetlands Study Project Water/ Watersheds
Great Lakes Aquatic Habitat Network and Fund Water/ watersheds
The Great Lakes Wetlands, Conservation Action Plan Wetlands
Great Lakes Wetland Policy Consortium Wetlands
Hampton Township, Bay County Coastal land
The Nature Conservancy Wildlife habitat

The National Wildlife Federation Great Lakes Field Office Water/ watersheds

 

Programs associated with the category “wetlands” refer to those with foci such as
protecting environmentally sensitive areas, maintaining environmental integrity,
protecting economic“ and aesthetic values, increasing public awareness, protecting,
maintaining, or restoring chemical, biological, and physical habitats, etcetera. The Great
Lakes Wetlands Conservation Action Plan is one program with the goal of securing,
protecting and rehabilitating coastal wetlands in the lower Great Lakes, and does so
through the purchase of land or easements as well as by working with private landowners

(Ontario Nature, 2003). Other programs in this category involve land and easement

14

purchases, although the Farm Service Agency’s Conservation Reserve Program uses ten
to fifteen year contracts with cost sharing (Bailey, 2003). All programs identified as
“wetland” programs deal with private landowners.

Programs focusing on preserving, restoring, and recreating waterfowl habitat are
categorized under “waterfowl”. Priorities include waterfowl migration, wintering habitat,
enhancing wetland habitat diversity, and benefiting waterfowl. The programs listed deal
with private landowners. Ducks Unlimited uses ten year agreements and provides
financial incentives, and its North American Wetlands Conservation Act uses cost-share
funding (Ducks Unlimited, 2003). Programs listed as “Waterfowl” all involve
purchasing of land or easements, and the US. Fish and Wildlife NAWCA deals with the
purchase of land or easements exclusively (Bailey, 2003).

“Natural resource” programs focus on the protection of natural resources and
open space. Priorities include promotion of better understanding, conservation and use of
the Great Lakes resources, and use of community based projects in achieving results in
clean air, climate change, clean water, and nature. The MNTRF requires 25% matching
funding on acquisition or development projects and involves the purchase of land
(Bailey, 2003).

Programs focusing on restoring, recreating, and protecting “water/watersheds”
and wetlands include priorities such as addressing the threat of water diversion, education
of the public about effects of toxic substances, and protection of endangered species and
habitats. Over half of these programs deal with private landowners, and several programs
deal primarily with education. The Army Corp of Engineers Continuing Authority

Programs vary from year to year, and was intended to allow the Secretary of the Army,

15

through the Corps, to conduct certain types of water resources improvements as
necessary (Bailey, 2003). One of the “water/watershed” programs uses fourteen to
fifteen year contracts, sign-up incentives, and a set level of cost-sharing. Other programs
are local grassroots initiatives with limited fimding, focusing more on education than
purchase of land.

“Wildlife habitat” programs in Michigan primarily focus their attention on
rehabilitation, conservation, and protection. Priorities include: all natural resources and
working lands, especially with greatest wetland function and optimal wildlife habitat;
coastal wetlands and lake plain, prairies, open dunes, fens, etcetera; coastal habitat to
conserve fish and wildlife; and contaminated sediment remediation, stewardship and
runoff control to protect habitat. Three programs deal with private land owners, and the
Wetland Reserve Program (USDA, NRCS) program utilizes ten year cost-sharing, thirty
year conservation easements, and permanent easement restoration (landowners control
access). The Wildlife Habitat Incentives Program utilizes five to ten year cost-share
payments (Bailey, 2003). 1

Programs in Michigan that protect “coastal lands” focus on advancing coastal
management objectives including: maintaining and strengthening coastal management
capabilities of coastal lakes, rivers mouths, bays, floodplains, etcetera; high quality
coastal land acquisition; and conservation sites of high ecological significance. Two
programs included in this category are the Great Lakes Revolving Loan Fund and the
Michigan Dune Alliance, both run by the Conservation Fund (The Conservation Fund,
2003 (a); The Conservation Fund, 2003 (b)). Most programs listed as “coastal lands”

utilize land purchases; two involve the purchase of easements, while three deal with

16

private landowners. The MDEQ’s Coastal Management Program gives one-third of the
money from the Federal Government (given to the state for DEQ programs) to local
communities of the Great Lakes shoreline (Michigfl Department of Environmenta_l
Quality, 2003 I. The DEQ receives $2.4 mill per year that must be matched 1:1 with state
and local funds. Enforcement is given to local communities (as with many of the
projects). Another program, MDEQ’s Coastal Restoration Grants, only requires 25%

matching funds (Bailey, 2003).

PROGRAM ELEMENTS

The programs discussed in the previous section utilize an assortment of methods
for preserving or restoring wetlands with a range of main objectives. While the previous
section presented programs that all protect wetlands in Michigan, they vary on the
environmental service that they focus on preserving or protecting. Here, the primary
focus of a program refers to the wetland function that is prioritized by the program. The
programs in the previous section were categorized by their primary focus, and will not be
re-visited here.

The programs presented above also differ from each other in their approach to
protecting wetlands, whether it is preservation or restoration (both described in the
following section) of wetlands or some combination of the two. Aside from the focus of
the program and the type of protection, the programs differ in the methods they use for
acquiring land that will be preserved or restored. Many of the programs acquire land
through the purchase of property, while others utilize easements or land contracts. Most

utilize a combination of methods to secure wetland acreage. In what follows,

17

descriptions of various program elements seen in actual programs in Michigan are
discussed in order to firrther lay the ground work for the use of these program elements to

create new coastal wetland protection programs in the state of Michigan.

Preservation and Restoration

Programs protecting wetlands can focus on either preserving the current condition
of a wetland by preventing its degradation, or on restoring wetlands that have previously
been harmed. Preservation of wetlands is just what it implies, preserving a wetland in its
current state, preventing further degradation. Wetlands of varying size and with varying
functions are preserved at their current levels. No effort is inputted into increasing the
wetlands’ functions.

Restoration of wetlands involves taking a degraded or destroyed wetland and
restoring its functions, or some of its functions, to some previous level of quality. One
example of restoration of wetlands is its use in mitigation projects (Mitsch and Wilson,
1996). Mitigation involves restoring a wetland (or creating a new one) in one location to
compensate for the loss or degradation of a wetland in another location. Restoration
projects take time and should be critiqued as success or failure only after a sufficient
amount of time has passed, such as fifteen to twenty years or more (Mitsch and Wilson,
1996)

There is currently a debate over what entails proper restoration efforts, such as
what the ending point should be, what sort of functions need to be included and at what
level (Mitsch and Wilson, 1996; Higgs, 1997). Multiple definitions of wetland

restoration are available, from the very general to the very specific, but a consensus does

18

not exist on what exactly defines proper restoration (Higgs, 1997). Both preservation and
restoration programs typically involve some sort of monitoring to make sure the wetland
in not being damaged or degraded any further.

Currently in the US. efforts from the Great Lakes Commission are attempting to
influence Congress to accept nine specific goals to guide legislative action, enforce
current laws and appropriate funding accordingly (Great Lakes Commission, 2005).
Among other things, the Great Lake Commission would like to see funding approved for
activities such as those to restore and enhance coastal lands.

Michigan Senators Carl Levin and Debbie Stabenow proposed new legislation in
the 109th Congress on January 31, 2005. They, along with support of other senators,
propose a bill to amend the Federal Water Pollution Control Act, the Great Lakes Water
Quality Indicators and Monitoring Act, which would result in science-based assessment
of the health of the Great Lakes. The proposal might also increase the focus on

ecological restoration activities in the Great Lakes.

Land Acquisition

Projects attempting to protect the environment, by their nature, can only target a
small number of people, in a small area, for a limited time frame (Kiss, 2002). Non-
project approaches, as they are sometimes referred, are those that offer something, such
as a financial benefit, to individuals to enable them to attain specific goals without telling
them how to go about it (Kiss, 2002). Financial incentives may be in the form of the

purchase of land or specific right to the land, as well as tax breaks for preservation. Non-

19

project approaches to environmental conservation include land purchase, conservation
easements, tax-credits, and subsidies.

In the United States, conservation organizations such as The Nature Conservancy
and the Conservation Fund preserve land primarily through direct purchase, leasing, or
easements of the land (Ferraro and Kiss, 2002). There are many methods that can be
utilized for land acquisition for the purpose of conservation. In the US and most
industrialized countries, the purchase of land, easements, leasing, and subsidies are the
most commonly used methods for conservation (Ferraro and Kiss, 2002). Below are a

few options program designers have, and the benefits and drawbacks of each.

Purchase of Property

The most widely known method of land acquisition for conservation purposes is
the purchase of property. Market value of the land is typically paid, the land changes
hands, and becomes controlled by an agency or organization. Becausefull value of the
land is required to purchase it for conservation, the costs are relatively high. There are,
however, relatively few subsequent costs. Enforcement and monitoring costs are

relatively nil since the land is now owned by the conserving group (Boyd, et al., 1999).

Conservation Easements

Conservation easements are the purchase of particular rights for an area of land
(Ferraro and Simpson, 2001). Public agencies, as well as private organizations, can

purchase specific rights to land but not purchase the land itself. An easement may

20

purchase development rights, limit possible development by the owner, or have
provisions to allow for public access (Parker, 2004). Conservation easements are
typically a permanent purchase of specific rights to the land. The landowner is usually
bound by contract to maintain the property, and prevent ecological degradation. The
group or agency purchasing the easement is often contractually free from pre-existing
liabilities, such as previous environmental contamination of the land work (Boyd, etal.,
1999)

According to Morrisette, conservation easements offer the best, and sometimes
only instrument, for preservation of private lands (Morrisette, 2001). Because easements
are voluntary, they are easier for land owners to accept than direct regulations restricting
land use (Boyd, et al., 1999). Conservation easements are less expensive than out-right
purchase of the land, and allows the current owner (and subsequent owners, depending on
the contract) to continue using the land in accordance with the provisions of the
conservation easement contact. The costs associated with conservation easements, above
the price for the specific rights, are in monitoring and enforcement as well as the cost of
contracting for the land. Administrative burdens for conservation easement, while
greater than for the out-right purchase of land, are lower than for most other methods,
such as tax incentives or tradable development rights (Boyd, et al., 1999).

Conservation easements are becoming increasingly preferred as the primary
preservation method, over the direct purchase of land. Land trusts, protecting various
elements of our environment, currently preserve more land by conservation easements
than outright land ownership (Parker, 2004). Land trusts are more likely to obtain land

through conservation easements when the goal of the conservation is for services such as

21

creating scenic amenities, working lands, or conserving large areas of land (Parker,
2004). Parker also notes that land trusts preserving or enhancing wetlands, water quality,
and rare species are just as likely to use conservation easements as purchase of land

(Parker, 2004).

Land Contracts

While the purchase of land is fairly straight-forward, and permanent easements
are the purchase of specific rights to land without actual possession of it, land contracts
vary greatly in their scope and duration. Land contracts are less expensive than the land
acquisition methods previously mentioned because they are usually for a specific period
of time (such at ten or fifteen years), and do not go on indefinitely. Private organizations
utilizing land contracts are only purchasing specific rights to the land for a specific period
of time. The landowner agrees to restrict their use of the land for a specific period of
time, in exchange for financial re-imbursement. In regards to preventing the conversion
or degradation of a wetland, there is no guarantee of what will happen to the wetland
once the land contract has expired.

Land contracts can also involve cost-sharing between the land owner and the
organization or agency holding the land contract. For example, the US. Department of
Agriculture has a voluntary Conservation Reserve Program (CRP) that provides rental
payments and cost-share assistance (for approved conservation practices) to conserve

resources on private land (Bailey, 2003). The contracts last ten to fifteen years.

22

RANKING AND VALUING ENVIRONMENTAL SERVICES

Michigan Coastal Wetlands

Federal and State laws and regulations, as well as wetland protection programs,
attempt to protect wetlands (Bailey, 2003; Dennison and Schmid, 1997; Gaddie and
Regens, 2000; Michigan Department of Environmental Quality, 2005 (b)). When
developing their programming, organizations seeking to protect wetlands often pick a
primary focus for wetland protection that reflects what they value most, or the fimctions
they perceive to be undervalued in the current market and in need of protection. The
optimal level of protection of wetlands depends upon the values society places upon the
wetlands and their functions. If wetlands had no values (ecologically, in use or non-use)
there would be little desire to protect them. Undervaluation of environmental amenities
can lead to overexploitation or conversion of the resource.

Coastal wetlands have many uses and values that are often difficult to estimate
(Woodward and Wui, 2001). As previously mentioned, wetlands provide water quality
improvement, flood storage, water supply, and erosion protection, as well as other
important fiinctions (Keating, 1995). Michigan’s coastal wetlands have recreational
fishing and hunting values that can be estimated through the willingness of the sportsmen
to pay for licenses and fees. Revenue from bird hunting licenses in a particular area is
one way to try and estimate the wetlands value for waterfowl, but this approach could
exclude non-hunted birds, bird-watching values, and values for photography. The value
of water quality improvements can be estimated by replacement cost, such as considering

how much it would cost to build and run a water treatment facility that provides the same

23

level of clean water, but this method does not include values that wetland provides for
things such as wildlife habitat.

Michigan’s coastal wetlands also have value not necessarily captured in the price
of land, in the value of the licenses people buy, or related to the cost to re-create the
functions lost. People often times do not have full information about the benefits of
wetlands, and even when the benefits are known, they are very difficult to value
accurately (Farber and Costanza, 1987). Non-use values, such as existence and option
value, are also an important part of the total value of a wetland, or any environmental
amenity. Values, such as non-consumptive recreation (bird-watching, nature study, etc.)
are an essential element in the total value of a wetland, even though they are difficult to
quantify (J aworski and Raphal, 1978). Non-use values such as existence value and
bequest value are even more difficult to estimate because many people who value
wetlands for these reasons may not even visit a wetland in their lifetime. Non-use values
cannot be estimated using market valuation techniques, because an individual’s value for
non-consumptive uses may not be noticed in the marketplace. Other methods are

required to obtain a more accurate estimate of coastal wetlands value.

Environmental Valuation

There are multiple methods for valuing or ranking environmental amenities.
Many valuation methods focus on market values. Others focus on non-market values
such as use values which can be estimated by travel cost method, hedonic price analysis,
and replacement cost estimation, among others. Revealed Preference (RP) methods look

at actual real-world activity by individuals in determining people’s preferences and

24

willingness to pay for particular amenities or services (Adamowicz, et al., 1998 (a)).
Hedonic price method and travel cost method studies fall under the category of RP.

RP methods address actual, rather than stated, activity and are beneficial because
the values and preferences revealed are based on observable (behavior, and not based on
hypothetical situations. This, however, is also a shortcoming as many interesting areas of
research are hypothetical or can not be seen in the marketplace, and therefore no data

from actual activity can be collected.

Stated Preference Modeling

Estimation of non-use values is a relatively new (around three decades) addition
to environmental valuation (Loomis, et al., 2000; Woodward and Wui, 2001). Non-use
values of environmental amenities or services are not revealed through RP studies. A
person may value a wetland for its existence alone, but because they never visit it, the
value they place upon it might not be counted. Stated Preference (SP) "methods can
incorporate total value, including the non-use value, by surveying a wide range of
individuals, not necessarily just those who use the wetland. Preferences for
environmental amenities, revealed by various types of SP studies, indicate relative value
for those amenities. SP methods allow researchers to look at hypothetical situations, in a
controlled environment.

Introduction to Attribute-Based Stated Choice Methods by Wiktor Adamowicz, et
a1, 1998, looks at how SP methods are used in the valuation of environmental services.
Estimating non-use values for environmental services results in a “ranking” of services.

SP methods include a wide range of research, from Open-ended Contingent Valuation to

25

Stated Choice (SC) methods, which includes Contingent Valuation (CV) and attributed
based methods (Adamowicz, et al., 1998 (a)). CV is a SP method which incorporates a
monetary value into the presented scenario (Adamowicz, et al., 1998 (a)). It is the most

widely used SP method for valuing the environment (Carson, et al., 1994).

Stated Choice Method

There are multiple methods for valuing and ranking environmental amenities.
Many CV studies measuring the value of particular services provided by wetlands (such
as recreational values) are available, however, numerous studies value only a single
function (Whitehead, 1993). CV is very useful in determining the monetary value of the
damage, but Adamowicz et al. state it may not be as appropriate for determining the most
preferred combination of resource attributes. Adamowicz et al. look at attribute based SC
methods for valuing environmental damage, and for determining the attributes that could
compensate for a lost or degraded resource (Adamowicz, etal., 1998 (a)). Choice
methods are different than CV methods in that environmental attributes are varied, based
on an experimental design (Blarney, et al., 1999; Boxall, et al., 1996). Attribute based SC
methods provide individuals with two or more sets of attributes to choose among, and
their choices can be used to rank the attributes presented. A SC ranking of preferences
for environmental attributes provides and indication of what people value most, without
the need to include a monetary value.

Attribute based SC method allows researchers to determine the relative
importance of the program attributes based on the choices made by individuals

(Morrison, et al., 1999). Choice experiments are based on random utility theory

26

(Morikawa, et al., 2002), and utility maximization (Adamowicz, et al., 1998 (a)) which
assumes that individuals will select a set of attributes that provides them with the most
utility. SC studies collect data from individuals by asking them to choose between a set
of alternatives. Their choices are based on the attributes presented in the alternatives.
Through the summation of individual choices and the experimental design, ranking of the
attributes becomes possible. Ranking of attributes may be preferred by some researchers
or policy-makers who are opposed to monetarily valuing the environment (Blamey, et al.,
1999)

SC methods avoid many of the problems present in other methods that look at
actual real world activity (observations of choices that are actually made). First, the
choice set is clearly identified in the questionnaire. Second, the design of the model
avoids the problem of co-linearity between variables by generating orthogonal attribute
data. Third, the design of the survey produces a sufficient number of observations for all
attributes (Adamowicz, et al., 1998 (b); Eamhart, 2001).

Choice modeling is ideal for estimating respondent’s ranking of various attributes.
When a monetary attribute is involved, it provides a technique for estimation of use
and/or non-use values that is both flexible and cost-effective, especially when involving
multiple alternatives (Blarney, et al., 1999). The SC method provides one method for
ranking attributes. Values people place on particular environmental amenities indicates
what their preferences are for those amenities. Monetary values give an indication of

preference, and preferences give an indication of relative value.

27

Examples of the Stated Choice Method

Studies valuing freshwater ecosystems are somewhat limited in scientific
literature (Costanza, et al., 1998) and additional research in valuing freshwater amenities
would provide firrther insight into the value that individuals have for our environment. In
recent years more studies valuing environmental non-use values using SP methods have
been conducted. Stated Choice methods were first developed in marketing and
transportation literature (Louviere and Hensher, 1982; Louviere and Woodworth, 1983).
Not until recently have Stated Choice methods been used in environmental valuation
(Adamowicz, et al., 1998 (a); Adamowicz, et al., 1998 (b); Adamowicz, et al., 1994;
Boxall, et al., 1996; Swallow, et al., 1994).

Projects comparing SP methods to RP methods have demonstrated that SP
methods provide accurate estimates of value. Adamowicz et al., compared results from a
SP study to those of a RP study to evaluate the effect on recreational fishing values from
water resource development. After correcting for error variances, both methods resulted
in similar preference structures (Adamowicz, et al., 1994). Another study conducted by
Adamowicz, in combination with Swait, Boxall, Louviere and Williams, estimates RP,
SP and a combined RP and SP model of recreational site selection (Adamowicz, et al.,
1997). After accounting for variance heterogeneity, it was found that RP and SP methods
were similar and that the combined method of RP and SP performed the best (much
better than RP by itself). There are, however, high costs associated with the joint RP and
SP method, and that cost needs to be weighed against the possible benefits from joint

analysis.

28

As previously mentioned, CV is the most widely used SP method. Adamowicz et
al. evaluated SC experiments by comparing the results fiom a study estimating the
passive use values for an environmental amenity using SC methods with the results from
the CV method (Adamowicz, et al., 1998 (b)). The estimated models for the choice
experiment performed well (Adamowicz, et al., 1998 (b)). When comparing the choice
experiment and CV method, while accounting for error variances, there is not a
significant difference in preference over income between the two methods. This gives
credence to both methods, indicating that they perform equally well under the given
circumstances.

Boxall et al. examined the effect a change in environmental quality on non-market
recreational use values (Boxall and et al., 1996). The estimation of the CV approach and
the Choice approach resulted in different estimates at first, but after corrections were
made for respondents not taking into account substitute possibilities, the estimates in both
methods were similar (Boxall and et al., 1996).

A choice experiment on wetland mitigation, conducted by Lupi, Kaplowitz and
Hoehn, related the characteristics of a drained and a restored wetland to the acceptance of
a mitigation project (Lupi, et al., 2002). It revealed that the public is concerned with the
attributes of wetlands. Among other things, they discovered that habitat was and
important attribute to the public (Lupi, et al., 2002).

A study conducted by Morrison et al. estimates the non-use values provided by
Macquarie Marshes in Australia using choice modeling with costs attached to each option
(Morrison, etal., 1999). The results of their study were statistically significant variables

with high explanatory power, consistent with expectations, giving credibility to the

29

employment of choice modeling in valuing environmental non-use values. Another study
found utilized a choice experiment valued the water supply in the Australian Capital

Territory (Blamey, et al., 1999). This is study provides a ranking of program options.

WETLAND VALUATION

While the literature is somewhat limited, studies valuing wetlands are growing in
number (Woodward and Wui, 2001). Valuation of specific wetlands has been conducted
by multiple studies, but most do not provide insight into the relative values of wetland
types, functions or services (Hoehn, et al., 2003). Estimated wetland values vary greatly
across studies, and estimate various services (Woodward and Wui, 2001).

Farber and Costanza state that the cost of slowing or reversing wetland loss must
be weighed against the social value of the wetlands they are trying to protect (Farber and
Costanza, 1987). They estimate the willingness-to-pay for wetland services in
Louisiana’s coastal zone, attempting to value the entire wetland system. Their research
estimates the value per acre people are willing to pay at above the current price of a
wetland acre, $590 compared to $200.

A study conducted by Azevedo, Herriges and Kling models the recreational
demand for wetlands in Iowa, and combines the use of revealed and stated preference
methods (Azevedo, et al., 2003). They utilized the travel cost method, and a contingent
behavior question by asking people how many times they would visit a wetland if the
cost for visiting increased to X dollars. Their results found no consistency between the

RP and SP methods. Data between the two methods was inconsistent, with results on

30

wetland value from each model being different. The source of the inconsistency is not
known, and could be the fault of either RP or SP data, as well as differences in their
estimation methods.

Woodward and Wui estimated the value of wetland services (a meta-analysis)
and found that wetlands have a high value of consumptive (such as recreational fishing)
and non-consumptive (such as flood control) services (Woodward and Wui, 2001).

In reference to coastal wetland valuation studies, those found were for the
valuation of saltwater wetlands, and these dealt primarily with only one aspect, such as
the value of oyster production (Batie and Wilson, 1978), the value of hurricane wind
damage protection (F arber, 1987), and the value for waste treatment (Gosslink, et al.,
1974)

A compilation of studies, by Wilson, showed that there have been 30 published
studies from 1971 to 1997 valuing non-market goods for freshwater systems in the US.
(Wilson, 1999). Valuation methods utilized by the various studies were the Travel Cost
method, Hedonic method and CV method (Wilson, 1999). Whitehead looked at multiple
methods that have been used for wetland valuation. Methods included in the study were
the Net Factor Income approach, CV, Travel Cost, Hedonic Price, and Damage Cost
methods (Whitehead, 1993). None of the studies or methods reviewed estimate value for
freshwater systems using Stated Choice methods. While none of the studies utilized SC
methods, their value estimates provide an indication of the public’s preferences for the
amenities. Those amenities with the highest valued would be the most preferred.
However, because many studies are narrowly constructed to value one or only a few

functions, it is difficult to rank a range of attributes.

31

More recently, Morrison conducted a study estimating the value of improved
wetland quality in the Macquarie Marshes through Choice Modeling (Morrison, 2002).
Morrison asserts that it is possible to estimate preferences for various wetland attributes,
to estimate how these preferences vary across different groups, and to estimate what level
of community support might be received for particular management programs (Morrison,
2002)

Because it is difficult to estimate values for an entire wetland system, there is
little information about wetlands overall value (F arber and Costanza, 1987). Farber and
Costanza state that more research into values of wetlands is needed to increase the

accuracy of estimates (Farber and Costanza, 1987).

Great Lakes Wetland Valuation

Currently, there are no known studies (to the best of my knowledge) estimating
preferences for Great Lakes coastal wetland program characteristics uSing the Stated
Choice method. In fact, no studies were found utilizing the State Choice method in
regard to any Great Lakes coastal wetland functions or values.

Adamowicz et al. assert that estimating people’s preferences for particular
resource attributes is necessary to determining adequate compensation for resource
degradation or loss, by compensating for the loss or degradation in ways the public most
values (Adamowicz, et al., 1998 (a)). Research determining preferences for wetland
protection, and estimating the value, including non-use value, of wetlands is important
and necessary to developing programs that will further protect our Great Lakes coastal

wetlands. There is a large gap in information regarding preferences for Great Lakes

32

coastal wetland (or really any wetland) programs. Stated Choice studies have only
recently been utilized for environmental valuation, and have yet to be used in Great

Lakes coastal wetland studies. This study attempts to fill in some of that gap.

33

CHAPTER 3

SURVEY METHODS

Survey Questionnaire Design

The survey questionnaire was designed to be as “user-friendly” as possible, so
respondents could easily understand and follow it (See Appendix for the complete
questionnaire). The questionnaire consisted of fifty-three questions, organized under
twenty-four headings. The first section of the questionnaire was designed to gather
information on respondents’ knowledge of and experience with Great Lakes coastal
wetlands. Questions also collected information on respondents’ knowledge of the types of
program characteristics associated with wetland protection and restoration.

Survey booklet question number eight, the focus of this research, was a “stated
choice question,” which will be explained in detail later. This question was designed to
gain information on the respondents’ preferences for the characteristics of Great Lakes
coastal wetland programs by asking respondents to choose between two program
alternatives. The previous pages in the questionnaire provided the respondents with the
information on the various parts of the program alternatives, so that they were able to
differentiate between the two programs.

Question nine of the survey booklet was a contingent valuation question, designed
to gather information on the values respondents place on particular wetland program
attributes. Question nine took the program presented as Program A in question eight and

included varying monetary values. Survey booklet question ten prompts respondents for

34

input as to their reasoning behind their answer in question nine. This information can be
used to gain insight as to the respondent’s basis for their decision. The remainder of the
survey booklet questions asked for background information on the respondents (e.g.,
demographic characteristics). This information is a valuable tool for indicating validity
of the sample, by determining if the respondents are a representative sample of the

statewide population.

SURVEY QUESTIONNAIRE PRE- TESTING

High quality survey instruments typically go through pre-testing as part of the
design and development process (Kaplowitz, et al., 2004). Pre-testing helps to ensure
that the questionnaire language is understandable and that the questions asked are
realistic and meaningful. For the Great Lakes coastal wetland survey, two different types
of pre-testing were used.

First, pretest interviews were conducted using intercept interviews with
participants randomly selected in public venues. For this, the researchers used two
different “food courts” at shopping malls as the pre-testing locations and successfully
completed twenty pre-testing sessions. Interviewers approached individuals in the food
court areas, and asked them if they would be interested in filling out a public policy
survey. Potential participants were told that they would receive an honorarium of ten
dollars for participating. After participants agreed to the pretest interview, they received
the ten dollar stipend as a good faith gesture. Participants next completed a draft of the

survey questionnaire. After they finished the survey questionnaire, they were asked a

35

series of questions (see Table 2 below) pertaining to the questionnaire, its design, and

their opinion of it.

Table 2. Interview Discussion Guide.

 

Survey Discussion Guide

1. Overall impression of the questionnaire

A.
B.

C.

What did you think of the questionnaire?

Was there any information in the questionnaire that seemed odd or
awkward to you?

Did the information that was provided help you to answer the questions?
Describe how.

D. Were there questions where you would have liked more information?

B.

What information?

Were there any areas where too much information was provided?

2. Wetland program
A. Now, I want to talk about the question near the end that asked how you

0

would vote on the wetland program. Tell me about what went through
your mind as you made your decision.

. In your own words, describe to me what the program would provide.
. How would coastal wetlands change if the program were implemented?

. How would the services provided by coastal wetlands change with the

program?

. Talk me through the table describing the program. What do each of these

mean to you?

How would the program affect you?

G. Was the program realistic to you?

3. Other

The questions asked of respondents during the pre-test interviews were designed

to determine if respondents understood the information and questions in the

questionnaire, as well as to determine if the correct amount and type of information was

36

provided to respondents. The pro-testing surveys were also intended to evaluate whether
the stated choice and non-market valuation were designed properly and resulted in a
range of useful feedback.

The second pretest interviews were conducted using telephone recruitment of
randomly selected individuals in mid-Michigan. The interviewees were invited to the
Michigan State University campus for scheduled individual interviews. Three sessions of
individual pre-test interview were used, the first on November 20, 2003, the second on
December 4, 2003 and the third on December 10, 2003. The week before the individual
interview sessions, interviewers made telephone contact with randomly selected residents
in mid-Michigan. Interviewers asked each potential interviewee their age, education
level, and gender. They then terminated any calls for which the demographic had been
filled, to avoid over sampling any one particular group of individuals. Targets were set
so to that only one person fi'om the eighteen to twenty-five year old age group, and no
one with an education level of a Masters degree or beyond were recruited. Individuals
who possess a masters degree or beyond were not recruited because they were not a
group of concern when testing the questionnaire for respondent comprehension, as well
as the fact that the sampling was conducted in a university town, where there was more
likely to be individuals with advanced degrees. Recruiting also sought to achieve an even
distribution of males to females, to correspond with the close to even ratio of males to
females in the state’s population.

Each person contacted by phone was asked if they would be interested in
participating in an interview regarding public policy at Michigan State University. They

were told that they would be asked to complete a survey questionnaire and then

37

individually discuss their answers. Potential participants in the on campus pretest
interviews were offered a twenty dollar honorarium for their time and effort. As in the
case of the intercept pretest interviews, on campus interviewees were asked to complete a
self administered written questionnaire and were then interviewed about their experience
with the survey instrument. In total, twenty-three on campus pretest interviews took
place, with questionnaire administration taking about ten minutes and with each interview
lasting roughly twenty minutes. Based on the pre-testing interviews, modifications to the
survey questionnaire were made for the purpose of clarity and to improve the overall

design.

IMPLEMENTATION

The initial sample for the mail survey consisted of 1,650 individuals randomly
drawn from a list of names and addresses obtained from the Michigan office of the
Secretary of the State. The list was made up of a random selection of Michigan residents,
twenty-one years or older, who possess a valid drivers license or State identification card.
A stratified random sample was drawn from this list, with sample strata that corresponded
to the seven most populous counties in Michigan (See Table 3). The strata ensured that
specific counties with large percentages of Michigan’s population were accurately
represented, so as not to over or under sample them. The number drawn from each
county directly corresponds to the percentage of the state population that county
comprises. The eighth strata contained individuals from the remaining counties in
Michigan and represented 43.21% of the sample. Because the sample strata proportions

correspond to the actual proportions in the population, no weights are needed in the

38

analysis of the stratified sample data (i.e., the weights to adjust for the stratified sampling

all equal one).

Table 3. Stratified Random Sample, Depicting the Seven Most Populous Counties.

 

 

County % of States Stratified % of overall
County . Sample

Number Population D Sample

rawn

Genesee ' 25 4.31% 71 4.30%
Ingham 33 2.91% 48 2.91%
Kent 41 5.61% 92 5.58%
Macomb 50 8.15% 135 8.18%
Oakland 63 12.17% 201 12.18%
Washtenaw 81 ‘ 3.43% 56 3.39%
Wayne 82 20.2% 334 20.24%
A“ (”fer 43.21% , 713 - 43.21%
counties
Total 100% 1,650 100%

 

Dilman’s survey steps were utilized by this survey to achieve the best response
rate possible (Dillman, 2000). The self-administered survey was implemented through
the mail using the Tailored Design Method (Dillman, 2000). As advocated by the
Tailored Design Method, a respondent-friendly questionnaire was developed, with the
survey utilizing a five contact format.

The mail survey implementation consisted of five contacts: a pre-notice letter, a

questionnaire, a reminder postcard, a replacement questionnaire, and a final contact with

39

an additional replacement questionnaire. To the extent possible, each contact was
personalized. All letters were addressed directly to respondents and printed on high
quality, watermarked paper. Letters and postcards were hand signed, and the large
format questionnaires were printed in color on glossy paper. An incentive, consisting of
three first class stamps, was included with the first questionnaire for half of the
respondents for the purpose of achieving higher response rate (Dillman, 2000).

The first of the five mailing contacts was a pre-notice letter designed to let the
respondent know that they would soon be receiving a questionnaire from Michigan State
University. The personally addressed, hand-signed, pre-notice letters were mailed first
class on March 29, 2004.

The second contact was a survey questionnaire packet, which consisted of a
questionnaire, a hand-signed cover letter, a business reply envelope for respondents to
return the questionnaire, and an incentive of three first class stamps included in half of
the sample. This second wave was mailed to participants first class on April 6, 2004.

The third contact with respondents was a hand signed reminder postcard, and was
mailed to respondents on April 13, 2004. Dillman recommends that this contact be
mailed a few days to a week after the questionnaire (Dillman, 2000). The postcard asked
them to please fill out and return the questionnaire, and thanked them if they had already
done so.

The fourth contact was mailed April 19th. It consisted of a questionnaire, a
business reply envelope, and a cover letter that was changed to reflect the fact that this
was a “replacement survey questionnaire” in case they had misplaced their first copy. No

incentive was included in this wave. Dillman suggests this wave be mailed two to four

40

weeks after the third wave, but this wave was mailed earlier because the previous booklet
was found to contain an error in the valuation section of the questionnaire (Dillman,
2000). The error involved the dollar amounts that appeared several times in the booklet,
and the values did not always match when they were supposed to in some versions of the
booklet. A new booklet was sent as soon as possible to correct for the error.

The fifth and final mailing included a questionnaire, business reply envelope, and
a cover letter indicating that this was the final contact. The entire sample was split into
two groups, half that would receive the final mailing with the regular postage, and half
that would receive the questionnaire packet in a priority envelope, in an attempt to
increase the response rate. The fifth wave was mailed May 11‘“, within the two to four
weeks after the fourth wave that is recommended by the Tailored Design Method

(Dillman, 2000).

DATA ENTRYAND RESPONSE RATE

Each of the five mailings were carefully documented. Mailing dates, mail
returned dates and post-mark dates were all recorded, and information on returned mail
and responses were all recorded and coded.

The American Association for Public Opinion Research (AAPOR) standards were
followed in data entry and in determining response rates (AAPOR, 2004). The AAPOR
code system was utilized by our survey, with some minor modifications and condensing
of categories. A Final Disposition code was recorded for each individual in the survey
sample. All categories, with the exception of 2.31, were condensed to the tenth decimal

place. A coding of 2.31 indicated that the individual was deceased, and for the purposes

41

of this study, the deceased were not counted as being valid possible responses. Table 4,

on the following page, depicts the condensed categories from the AAPOR final

disposition codes for mail surveys that were utilized in determining the response rate.

The minimum response rate is determined by the formula presented in table 4.

The response rate presented here is Response Rate (RR2), which includes partial

interviews. It is the number of completed questionnaires divided by the total completed

plus the total number of non-interviews (minus the deceased) plus the number of cases of

unknown eligibility.

Table 4. Response Rate Coding.

 

RR2 = (I+P)./ ((I+P) + (R+NC+O) + (UH+UO))

 

 

 

RR Response rate # of Sample
I Complete interview (coded 1.1) 540
P Partial Interview (coded 1.2) O

Refusal and break-off (coded 2.1) 22

NC Non-contact (coded 2.2) .0
o Other (coded 2.32)2 9

UH 5213161333? household occupied HU 645

UO Unknown, other (coded 3.2) 4

Invalid (coded 3.3, 3.4 or 4.1 430
including 2.31-the deceased)
Total 1650

 

 

2 “Other” includes individuals with language problems or incapacitated persons ( both coded as 2.32 here).
AAPOR includes those that have passed on as being valid possible respondents under 0 (coded 2.31). For
this study, those who are discovered to be dead are not counted as being valid for response rate purposes.

42

Of the 1,650 persons in the initial sample, 430 had invalid addresses. Invalid
means that the researchers received at least one notification that the address was not
correct (e. g. wrong address, moved). Removing these invalid yielded a sample with
1,220 valid addresses. In total, 540 completed questionnaires were returned yielding an
overall response rate for the survey of 44.26%. Of the 540 completed questionnaires,
only 538 were available for data analysis (2 questionnaires that were entered as received
in incoming mail were subsequently lost), hence, results are based on 538 completed

questionnaires.

The details of the survey questionnaire responses by mailing are as follows:

0 1,650 survey questionnaires were mailed in wave two (the first mailing of the
questionnaire). Of those mailed, 356 questionnaires were returned completed,
and 408 questionnaires, pre-notice letters or post cards (wave one and three) were
returned with a bad address (or were found to be deceased) and coded as invalid,

. which give us a 28.66% response rate for wave two.

0 1,239 questionnaires were mailed in the fourth wave (the second mailing of the
survey questionnaire). Not all of the notices of bad address from wave one, two
and three had been received prior to mailing out the second survey questionnaire.
In addition to the 408 invalid questionnaires from wave one through three, 12
questionnaires were returned invalid in this wave. 137 additional questionnaires
were returned completed. The response rate for wave four was 15.91%.

0 813 questionnaires were mailed in wave five, the final mailing of the survey

questionnaire to non-respondents. This wave resulted in an additional 7 addresses

43

being invalid and 47 additional questionnaires returned as completed. The
response rate for wave 5 was 6.69%

0 Overall response rate for the survey questionnaires was 44.26% (540/ 1220)

Demographic Characteristics of the Sample

Survey samples should be representative of the population for which they are
claiming to estimate values. Table 5 depicts how respondents to this survey
questionnaire compare demographically to the general population of Michigan (US.
Census Bureau, 2000).

As can be seen in Table 5, the demographic characteristics of the survey sample
respondents are reasonably similar to those of Michigan’s general population (Lupi, et
al., 2004). Respondents to the questionnaire did not have a substantially different
average household size than the State of Michigan’s average (2.76% and 2.56% persons
respectively). The average age of respondents (49.8%) is slightly higher than the average
for Michigan residents (46.4%). Survey respondents had a lower median household
income ($42,500) than the States median household income ($46,986).

Survey respondents, on average, were slightly more educated than Michigan’s
population, 63.7% compared to 61.6% respectively for attainment of a high school
education or higher, and 30.7% compared to 21.8% respectively with a bachelor’s degree

or higher.

44

Table 5. Comparison between Survey Respondents and Census Results for Michigan

 

Survey Sample State of Michigan

 

Average Household size 2.76 2.563
Average age 49.8 46.44
Household Median income $42,500 ' $46,9865
Education

High school graduate or higher 63.7% 61.6%6

Bachelor’s degree or higher 30.7% 21.8%
Ethnicity

White 89.3% 80.7%7

African-American 4.7% l 3.0%

Hispanic, Latino or Spanish 1.2% 2.7%
Female 52.4% 51.8%8

 

Most survey respondents reported being ‘white’, 89.3%, more than the States’
population of 80.7%. Minorities were under represented in the sample. Fewer
respondents indicated they are ‘African-American’ (4.7% compared to 13.0% for

Michigan) or ‘Hispanic, Latino or Spanish’ (1.2% compared to 2.7% in the state). Other

 

3 US. Census Bureau, Census 2000. Table DP-l Profile of General Demographic Characteristics: 2000.
Geographic area: Michigan.
" Computed from: US. Census Bureau, Population Estimates Branch, "Estimated Population of States by
Age Group and Sex, 2000-2003" as released by Census Bureau on March 10, 2004; for population over 20
ears.
US. Department of Commerce, Economics and Statistics Administration. US. Census Bureau. Money
Income in the United States: 2000, page 12; for population 25 years and over.
6 Educational information obtained from US. Bureau of the Census, Census 2000. Table DP-2 Profile of
Selected Social Characteristics: 2000. Geographic area: Michigan; for population 25 years and over.
7 US. Bureau of the Census. Census 2000. Table PL4. Hispanic or Latino and Not Hispanic or Latino by
Race for the population 18 years and over [73] - Universe: Total population 18 years and over. Data Set:
Census 2000 Redistricting Data (Public Law 94-171) Summary File. http://factfinder.census.gov/.
8 US. Bureau of the Census. Census 2000. Table DP—1.Profile of General Demographic Characteristics:
2000. Data Set: Census 2000 Summary File 4 (SF 4). Summary File. Geographic Area: Michigan. For
population 18 years and over. http://factfinder.census.gov/

45

minority groups, not presented in the table above, include: ‘Hawaiian or other Pacific
Islander’ (0.6%), ‘Asian’ (0.8%), ‘American-Indian or Alaska native’ (0.6%) and ‘other’
(2.9%) (Lupi, et al., 2004).

The distribution of male to female respondents is very similar to the State
population (52.4% and 51.8% respectively). Overall, respondents to the survey
questionnaire measure up reasonably well with census data for Michigan residents (Lupi,

etaL,2004)

46

CHAPTER 4

STATED CHOICE MODEL

Stated Choice

Trade-offs between environmental services must always be made, as funds are not
unlimited. Programs requiring the financial inputs need to be protecting or restoring a
service that is valued by society in an amount equal to or greater than the cost of the
program to make it economically justified (Barbier, 1997). Discovering people’s
preferences for a service gives an indication to the relative value of the services.

It can be very difficult for policy makers and program planners to prioritize
environmental services when designing a wetland policies or programs. To address this
need for prioritizing services, a stated choice questionnaire was designed to elicit
people’s preferences over alternative coastal wetland programs. Stated choice methods
develop statistical relationships between people’s choices and the attributes of the
alternatives they choose among. This research sought to identify the particular
environmental services provided by Great Lakes Coastal wetlands which were top
priorities to Michigan residents, what sort of mix people prefer between preservation and
restoration activities, and the desirability of alternative mechanisms for acquiring land for
wetland preservation and restoration.

For the stated choice question, respondents were presented with a scenario, or
choice set. In order to obtain responses that are as accurate as possible, respondents must

believe that the program offered is realistic and that they could actually be voting for or

47

against a particular program. If the program is not realistic, responses to the questions
might not be realistic either. The scenario consisted of two program alternatives from
which respondents were asked to choose. The two program alternatives were presented
side by side for easy comparison and labeled program A and program B.

The choice question was a binary discrete choice question in which respondents
were asked to choose whether they preferred program A or program B, based on the
attribute levels given. Table 6, below, depicts the general scenario given to respondents.
Each program alternative contained three attributes, with attribute levels that varied from
program to program. In total, there were ninety-eight different programs, making up
forty-nine different scenarios, evenly distributed across the sample. The number of
programs needed was determined by the number of attributes in the model, and will be
discussed further in the section on the experimental design.

The stated choice method provides a means of turning choices, such as in table 6,
into quantitative estimates of peoples’ preferences. By varying the attributes across
people, their effect on peoples’ choices can be determined. The stated choice method is
based on random utility theory (Adamowicz, et al., 1998 (a)), discussed below. Random
utility theory is utilized to model individual’s preferences for Great Lakes coastal wetland

programs, including modeling preferences for the individual attributes of the programs.

48

Table 6. Example Survey Questionnaire Page Showing One Version of the Stated

Choice Question (Scenario).

WHICH PROGRAM WOULD YOU PREFER?

The two programs described below take different approaches to coastal wetland
protection. Please compare the two programs before answering Question 8.

 

 

 

 

 

PROGRAM A PROGRAM B
01' the many features coastal
wetland provide, which is the
program’s highest priority?
. Water Quality and
Primary Focus { Flood Control Non-Game Species

 

How the program effort and
resources are divided

Preservation of high quality
coastal wetlands

 

 

40% Preservation

60% Preservation

 

 

 

easement, Ol‘ ten-year contract

 

 

 

 

Pays for purchase, permanent {

 

 

Restoration of coastal , . ,
wetlands in poor condition { 60 /0 Restoration 40 /o Restoration
How the program gets
coastal wetland sites from
volunteers
Purchase Property Permanent Easements

 

1. If the state had to choose one of these two programs for protecting coastal

wetlands, which of these programs would you prefer?

(Circle one response)

a. Program A
b. Program B

49

 

In any survey, there are a limited number of options that can be offered to
respondents. In this survey questionnaire there were three attributes offered in each of
the two programs. The three attributes that made up each program were program
“priority”, program “mix” of effort divided between preservation and restoration of
coastal wetlands, and the program “tool” used to implement the program. Each attribute
was made up of levels: “priority” contained six levels, “mix” had seven levels, and “tool”

had three levels. Table 7 shows the different wetland priorities.

Table 7. Priority: The Program’s Primary Focus.

 

Priority:

 

Non-Game Species

Open Space Near Cities

Water Quality & Flood Control
Fish Habitat

Waterfowl Habitat
Biodiversity

 

“Priority” refers to the six different program priorities presented in the programs.
It is the priority offered to respondents as the program’s primary focus. Each individual
has two programs from to choose between, and each program contains one of the six
priorities presented above. The experimental design of the choice model pre-determines

which priorities are in a given scenario.

50

Table 8. Mix: Effort Devoted to Preservation and Restoration.

 

Mix:

 

10 % Preservation 90% Restoration
25 % Preservation 75% Restoration
40 % Preservation 60% Restoration -
50 % Preservation 50% Restoration
60 % Preservation 40% Restoration
75 % Preservation 25% Restoration

90 % Preservation 10% Restoration

 

Table 8 shows the preservation / restoration mix. “Mix” refers to the division of
effort between preservation and restoration activates offered in each program. The
programs vary between the seven levels presented. The “mix” shows how each

program’s resources would be split between the two activities.

Table 9. Tool: Land Acquisition Method.

 

Tool:

 

Purchase Property
Permanent Easements

Ten Year Contracts

 

Table 9 shows the three different tools that could be utilized for implementing the
program. “Tool” refers to the method utilized for acquiring acreage for coastal wetland
preservation or restoration. The “tool” is how the program gets coastal wetland sites

from volunteers.

51

Experimental Design of the Stated Choice Question

The experimental design of a stated choice question refers to the method for
combining the attribute levels into alternative programs, and how these are then
combined to create the choice scenarios to be presented in the questionnaire. In order to
estimate the response of the choice probability to the X’s (the attribute levels), the
attribute levels must vary within the survey. A full factorial design is one in which every
level of every attribute can be combined with every level of all other attributes. Program
“priority” contained six levels inputted into both program A and program B options.
Program “mix” contained seven levels, and “tool” was made up of 3 levels, which also
appear in both program options. To determine the number of combinations that would be
necessary for a full factorial design, the number of levels of each attribute are squared (to
account for them being included in both program A and B) and then multiplied by the
other squared attribute level counts. The full factorial design for this study would result
in 62x72x32, or 15,876 combinations. Such a design results in far too many options for a
sample (Louviere, et al., 2000).

Because this number of combinations is far too large to manage, a main effects
design was utilized. A main effects design ensured that the attributes, and their levels,
were independent of each other. A main effects plan generates variables that are linearly
independent, within and across alternatives. In a main effects design, only the first order
interactions among variables can be identified. The main effects design resulted in 49
pairs of alternatives requiring 49 questionnaire versions, or 49 different choice sets, far

fewer and easier to manage than the 15,876 needed for a factorial design.

52

THE MODEL

Next, a little math background is needed. In order to estimate a choice
experiment, some initial groundwork must be laid, to set out the theory and basic
assumptions behind a choice model. The section begins with random utility theory and
its assumptions, and then presents the statistical model used to estimate preference
parameters.

First, the stated choice model is based on the standard economic assumption that
individuals will try to maximize their utility, U (Adamowicz, et al., 1998 (a)). Put
differently, it is assumed that individuals will pick the alternative they like best, and each
level of each attribute contributes to the total utility level in its own way. That is, an

individual, i (i = 1,. . .,N) will choose alternative 0 over alternative b if:

Ua; > Ubi- . (1)

The subscripts a and b refer to program alternative a and program alternative b in
the choice experiment, and Um- is the utility of the 0th alternative for the 1‘“ individual
(Louviere, et al., 2000). Based on the attribute levels given in each program, one program
will provide the individual with a higher level of utility than the other. The model also
assumes that U...- has two components, a systematic or ‘representative utility’ component,
Va, which is deterministic, and an additive random component, ea, which is the stochastic
term, as follows:

Um- = Va + ea for alternative a, and
Ub, = Vb + 8;, for alternative b. (2)

53

By combining equations, it can now be seen that alternative 0 is chosen when

(Va + ea) > We + 8b) . (3)
which can also be written as

(Va+80)-(Vb+8b)>0. (4)
Va, the representative utility component is an expression in terms of the attributes of the
question, a vector of all attributes in a choice.

Since, the choice an individual makes is based on the vectors of attributes given

them (X), the usual specification is followed by writing V as a linear-in-parameters

firnction of the attributes (Louviere, etal., 2000) as follows:

Va = BXa and V], = Ma (5)

where B is the vector of the preference parameters associated with the attributes, X, that

make up alternatives a and b. Substituting the BX’s into equation (4) gives:

(BXa + 8a) - (BX!) + 8b) > 0- (6)

Then, collecting terms results in,

54

B(Xa - Xb) + 8,, - 8;, > 0. (7)

Again, the above equation would be true if, and only if, option a is chosen.
At this point, the probability that option a is chosen (over option b can be written

as

PI’Ob(Ua > Ub) = PI’Ob (BUG ' Xb) + 80 ' 3b > O) = P0: (8)

where Pa equals the probability that alternative a is preferred to b. A dummy variable is
used to indicate choice, equaling 1 when alternative a is chosen, and 0 when alternative b
is chosen (a binary choice).

To achieve the desired goal of estimating the B’s, an assumption about the errors
must be made. A common approach involves assuming that the errors, e's, are
independent and identically distributed from a type I extreme value distribution
(Louviere, et al., 2000) with a mean of zero and scale factor a, which yields the logit

form for the choice probabilities as follows:

P.=F(13X)=1/(1+e"‘”’“’), (9)

with X = X, - X1, in our model. This logit form is the basic choice model estimated, and
this is the most widely used form of the random utility model for stated choice data. In a
logit model, as a consequence of the errors assumptions (McFadden, 1973), there is an

assumption of ‘positivity,’ meaning that the probability that a given program alternative

55

is chosen must be greater than zero (Louviere, et al., 2000). When estimating B’s using
the logit form, it is not possible to separately identify the scale factor (due in part to the
discrete nature of the dependent variable). The estimated B’s from equation (9) are thus
estimates of B/a. For this reason, it is often assumed that (i=1. Consequently, one cannot
directly compare B’s across data sets because a is directly connected to the estimated B’s
in the data set. However, it would be appropriate to compare ratios of the B’s among

alternate data sets because the scale factor cancels out of such calculations.

ESTIMATION OF THE MODEL

The model is estimated by Maximum Likelihood estimation (MLE). Theory
behind MLE for random samples implies that MLE is consistent, asymptotically normal,
and asymptotically efficient, and MLE is the estimation method most commonly used for
logit models (Louviere, et al., 2000; Wooldridge, 2000).

The exact specification of the model in terms of the specific variables is given by:

V = BX = BlNon Game Species + B2 Water Quality and Flood Control +
B3Fish Habitat + B4 Waterfowl Habitat + BsBiodiversity + (10)
B6Purchase Property + B7Permanent Easements +
BgPercent Preservation + BgPercent Preservation" .

The specific variables in equation (10) are all defined in Table 10. The goal was to
estimate the B’s which indicate the effect that each of the variables has on utility, and
consequently, through equation (9), relate the probability that individuals prefer one

program to another given the attributes of the two programs. To estimate the model, the

56

levels for the “priority” and “tool” variables were converted to binary variables due to the

fact that they were not continuous (see Table 10).

Table 10. Variables Used in the Estimation of the Utility Function.

 

Variable

Variable Definition

 

Non Game Species

A dummy variable indicating whether Non
Game Species is the program’s top priority.

A dummy variable indicating whether Water

 

Water Quality and . . ,
Flood Control Quality and Flood Control rs the program stop
priority.
. . A dummy variable indicating whether Fish
“Priority” FISh Habitat Habitat is the program’s top priority.
. A dummy variable indicating whether
Waterfowl Habitat Waterfowl Habitat is the program’s top priority.
. . . A dummy variable indicating whether
Biodiversny Biodiversity is the program’s top priority.
Open Space Near The baseline for comparison of priority
Cities variables. Not reported in Table 10.
A dummy variable indicating whether or not
Purchase Property Purchase Property was the approach used to
acquire wetland acreage.
“T 1,, Permanent A dummy variable indicating whether or not
00 E Permanent Easements was the approach used to
asements

Ten Year Contracts

acquire wetland acreage.

The baseline for comparison of tool variables.
Not reported in Table 10.

 

66Mix,,

Percent
Preservation

Percent
. 2
Preservation

A continuous variable representing the
percentage of program effort devoted to
preservation.

A continuous variable representing the
percentage of program effort devoted to
preservation, squared, to show the non-linear
effects.

 

57

When running a logit regression with binary variables, one level of each of the
attributes must serve as the baseline for the remaining levels of that attribute. The
“priority” and “tool” attribute levels that are included in the regression must be
interpreted relative to the levels of the variable that served as the baseline. Open Space
near Cities was the level omitted from the regression for program priorities, and was
therefore the baseline for program priorities. Coefficients resulting from the logit
regression for the remaining program priorities are interpreted as the effect on the choice
probability, relative to the effect on the choice probability that Open Space near Cities
has. Ten Year Contracts was the variable omitted fi'om the regression for program tools,
and was likewise the baseline for program tools. All remaining attribute levels for
program priorities and program tools were included as variables, and their estimated
parameters must be interpreted as the effect they have on utility relative to the omitted
attribute of their group.

The attribute “mix” (see Table 10) was represented by the continuous variable,
Percent Preservation. Percent Preservation is based solely on the preservation level of
the mix attribute because the percent preservation and percent restoration levels for a
program always add to 100%. The variable Percent Preservation takes on seven levels,
ranging from 10% to 90% (see Table 8). Percent Preservationz was also included in the
logit regression to estimate any non-linear effects the preservation level might have had
on the choice probability. Estimating non-linear effects enables us to view any peaks in
preference for the level of preservation; it may not be the case that more preservation is

always preferred to less.

58

ESTIMATION RESULTS

Out of the 540 responses to the questionnaire, only 512 were valid for estimation.
Two questionnaires were lost, and 26 respondents did not answer the stated choice
question (question number 8), hence, the estimation of the lOgit choice model was based
on the 512 observations that had valid responses for the choice question. The estimation

results for the logit model are presented in Table 11. The estimated coefficients (the B’s)

associated with each of these variables (the attributes of X) tell about the effect the

variable has on the choice probability (the utility level). A positive B indicates that,

holding other variables constant, the variable has a positive effect on the choice

probability.

Table 11. Estimated Parameters from the Stated Choice Model.

 

 

Variable (X) Coe$;ient t-stat p-value 2132?:
Purchase Property 0.662 3.768 0.0002 0.09

Permanent Easement 0.342 2.084 0.0371 -0.063
Percent Preservation 0.044 3.941 0.0001 -0.078
Percent Preservation Squared -0.0003 -3. 152 0.0016 20.801
Non-Game Species 0.234 1.129 0.2590 0.043

Water Quality and Flood Control 0.938 3.713 0.0002 -0.002
Fish Habitat 0.299 1.191 0.2335 -0.004
Waterfowl Habitat 0.537 2.260 0.0238 -0.03 l
Biodiversity 0.873 3.561 0.0004 0.021

 

The means presented in Table 11 represent the differences between the attributes

of program A and B presented to respondents (as in equation 9). Each program is made

59

up of different characteristics, and the means show the average difference of the
individual attributes use in the choice set. For example, the mean of 0.09 for Purchase
Property indicates that it was presented as the primary focus for respondents for program
A slightly more often than for program B.

One way to assess the logit model performance is to examine the percent correctly
predicted, depicted in Tables 12 and 13. Table 12 compares the actual choice made
(selection of program A or program B) with the predicted choice using the estimated
parameters. One indicates program A has been selected (or predicted to be selected) and

0 indicates program B.

Table 12. Frequencies of Actual and Predicted Outcomes.

 

 

 

Predicted Total
0 l
0 182 77 259
Actual
92 161 253
Total 274 238 5 12

 

The percent correctly predicted classification, presented in table 13 below, refers
to the percentage of times in the actual data that the predicted choice corresponds with
what the respondent actually selected. For the estimated logit model, the percent
correctly predicted classification is 67% overall and its performance is well-balanced for
both Program A and Program B predictions. Moreover, with a Chi2 estimate of 64.55
with 8 degrees of freedom, the hypothesis that all parameters are 0 is rejected, indicating

the model is highly significant.

60

Table 13. Percent Predicted Correctly.

 

Prediction Success:

 

Sensitivity = actual ls correctly predicted 66.64%
Specificity = actual 0s correctly predicted 70.27%
Correct prediction = actual 1s and 0s correctly predicted - 66.99%

Analysis of Binary Choice Model Predictions Based on Threshold 0.50

 

An R2 is a measure of fit in a linear regression model and is based on comparing the fitted
values with the actual values (it is the percent variance explained by the regression).
However, in discrete choice models there are no residuals to compute due to the binary
nature of the dependent variable, and it is not possible to estimate variances (because they
are directly related to the unidentified scale factor). Hence, there is no measure that can
be interpreted as explaining a percentage of the variance. An alternative method for
measuring the goodness of fit in the binary discrete model is the pseudo R2, p, or

‘likelihood ratio index’ or the “McFadden R2”

P = 1 - 1403') / L(0)-

L(0) is the value of the log-likelihood fimction when all variables are equal to 0
and L(B') is the value when the log-likelihood function has been maximized (Pindyck
and Rubinfeld, 1998; Wooldridge, 2000). A pseudo R2 is only good for comparisons
within a data set. The pseudo R2 is not a comparable measure fi'om one data set to
another, because the pseudo R2 is a function of the size and variance of the particular data
set (L(0)) and alternative data sets have different sizes and variances (Wooldridge, 2000).

The estimated model above has a McFadden’s R2 of 0.091.

61

Interpretation of Estimated Parameters

For each estimated parameter, Table 11 also presents the t-values and p-values for
the hypothesis test that the parameter is equal to zero. Seven of the nine variables are
statistically significant at the 5% level (that is, they have p-values smaller than 0.05).
Only Non-Game Species and Fish Habitat are not statistically significant. However, the
low statistical significance of Fish Habitat and Non-Game Species does not reflect a lack
of effect on the probability of selecting A. It just means, that they do not have a
statistically different effect on the choice probability than does the baseline level for
program priority, Open Space Near Cities. The three priorities are not statistically
different fi'om each other in their effect on choice probabilities.

A positive coefficient has a positive effect on the representative utility function,
which affects the probability of program A being selected. Both Purchase Property and
Permanent Easement have positive coefficients, 0.662 and 0.342 respectively. The
positive coefficients imply that each variable has a positive effect on the choice
probability, compared to the baseline tool of Ten Year Contracts, holding all other
variables constant. Non-Game Species, Water Quality and Flood Control, Fish Habitat,
Waterfowl Habitat, and Biodiversity all have positive coefficients of 0.234, 0.93 8, 0.299,
0.537, and 0.873 respectively. Their positive coefficients imply that each variable has a
positive effect on the choice probability, compared to the baseline priority of Open Space
Near Cities, holding all other variables constant.

Percent of effort devoted to preservation, Percent Preservation, has a positive
coefficient of 0.044, implying a positive effect on the choice probability, holding all other

variables constant. Increasing the Percent Preservation has a positive effect on the

62

choice probability. People prefer preservation over restoration, and tend to vote for
programs that will provide for more preservation of Great Lakes coastal wetlands, but
only up to a point.

Percent Preservationz is a nonlinear form of the variable Percent Preservation. It
is used to capture the potentially different marginal effects of preservation effort. Percent
Preservation is graphed in Figure l, to show how the percent devoted to preservation

begins to diminish in its increasingly positive effect on estimated utility.

Figure 1. The Nonlinear Effect that Percent Preservation has on Estimated Utility

2
13X + 13X y = 0.044x- 0.0003 x2

 

 

 

 

0 25 50 75 100
% Preservation

Every level of preservation is preferred to 100% restoration, but the most
preferred level of preservation is 73%. A program with a 73% preservation mix would
be preferred over all other preservation mixes (all else equal), but beyond 73%, the effect

is declining (yet still positive). One hundred percent preservation and 47% preservation

63

were equally preferred by respondents. Below a 47% preservation level, 100%
preservation would be preferred; however, 100% preservation would not be preferred to
levels of preservation between 47% and 100%.

In interpreting the regression results presented in Table 11, the magnitude of the
coefficient is in direct proportion to the effect that a change in the variable will have on
the choice probability (the utility level). The larger the variable’s coefficient, the greater
the effect on the choice probability. However, only when the variables are significantly
different from each other should they be determined to have a significantly greater or
smaller effect on the choice probability. Variables that are not significantly different
from each other should not be regarded as having a significantly different effect on the
choice probability.

To start interpreting the variables and their relative significance, look first at the
“base line” variables. Because Open Space Near Cities and Ten Year Contracts are the
“base line” for their respective attributes, and because all their respective remaining
levels have positive coefficients, the two base lines are of the lowest ranking order;
however, variables that are not significantly different than them may also be of the lowest
ranking order.

All “tool" variables are significantly different from each other, and can be ranked
based on their effect on the choice probability. Purchase Property is most preferred as a
program tool, followed by Permanent Easements, then Ten Year Contracts. In order to
determine if the model parameters were significantly different from each other in their
effect on the choice probability, a hypothesis test asserting that the attribute levels

Purchase Property and Permanent Easements are equal to each other was conducted.

64

The result was p-value of 0.0415, so the hypothesis that they are equal was rejected at a
5% level of significance.

The order for “priorities” is complicated by the fact that several of the variables
are not significantly different from one another. To illustrate this, table 14 displays the p-
values for the pair wise tests of parameter equality for the “priority” variables. It depicts
which pairs of attribute levels are significantly different from each other in their effect on

the utility level, and at what level.

Table 14. Choice Model Parameters Tests of Equality: P-values for Priority Variables
and Levels of Significance.

 

 

E: 4..
N” ”B b N S. q, 8 a
t: o '3 I; 2 a g a, c a:
61% s s 2% E S °§ as
S E L3 S 3 ~= é Sf § §
§ ° 5 & L3 ‘ 2 8‘2
Water Quality and Flood "
Control
Biodiversity 0.8034 --
(NS)
Waterfowl 0.0910 0. 1703 --
(*) (NS)
Fish Habitat 0.0158 0.0211 0.3379 --
("I (") (NS)
Non-Game Species 0.0013 0.0027 0.1419 0.7648 --

(""l ("‘9 (NS) (NS)

Open Space Near Cities 0.0002 0.0004 0.0238 0.2335 0.2590 --
0'") ("'9 C") (NS) (NS)

NS = Not significant at 10% or lower, "‘ = significant at 10%, ** = significant at 5%,
and "‘ = significant at 1%.

 

NS stands for not significant at a 10% level or lower, which means the two

attribute levels are not significantly different from each other in their effect on the choice

65

probability. One asterisk indicates that the two attribute levels are significantly different
from each other at the 10% level; two asterisks indicate they are significantly different
from each other at the 5% level, and three asterisks indicate significance at the 1% level.
The lower the significance level percentage, the more accurate the measures, meaning the
more confident one has that attribute levels are significantly different from each other.

Based on the model parameters, the pair wise tests of equality and the
corresponding significance levels, the rank of preference for “priority” variables can be
determined. Water Quality and Flood Control and Biodiversity are the two top program
priorities. Water Quality and Flood Control and Biodiversity are significantly better than
Non-Game Species, Fish Habitat, and Open Space Near Cities as program priorities at
5% and 10% significance levels (shown in table 14 above). Water Quality and Flood
Control is also significantly better than Waterfowl at a 10% significance level.

Waterfowl, next in order of preference (third in ranking), is significantly more
preferred than Open Space Near Cities as a program priority, at a significance level of
5%. However, Waterfowl is not significantly different than Biodiversity, Fish Habitat,
and Non-Game Species in its effect on the choice probability.

Fish Habitat, Non-Game Species, and Open Space Near Cities, would be next in
order of preference, based solely on their attribute levels coefficient; however, they are
not significantly different from each other (as can be seen in table 14), so no exact order
can be made among the three. They all have the lowest effect on the choice probability.
The low significance levels of Fish and Non-Game Species imply they are not

statistically different from the priority Open Space Near Cities, it does not mean their

66

effect on the choice probability is not significant when compared to other possible
priorities.

As mentioned previously, the variance in the data is not identified. Because of
this, it is not possible to determine the standard deviations in this model. This makes the
B’s that result from the logit regression incomparable to [3’s in other studies. While the
B’s are incomparable, the marginal effects of the individual variables can be compared to
other studies marginal effects. The marginal effects, presented in Table 15, are the effect
each variable has on the probability of program A being selected (the probability that Y =
1) from a one unit change in the variable (such as the given variable being utilized by a
program). If a given variable is included in program A, the presented marginal effect is
how that variable, all else equal, affects the probability of program A being selected

evaluated at the mean of the X’s for all other variables.

Table 15. Marginal Effects on the Probability of A.

 

 

Variable (X) Marginal effects

“Tool” Purchase Property 0. 16534
Permanent Easement 0.08535

“ . ” Percent Preservation 0.01098

MIX Percent Preservation Squared -0.00008

Non-Game Species 0.05853
Water Quality and Flood Control 0.23434

“Priority” Fish Habitat 0.07469
Waterfowl Habitat 0. l 3421
Biodiversity 0.2 1 824

 

67

The marginal effect of 0.05853 (see table 15) for Non Game Species signifies that,
all else equal, the inclusion (or removal) of Non Game Species in a program (such as in
program A) increases (or decreases) the probability that program A will be selected by
5.9%. The overall probability changes in probability from one attribute level being
removed from a program and replaced with an alternative attribute level depends on
which variable is being inputted and which is being replaced. For example, if program A
contained Water Quality and Flood Control as its program priority and it was replaced
with Fish Habitat as the priority, based on the marginal effects of 0.23434 and 0.07469,
respectively, the probability of program A being selected, all else equal, would increase
to about 16% (0.23434-0.07469).

It is worth while to point out that even though the marginal effect for Percent
Preservation and Percent Preservation Squared seem small comparatively, the “Mix”
variable is a continuous variable, unlike the binary variables of the remaining attributes.
This means that there are many units of change possible for the “Mix” variable. For
example, a 20% increase in the relative level of preservation can have a large effect on
the choice probability. For example, using the average difference for all variables,
increasing program A’s preservation level by 20 percentage points over that of program
B increases the probability of program A being select by 0.18. The effect of this 20 point
increase in preservation would be similar to the effect Biodiversity has on the choice

probability).

68

CHOICE EXAMPLES

The estimated coefficients presented above can be used to get an indication of the
relative importance of each attribute level, as can also be seen by the marginal effects of
each variable. By inputting the relevant coefficients and attribute levels into the
probability model, a comparison of two separate programs becomes possible. The model
indicates the probability of one program being selected over another, given the attributes
of each.

The predicted probability of selecting program A is determined by the random
utility probability model (equation 9), Pa = F(BX) = 1 / (l + em”), with X = X, - Xb in our
model. By simply inputting the values for the attribute levels (in their form as the binary
variables for “priority” and “tool” from Table 10) and their corresponding coefficient
from the logit regression, the probability of selecting program A results (Louviere, et al.,
2000). Changes in probability can also be indicated by the variable being omitted and the
variable being inputted marginal effects. To more completely show how the probability
of selecting a program changes with the various attribute levels, several example choice

sets are presented, along with predicted probability of selecting a program.

Table 16. Predicted Choice Example 1.

 

 

Program A Program B
Priority Water Quality and Flood Control Open Space Near Cities
Mix 73 % Preservation 0% Preservation
2 7% Restoration 100% Restoration
Tool Purchase Property Ten Year Contracts

 

69

Table 16 presents two possible programs alternatives and their attributes. Based
on the attribute levels presented, and their corresponding estimated coefficients from
Table 11, it is apparent that each attribute level presented in program A are the most
preferred attributes levels, while program B is comprised of the least preferred attribute
levels. Here, program A will be the most preferred program because each of its attributes
are more preferred to those presented in program B. As previously mentioned, the
probability of selecting program A is determined by inputting the attribute levels and
their corresponding coefficients, as presented in Table 11, into the random utility
probability model. Based on the attribute levels and their corresponding coefficients, the
probability of selecting program A is about 0.96. This indicates that about 96% of
respondents would prefer program A to program B. The high likelihood of program A
being selected is the result of the most preferred program attribute levels being compared
with the least preferred program attribute levels.

Table 17 presents two alternative program options. As with the previous
example, the probability of program A being selected by a respondent can be determined
by utilizing the probability model. This example differs from the previous in that the
attribute levels presented in program A are not all clearly preferred to those attributes
presented in program B. in this way, the model can be usefirl in predicting which
programs are most preferred, based on the attribute levels within a program as a whole,

and not individually.

70

Table 17. Predicted Choice Example 2.

 

 

. Program A Program B
Priority Non-Game Species Water Quality and Flood Control
. 25 % Preservation 10% Preservation
MIX 75 % Restoration . 90% Restoration
Tool Purchase Property Permanent Easement

 

By evaluating the program attributes in Table 17 using the probability model, the
probability of a respondent selecting program A is about 0.53, or about a 53% chance
program A is preferred. Here, neither program is clearly preferred, and neither program
is comprised entirely of attribute levels clearly preferred to the other.

The example presented in Table 17 can also be used to demonstrate how the
choice probability changes when a particular attribute changes. If program B had Ten
Year Contracts, instead of Permanent Easements, for its “tool” in Table 17, the
probability of choosing program A, holding all other variables constant, would be about
0.61. The example demonstrates how use of Permanent Easements, as opposed to Ten
Year Contracts, increases the probability that program A is preferred by about 8%, all
else equal. This change in probability can also be seen in the marginal effect of
Permanent Easements of 0.08535, or an effect of about 8.5% (see table 15). Ten Year
Contracts is the baseline attribute level for land acquisition method, hence, does not have
a marginal effect that would need to be taken into account in determining probability
changes. By changing only one variable, from Ten Year Contracts to Permanent
Easements, the resulting change in probability is about 8.5%.

If, instead of the attributes presented in Table 17, program A had Ten Year

Contracts (which has no marginal effect as it is the baseline tool) rather than Purchase

71

Property (with a 0.16534 marginal effect) for its “tool” (while program B retained
Permanent Easements for its tool), the predicted probability that program A would be
preferred would drop to about 37%, holding all other variables constant. So, while
Permanent Easements are around 8% more likely to be chosen as a program “tool” than
Ten Year Contracts, and Purchase Property is about 8% more likely to be chosen as a
program “tool” than Permanent Easements (see the marginal effect reported in Table 15),
the net effect of a change from Ten Year Contracts to Purchase Property is a about 0.16,
or 16% (compared to the marginal effect for Purchase Property of0. 16534), all else

equal

Table 18. Predicted Choice Example 3.

 

 

Program A Program B
Priority Open Space Near Cities Open Space Near Cities
. 25 % Preservation 4 0% Preservation
MIX 75 % Restoration 60% Restoration
Tool Permanent Easement Ten Year Contracts

 

Table 18 presents another set of possible programs. The probability of selecting
program A for the set of choices presented is about 0.5, or about 50% chance that
program A is preferred to program B. Both programs provide respondents with the same
level of utility; hence, respondents would tend to be indifferent between the two. When
looking at the individual attributes, both programs contain the same program “priority”,
while program B has a preferred level as its “mix” and program A has a preferred level
for its “tool”. Based on the individual attributes presented, there is no clear indication of

which program would be preferred. The utilization of the probability model allows us to

72

see which would be preferred, and in this created scenario the two programs are equally
preferred even though the individual attributes are not. Preferences are determined by the
entire mix of attributes within the programs, not on the individual attribute comparisons.
Using Table 18, and the attribute levels it presents, as a starting point, the effect
that changing the program “priority” has on the choice probability can be illustrated.
First, if program A focused on Water Quality and Flood Control instead of Open Space
Near Cities, the probability that program A is preferred would increase to about 0.71
(about a 21% increase in the probability of program A being selected). This corresponds
to the marginal effects of 0.23434 for Water Quality and Flood Control, and because
“Open Space Near Cities ” is the baseline priority and has no marginal effect in this
model, the full effect of Water Quality and Flood Control ’s marginal effect would be
seen by its inclusion. A second option would be changing the “priority” of program A to
Biodiversity (with a marginal effect of 0.2 l 824) instead of Open Space Near Cities (with
no marginal effect), resulting in about a 20% increase in the probability of program A
being selected, from a 0.5 probability, to a 0.7 probability. if, on the other hand, program
A utilized Waterfowl Habitat (with a marginal effect of0. 13421) as the program
“priority” as opposed to Open Space Near Cities, then the probability of program A being
chosen would increase to about 0.62 (about a 12% increase in the probability of program
A being selected). If, for another comparison, program A utilizes Fish Habitat instead of
Open Space Near Cities, the probability of program being selected increases to only 0.56,
or about a 6% increase in the chance that program A is preferred. Finally, if program A

focused on Non Game Species instead of Open Space Near Cities, the probability of

73

program A being preferred would increase to about 0.55, indicating about a 5% increase
in the chance that program A is preferred.

The few examples presented are simply illustrating the point that changing one
attribute effects the probability of program A being selected, based on the estimated
coefficients and in a non-linear fashion. For example, when a program “priority” that is
preferred to the priorities currently offered in the choice set is substituted into program A,
the probability that program A will be selected will increase based on the variables
parameter estimates, all else equal. This is simply another way of illustrating the
information presented above in Table 15 on the marginal effects of each of the model
variables.

By utilizing the model, program attributes can be modified in order to create
programs more in line with public preferences. These choice examples illustrate how the

model results can be applied to evaluate potential support for programs.

74

CHAPTER 5

SUMMARYAND CONCLUSIONS

The reported research asked Michigan residents about their experience with Great
Lake coastal wetlands and asked them to make trade-offs between protection programs
with different protection foci, levels of preservation and restoration effort, and land
acquisition methods. The results reveal, among other things, public preference for
protection program characteristics. Since estimation of preferences for Great Lakes
coastal wetland programs has not previously been reported in the literature, this research
provides a fist glance into peoples’ preferences for Great Lakes coastal wetland programs

Based on the estimated parameters for the model, it is possible to rank
respondents preferences for particular Great Lakes coastal wetland program attributes.
Results indicate that respondents prefer the purchase of property to the use of permanent
easements, and both the purchase of property and permanent easements are preferred to
land contracts. The exact reason for this preference was not elicited, but two possibilities
are noted here. First, in stating a preference over wetland programs, there was no cost to
the respondent for the programs. If purchase of property is viewed as a less risky
alternative compared to the use of easements and these are viewed as less risky than the
use of lO-year land contracts, then the preference ordering might be revealing a
preference for the more secure land acquisition method given that there is no difference
in cost to the respondent. Second, respondents were not equally familiar with the three
land acquisition methods. Information gained through the questionnaire and reported in

Lupi et al. (2004) indicated that respondents were more familiar with the purchase of

75

property than permanent easements and least familiar with land contracts as land
acquisition methods. Specifically, 33% of respondents were familiar with purchase of
property, 25% were familiar with permanent easements, and 16% were familiar with 10-
year contracts. The degree to which respondents are familiar with the land acquisition
methods may be a contributing factor to the order of preference. Future research efforts
might seek to explain some of the reasons for the preference ordering for the land
acquisition approaches that was revealed by this research.

The findings also demonstrate a trend in preferences for the method of
conservation. People tended to prefer programs which devoted more effort towards the
preservation of high quality coastal wetlands, as opposed to the restoration of low quality
wetlands. However, this effect was highly non-linear with the most preferred level being
a mix of 73% preservation and 27% restoration. In other words, people preferred
programs that were not focused solely on either of these two approaches but favored
more preservation in the mix. Survey results on people’s familiarity with these two
approaches were somewhat similar with 49% of respondents indicating they were
familiar with preservation and 42% indicating familiarity with restoration (Lupi et al,
2004). Thus, the familiarity with the approaches is unlikely to explain the preference. It
may be that some of the preference for preservation over restoration is due to people’s
uncertainty about restoration activities, such as what the end result will actually be. If the
focus of Great Lakes wetland programs is to be on restoration, educating the public about
such activities could increase the acceptance and support for restoring wetlands.

The research also allowed wetland programs to focus on the provision of

alternative environmental services. Programs were given a specified environmental

76

service which served as the program’s primary focus. The findings reveal that Water
Quality and Flood Control and Biodiversity are the two program priorities most preferred
by respondents. Waterfowl Habitat is third in order of preference, being statistically
significantly less preferred than Water Quality and Flood Control. Because they are not
significantly different from each other (see table 14), the remaining variables are
considered to be of similar order; that is, no statistically significant ordering can be made
between Fish Habitat, Non-Game Species, and Open Space Near Cities.

The review of current wetland programs in the Great Lakes found that there is a
large supply of funding and effort devoted to waterfowl habitat. While the findings
indicate that waterfowl habitat was a highly ranked program priority, people actually
preferred programs that prioritize water quality and flood control, as well as biodiversity,
to waterfowl habitat. NGO’s such as Ducks Unlimited that focus on waterfowl habitat
have been very successful in protecting wetlands and serve as an example of private
provision of a public good. The findings here might suggest that there is potential for
NGO’s to build on the example of Ducks Unlimited and develop programming stressing
I the provision of features such as biodiversity or water quality and flood control. That
said, it may be that it is easier to organize and firndraise for the provision of waterfowl
habitat because users and beneficiaries are easier to identify and may have larger values
for the feature than does the general public, whereas values for other features might be
more dispersed and more susceptible to free rider effects. Alternatively, the findings may
indicate that the public sees features such as water quality and flood control or

biodiversity as being relatively scarce as program priorities, when compared to the

77

exiting suite of wetland programs, and therefore would prefer programs that focus on
these features.

One question that has arisen during the course of the research has to do with
whether or not respondents favored water quality and flood control as a program priority
because it seemed to be providing more services than priorities composed of a single
feature (e.g., fish habitat). No indication of this was revealed in the qualitative pretest
interviews but the issue is unresolved. Future work might consider breaking down the
program priority categories even firrther by separating water quality and flood control
into two separate priorities.

Another area for future research involves identifying segments of the population
with different preferences. The present research provides estimates of the average
preference revealed by respondents’ choices. With the use of the multiple questions
presented in the survey, as well as the demographic information, it would be possible to
determine if any particular group of individuals (such as those who never visited a
wetland, live in the city, or are of a particular political persuasion) have identifiable
differences in preferences. Such efforts can aid decision makers interested in targeting
programs to gain support of particular groups and may also identify groups where
communication and education effort might best be focused.

The model and findings reported here can be used to estimate what residents
“really want” in protection programs. Due to limited resources available for wetland
programs, it is important to know what coastal wetland services are most prized by the
public so program funding can be focused on the areas the public values most. Policy

makers can utilize this model to evaluate how current Great Lakes coastal wetland

78

programs or regulations measure up to resident’s desires, as well as to develop programs
and regulations that are aligned with resident’s preferences. Decision-making concerning
the Great Lakes coastal ecosystems can benefit from incorporating public preferences for
coastal wetlands and wetland services. The reported research provides a first step
towards a better understanding of what the public values about Great Lakes coastal

wetlands and those programs designed to protect and restore them.

79

APPENDIX

This appendix contains:
1. Pre-notice Letter —- used in the first mailing
2. First Cover Letter - used in the second mailing, first questionnaire mailing
3. First Cover Letter — With Incentive — used in second mailing, first questionnaire
4. Reminder Post Card — used in the third mailing
5. Second Cover Letter — used in the fourth mailing, second questionnaire mailing
6. Final Cover Letter -— used in fifth mailing, third and final questionnaire mailing
7. Survey Questionnaire

All documents except the postcard were printed on 8.5 by l 1 inch paper and have been
scaled down to meet thesis formatting requirements.

80

Date

Name Surname
Address
City, State, Zip

Dear Name Surname:

You have been selected to participate in a study of Michigan’s Great Lakes Coastal
Wetlands. The study is part of research at Michigan State University. The project will
provide needed information to local, regional, and state agencies about residents’
opinions and concerns about Michigan’s wetlands along the Great Lakes.

All that we ask is that you complete a brief survey booklet that you will receive in the
° ' ' ' ' ' ’ mail in about a week. We are writing to you now since many people like to receive
advance notice of the survey booklet.

Thank you very much.

Sincerely yours,
courses OF
AGRICULTURE
AND NATURAL
RESOURCES . .
Dr. Michael D. Kaplowrtz

00de Principal Investigator

Helium State University
323 Natural We»
East Lansing, “than

48824-1222
517353-1919
Fax: 517/353-8994
kaplowitQmsucdu

Malia-um
WWW.

81

COLLEGEOF

AGRICULTURE
AND NATURAL
RESOURCES

Rom

W

“chum State Un’vcnity
323 Natrnl Marcos
East Lam W
48824-1222
517353-1919

Fax: 517/353-8994
Wadi:

uwan
WWW.

Name Surname
Address
City, State Zip

Dear Name Surname:

You have been selected to participate in a study about Michigan’s Great Lakes Coastal
Wetlands. You may recall receiving a letter about this study about a week ago. The
study is part of an effort by Michigan State University to learn about citizens’ opinions
and concerns regarding the wetlands of Michigan’s Great Lakes.

Your input is important because managing Michigan’s wetlands, including decisions
about restoration or preservation, involves trade-offs that affect you. Results of the
questionnaire will provide needed guidance to local, regional, and state agencies about
residents’ opinions and concerns about Michigan’s wetlands along the Great Lakes.

You have been selected as part of a scientific sample of Michigan residents. That is why
the survey asks a few questions about you and your household--so we can make sure
that we get a scientific cross-section of Michigan residents. Your participation is vital to
make sure that the information collected represents everyone.

By completing and returning this survey, you indicate your voluntary consent to
participate in this study and have your answers included in the project data set. The
answers are anonymous, and we will keep your individual views entirely confidential.
Rest assured, your privacy will be protected to the maximum extent allowable by law.

If you have any questions or comments about this study feel flee to call me by phone at:
(517) 353-1919, fax at: (517) 353-8994, or e-mail at: kaplowit@msu.edu. If you have
questions concerning your rights as a survey participant, please contact Dr. Peter
Vasilenko, Chair of the MSU Committee on Research Involving Human Subjects, by
phone at (517) 355-2180, fax at: (517) 432-4503, or e-mail at: ucrihs@msu.edu.
Thanks for participating in this study.

Sincerely,

Dr. Michael Kaplowitz
Principal Investigator

82

 

COLLEGE OF
AGRKZULTURE
AND NATURAL

RESOURCES

Mira

mm

“diam Stats My
323 Natural Rome»
East Lamina, Helium
48824-1222
517/353-1919

Fax: 51713539994
kWitQmsuedu

awn-um
WWW

Name Surname
Address
City, State Zip

Dear Name Surname:

You have been selected to participate in a study about Michigan’s Great Lakes Coastal
Wetlands. You may recall receiving a letter about this study about a week ago. The
study is part of an effort by Michigan State University to learn about citizens’ opinions
and concerns regarding the wetlands of Michigan’s Great Lakes.

Your input is important because managing Michigan’s wetlands, including decisions
about restoration or preservation, involves trade-offs that affect you. Results of the
questionnaire will provide needed guidance to local, regional, and state agencies about
residents’ opinions and concerns about Michigan’s wetlands along the Great Lakes.

You have been selected as part of a scientific sample of Michigan residents. That is why
the survey asks a few questions about you and your household--so we can make sure
that we get a scientific cross-section of Michigan residents. Your participation is vital to
make sure that the information collected represents everyone.

We realize that it takes time out of your day to fill out this survey and have enclosed
three first class stamps as a way of saying thank you for your help.

By completing and returning this survey, you indicate your voluntary consent to
participate in this study and have your answers included in the project data set. The
answers are anonymous, and we will keep your individual views entirely confidential.
Rest assured, your privacy will be protected to the maximum extent allowable by law.

If you have any questions or comments about this study feel free to call me by phone at:
(517) 353-1919, fax at: (517) 353-8994, or e-mail at: kaplowit@msu.edu. If you have
questions concerning your rights as a survey participant, please contact Dr. Peter
Vasilenko, Chair of the MSU Committee on Research Involving Human Subjects, by
phone at (517) 355-2180, fax at: (517)432-4503, or e-mail at: ucrihs@msu.edu.
Thanks for participating in this study.

Sincerely,

Dr. Michael Kaplowitz
Principal Investigator

83

 

 

MICHIGAN STATE

UNIVERSITY

COASTAL WETLANDS STUDY
clo Dr. Michael D. Kaplowitz
Michigan State University

323 Natural Resources Building
East-Lansing, MI 48824-1222

Name Surname

Address
City, State Zip

 

 

 

Dear Sir or Madam:

We recently sent you a booklet and request to participate in a study of
Michigan’s Great Lakes Coastal Wetlands. If you returned the booklet,
thank you very much. If you have not yet completed the booklet, please
take some time to do so now. Your input is important to make sure that
policy decisions reflect the views of Michigan citizens. Thank you very
much.

Sincerely,

Dr. Michael D. Kaplowitz

Principal Investigator MICHIGAN STATE
Michigan State University U I R Y
(517)343-1919

 

84

 

 

COLLEGE OF

AGRICULTURE
AND NATURAL
RESOURCES

Resales

Development

W State University
323 Natural Resources
East Lans'ng, man
40524-1222

5170531919
. Fax: 517853-8994

kaplowit@msu.edu

“Sills-ram.

Name Surname
Address
City, State Zip

Dear Name Surname:

About two weeks ago, we sent you a short survey about Michigan’s Great Lakes Coastal
Wetlands. While we’ve heard from some people, to the best of our knowledge, we have
not yet heard from you. We are enclosing another copy of the survey booklet for you
convenience.

Please mail the completed booklet back to us In the enclosed pre-pald envelope.

m input is important because managing Michigan’s wetlands, including decisions
about restoration or preservation, involves trade-offs that affect you. Results of the
questionnaire will provide needed guidance to local, regional, and state agencies about
residents’ opinions and concerns about Michigan’s wetlands along the Great Lakes.

You have been selected as part of a scientific sample of Michigan residents. That is why
the survey asks a few questions about you and your household--so we can make sure
that we get a scientific cross-section of Michigan residents. Your participation is vital to
make sure that the information collected represents everyone.

By completing and returning this survey, you indicate your voluntary consent to
participate in this study and have your answers included in the project data set. The
answers are anonymous, and we will keep your individual views entirely confidential.
Rest assured, your privacy will be protected to the maximum extent allowable by law.

If you have any questions or comments about this study feel free to call me by phone at:
(517) 353-1919, fax at: (517) 353-8994, or e-mail at: kaplowit@msu.edu. If you have
questions concerning your rights as a survey participant, please contact Dr. Peter
Vasilenko, Chair of the MSU Committee on Research Involving Human Subjects, by
phone at (517) 355-2180, fax at: (517) 432-4503, or e-mail at: ucrihs@msu.edu.
Thanks for participating in this study.

Sincerely,

Dr. Michael Kaplowitz
Principal Investigator

85

COLLEGE OF

AGRICULTURE
AND NATURAL

RESOURCES

Reeouee

Development

lichigan State University
323 Nattral Resources
East Lane'ng, Mchigan
48824-1222

517/353-1919
Fax: 517353-8994

kaplowit@msu.edu

MSUismaIbmatM-aoom.

Name Surname
Address
City, State Zip

Dear Name Surname:

During the last two months we have sent you several mailings about Michigan’s
Great Lakes coastal wetlands. Our study is drawing to a close, but we are making

- one final attempt to obtain your input.

Your completed survey will provide needed guidance to local, regional, and state
agencies concerning wetlands along Michigan’s Great Lakes. You have been
selected as part of a scientific sample of Michigan residents, and YOUR
PARTICIPATION IS ESSENTIAL to make sure that the information collected represents
everyone.

Please mail the completed booklet back to us In the enclosed pre-pald envelope.

By completing and returning this survey, you indicate your voluntary consent to
participate in this study and have your answers included in the project data set. The
answers are anonymous, and we will keep your individual views entirely
confidential. Rest assured, your privacy will be protected to the maximum extent
allowable by law.

If you have any questions or comments about this study feel free to contact me by
phone at: (517) 353-1919, fax at: (517) 353-8994, or e-mail at: kaplowit@msu.edu.
If you have questions concerning your rights as a survey participant, please contact
Dr. Peter Vasilenko, Chair of the MSU Committee on Research Involving Human
Subjects, by phone at (517) 355-2180, fax at: (517) 432-4503, or e-mail at:
ucrihs@msu.edu.

Thanks for participating in this study.

Sincerely,

Dr. Michael Kaplowitz
Principal Investigator

86

 

 

MICHJ93;‘\N STAKE,

l\ I’\ this I ’l r

 

Alichigan Great Lakes Wetlands Trust:
Citizens’ Survey Questionnaire

 

This is an opportunity to provide your opinion and feedback on natural resource
policy and management in Michigan. Your input will help decision makers make
more informed choices on how to restore and preserve Michigan‘s coastal wetlands.

This booklet begins by explaining about Great Lakes coastal wetlands and what they
do. It then focuses on how programs can preserve high quality coastal wetlands and
restore poorly fmrctioning coastal wetlands. You are then asked about a proposed
wetland protection program.

Your answers will help ensure that state level decisions reflect the views of the
citizens of Michigan Thank you for your participation. This booklet‘s several
sections of brief questions should take about 15 minutes to complete.

Please return your completed questionnaire in the enclosed envelope to:
CL Wetland Study, Dept. of Resource Development, 323 Natural Resources
Building, Michigan State University, East Lansing, MI 48824-1222

 

87

 

 

 

THE GREAT LAKES AND THE GREAT LAKES
WATERSng
TYPICAL MICI-HGAN COASTAL WETLANDS

 

 

 

 

 

88

 

 

 

\\
\\

 

 

Some of Michigan’s important natural resources are Great Lakes coastal wetlands.

ll
K‘
\

GREAT LAKES COASTAL WETLANDS

These wetlands support flre health and (fiversity of the Great Lakes ecosystem

Coastal wetlands provide ecological functions and critical habitat for a wide
range of plants, fish, and animals. They also provide storm water retmdon,
erosion prevention, water filtration, and other services.

In Michigan, Great Lakes coastal wetlands range from shoreline wetlands and
marshes along the northern coastline, to extensive wetlands along Saginaw Bay,
and to the delta marshes of the St. Clair River.

Healthy Great Lakes coastal wetlands also support a range of recreational
activities inchrding fishing, hunting, bird watching, and hiking.

1. Have you ever visited a Great Lakes coastal wetland? (Circle one response)
a. Yes
I). No (Skip to Question 3)
c. I don’t know

2. If you have visited a Great Lakes coastal wetland during the last year,
what did you do there? (Circle all that apply)

Fish or hunt

Watch for birds or other wildlife

Enjoy the outdoors

Other (If selected, fill in

blank)

999‘?

3. How important do you think Great Lakes coastal wetlands are for the
following activities? (Mark E one response for each activity)

 

 

 

 

 

2‘...» g u -e-I a .u
l Importance of Great Lakes g g g E g ,3 g :81 g
; wetlands for... g ‘6 ’6’ Q g

I E on El- Z E Z S
a. Hunting? C] U Cl C] CI
b. Fishing? [3 CI El E] El
c. Bird watching? Cl Cl Cl Cl C]
d Hiking? Cl Cl Cl Cl C]

 

 

 

 

 

 

 

 

 

 

~.._ ._. .. _.._,. -

 

89

 

 

 

WHAT COASTAL WETLANDS D

 

——~.—— Water Quality & Flood Control
. W , Wetlands absorb nutrients and chemicals as well as filter sediments.

 

 

 

They control flood damage by intercepting and storing storm water.

Open Space near Shore
Wetlands provide areas of open, undeveloped natural areas. They may
provide opportunities for public enjoyment and education.

Waterfowl Habitat
Wetlands provide food and breeding habitat that benefits waterfowl.
Wetlands support waterfowl as wood ducks, mallards, and geese.

Fish Habitat
Coastal wetlands provide shallow water areas with vegetative cover
that serve as spawning younds, nursery areas, and adult fish habitat.

Non Game Species
Wetlands provide habitat for animals not fished or hunted Such
species include shorebirds, amphibians, and wading birds.

Biodiversity
Wetlands’ complex ecosystems support diverse and unique plants and
animals. Such biodiversity is important for ecosystem health.

 

 

In your opinion, how important are Great Lakes coastal wetlands for
helping to provide and maintain... (Mark one response for each function)

 

 

 

 

 

 

 

Importance of Great Lakes wetlands g; g g E E 2 1g ’é’ g
atprovrdrng... E E. g g. ‘25 a. g E. Q j
a. Water quality 8:. Flood control? E] Cl E] El E]
b. Open space? [3 E] C] E] C]
c. Waterfowl habitat? [I E] D C] D
(i Fish habitat? [:1 Cl C! E] E]
e. Non-game species? E] Cl Cl E] El
f. Biodiversity? E] l] E] El El

 

 

 

 

 

 

 

 

 

 

90

 

 

 

COASTAL WETLANDS STATUS

Wetlands historically stretched flow the western edge of Lake Erie across Ohio int
Indiana, and covered the southern edge of Ontario. However, more than two-thirds
of the Great Lakes coastal wetlands have been lost.

Three ways wetlands are typically damaged and destroyed:

Loss of Wetland Acres

Acres of coastal wetlands are destroyed for land development purposes
including agricultm'e, harbor facilities, shoreline development, resource
extraction (such as peat mining), and urban expansion

Interference with Water Flows

Coastal wetlands are impaired and lost when activities disturb water flows
essential to support healthy wetland plants and animals. Adequate water
flows are essential for coastal wetlands.

Loss of Plant Diversity

Loss of wetland vegetation increases the danger of wetland loss. Wetlands
dominated by a single type of plant, including invasive species, typically
lead to decreased animal diversity because food and habitat are unavailable.

 

5. In your opinion, how serious a threat to Great Lakes coastal wetlands are
the following activities... (Mark [3 one response for each activity)

 

 

 

 

 

>~ a m
Seriousness of adverse impact on g 5 E § E 3 3 § ’§ 3
Great Lakes coastal wetlands of... a '5 a '5 ‘5 '5 3 '5 g :2

m o m z m z m

Ill-l U)
21 Agriculture? Cl E] El D D
b. New harbor facilities? Cl C] E] Cl C!
c. Shoreline development? C] D Cl Cl C]
d. Urban expansion? Cl E] E] [:1 Cl

 

 

 

 

 

 

 

 

 

 

91

 

 

 

  
 

  
 

   

 

   

State and federal laws try to prevent loss of wetland acreage by placing
conditions on wetland permits to avoid. minimize. and mitigate wetland loss.
However. existing laws do not always protect wetland quality.

 

    
 

Wetland protection programs can preserve high quality wetlands already in
the Great Lakes ecosystem. Wetland protection prog'ams can also restore
damaged or impaired wetlands to improve Great Lakes ecosystem health.

0 Preservation of High Quality Wetlands
Wetland programs can protect wetlands that are high quality. ecologically
rich. and hydrologically sound. Wetland preservation is accomplished by

agreements with landowners that legally and physically protect wetlands

and their surrounding uplands from adverse changes.

 
   
     
   
      
  

Restoration of Wetlands and Wetland Functions
Wetland programs can improve wetland ecosystems and wetland functions '
by restoring wetlands that have been degraded to a pre-existing condition
The restored wetlands would then be preserved. Coastal wetlands may be
restored by improving water flows. planting native species. and removing
invasive species.

 
   
     
       

6. How familiar are you with the following approaches used in wetland
protection programs? (Mark [2'] one response for each activity)

 

 

 

. 5‘33 E5 E2: 2:; E3
Familiaritywnh... :> In gm :3 333 Q g
a Wetland preservation? C]

b. Restoration of wetlands? Cl

 

 

 

 

 

 

 

 

 

 

92

 

 

 

GETTING WETLANDS TO PRESERVE AND RESTORE

When a wetland program is voluntary, the wetland program needs to negotiate
details of wetland access, restoration efforts, and protection plans with interested
property owners who receive compensation.

Written agreements restrict and prevent incompatible land uses and activities
around wetlands being preserved and restored. Here are three ways for
voluntarily programs to get wetlands for preservation and restoration:

0 Purchase Property from Willing Owners
Wetland programs may buy Great Lakes coastal properties that contain wetlan- ‘-
and that may be well suited for restoration or preservation Purchases are made
from willing property owners.

Voiun tary Permanent Easements

Programs may pay coastal property owners to place permanent restrictions,
called easements, on their wetland property. Permanent wetland easements are
voluntary and cost less than the purchase price of the land Property owners
retain their land but are required to permanently protect the wetlands and allow
any negotiated restoration eflofls to occur.

Short-term Contracts (10 Years) 1
Programs may contract with owners to protect wetlands on their property.
Contracts cost less than easements and land purchases. When the contract perio .
is over, the owner has no obligation to continue protecting wetlands.

 

7. In your opinion, how familiar are you with the idea of programs or agencies
purchasing property? (Mark IZI one response)

 

F:
;>

Don’t
Know

Familiarity with protecting natrn'al
resources through

at
Somewhat

familiar

Not Very

familiar

Not at all

familiar

F

 

C]

a. Purchasing property fi°om volunteers?

 

b. Voluntary permanent easement?

 

c. Short-term contracts (lfi-years)?

 

 

 

 

 

 

 

 

 

 

93

 

 

 

ALTERNATIVE COASTAL WETLAND PROGRAMS

We would like your input on alternative approaches to protecting Michigan’s coastal
wetlands. Please review the elements of the two alternative programs before answering

Question 8.

MICHIGAN ’S COASTAL WETLANDS
These programs target the preservation and restoration of Michigan’s coastal
wetland resources. Today, Michigan’s coastal wetlands account for 4% of
Michigan’s wetlands and cover 105,855 acres statewide. The number and health 0
coastal wetlands has decreased by 70% in the last 150 years.

 

   

 

PROGRAM PRIORITIES
Wetland protection programs can not do everything. They must prioritize their
efi'orts in order to accomplish their goals with their limited resources. Michigan
coastal wetland programs may prioritize:

Non Game Species Fish Habitat

Open Space near Cities Waterfowl Habitat

Water Quality & Flood Biodiversity
Control

When wetland pro yams prioritize a particular feature, other wetland features
associated with the primary focus will also be provided

Coastal wetland programs protect wetlands in two ways. These prog'ams divide
their effomnd resources between two types of activities:

0 Preservation of high quality coastal wetlands .
0 Restoration of coastal wetlands in poor condition J

Wetlands restored by the program are then preserved by the pro g‘am

 

 

is? _,

 

94

 

 

WHICH PROGRAM WOULD YOU PREFER?

The two programs described below take difi'erent approaches to coastal wetland
protection Please compare the two programs before answering Question 8.

 

PROGRAM A

PROGRAM B

 

How coastal wetland services
are prioritized by program

 

Primary Focus {

Non-Game Spades

Waterfowl Habitat

 

How the program effort and
resources are divided

 

Preservation of high quality
coastal wetlands

70% Preservation

50% Preservation

 

Restoration of costal
wetlands in poor condition

30 % Restoration

50% Restoration

 

 

How the program gets
coastal wetland sites from
volunteers

 

Purchase, voluntary easern cut,
or short term contract

 

Purchase Property

 

 

Ten-year Contracts

 

 

8. Ifthe state had to choose one of these two programs for protecting coastal

wetlands, which of these programs would you prefer?

(Circle one response)

a. ProgramA
b. ProgramB

 

95

 

 

 

 

PROPOSED STATE-WIDE PROGRAM

Now, we would like your input on one of the two proposals you have just considered
This proposed program would go into effect if voters approve it in a special referendum.
Please review the elements of the ballot proposal before answering Question 9.

MICHIGAN ’8 CURRENT WETLAND LAWS
State and federal laws try to prevent loss of wetland acreage by placing conditions
on wetland permits to avoid, minimize, and mitigate wetland loss. However, I
existing wetland laws do not always protect wetland quality. For example, i
I
!

      

 

protecting wetland acreage does not always protect the things that wetlands do.
Also, existing wetland laws do not fund preservation and restoration of wetlands.

 

____________._..__.-._________...n

PROGRAM IMPLEMENTATION ii
The proposed program would build on existing wetland programs by providing 1 i
funds for the preservation and restoration of Michigan’s coastal wetlands. ;
|

0 The proposed program would be implemented if voters approve it in a
special referendum.

0 The program would not be implemented if voters do not approve it.

 

l
l
l
l
I
fil

PROGRAM FUNDING

Wetland preservation and restoration cost money. The proposed program will be
funded by using a one-time only, income tax payment that would depend on a
household’s income.

The cost to your household will be a one-time. $90 payment that will be added to
yom' 2005 state income tax

 

All monies collected will he spent on coastal wetland protection.

\_

u_.__.__..__.W,M_.-a__.— Wu—Mb—u“ WW‘—a—-~_u——- __.___.___-. -...._—_.—-

 

96

 

 

 

PROPOSED COASTAL WETLAND PROGRAM AND COST

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PROGRAM A
Primary Focrs Non—Game Species
Preservation effort 70 %
Restoration effort 30%
Get Program Sites Purchase Property
Total acres a 5000 acres
in program
|
One-time only payment {I S 90, one time

 

. HOW WOULD YOU VOTE?

9. While there may be other possible program, we want to know how you would
vote on the ABOVE program.

If asked to vote in a state-wide election on a ballot issue for the coastal wetland
program described above, how would you vote? (Circle one response)

a Yes
b. No
c. Idon't know

 

97

 

 

 

HELP US UNDERSTAND YOUR VOTE

10. Please help us understand some of the reasons you voted the way you did.

a. If you answered “Yes” to the proposed program in Question 8, please
share with us some of the reasons you did so

b. Ifyou answered “No” to the proposed program in Question 8, please
share with us some of the reasons you did so.

c. If answered “I Don’t Know” to the proposed program in Question 8,
please share with us some of the reasons you did so.

 

98

 

 

INFORMATION ABOUT YOU

This section asks a few questions about your background so that we can

compare our results with the makeup of the state population Your responses are
completely confidential and will not be linked to your identity in any way.

 

11. Which best describes the area where you live? (Circle one answer)

a. An urban area
b. A suburban area
c. A rural area

12. How strongly do you agree with each of the following statements?
(Mark 12] one response for each function)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>‘ ‘5 I- ‘5
How strongly do you agree with the E g E g g g g g g a a
following statements? 3 (f a < g Sig a .53} 3 .§
m {g <: Q (g Q m G
a I contribute to/am a member of E] E] Cl Cl C]
an envn'onmental group.
b. I go fishing and/or hunting. Cl Cl E] Cl C]
c. I go camping and/or hiking. Cl [3 Cl E] El
’d. Ivote in all state and local elections. C] D D C] C]
e. Iwrrte letters tonewspapers/agencies Cl E] El [3 C]
to express my vrews
I follow community issues
f (e.g, watch public meetings). D D D D D
I manage my time well. D D Cl C] D
I am often late for appointments. C] D C] E] C]
i. I never seem to have enough time. D E] Cl C] E]
j. I have a hectic schedule. Cl D Cl Cl C]

 

 

 

 

 

 

 

 

 

 

99

 

 

 

l3.

14.

16.

17.

18.

How important is it to know that wetlands are there? (Circle one
response)

a. Extremely Important
b. Somewhat Important
c. Not Very Important
d. Not at All Important
e. Do Not Know

How important it is to know that wetlands will be there for future
generations? (Circle one response)

Extremely Important
Somewhat Important
Not Very Important
Not at All Important
Do Not Know

9.9-9.7?

. What is your age in years? (Circle one response)

a. 18 to 25 years e. 46 to 55 years
b. 26 to 35 years f. 46 to 65 years
c. 36 to 45 years g. More the 65 years

What is your gender? (Circle one response)
a. Female b. Male

What is your ethnicity and race? (Circle one response)

a. Hispanic, Latino, or Spanish origin
b. White

c. Afiican American or Black

d. Hawaiian or other Pacific Islander

e. Asian
f American Indian or Alaska Native
3 Other:

 

Generally speaking, do you think of yourself as a... (Circle one response)

a Democrat c. Independent
b. Republican (1. Other:

 

 

100

 

 

 

19.

20.

21.

22.

24.

What is the highest level of schooling that you have completed?
(Circle one response)

a. Some high school d. Associate degree, 2 year college
b. High school degree e. College degree, 4 year college
c. Some college f Advanced degree (MBA, MD, etc.)

We are interested in learning about the different ways people may earn
their living. Last week, were you? (Circle one response)

a. Working full-time d A homemaker
b. Working part-time e. Retired
c. Going to school f. Other:

 

What was your gross household income in 2003? (Circle one response)
$50,000 - $74,999

a. 0 to $14,999 e.

lg. $15,000 - $24,999 f. $75,000 - $99,999

9,. $25,000 - $34,999 3 $100,000 - $149,999
g. $35,000 - $49,999 h. $150,000 or more

Which category best describes your primary income source? (Circle one
response)

a Construction e. Farming

b. Forestry £ Government

c. Education g. Transportation
d Real estate h. None of these

. What is your household size?

 

a Number of adults (18 years old and up):
b. Number of children (under 18 years old):

 

.Which of the following best describes you? (Circle one response)

a. The person to whom the letter is addressed

b. A spouse or relative of the person to whom the letter is addressed.
c. A close friend

d. Someone else (fill in please)

 

Thank You! This completes our questions for you.

 

101

 

 

 

You may use the back cover to share your ideas or opinions with us.

Please place the survey in the envelope provided and return it to:

GL Wetland Study, Dept. of Resource Development, 323 Natural Resources
Building, Michigan State University, East Lansing, MI 48824— 1222

WEEHCAN STATE

UNIVERSITY

 

 

102

 

BIBLIOGRAPHY

AAPOR. Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for
Surveys. 3rd ed. Edited by American Association for Public Opinion Research.
Lenexa, Kansas, 2004.

Adamowicz, W., J. Louviere, and M. Williams. "Combining Revealed and Stated
Preference Methods for Valuing Environmental Amenities." Journal of
Environmental Economics and Management 26, no. 3(1994): 271-92.

Adamowicz, W., J. Swait, P. Boxall, J. Louviere, and M. Williams. "Perceptions Versus
Objective Measures of Environmental Quality in Combined Revealed and Stated
Preference Models of Environmental Valuation." Journal of Environmental
Economics and Management 32, no. 1(1997): 65-84.

Adamowicz, W., J. Louviere, and J. Swait. Introduction to Attribute-Based Stated Choice
Methods. Advanis, Alberta Canada, 1998 (a).

Adamowicz, W., J. Louviere, and M. Williams. "Stated Preference Approaches for
Measuring Passive Use Values: Choice Experiments and Contingent Valuation."
American Journal of Agricultural Economics 80, no. 1(1998): 64-75 (b).

Albert, D. A. Between Land and Lakes: Michigan's Great Lakes Coastal Wetlands.
Michigan State University Extension, December, 2003.

Azevedo, C. D., J. A. Hirriges, and C. L. Kling. "Combining Revealed and Stated
Preferences: Consistency Tests and Their Interpretations." American Journal of
Agricultural Economics 85, no. 3(2003): 525-537.

Bailey, D. Great Lakes Coastal Wetland Conservation/Preservation Programs within
Michigan. MSU Department of Resource Development, January 9th, 2003.

Barbier, E. B., Acreman, MC. and Knowler, D. Economic Valuation of Wetlands: A
Guide for Policy Makers and Planners. Gland, Switzerland: Ramsar Convention
Bureau, 1997.

Batie, S. 8., and J. R. Wilson. "Economic Values Attributable to Virginia's Coastal
Wetlands as Inputs in Oyster Production." Southern Journal of Agricultural
Economics (1978): 111-118.

Blamey, R., J. Gordon, and R. Chapman. "Choice Modelling: Assessing the

Environmental Values of Water Supply Options." The Australian Journal of
Agricultural and Resource Economics 43, no. 3(1999): 337-357.

103

Boxall, P. C., W. L. Adamowicz, J. Swait, M. Williams, and J. Louviere. "A Comparison
of Stated Preference Methods for Environmental Valuation." Ecological
Economics 18, no. 3(1996): 243-53.

Boyd, J ., K. Caballero, and R. D. Simpon. "Carving Out Some Space: A Guide to Land
Preservation Strategies." Resources for the Future, no. 136(1999): 10-13.

Canadian Wildlife Service. Where Land Meets Water: Understanding Wetlands of the
Great Lakes. Ottawa, Canada: Queen's Printer, 2002.

Carson, R. T., J. Wright, N. Carson, A. Alberini, and N. Flores. A Bibliography of
Contingent Valuation Studies and Papers. La Jolla, Calif.: Natural Resource
Damage Assessment, Inc, 1994.

Conservation Fund, The (2003) (a) Great Lakes Revolving Loan Fund, September 2003.
URL:
http://www.conservationfund.org/pagespinner.asp?article=2454&back=true.

Conservation Fund, The (2003) (b) Michigan Dune Alliance, September 2003. URL:
http://www.conservationfund.org/pagespinner.asp?article=2566&back=true

Costanza, R., R. d'Arge, R. de-Groot, S. F arber, M,, Grasso, B. Hannon, K. Limburg, S.
Naeem, R. O'Neil, J. Paruelo, R. Raskin, P. Sutton, and J. van den Belt . "The
Value of the World's Ecosystem Services and Natural Capital." Ecological
Economics 25(1998): 3-15.

Denison, M. S., and J. A. Schmid. Wetland Mitigation: Mitigation Banking and other
Strategies for Development and Compliance. Rockville, MD: Government
Institutes Inc., 1997.

Dempsey, D., M. D. Kaplowitz, "The Persistence of Local Wetland Ordinances in
Michigan." Michigan Environmental Law Review. 19(2002): 18-23.

Dillman, D. A. Mail and Internet survey: The Tailored Design Method. New York: John
Wiley & Sons, 2000.

Ducks Unlimited (2003) Ducks Unlimited North American Wetland Conservation Act,
August, 2003. URL: http://www.ducks.org/conservation/nawca.asp

Eamhart, D. "Combining Revealed and Stated Preference Methods to Value
Environmental Amenities at Residential Locations." Land Economics 77, no.
1(2001): 12-29.

Farber, S. "The Value of Coastal Wetlands for Protection of Property Against Hurricane

Wind Damage." Journal of Environmental Economics and Management 14, no.
2(1987): 143-151.

104

Farber, S., and R. Costanza. "The Economic Value of Wetland Systems." Journal of
Environmental Management 24(1987): 41 -5 l .

Ferraro, P. J ., and A. Kiss. "Direct Payments to Conserve Biodiversity." Science
298(2002): 1718-1719.

Ferraro, P. J ., and R. D. Simpson. "Cost-Effective Conservation: A Review of What
Works to Preserve Biodiversity." Resources for the Future, no. 143(2001): 17-20.

Gaddie, R. K., and J. L. Regens. Regulating Wetlands Protection: Environmental
Federalism and the States. Albany, NY: State University of New York Press,
2000.

Gosslink, J. G., E. P. Odum, and R. M.~Pope. The Value of the Tidal Marsh. Baton
Rouge: Louisiana State University, 1974.

Great Lakes Commission (2005) Restore the Greatness! The 2005 Great Lakes Program
to Ensure Environmental and Economic Prosperity, April 2005. URL:
http://www.glc.org/restore.

Hey, D., L. Nancy, and S. Philippi. A Case for Wetland Restoration. New York: John
Wiley & Sons, Inc., 1999.

Higgs, E. S. "What is Good Ecological Restoration." Conservation Biology 11, no.
2(1997): 338-348.

Hoehn, J. P., F. Lupi, and M. D. Kaplowitz. "Untying a Lancastrian Bundle: Valuing
Ecosystems and Ecosystem Services for Wetland Mitigation." Journal of
Environmental Management 68, no. 3(2003): 263-272.

Jaworski, E., and C. N. Raphal. Fish, Wildlife, and Recreational Values of Michigan's
Coastal Wetlands. Twin Cities, Minnesota, 1978.

Kaplowitz, M. D., F. Lupi, and J. P. Hoehn. Multiple-Methods for Developing and
Evaluating a Stated Choice Survey for Valuing Wetland Ecosystems.
Questionnaire Development, Evaluation, and Testing Methods. Edited by A. S.
Presser. New York: John Wiley & Sons, Inc., 2004.

Kcating, D. M. The Valuation of Wetlands: An Appraisal Institute Handbook. Chicago,
IL: Appraisal Institute, 1995.

Kiss, A. Making Biodiversity Conservation a Land Use Priority, ed. T. McShane, and M.
Wells. D.C., Colombia University Press, 2002.

105

Loomis, J. B., P. Kent, L. Strange, K. Fausch, and A. Covich. "Measuring the Total
Economic Value of Restoring Ecosystem Services in an Impaired River Basin:
Results from a Contingent Valuation Survey," Ecological Economics 33, no.

1(2000): 103-117.

Louviere, J ., and D. A. Hensher. "On the Design and Analysis of Simulated Choice or
Allocation Experiments in Travel Choice Modelling." Trans. Res. Rec.
890(1982): 11-17. '

Louviere, J ., and G. Woodworth. "Design and Analysis of Simulated Choice or
Allocation Experiments: An Approach Based on Aggregate Data." Journal of
Marketing Research 20(1983): 350-367.

Louviere, J. J ., D. A. Hensher, and J. D. Swait. Stated choice methods: Analysis and
applications. Cambridge: Cambridge University Press, 2000.

Lupi, F ., M. Kaplowitz, and J. P. Hoehn. "The Economic Equivalency of Drained and
Restored. Wetlands in Michigan." American Journal of Agricultural Economics
84:1355-1361, 2002.

Lupi, F., M. Gibson, 0. Arreola, M. Kaplowitz, and J. P. Hoehn. Michigan Great Lakes
Coastal Wetlands: Survey Results and Preferences for Wetland Programs.
Michigan State University, October, 2004.

McFadden, D. Conditional Logit Analysis of Qualitative Choice Behavior, ed. P.
Zaremba. New York, Academic Press, 1973.

Michigan Department of Environmental Quality (2003) Coastal Management Program,
State of Michigan, August 2003. URL: http://www.michigan.gov/deq/O,1607,7-
135-3313__3677_3696-1 1 188--,00.htm1

Michigan Department of Environmental Quality (2005) (a) Wetlands Protection, State of
Michigan, January 2005. URL: http://www.michigan.gov/deq/O,1607,7-135-
3313_3687---,00.html

Michigan Department of Environmental Quality (2005) (b) State, Federal and Local
Wetland Regulations, April, 2005. URL: http://www.michigan.gov/deq/0,1607,7-
135-33 1 3_3687-10801--,00.html.

Michigan Sea Grant. "Treasures of the Great Lakes: Recognizing the Value of Michigan's
Coastal Wetlands." Upwellings 25, no. 3(2002): 10.

Mitsch, W. J., and R. F. Wilson. "Improving the Success of Wetland Creation and

Restoration with Know-How, Time, and Self-Design." Ecological Applications 6,
no. 1(1996): 77-83.

106

Mitsch, W. J., and James G. Gosselink. Wetlands. Third Edition ed. New York: John
Wiley & Sons, Inc., 2000.

Morikawa, T., M. Ben-Akiva, and D. McFadden. "Discrete Choice Models Incorporating
Revealed Preferences and Psychometric Data", ed. P. H. Franses, and A. L.
Montgomery. Advances in Econometrics, vol. 16. Amsterdam; London and New
York, Elsevier Science JAI, 2002, pp. 29-55.

Morrisette, P. M. "Conservation Basement and the Public Good: Preserving the
Environment on Private Lands." Natural Resources Journal 41, no. 2(2001): 373-
426.

Morrison, M., J. Bennett, and R. Blarney. "Valuing Improved Wetland Quality Using
Choice Modeling." Water Resources Research 35, no. 9(1999): 2805-2814.

Morrison, M. "Understanding Local Community Preferences for Wetland Quality."
Blackwell Science Asia Pty. Lt. 3, no. 2(2002): 127.

National Research Council. Compensating for Wetland Losses Under the Clean Water
Act. Washington, DC: National Academy Press, 2001.

Ontario Nature (2003) The Great Lakes Wetlands Conservation Action Plan (GLWCAP),
August 2003, Federation of Ontario Naturalists. URL:
http://www.ontarionature.org/enviroandcons/wetlands/wet1 .html

Parker, D. P. "Cost-Effective Strategies for Conserving Private Land: An Economic
Analysis for Land Trusts and Policy Makers." Natural Resources Journal, 2004.

Pindyck, R. S., and D. L. Rubinfeld. Econometric Models and Economic Forecasts. 4th
ed: Irwin McGraw-Hill, 1998.

Swallow, S. K., T. Weaver, J. J. Opaluch, and TS. Michelman. "Heterogeneous

Preferences and Aggregation in Environmental Policy Analysis: A Landfill
Sitting Case." American Journal of Agricultural Economics 76(1994): 431-443.

US, Census Bureau, Census 2000. Geographic Area: Michigan, July 2004. URL:
http://factfinder.census.gov/

United States Senate (2005) Statements on Introduced Bills and Joint Resolutions, April
2005, The United States Senate. URL: GovTrack.us.

Vernberg, F., J. Winona, and B. Vernberg. The Coastal Zone: Past, Present, and Future.
Columbia, South Carolina: University of South Carolina Press, 2001.

Whitehead, J. C. "Economic Valuation of Wetland Resources: A Review of the Value
Estimates." Journal of Environmental Systems 22(1993): 151-161.

107

Wilson, M. A. "Economic Valuation of Freshwater Ecosystem Services in the United
States: 1971-1997." Ecological Applications 9(1999): 772-783.

Woodward, R. T., and Y.-S. Wui. "The Economic Value of Wetland Services: A Meta-
Analysis." Ecological Economics 37, no. 2(2001): 257-270.

Wooldridge, J. M. Introductory Economics: A Modern Approach: South- Western
College Publishing, USA, 2000.

World Atlas (2004) United States: Land Statistics, September 2004, Graphic Maps. URL:
http://worldatlas.com/webimage/countrys/namerica/usland.htm.

108

I-

lit.

TO