MEASURING THE RECREATIONAL VALUE OF CHANGES IN CORAL REEF

ECOSYSTEM QUALITY IN JAMAICA: THE APPLICATION OF TWO

STATED PREFERENCE METHODS

by

Peter E.T. Edwards

A dissertation submitted to the Faculty of the University of Delaware in partial

fulfillment of the requirements for the degree of Doctor of Philosophy in Marine Studies

Spring 2009

Copyright 2009 Peter E.T. Edwards

All Rights Reserved
 UMI Number: 3360224

Copyright 2009 by
Edwards, Peter E.T.

All rights reserved

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.

______________________________________________________________

UMI Microform 3360224

Copyright 2009 by ProQuest LLC
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
_______________________________________________________________

ProQuest LLC
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106-1346
MEASURING THE RECREATIONAL VALUE OF CHANGES IN CORAL REEF

ECOSYSTEM QUALITY IN JAMAICA: THE APPLICATION OF TWO

STATED PREFERENCE METHODS

by

Peter E.T. Edwards

Approved: ____________________________________________________________
Nancy M. Targett, Ph.D.
Dean of the College of Marine and Earth Studies

Approved: ____________________________________________________________
Debra Hess Norris, M.S.
Vice Provost for Graduate and Professional Education
 I certify that I have read this dissertation and that in my opinion it meets
the academic and professional standard required by the University as a
dissertation for the degree of Doctor of Philosophy.

Signed: ________________________________________________________
George R. Parsons, Ph.D.
Professor in charge of dissertation

I certify that I have read this dissertation and that in my opinion it meets
the academic and professional standard required by the University as a
dissertation for the degree of Doctor of Philosophy.

Signed: ________________________________________________________
Lee G. Anderson, Ph.D.
Member of dissertation committee

I certify that I have read this dissertation and that in my opinion it meets
the academic and professional standard required by the University as a
dissertation for the degree of Doctor of Philosophy.

Signed: ________________________________________________________
Joshua Duke, Ph.D.
Member of dissertation committee

I certify that I have read this dissertation and that in my opinion it meets
the academic and professional standard required by the University as a
dissertation for the degree of Doctor of Philosophy.

Signed: ________________________________________________________
Steve M. Thur, Ph.D.
Member of dissertation committee
 ACKNOWLEDGEMENTS

I would wish to thank my advisor George Parsons for his guidance and

mentorship during my time here at CMES. Also appreciated was the atmosphere of

collegiality and mentorship experienced from other faculty and students in the

Department. Dr Biliana Cicin-Sain must be specially mentioned for her role in

exposing me to the colleges interdisciplinary Marine Policy program. Valuable

advice, comments and suggestions from, Dr. Joshua Duke (Faculty of Resource

Agriculture, UD), Dr Andy Krueger (US Department of the Interior) and fellow

graduate students Ami Kang, Kelley Appleman, Allison Borchers and Stela Stefanova

were greatly appreciated.

I would like to acknowledge the support of the Environmental Foundation of

Jamaica for their instrumental role in providing initial scholarship support and funding

for some of the research activities. I would also like to thank the Latin American and

Caribbean Environmental Economics Program (LACEEP) for providing technical

support and funding for the study. Invaluable assistance and advice from Dr.

Francisco Alpizar, Dr. Dale Whittington and Dr. Fredrick Carlsson along with other

members of the LACEEP technical committee were greatly appreciated. I would also

like to extend heartfelt thanks to the Montego Bay Jamaica Airport, in particular Mr.

iv
Preston Jennings and staff. Without their permission the data collection would have

not been possible.

I would like to thank my family for their unconditional support as I chose to

embark upon this new journey into the uncharted world of environmental policy

analysis and natural resource economics. I would also like to thank my friend and

partner Aideen Murphy for her unwavering support and love over the entire course of

my study. I am also grateful for her invaluable assistance with the preliminary

sampling effort in the Montego Bay airport.

Finally I would like to give thanks to the Almighty God through which all

things are possible.

v
 DEDICATION

This dissertation is dedicated to my father, the late Errol Anthony Edwards (1937-

1987), and to my mother, Eppie Delorie Edwards.

My father for awakening my appreciation for nature and the outdoors, and especially

to my mother who single handedly raised her three children after my fathers passing.

Thank you for your continued love and prayers.

vi
 TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................... xii

LIST OF FIGURES ..................................................................................................... xiv
LIST OF ACRONYMS ............................................................................................... xvi
ABSTRACT ............................................................................................................... xvii

CHAPTER
1 NON-MARKET VALUATION OF ECOSYSTEM SERVICES ...................... 1

1.1 Introduction ............................................................................. 1

1.2 Jamaican Public Goods: The Economic Problem ................... 2

1.2.1 Market failure and externalities .................................................. 4

1.3 Study Objectives ...................................................................... 8

1.4 Outline ................................................................................... 11

2 CORAL REEFS, PROTECTED AREAS POLICY AND TOURISM ............ 13

2.1 Introduction ........................................................................... 13

2.2 Threats to Coral Reef Ecosystems ......................................... 14
2.3 The Impact of Coral Reef Decline on Tourism ..................... 16
2.4 National Protected Areas and Resource Management Policy 17
2.5 Sustainable Financing for Protected Areas............................ 20
2.6 The Jamaican Tourism Product ............................................. 22
2.7 Characteristics of Tourism Locations .................................... 23
2.8 Impact of Tourism on Coastal Ecosystems ........................... 25

3 NON-MARKET VALUATION AND RECREATIONAL DEMAND ........... 29

3.1 Introduction ........................................................................... 29

3.2 Contingent Valuation............................................................. 30

vii
 3.2.1 Contingent behavior ................................................................. 32

3.3 Choice Experiments............................................................... 33

3.4 Theoretical Rationale for CV and CE ................................... 34
3.5 Coral Reef Ecosystem Valuation Literature .......................... 35

3.5.1 Diving and marine protected areas ........................................... 36

3.6 Jamaican Non-Market Valuation Studies .............................. 40

4 CONTINGENT BEHAVIOUR STUDY ......................................................... 43

Feasibility of Sustainable Funding from Tourist User Fees ............................. 43

4.1 Introduction ........................................................................... 43

4.2 Contingent Behavior, Welfare Estimation and Sustainable
Revenues................................................................................ 43
4.3 Study Population ................................................................... 45
4.4 Survey Development and Sampling ...................................... 46
4.5 Survey Design ....................................................................... 49
4.6 The Impact of Different Payment Vehicles on Consumer
Surplus ................................................................................... 53

4.6.1 Consumer surplus for what? ..................................................... 56

4.6.2 Validity tests ............................................................................. 57

4.7 Theory and Econometric Models .......................................... 59

4.7.1 Choosing the probability response model - empirical

assumptions .............................................................................. 61

4.8 Parametric Valuation of Contingent Behavior ...................... 68

4.9 Non-parametric Valuation of Contingent Behavior .............. 71
4.10 Results ................................................................................... 74

4.10.1 Descriptive statistics ................................................................. 74

viii
 4.10.2 Summary of descriptive statistics Contingent behavior survey93
4.10.3 Response frequencies and non parametric welfare estimation . 94
4.10.4 Parametric analysis Tourism and environmental surveys ..... 98
4.10.5 Combined surveys parameter estimates ................................. 102

4.11 Policy Application - Sustainable Financing for Coastal

Resource Management ........................................................ 106

4.11.1 Policy steps ............................................................................. 106

4.11.2 An example: Towards implementation .................................. 109

4.12 Summary of Findings Contingent Behavior Survey ........ 113

5 STATED CHOICE STUDY........................................................................... 114

Measuring the Value of Changes in Coral Reef Ecosystem Quality
in Jamaica ...................................................................................................... 114

5.1 Introduction ......................................................................... 114

5.2 Choice Experiments and Welfare Estimation...................... 115
5.3 Survey Design ..................................................................... 119
5.4 Sampling .............................................................................. 127
5.5 Stated Choice Theory .......................................................... 127
5.6 Choice Models ..................................................................... 129

5.6.1 Stated Choice Model I ............................................................ 130

5.6.2 Valuation Stated choice model I.......................................... 132
5.6.3 Stated Choice Model II ........................................................... 135
5.6.4 Valuation Stated Choice Model II ....................................... 136

5.7 Results ................................................................................. 138

5.7.1 Descriptive statistics ............................................................... 138

5.7.2 Summary of descriptive statistics Stated choice survey ...... 144

5.8 Stated Choice Model I Parameter Estimates ....................... 144

ix
 5.8.1 Basic model ............................................................................ 145
5.8.2 Expanded model socioeconomic interactions ........................ 148

5.9 Stated Choice Model II Parameter Estimates ...................... 151

5.10 Welfare Estimates ................................................................ 154

5.10.1 Model I Welfare estimates................................................... 155

5.10.2 Model II Part-worth calculations ............................................ 159

5.11 Model Simulations............................................................... 161

5.11.1 Model I simulation Changes in surcharge price .................. 161

5.11.2 Model II simulation Changes in attribute quality ................ 163

5.12 Policy Application ............................................................... 168

5.13 Summary of Findings Stated Choice Survey.................... 170

6 POLICY IMPLICATIONS ............................................................................ 172

6.1 Policy Relevance ................................................................. 172

6.2 Use of Welfare Estimates in Resource Management Policy.172
6.3 Does the Method of Welfare Estimation Matter?................ 174
6.4 The actual impact of taxes on visitation rates ..................... 177
6.5 Does the Type of Tax Matter? ............................................. 178
6.6 How Should Revenues be Allocated? ................................. 180

7 CONCLUSIONS AND FINAL THOUGHTS ............................................... 182

7.1 Final Summary of Findings ................................................. 182
7.2 Study Limitations and Recommended Future Work ........... 184
7.3 From Policy Analysis to Policy Implementation................. 189

APPENDICES ............................................................................................................ 193

Appendix 1 Descriptive Statistics ........................................................... 194
Appendix 2 Non Parametric Results ....................................................... 202
Appendix 3 Parametric Results ............................................................... 203

x
 Appendix 4
Sample Copies of Surveys ................................................... 213
Appendix 5
Jamaican Tourism Statistics (JTB 2008) ............................. 263
Appendix 6
Respondent Feedback: Likes and dislikes
of Jamaica visit .................................................................... 264
REFERENCES ............................................................................................. 275

xi
 LIST OF TABLES
Table 1 Selected coral reef valuation studies ....................................................... 38
Table 2 Reasons for divergence between WTPTOUR and WTPENV. ...................... 58

Table 3 Descriptive statistics Contingent behavior survey ............................... 76

Table 4 Stop over arrivals by intended area of stay (JTB, 2008) ......................... 79

Table 5 Respondent rating of coastal attributes.................................................... 81

Table 6 Respondent rating of underwater attributes. ............................................ 84

Table 7 Respondents reported level of awareness of environmental issues ......... 88

Table 8 Kaplan-Meier-Turnbull calculations of WTP ......................................... 96

Table 9 Linear regression for tourism and environmental surveys (Basic and
expanded models) .................................................................................... 99

Table 10 Comparisons of expected willingness to pay

Tourism vs Environmental tax .............................................................. 101

Table 11 Combined Data - Linear regression output ........................................... 103

Table 12 Combined Data - Linear regression output with categorical dummy

variables................................................................................................. 104

Table 13 Estimated annual coastal resource management costs Jamaica. ........ 110

Table 14 The impact of environmental taxes on visitation rate and potential

revenues. ................................................................................................ 111

Table 15 Design stages for choice experiments ................................................... 116

Table 16 Possible levels of quality change in Jamaican coastal ecosystems. ...... 122

Table 17 Attributes, levels and their corresponding variables used in choice

scenarios ................................................................................................ 124

xii
Table 18 Descriptive statistics State choice survey ........................................... 139

Table 19 Fixed parameter MNL basic Model I .................................................... 146

Table 20 Random parameter Mixed Logit basic Model I .................................... 147

Table 21 Fixed parameter MNL, expanded Model I ............................................ 149

Table 22 Random Parameters Mixed Logit expanded Model I............................ 150

Table 23 Fixed parameter MNL Model II ............................................................ 152

Table 24 Random Parameters Mixed Logit Model II........................................... 153

Table 25 Welfare associated with coastal ecosystem quality changes ................. 156

Table 26 Annual welfare values associated with quality changes and total loss . 158

Table 27 Implicit Prices for each Attribute (Mixed Logit Model) ....................... 159

Table 28 Four scenarios of quality change used in the Model II simulation........ 165

Table 29 Simulated welfare estimates for the four scenarios of quality change .. 166

Table 30 Aggregate welfare for changes in attribute quality. .............................. 167

Table 31 Comparison of welfare estimates based on the type of survey and

econometric analysis ............................................................................. 175

xiii
 LIST OF FIGURES
Figure 1 Island of Jamaica, major towns and cities ............................................... 24

Figure 2 Environmental information presented in the coral reef survey. .............. 50

Figure 3 Example of the policy information presented in both versions of the

survey ...................................................................................................... 51

Figure 4 Censored response model (adapted from Hanemann and

Kaninnen, 1999) ...................................................................................... 63

Figure 5 Truncated canonical response model (by shrinking) ............................... 64

Figure 6 Response probability models with indifference and negative

preferences.............................................................................................. 66

Figure 7 Graphical representations of WTP distribution and measures of central

tendency (Hanemann and Kaninnen, 1999, Pg 327) ............................... 69

Figure 8 Location of hotels/accomodation stayed at by sampled respondents ...... 78

Figure 9 Frequency distribution of country of origin Contingent

Behavior Survey. ..................................................................................... 80

Figure 10 Respondent rating of beach sand. ............................................................ 82

Figure 11 Respondent rating of beach cleanliness. .................................................. 82

Figure 12 Respondent rating of swimming water clarity......................................... 83

Figure 13 Respondent rating of coral reef health..................................................... 85

Figure 14 Respondent rating of fish abundance. ..................................................... 85

Figure 15 Respondent rating of abundance of other sea creatures. ......................... 86

Figure 16 Respondent rating of underwater visibility. ............................................ 86

Figure 17 Respondent awareness of air pollution issues. ........................................ 89

xiv
Figure 18 Respondent awareness of water pollution issues..................................... 89

Figure 19 Respondent awareness of solidwaste pollution issues............................. 90

Figure 20 Respondent awareness of nature conservation issues. ............................ 90

Figure 21 Respondent awareness of climate change issues. .................................... 91

Figure 22 Respondent awareness of deforestation issues. ....................................... 91

Figure 23 Respondent awareness of reef management issues ................................. 92

Figure 24 Respondent awareness of population issues. ........................................... 92

Figure 25 Comparison of response frequencies between survey types. .................. 95

Figure 26 Flow diagram Policy steps for sustainable revenue feasibility

assessment. ............................................................................................ 108

Figure 27 Impact of environmental taxes on visitation rate and

potential revenues. ................................................................................. 112

Figure 28 Stated choice question presented to respondents................................... 126

Figure 29 Ratio of management options selected by all respondents. ................... 141

Figure 30 Frequency distribution of country of origin Stated choice survey. .... 142

Figure 31 Respondent awareness of Jamaica's coral reef management issues. ..... 143

Figure 32 Simulated respondent support for an environmental surcharge. ........... 163

xv
LIST OF ACRONYMS

CV Contingent Valuation
CB Contingent Behavior
CBA Cost Benefit Analysis
CBD Convention on Biological Diversity
CE Choice Experiments
GDP Gross Domestic Product
IIA Independence of Irrelevant Alternatives
JCDT Jamaica Conservation and Development Trust
JTB Jamaica Tourist Board
MBMP Montego Bay Marine Park
MBMPT Montego Bay Marine Part Trust
MNL Multinomial Logit
MXL Mixed Logit
NCRPS Negril Coral Reef Preservation Society
NEPA National Environment and Planning Agency
NEPT Negril Environmental Protection Trust
NJCA Northern Jamaica Conservation Association
NOAA National Ocean and Atmospheric Agency
NRCA Natural Resources and Conservation Authority
PARC Protected Areas Resources Conservation
RP Revealed Preference
RPL Random Parameter Logit
SIDS Small Island Developing State
SP Stated Preference
TEF Tourism Enhancement Fund
USAID United States Agency for International Development

xvi
 ABSTRACT

Coral reefs and their associated ecosystems provide important ecological and

economic benefits for small island developing states such as Jamaica. They are

however threatened by natural and anthropogenic impacts such as: coastal pollution,

rapid coastal development, overfishing and global warming. These threats require

environmental management in order to mitigate the negative impacts. Management

activities are, however, dependent on funding which is often limited or absent in island

nations such as Jamaica. Given the limitations on financing for coastal protection and

resource management it is understandable that policymakers might consider

alternative mechanisms of generating funding. These funds could be raised in the

form of taxes on the primary resource users. In the case of Jamaica this would be

tourists visiting the island.

This study applies two types of stated preference techniques to provide

estimates of the recreational values associated with quality changes to Jamaicas

coastal ecosystem. The results of the study are used to discuss the feasibility of

generating revenues for the sustainable financing of ocean and coastal management in

Jamaica. It is hoped that the findings of this dissertation will contribute to policy

development, implementation leading to better management of coastal resources in

Jamaica. A random airport intercept of tourists who visited Jamaica was conducted in

xvii
January 2008. Two separate groups of respondents received either a self administered

contingent behavior survey or a stated choice survey.

The findings of the contingent behavior study show that the label of the tax

as well as the respondents awareness of the institutional mechanisms for

environmental protection and tourism are important to their decision framework. The

results show that current coastal zone management activities could be financed from

the introduction of a $2 per person environmental tax in addition to the existing $10

tourism surcharge. The potential negative impact on the annual visitation rate to

Jamaica from the introduction of this additional tax appears to be negligible ( 0.9%).

However any decline in visitation rate could be mitigated by providing visitors with

information on how tax revenues are allocated. It is important that the provision of

this information is complimented by ensuring that the funds are indeed used for the

purposes specified.

The findings of the stated choice study confirm a priori expectations that a

decline in present day beach, water and coral reef quality is least preferred while

improvements to the existing environmental quality is most desired by respondents.

For a hypothetical decline in quality from the status quo, that is, good beach and water

quality and fair marine life, the mean welfare loss for each individual was calculated at

US$97(2008). Mean welfare gain for an improvement in quality from the status quo

was estimated at $22 per individual. Access value (welfare loss from removing

Jamaica from respondents choice set) calculations of $128 per-person also confirm

xviii
that there is a significant consumer surplus associated with a typical coastal vacation

in Jamaica.

Both studies show that tourists have consumer surpluses associated with a

beach and coral reef vacation that greatly exceed the costs of coastal management.

The studies also show that environmental user fees could be an attractive sustainable

financing option. The findings also seem to refute the oft touted concern that the

imposition of additional taxes would ruin the tourism industry in Jamaica.

This study demonstrates an approach that could be used as part of the policy

framework for resource protection and sustainable management of important coastal

ecosystems and natural resources in other countries dependent on coral reef based

tourism.

xix
 CHAPTER 1

NON-MARKET VALUATION OF ECOSYSTEM SERVICES

1.1 Introduction
Coral reef ecosystems hug tropical coastlines and offer protection from the

pounding of waves and scouring currents on a daily basis but, more importantly, protect

against the worst ravages of storms and hurricanes. They are able to grow in high-energy

environments and reef growth gradually builds up huge limestone structures, which

buffer and defend the coastline. In addition reefs also provide the major source of sand,

which builds land and replenishes beaches (Edwards, 2002). Coral reefs are also well

known for their biological diversity and high ecological productivity (Spurgeon, 1992).

They provide humans with a source of livelihood, food, recreation, and medicinal

compounds. Coral reefs and their associated ecosystems are, however, threatened by

natural and anthropogenic impacts such as: coastal pollution, rapid coastal development,

overfishing and global warming. Mitigating the anthropogenic threats to these natural

resources requires management and this in turn is dependent on funding which is often

limited or absent in an island nation such as Jamaica.In light of the existing limitations on

financing for coastal protection and resource management in addition to significant

resistance from tourism industry stakeholders towards increases in corporate taxes,

policymakers may consider alternative mechanisms to generate funding such as taxes on

1
resource users. In the case of Jamaica this would be tourists and the tourism industry. In

order to justify imposing additional costs on key stakeholders it is critical that the benefits

associated with the resource are understood. These benefits include market and non-

market benefits.

It is important that when considering tourism and the environment there is a

fundamental question that should be asked. That is; what are the net social benefits (to a

nation and its people) provided by the development of the environmental assets (sun, sea,

sand, vegetation) on which tourism depends (Dixon et. al., 2001)? Caribbean

governments typically incur costs in order to promote and provide infrastructure for the

tourism sector, while economic benefits for residents arise from tourists expenditures

and profits for businesses (hotels and attractions). Tax revenues are a potential benefit

for the government and these might be generated through general consumption and

corporation taxes associated with tourism, or from taxes specifically targeted at tourists

and the tourism sector (Dixon et. al., 2001).

1.2 Jamaican Public Goods: The Economic Problem

Environmental goods are, generally, a subset of a general class of goods called

"public goods. One definition of a public good is one for which it is not practical to

exclude consumers (Boardman et. al., 2001). A pure public good is one where it is

impossible to exclude some consumers, that is, if the good is provided for one person, it

is provided for all (Boardman et. al., 2001). In other words a public good is non

excludable and non rival. Two important problems stem from these properties of public

2
goods. First, it will never be profitable to produce public goods privately, because the

producer who incurs the cost of production cannot prevent the consumer from using the

good freely (it is non-excludable). So the responsibility for the provision of public goods

falls on the government (e.g. national defense and environmental goods). The

government must be in charge of supplying public goods, since it will not be profitable

for the private market to provide them (Boardman et. al., 2001, p82).

For the Jamaican case, the environmental or public good that is considered for this

study is the coastal and nearshore marine environment. This includes coastal vegetation,

beaches and rocky intertidal zones, seagrass beds and other submerged aquatic vegetation

as well as the coral reef ecosystem (corals, fish, macro and micro invertebrates etc).

There are of course other environmental goods that may be directly connected to marine

ecosystems such as forests (watersheds), riverine systems and possibly ambient air

quality. The government of Jamaica therefore has a responsibility to provide these public

goods.

However as has been mentioned above, pure public goods because of their

characteristics of non-excludability and non-rivalry are prone to market failure, where

uncoordinated markets driven by parties working in their own self interest are unable to

provide these goods in desired quantities. The optimal use of coral reef resources is

distorted by the fact that reefs are often open access in nature and that many of the

products and services provided by coral reefs have (quasi-) public good characteristics

resulting in markets for coral reef services being absent or under-developed (Dixon et. al.,

2001; Brander et. al., 2007). For example over exploitation of fisheries due to its open

3
access nature or coastal pollution from unregulated industrial and agricultural activities.

The quasi public good nature of coral reefs and associated ecosystems often result in

them being undervalued in decisions relating to their use and conservation.

While recognizing that there are various types of economic services provided by

coral reefs, this study focuses on the recreational values associated with this resource.

The aim being, to provide results that can be used to address specific policy issues related

to the management of coral reefs and beaches, such as the feasibility of charging user

fees. Furthermore, recreation and tourism values are often the most important direct and

indirect use values of coral reefs. Since Jamaicas tourism model is largely based on

coastal tourism its success depends to a large extent on the quality of the coral reefs and

beaches. The quasi public good described here is subject to market failures, some of

which are related to tourism activities and some to wider environmental threats. These

will be discussed below.

1.2.1 Market failure and externalities

The quasi public good defined above are the beaches, swimming waters and coral

reefs which are critical for the Jamaica coastal tourism product. The negative impacts

from tourism activities are discussed in more detail in section 2.8 but include impacts

such as, coastal construction, loss of sea grass beds and reefs, beach erosion, use of water,

congestion effects and coastal pollution. By and large these negative externalities are not

accounted for by Jamaican tourism operations.

4
 The pressure placed on these coastal resources by activities associated with the

tourism industry creates the need for funding for environmental conservation and

management. These negative externalities associated with the tourism industry are

evidence of market failure. The protection of beaches, pollution prevention (water or air)

and adequate solid waste management require financial resources that are usually

provided by the national budget. Given Jamaicas high debt-to-GDP ratio of almost 130

percent (CIA, 2008), budgetary allocations for environmental management in Jamaica are

very limited and inadequate for providing protection and management.

There are other issues of market failure or distortion that are relevant to the

Jamaican (and to a wider extent the Caribbean) tourism industry. This has to do with the

capture of the economic rent associated with the coastal environmental good or more

specifically rent seeking behavior by tourism industry stakeholders. First of all it should

be noted that a visitor to the island who enjoys the beaches, green mountains and idyllic

towns experiences a benefit from these environmental assets and this stays with them.

The struggle therefore is who captures the rents, the country, the tourism developers or

the visitors (Dixon et. al., 2001).

Rent capture can be described as the process whereby an economic agent, either

the government or a private concern or individual, secures or "captures" part of the

economic rent associated with a site (Tietenberg, 2003, p 78). The lobbying and rent

seeking behavior of the tourism sector highlights issues of taxation and investment

incentives, which are not, strictly-speaking, environmental issues (Dixon et. al., 2001) but

can result in market failure. Investment incentives are one example of direct transfers to

5
developers. Sometimes, however, large transfers take place without any money changing

hands. For example when a developer is allocated a prime site, say a beach front

property, without competitive bidding, the government has in effect given part of the

economic rent to the developer (Dixon et al., 2001).

Rent seeking and misallocation of rents therefore create the potential for

environmental impacts to arise from the tax and incentive structure. For example

subsidized investment can lead to congestion in prime tourist locations, stressing local

resources and ultimately dissipating resource rents by making the site less attractive,

ultimately lowering the amount a visitor will be willing to pay (Dixon et. al., 2001).

Presently Jamaican tourism developers currently receive economic incentives such as

reduced barriers (fast tracking) for environmental permits as well as tax breaks (15-20

years) and reduced general consumption taxes (Jamaica Trade and Invest, 2007).

Tourism areas also benefit from large scale infrastructure development by the

government of Jamaica such as Highway 2000 (Ministry of Transport and Works, 2006).

There have been instances where the government pays for the completion of the

construction of the hotel and then hands it to a resort company to manage (Office of the

Contractor General, 2006). These examples are further evidence of market failure.

This is not to say that there are no positive externalities that arise from tourism

development such as employment and skills generation. However the size of these

externalities has been poorly documented. There are also questions about the real

economic impact of tourism to the local economy (Boadway and Shah, 1992). That is the

repatriation of profits from local or internationally owned hotels result in leakage of

6
tourism receipts out of the country. Therefore, even with some positive externalities,

there are strong reasons to think that the current incentive regime may not be the most

effective tool to stimulate tourism development (Dixon et. al., 2001; Boxhill, 2004). The

subjective argument that could be made here is that as the owners of the sun, sea, and

sand resources, the Jamaican people (through their government) should be the

beneficiaries of these rents. The policy question for the government is how to most

effectively capture and use these rents. The two most obvious methods are tourist user

fees or taxation or incentive schemes.

As outlined above, incentive schemes suggest that some inefficiency or market

failure will occur. However when access to a particular environmental resource can be

controlled (for example through a port of entry), charging user fees provides a simple

mechanism to capture part of the rents being generated. It should be noted that user fees

do not provide a practical means of capturing (all) the rents being generated. More

general taxation schemes are required so that tourism entities internalize the costs of the

negative externalities. However, given the extremely influential and powerful lobbying

(rent seeking) by the tourism industry, the feasibility of implementing general taxation

mechanisms is unlikely.

Dixon et al. (2001) provide a more detailed and comprehensive treatment of the

issue of rent capture in Caribbean tourism. They also discuss the issue of using the rents

(or a portion thereof) to maintain the very rents that provides the basis for the tourism

product. It is with this in mind that I attempt to provide information can be used to guide

7
policymakers about what instruments could be used to capture a portion of the rents that

visitors to the island receive.

1.3 Study Objectives

The main objective of this dissertation is to provide policy relevant information

based on estimates of the economic values associated with quality changes to Jamaicas

coastal ecosystem. It is hoped that the results of this research will provide information

that will assist the development of the Protected Area System Master Plan (PASMP) and

assist the policy process. It is envisaged that welfare estimates for the coastal ecosystems

can be incorporated into the work of the Financial Sustainability working group, a sub-

group of Jamaicas protected areas policy program. The findings can therefore be used to

guide the possible development of revenue generation instruments for the sustainable

management of the resources of protected and multi-use areas across the island. This

research can potentially contribute to the development of a national system of access fees

and general funding for natural resource management. It can also inform the

development and implementation of Jamaicas National Park System policy across the

island.

To achieve these goals, I conducted two separate stated preference (SP) surveys.

The first was a contingent behavior survey. I randomly intercepted persons in the

departure terminals at the Donald Sangster International Airport in Montego Bay,

Jamaica and distributed a self-administered written survey. The survey posed questions

to people about their willingness to travel to Jamaica if their travel costs were increased

8
as a result of higher tourism surcharges. A key part of this survey was the test of the

payment vehicle. I considered two types of tourism surcharges (taxes) namely a tourism

tax and an environmental tax. This study attempts to also add to the literature on the

effect of providing information on the respondents willingness to pay for improvements

in environmental quality in the coastal and marine environment (Blomquist and

Whitehead 1998) but was ostensibly focused on the goal of evaluating the feasibility of

implementing a special tourism environmental tax.

The two types of taxes presented to respondents are described within different

types of institutional frameworks. The type of institutional mechanism that ensures the

protection (preservation) of the given non-market good may also be important to the

respondents decision framework. This dissertation therefore provides a test of the

potential effects of different types of institutional frameworks on respondents consumer

demand. I model contingent behavior for tourists that receive two slightly different

scenarios and formulate hypotheses about how consumer demand may differ across

individuals. Based on the results of the contingent behavior predictions and the

differences between the payment vehicles, I then discuss the feasibility of generating

revenues for the sustainable financing of ocean and coastal management in Jamaica.

The second survey is choice experiment (CE) survey that was used to estimate the

recreational value of changes to key coral reef ecosystem attributes to tourists. The use

of a choice experiment methodology allows for the estimation of values associated with a

decline or an improvement in coastal environmental quality. The CE survey was

administered to persons in the departure concourses at the airport in Montego Bay.

9
Respondents were faced with a series of choice scenarios where they were asked to

choose between hypothetical environmental management scenarios that would result in

either a change to the quality of the coastal environment or keeping it in its current

condition. The estimated values can be used as part of the policy decision making

process by considering the benefits and costs of implementing management actions that

can facilitate ecosystem resilience (increased environmental carrying capacity) at the

local level. It should be noted that this is the first use of the stated choice technique to

value coral reef ecosystem quality changes in Jamaica. The willingness to pay for an

improvement or to prevent a decline in environmental quality will be estimated. These

results can therefore be used as part of the decision framework for resource protection

and management.

The combined results of the two studies could also be applied to tourism

development planning particularly with respect to deciding appropriate levels of resource

use given the carrying capacity of the associated coral reef ecosystems. The study also

has relevance to the wider Caribbean and other tropical regions as nations continue to

place greater emphasis on coastal or nature based tourism development within and

adjacent to protected areas. The estimations of welfare value from this study can also be

used in benefit transfer studies for locations with similar attributes.

It should be noted that this study is not a benefit cost analysis study per se. As

part of the overall research objectives I attempt to estimate one portion of the economic

benefits associated with coral reefs and beaches (recreational) but do not attempt to

present costs in order to facilitate optimizing the allocation of benefits and costs. It

10
merely attempts to provide information that can be used to inform decisions on how best

to allocate funding for coastal ecosystem management and feasibility of implementing

one type of welfare capture that is via tourist user fees.

1.4 Outline
Chapter two begins with a description of the Jamaican tourism product and the

role that the coral reef and associated ecosystems play in supporting the islands tourism

industry.

Chapter three provides an overview of non market valuation literature with a

specific focus on stated preference methods as they are applied to coral reef and coastal

ecosystem valuation studies.

Chapter four presents findings of the contingent behavior study. This includes

background information and literature on the economic theory that guides the method and

demonstrates how this theory is used to provide the basis for welfare estimation and

prediction of contingent behavior. It also outlines the econometric models that were used

in the analysis of the survey data. This chapter also includes details on the study design

including the sample frame and development of the survey instrument. It also presents

descriptive statistics and results of the analyses. It ends with the presentation of an

example of a policy application and a summary of the findings.

Chapter five presents the findings of the stated choice survey. This includes

background information and literature on the economic theory that guides the CE

method. It also discusses how this theory is used to provide the basis for welfare

11
estimation based on changes in environmental quality. The models associated with the

choice experiment design are also outlined. The chapter also presents the results, namely

the descriptive statistics and the parameter estimates from the econometric analyses as

well as welfare calculations. It ends with a discussion of possible policy applications and

a summary of the findings.

Chapter six discusses some overall policy implications of the results from the two

studies and examines the role of welfare estimation in resource management policy. It

also compares the findings of the two studies and discusses how they can be incorporated

into a wider policy framework.

Finally, chapter 7 offers some conclusions based on the overall findings and

provides final comments and recommendations for future studies.

12
 CHAPTER 2

CORAL REEFS, PROTECTED AREAS POLICY AND TOURISM

2.1 Introduction
Non-market valuation techniques are widely applied over a wide range of goods

and services and their use as a tool for natural resource management policy is now fairly

common across several countries. In many instances these studies are used to support

decisions on the implementation of user fees for national parks and marine protected

areas (Chase et al., 1998). Common to most of these studies is the estimation of

consumer surplus or welfare often expressed as willingness to pay (WTP). It is often

applied in the context of public goods such as air and noise pollution. It is also used in

damage assessments and cost benefit analyses for various types of development projects

(Bateman et al., 2002).

Non-market valuation techniques have also gained traction in valuing quasi public

goods in particular determining values associated with recreation. These studies typically

estimate the recreational values associated with a range of environments and activities.

These include hiking, mountain climbing, boating, river rafting. There has also been

research conducted on marine based recreational activities such as beach use, snorkeling,

scuba diving and sport fishing. However, many of these studies often tend to focus on

13
one activity for example scuba diving (Arin and Kramer, 2002; Weiglus et al., 2003).

There are very few these studies that attempt to look at contingent market participant

values for ecosystem services as a whole. This study applies two non-market valuation

techniques in order to provide estimates of value for the recreational services associated

with coral reefs and beaches located within managed coastal areas.

2.2 Threats to Coral Reef Ecosystems

Coral reefs and their associated ecosystems are threatened by direct anthropogenic

impacts such as; coastal pollution, rapid coastal development and overexploitation

(Carpenter et al., 2008). As development continues to alter the landscape the volume of

freshwater runoff will increase. This terriginous runoff may carry large amounts of

sediment from deforested land, high levels of nutrients from agricultural areas or septic

systems, as well as pollutants such as petroleum products or insecticides. The combined

impacts of direct sedimentation onto the reef and increased turbidity of coastal water due

to eutrophication, will result in decreased light availability for corals which in turn may

cause bleaching (Brown and Ogden 1993; Bryant et al., 1998). In addition, increases in

the amounts of nutrients enhance the growth of other reef organisms such as fleshy

macro-algae which may out-compete the corals for space on crowded reefs.

In addition to runoff, outflows from water treatment plants and large power plants

are additional threats to coral reefs. Sewage treatment facilities greatly increase the

nutrient levels in the receiving waters adjacent to their outflow pipes while large power

14
plants alter water temperatures by discharging hot water into the coastal waters (UVI,

2001).

In order to feed a growing global population, overfishing of reef fish stocks have

continued to increase thus threatening the future of sustainable fisheries. The removal of

large numbers of reef fish has caused the coral reef ecosystems to become unbalanced

and allowed more competitive organisms, such as algae, which were once controlled by

large fish populations, to become dominant on reefs in many regions (UVI, 2001).

Decreased fishing yields also cause drastic changes in fishing techniques including the

use of smaller mesh diameters for fish pots (traps) as well as the use of explosives or

poisons (Richmond, 1993). These fishing practices result in reduced numbers of viable

reproductive recruits and cause severe structural damage to reefs (respectively).

Harvesting of corals and other reef associated organisms occur as a result of the

marine ornamental and tourism trades. This also poses a threat to coral reefs worldwide.

Coral reefs also face threats from commercial and private vessels for example from the

leakage of fuels into the water and the occurrences of spills by large tankers. Anchor

damage is also very common from small and medium sized vessels while the grounding

of large sea-going vessels also results in large sections of coral reefs being destroyed.

In addition to the localized anthropogenic impacts affecting coral reef ecosystems,

the impact of natural weather events, global climate change and ocean acidification needs

to be taken into account. The most recognized of these natural events are hurricanes, or

typhoons, which bring large and powerful waves to the tropics (UVI, 2001). These storm

waves can cause large scale physical damage to coral reefs as well as scouring of

15
seagrass beds and displacement of sand from beaches. Following a storm, slow growing

corals are easily overgrown by faster growing algae. In addition, these weather events

are often associated with heavy rainfall which increases runoff and sedimentation. It is

postulated that global climate change may result in the increase in frequency and

intensity of hurricanes and typhoons (Pielke and Landsea, 1998; Emanuel, 2005).

Under conditions expected in the 21st century, global warming and ocean

acidification will compromise carbonate accretion, with corals becoming increasingly

rare on reef systems. This is expected to result in reef communities that have lower

diversity and carbonate reef structures that fail to be maintained. Climate change also

exacerbates the aforementioned local stresses from declining water quality and

overexploitation of key species, driving reefs increasingly toward the tipping point for

functional collapse (Hoegh-Guldberg et al., 2007).

2.3 The Impact of Coral Reef Decline on Tourism

The combined effects of local anthropogenic activities and climate change are

likely to fundamentally alter the attractiveness of coral reefs to tourists which is an

important consideration for the many low-income coastal countries and developing small

island states lying within coral reef regions (Hoegh-Guldberg et al., 2007). Under-

resourced and developing countries (for example small island developing states) have the

lowest capacity to respond to climate change, but many have tourism as their sole income

earner and thus are at risk economically if their coral reefs deteriorate. For instance,

16
tourism is a major foreign exchange earner in the Caribbean basin and in some countries

accounts for up to half the gross domestic product (Bryant et al., 1998).

Mitigating the global climate change and ocean acidification impacts will require

building ecosystem resilience at the local level. In other words resource managers and

coastal resource policies must first reduce the influence of local stressors such as

declining water quality, coastal pollution, and overexploitation of key functional groups

such as herbivores (Hughes et al., 2003). These types of action are most likely to assist

coral reefs through the decades of stress that inevitably face them.

The incorporation of non-market valuation techniques should play an important

role in the policy decision making process. Non-market valuation can also provide

information that might be useful for sustainable tourism development. For example

information on consumer surplus associated with coastal ecosystems can be used to set

levels of carrying capacity that will meet sustainable development goals for the tourism

industry. The current practice of providing tax exemptions for tourism development

projects which in turns leads to over congestions is another reason for the use of non-

market valuation techniques as part of any sustainable development policy.

2.4 National Protected Areas and Resource Management Policy

As mentioned previously coral reefs are threatened by natural and anthropogenic

impacts such as coastal pollution, rapid urban development, overfishing and global

warming. Mitigating the anthropogenic threats to these natural resources requires

management. One way of preserving important coastal resources is through the creation

17
of national parks and protected areas. The Government of Jamaicas Policy for the

National System of Protected Areas of November, 1997 (page 5) states that:

1. "the system is an expression of Jamaicas commitment to protecting its

environment and resources that are of recognized national and
international importance"

2. "the system should be an essential tool for environmental protection,

conserving essential resources for sustainable use, helping to expand and
diversify economic development, and contribute to public recreation and
education".

In 1989, the Government of Jamaica, with support from United States Agency for

International Development (USAID), launched the Protected Areas Resources

Conservation (PARC) Project. The Natural Resources and Conservation Authority

(NRCA) was the executing agency for the project, which was implemented in

collaboration with the Planning Institute of Jamaica, and the Jamaica Conservation and

Development Trust (JCDT), a local nongovernmental organization (NGO). Under this

project, two National Parks were established, the Blue and John Crow Mountains

National Park and the Montego Bay Marine Park. These parks allow for the effective

management, close monitoring and evaluation of protected areas.

Under the NRCA Act, management of these two parks was delegated to two

NGOs in 1996. Management of the Blue and John Crow Mountains National Parks was

delegated to the JCDT while management of the MBMP was delegated to the Montego

Bay Marine Park Trust (MBMPT), an NGO established for the sole purpose of managing

that Park. The delegation process involved the NGOs preparing a comprehensive

management plan for all aspects of the areas to be managed in collaboration with the

18
NRCA (now renamed National Environment and Planning Agency NEPA) and any

other agencies with major responsibility in the area. The potential management or co-

management organization prepares a proposal for managing the area in accordance with

the management plan. Once the entity has met the criteria outlined by NEPA it is

delegated management authority. Delegation instruments outlining roles and

responsibilities of each party are agreed upon and signed.

A second national marine park was declared in 1998, the Negril Marine Park

(NMP) and was established under the NRCA Act. The management of that Park is

delegated to the Negril Coral Reef Preservation Society (NCRPS). Montego Bay and

Negril therefore share similar characteristics with respect to the institutional framework

for coastal zone management. The third major resort town of Ocho Rios also has a

protected marine park area. The Ocho Rios marine park, was however, not officially

designated and so the environmental advocacy and on the ground activities are

undertaken primarily by two NGOs, Friends of the Sea (Ocho Rios) and the Northern

Jamaica Conservation Association (NJCA). NEPA is responsible for overall coastal zone

monitoring and management across the entire island.

The Forestry Department, Fisheries Division, National Environment and Planning

Agency (NEPA), and Jamaica National Heritage Trust are responsible for protected areas

declared under the following pieces of legislation: Forest Act, Beach Control Act, Fishing

Industry Act, Natural Resources Conservation Authority Act, the Wild Life Protection

Act and the Jamaica National Heritage Trust Act. Jamaica is also a party to international

agreements such as the Convention on Biological Diversity (CBD), and is therefore

19
obligated to conserve, sustainably use biodiversity, and equitably share the benefits

arising from the use of biodiversity (Todd, 2006).

2.5 Sustainable Financing for Protected Areas

One major challenge facing protected area management is the issue of the

financial sustainability of NGO co-managed parks and multi-use areas. The national

parks and protected areas policy recommends that Local Management NGOs engage in

collecting user and entrance fees as a cost recovery system to be used to finance the

operations of the parks but the legal framework within which this can be done has not

been completed (OAS, 2000; Todd, 2006). Jamaica faces many economic challenges and

these difficulties have meant necessary budgetary cuts by the central government. This

has therefore resulted in a reduction of the amount of money received by the park staff

through the NRCA/NEPA, and so parks such as MBMP and Negril Marine Park have

found it extremely difficult to effect their management plans. Assistance through a

USAID debt relief facility administered by the Environmental Foundation of Jamaica has

also played an important role in financing several of the management activities carried

out in the islands National Parks. However, this debt relief facility is not permanent has

a limited timeframe for completion.

The government is currently in the process of revising the Protected Areas System

Master Plan (PASMP), which is part of the National Protected Areas policy. As a party

to the CBD, Jamaica was obligated to complete this master plan in December 2006.

However, based on consultations between the relevant government agencies, it was

20
recommended that March 2007, would be a more realistic timeframe within which to

complete the plan (Todd, 2006). At the time of writing this dissertation, the plan has not

been completed.

The current activities for completing a system master plan are based on a number

of prior initiatives. For instance, the three components of the project preparation of a

National System Plan for Protected Areas in Jamaica will be integrated into the PASMP.

Based on existing gaps and recommendations, the following process was proposed to

complete Jamaica's protected areas master plan. The process included the formation of

three working groups: i) Ecological Working Group; ii) Capacity Development and

Management Effectiveness Working Group; and iii) the Financial Sustainability Working

Group. Each working group consists of representatives from government agencies, non-

governmental organizations and academia. From these working groups, four additional

components were identified to finalize the Master Plan. These were: Ecological Gap

Analysis, Management Effectiveness Assessment, Capacity Development Plan and

Sustainable Finance Plan.

With respect to the finance working group, their two primary deliverables are

(Todd, 2006):

To develop cost estimates for protected area creation and management needs
over 10 years (minimum, medium and ideal scenarios)

To identify existing funding sources and financial gaps, and the supporting
policy and legislative reforms where necessary

The final product from the working group should be an iterative and broadly

owned Sustainable Finance Plan, which will attract sufficient and sustainable financial

21
resources to effectively manage the protected areas system. To this end, the Protected

Areas System Master Plan is being developed in collaboration with a number of entities,

thus fulfilling Jamaicas international obligations under the Convention of Biological

Diversity (Todd, 2006).

2.6 The Jamaican Tourism Product

The Jamaican tourism industry accounts for 32% of total employment and 36% of

the countrys gross domestic product or GDP (WTTC, 2004). It is largely based on the

sun, sand and calm seas, the last two of these attributes being dependent on healthy coral

reefs. Jamaica is primarily a Sun, Sea and Sand destination and therefore the primary

recreational activities of visitors include sun and sea bathing on the beaches. The

recreational users of the coastal resources are primarily: beach goers, snorkelers, scuba

divers, and glass-bottom boaters (Bunce et al., 1999). The coastal tourism industry

continues to be a major growth area in Jamaica. For example, over the last three years,

the number of hotel rooms in Jamaica has been increasing at a rapid rate, with the room

stock primarily associated with large and more luxurious developments along the islands

North Coast. This is expected to continue and to increase at an average of 4,600 rooms

per year, bringing Jamaicas room stock to 75,000 by 2015 (Bartlett, 2008).

It is important to understand that Jamaicas tourism product over the last 15 to 20

years has largely been driven by the All Inclusive vacation concept (Issa and

Jayawardena, 2003; Boxhill, 2004). An all inclusive vacation can be defined as a total

service package where the visitor pays one fee that covers accommodation, meals, drinks

22
and recreational activities at a hotel or resort. Quite often the visitor is often not

encouraged to leave the resort and thus spends their entire vacation within the boundaries

of the property. One could say therefore that the tourism industry actually markets the

Resort as the destination instead of the total island. The typical tourist therefore spends

much of their time sunbathing and engaging in watersports related activities. The island

is primarily marketed as a couples or romantic destination and hotels typically offer

wedding, honeymoon and anniversary packages. There are however a few large hotel

chains that cater to family oriented vacationers.

The large resorts and hotels as well as coastal attractions such as dolphinaria and

water parks in general are responsible for maintaining the aesthetics of their beaches and

surroundings. However, the overall protection and management of the coastal resources

are not the tourism industrys primary concern.

2.7 Characteristics of Tourism Locations

The major coastal locations characterized by heavy tourism related activities are

the city of Montego Bay and the towns of Ocho Rios and Negril. These three locations

are the islands premier tourist destinations and are all located on the north and

northwestern coasts of the island of Jamaica (see Figure 1). Notably the coastal waters

surrounding these towns are designated as protected areas and are co-managed by NGOs

and the National Environment and Planning Agency (NEPA). These three coastal

tourism locations all have similar characteristics. They are all bounded by relatively

large coastal areas used by multiple groups, including fishers, divers, snorkelers,

23
swimmers, and other recreational users. These three locations are also impacted by a

wide range of anthropogenic issues, including sewage, solid waste disposal and

deforestation, as well as activities associated with their direct use, such as hotel

development, watersports activities and artisanal fishing.

Figure 1 Island of Jamaica, major towns and cities

Hotels and other attractions may sometimes be situated outside of the boundaries

of the marine protected areas but many of their watersports related activities such as

24
snorkelling and diving occur within the boundaries of the relevant parks and protected

areas.

For example, a study by Bunce et al. (1999) found that the watersports operations

in Montego Bay were typically dive operations, snorkel and glass bottom boats, party

cruisers, and small scale watercraft businesses (including parasailing, paddle boating,

jetskiing and waterskiing facilities). The study showed that the watersports operations

were responsible for taking over 3,100 tourists and making nearly 220 trips into the

marine park waters each week. Extrapolated over the year, approximately 163,000

tourists utilized the parks waters and over 11,000 watersports trips were taken into the

Park (Bunce et al., 1999). The study by Bunce et al. found that, of the 28 watersports

entities, 6 were dive shops taking approximately 675 guests diving per week, 12 were

snorkel operations and took approximately 860 guests snorkeling per week while the

small scale watercraft operators service catered to approximately 390 persons per week.

2.8 Impact of Tourism on Coastal Ecosystems

Tourism often brings large groups of people to destinations, and they, and the

supporting infrastructure they require, create a variety of threats for the environment

(Dixon et al., 2001). Firstly, tourists generate substantial amounts of solid waste. For

example, stopover tourists in the Caribbean have been estimated to generate twice as

much solid waste per capita as local residents (Dixon et al., 2001), while cruise ship

passengers are estimated to produce as much as four times as much garbage per day as

25
local populations. Tourists also generate substantial amounts of liquid waste, much of

which goes untreated.

Secondly, tourist facilities such as hotels and other attractions are often built in

environmentally sensitive areas. In the Caribbean, the majority of tourism facilities are

located within 800 meters of the high water mark, and most tourist activity takes place in

the area between the back reef and fore reef areas (Dixon et al., 2001). These facilities

can disrupt the sensitive ecological processes that often occur in these areas, especially if

they displace mangrove forests, salt ponds, or other ecologically sensitive areas (Bruner

et al., 1998). The construction process itself can generate problems such as high noise

levels and waste generation. The impact of construction can also often be felt far from

the resorts themselves, through the impact of mining for construction materials and the

access roads created for the removal of said materials.

Thirdly tourism facilities and activities place high demands on energy and

freshwater resources. The typical tourist uses much higher amounts of water and

electrical energy and this places an additional burden on existing energy and water supply

systems in the host country. Other impacts on the coastal ecosystems can be traced to the

recreational activities that visitors may be engaged in. For example the presence of high

numbers of snorkelers and divers can have negative impacts on coral reefs (Dixon, 1993;

Thur, 2003).

Coastal tourism infrastructure and tourist activities can have direct impacts on the

key resources that visitors come to the region to enjoy. These are in particular the sand

and the sea, these being two of the components of the three Ss.

26
 The sand resource (i.e. beautiful white sand beaches) is vulnerable to a number of

impacts. Firstly congestion can be an issue, here overcrowding creates a dis-amenity thus

reducing the enjoyment of the resource. Then there is the problem of pollution which can

occur from overcrowding as well as from non point sources of pollution. This includes

both liquid and solid waste and can cause loss of use of the beaches. Beaches are also

subject to erosion, which can reduce or significantly modify their physical extent. Beach

erosion occurs naturally but can be exacerbated by changes in currents and depositional

patterns due to construction related activities in the coastal zone. These include mining

of construction materials (sand mining), jetty and groyne construction and the removal of

natural sea defenses such as coral reefs, seagrass beds and mangrove forests (Edwards,

2002).

The sea resource is used in a variety of ways including swimming, boating,

diving, waterskiing, glass-bottom boating among others. The threats to the sea can arise

from a multitude of sources. Some of these sources can be traced directly to tourism

related activities while others are from external sources. Probably the most common

form of degradation is by pollution. Pollution can arise from a variety of sources.

Untreated sewage from hotels is often an important source, but this is far from being the

only source. Residential and industrial sources can also be major sources of sewage.

Runoff from agricultural areas can be contaminated with fertilizers, pesticides, and

sediment. Off-shore sources such as cruise ships and transport ships also contribute to

pollution. Water pollution, in turn, can cause a variety of problems, such as making

bathing unpleasant, as well as increased potential for serious health hazards. Congestion

27
of swimming areas because of excessive numbers of users as well as mixing incompatible

uses such as swimming and jet-ski operation (all within the same area) can also be

considered a threat to the sea resource (Dixon et al., 2001). Beach erosion and

changing existing water current patterns can also significantly alter swimming water

depth as well as increase the level of suspended solids in the water column. Increased

sedimentation can therefore result in cloudy waters making it unattractive for sea bathing

or snorkeling.

The direct anthropogenic threats on the coral reef ecosystem were discussed

previously and will not be repeated here. However, the close proximity of tourism

activities and infrastructure to the resource means that overuse and inappropriate use can

impair the very benefits that are currently derived from reefs and beaches.

Jamaicas economy is very dependent on the coastal tourism industry, which is in

turn dependent on a healthy coastal ecosystem. This healthy ecosystem is defined by the

key attributes which include white sandy beaches, clean and clear swimming waters,

healthy coral reefs and abundant and diverse fish and invertebrate species. This means

that a potential decline in environmental quality would negatively impact earnings from

tourism while an improvement in quality may increase the flows of well needed revenue.

The existing policy and institutional framework in Jamaica is essentially one

where there is inadequate funding for coastal management (high debt to GDP ratio, other

social pressures). This therefore suggests that alternative sources of revenue for the

purpose of natural resource management need to be identified. One possible source is

from the major users of the coastal resource, tourists who visit Jamaica.

28
 CHAPTER 3

NON-MARKET VALUATION AND RECREATIONAL DEMAND

3.1 Introduction
There are a variety of techniques used to elicit non-market values for

environmental amenities and can be classified in general terms as stated preference (SP)

or revealed preference (RP) techniques. Much of the recreational economic valuation

literature is dominated by RP methods, for example travel cost studies. These studies for,

example, use observable data such as travel and time costs to estimate demand curves

that allow for the determination of economic benefits of a particular location (Bockstael,

1995; Boyle, 2003; Parsons, 2003).

By contrast, SP techniques are useful in the absence of observable data for

hypothetical or real changes in quality of a particular environmental good (Adamowicz et

al., 2001; Mercado, 2001; Bateman et al., 2002; Smith, 2006). The valuation of a

resource with multiple attributes is probably best captured through the use of a survey

instrument with a contingent valuation (CV) and or choice experiment (CE) framework

(Mercado 2001). This technique has been utilized for a number of years and the method

has evolved and has become increasingly accepted as a valid method of environmental

29
valuation. SP techniques are also useful when considering non-use values associated

with a resource or public good (Bateman et. Al., 2002).

3.2 Contingent Valuation

Contingent valuation involves constructing a hypothetical market for the purpose

of eliciting peoples preferences for public goods. The market typically defines the good

of interest, the status quo level of provision and the offered improvement or decline

therein, the institutional structure under which the good is provided and payment vehicles

for said good. In the survey the respondent is asked to reveal their willingness to pay

(Mitchell and Carson, 1989; Mercado, 2001).

Contingent valuation has been utilized for a number of years and the first

recognized use of this kind of SP technique is credited to Robert Davis who investigated

the benefits of outdoor experiences to recreational users. This study was conducted in

1963. Since then the method has evolved and has become increasingly accepted as a

valid method of environmental valuation. Its acceptance as a valid method gained

credibility following its use to value environmental damage after the Exxon Valdez oil

spill in Alaska (Boyle, 2003). Following this event a number of books and articles were

published on this issue (Cummings et al., 1986; Mitchell and Carson 1989). In fact, the

National Oceanic and Atmospheric Administration conducted a blue ribbon panel in 1993

to review the method and came up with a set of recommendations which have become

almost an industry standard (NOAA, 1993; Bateman and Willis, 1999). The panel

received several opinions on studies for and against the use of CV in measuring non-use

30
values associated with oil spills. The panel concluded that carefully designed and

implemented CV studies convey useful information for judicial and administrative

decisions involving non-use and existence values (Loomis, 1999).

A proper CV study should at least have the following basic components (Kolstad,

2000, pg 297). Firstly, it requires that there is a carefully defined market scenario with a

well defined good. Critical to this is the selection of an appropriate payment vehicle that

should be directly tied to the good or service being valued (Boardman et al., 2001; Thur,

2003). Secondly, an appropriate method to elicit the respondents value must be selected.

This could be conducted using a variety of formats including open-ended questions,

payment cards, bidding games and referendum or voting questions. Studies have shown

that the dichotomous choice (DC) referendum question format to be very effective at

providing the data that can be used to generate estimates of welfare (Haab and

McConnell 2002; Bateman et al., 2002).

One criticism of SP techniques or the CV method is that hypothetical referenda

may not be incentive compatible. Incentive compatible institutional mechanisms should

provide individuals with incentives to truthfully and fully reveal their preferences

(Cummings et al., 1997). Therefore it is important that the valuation scenario be

carefully crafted to reduce the effects of hypothetical bias. There are methods that have

been used to mitigate this problem. These include using laboratory experiments to

conduct calibration of hypothetical bias (Cummings et al., 1997; Blackburn et al.,

1994; Fox et al., 1999). Another way of trying to reduce this bias is to use a cheap talk

design for the CV questions (Cummings and Taylor, 1999). This design involves

31
providing the respondent with an explicit discussion on what hypothetical bias is and why

it might occur. In the study by Cummings and Taylor they were able to use this design to

reduce bias in their experimental scenario. For this dissertation these options were

utilized in the design of the valuation scenarios for both the CB and CE surveys.

Thirdly a reasonable and effective method of administering the survey must be

selected. This may vary with the requirements of the researcher, the budget and time

limitations of the study. Typical methods of data collection include, in-person

interviews, telephone surveys, mail and internet surveys. These different methods have

different rates of success and often depend on the budget and personnel limitations of the

research team (Dillman, 2007).

The fourth component involves actually determining the sample population and

randomly selecting respondents to survey. Finally once the data are collected and

organized, the necessary statistical analyses are conducted including the estimation of

willingness to pay so that economic valuation can be reported.

3.2.1 Contingent behavior

A modification of the contingent valuation method can be used to derive estimates

that can explain how changes in resource quality can impact respondent behavior in

addition to estimating the economic value of the resource. This modification of the CV

method is termed contingent behavior (CB). The contingent behavior methodology

involves constructing a hypothetical market for the purpose of eliciting peoples

preferences for public goods. The goal of this approach is to estimate a demand function

32
for the good but in this instance conventional data on prices and quantities consumed

(RP) is supplemented by responses to a survey question in a discrete-response format.

The survey question focuses on demand behavior rather than willingness to pay per se

(Carson and Hanemann, 2005), however, the estimation of a demand function allows for

the calculation of consumer surplus.

CB is commonly used to assess quality or price changes at a recreational site. In

the CB framework, respondents are asked to make statements about their intended

behavior (e.g., visitation to a site) given a proposed change (e.g., in site quality, access, or

price). Whereas CV elicits a value statement, CB is used to estimate changes in behavior

or levels of use for a nonmarket good (Grijalva et. al., 2002).

3.3 Choice Experiments

Stated preference surveys can also utilize attribute based methods to collect

valuation information. These are often in the form of choice experiments (CE) that are

used to represent individual judgments of multi-attribute stimuli (Batsell and Louviere,

1991). For example the possible effects that an institutional program can have on various

"attributes" of a given environmental good can be determined through a CE framework.

Different levels of environmental management for example can generate different levels

of these attributes. By valuing the attributes of various programs, the benefits of these

programs can be estimated. This type of analysis is common for comparing market

goods in an effort to understand the tradeoffs that consumers are willing to make, with

respect to a products attributes. In market research this is called conjoint analysis and

33
has recently gained popularity for valuation of non-market goods because of its intuitive

applicability when comparing policy alternatives (Huang et al., 2007).

Its use in non-market valuation has increased over time and has been applied to a

variety of issues in the developed and developing world. This includes use of CE for

modeling the recreation demand for rock climbing in Scotland (Hanley et al., 2002),

assessing the role of public access in willingness to pay for coastal land conservation in

Rhode Island, USA (McGonagle and Swallow, 2005) as well as valuing renewable

energy investments and the impact on landscape quality, wildlife and air quality in

Scotland (Bergmann et al., 2006). There have been other CE studies that have looked at

marine related issues but these will be discussed in more detail in following sections.

3.4 Theoretical Rationale for CV and CE

The design of CV and CE experiments can follow a random utility model (RUM)

framework (McFadden, 1974) and thus can also be compared with RP data. Combining

revealed and stated techniques is another possible approach for obtaining such values.

Using both RP and SP data may increase the efficiency of estimation because it can

exploit the advantages of each data source while at the same time compensating for each

others weaknesses (Cameron, 1992; Carson et al., 1996; Adamowicz et al., 1994;

Adamowicz et al., 1997; Whitehead et al., 2005).

The general theoretical rationale for this study is the application of a valuation

method that is appropriate for providing information that has the potential to contribute to

34
policy development and implementation, management for protected areas and general

coastal resource management.

In this case, the policy scenario being valued is the prevention of a decline and/or

resultant improvement in the coastal ecosystem as a result of adequate environmental

management. The management and protection of coral reefs is expected to result in

improvements to marine life (biodiversity), swimming water clarity and beach quality.

These improvements are therefore likely to increase the utility of the individual user and

result in an increase in consumer surplus. The assumption therefore is a policy outcome

of better management, resulting in improved environmental quality that will likely lead to

an increased demand for coral reef recreational services, ultimately contributing

increased revenue to the tourism sector.

As stated in the previous section, a hypothetical choice experiment is offered to

respondents in which they face a trade-off between environmental goods, other goods and

services, and cost. If constructed carefully, the valuation scenario can reveal individuals

willingness to trade off environmental goods with other goods and services and provide

insight into their relative value.

3.5 Coral Reef Ecosystem Valuation Literature

In recent years the body of literature on coral reef ecosystem resource valuation

has increased. These studies have however primarily focused on scuba diving and are

therefore limited to a unique set of resource users. There are a small number of studies

35
that have attempted to estimate values for multi-use tropical marine areas that support a

wide range of recreational and non recreational uses of the natural resources.

3.5.1 Diving and marine protected areas

Table 1 identifies selected studies that valued access value and quality changes for

diving at various coral reef sites and is adapted from Parsons and Thur (2008). One of

the earliest studies by Dixon et al. (1993) was conducted prior to the implementation of

the Bonaire Marine Parks SCUBA diving fee. A survey of 79 divers was conducted to

elicit reactions to the proposed fee (Dixon et al., 1993; Dixon et al., 2000). Data was

collected using a convenience sample, in-person survey in order to test the feasibility of a

implementing a fee for raising revenue. They used a dichotomous choice question with a

$10 dive tag price followed by a payment card question with $20, $30, $50, and $100

options. They found that the average annual per person WTP for access was $27.40

(1991US$).

A similar study was conducted at the Tubbataha Reef National Marine Park in the

Philippines by Tongson and Dygico (2004). The authors used a convenience sample of

divers and conducted in-person interviews on several charter vessels. They used a

dichotomous choice elicitation question with fixed fees ranging from $25 to $75

(1999US$) per trip. Mean willingness to pay (WTP) was $41.11 (1999US$) per trip,

with the average trip including three days of SCUBA diving. In another study from the

Philippines Arin and Kramer (2002) used a nonrandom, convenience sample to assess

visitors WTP for access to three hypothetical MPAs in the Philippines. The survey was

36
conducted during the summer of 1997 and was administered to respondents planning to

SCUBA dive or snorkel. The studys estimates were based on data collected using a

payment card format question with $0, $1, $3, $5, and $10 options (1997US$). They

calculated estimates of mean daily WTP for access ranging from $3.40 to $5.50

(1998US$).

Lindsey and Holmes (2002) estimated WTP for entry to a proposed MPA in Nha

Trang Bay, Vietnam. A nonrandom convenience sample conducted in 1999 was used to

collect the survey data. The payment vehicle was a proposed entrance fee which allowed

respondents to visit an aquarium located on an island surrounded by the MPA. The

survey first used a dichotomous choice question to screen for respondents with any

positive WTP. The respondents were then presented with a payment card with bids

ranging from $0.07 to $5.00 (1999US$). Vietnamese citizens comprised the majority of

the respondents (89%), and had a mean WTP of $0.51 (1999US$) per day. Foreign

tourists reported a mean WTP of $1.48 (1999US$).

37
Table 1 Selected coral reef valuation studies

Author(s) Resource Year of Study Values per diver in year of study dollars
Dixon et al. (1993) Bonaire Marine 1991 $27.40 mean annual WTP for access to scuba dive
Park
Tongson and Dygico (2004) Philippines 1999 $41.11 mean WTP per live-a-board boat trip (avg. 3
Marine Park dive days) for access to scuba dive
Spash (2000) Jamaica Marine 1998 $25.89 mean annual donation for five years to trust
Park fund to operate marine park to improve
environmental quality from 40% to 100% of its
potential
Spash (2000) Hypothetical 1998 $25.21 mean annual donation for five years to trust
Curaao Marine fund to operate marine park to improve
Park environmental quality from 35% to 75% of its
potential
Lindsey and Holmes (2002) Proposed 1999 $0.51 - $1.48 mean WTP for daily access for any
Vietnam Marine activity
Park
Mathieu et al. (2003) Six Seychelles 1998 $5.20 - $14.40 mean WTP for daily access for any
Marine Parks activity (range is for different areas). $19.80 mean
WTP for daily access to scuba dive
Arin and Kramer (2002) Hypothetical 1997 $3.40 - $5.50 mean WTP for daily access to scuba
Philippines dive
Marine Parks
Wielgus, et al. (2003) Eilat Coral 2001-2 $1-$3 mean WTP per dive for moderate
Beach Nature improvements in quality
Reserve in Israel
*Adapted from Parsons and Thur, 2008

38
 Mathieu et al. (2003) interviewed 300 tourists in the Seychelles across six

designated MPAs in June 1998. MPA visitors and general tourists alike were both asked

if they thought that it was acceptable to be asked to pay a fee to enter Marine Parks.

They were then offered a payment card with bids ranging from $0 to $40 (1998US$).

This study utilized a daily entrance fee for the payment vehicle. The total sample mean

WTP was estimated to be $12.20 (1998US$). Mean WTP was highly dependent on

which, if any, of the six MPAs the respondent had visited. Mean WTP for the five

functioning marine parks varied from $5.20 to $14.40. SCUBA divers were shown to

have an average WTP of $19.80, which was significantly above the whole sample mean.

Wielgus et al. (2003) estimated the economic value of coral reef damage at Eilat

Coral Beach Nature Reserve in Israel. This study employed the use of videos in

conjunction with a stated preference choice model and to convey quality changes. The

authors estimated values for changes in a biological index as well as for changes in

quality of water, coral cover, and diversity of species. Wielgus et al. conducted a

convenience sample of divers at dive shops located at the nature reserve. Moderate

changes in quality were valued in the range of $1 to $3 per dive per person.

Parsons and Thur (2008) estimated the economic value of changes in the quality

of a coral reef ecosystem to SCUBA divers in Bonaire using a stated choice mail survey.

From a survey of 211 respondents they used mixed logit to analyze the data. They

estimated annual per person losses at $45 for modest declines in quality while for larger

losses the value was $192 per person.

39
3.6 Jamaican Non-Market Valuation Studies
In addition to the study cited in Table 1 there are a few other non-market

valuation studies that have been conducted in Jamaica. A study by Wright (1995)

attempted to estimate non-market values of coral reefs. He used both CV and zonal

travel cost methods to estimate welfare for coral reefs in Negril. The travel cost estimate

of welfare was US$121. Using an open-ended CV method Wright also estimated WTP

values ranging from US$31-49 to pay to preserve the coral reefs in Negril. The payment

vehicle was a one-time contribution to a reef improvement fund.

The largest study using a stated preference method in Jamaica was a World Bank

funded project that looked at the values associated with the coral reef ecosystem in

Montego Bay as well as a companion study in Curacao, Netherlands Antilles. The main

focus of the study was to provide a valuation of biodiversity with respect to the potential

for bioprospecting of coral reef resources (Ruitenbeek and Cartier, 1999). A number of

articles have been published on various aspects of this study including the paper by Spash

(2000) shown in Table 1. The World Bank study primarily examined the valuation of

Montego Bays coral reefs as a biotechnology resource (Gustavson, 1998; Bunce et al.,

1999). As part of the larger study, a semi random intercept survey of visitors and

residents was conducted, welfare values were assessed for the reefs in the Montego Bay

Marine Park and a hypothetical marine park in Curacao. Street intercept, in-person

interviews of locals and tourists were conducted using open-ended elicitation questions.

In Jamaica, respondents were asked for an annual contribution over a period of

five years to a trust fund for the existing Montego Bay Marine Park. Respondents were

40
informed that the marine environment was at 75% of its quality potential. They were also

informed that if they contributed to the fund, managers would be able to fund initiatives

that would raise quality to 100%, while a decline to 60% of quality potential would occur

without the trust fund. Using OLS techniques mean WTP was estimated at $25.89 per

year, and there was no statistically significant difference between locals and tourists

WTP. A comparative study was conducted in Curacao, using a similar payment vehicle

(trust fund) and quality scenarios. Mean WTP was estimated to be $25.21 annually,

again with no significant differences between locals and tourists.

A study by Dharmaratne et al., (2000) looked at tourism potentials for financing

protected areas. They estimated use and non use values for the Montego Bay Marine

Park and a proposed Barbados National Park. A payment card format was used and the

payment vehicle was an entrance fee to enter the respective parks. For Jamaica, they

estimated WTP for access to the resources of the Marine Park at $20 for first time users

and $10 for repeat users. Non use value was estimated at $1.45. For Barbados they

estimated WTP for access to the park at $109 for first timers and $66 for repeat users.

A study in 2001 by the Environmental Management Unit (EMU) of the University

of the West Indies used a zonal travel cost model to estimate consumer surplus for the

Ocho Rios Marine Park of US$132 per person who stayed for an average of 10.5 nights

(EMU, 2001).

The studies described above were either site specific (marine parks) and/or

activity specific (scuba diving). It should also be noted that SCUBA diving is not one of

the primary recreational activities associated with a vacation in Jamaica. This study is

41
therefore different because firstly, it focuses on estimating recreational values from a

wide range of resource users which include but are not limited to beach goers, snorkelers,

glass bottom boaters as well as SCUBA divers. Secondly the recreational activities

associated with the coastal ecosystem are situated across all the tourism locations in

Jamaica including the three major tourist towns. The estimates of value are therefore not

site-specific per se but are for the quality of the coral reef ecosystem that is associated

with the coastal tourism industry. Therefore, the estimates of value presented here are

relevant to the tourism product that is the entire western and northern coast of the island

of Jamaica.

42
 CHAPTER 4

CONTINGENT BEHAVIOUR STUDY

Feasibility of Sustainable Funding from Tourist User Fees

4.1 Introduction
This study focuses on the feasibility of obtaining sustainable financing (from

tourist user fees) for the preservation of the ecosystem and recreational services provided

by the coral reefs and beaches. The rationale being, the near shore coastal ecosystem is

largely responsible for the existence of coastal tourism, an important economic activity

for Jamaica. As a result data were gathered from recreational users (tourists) in order to

provide policy relevant information. In particular, the information was collected from

stopover tourist visitors to the island.

4.2 Contingent Behavior, Welfare Estimation and Sustainable Revenues

The primary objective of the contingent behavior survey is to provide the policy

relevant information that can guide the development of sustainable financing mechanisms

for natural resource protection. This will be achieved through a survey that poses

different levels of user fees (taxes) to tourists visiting the island. Each respondent is

asked if he or she would have made a trip to Jamaica at different fee levels. A split

43
sample method was used to assess differences in respondents WTP for an environmental

tax versus a general tourism development tax. Using data generated from the

environmental tax version of the survey, I estimate the responsiveness of households to

these different fee levels. This behavior is then used to predict the potential revenue that

may be raised from the implementation of a per-person (environmental) user fee as well

as the concomitant decline in tourist visitation at the different fee levels.

The CB approach has been applied in other studies (Ward, 1987; Loomis, 1993;

Chase et al., 1998; Dharmaratne et al., 2000). These studies used stated preference

approaches with a random utility model or direct estimation of visitation demand changes

in response to entrance fees. The most common use of the CB method is to complement

recreation demand modeling where CB data is combined with RP trip data (Englin and

Cameron, 1996). In the combined RP-SP recreation demand framework, individuals are

asked to provide information on actual trips taken to a site under existing resource

conditions or management rules (i.e., RP data) and subsequently are asked to indicate the

number of trips they would take to the site under alternative, hypothetical management

rules. However this was not the case for Englin and Camerons paper or the other studies

cited above.

There have been other Jamaican non-market valuation studies that have used

stated preference techniques to provide estimates of welfare that are then used to support

arguments for the implementation of tourist user fees (Wright, 1995; Reid-Grant and

Bhat, 2009). However these studies do not present models of visitation demand in

response to changes in fee price. This dissertation therefore attempts to contribute to the

44
body of knowledge by presenting a CB demand model that can be used to predict

changes in visitor demand in response to increases in tourism surcharges.

There have been studies that utilized open ended and payment card CV formats to

provide information that is then used to estimate revenue maximizing fees or model

recreational demand in response to increased prices. A study by Chase et al., (1998) used

a CB approach to assess the effects of differential pricing of entrance fees to national

parks in Costa Rica. They used a payment card method to collect WTP data and used

estimated demand elasticities to calculate entrance fee structures that would maximize

park revenues. They also presented estimates of corresponding visitation rate changes in

response to changes in entrance fees. The other study by Dahrmaratne et al. (2000) used

an open-ended CV method to collect data on WTP to access a marine park in Jamaica and

a proposed marine park in Barbados. They used the payment card responses to create a

referendum data set and then estimated a logit model which was used to estimate the

percentage of respondents who were willing to pay different prices (user fees) for

admission (for a seven day visit) to the Montego Bay Marine Park via a particular beach

(Doctors Cave beach).

4.3 Study Population

The sampling frame for this study is visitors who have spent one or more nights

vacationing in the island. According to the Jamaica Tourist Boards annual report (Jan

Dec 2007), Jamaica received 1,700,785 stopover visitors (persons spending one or more

nights), this representing an increase of 1.3 per cent over the previous year. Of the

45
stopover visitors, 1,573,267 were foreign nationals and 127,518 were non-resident

Jamaicans. The majority of all visitors to Jamaica were from North America (USA and

Canada) and they comprised 77.8% of total visitors to the island in 2007. Stopover

visitors typically spend 5-10 nights in the island and therefore are able to experience a

wide range of coastal recreational activities. More recently released statistics for 2008

show similar trends (JTB, December 2008). That is, Jamaica received 1,767,271 visitors

of whom 1,623,675 where foreign nationals and 143, 596 were non-resident Jamaicans.

The rationale for using this sampling frame is that these visitors are more likely to

have spent some portion of their vacation time at the beach and may have engaged in

activities such as swimming, sun bathing, snorkeling, glass bottom boating or scuba

diving.

4.4 Survey Development and Sampling

The design of the survey instrument began with pre-tests of the questionnaire

design with colleagues and twelve visitors to Jamaica. Input from this exercise and

interviews with some of the visitors paved the way for a pilot of the survey instrument.

In order to feasibly sample the wide range of visitors to Jamaica who may have

stayed in various locations, I decided to use the international airport in Montego Bay.

The Montego Bay International Airport is one of Jamaicas two international airports and

operates as a hub for major airlines entering and transiting the island. Of the stopover

visitors to Jamaica, the majority enters and leaves the island by air via this Airport. The

choice of using the departure terminals in the airport was based on two key factors.

46
Firstly, the airport was the most feasible location that provided access to the sample

population. Visitors to the island vary in their characteristics and the locations that they

spend their vacation. Some visitors may have stayed in super all-inclusive hotels while

others may have stayed in lower cost inns or alternatively, high end private villas. There

are also visitors such as student groups and service club members who spend a part of

their vacation at the beach enjoying the coastal attributes and another part engaged in

some social activity for example painting elementary schools or building churches.

Finally there are also visitors who may have stayed with friends and relatives while they

were on holidays. Sampling these disparate types of visitors randomly at their given

locations would not have been feasible. Additionally my attempts to distribute survey

packets (mail return) randomly in various hotel rooms (in various tourist towns) proved

unsuccessful as the hoteliers were not willing to allow access, many citing concerns about

protecting the privacy of their guests. The second factor of using the departure areas

rather than arrivals is linked to the rationale for using this sampling frame discussed in

the previous section. That is respondents would have spent some portion of their

vacation time at the beach and may have engaged in coastal recreational activities.

As a result of these issues a simple random intercept method was used to collect

data in the departure terminals of the Montego Bay International Airport. As stated

above the majority of coastal tourists travel through this airport making it feasible to

intercept different types of tourists who stayed in different locations. The sampling

approach in this study differs significantly from the previous CV studies conducted in

Jamaica. For those studies the methods of data collection involved either street intercepts

47
in Montego Bay (Gustavson 1998, Spash et. al., 2000; Dahrmaratne et. al., 2000) or

intercepts at dive shops in Negril (Wright, 1995).

In order to gain entry to the departure concourses, special permission was granted

by the Montego Bay Jamaica Airport management company (MBJ Limited) only after

adequate security clearance for the survey personnel was obtained. Based on security and

customer service considerations a narrow window of opportunity for the pilot test and

final survey effort was provided by MBJ Limited. Sampling was only allowed for two

brief periods as the MBJ officials were concerned about the potential discomfort of the

clients travelling through their terminals may have faced as a result of the in person

sampling effort.

A pilot survey was conducted in August 2007 and based on the results from this

pilot exercise, adjustments were made to the survey instrument such as the bid design,

refining the payment vehicle, refining the cheap talk script etc. At this stage a decision

was also made to utilize a split sample design where half of the respondents would

receive an environmental tax version while the other half would receive a general tourism

tax version. The development of a stated choice version of the survey was based in part

on the results of this pilot study. A similar intercept protocol was used for the main

survey exercise which was conducted over 6 days in January 2008.

Surveys were distributed to tourists who were randomly intercepted while they

waited on their departing flights. Respondents were selected from individuals who were

seated in departure concourses. Every nth person seated in a given row was intercepted

and asked if they were willing to participate in the study. For example if 15 persons were

48
seated in a row, a surveyor might select the fourth person from the end of the row and

then the fourth person further down that given row (and so on). Respondents were first

informed of the study and asked if they would be willing to complete a self administered

survey. The respondent was left to answer the survey on their own (no influence from

the surveyor) and upon completion they placed the survey booklet in a sealed envelope

that was collected by the surveyor. Sampling of visitors took place over ten hours each

day and surveyors were deployed across the nineteen gates in the airport terminal.

Sampling began on a Friday afternoon and continued daily through to the following

Friday afternoon.

4.5 Survey Design

The contingent behavior questionnaire was structured as follows. Respondents

were first asked questions related to their trip characteristics, environmental awareness,

their rating of beach quality among others. They were then provided with information

before proceeding to the contingent market scenario. For the environmental tax version

of the survey (which appears in Appendix 4) the respondent was provided with

information on the benefits of coral reef ecosystems services such as: coastal protection,

habitat and breeding areas for marine life and production of white sand for beaches

(Figure 2). They were informed that the presence of the coastal ecosystem provided

food, jobs (for locals), and recreational activities (locals and tourists).

49
 Coral reefs, seagrass beds and beaches provide a number of benefits and environmental
services such as:
Coastal protection
Habitat and breeding areas for fish and other marine life
White sandy beaches
Food and jobs
Recreational activities (swimming, scuba diving etc).

Preserving the health of the coral reef ecosystem and beaches in Jamaica requires active
management to address threats such as:
Overfishing
Damage from: boats, divers and snorkelers,
Water pollution from; sewage, garbage and stormwater
Beach erosion
Overgrowth of coral reefs with algae (sea plants)

Environmental management is the responsibility of local Marine Park Management

Organisations along with the National Environment and Planning Agency. These
organisations require funding to sustain the following activities:
Marine Patrols - To: monitor anchor damage, beach erosion and coastal water
quality; work with fishermen to reduce overfishing; monitor diving and
snorkeling behaviour
Public Education - with the public, schools and businesses; beach clean-up
programs
Joint Programs with other agencies - TO: reduce deforestation; address river
pollution; improve agricultural practices; improve garbage and sewage
collection.

Figure 2 Environmental information presented in the coral reef survey.

The respondent also received information about threats such as; overfishing,

snorkel and diver damage, sewage, garbage pollution, beach erosion and overgrowth of

coral reefs with algae. The respondents were told what agencies were responsible for

environmental management (NEPA & NGOs). Additionally they were informed that

these agencies require funding to sustain management activities such as; marine patrols,

fisheries management, public education programs. They were also informed that some of

50
the revenues would be used to assist in funding cooperative programs with other agencies

to address issues such as deforestation; river pollution; sustainable agricultural and

improved solidwaste management.

By contrast, no information on ecosystem services, threats and the required

environmental management activities for protection was provided to respondents in the

tourism tax version of the survey (titled Tourism and Beach Recreational Survey). The

only information that these respondents received is shown below. This information was

provided in both versions of the survey (see Appendix 4);

The Jamaican government currently charges a tourism surcharge (tax) of US$10 per person to all

overnight visitors to Jamaica. This surcharge is typically included in your overall travel costs

(airfare).

The proceeds from this tourism tax go towards a general tourism development fund and are to be

used to support management of the local tourist municipalities with activities such as;

General beautification

Human resource training.

The purpose of the tourism development fund is to help maintain the tourism industry and assist in

Jamaicas social and economic development.

Figure 3 Example of the policy information presented in both versions of the

survey

Following this the respondent proceeded to the contingent market scenario. The

payment scenarios varied between survey versions as discussed previously and are shown

below. It should be noted that the labels of the taxes differed in the two versions of the

51
survey, i.e. tourism tax versus environmental tax. All respondents were faced with a

single bounded dichotomous choice question that asked them to indicate whether or not

they would be willing to pay a per-person surcharge (tax) or choose not to visit Jamaica.

The contingent behavior scenario for the tourism tax version is shown below:

Suppose that prior to your most recent trip to Jamaica, the Jamaican
government decided to increase this tourism tax. This new tax would
result in an increase in your overall travel costs. The extra revenue from
this tax would go to the Government of Jamaica to be used to support
necessary government programs.

Q If, because of the increased tax, you now had to pay a per person
surcharge of US$100 (in other words an additional $90 on top of the
existing $10) as part of your overall travel expenses, would you still have
decided to visit Jamaica?

This can be compared with the contingent behavior scenario for the

environmental tax version shown below;

Suppose that prior to your most recent trip to Jamaica, the Jamaican
government decided to add an environmental tax to the existing US$10
surcharge, as part of its efforts to provide funding for the management of
the coastal environment. These funds would go directly to the relevant
environmental management agencies for activities such as; marine patrols,
public education and joint environmental programs and therefore preserve
the existing conditions and prevent a decline in environmental quality.

Q. If, because of this mandatory environmental tax, you now had to

pay a per person surcharge of US$100 (in other words an additional $90
on top of the existing $10) as part of your overall travel expenses, would
you still have decided to visit Jamaica?

Five different bid amounts for the per person tax were evenly distributed across

both versions of the survey, ($10, $50, $90, $490 and $990). Standard demographic data

were also collected from all individuals. Debriefing questions were also asked based on

52
their answer to the DC question. Respondents were asked to make written responses to

questions asking for their reasons for yes or no to the proposed tax presented in the DC

question. The open ended comments and answers to the debriefing questions were used

to identify protest responses.

4.6 The Impact of Different Payment Vehicles on Consumer Surplus

This study aims to estimate the value of the recreational services provided by the

coral reefs and beaches to the typical tourist that visits Jamaica. In order to elicit and

construct valid measures of willingness to pay for preventing the decline or preserving

the existing environmental conditions using the CV method, contingent market

participants must be familiar with the quality of the resource (Blomquist and Whitehead,

1998). Tests for construct validity of WTP are of two types (Mitchell and Carson, 1989).

Convergent validity tests are used for testing correlation in WTP measured with CV

methods compared with implicit market (RP) methods (Loomis et al., 1991). Convergent

validity tests may also look at correlations in WTP measured with different types of SP

methods (CV versus CE) as well as different payment vehicles for example comparing

taxes versus donations (Bateman et al., 1995).

Theoretical validity tests determine whether WTP is related to measures such as

environmental resource quality which theoretically affect WTP. Information presented in

contingent markets can be used to test for theoretical validity because information that

changes survey respondents true WTP should change their stated WTP (Randall et al.

1983; Randall 1986). Hoehn and Randall (1987) develop a theory which shows that

53
information presented in contingent markets enhances information about policy effects,

reduces respondent uncertainty and the divergence between formulated and true WTP.

Furthermore, additional information presented in contingent markets if true, generates

desirable information effects (Bergstrom et al., 1989; Blomquist and Whitehead, 1998).

Some contingent market participants will have little prior information about

ecosystem services and quality and would therefore be expected to formulate WTP based

on less (or one could argue inadequate) information. Additionally contingent market

participants can be presented with distinct incentive compatible institutional mechanisms

associated with the payment vehicle. For example the payment vehicle may remain the

same (tourism tax) but the institutional framework for using the funds may be different

for different groups of contingent market participants. This study will examine the

combined effects of providing ecosystem service information as well as offering distinct

institutional contexts on WTP in the context of a convergent validity test. The current

study design does not allow for the test of theoretical validity but the impact of

information provision will be considered in the discussion of the results.

This study considers the findings of Blomquist and Whiteheads (1998) paper that

looked at the role of providing resource quality information (to the respondent) and its

impact on generating valid willingness to pay estimates. In their paper they show that

true willingness to pay is the difference between consumer expenditures with a decline in

resource quality and expenditures without a decline in quality:

(1) ", ,

54
where WTP is the ex ante willingness to pay (value of avoiding a decline in quality), e ()

is the planned expenditure function and " is resource quality after a decline, is

preserved quality and u is expected utility.

Blomquist and Whiteheads model of information effects on willingness to pay

utilize an adaptation of prospective reference theory and examine the divergence between

perceived and objective resource quality. Willingness to pay statements made in

contingent markets depend on the resource quality perceived by the respondent

(Blomquist and Whitehead, 1998). Perceived quality (q) in my context includes the

perceived quality, ecosystem services and threats to the resource. It depends on the

objective quality , along with information about ecosystem services and institutional

mechanisms to preserve quality, I. This relationship can be assumed linear and can be

written as

(2) ,

where > 0 and are learning parameters, for prior information and for information

contained in contingent markets. The assumption that is positive implies that perceived

quality and objective quality are positively correlated.

If the respondents perception of ecosystem benefits and threats is less than

objective ecosystem benefits and threats, information about ecosystem benefits and

threats will increase perceived ecosystem benefits and threats and increase stated WTP

toward the true WTP. If perceived ecosystem benefits are greater than objective benefits,

information about ecosystem services and benefits will decrease perceived benefits, and

55
decrease stated WTP toward the true WTP. In either case, the information effect is

desirable in that WTP with perceived quality is closer to WTP with objective quality with

additional information provided in the contingent market. It is with this in mind that the

results from the two types of surveys should be examined.

4.6.1 Consumer surplus for what?

It is important to note that there is a distinct difference between the two welfare

measures. The tourism tax scenario does not force the respondent to make a trade-off

between income and environmental quality. There is no explicit resource change so in

effect respondents are being asked their willingness to pay for maintaining the existing

tourism product. This may include, but is not limited to environmental quality. Thus the

welfare measure for the tourism tax must be understood within this context.

However the environmental tax scenario forces the trade-off between higher user

fees (travel costs) and preserving the existing environmental quality or in other words

preventing a decline in environmental quality. In this instance the welfare measure is

comparable to the more traditional welfare measure that is based on monetizing the utility

difference that individuals experience due to a change in environmental quality

It should be noted that differences in welfare will also depend on what people

(respondents) perceive will happen as a result of them paying this hypothetical tax. For

example, the level of decline was not explicitly defined so the respondents so their

perception of what might happen in the absence of management will vary. . As a result

56
of the potential differences in perception, the respondent may have a different perception

of how their tourism tax dollars will be spent as opposed to the environmental tax.

4.6.2 Validity tests

For this study, I examine the validity tests of information on ecosystem services

(and institutional mechanisms) for preventing a decline in quality by considering the

trade-off between paying a tax, or choosing not visiting Jamaica. This is done by

considering two scenarios.

In one case, the respondent is provided with the relevant (objective) information

as described in the previous section. Thus, the respondent faces a contingent market

(choice of a vacation to Jamaica) where they are being asked to state their WTP for the

prevention of a decline in coastal ecosystem quality, given specific institutional

framework (WTPENV). In my example the institutional framework are environmental

management agencies (NGOs and NEPA). The payment vehicle is an additional tax that

results in an increase to their current travel expenses.

In the second case, the respondent is provided no information about ecosystem

services and institutional mechanisms that are required to preserve the status quo. The

respondent faces the same contingent market as above (vacation in Jamaica) however

they are being asked to state their WTP to maintain the status quo or tourism product,

that is, sun, sea and sand (WTPTOUR). They are also presented with a slightly different

institutional framework. In the second case the institutional framework is a generic

tourism development program (central government). It should be noted that this

57
institutional framework currently exists in Jamaica. For the second case the payment

vehicle is an increased tax (surcharge) that also increases their travel expenses.

I assume a priori that the respondents perception of the ecosystem benefits and

threats is less than the objective information provided. Therefore the provision of

information is expected increase WTP. There may be other reasons for such a divergence

and Table 2 below highlights some of these possible reasons.

Table 2 Reasons for divergence between WTPTOUR and WTPENV.

Reasons WTPTOUR why might be less than WTPENV

1. No information about ecosystem benefits, threats and mechanisms for preservation

2. Institutional mechanisms. Lack of trust that the funds will be used appropriately
3. Respondents have a different perception of how tourism funds will be used (vague vs
concrete perceptions).

For example it could be expected that WTPTOUR will be lower than WTPENV

because of a lack of belief that the institutional mechanism will be used appropriately.

For example if there is no clear description of how tax revenues will be spent (for

example environmental protection, poverty alleviation, solidwaste management) then this

will result in lower WTP. The non-provision of information could also mean that

respondents may have a different perception of how tourism funds will be used. Given

the lack of concrete information regarding the use of the funds for specific management

activities, it is likely that the respondent will tend to understate their WTP. Based on

these assumptions the general hypothesis of this study is H1: WTPTOUR < WTPENV, and

thus the null hypothesis is that WTP in each scenario is equal.

58
4.7 Theory and Econometric Models
The CV (or CB) data were analyzed using a random utility model. This is

essentially an econometric analysis of the binary choice data from the valuation questions

on the respondents decision to make a trip to Jamaica based on the imposition of a user

fee. This would be a YES or NO answer question that could also be followed up with an

open ended question if the respondent said NO to the bid price (Appendix 4). Please note

for this study, only the dichotomous choice response was used in the analysis.

Random utility theory, in this context, models an individuals choice of

recreational options in this instance to take a trip or not. Implicit in these choices are the

relative values of the environmental characteristics of the destination.

In my application an individual i faces J alternatives where j = 1, 2. Each

alternative gives an individual some utility defined as by:

(3) Uij = zij + j

where the term zij is a vector which represents a combination of the individuals

characteristics and some quality measure or ecosystem attribute (beaches, reefs etc) of

Jamaica, is a vector of parameters to be estimated and ij, is a random component of

preferences known to the respondent but unknown to the researcher. The determinants of

utility are yj, the jth respondents income and Aj are the environmental tax presented in the

DC question, while zij, and ij are as described above. The indirect utility is therefore

deterministic to individuals but random to the researcher.

59
 The choice situation based on the model cited above can be explained by the

equations below.

(4) U1 = zj + (yj-Aj) + 1

U0 = zj + yj + 0

where U1 is the utility derived from choosing to pay the environmental tax and visit

Jamaica and U0 is the utility derived from not paying the tax and staying home. Note that

paying the tax will result in the quality being preserved (q1) while not paying the tax will

result in a reduction in quality (q0). is the marginal utility of income and is the

marginal utility of individual characteristics and environmental quality. The utility

difference (U1 U0) can therefore be used as a proxy for estimating the WTP (Haab and

McConnell 2002).

Given the model outlined above each respondent will provide a yes response to

the DC question if and only if the perceived utility derived from paying the tax to visit

Jamaica exceeds the utility of not paying the fee and staying at home. Take for example

the case of paying the tax; to derive an expression for the probability that the jth

respondent answers yes to the DC question you first have to separate the utility function

into additively separable deterministic and stochastic components. This is shown below

as,

(5) Pr (yes j) = Pr[v1(yj Aj, zj,) + 1 > (v1(yj, zj,) + 0j)]

60
 If you assume that the error terms (ij) are distributed type I extreme values then

equation (5) takes the forms,

(6) Pr

where 0 is an estimated intercept is an estimated parameter on the monetizing variable

and zj is a vector of all other relevant and observed determinants (Hanemann and

Kanninen, 1999; Haab and McConnell, 2002). Equation 6 describes the linear logistic

form which can be used to estimate mean WTP in the sample as well as to inform the

effects of various characteristics on the probability of providing a yes (affirmative)

response to the DC question (Freeman, 2003; Haab and McConnell 2002; Neter et al.,

1996).

The linear model described above has been widely used in CV because of its

simplicity (Hanemann and Kanninen, 1999). It is also readily estimated with standard

econometric software to produce parameter estimates that can be used in welfare

estimation and predicting behavior.

4.7.1 Choosing the probability response model - empirical assumptions

It should be noted that the valuation of a change from q0 to q1 is based on a

number of empirical assumptions. The first assumption is that a persons maximum

willingness to pay for an item is bounded by income (y). In order to satisfy this utility

theoretic restriction the WTP function has to be truncated at y max (see figures 4 and 5).

The second assumption in this analysis is that an increase in q (or preventing a decline in

61
q) is viewed as an improvement. This however is an empirical assumption about how

people feel about the change in quality (Hanemann and Kanninen, 1999). Generally

speaking we could consider three cases: (1) the individual views the change as an

improvement; (2) the individual views the change as an improvement or is indifferent to

it (that is they place no value on it); (3) the individual could view the change as a good

thing or as a bad thing(negative preferences), or be indifferent.

Based on the objectives of this study the application here is that the individual

views the change as an improvement. This assumption therefore requires the use of a

probabilistic model for the purposes of predicting behavior, that is, the individual views

maintaining the status quo as an improvement over the decline in quality. The requires

using a canonical response model (Hanemann and Kanninen, 1999) which is based on the

assumption that a respondent views the quality change as an improvement (increase in q)

with a probability of 1. Therefore in terms of a probability response formula can be

described as

(7) Pr (response is yes) = 1, when A = 0

In other words, given an improvement in quality and no requirement to pay an

environmental fee then all (100%) respondents would still choose to take a trip to

Jamaica. This implies in terms of the RUM indirect utility function, that this is

equivalent to requiring that U1 stochastically dominates U0 (see figure 4 below).

62
 Pr{Yes}

100%

emax

0
A=Y

Figure 4 Censored response model (adapted from Hanemann and Kaninnen,

1999)

The logistic response model described above in equation 6 can be modified to

satisfy this stochastic dominance thus making it appropriate for contingent behavior

econometric modeling. This requires the distribution to be censored via the insertion of a

spike at A = y truncated at emax on the horizontal axis such that the bid A = y. The

canonical probability response model is shown in equation 8 below and figure 4 above

(Hanemann and Kanninen, 1999).

(8) Pr

63
 It should be noted that inserting a spike at emax is a simple adjustment that can be

performed after the original response model has been estimated (Hanneman and

Kanninen, 1999). This means that a re-estimation of the modified response probability

model is not required (as long as the As never exceed respondents incomes). Another

approach for generating a canonical response function is to truncate the distribution at

emax by shrinking the response probability graph to the horizontal axis such that the bid A

= y (as per figure 5 below). However truncation of the distribution using this method

would require re-estimation of the modified response probability model.

Pr{Yes}

100%

emax

0 Y

Figure 5 Truncated canonical response model (by shrinking)

64
 Now consider the other two cases cited earlier, that is the individual views the

change as an improvement or is indifferent to it and the individual views the change as a

good or bad thing, or indifferent. These empirical assumptions now require that the

probabilistic models be different from the canonical case above. The probabilistic

response curve for the second case can be interpreted as representing a population with

two types of individuals; a group of people amounting to 100 % of the population who

are simply indifferent for whom Pr (response is yes) = 1, when A = 0; and another

group which has a varying but positive WTP for the change in q (see figure 6a below

adapted from Hanemann and Kanninen, 1999, pg 320).

The probabilistic response curve for the third case incorporates the assumption

that some individuals will dislike the change, therefore -A as defined in equation 5

represents minimum willingness to accept to suffer the change, rather than maximum

WTP to avoid it (see figure 6b below).

65
 Pr{Yes

100%

0 Y A

Pr{Yes

100%

0 Y A

Figure 6 Response probability models with indifference and negative

preferences

66
 Another way of generating response probability distribution that satisfies the

empirical assumption of non negative preferences is to use a nonlinear model. Again this

would require inserting a spike at emax (ie censoring the distribution). The probability

response formula can be represented as:

(9) Pr (response is yes) = 1 Ge (- + ln A)

where Ge is the probability distribution. The response distribution can either take

the standard normal (probit) or logistic form, and produces a distribution similar to figure

4. The non-linear form can however have implications for estimation of welfare and will

be discussed briefly in the following section.

It is therefore important to modify statistical models for CV responses to

incorporate restrictions imposed by economic theory and the utility theoretic assumptions

made by the researcher (Hanemann and Kanninen, 1999). For the policy purposes

outlined in the beginning of the chapter, the linear canonical response model is the most

appropriate in this case. This response model facilitates the use of the parametric

estimates in predicting the probability of yes (or no) responses to increases in the

hypothetical tourism surcharge. The response model therefore facilitates use of the

predictions as part of the decision process for determining the feasibility of user fee

implementation.

67
4.8 Parametric Valuation of Contingent Behavior
When survey data come from questioning different individuals in a population the

summary measure of the WTP distributions would be multiplied by the number of people

in the population to produce an estimate of aggregate value. It is therefore important that

the choice of the summary statistic implies a particular approach to the aggregation of

welfare across the population. Using mean WTP (C+) is equivalent to adopting the

Kaldor-Hicks potential compensation principle but can lead to logical inconsistencies

particularly if you allow for negative preferences for the change in quality (Hanemann

and Kanninen, 1999). The use of the median WTP (C*) is equivalent to applying the

principle of majority voting that is the change is desirable if a majority of the population

would vote for it. While the use of median WTP does not satisfy Pareto efficiency, it

may still be considered to be ethically preferred. Choosing a measure of central tendency

is often a value judgment and different circumstances may require using different welfare

measures (Hanemann and Kanninen, 1999).

Another consideration for deciding which measure of central tendency to use is

that C+ is more sensitive than C* to skewness in the WTP distribution. This is

demonstrated graphically in figure 7 below where WTP corresponds to the shaded area

under the response probability function. The median WTP can be read directly from the

empirical response probability function and is essentially the dollar amount that

corresponds to a 50% probability of saying yes. C+ on the other hand is the result

between the relation of the mean of a random variable and the integral of its cumulative

68
distribution function, or in other words the area under the empirical response probability

function.

Figure 7 Graphical representations of WTP distribution and measures of

central tendency (Hanemann and Kaninnen, 1999, Pg 327)

69
 In the case where WTP 0 (non negative preferences), small differences in the

right tail have essentially no effect on the median value, these differences can however

affect the mean value greatly. This is why the specification of the WTP distribution can

impact the relationship between the mean and median.

Note that in the case of linear utility functions (such as the one utilized in this

study) the mean and median WTP with respect to random preferences are assumed to be

equal (Haab and McConnell, 2002). Using the linear model as described in equation 6

the willingness to pay would be described by;

(10)

However for the lognormal canonical model the willingness to pay can be

described by the following equations;

(11a) Median WTP; C* = e/

(11b) Mean WTP; C+ = e/ exp [1/2]

The equations for the non linear (exponential) model above demonstrate that the

mean WTP estimate can be sensitive to the right tail of the log normal (or log logistic)

distribution (Hanemann and Kanninen, 1999, pg 326). Given their sensitivity to the

assumed form of the error term, exponential models may assign too much weight to WTP

values in the upper tail of the distribution when calculating mean WTP. The linear model

accurately predicts the percentage of no responses at the highest bid (Haab and

70
McConnell, 2002, pg 95), thus making it more appropriate for the CB application of this

study.

Based on the discussion above and because of the major aim of the CB study, the

linear canonical distribution was used to predict the percentage of (Yes/No) responses for

paying the environmental surcharge. Welfare estimates from linear, lognormal and non

parametric methods can however be compared if necessary. More details on non-

parametric welfare estimation will be discussed in the following section.

4.9 Non-parametric Valuation of Contingent Behavior

As mentioned above estimates of willingness to pay can be extremely sensitive to

the distribution or functional form of the preference function. Because of this it can be

useful to develop a less restrictive approach of estimating willingness to pay (Haab and

McConnell, 2002). Responses to discrete choice questions offer the researcher limited

information regarding each respondents true willingness to pay. If the respondent

answers yes to an offered price then their WTP is greater than or equal to the offered

price. If their answer is no then their WTP is less than the offered price. Since WTP is

unobservable to the researcher it can be thought of as random variable with a cumulative

distribution function FW (W), the probability that willingness to pay is less than W. The

probability of a randomly chosen respondent having willingness to pay less than $tj can

therefore be expressed as;

(12) Pr (WTPi < $tj) = FW (tj).

71
 This probability can be denoted as Pr (WTPi < $tj) = Fj., and is the probability that

the respondent will say no to the bid price of tj. Based on this the researcher can

randomly assign different prices to a sample population. Each price would represent a

sub-sample of the total population from which the probability of a no response can be

estimated (Haab and McConnell, 2002). Since the distribution of WTP is unknown, the

series of yes/no responses from the sample represents a series of binary outcomes from an

unknown data generating process and can therefore be used to estimate the probability Fj.

This probability can be treated as an unknown quantity and maximum likelihood

estimators of Fj can be found by using the yes/no responses as well as the number of

respondents per sub-sample in the following likelihood function

(13) ln ln ln 1

where N is the number of no responses, Y is the number of yes responses, T is the

number of respondents (per sub sample) and M is the number of sub-samples.

Maximizing the log of the likelihood function and solving the first order conditions for Fj

yields the maximum likelihood estimates of F

(14)

This means that since Nj is the number of people responding no to the offered

price and Tj is the total number of people offered that particular price then Fj is the

sample proportion of no responses to the offered price. In short the maximum likelihood

72
estimate of the probability that a randomly selected respondent will not be willing to pay

$tj is equal to the sample proportion of individuals that respond no to the offered price.

Therefore if you have a reasonably large sample size and the offered price

increases, the proportion of observed no responses to each bid should increase. The non-

parametric estimation therefore requires monotonically increasing probabilities as the bid

price increases. However in the case of random samples this does not occur and

monotonicity restrictions have to be placed on the maximum likelihood estimation of the

data. This approach is known as the Kaplan-Meier-Turnbull estimator (Hanemann and

Kanninen, 1999; Haab and McConnell, 2002).

The Kaplan-Meier-Turnbull estimator only uses the information contained in

responses to provide estimates of the distribution function. It does not include other

covariates such as age or income. Because of this any estimate of mean or median

willingness to pay is based solely on the yes/no responses. The Kaplan-Meier-Turnbull is

an estimator of the lower bound on the mean of the WTP distribution, C+. This makes it

a conservative estimate if the mean is the chosen welfare from which a confidence

interval can be obtained (Carson et al., 1994; Hanemann and Kanninen, 1999)

Compared to parametric models non-parametric approaches are more robust

against possible misspecification of the response probability distribution. They also offer

the least restricted characterization of what the data have to say and lower bounds of

sample mean WTP can be calculated without assuming any distributions for the

unobserved component of preferences. The simplicity of the method also means the

calculations can be conducted without the use of computers. However since there is no

73
parametric specification of the underlying RUM model, it is not possible to extrapolate

from the observed responses to the measurement of other covariate effects (Haab and

McConnell, 2002).

4.10 Results

4.10.1 Descriptive statistics

Five hundred and sixty-six (566) individuals were approached over the six days of

sampling. Of the persons approached, four hundred and eighty-one (481) agreed to

participate in the survey yielding an overall response rate of 85%. Most respondents

were observed to take approximately 9 to 15 minutes to complete the survey. Of the 481

surveys distributed, 239 contained the environmental tax payment scenario as described

above while 242 contained the tourism tax scenario.

Table 3 below compares the demographics of the two samples after data cleaning

and removal surveys that could not be used in the analysis. At a glance, the descriptive

statistics suggest that respondents for the two samples were similar. From the results

above we can infer that on average persons spent approximately 8 nights in Jamaica,

were between 41 42 years old. Persons also spent on average approximately US$3,000

on travel expenses (airfare, accommodation and other) which typically covered 2 persons

($1,463 per person). Mean household incomes were reported at just above US$120,000

(median value of $90,000). Just under half of both sample population were female, while

US and Canadian visitors comprised the majority of respondents sampled.

74
 The findings also show that respondents generally rated their vacation as good,

the beaches and swimming waters as good or better. Respondents rated underwater life

as fair to good and most stated that they were only somewhat aware about key

environmental and social issues affecting the globe.

Detailed frequency distributions for key demographic statistics were based on

combined sample data. These distributions for individual and trip characteristics can be

found in Appendix 1. Other statistics are shown below including other respondent

characteristics such as the location of their accommodation in Jamaica, nationality as well

as qualitative information on trip satisfaction and ratings on environmental quality.

75
Table 3 Descriptive statistics Contingent behavior survey

Tourism Environmental Combined

Survey Survey Sample
Individual Characteristics

Age (Mean years) 43 41 42

Female visitors 45.6% 49.1% 47.9%

Household income (mean $US) $125,832 $121,586 $123,195

Education (2 High School, 3 College) 2.87 2.81 2.84

Trip Characteristics

Nights in Jamaica (mean) 8 8 8

Per-person trip cost $1,536 $1,579 $1,555

US and Canadians 84.3% 86.9% 85.7%

Other Respondent Characteristics

Trip Satisfaction (5=excellent to 1=poor) 4.66 4.63 4.65

Rating of beaches (5=best to 1=worst) 4.34 4.36 4.35

Rating of marine life (5=best 1=worst) 3.49 3.67 3.58

Environmental awareness
2.12 2.18 2.16
(3 =very 0=none)
*Values for household income and trip cost were taken from the mid-point of the range

76
 The results when compared to the annual tourism statistics (JTB, 2007) suggest

that the sample is representative of the population of tourists who visit the island. For

example the monthly tourism statistics for 2008 show that on average there were 147,273

visitors per month with peak in July (185,447) and the lowest in September (92,037).

There were 142,861 visitors reported for January 2008, which is close to the annual

monthly average for 2008.

Respondents were asked to indicate where possible, the location of their

accommodation while on their visit to Jamaica. Of those who indicated where their

accommodation was located the results show that 27% stayed in Montego Bay, 25%

stayed in Ocho Rios while 26% stayed in Negril (see figure 8 below). The remaining

22% reported staying in other locations.

77
 LocationofRespondents'Hotel

MontegoBay
15%
OchoRios
27%
Negril
7%
RunawayBay
Other

26%
25%

Figure 8 Location of hotels/accomodation stayed at by sampled respondents

These results can be compared with the Jamaica Tourist Board 2008 statistics

(JTB, December 2008). The table below shows the distribution of stopover arrivals by

their intended area of stay for and these are comparable with the findings of this study.

Refer to figure 1 (map of Jamaica) for a reminder of where these towns and cities are

located.

78
Table 4 Stop over arrivals by intended area of stay (JTB, 2008)

Location January to December % Share Length of Stay

Montego Bay 474,755 26.9% 7.9

Ocho Rios 441,398 25.0% 8.7

Negril 351,404 19.9% 8.4

Kingston 226,502 12.8% 12.1

Mandeville/South
118,560 6.7% 13.9
Coast

Port Antonio 17,815 1.0% 17.2

Other 136,837 7.7% 16.5

Total 1,767,271 100% 9.9 (Average)

Note length of stay is based on intended length of stay for total stopover arrivals

Comparing the JTB data with the sampled population it appears that visitors to

Negril appear to have been somewhat over sampled. Montego Bay and Ocho Rios

however show comparable percentages to the JTB data and this suggests that overall, the

demographic data appears to be fairly representative of the general tourist population that

makes visits to Jamaica.

79
Country of Residence

The figure below shows the frequency distribution of the number of respondents

who indicated what country they resided in permanently. The graph below confirms that

the sampled population was comprised mainly of persons who reside in the USA then

followed by Canada.

CountryofResidence
250
235

200

150
Respondents

100

67

50
31

1 2 1 2 1 1 1 2 1 1 5 1 1 1 1
0

Country

Figure 9 Frequency distribution of country of origin Contingent Behavior

Survey.

80
Beach and swimming water quality rating

Respondents were asked to give their opinion of the quality of the beaches and

nearshore swimming waters that they may have encountered on their vacation. They

were asked the following question:

How would you rate the overall quality of the beaches and swimming
waters you experienced in Jamaica?

Below is an outline of the results for this question.

Table 5 Respondent rating of coastal attributes.

Attribute Average Quality (1=Very Poor 5=Very Good)

Beach Sand Quality 4.47

Beach Cleanliness 4.41

Water Clarity 4.56

The table above shows that all attributes had average ratings somewhat above

good but below very good. The figures below give a more detailed breakdown of the

responses provided by respondents. The graphs show that the majority of the respondents

rated the beach sand, beach cleanliness and swimming water clarity as good or very

good. This suggests that in general respondents have an overall positive rating of the

major tourism attribute (sand and sea) that Jamaica offers.

81
 BeachSand
250
204
200
Respondents

150
121

100

50
23
3 3
0
VeryPoor Poor Fair Good VeryGood
QualityRating

Figure 10 Respondent rating of beach sand.

BeachCleanliness
250
206
200
Respondents

150
110
100

50
20
8 8
0
VeryPoor Poor Fair Good VeryGood
QualityRating

Figure 11 Respondent rating of beach cleanliness.

82
 WaterClarity
250 227

200
Respondents

150
108
100

50
16
2 3
0
VeryPoor Poor Fair Good VeryGood
QualityRating

Figure 12 Respondent rating of swimming water clarity

Consistent with the findings in Table 5 above, most respondents rated beach sand,

beach cleanliness and water clarity as good or very good.

Underwater life and visibility quality rating

If the respondent participated in any underwater viewing activity such as

snorkeling, diving or glass-bottom boating while on their vacation they were asked the

following question:

How would you rate the overall (or average) quality of the underwater life
you observed during these activities?

Below is an outline of the results for this question.

83
Table 6 Respondent rating of underwater attributes.

Attribute Average Quality (1=Very Poor 5=Very Good)

Coral Reef Health 3.81

Fish Abundance 3.62

Other Sea Creatures 3.42

Underwater Visibility 4.35

The table above shows that in general respondents had rated the underwater

attributes as fair or above fair. If the average values were rounded to whole numbers and

then assigned the qualitative rating (where, very poor = 1, poor =2, fair =3, good =4 and

very good=5), then coral reefs were rated by respondents as good, fish abundance as

good, other sea creatures as fair and underwater visibility as good.

The figures below show detailed qualitative responses to the question. As the

average values shown in the table above suggest, most of the respondents rated each

attribute as being of fair quality or better. Interestingly for the underwater visibility

attribute there were no ratings of very poor or poor underwater visibility. It should be

noted that there were fewer responses to this question which suggest fewer respondents

engaged in underwater viewing activities than general beach and sunbathing typically

associated with a beach vacation.

84
 CoralReefHealth
90
80
70
60 57
Respondents

52
50
41
40
30
20
11
10 7

0
VeryPoor Poor Fair Good VeryGood
QualityRating

Figure 13 Respondent rating of coral reef health.

FishAbundance
90
80
70 66

60
Respondents

52
50
40 33
30
20 16

10 5
0
VeryPoor Poor Fair Good VeryGood
QualityRating

Figure 14 Respondent rating of fish abundance.

85
 OtherSeaCreatures
90
80
70
60
60 53
Respondents

50
40
30 25
20
20
10
10
0
VeryPoor Poor Fair Good VeryGood
QualityRating

Figure 15 Respondent rating of abundance of other sea creatures.

UnderwaterVisibility
90
80
80 72
70
60
Respondents

50
40
30
20
20
10
0 0
0
VeryPoor Poor Fair Good VeryGood
QualityRating

Figure 16 Respondent rating of underwater visibility.

86
 Consistent with the findings in Table 6, most attributes were rated as good or

very good by a majority of the respondents. Other sea creatures was the attribute

with the lowest rating with a significant number of respondents rating this attribute as

fair.

General level of environmental awareness

Respondents were also asked to indicate their level of awareness on a broad range

of environmental and social issues relevant to todays world. They were faced with the

following question:

What would you say is your general level of awareness about the
following issues affecting the planet as a whole?

The four point Likert scale that was used code the level of awareness to the issues

were; 0=not at all, 1=very little, somewhat=2 and very aware=3. Below is an outline of

the results from the question.

87
Table 7 Respondents reported level of awareness of environmental issues

Issue Awareness (Very Aware = 3 Not at all Aware= 0)

Air Pollution 2.33
Water Pollution 2.67
Solid Waste Pollution 2.09
Nature Conservation 2.20
Climate Change 2.34
Deforestation 2.00
Coral Reef Protection 1.79
Population Growth 2.05

The table above shows that in general respondents were somewhat aware of most

of the issues highlighted above. Of note, respondents were most aware of water pollution

issues while they were least aware of coral reef protection and management issues. The

figures below give a more detailed breakdown of the responses to the question.

88
 AirPollution
200
181
180
160 144
140
Respondents

120
100
80
60
40 23
15
20
0
NotatAll VeryLittle Somewhat Very
LevelofAwareness

Figure 17 Respondent awareness of air pollution issues.

WaterPollution
200
180 168
160 146
140
Respondents

120
100
80
60
40 28
21
20
0
NotatAll VeryLittle Somewhat Very
LevelofAwareness

Figure 18 Respondent awareness of water pollution issues.

89
 Solidwaste
200
180
160 148
138
140
Respondents

120
100
80
60 48
40 29
20
0
NotatAll VeryLittle Somewhat Very

LeveofAwareness

Figure 19 Respondent awareness of solidwaste pollution issues.

NatureConservation
200
180
154 154
160
140
Respondents

120
100
80
60
40 26 29
20
0
NotatAll VeryLittle Somewhat Very
LevelofAwareness

Figure 20 Respondent awareness of nature conservation issues.

90
 ClimateChange
200
180 163
160 149
140
Responents

120
100
80
60
40 25 26
20
0
NotatAll VeryLittle Somewhat Very
LevelofAwareness

Figure 21 Respondent awareness of climate change issues.

Deforestation
200
180
160 153

140 123
Respondents

120
100
80
60 52
35
40
20
0
NotatAll VeryLittle Somewhat Very
LevelofAwareness

Figure 22 Respondent awareness of deforestation issues.

91
 ReefManagement
200
180
160
140 126
Respondents

120 105
100 83
80
60 49
40
20
0
NotatAll VeryLittle Somewhat Very
LevelofAwareness

Figure 23 Respondent awareness of reef management issues

PopulationGrowth
200
180
160 153

140 128
Respondents

120
100
80
60 54

40 28
20
0
NotatAll VeryLittle Somewhat Very
LevelofAwareness

Figure 24 Respondent awareness of population issues.

92
 The figures above suggest that majority of respondents claim to have some

knowledge on the general issues that were outlined. However, they appear to be least

knowledgeable on issues related to coral protection and management.

4.10.2 Summary of descriptive statistics Contingent behavior survey

The descriptive statistics of the sampled population when compared to the annual

tourism statistics (JTB, 2007; JTB, 2008) suggest that the sample is representative of the

population of tourists who visit the island. Persons spend an average of eight nights in

Jamaica and spend between $1500 and $1750 per person. They are typically college

educated with a median age of 42 years old.

The survey data also show that the average tourist has a generally good rating of

the beach sand and cleanliness and swimming water clarity in Jamaica. Fewer visitors

were able to answer questions related to the quality of the underwater attributes (reefs,

fish abundance etc). Those who did provide an answer rated the underwater attributes as

fair or above fair. Respondents were also observed to be somewhat aware of typical

global environmental and social issues.

These results confirm a priori expectations that the average tourist does not have

a high level of detailed knowledge on coral reef ecosystems services and threats. They

are also more familiar with the beach and swimming water attributes than underwater

attributes and do not consider that they possess a particularly high level of environmental

awareness.

93
4.10.3 Response frequencies and non parametric welfare estimation

A non-parametric estimation of WTP was conducted using the Kaplan-Meier-

Turnbull estimator (Hanemann and Kanninen, 1999). As stated previously, this produces

a conservative (lower bound) estimate of consumer surplus and also allows for the

calculation of confidence intervals on the means as well as tests for convergent validity

(Haab and McConnell, 2002).

Figure 25 compares the frequency of the actual Yes dichotomous choice

question responses for both survey versions and confirms that in general, the percent of

yes responses decreases as the level of the bid increases (evidence of a downward sloping

demand curve). The figure shows that as expected, for most bid levels the percentage of

yes responses to the environmental tax is slightly greater than the tourism tax (please

note, the x axis is not drawn to scale) .

94
 100%
Tourism Tax
90%
Environmental Tax

80%

70%
% Yes (Tax)

60%

50%

40%

30%

20%

10%

0%
10 50 90 490 990
BID ($)

Figure 25 Comparison of response frequencies between survey types.

Using the Kaplan-Meier-Turnbull method to analyze the frequency data, mean

lower bound estimates for WTP for the tourism tax were US$130.07 (95% C.I. $49.98)

and $165.15 (95% C.I. $81.50) for the environmental tax. Using the average of nights

stay per visitor and the estimates of WTP we can infer that tourists have a mean per

person value of $16.16 per day for the tourism product, and per person value of $20.52

per day for preserving coastal ecosystem services. Detailed Turnbull calculations are

presented below in Table 8.

95
Table 8 Kaplan-Meier-Turnbull calculations of WTP

Tourism Tax
BID (t) YES NO %Yes (t-t-1)* % Yes
10 31 5 0.8611 8.6111
50 26 12 0.6842 27.368
90 25 24 0.5102 20.408
490 7 31 0.1842 73.684
990 0 11 0 0

WTP 130.07

F* 1-F* T* (F*(1-F*))/T* (t-t-1)^2 ((F*(1-F*))/T*)*(t-t-1)^2

0.138889 0.861111 36 0.003322 100 0.332219
0.315789 0.684211 38 0.005686 1600 9.097536
0.489796 0.510204 49 0.005100 1600 8.159865
0.815789 0.184211 38 0.003955 160000 632.7453
1.000000 0.000000 11 0.000000 250000 0.000000
Variance 650.3349
SE 25.50167
Environmental Tax
BID No Yes Pooled % Yes (t-t-1)* % Yes
10 3 41 41 0.9318 9.3181
50 13 23 56 0.7 28
90 11 33 pooled pooled pooled
490 32 9 9 0.2195 96.585
990 15 1 1 0.0625 31.25

WTP 165.15

F* 1-F* T* (F*(1-F*))/T* (t-t-1)^2 ((F*(1-F*))/T*)*(t-t-1)^2

0.068182 0.931818 44 0.001444 100 0.144393
0.3 0.7 80 0.002625 1600 4.2

0.780488 0.219512 41 0.004179 193600 808.9958

0.9375 0.0625 16 0.003662 250000 915.5273
Variance 1728.868
SE 41.57965

96
 These results can be compared with some of the previously mentioned studies of

recreational and existence value in Jamaica (section 3.7). The study by Wright (1995)

used a zonal travel cost method to estimate consumer surplus (recreational values) for

coral reefs in Negril of US$121 (9 nights). While in 2001, a similar travel cost study for

Ocho Rios Marine Park (EMU 2001) estimated a welfare value of US$132 (10.5 nights).

The semi random street intercept 1998 World Bank survey that asked respondents their

willingness to pay a contribution to a reef management fund produced a WTP estimate of

US$26 (one time contribution). The study by Dharmaratne et al. (2000) calculated WTP

estimates of $10 and $20 per person per day for entrance fees to the beach in Montego

Bay (for repeat and first time visitors respectively). These related studies (along with the

Turnbull lower bound estimates for this study) support the finding that in general there is

a comparatively high level of consumer surplus associated with coral reef ecosystem

services in Jamaica.

The statistical analysis of the dichotomous data allows for the convergent validity

tests of the null hypotheses (H0): WTPTOUR = WTPENV. The null hypothesis was tested

using a standard t-test of the comparison of the means. The test statistic of 4.28 falls well

within the 1% critical region for rejecting the null hypothesis (see Appendix 2). We can

therefore conclude that the mean willingness to pay for the tourism tax is statistically

different from the environmental tax. This confirms our a priori expectations that

WTPTOUR is less than WTPENV.

97
4.10.4 Parametric analysis Tourism and environmental surveys

Table 9 shows the results from the parametric analysis of the two survey data sets.

A linear logistic regression was conducted for each survey data set. A basic model was

estimated where the dependent variable was regressed against the bid coefficient as well

as an expanded model which included key explanatory covariates. This was done for the

tourism and environmental tax survey data.

The estimation results presented in Table 9 are generally consistent with empirical

findings suggesting that the internal validity of the study is sound. In particular the bid

coefficient (Bid) is negative and highly significant in Models I and II for both tourism

and environmental tax surveys. This confirms a priori expectations of a downward

sloping demand relationship between increasing bid levels and the probability of a yes

response. The only other highly statistically significant parameter was the coefficient on

gender for the tourism tax model (95%). For the other variables the coefficients were

shown to have a low level of statistical significance and there was no consistency with

respect to the signs on the coefficients. For example the signs on the coefficients for

nights in Jamaica and age were negative for the tourism tax sample but positive for the

environmental tax sample. While the coefficient for gender in the tourism tax sample

was positive and significant at 90% while it was negative and insignificant in the

environmental tax sample. The coefficients on household income in each survey type

were both negative but not statistically significant. For the environmental survey

regression only the bid coefficient was statistically significant. These findings again

suggest that selected covariates have a low level of explanatory power.

98
Table 9 Linear regression for tourism and environmental surveys (Basic and expanded models)

Tourism Tax Survey Regression Output Environmental Tax Survey Regression Output

Variable Model I t-value Model II t-value Model I t-value Model II t-value

Intercept 1.0752 4.79 1.3009 1.25 1.5457 6.39 07867 0.75

Bid -0.0057 -5.73 -0.0061 -5.46 -0.0055 -6.23 -0.0056 -6.22

Nights in Jamaica -0.0512 -1.20 0.0101 0.30

Age -0.0157 -1.11 0.00113 0.081

Education 0.2844 1.06 0.3057 1.15

Gender (1 = Female, 0 = male) 0.8079 2.05 -0.2535 -0.67

Household Income ($10,000) -0.0066 -0.325 -0.0079 -0.43

USA_Canada -0.1806 -0.348 0.0061 0.01

Log Likelihood -92.73 -87.42 -89.43 -88.29

No. Observations 171 171 181 181

99
 Given that only the constant, bid and gender (tourism tax) coefficients are

significant, the general low level of statistical significance on the other coefficients

suggests that these variables have a fairly low level of explanatory power. Likelihood

ratio (LR) tests were conducted to compare the basic and expanded models for each data

set. That is, the null hypothesis imposes the restriction that all other covariates (apart

from the constant and the bid) are equal to zero. For the tourism survey regression if the

test statistic of 10.6 (6 degrees of freedom) is compared to the chi squared critical value

of 12.59, then you cannot reject the null hypothesis that the other covariates have no

effect. Likewise the chi squared value of 2.28 for the environmental survey LR test

suggests that you cannot reject the null hypothesis. This confirms that the covariates do

not have a high level of explanatory power.

The WTP estimates for both tourism and environmental tax survey data are shown

below in Table 10. These results are based on the estimated coefficients from the linear

logistic regressions. Confidence intervals were calculated using the Krinsky-Robb

procedure (5000 iterations). These were used to calculate the lower and upper bound

values for both the mean and median values of welfare. The welfare estimates for the

basic tourism survey econometric model was $188.60 (95% C.I. $69.90) while for the

expanded model with covariates the mean WTP was $191.67 (95% C.I. $66.03). The

estimated consumer surplus for the environmental tax was $282.97 (95% C.I. $77.95)

for the basic model and $282.27 (95% C.I. $76.29). Like the non-parametric analysis,

the parametric analysis also shows that WTPTOUR is less than WTPENV.

100
Table 10 Comparisons of expected willingness to pay Tourism vs
Environmental tax

Tourism Tax WTP Model I Model II

Mean $188.60 $191.67
Lower Bound $118.71 $125.63
Upper Bound $258.50 $257.70
Environmental Tax WTP Model I Model II
Mean $282.97 $282.27
Lower Bound $205.03 $205.98
Upper Bound $360.92 $358.56

By using the standard errors generated from the Krinksy-Robb procedure a t-test

was conducted to test the null hypothesis (H0): WTPTOUR = WTPENV. Comparisons

between mean WTP for the basic and expanded models give t test statistics that fall well

within the 99% critical region for rejecting the null hypothesis. The test statistics of

Model I (t =10.87) and Model II (t = 10.63) can be compared to the critical value of 1.96.

This leads us to reject the null hypothesis that WTPTOUR is equal to WTPENV. We can

therefore conclude that the mean willingness to pay for the tourism tax is statistically

different from the tourism tax. This again confirms our a priori expectations that

WTPTOUR is less than WTPENV.

101
4.10.5 Combined surveys parameter estimates

The data from the two samples were also combined and a multivariate logit

regression was conducted in order to evaluate the effect of the different treatments

(tourism versus environmental tax). A dummy variable for the environmental tax survey

was created to test the statistical difference between the samples (Table 11). The

coefficients and standard errors from the regression are shown in Table 11.

As expected the bid coefficient (Bid) has a negative sign and is highly significant.

In regression I the dummy variable for the environmental tax version (EnviroTax) is

positive and is significant at the 90% level (p =0.065) and this suggests there is a

significantly different and higher willingness to pay for an environmental tax than a

tourism tax. This again confirms the rejection of the null hypothesis (H0: WTPTOUR =

WTPENV).

In general the other covariates were not statistically significant and hence cannot

be used to make generalized conclusions about the behavior of the wider tourist

population. However the coefficients for Nights in Jamaica and Age are negative. This

implies that the more nights spent in Jamaica and the older a respondent is the less likely

they would be willing to pay any type of additional tax. North American visitors are also

less likely to be willing to pay any form of tax. The estimates suggest that more educated

persons and women have a higher probability of saying yes to any form of tourism

surcharge. While the negative sign on the income parameter suggests that the higher the

income the lower the probability of saying yes to the surcharge.

102
Table 11 Combined Data - Linear regression output

Model I
Parameters Coefficients t-value
Intercept 0.9693 1.33
Bid -0.0056 -8.23
Enviro Tax ( 1= Env, 0 = Tourism) 0.4854 1.85

Nights in Jamaica -0.0211 -0.996

Age -0.0075 -0.782
Education 0.2359 1.24
Female 0.2679 1.01
Household Income (US$10,000/year) -0.0101 -0.739
USA_Canada -0.0869 -0.228
Log Likelihood -179.56
No. Observations 352

Two other logit regressions were estimated using categorical dummy variables for

testing if there were subtle differences not accounted for by the Model I combined

regression. Firstly categorical variables for income were regressed on the dependent

variable to test the possible impact of income level on the probability of saying yes to

tourism the surcharge. Secondly dummy variables were created to test the impact of the

location of the respondents accommodation (hotel or other) may have had on the

probability that they would say yes to any form of tourism surcharge. The results are

shown in the table below.

103
Table 12 Combined Data - Linear regression output with categorical dummy
variables
Model II Model III
Coefficients t-value Coefficients t-value
Intercept 1.2942 0.817 0.9060 1.07
Bid -0.0058 -8.27 -0.0056 -8.20
Enviro Tax ( 1= Env, 0 = Tourism) 0.5342 2.01 0.5060 1.79
Nights in Jamaica -0.0191 -0.846 -0.0232 -1.07
Age -0.0116 -1.17 -0.0069 -0.702
Education 0.2352 1.19 0.2253 1.16
Female 0.2866 1.06 0.2755 1.04
Household Income (US$10,000/year) -0.0074 -0.158 -0.0093 -0.675
USA_Canada -0.0494 -0.126 -0.0807 -0.211

Income Dummy Variables

Income $50,000 -0.4532 -0.770
Income $50,001 $100,000 0.2480 0.255
Income $100,001 $$200,000 0.2715 0.348
Income $200,001 $300,000 -0.4241 -0.778
Income $300,001 -0.7190 -0.717

Location Dummy Variables

Montego Bay 0.0191 0.043
Ocho Rios -0.1692 -0.037
Negril 0.1628 0.360
Runaway Bay -0.0619 -0.102
Other Location 0.1127 0.234
Log Likelihood -177.20 -179.39
No. Observations 352 352

104
 The signs on the parameter estimates for the Bid variable for both Model II and II

remain the same as the combined Model I regression, that is, negative and highly

statistically significant. For the combined regression model II, the level of significance

for the environmental tax dummy variable is now significant at the 95% level (p = 0.045)

while for Model II it remains at the 90% level of significance (p = 0.073). The signs on

the parameter estimates for the other demographic variables remain the same and again

are not observed to be statistically significant.

There were no statistically significant income dummy variables possibly due to

the overall small sample size (N=352). However the signs on the parameter estimates

suggest that the respondent has a higher probability of saying yes to any form of tourism

tax if their household income falls within the range of $50,000 to $200,000 per annum.

The reverse is true for persons with incomes lower than $50,000 and higher than 200,000

per annum. The low levels of statistical significance on the parameter estimates do not

allow for the extrapolation of this finding to the general population of visitors to the

island.

The parameter estimates for the location dummy variables were also not observed

to be statistically significant. However the signs on the parameter estimates suggest that

respondents who stayed in Ocho Rios and Runaway Bay have a lower probability of

saying yes to any tax. While those who stayed in Negril had the highest probability of

saying yes to any form of tourism surcharge. While the level of statistical significance on

the location dummy variables do not allow for conclusive predictions of WTP for a

surcharge, their signs suggest that there may be quality differences (environmental,

105
customer service, congestion effects etc) between locations that might enter respondents

preferences. For example some of comments provided by respondents would suggest

that visitors to Ocho Rios seem to be prone to tourist harassment. The lack of statistical

significance however across locations supports the decision to combine data from

different tourism locations for estimating recreational benefits associated with the entire

coastal tourism product of Jamaica.

4.11 Policy Application - Sustainable Financing for Coastal Resource

Management
One of the objectives of this study is to provide policy relevant information that

can guide the development of a user fee system for that can provide funding for

environmental management and protection. Results from this study can also be used to

generate optimal pricing values for the environmental tax. Based on the fact that tourists

have a significant consumer surplus associated with the beaches and coral reefs of

Jamaica then it would be reasonable to think that they might be willing to contribute to

the sustainable financing of coastal ecosystem protection.

4.11.1 Policy steps

To determine the feasibility of implementing an environmental user fee would

require the relevant policymakers to take the following set of actions or policy steps. The

first step would be to identify the costs of ocean and coastal zone management programs.

Then based on these costs policy makers can determine the necessary ranges for the per-

person tax that would cover the annual costs of resource management.

106
 The second step would be to incorporate the use of the statistical models of

contingent behavior to estimate the impact that these price ranges would have on tourist

visitor rates. This would enable policy makers to make informed decisions between the

trade-off between possible reductions in visitation rates versus the protection and

management of the critical coastal ecosystems, such as coral reefs and beaches.

Lastly, after considering all of the above and consulting with the relevant

stakeholders (hotels, environmental agencies, NGOs, municipalities etc.) a decision can

be made to select the price that would meet the goals of environmental protection and

sustainable development of the coastal tourism industry (Edwards, 2009).

Figure 26 below is a flow diagram detailing the steps mentioned in the previous

paragraph. It should be noted that the estimated welfare values can also contribute to the

decision making process regarding what amounts of taxes to levy on resource users. That

is the amount of consumer surplus that is estimated can be used to determine what an

equitable (fair) portion of the surplus can be captured through user fees or taxes (not just

on tourists but hotels and attractions as well). This process is represented by the dashed

line between the boxes labeled welfare calculation and set tax price.

I would like to emphasize however, as was highlighted in the introductory

chapters, this study is not a benefit cost analysis. The recreational benefit estimation is

used here as part of the process for assessing the feasibility of capturing a portion of the

economic benefits associated with the beaches and coral reef ecosystem.

107
 Identify Natural Resource
Management Costs

Select a Range of User Fees

Conduct Contingent
Behavior Survey

Welfare Calculation Response Probability Model

(Parametric or Non- (Canonical)
Parametric)

Set price of tax that covers Predict Visitor Declines

cost of resource Calculate Potential Revenues
management

Figure 26 Flow diagram Policy steps for sustainable revenue feasibility

assessment.

108
4.11.2 An example: Towards implementation

The information from the canonical probability response model can be used to

guide the development of surcharges or user fees for funding ocean and coastal resource

management. The approximate annual costs of environmental and coastal zone

management for Jamaica were obtained from personal communication with marine park

managers and officials in the coastal zone management branch of NEPA and are shown

in Table 13 below. The actual 2008 budgetary allocation from central government for

environmental management (US$185,133) is also included. It should be noted that the

costs outlined in Table 13 below are overestimates of actual management costs and

represent what would be the best case scenario for resource managers. This represents

the cost of management to maintain or improve current environmental quality in the

coastal areas.

The visitation rate and revenue calculations shown in Table 14 (overleaf) are

based on the canonical probability response model for the environmental tax sample and

the total number of stopover visitors in 2007 (i.e. 1,700,785).

109
Table 13 Estimated annual coastal resource management costs Jamaica.

Estimated Annual Costs

Jamaican $ $1 US = J$71.30
Central Government $13,200,000 $185,133
NEPA $50,000,000 $701,262
5 NGO's (J$15M each) $75,000,000 $1,051,893

TOTAL $138,200,000 $1,938,289

Using the canonical response model we can enter different bid amounts to predict

contingent behavior. The model predicts the probability of a respondent saying yes (or

no) to the given bid amount. From this we can use the predicted percentage of

respondents who would say yes to the tax and multiply this percentage by the number of

stopover visitors to the island in 2007 (1,700,785). This number is then multiplied by the

given bid amount and thus the potential revenue that could be generated is estimated.

The table below outlines the predictions of the reduction in visitation rate (% No to the

tax) for varying levels of the hypothetical environmental tax as well as the potential

revenue that could be generated from those visitors who would say yes to the given bid

amount.

110
Table 14 The impact of environmental taxes on visitation rate and potential
revenues.

Per Person Tax (US) Potential Revenue (US) Rate of Annual Decline in Visitors
$1 $1,692,939 0.5%
$2 $3,370,243 0.9%
$10 $16,236,147 4.5%
$50 $67,143,571 21%
$100 $105,064,272 38.2%
$165.15* $124,665,223 55.6%
* Turnbull consumer welfare estimate

Table 14 above shows that if an environmental tax of $1 per person were

introduced it would not cause a significant decline in visitation rate (0.5%) and would

generate revenues of $1.7M. This would be somewhat lower than the cost estimate

shown in Table 13 of $1.9M for natural resource protection. A $2 per person

environmental tax (in addition to the existing $10 tourism tax) would cause a decline in

visitors of 0.9% and generate revenues of $3.4M. Higher amounts for the environmental

tax are also shown with their corresponding rates of decline. Note that implementing a

tax of $165.15 (Turnbull estimated welfare) would result in a 56% decline in visitors, this

demonstrates why it would not be feasible to try and capture the total consumer surplus

associated with the quasi public good. The figure below shows that environmental tax

revenues could be maximized at $200 per person ($126M per annum). However this

price for the environmental tax would result in 63% rate of visitor decline.

111
 PotentialRevenues EnvironmentalTax
1 140
0.9
120

AnnualRevenue(MillionsUS$)
0.8
0.7 100
Prob(YesTax)

0.6
80
0.5
60
0.4
0.3 40
0.2
20
0.1
0 0
1 5 10 20 30 40 50 100 200 400 500 600 800 1000
EnvironmentalTax

ProbYes Revenue

Figure 27 Impact of environmental taxes on visitation rate and potential

revenues.

It should be noted that the graph above is presented as an illustration and is by no

means a suggestion that the goal of policy makers should be to maximize revenue.

However the response probability model can be used by policy makers to make decisions

of the levels of visitor reduction that can be tolerated based on the goals of preserving the

environment. Table 14 above shows that adequate funds can be generated from an

additional $2 per person environmental tax (on top of the existing $10) with minimal

negative impacts on visitation rates to Jamaica.

112
4.12 Summary of Findings Contingent Behavior Survey
A random airport intercept contingent behavior survey with an 85% response rate

was used to compare estimates of two groups of tourists willingness to pay additional

tourist fees. The nonparametric and parametric estimation of the survey data show that

tourists have a high consumer surplus associated with a vacation in Jamaica, and have a

lower willingness to pay for a tourism tax when compared to an environmental tax.

Statistical comparisons of the median WTP between the two survey data confirms the a

priori expectation that WTPTOUR less than WTPENV.

Some of the covariates such as age, education and nights in Jamaica do not appear

to have much explanatory power. However regression results from the combined data

show that persons with incomes below the median income have a greater probability of

saying yes to the proposed tax and women seem to have a lower WTP than men. The

data also show that North Americans (US and Canadian visitors) are less likely to be

willing to pay any form of additional taxes to visit Jamaica when compared to visitors

from other countries.

The parameter estimates obtained from the linear regression for the environmental

tax survey data were used in a canonical probability response model to predict visitor

behavior in response to varying amounts of environmental taxes. Using this model

makes it applicable for contingent behavior survey analysis. The results show that a two

dollar ($2) environmental fee added to the existing $10 surcharge would cause a

negligible decline in annual visitation rate (-0.9%) while generating enough revenue

($3.4M) to cover annual costs for coastal zone management in Jamaica.

113
 CHAPTER 5

STATED CHOICE STUDY

Measuring the Value of Changes in Coral Reef Ecosystem

Quality in Jamaica

5.1 Introduction
The purpose of the stated choice study is to estimate the recreational benefits or

economic value associated with a quality change in the coastal ecosystem attributes

relevant to Jamaicas tourism product. The study attempts to estimate the benefits

associated with an improvement in quality as well the welfare loss associated with a

decline in quality of four key attributes. These attributes are beach quality, water clarity,

coral reef quality and fish abundance. In addition to the welfare loss or gains associated

with the quality changes, this study also provides an estimate of the access value

associated with the removal of Jamaica from the respondents choice set (for example as

a result of a massive oil spill). The study also aims to provide information on the relative

importance of the key ecosystem attributes to the respondents valuation framework.

114
5.2 Choice Experiments and Welfare Estimation
Choice modeling (CM) techniques include choice experiments, contingent

ranking, contingent rating and paired comparisons (Batsell and Louviere, 1991; Krueger

2007). This study utilizes the choice experiment approach. Bateman et al. (2002) provide

further details on the other CM techniques.

Choice experiments are based on two theories; Lancasters characteristics theory

and random utility theory (Bergmann et al., 2006). Lancasters consumer theory states

that individuals derive their utility from the characteristics of a given good, not from the

consumption of the good itself (Lancaster, 1966). This suggests that the value of the

good is represented by the sum of the goods characteristics. Random utility theory

recognizes that the individual has observable and unobservable components that they

derive from consumption of the given good (McFadden, 1974). Survey response data can

therefore provide the observable portion of an individuals utility but the researcher must

make assumptions about the distribution of the unobserved components of utility in

modeling the probability function to predict which alternatives will be most preferred

over the sample population (Hensher et al., 2005).

In general it is assumed that individuals exhibit utility maximizing behavior

which essentially means that they will choose alternatives that will yield the highest level

of satisfaction given certain constraints such as income, leisure time etc. (Hensher et al.,

2005). If a good can be described in terms of its attribute levels then changing those

levels will result in a different good being produced (Bateman et al., 2002).

115
 Choice experiments tend to focus on the value of changes in the attributes and are

therefore particularly useful in the context of policy analysis because it allows for the

identification of the following (Bateman et al., 2002):

The attributes that are significant determinants of the values people place
on non-market goods.

The implied ranking of these attributes

The value of changing more than one of these attributes at once

The total economic value of a resource or a good

There are five general design stages for choice experiments as outlined by

Bateman et al. (2002). Table 15 below describes these stages.

Table 15 Design stages for choice experiments

Stage Description

1. Selection of attributes The good being evaluated in the choice scenario

2. Assignment of levels
3. Experimental design
4. Construction of choice sets
must be described by a number of characteristics,
or attributes.
Each attribute must be assigned a range of levels.
Levels should be realistic and relate to respondents
preferences.
Statistical design theory is used to generate a
number of alternatives to be presented to
respondents.
Alternatives are paired to create choice sets.

5. Measuring preferences A method is chosen to collect the appropriate data,

and the survey is conducted.
Adapted from: Bateman et al., 2002, pg 259

116
 According to Bateman et al. (2002), stage 1 is the selection of the relevant

attributes of the good to be valued. At this stage attributes are typically chosen through

literature reviews, focus groups, direct questioning, or from the nature of the problem

itself. Typically a monetary cost or payment vehicle is included as one of the

attributes to allow for the estimation of willingness to pay (WTP). Stage 2 is when

attribute levels are assigned. It is important that levels are realistic and should span the

range over which respondents are expected to have preferences for the good being

valued. Stage 3 is the experimental design, where statistical design theory is used to

combine the levels of the attributes into a number of alternatives or profiles. Stage 4 is

the construction of the choice sets. Here the various profiles created during the

experimental design are grouped into choice sets to be presented to the respondents. The

final stage (5) involves the administration of the survey and subsequent the measuring of

preferences.

Choice experiments present a hypothetical scenario where respondents choose the

most preferred option among two or more alternatives. Respondents usually face a series

of three to nine choice scenarios (sometimes more), depending on the complexity of the

survey design and goals of the researcher. The levels of the attributes are varied across

alternatives and across choice sets to allow respondents to face a variety of choice

scenarios (Hensher et al., 2005; Krueger, 2007). It is typically advisable to include a

baseline alternative that corresponds to the status quo in order to ensure welfare

consistent estimates.

117
 There are a number of advantages as well as disadvantages of using CE as

opposed to using other SP methods. One advantage is that choice experiments provide a

realistic way of valuing changes in a good or policy since they force the respondent to

examine more than two alternatives at one time (Bateman et al., 2002). This is possible

with a CV study but would require designing different valuation scenarios for each level

of attribute. Also, because of the nature of the hypothetical scenarios, CE can value

changes in the levels of attributes or goods that do not currently exist, but that may be of

interest for some future policy (Bateman et al., 2002; Krueger, 2007). CE also may avoid

some of the response difficulties in CV, and thus reduce yea-saying. Another important

advantage is that choice experiments typically do a better job than CV at measuring the

marginal values associated with changes in the characteristics of the good. This is

particularly useful for the purposes of policy development.

A number of disadvantages to CE are also recognized. Although CE does derive

welfare estimates, these estimates are sensitive to study design. In other words, the

selection of attributes, attribute levels, and format of the survey itself can all impact the

values of the welfare estimates. Also, because of the nature of the design, attributes that

are important in a respondents decision process yet unknown to the researcher may be

excluded. Choice and ranking complexity also can be a problem for respondents,

especially if the scenarios contain multiple attributes with multiple levels. They can

therefore become complex for the researcher in the design phase and unwieldy for the

respondent when attempting to answer (Thur, 2003).

118
5.3 Survey Design
The stated choice scenario for this study was developed based on feedback from

the pre-test of the contingent behavior survey. The development of the management

alternatives and payment vehicle in the stated choice model were also based in part on the

results from the pretest of the companion contingent behavior survey conducted on the

same population of visitors to the island. This also meant that the same population (of

tourists) and same sampling strategy were used to collect the data.

A critical issue in stated choice models is the adequate presentation of plausible

valuation scenarios and accuracy of descriptions (Mitchell and Carson, 1995). The

valuation exercise should present alternatives in a manner that is understood by

respondents and the payment vehicle must be realistic and incentive compatible.

The stated choice questionnaire was generally structured as follows. Respondents

were first asked questions related to their trip characteristics, trip satisfaction, expenditure

etc. The respondent was then provided with information on the benefits of coral reef

ecosystems services such as: coastal protection, habitat and breeding areas for marine life

and production of white sand for beaches. They were informed that the presence of the

coral reef ecosystem provided food, jobs (for locals), and recreational activities (locals

and tourists). The respondent also received information about threats such as;

overfishing, snorkel and diver damage, sewage, garbage pollution, beach erosion and

overgrowth of coral reefs with algae. The respondents were also asked to answer a Likert

scale question that required them to indicate their previous level of awareness of the

environmental issues (namely ecosystem services and threats) facing coral reefs.

119
 The respondent was also informed that the Jamaican government charges a

tourism surcharge (tax) of US$10 per person to all overnight visitors to Jamaica, typically

included in their overall travel costs (airfare). They were told that these funds are to be

used to; support the maintenance and beautification of the local tourist municipalities,

help maintain the tourism industry and assist in Jamaicas social and economic

development. They were also asked to indicate (yes or no) if they were previously aware

that this surcharge was included in their travel costs.

Respondents were then asked to consider three potential management scenarios.

First a Low Management scenario where minimal management activities would occur

(sporadic beach cleaning) and no additional tourism surcharges would be added to the

existing US$10 per person surcharge. They were informed that this option would result

in an eventual decline in environmental quality.

The second hypothetical management plan was called Basic Management. This

was defined as activities such as marine patrols, beach protection & fisheries

management). In this case there would be an increase on the existing $10 surcharge.

These extras funds would be used to fund the basic management activities. With this

option it is expected that the current environmental quality will be preserved.

The third management plan was labeled Advanced Management and this

included the above described activities as well as additional programs such as solid waste

management, forest preservation and environmental education. Respondents were

informed that with this option it is expected that the environmental quality will be

improved. They were told that all the management options described would affect the

120
quality of Jamaicas beaches, swimming waters, coral reefs and fish life. They were also

informed that noticeable changes (in the case of an improvement or decline) would take

effect in about one year. An example of the complete survey is provided in Appendix 4.

Respondents were provided with information of the different types of attribute

levels that could be expected depending on the level of management. Table 16 below

shows the quality attributes and the corresponding (four) levels that respondents were

asked to consider. This table was to be used by respondents as a reference in order to

assist them in answering the choice questions. The payment vehicle that was used was an

increase in the existing (per person) tourism surcharge that would be included as part of

their overall travel expenses.

121
Table 16 Possible levels of quality change in Jamaican coastal ecosystems.

Attribute Levels of environmental quality

Poor Fair Good Excellent

Beach Eroded, poor sand Some erosion, slightly No erosion, with mainly Wide, with very white
quality improved sand quality white sand sand

Water Poor underwater Variable visibility, Good underwater Crystal clear underwater
visibility (cloudy) (sometimes cloudy visibility (clear) visibility

Coral Reef 90% of corals dead, 15% live corals (85% dead), 40% live corals, other 75% live corals, other
other marine life other marine life (eg marine life seen marine life seen
absent lobsters) rarely seen sometimes (lobsters, octopi)
Fish Only few small fish Moderate number of small Moderate number of Many large and small
seen fish live on the reefs small and few large fish fish seen on reefs
seen

122
 Fractional factorial designs were used to assign attribute levels to alternatives and

reduce the number of alternatives that respondents would face to a manageable level

(Bennett and Blamey, 2002; Hensher et al., 2005). In order to eliminate implausible

choice scenarios, the choice sets were designed so that attribute levels improved as the

management level increased. For example, low management was constrained to have

only fair or poor quality attributes. The fee was also constant at $10. For the basic

management option, the attributes were constrained to be the status quo (current rating of

quality in Jamaica) but the environmental surcharge was randomly allocated across

choice sets ($20 to $150). For advanced management all the ecosystem attributes were

either rated at good or excellent quality and the surcharge applied was always more

expensive than basic management for any given choice set. The statistical program

SPSS was used to generate 50 choice sets from the orthogonal design. The elimination

of implausible choice scenarios resulted in the 50 choice sets being reduced to 18. These

were blocked so that each respondent faced 3 choice sets. Table 17 below demonstrates

how the attributes and levels were organized.

This manipulation of the factorial design means that perfect orthogonality is lost,

however this design is realistic as coral reef degradation occurs in a collinear fashion and

it is expected that all attributes will respond similarly to changes in quality (Thur, 2003;

Parsons and Thur, 2008). Qualitative coding was used to code the ecosystem attributes,

that is Poor = 0, Fair = 1, Good = 2 and Excellent = 3.

123
Table 17 Attributes, levels and their corresponding variables used in choice scenarios

Attributes Hypothetical Management Plans

Low Basic Advanced

Fee $10 (existing surcharge) $20,$40,$60,$80,$100,$150 $40,$60,$80,$100,$150, $200

Beach Quality Fair, Poor Good Good, Excellent

Water Clarity Fair, Poor Good Good, Excellent

Coral Reef Quality Fair, Poor Fair Good, Excellent

Fish (Size and Abundance) Fair, Poor Fair Good, Excellent

124
In addition to the three types of management options respondents were also given the

option of choosing not to visit Jamaica (choose a substitute location or stay home) given

a deterioration in environmental attributes. This is an opt-out alternative needed to

measure access value. An example of the introduction of the policy scenario and choice

set is provided below.

For each choice set, respondents were reminded that the environmental

characteristics for Basic Management were based on current estimates of quality in

Jamaica. After they answered the three stated choice questions a follow up (Likert scale)

debriefing question was used to determine if respondents found the choice scenarios

confusing. Standard demographic information such as gender, age income etc was also

collected.

125
 Now I am going to ask you three hypothetical questions each has the
same format.

Suppose you were planning to take a trip to Jamaica and one of the previously
mentioned management options were already implemented. Suppose also that each
option would vary by the cost of the tourism surcharge as well as environmental quality.

Q. First, assume you were faced with one of the following three options

LowManagement BasicManagement AdvancedManagement

PerPerson
TourismSurcharge $10 $80 $100
BeachQuality Fair Good Excellent
WaterClarity Fair Good Excellent
CoralReef Fair Fair Good
Fish Poor Fair Good
(Feel free to refer to Page 8 for a review of quality ratings)

which of the options (if any) would you prefer? (Check one box below)

Before you answer, please remember that the tourism surcharge is currently US$10
per person. Although this is a hypothetical situation I am asking you to consider
what decision you would really make if you had to spend the extra money, given
your current budget.

Low Management

Basic Management

Advanced Management

I would choose none of the options, instead I would

(If none check one below)

Stay home (cancel my vacation)

Take a vacation in my home country

Visit some other island or country

Figure 28 Stated choice question presented to respondents.

126
5.4 Sampling
A simple random intercept method as was described for the contingent behavior

survey was used to distribute the self administered stated choice survey in the departure

terminals of the Montego Bay International Airport. If the respondent agreed to take the

survey they were left to answer the survey on their own (no influence from the surveyor)

and upon completion they placed the survey booklet in a sealed envelope that was

collected by the surveyor. The stated choice survey exercise was distributed at the same

time as the contingent behavior study over 6 days January 2008. Like the CB survey,

each person was approached randomly and asked if they wished to participate in the

survey. If they agreed they were given the survey to fill out. On completion respondents

were instructed to place the survey booklet in a sealed envelope, which was subsequently

collected by the surveyor.

5.5 Stated Choice Theory

Stated choice theory is based on the same random utility theory outlined in the

previous chapter. In the stated preference approach the respondent is asked to choose

their preferred alternative as described by different attributes at their varied levels. Each

alternative gives an individual some utility defined as by:

(15)

This utility is composed of an objective component (Vi) and an error component

(i) and is also known as a conditional indirect utility function (Adamowicz et al., 1999).

This is because it is conditional on the choice of the object (i) which in the case of choice

127
models is the package of attributes that the respondent faces given the particular

alternative. Selection of one object over another implies that the utility (Ui) of that object

is greater that another (Uj). Overall utility is random therefore you can only analyze the

probability of one choice over another. This can be expressed as shown below

(Adamowicz et al, 1999),

(16) Pr

where C is the choice set. An assumption of Type I extreme value distributed errors

results in the specification of the probability of the respondents choice, or

(17) Pr

Therefore the random utility model provides the basis for the choice experiment

process and can be analyzed using a standard multinomial logit (MNL) framework. The

MNL framework assumes the independence of irrelevant alternatives (IIA) property,

which states that the relative probabilities of two options being chosen are unaffected by

introduction or removal of other alternatives (Greene, 1997; Hensher et. al., 2005). If the

IIA property is violated then MNL results will be biased and hence a discrete choice

model that does not require the IIA property, such as random parameter logit (RPL) or

mixed logit (MXL) model, should be used. Although the MNL model may not violate

the IIA property, it assumes homogeneous preferences across respondents. Preferences

however are in fact heterogeneous and accounting for this heterogeneity enables

estimation of unbiased estimates of individual preferences and enhances the accuracy and

128
reliability of estimates of demand, participation, marginal and total welfare (Greene,

1997; Birol et. al., 2006). Additionally, accounting for heterogeneity enables a better

understanding of who will be affected by a policy change in addition to understanding the

aggregate economic value associated with such changes (Boxall and Adamowicz, 2002).

The mixed logit model (Train, 2003), which accounts for unobserved, unconditional

heterogeneity, should be used in order to account for preference heterogeneity in pure

public goods for example the recreational coastal ecosystem benefits examined in this

study.

In order to account for a general correlation of error terms across the alternatives

as well as to avoid independence of irrelevant alternatives that plagues simpler RUM

models, the choice data were estimated firstly with multinomial and then mixed logit

versions of the model. For more on mixed logit see Train (2003).

5.6 Choice Models

In my application the stated preference/choice experiment places respondents in

three choice situations. They face four distinct alternatives, namely three hypothetical

management options and the no-trip alternative. These four alternatives are also expected

to have an impact on the four key attributes discussed previously.

The data are analyzed using two types of choice models. Firstly, I analyze the

data using the alternative specific constants on the management alternatives and the

environmental surcharge. The second model estimates the welfare changes associated

with the impact of the management alternatives on the environmental attributes. The

129
alternative specific constants on the management alternatives are not included in this

second model.

5.6.1 Stated Choice Model I

Using random utility theory I assume that each utility gives an individual some

trip utility UB for the maintaining the status quo through basic environmental

management, UL for a decline due to low management and UA for an improvement in

quality from advanced management activities, and UN for choosing no trip to Jamaica.

Based on the discussion above, the utilities would take the following form show in the

model below.

(18)

where 1 is the coefficient on the environmental surcharge (fee), 1, 2 and 3 are

alternative specific constants that correspond to the different levels of quality associated

with each management alternative and L, B A and N represent the error terms of the

four options.

As mentioned previously the model could possibly suffer from violations of the

IIA assumption. To ensure that random taste variations are captured requires estimating a

second model that included socioeconomic interactions. In order to do this, key

130
demographic variables were interacted with each alternative specific constant. These

variables were income, age and education and show the effect of various characteristics

on the probability that a respondent will chose a particular option. The utilities take the

following form shown in the expanded model below.

(19)

where 1, 2, 3 and 1 are as described above while 2 through 10 are coefficients on

the aforementioned alternative specific constant interacted socioeconomic variables.

I expect 1 < 0 which should confirm that respondents will dislike higher

environmental fees. Additionally I expect 1 < 2 < 3 thus aligning utility rankings with

improvements in ecosystem quality. The alternative specific constant for the No Trip

option is set to 0 therefore a i > 0 indicates a preference ranking above the neither

option and a i < 0 indicates a preference below the No Trip option (Parsons and Thur,

2008). Based on the literature on the price coefficient and identification, the coefficient

on the fee is assumed to be fixed in the mixed logit (Train and Weeks, 2005). The

interacted socioeconomic variables are also assumed to be fixed in this model.

131
5.6.2 Valuation Stated choice model I

The primary purpose in estimating the choice model is to value changes in the

quality in the coastal ecosystem of Jamaica in key coastal tourism locations. The

parameters of the model are estimated using a conventional mixed logit model (Train,

2003). Each respondent has an expected utility given by EU = E{max (UB, UN)}. EU is

the expected utility of the maximum of choosing Basic Management (the status quo) or

No Trip. The other two options are not included in the valuation simulation because it is

not in the individuals actual choice set. A tourist will either take a trip to Jamaica or use

their funds to select another destination or stay at home. Utility is therefore expressed as

a random value because the utility is random from the perspective of the researcher

(Parsons and Thur, 2008).

To conduct the valuation procedure we compare each individuals expected utility

of a trip to Jamaica when the quality of the coastal ecosystem in Jamaica is in its current

condition with the expected utility of a vacation when the conditions are deteriorated (as

a result of low management). We also compare individuals utilities when conditions are

improved (through improved management).

The expected utility of a Jamaica vacation with deteriorated conditions is

max , where the superscript D indicates a deterioration of quality

(Parsons and Thur, 2008). Conversely the expected utility of a vacation with improved

conditions is max , where the superscript I indicates an improvement

of quality. If there is a lower utility for a trip to Jamaica because of a decline then UN had

132
a greater likelihood of being the chosen alternative in E{max (UB, UN)}. The individuals

decline in utility from a decline in the quality of attributes can be expressed as;

(20) max , max , .

While the individuals increase in utility from quality improvements can be expressed as;

(21) max , max , .

It is these respective changes in quality (decline and improvement) that I attempt

to monetize using the estimated choice model. The decline in quality can be examined

using the expression derived by Hanemann (1984) which is the log-sum of two utilities.

This term takes the form

(22) 1 1

While the improvement in quality can be described by

(23) 1 1

where W is compensating variation and dividing by the coefficient on the fee (1), the

marginal utility of income, monetizes the utility change. This gives a compensating

variation measure of the change in quality in coastal ecosystem attributes. The random

parameters that are estimated result in the calculation of a random compensation

variation. As a result this requires the simulation of a mean compensating variation that

takes the following form:

133
(24) exp 10 1 exp 10 1 /3000

where j denotes one of 3000 draws from the estimated normal distributions for (status

quo) and and where i denotes the quality change in question (decline or

improvement). The model was estimated using LIMDEP econometric software.

It is possible to derive an estimate of access value from the results of the choice

experiment. The difference between the utility derived from visiting Jamaica in its

current state and the utility of not taking a trip (the neither option) appropriately

monetized is a measure of access value (Thur, 2003). The difference between the

monetized utilities can be considered to be an upper bound on the loss that would be

associated with a significant decline in coastal ecosystem quality. In other words this is

effectively removing a Jamaica vacation from the individuals choice set (Thur, 2003).

This willingness to pay can be expressed as:

(25) 1

The estimates of access value may be compared to welfare measures derived from

comparable CV data (Thur, 2003). In this instance the access value can be compared

with the WTP estimates from the environmental survey data presented in the previous

chapter. This will be discussed in more detail in chapter 6.

134
5.6.3 Stated Choice Model II

We can also derive policy relevant information from the relative values of the key

environmental attributes to the respondent. Random utility theory allows us to assume

that if the levels of quality for the particular attribute is preserved through basic

environmental management then this gives an individual some trip utility UB. The

individual will likely experience some trip utility UL for a decline in environmental

attributes given low management and UA for an improvement in quality of the key

attributes from advanced management activities, and like the previous model UN for

choosing no trip to Jamaica. Based on the discussion above, the utilities would take the

following form shown in Model II below.

(26)

where Ui = UB, UL, or UA and UN is again = 0. Like the previous model 1 is the

coefficient on the environmental surcharge (fee), while 2 through to 5 are coefficients

that correspond to the different levels of attribute quality for each alternative. Like the

previous model (i) represents the error terms of the four options.

It should be noted that the equation above assumes the marginal effects of each

attribute are constant at each management option. An alternative model would allow the

marginal effects of each attribute level to vary at each management option (poor, fair,

good, excellent). This model was not explored in this study but could possibly be used

for future studies.

135
5.6.4 Valuation Stated Choice Model II

A linear statistical model such as the one presented above in equation 26 can

generate coefficients that can be used to estimate the rate at which respondents are

willing to trade off one attribute for another. For policy analysis and welfare estimation

purposes the trade off occurs between a particular attribute and a monetary attribute.

This trade-off is known as a part-worth or implicit price. The implicit price demonstrates

the amount respondents are willing to pay in order to receive more of the non-market

environmental attribute (Bennett and Adamowicz, 2001).

(27) Part-worth = (nonmarket attribute/monetary attribute)

Estimates of implicit prices are made ceteris paribus and are estimates of the

willingness to pay an increase for the attribute of concern given that everything else is

held constant. It should be noted that part-worths can be applied to derive the willingness

to trade off between any pairs of attributes. This means that these estimates of marginal

rates of substitutions can be applied in case of non-monetary compensation for

environmental changes (Bennett and Adamowicz, 2001). For example trading

determining how much coral reef re-planting will compensate for seagrass removal for

creating a swimming beach in front of a hotel.

Implicit prices for policy purposes are useful because they allow for an analysis of

the composition of potential alternative allocations of resources. The comparison of the

relative attributes allows for some understanding of the relative importance the

respondents hold for them. For example, respondents may value the beach more than the

136
underwater life. By being able to make these comparisons, policy makers can better

allocate resources to protection and preservation of particular attributes. One should be

careful however that comparison of implicit prices of attributes are not undertaken

without recognizing that different units could have been used to define the attributes (this

is not so for this study). It is also very important to note that implicit prices are generally

not welfare measures and therefore should not be used in benefit-cost analyses.

Given the limitations of part worth estimation it is also possible to simulate

compensating variation measures of welfare for changes in quality of selected attributes.

The simulations are based on the aforementioned qualitative ratings of attribute quality

(i.e. 0 = poor, 1 = fair, 2 = good, 3= excellent). This requires the simulation of a mean

compensating variation that takes the following form:

(28) exp 10

1 exp 10 2 2 1 1 1

It should be noted in equation 28 above, that the status quo ratings for the

attributes are beach = 2, water = 2, reef = 1 and fish =1. While the attribute quality

ratings for the utility function that represents simulated quality changes (i.e. attribute*)

will depend on whether there is a decline or improvement in attribute quality. More

details on how this was modeled are provided in section 5.11.2. Based on the discussion

above, these simulated compensating variation welfare measures have more policy

applicability in comparison to implicit prices.

137
5.7 Results
The results of the stated choice survey are presented below. Some key descriptive

statistics of the sample population are presented. The parameter estimates of the

multinomial and mixed logit regressions as well as simulations are then presented along

with a discussion of the results.

5.7.1 Descriptive statistics

Two hundred (200) individuals were approached over the six days of sampling.

Of the persons approached, one hundred and eighty (180) agreed to participate in the

survey yielding an overall response rate of 90%. Most respondents were observed to take

approximately 15 minutes to complete the survey. Table 18 below shows the

characteristics of the respondents in the sample.

138
Table 18 Descriptive statistics State choice survey

VARIABLES Values
Individual Characteristics
Age (Years) 41.8
Household income (US$) $121,522
Female visitors 58.05%
Elementary school 1.15%
High School 25.86%
College 54.02%
Graduate degree 18.39%

Trip Characteristics
Nights in Jamaica 7.06
Per-person trip costs $1,630.42
US visitors 64.37%
Canadian visitors 28.74%
Other visitors 6.32%

Other Respondent Characteristics

Aware of Surcharge 32.9%
Aware reef issues (1 not aware 4 very aware) 2.8
*Values for household income and trip cost were taken from the mid-point of the range

On average respondents were 42 years old and earned a mean annual household

income (using the mid-point of the range) of approximately US$122,000. Just over half

the respondents were female and the majority of the sampled population had college

degrees. Respondents were shown to have spent an average of 7 nights on vacation with

a per-person cost of about $1,600. Of the visitors sampled 64% were from the USA and

28% were from Canada. This sub-sample again coincides with the Jamaica Tourist

139
Boards statistics that North American visitors comprise the majority of vacationers to the

island.

Figure 29 shows a bar chart of the combined results for the stated choice

questions. It shows the percentage of management options that were selected by

respondents. Figure 29 shows that of the management options that were selected, basic

management had the highest proportion (36%), while advanced management and the

second highest proportion (33%). The results show that 69% of the respondents chose an

option that would either maintain current environmental quality or cause an improvement

in the attributes. Or in other words almost 70% of the choices showed a willingness to

pay extra for improved environmental management.

The low management option was selected 12% of the time by respondents while

19% for the no trip option. Respondents who choose the no-trip option were asked to

indicate what they would do instead of choosing not to visit Jamaica (refer to figure 28).

Of those who provided this extra information 1% indicated that they would stay home

(cancel their vacation) while 16% said that they would choose to take a vacation in their

home country instead. Interestingly 83% indicated that they would visit some other

island or country for their vacation thus confirming that there are similar substitutes to a

Jamaica beach vacation.

140
 %Respondentschoosingeachoption
100%
90%
80%
70%
60%
50%
40% 36%
33%
30%
19%
20%
12%
10%
0%
NoTrip Low Basic Advanced
ManagementOptions

Figure 29 Ratio of management options selected by all respondents.

The detailed frequency distributions for key demographic statistics such as age,

household income, travel costs and nights spent in Jamaica are found in Appendix 1.

Other statistics are shown below including other respondent characteristics such as

nationality as well as the respondents awareness of the Jamaican coral reef issues.

141
Country of Residence

The graph below shows the frequency distribution of the number of respondents

who indicated what country they resided in permanently.

CountryofResidence
120
112

100

80
Respondents

60
49

40

20
8
1 1 1 2
0
Canada England France Israel Nigeria Northern USA
Ireland
Country

Figure 30 Frequency distribution of country of origin Stated choice survey.

The results again confirm that the sampled population was comprised mainly of

persons who reside in the USA then followed by Canada. This shows that the sample is

142
comparable to the tourist boards statistics and is therefore representative of the typical

tourist population that visits the island.

Awareness of Jamaican Reef Issues

Respondents were asked to indicate their level of awareness of Jamaican coral

reef management issues. A four point likert scale was used to code the answers, these

were; not at all aware =1, a little aware = 2, somewhat aware =3, very aware =4. The

graph below shows the results of those respondents who answered the question.

JamaicanCoralReefIssues
90
78
80

70

60
Respondents

50
43
40 35

30

20
12
10

0
Notatall Alittle Somewhat Veryaware
LevelofAwareness

Figure 31 Respondent awareness of Jamaica's coral reef management issues.

143
 These results show that the respondents who completed the stated choice survey

are similar to those who completed the contingent behavior survey. The above figure

confirms that the sample population was comprised largely of tourists that do not have

particularly high levels of knowledge on coral reef ecosystems services and threats. Most

expressed some awareness of the coral reef management issues but those who considered

themselves very aware were not the majority of the sampled respondents.

5.7.2 Summary of descriptive statistics Stated choice survey

Like the contingent behavior sample, the descriptive statistics suggest that the

sample is representative of the population of tourists who visit the island (see appendix

5). That is the majority of tourists are from the USA, they stay for an average of seven to

eight nights in Jamaica and spend approximately $1,600 per person. They are typically

college educated with a median age of 42 years old.

Approximately seventy percent of the respondents were unaware that they paid a

$10 per person surcharge to visit Jamaica. Respondents were also somewhat aware of

coral reef management issues affecting Jamaica, and this confirms a priori expectations

that the average tourist does not have a high level of detailed knowledge on coral reef

ecosystems services and threats.

5.8 Stated Choice Model I Parameter Estimates

Below are the results for the mixed logit regressions of the stated choice version

of the survey. The models presented are the basic, expanded (with socioeconomic

interactions) and the attributes only model as outlined in sections 5.6 and 5.7.

144
5.8.1 Basic model

The estimates for the basic multinomial and mixed logit models are shown below

in Tables 19 and 20. For both the MNL and MXL models the alternative specific

constants satisfy the a priori expectations (1[Low] < 2[Basic] < 3[Advanced]). .

For the MNL (fixed parameter) model (Table 19) the mean coefficient is negative

for the low management option and indicates a lower preference ranking than the no trip

option. The basic management option is positive while the mean coefficient is highest

for the advanced management option. The t-values also show that the basic and

advanced management options are statistically significant at the 0.01 level or lower. The

coefficient on the tourism surcharge is negative and highly significant as expected. This

confirms expectations of a downward sloping demand relationship between increasing

bid levels and the probability of a yes response.

145
Table 19 Fixed parameter MNL basic Model I

Variable Fixed Coefficient Model t value

Low Management -0.3126 -1.938

Basic Management 1.672 8.005
Advanced Management 2.162 7.669
Surcharge (Fee) -0.0134 -6.658
Log Likelihood -658.1
Number of Respondents 174
Total No. of Responses (Nx3) 522

For the MXL model (Table 20) the mean coefficient for the low management

option is also negative, the basic management option is positive while the mean

coefficient is highest for the advanced management option. The t-values for the low and

basic management options are not statistically significant while the advanced

management option is statistically significant at the 0.01 level or lower. The coefficient

on the tourism surcharge is negative and highly significant as expected.

The standard deviations for the random parameters mixed logit model shown in

Table 20 below suggest that there is the possibility of unobserved heterogeneity in the

sample of respondents. The mean variables for Low and Basic Management have

standard deviations that are larger in value. However the t-values of the coefficients

suggest that the standard deviations are not statistically significant. It is expected that the

expanded model will provide more information on possible effects of heterogeneity in the

sample.

146
Table 20 Random parameter Mixed Logit basic Model I

Variable Random Parameters Standard Deviation

Low Management -0. 6989 (-0.639) 1.216 (0.681)
Basic Management 1.039 (0.811) 3.413 (0.881)
Advanced Management 2.362 (6.620) 0.0797 (0.052)
Fixed Parameters
Surcharge (Fee) -0.015 (-5.716)
Log Likelihood -656.76
Number of Respondents 174
Total No. of Responses (Nx3) 522
(t-values in parentheses)

In order to test for the accuracy of the assumption of IID error terms a mother-

logit model (or base model) was estimated. The base model is equivalent to a model

estimated with the alternative specific constants only. To determine the estimated model

is significant you compare the log likelihood functions of the estimated model with that

of the base model. That is, if the LL function of the estimated model is shown to be an

improvement over the base model then the model is considered to be statistically

significant overall (Hensher et. al., 2005, pg 330). The likelihood ratio test between the

base model and both the fixed coefficient (MNL) and random parameters (MXL) models

indicate that at the 5 percent significance level the estimated model was the true model.

MNL and MXL regressions were also conducted with the Likert scale confused

parameter (i.e. if the respondent found the three choice scenarios confusing). This

parameter was not shown to be statistically significant.

147
5.8.2 Expanded model socioeconomic interactions

The estimates for the expanded MNL model are shown in Table 21. The mean

coefficient is negative for the low management option and indicates a lower preference

ranking than the no trip option. The basic management option is positive while the mean

coefficient is highest for the advanced management option. The t-values also show that

the basic and advanced management options are statistically significant at the 0.05 level

or lower. The coefficient on the tourism surcharge is negative and highly significant as

expected. This confirms expectations of a downward sloping demand relationship

between increasing bid levels and the probability of a yes response.

No a priori expectations were held for the other socio-economic variables. For

both models the interacted income variable was statistically significant at the 95% level

for all options and was positive for low management while negative for basic and

advanced options. The interacted variable for age for the basic option was observed to be

significant at the 90% level. While not much can be inferred from the age and education

variables the results suggest that the lower a respondents income the higher their

probability of selecting the low management option.

148
Table 21 Fixed parameter MNL, expanded Model I

Variable Fixed Coefficient Model t values

Low Management -0.670 -0.773

Basic Management 2.57 3.607

Advanced Management 3.35 4.413

Tourism Surcharge (Fee) -0.014 -6.755

Low*Income 0.004 3.602

Basic*Income -0.002 -2.64

Advanced*Income -0.006 -2.052

Low*Age -0.003 -0.508

Basic*Age -0.010 -1.62

Advanced*Age -0.023 -1.061

Low*Education -0.204 -1.59

Basic*Education 0.121 1.014

Advanced*Education 0.010 0.024

Log Likelihood -599.737

Number of Respondents 174

Total No. of Responses (Nx3) 522

149
 Based on the results of the MNL regression above the age and education

interacted variables were omitted and a mixed logit regression with only the management

options, surcharge and interacted income variables was estimated. The results from the

500 Halton draw regression are shown below.

Table 22 Random Parameters Mixed Logit expanded Model I

Random Parameters Standard Deviation

Low Management 0.1626 (0.183) 0.290 (0.715)
Basic Management 1.712 (0.449) 0.564 (1.373)
Advanced Management 2.802 (0.453) 0.062 (0.193)
Fixed Parameters
Tourism Surcharge (Fee) -0.0156 (-5.39)
Low*Income 0.0245 (2.29)
Basic*Income -0.0356 (-1.409)
Advanced*Income -0.0014 (-2.134)
Log Likelihood -649.76
Number of Respondents 174
Total No. of Responses (Nx3) 522
(t-values in parentheses)

Although the coefficients are not statistically significant, the alternative specific

constants once again satisfy the a priori expectations (1[Low] < 2[Basic] < 3[Advanced]).

The mean coefficient is lowest for the low management option but now indicates a

slightly higher preference ranking than the no trip option (p =0.079). Like the previous

models the basic management option is also positive while the mean coefficient is highest

150
for the advanced management option. The standard deviations for the random parameters

shown in the table above suggest that there is little unobserved heterogeneity when

individual characteristics are added to the sample. None of the mean random parameters

or their standard deviations was statistically significant. For the fixed parameters the sign

on the coefficient for the tourism surcharge is negative and highly significant as expected.

The t values for the alternative specific constant (ASC) interacted income variables for

low and advanced management options are significant while the basic*income variable

was not statistically significant.

The standard deviations of the mean random parameter estimates from the MXL

regressions show that heterogeneity is present in the sample population with respect to

preferences for management options and attribute quality. This variation across the

sample population will be considered when discussing the welfare estimates.

5.9 Stated Choice Model II Parameter Estimates

Coefficients on the environmental attributes and fee variables were estimated in

stated choice model (II) using both MNL and MXL regressions. The results for the MNL

regression are shown below in Table 23. As expected the coefficient on the tourism

surcharge is negative and significant. The sign on the coefficients for beach and water

quality are positive and statistically significant. While the signs on the reef and fish

attributes are negative and not statistically significant. The positive sign on the beach and

water quality coefficients suggests that holding all other attributes constant respondents

would be willing to pay for an improvement in beach and water quality respectively. The

151
opposite is true for the reef and fish attributes, that is, for a given parameter (reef or fish)

if you hold all other parameters constant respondents are not willing to pay for an

improvement in that parameter.

Table 23 Fixed parameter MNL Model II

Parameters Coefficient t value

Beach 0.4152 3.00
Water 0.4198 2.94
Reef -0.1117 -0.964
Fish -0.0040 -0.030
Surcharge (Fee) -0.0092 -4.47
Log Likelihood -674.6
Number of Respondents 174
Total No. of Responses (Nx3) 522

The mixed logit regression output is shown Table 24 below, the parameter

estimates are the result of 500 Halton draws. Like the MNL regression the signs on the

attribute and fee parameters remain the same. However the reef attribute is now

statistically significant at the 90% level of statistical significance (p = 0.065). As a

reminder, Model II assumes the marginal effects of each attribute are constant at each

management option, and does not allow for estimating the marginal effects of the

different levels for each attribute.

152
Table 24 Random Parameters Mixed Logit Model II

Random Parameters Coefficient t value Standard Deviation t value

Beach 0.7377 3.599 0.0093 0.027
Water 0.5754 3.104 0.7286 3.077
Reef -0.2954 -1.844 0.0090 0.038
Fish -0.0848 -0.482 0.8333 2.641
Fixed Parameters
Surcharge (Fee) -0.0123 -3.938
Log Likelihood -668.98
Number of Respondents 174
Total No. of responses
522
(Nx3)

The signs and statistical significance on the coefficients suggest that changes in

the quality of the beaches, swimming water quality and (possibly) coral reef health

contribute to the respondents valuation framework. The lack of statistical significance

and the negative sign on the fish attribute suggest that this quality attribute may not enter

their valuation framework. The negative signs on the coefficients for reef and fish

quality are probably due the sample populations general lack of familiarity with the

attributes. For example of the respondents surveyed, 3% stated that they were certified

SCUBA divers while 10% stated that they had engaged in SCUBA diving activities while

on vacation in Jamaica. This 10% would include persons who were not certified but

perhaps engaged in try dive excursions. The relatively low level of statistical

significance and the signs on the fish and reef attributes may have been different

(positive) if scuba divers were the target sample population. This is because divers are

153
likely to be more comfortable underwater and more familiar with underwater marine life.

Their rating of the underwater attributes is likely to be greater than that of the average

beach tourist.

The standard deviations of the random parameters suggest that there is little

unobserved heterogeneity for the beach and reef parameters while there is some

unobserved heterogeneity, with respect to the water quality and fish abundance

parameters. These two parameters have standard deviations that are somewhat larger

than the mean value and are also statistically significant (95%). These results suggest

that there is some heterogeneity with respect to how respondents rate water quality and

fish abundance, while there are no significant differences in preferences for beach and

reef quality. In other words, the relative magnitude of the standard deviations implies

that there is a probability that visitors may have reverse preferences for water clarity and

fish abundance.

A comparison between the MNL and MXL models show that there are similar

trends regarding the signs and relative sizes of the parameter estimates. The important

additional information that the mixed logit model gives is perhaps mainly that there is a

strong heterogeneity in the preferences for the water quality and fish attributes.

5.10 Welfare Estimates

As a reminder the welfare estimates that are estimated are based on the quality

changes associated with the management options. The status quo or existing

environmental conditions can be considered to be; beach quality as good, water clarity

154
as good, coral reef health as fair and fish abundance also fair. Therefore, low

management is expected to result in a reduction in quality from the status quo while the

advanced management option is expected to increase the current environmental quality.

Basic management is expected to maintain the status quo environmental quality. As

stated previously because of the ecological relationships in the tropical coastal

environment, these attributes are expected to move collinearly.

The welfare estimates for Model I will be discussed below followed by the

implicit price estimates for Model II. Welfare estimates from model simulations will be

discussed in section 5.11.

5.10.1 Model I Welfare estimates

The welfare changes associated with the hypothetical quality changes are shown

in table 25 below. The changes are based on the expected quality changes as discussed

above (section 5.10). The estimates are presented as mean per person trip values for the

sample of respondents. Welfare calculations shown below are based on the basic MXL

model outlined in Table 20 above. The results show that respondents would experience a

welfare loss of approximately $97 if coral reef ecosystem quality were to decline from

current conditions. In compensating variation terms this means individuals income

would have to be increased by about $97 to maintain their utility in lieu of a decline in

quality. The welfare gain associated with an improvement from the status quo was

estimated at $22 per person. In other words respondents are willing to pay this amount to

move from the status quo to an improvement in coastal ecosystem quality. It should be

155
noted that the estimates of mean welfare for have large standard deviations which

confirm the heterogeneity of preferences in the sample population.

Table 25 Welfare associated with coastal ecosystem quality changes

Mixed Logit Welfare Estimates (Compensating Variation)

Mean
Scenario 25th Percentile Median 75th Percentile
(Std Dev)

Quality Decline -$2.90 -$54.16 -$96.61 -162.27

(110.79)

Quality Improvement $133.42 $73.77 $22.39 -$51.80

(134.73)

Access Value -$14.09 -$76.78 -$128.21 -$205.52

(138.95)

Table 25 also shows the estimates for the access value a Jamaica vacation. The

access value is calculated from the difference between the utility from a trip to Jamaica in

its current environmental state and the utility of not taking a trip to the island. Using the

MXL regression, the results show that the estimated access value for Jamaica is $128.

This means that the removal of a Jamaica vacation represents a welfare loss of

approximately $128 per person. This value can be compared to the range of WTP values

that were estimated for the companion contingent behavior survey in the previous

156
chapter, that is, the respondents were faced with the choice of a vacation to Jamaica or

not (depending on increased surcharges).

The value of the welfare loss from the quality decline can be used to calculate an

annual aggregate recreational value of the beaches, and coral reef ecosystems that support

the tourism industry in Jamaica. The table below outlines the annual consumer surplus

associated with a decline, an improvement or total removal of a Jamaican vacation. The

aggregate values are based on multiplying the mean per person welfare estimates by the

number of stopover visitors to Jamaica in 2007 (that is 1,700,785 persons).

The results show that visitors have a significant consumer surplus associated with

a typical beach vacation (sun, sea and sand) in Jamaica. The estimates of the recreational

benefits of quality change can be a key component of any thorough benefit cost analysis

that may need to be conducted for coastal resource management of Jamaicas coastal

towns and tourism locations.

157
Table 26 Annual welfare values associated with quality changes and total loss

Scenario Mean Annual Aggregate Surplus

Quality Decline -$96.61 -$164,312,839

Quality Improvement $22.39 $38,080,576

Access Value -$128 -$217,700,480

Table 26 above shows that if you were to use the annual number of stopover

visitors to Jamaica in 2007 to derive aggregate consumer surplus you would confirm that

Jamaica is highly valued as a vacation destination. Annual aggregate welfare loss from a

decline from current quality would be equal to US$164 million. However if the current

quality were to be improved this would result in a gain in consumer surplus of $38

million per year. This implies that this is what an improvement in quality would be

worth to tourists that visit Jamaica. However if the quality of the coastal ecosystem were

to decline to the point that it was no longer worth to visit, or if a natural disaster such as a

massive hurricane or tsunami were to completely destroy the coastal environment this

would result in an annual welfare loss of $217 million.

The annual surplus values for the Jamaica coastal ecosystem shown above may be

incorporated into a larger benefit cost analysis study or the findings here may be used to

justifying the implementation of user fees or other economic instruments as a means to

generate funds for resource management. The next step in the process would be to decide

how to allocate funding for resource management.

158
5.10.2 Model II Part-worth calculations

Implicit price calculations from the mixed logit coefficient estimates are shown

below (table 27). These estimates indicate that respondents are willing to pay $60 for an

improvement in the quality of the beaches (holding all other attributes constant). While if

water quality were to be improved, holding all other attributes constant, respondents

would be willing to pay an amount equivalent to $45. However if reef quality were to be

improved holding all other attributes constant that respondents would have to be

compensated with an amount equivalent to $24. The implicit price for fish abundance

was the lowest at $9 and this suggests that if you were to hold all other attributes constant

and increase fish abundance, respondents would have to be compensated by this amount.

The comparatively larger standard deviation on the means for the water and fish

attributes again confirms that heterogeneity of preferences exists among the sample

population.

Table 27 Implicit Prices for each Attribute (Mixed Logit Model)

Quality 25th Percentile Median Mean (Std Dev) 75th Percentile

Attribute
Beach $59.40 $60.14 $60.02 60.53
(0.75)
Water $6.28 $44.61 $45.18 $84.71
(58.14)
Reef -$24.64 -$24.06 -$24.05 -$23.57
(0.72)
Fish -$53.28 -$9.44 -$8.79 $36.43
(66.50)

159
 As discussed previously these results may be indicative of the sample population.

The average tourist may not have a full appreciation of the role of ecosystem services, as

a result attributes such as reef and fish abundance may not enter their valuation

framework. The high statistical significance and positive implicit prices for beach and

water quality improvement suggest that these two attributes are very important to the

respondents valuation framework. While the comparatively low level of statistical

significance on the underwater attributes suggests that respondents have a lack of

familiarity with these two attributes (reef and fish). This is not surprising given Jamaicas

reputation as a sun, sand and sea destination. It should be noted that the estimated

implicit prices for the reef and fish attributes are based on parameter estimates with low

levels of statistical significance and for policy purposes would not carry much

explanatory weight.

It should also be reiterated that implicit prices are not estimates of compensating

surplus that are typically used in benefit-cost analyses. This is because the attributes of

change from the status quo do not capture all the reasons why respondents might choose

an alternative that would result in an improvement in environmental quality (Bennett and

Blamey, 2001). However the implicit prices allow for a better understanding of the

relative importance that respondents hold for each attribute. This fact can be useful for

designing policy for management and protection of key environmental attributes and will

be discussed in a later section of this chapter.

160
5.11 Model Simulations
Model simulations were run in NLOGIT using the MXL results from Model I

(basic). The simulations were used to determine how changes in the price attribute

impacted the choice probability for the management alternatives.

Model simulations were also run using MS Excel using the MNL results from

Model II to determine how changes in the quality of the attributes would affect

respondents welfare. The results of these simulations are presented below.

5.11.1 Model I simulation Changes in surcharge price

Simulations were run so that respondents would choose between the basic

management alternative or the no trip option as well as the advanced management or no

trip option. Each model was simulated using the following constraints:

1. The low management option was not provided as an option

2. The status quo would be maintained with basic management

3. Quality improvements would be expected with advanced management

4. The per-person environmental surcharge varied from an additional $10 to

$1,000.

These scenarios mimic what is reasonably expected given the proposed

management alternatives presented to the respondents. It should be noted that for the

actual choice scenarios in the survey, respondents were presented with bids ranging from

$20 to $200 (see Table 17).

The simulated results presented in figure 32 below show that as expected, across

the range of prices, the percentage of respondent support for the basic management

161
scenario is lower than that for the advanced management scenario. For each option as the

price of the surcharge increases the percent of respondents that are willing to pay for

either basic or advanced management options decreases. This confirms the downward

sloping demand for the non-market good as price increases.

For a bid of $200 (the maximum bid presented in the surveys) for basic

management the simulated results show that 70% of the respondents would have chosen

not to take a trip to Jamaica, or in other words only 30% would pay this amount for basic

management. While for advanced management 67% would choose not to take a trip

while 33% would be willing to pay $200 for advanced management. At a simulated

maximum bid of $1,000 (per person) for both management options the results show that

100% of respondents would choose not to take a trip to Jamaica.

162
 %YesEnvironmentalSurcharge
100%
90%
80%
Percent(YesSurcharge)

70%
60%
50%
40%
30%
20%
10%
0%
10 20 40 60 80 100 150 200 500 1000
Surcharge($)
BasicManagement AdvancedManagement

Figure 32 Simulated respondent support for an environmental surcharge.

5.11.2 Model II simulation Changes in attribute quality

Simulations for four possible scenarios of changes in attribute quality were

conducted based on Model II parameter estimates. Table 28 below describes the

scenarios of quality change that were used in the simulation. The quality attributes for

the status quo remained the same while the environmental quality attributes either

improved or declined depending on the management option (advanced or low). The

surcharge remained fixed at $10 for all scenarios and management options.

The scenarios shown in below allow for the calculation of the welfare changes

associated with either an improvement in quality given advanced management or a

163
decline in environmental quality with the low management scenario. Equation 28

(section 5.6.3) was used to calculate the estimates of welfare for a decline and

improvement in quality. Scenario 1 is used to simulate the change in welfare for changes

in beach quality only. That is attribute quality declines from good to fair for low

management, while beach quality improves from good to excellent for the advanced

management option. Scenario 2 models the welfare change associated with a similar

improvement or decline in water quality (only) from the status quo. Scenario 3 models

the welfare changes when both beach and water quality either improves or declines.

Finally scenario 4 models the changes when both beach and water quality declines from

the status quo to poor (the lowest level of the attributes). It should be noted that because

of the comparatively lower levels of statistical significance on the mean parameters for

the coral reef and fish quality variables these levels were not changed from the status quo

in the simulations (see table 28 below).

164
Table 28 Four scenarios of quality change used in the Model II simulation.

Management Options

Status Quo Scenario 1 Scenario 2 Scenario 3 Scenario 4

Attributes

Basic Low Advanced Low Advanced Low Advanced Low

Beach Good Fair Excellent Good Good Fair Excellent Poor

Water Good Good Good Fair Excellent Fair Excellent Poor

Coral Reef Fair Fair Fair Fair Fair Fair Fair Fair

Fish Fair Fair Fair Fair Fair Fair Fair Fair

165
 Table 29 below outlines the welfare changes associated with each of the

scenarios.

Table 29 Simulated welfare estimates for the four scenarios of quality change

Scenario (Quality Change) Welfare Change

Beach and Water: Good to Poor -$116.67

Beach: Good to Fair -$35.00

Water : Good to Fair -$35.36

Beach and Water: Good to Fair -$66.74

Beach: Good to Excellent $37.85

Water: Good to Excellent $38.29

Beach and Water: Good to Excellent $78.33

The estimates in Table 29 show that the per person welfare loss associated with a

decline in beach or water quality is approximately $35 for each attribute. The results

show that a combined decline in both beach and water quality would result in a welfare

loss of $67. The largest welfare loss of $117 is observed when combined beach and

water quality declines from the status quo (good) to poor.

The welfare gain associated with an improvement in beach or water quality (only)

was estimated to be approximately $38 each. The welfare gain when both attributes

improved (together) from the status quo to a quality level of excellent was estimated to be

approximately $78.
 Like previous examples, the aggregate values were estimated by multiplying the

mean welfare by the number of stopover visitors to Jamaica in 2007 (1,700,785). The

aggregate annual welfare change for improvements or declines in quality are shown in

Table 30 below.

Table 30 Aggregate welfare for changes in attribute quality.

Scenario Aggregate Welfare Change

Beach and Water (Good to Poor) $198,430,586 Loss

Beach (Good to Fair) $59,527,475 Loss

Water (Good to Fair) $60,139,758 Loss

Beach and Water (Good to Fair) $113,510,391 Loss

Beach (Good to Excellent) $64,374,712 Gain

Water (Good to Excellent) $65,123,058 Gain

Water and Beach (Good to Excellent) $133,222,489 Gain

The results in Table 30 show that considerable aggregate welfare values are

associated with changes in the quality of the beach and water quality attributes.

Aggregate welfare loss if both beach and water quality attributes decline from the status

quo to fair is estimated to be $114M per annum. Welfare loss when the attributes decline

to poor quality is show to be approximately $200M per annum. A welfare gain of $133M

is estimated from an improvement in the beach and water quality from the status quo.

167
5.12 Policy Application
As a reminder, the purpose of this stated choice study was to estimate the

recreational benefits or economic value associated with a quality change in the coastal

ecosystem attributes relevant to Jamaicas tourism product. This included estimating the

benefits associated with an improvement in quality as well the welfare loss associated

with a decline in quality of the key attributes. The access value associated with the

removal of Jamaica from the respondents choice set (for example a massive oil spill)

was also estimated. Information on the relative importance of the four key attributes to

the respondents valuation framework was also provided.

Based on the purposes highlighted above, the different pieces of information can

therefore be used by policy makers to establish that there is a significant consumer

surplus associated with a visit to Jamaica. The use of the stated choice framework also

allows policy makers to identify the consumer surplus associated with the ecosystem

quality of the tourist destination. This is because of the decision framework and payment

vehicle that was used in the stated choice survey. The study also allows for policy

makers to be aware of the welfare loss or gain from changes in quality. The access value

of $128 can be compared to other measures of welfare such as the WTP derived from the

companion contingent behavior study or the other previously mentioned Jamaican

revealed or stated preference studies.

The use of the part worth estimates (implicit prices) for determining how funds

can be allocated for management would be the next step in the policy process. Based on

the results discussed in the previous section, the key areas of concern for the major user

168
group appear to be beach quality and water quality. Resource managers can therefore

structure their management programs to address these priority issues. It should be noted

that beach quality with along with water clarity are attributes that are both dependent on

the health of other ecosystems. For example coastal water quality is impacted by the

level of environmental management in surrounding watersheds as well as land use and

agricultural practices. Likewise beach sand quality is affected by the health of the

adjacent coral reef and seagrass beds. Coral reef health is also dependent on good water

quality (low nutrients and suspended solids). That being said the average respondent may

not be fully aware of the importance (and inter-connectedness) of the other ecosystems to

the quality of the beach and water clarity attributes. However, the relative significance of

beach and water quality to the respondents valuation framework suggests that these

attributes are good proxies for lay-person measures of ecosystem health.

The findings from the model simulations can also be used in the policy decision

making process. Model I simulation results can be used as part of the cost benefit

analysis process when deciding what would be a feasible range of prices for a particular

management option (basic or advanced). This is synonymous with the policy steps that

would be taken as was outlined in the previous chapter (see flow diagram in figure 25).

The results from the Model I simulation exercise can provide information on the potential

decline in visitation as the price of the surcharge increases. Model II simulation results

could also be used as part of a cost benefit analysis (CBA) framework where the potential

welfare losses or gains could be compared to the cost of natural resource management

that would either preserve or improve current beach and water quality in Jamaica.

169
5.13 Summary of Findings Stated Choice Survey
A random airport intercept stated choice survey of visitors to Jamaica with a 90%

response rate was used to estimate the economic value (recreational benefits) of changes

in the quality of the Jamaican coastal ecosystem. Respondents were asked to consider

three choice scenarios where they had to decide on taking a hypothetical vacation to

Jamaica. Each choice question asked the respondent to consider three environmental

management options or a no trip (to Jamaica) option. An increase on the existing $10

tourism surcharge was used as the payment vehicle. The three management options were

associated with a decline, an improvement or maintaining current quality of the beaches,

swimming waters, coral reef and fish size and abundance.

A random utility model was estimated using both multinomial and mixed logit

techniques. The results show that the alternatives satisfy a priori expectations that a

decline in quality is least preferred and improved environmental quality is most desired.

The increases in the tourism surcharge had a negative and significant effect on

respondents, as was also expected. Annual aggregate welfare calculations (gains and

losses) show that there is a significant consumer surplus associated with a typical Jamaica

beach vacation.

The estimated coefficients and calculated implicit prices on the key environmental

attributes suggest that improvements to the quality of the beach and water quality are

very important to respondents valuation framework. Respondents were willing to pay

for improvements to beach and water quality and were conversely not willing to pay for

improvements to reef quality or fish abundance. Possible explanations for these

170
differences may be due to the fact that respondents may be less familiar with the

underwater quality attributes (coral reef health and fish abundance) as well as

heterogeneity of preferences among the sampled respondents.

Respondents prefer the advanced management option to basic management.

While model simulations confirm the downward sloping demand curve that is expected

with increases in the price of the hypothetical surcharge. Model simulations of changes

in beach and water quality show that simultaneous declines or improvements would result

in significant changes in aggregate welfare for visitors to Jamaica.

171
 CHAPTER 6

POLICY IMPLICATIONS

6.1 Policy Relevance

The results of this study are useful in the wider decision making process for coral

reef protection for vulnerable small island developing nations such as Jamaica. Given the

current threats to coral reefs and their associated ecosystems including the impact of

global climate change and ocean acidification, it is important that valuation estimates are

included in the policy process. This study is timely as it is the first time that a contingent

behavior model based on a dichotomous choice format with a canonical response

probability model has been used to predict the impact of increased taxes on tourist

visitation rates to the island of Jamaica. It is also the first stated choice (CE) study to be

conducted on coral reef ecosystem services in Jamaica. It therefore provides relevant

information on the overall consumer welfare as well as the potential losses (or gains)

from quality changes and the resultant impact on the tourism industry in Jamaica.

6.2 Use of Welfare Estimates in Resource Management Policy.

The welfare estimates derived from the CV and CE methods employed in this

study can be used to contribute to resource management policy in Jamaica. As discussed

previously (section 4.8) the type of welfare measure used is dependent on the economic

172
principles being applied to the CBA process. That is whether policy makers might adopt

either the Kaldor-Hicks potential compensation or the majority voting principles as part

of their decision making process. The findings from both studies suggest that either one

of these principles can be applied to the policy process for developing a sustainable

financing framework for resource management. Based on the policy goals of this

research the first step was to provide an estimate of the level of consumer surplus for the

coastal ecosystems associated with coastal tourism in Jamaica. Both studies show that

there is a relatively high consumer surplus associated with a Jamaican coastal vacation.

The next step after determining that there is significant welfare is to use this

information to make a decision about the equitable capture of a portion of this consumer

surplus. This capture of a portion of the welfare should then lead to the identification of a

sustainable source of funding for resource management. Both the contingent behavior

response models and simulated choice experiment probability model (Model I) provide

information on changes in visitation rate in response to different prices of tourism

surcharge. It should be noted that while it is possible to use the CE model parameter

estimates to provide direct behavioral interpretation when discussing utilities it is

however, not possible when discussing probabilities (Hensher et al., 2005 pg 341). As a

result of this I would recommend only using the canonical response models based on the

contingent behavior estimates for predicting the impact of increased surcharges on

visitation rates. In the absence of CB estimates the CE model results can however be

used to guide setting reasonable rates (levels) for environmental surcharges.

173
6.3 Does the Method of Welfare Estimation Matter?
This study shows that tourists visiting Jamaica appear to have a high consumer

surplus associated with a beach vacation to Jamaica. However, one could ask the

question, are the welfare estimations accurate? Firstly accuracy of welfare estimates is

a long debated topic and it is not the goal of this dissertation to contribute to the debate.

Secondly, the veracity of the reported estimates is also dependent on a host of other

factors including survey and bid design, sampling methodology and sample size. The

functional form of the RUM models chosen can also affect the parameter and welfare

estimates that are generated. These as well as other factors outside of the control of the

researcher can contribute to subtle or great variations in consumer surplus estimates.

However, is accuracy the ultimate goal and how is it determined? Or is it more

important to know how the estimates of welfare can be used as part of a wider policy

framework? I would argue for the latter. The table below shows a comparison of some

of the per person welfare estimates that were derived from the different methods

employed in this study. These include the non-parametric and parametric results for the

tourism and environmental tax as well as the choice experiment estimates. The table

below also contains direct estimates of the mean WTP from the canonical response

probability distributions. That is the calculated area under the curve and median WTP

(Pr {Yes} = 0.5) that were directly obtained from the response probability curves for each

model (tourism and environmental). The table also includes parametric estimates from a

non linear (log normal) regression. Details of the calculations of the graphical and

nonlinear estimates are presented in Appendix 3.

174
Table 31 Comparison of welfare estimates based on the type of survey and
econometric analysis

Survey Type and Econometric Analysis Tourism Environmental

Surcharge Surcharge
CB Non-Parametric Mean WTP $130 $165

CB Graphical Estimation Median WTP $137 $142

CB Graphical Estimation Mean WTP $151 $160

CB Linear Logistic Mean = Median WTP $192 $282

CB Log Normal Median WTP $95 $143

CB Log Normal Mean WTP $418 $528

CE MNL Mean Access Value $130

CE Mixed Logit Mean Access Value $128

The table above shows a range of values for the different types of WTP measures.

The non-parametric WTP values that can be assumed to be lower bound estimates show

consumer surplus estimates above $100 per person. The welfare values based on the

graphical estimation of the canonical response models also show welfare above $100 per

person. The table above also highlights the differences between estimates of WTP that

can occur depending on the econometric method or empirical assumptions made about

the distribution of the error terms. For example, as expected, there are significant

differences for estimated WTP between the linear and non-linear econometric estimates

(refer to the discussion in section 4.8). The comparatively high mean values for the log

normal models suggest a slight fat tail problem and confirms that the non-linear model

175
is assigning too much weight to WTP values in the upper tail of the distribution (Haab

and McConnell, 2002). However the median WTP estimates are clearly not as sensitive

to the distributional and functional form assumption and as a result show much lower

median WTP estimates when compared to the linear model (where mean and median are

assumed to be equal).

The estimated access values (MNL and MXL) from the stated choice survey are

also comparable surplus measures to the CB willingness to pay estimates. This is

because the contingent behavior scenario offers the respondent a choice between an

increase in cost and the option of not visiting Jamaica. The stated choice estimate of

access value is therefore analogous to the CB valuation scenario where the choice of a

Jamaican vacation is removed from the respondents choice set (this time because of

quality decline and price). It is notable that the CE welfare estimates were lower in

comparison to the CB environmental survey estimates.

The table above also shows the possible differences that can occur because of

differences in the type of valuation scenario and information provision (tourism versus

environmental tax), as well as the type of survey instrument (contingent behavior or

stated choice) and econometric model chosen by the researcher.

As stated previously, this study is not a benefit cost analysis which would require

the use of the mean in order to satisfy the Kaldor-Hicks potential compensation principle

(Hanemann and Kanninen, 1999, pg 330). Based on the possible differences in estimated

welfare highlighted above, any analysis and subsequent policy decisions should therefore

176
take into consideration the type of functional form, empirical assumptions that are chosen

as well as the overall policy goals of the study.

Although there are some clear differences with respect to the welfare estimates,

the primary policy goals of this study can still be achieved. These are: to estimate the

relative size of the consumer surplus (using the environmental CB and CE surveys) and

to use the models (CB survey) to predict respondent behavior to increases in surcharges

as part of a sustainable financing policy framework. The results of both stated preference

methodologies were able to provide the necessary information to achieve these goals.

The estimates of welfare loss and gain due to changes in attribute quality from the

stated choice models (including simulations) can also be used to inform policy makers of

how best to allocate funding for management of particular types of natural resources. For

example the data show that the beaches and swimming water quality are very important

to tourists and are therefore primary targets for management activity.

6.4 The actual impact of taxes on visitation rates

It is important to briefly discuss some subjective opinions on the use of the

canonical response model to predict visitation response to increases in surcharges. In my

opinion the response model provides higher than expected estimates of the potential

declines that might occur with corresponding tax increases. To illustrate my point, the

probability response model shows that a per person $2 increase as a result of an

environmental surcharge would result in approximately 1% decline per annum, while a

$10 increase would result in a 4.5%. One could however ask the question, would a $2

177
per person surcharge on a typical Jamaican vacation that, on average, costs $1600 per

person, really cause 1% of visitors to cancel their trip? Likewise, would an additional

$10 really cause an almost 5% decline in annual visitation? One could argue that the

response probability models provide conservative estimates of the probabilities of saying

yes to the surcharge. This could possibly be due to the inability to completely sort out

protest bids from the sample. In my opinion, where the models are probably most useful

are in the upper portions of the distribution ($100 or more) where one can demonstrate

the potential negative impact that revenue maximization could have on visitation rates to

the island (refer to figure 27).

6.5 Does the Type of Tax Matter?

There may be a number of reasons why a respondent may be motivated to respond

differently to different types of taxes. For example, it was clear from some of the

(written) debriefing comments from the CB surveys that there was a general lack of trust

that the government would use the tourism tax appropriately. In some cases those who

indicated a willingness to pay also suggested where they would like to see the money

being spent. These include areas such as: environmental management, employment

creation, road repairs and solid waste management. The findings of this study also

confirm that the type of institutional framework for resource management and how tax

revenues are allocated are both important to the respondents valuation context.

The data show that there are some differences between the probabilities of saying

yes to the two types of taxes (tourism and environmental). This suggests that the

178
label of the tax as well as the institutional mechanism for the provision of the good is

important to the respondents decision framework. The results of the contingent behavior

study support the assumption that when respondents obtain information on ecosystem

services, threats and management mechanisms as well as information on how their tax

dollars will be spent, they have a higher probability of saying yes to this type of tax than

those who did not receive any such information and were facing a different label for

the tax. Thus supporting the a priori expectations of the study (WTPTOUR < WTPENV).

Policy makers should therefore be careful about the label given to any revenue

generating instrument. This study shows that the type of label is important and that it is

also very critical that users are informed about how the funds will be allocated. It is also

important to demonstrate that the funds are indeed being used for the purposes specified.

How this is to be done is up to the implementers of the policy and this may include

appropriate use of signage and well as other public relations efforts.

For the stated choice survey, the results show that for those respondents who

indicated that they were willing to pay some type of surcharge the majority preferred the

advanced management option. This is expected given that the proposed management

scenarios also implied that the level of quality would have improved collinearly with

increased management. Expected utility will therefore be highest when the attribute

levels are at their maximum. The CE results suggest that providing the information on the

management activities and the expected outcomes of those management activities may

have influenced the respondents preference for a given option. This supports the

179
findings of the CB study that the label or type of tax and how the revenues are to be used

are important to the respondents decision framework.

6.6 How Should Revenues be Allocated?

Successful revenue generation is just the first step. The next step is the

appropriate allocation of funds in key areas of concern. As mentioned in section 6.3

above, the part worth estimates from the stated choice survey and model simulations can

be used to determine how funds should be allocated for resource management.

What are the key areas of concern for the user group? Based on the qualitative

findings from the contingent behavior study as well as the implicit price estimates and

simulations from the stated choice models, it appears that beach quality and water clarity

are very important to visitors to the island. However, as discussed in the previous chapter

these two attributes are linked to other key elements of the coastal ecosystem of a tropical

island such as Jamaica. So while the average tourist may not be very aware of the

benefits, threats and management options for coral reef management, it appears that

beach and water quality can serve as proxies for the other attributes associated with

coastal ecosystem health. High numbers of attributes can lead to increased complexity in

survey design and respondent fatigue. By using these two key attributes this can lead to

reducing survey complexity thus making the survey more respondent friendly. This

may be particularly useful in CE based designs where the target population may not be

very familiar with all of the key non-market or ecosystem attributes.

180
 We can also consider how revenues should be allocated for coastal resource

management in the Jamaican case. The findings suggest that information such as this can

be used by policy makers and resource managers for developing ecosystem based

management programs that will lead to preservation or improvements in beach and

coastal water quality. This may include activities such as watershed management,

pollution control, placing limits on coastal development as well as direct activities such

as beach restoration and solid waste management. These improvements in water and

beach quality may therefore result in an enhancement of the overall coastal tourism

product.

Based on the comments above and the general findings of this study, I would

recommend that a portion of the funds generated from the environmental user fees be

used to address the (direct) tourism related impacts to the environment as highlighted in

chapter 2 (section 2.6). This would most likely be implemented through the existing co-

management framework between relevant NGOs and central government (NEPA). The

remaining funds should therefore go towards wider environmental management issues.

This would include institutional strengthening and capacity building of other key

development and planning agencies. Funds could also be used to enhance the existing

tourism product by assisting in the development of a more sustainable tourism product.

For example shifting the emphasis from the current beach lust tourism model (sun,

sand and sea) toward wonder lust models (heritage, nature and adventure based

tourism), which typically have higher individual consumer surpluses associated with

them (Dixon et al., 2001).

181
 CHAPTER 7

CONCLUSIONS AND FINAL THOUGHTS

7.1 Final Summary of Findings

The findings of the contingent behavior study show that the label of the tax and

as well as the respondents awareness of the institutional mechanisms for environmental

protection and tourism are important to their decision framework. The findings show that

coastal zone management activities could be completely financed from the introduction

of a $2 per person environmental tax in addition to the existing tourism tax. The

potential negative impact on the annual visitation rate to Jamaica from the introduction of

this additional tax appears to be negligible ( 0.9%). However, any decline in visitation

rate could be mitigated by providing visitors with information on how tax revenues are

allocated. It is important that the provision of this information is complimented by

ensuring that the funds are indeed used for the purposes specified.

The findings of the choice experiment study confirm a priori expectations that

respondents utility is lowest with a decline in quality given low level environmental

management while utility is highest given advanced management and improved

environmental. The estimates of welfare from the models and simulations also confirm

that there is a significant consumer surplus associated with a typical coastal vacation in

182
Jamaica. Beach and water quality were shown to be important to the respondents

valuation context.

Both studies show that tourists have consumer surpluses that greatly exceed the

costs of coastal management that would prevent a decline in quality of the key coastal

attributes. The environmental tax CB model and CE simulations also suggest that

environmental user fees could be an attractive sustainable financing option. The findings

also seem to refute the oft touted concern that additional taxes would kill the tourism

industry in Jamaica. These findings are also supported by other studies that come to

similar conclusions that tourist spending is generally inelastic suggesting that there is

scope to raise taxes from tourists (Clarke and Ng, 1993). A study by Gooroochurn and

Sinclair (2005) showed that unlike other levies, tourism taxes can increase domestic

welfare since international tourists bear most of the welfare loss associated with higher

revenue. They used a computable general equilibrium analysis for Mauritius and found

that taxing tourism is relatively more efficient and equitable than levying other sectors.

Their study also found that a narrow policy that taxes the highly tourism-intensive

sectors, extracts significantly more revenue from tourists than a broader policy where all

tourism-related sectors are taxed.

The importance of coastal tourisms continued contribution to Jamaicas economy

rests upon the ability of key stakeholders to protect the coastal ecosystem that the

industry is so vitally dependent upon. In the absence of adequate government funding for

natural resource management, targeted taxes on major resource users of the coral reefs

and beaches such as tourists can generate income to support comprehensive management

183
of the ocean and coastal resources of Jamaica. This study demonstrates an approach that

could be used as part of the policy framework for resource protection and sustainable

management of important coastal ecosystems and natural resources in other countries

dependent on coral reef based tourism.

The welfare estimates presented in this study may be used in benefit transfer

studies to similar Caribbean islands or other coastal nations dependent on coral reef based

tourism. Any benefit transfers should take into consideration the possible differences

across countries with respect to environmental quality as well as institutional frameworks

governing coastal zone management and environmental protection.

7.2 Study Limitations and Recommended Future Work

The random airport sampling method was novel and provided a random selection

of a wide range of tourists however there were some sampling limitations that should be

highlighted. As a result of the sample frame chosen for both of these studies cruise ship

passengers were not targeted as part of the sample. This is primarily due to research

funding limitations as well as lack of accessibility to this population of resource users.

Based on the Jamaica Tourist Boards statistics, 1,179,504 cruise ship passengers visited

the island in 2007. This is a unique type of tourist who typically spends less than 24

hours on the island, with activities divided between beach use, shopping and dining. It

should also be noted that that the current regulation stipulates that this group be charged a

$2 tourism surcharge and I would recommend that any future stated preference surveys

on this topic should attempt to sub sample this user population.

184
 The study also explicitly targeted tourists visiting the island as the primary

recreational users of the coral reefs and beaches. Even though this included a small

percentage of Jamaicans who reside overseas, no data from local users of the resource

were collected for welfare estimation. This would require a different sampling approach.

One key difference would be the institutional framework and type of payment vehicle

that would be used as well as the utility theoretic framework governing the contingent

market. WTP may not be an appropriate welfare measure and locals may reject the idea

of paying additional taxes to preserve the environment. Conducting stated preference

surveys on local populations in developing countries will require a slightly different

approach to the study design. A paper by Whittington (1998) concluded that conducting

stated preference studies in developing countries can sometimes be simpler and less

expensive than in an industrialized country. He, however, goes on to state that this does

not mean that conducting CV surveys in developing countries is easy. Whittington also

reiterates in his paper that there are numerous issues that arise in CV work in developing

countries that demand careful attention in order to increase the probability that high-

quality results are obtained.

Another potential limitation with respect to sampling is the possible

representativeness of the sample. This again is related to the problem of access to the

sample population and the choice of the airport as the sampling location. Firstly, only

one week in the month of January was used to collect the data. This was mainly due to

funding limitations and the accessibility restrictions of the Montego Bay airport

management company. Ideally what would have been best was have sampling occasions

185
spread out over peak and off peak periods. The month of January was selected because

the winter tourism season traditionally begins in late December and ends in late spring.

The monthly tourist board data (JTB, December 2008) shows that the peak periods were

December, March and July. However stopover arrivals in January were shown to be

close to the annual average for 2008. Although sampling spread out over the year or

across seasons might have improved representativeness of the sample, if you compare the

demographic information collected from respondents (hotel location, nights spent etc)

with the JTB statistics for 2007 and 2008, it would suggest that the sample is fairly

representative. Of note, there may have been some oversampling of North American

respondents and this might be partially due to some instances where non native English

speakers refused to be sampled or their survey responses could not be used. This could

be addressed in any future survey effort by providing questionnaires in languages other

than English.

There may be some concern with regards to how random the method of

intercepting respondents actually was. Although there was the potential for subjective

interviewer bias (for example approaching persons who seemed friendly), the method of

selecting every nth person per row (given the density of said row) introduces randomness

into the data collection process. This can be considered an onsite simple random

sampling method. In the absence of a pre-determined list of names and addresses this

method was deemed to be most appropriate. This method can be compared to studies

where persons entering a state park have been randomly pulled over by park rangers

(every nth car) and asked to participate in a survey (Leeworthy, 1996). Of course the

186
onsite sampling nature of the survey efforts precludes using data from potential visitors to

the island, however based on the rationale that respondents were to have been familiar

with the attributes of a Jamaica vacation this justified choosing this method of data

collection.

It should be noted however, that because of the potential issues of non seasonal

sampling as well as the absence of a purely random sample, care must be taken when

extrapolating the findings of the study to the general population. One possible

recommendation could be to collect similar demographic data from respondents across

other seasons (winter, summer, spring break) and then use this information to weight the

data.

One final possible sampling limitation of this study is the fairly modest sample

size that was used to derive the econometric estimates. The sample size may influence

the robustness of the econometric estimates. The sample size used for this study should

not be viewed as a critical drawback however as other studies have used similar and even

smaller sample sizes to provide useful policy information (Dixon et al., 1993; Parsons

and Thur, 2008).

With respect to the survey design there are also some limitations that should be

highlighted. For the contingent behavior surveys the current design of the payment

scenarios did not allow for theoretical validity tests of the hypothesis that ecosystem

information provided to the respondent will influence their WTP. The correct treatment

for this would have been to present the identical choice scenario in the environmental

survey version to two groups of respondents. However one group would have been

187
provided with the relevant (objective) ecosystem information while another control group

would not have been treated with the information. This was indeed done between the

tourism and environmental survey versions however the slightly different choice

scenarios (and labels of the payment vehicles) that were presented did not allow for the

theoretical testing of this hypothesis. Depending on the type of econometric model used

(parametric, semi-parametric or non parametric) a comparison of the means show that the

level of statistical significance between the types of taxes varies. The raw frequency

response data on the other hand suggests that one has to be cautious about placing greater

emphasis on the label of the tax. Again however, based on some of the written debriefing

comments (that were asked after the WTP question) respondents seem to be very

responsive to the type of institutional framework or in other words how their tax dollars

will be spent. Thus my recommendation that the consideration be given to what any

additional tax is called (labeled) is therefore valid. Dixon et. al. (2001, pg 28), in their

paper on the tourism and the environment in the Caribbean, also provide this

recommendation.

There may have also been some issues with the bid design for the CB survey.

The bids range from $10 to $990 and may have resulted in some yea-saying for the

higher bid levels (fat tail problem). Comparison of the results of this dissertation with

other studies would imply however that that the WTP (welfare) values are not necessarily

unrealistic. It should be also noted that there were some issues with non-monotonicity in

the environmental survey version and this may have been avoided with a larger sample

size.

188
 For the stated choice study there were also some limitations with respect to the

study design. Firstly, the intentional collinear movement of attribute levels with the

increased management levels precluded estimating the implicit prices associated with

each attribute level (fair, good, excellent). The relatively modest sample size and the

number of attributes and levels may have also resulted in the loss of degrees of freedom

in the MNL and ML analyses. Any future studies should aim for a larger sample size,

use fewer attribute levels and also quantitatively define the attributes. That is instead of

using qualitative attributes such as poor, fair, good and excellent the respondent should

face attributes such as; 10, 20, 50ft wide beaches, varying percentage of live coral cover

(10,40, 50%) and actual numbers of fish seen per snorkel tour among others (Thur, 2003).

The use of photographs or other audiovisual aids could also be considered for similar

studies. Wieglus et al. (2003) incorporated the use of underwater video footage in their

survey of SCUBA divers in Israel. The use of pictures and audio visual aids in stated

preference studies is not without controversy but this could be considered for future

studies.

7.3 From Policy Analysis to Policy Implementation

What are the challenges to creating sustainable financing for resource

management in a small island developing state? What are some of the methods that that

can be used to provide answers to address these challenges? How can one integrate any

of these methods into policy development and implementation? These are three

overarching questions that guided the motivation for this research and it is hoped that the

189
findings here have helped to provide some answers to the first two questions. I will

attempt to provide an answer for the third question in the paragraphs below.

As was highlighted in the introductory chapters, budgetary constraints are a major

limiting factor in providing adequate environmental management in a small island

developing state (SIDS) like Jamaica. Given this reality one would assume that there

would be a high level of support for the implementation of mechanisms to capture a

portion of the consumer surplus associated with Jamaicas natural resources. In reality,

there is significant resistance from the tourism sector and even within some sections of

the Government to the implementation of this type of economic instrument.

The tourism industry in the Caribbean is already heavily subsidized through

various means such as airline subsidies, tax waivers to hotel chains and reduced

regulations to coastal development (Ramsaran, 1995). Sectors of the Jamaican tourism

and hospitality industry currently have access to the Tourism Enhancement Fund (TEF)

created from the existing $10 per person surcharge. However tourism industry

stakeholders would strongly object to any suggestion that an additional environmental

surcharge be levied on visitors to the island. Whenever this idea has been raised in the

past the key stakeholders make the claim that such measures would ruin the Jamaican

tourism sector and result in visitors going elsewhere. The findings of this study (and

others) can however be used to refute these claims (Hiemstra and Ismail, 1992;

Gooroochurn and Sinclair, 2005).

How then can one move from policy recommendation to implementation?

Incorporation of the findings into the existing Protected Areas System Master Plan

190
(PASMP) as was discussed in chapter 2 (section 2.4) is one such avenue for

implementation. However as was highlighted this process has been ongoing since prior

to 2000 and frankly does not have the type of prominence or profile that other sectors

such as education, energy and tourism seem to have.

One suggestion towards facilitation of the implementation of these policy

recommendations is to use the existing TEF legislation and mechanism of revenue

collection that is already in place. Environmental management agencies, both

governmental and NGOs should use the information presented in this dissertation or

information from a similar and more extensive study to lobby the central government and

relevant agencies such as the Ministries of Finance and Tourism to implement

environmental surcharges. The environmental agencies would also be in the best position

to present the actual costs of natural resource management for the island, thus

contributing to the CBA. The results of the valuation and CBA exercise would contribute

to the process of setting an appropriate price for the environmental user fee. These costs

should include both coastal and terrestrial environmental management given the islands

small size and relatively narrow coastal zone as well as the fact that activities in the

mountain ridges can impact the reef. The agencies would also be able to use the results

of this (or a similar) stated choice study to determine how best to allocate resources based

on the preferences of the user population (local or foreign).

The implementation of a system of user fees as recommended above would

therefore make funds available not only to the marine parks but to the wider protected

areas and national park system in Jamaica. These would include the Blue and John

191
Crow Mountain national parks as well as protected areas such as Portland Bight and the

Cockpit Country. This is especially so because almost all of these parks and protected

areas have no real entry point that would facilitate the collection of fees. This was also

highlighted in chapter 2 where the marine parks are located within and around tourism

locations. Capture of the consumer surplus associated with the countrys natural

resources must therefore occur at the ports of entry into the island (Dixon et al., 2001).

The capture of the consumer surplus associated with the natural resources of

Jamaica is the underlying rationale for using this kind of revenue instrument. Thus is

based on the assumption that capture of some of that surplus from the tourism industry

stakeholders such as hotel owners would be hugely controversial and practically

impossible. Even though it may be initially unpopular, welfare capture from tourists may

be more feasible in the long run. The fact that legislation already exists to collect a user

fee from visitors to the island should make it easier for the additional adjustments to be

made (such as providing information on how funds will be used). This would facilitate

the goals of providing funding environmental protection while facilitating sustainable

economic growth and development.

192
APPENDICES

193
Appendix 1 Descriptive Statistics

Contingent Behavior Survey Descriptive Statistics

Age

The figure below shows the frequency distribution of the age of respondents who

were surveyed. The median age was observed to be 42 years old.

36
Series: AGE
32 Sample 1 352
28 Observations 341

24 Mean 42.02639
Median 42.00000
Frequency

20
Maximum 80.00000
16 Minimum 19.00000
Std. Dev. 14.25779
12 Skewness 0.306183
Kurtosis 2.213341
8
4 Jarque-Bera 14.12061
Probability 0.000859
0
20 30 40 50 60 70 80
Age

Frequency distribution of Age

The figure shows that while the median age is 42 years old, tourists with a wide

range of ages visit the island. The youngest valid respondent was 19 years old while the

oldest respondent across both survey types was 80 years old.

194
Household Income

Median household income was recorded at US$90,000 per annum and the data

suggests that majority of respondents earned 200,000 per annum or lower. Figure 6

below shows the frequency distribution.

70
Series: HHINCUSD
60 Sample 1 352
Observations 322
50
Mean 123629.5
Median 90000.00
Frequency

40
Maximum 694400.0
Minimum 19000.00
30
Std. Dev. 104022.5
Skewness 2.077028
20
Kurtosis 8.714681
10 Jarque-Bera 669.6764
Probability 0.000000
0
0 125000 250000 375000 500000 625000

Household Income (US$)

Frequency distribution of household income.

195
Nights in Jamaica

The median number of nights spent in Jamaica was 7 days (one week). There are

a few outliers where persons spent a month or longer. These may represent individuals

with second homes in Jamaica, or persons with relatives that live in Jamaica.

200
Series: NIGHTSJ
Sample 1 352
160 Observations 348

Mean 8.025862
120 Median 7.000000
Frequency

Maximum 90.00000
Minimum 0.000000
80 Std. Dev. 6.842141
Skewness 6.493468
Kurtosis 66.67038
40
Jarque-Bera 61227.38
Probability 0.000000
0
0.0 12.5 25.0 37.5 50.0 62.5 75.0 87.5
Nights in Jamaica

Frequency distribution of nights spent in Jamacia.

196
Travel Cost

The figure below shows the frequency distribution for the per person travel cost

reported by respondents.

60
Series: TCSTPERS
Sample 1 352
50
Observations 328

40 Mean 1566.813
Median 1483.750
Frequency

Maximum 4475.000
30
Minimum 8.333300
Std. Dev. 784.1560
20 Skewness 0.543350
Kurtosis 3.403352
10
Jarque-Bera 18.36263
Probability 0.000103
0
0 1000 2000 3000 4000
Travel Cost Per Person

Frequency distribution of per person travel cost.

Median per person travel cost was observed to be $1484. Using the median

nights spent in Jamaica this cost translates to a median per person daily cost of $212.

197
Stated Choice Survey Descriptive Statistics
Age

The figure below outlines the frequency distribution of the age of the stated choice survey

sample. Median age was observed to be 39 years old while the oldest respondent was

observed to be 74 years old.

16
Series: AGE
Sample 1 174
Observations 172
12
Mean 41.80233
Median 39.50000
Frequency

Maximum 74.00000
8
Minimum 19.00000
Std. Dev. 14.07816
Skewness 0.265170
4 Kurtosis 2.025235

Jarque-Bera 8.825238
Probability 0.012123
0
20 30 40 50 60 70
Age

Frequency distribution of respondent Age.

198
Household Income

The figure below shows the frequency distribution of household income for the

stated choice survey sample. The median income for this sample was also $90,000 and

shows that the sample is similar to the contingent behavior survey sample.

32
Series: HHINCUSD
28 Sample 1 174
Observations 165
24
Mean 121522.0
20
Median 90000.00
Frequency

16 Maximum 400000.0
Minimum 20000.00
12 Std. Dev. 90978.48
Skewness 1.334054
8 Kurtosis 4.381789

4 Jarque-Bera 62.06848
Probability 0.000000
0
100000 200000 300000 400000
Household Income (US$)

Frequency distribution of respondent Household Income.

199
Nights in Jamaica

The figure below shows the frequency distribution of the nights spent in Jamaica

by respondents. The median number of nights spent was observed to be 7 nights. Again

like the previous sample there were a few individuals who spent over twenty days in the

island. The majority of respondents were observed to spend 7 nights or less visiting

Jamaica.

80
Series: NIGHTSJ
70 Sample 1 174
Observations 174
60
Mean 7.063218
50
Median 7.000000
Frequency

40 Maximum 31.00000
Minimum 0.000000
30 Std. Dev. 3.809256
Skewness 2.499756
20 Kurtosis 13.53203

10 Jarque-Bera 985.4115
Probability 0.000000
0
0 5 10 15 20 25 30
Nights in Jamaica

Frequency distribution of nights spent in Jamaica.

200
Travel Cost

The figure below shows the frequency distribution of the reported travel expenses

of respondents. Median per person travel costs was observed to be $1,750. This again

shows that persons pay relatively high amounts for their vacation to the island.

50
Series: PERTCST
Sample 1 174
40 Observations 169

Mean 1630.416
30 Median 1750.000
Frequency

Maximum 5000.000
Minimum 0.000000
20 Std. Dev. 835.5104
Skewness 0.989749
Kurtosis 6.030338
10
Jarque-Bera 92.25538
Probability 0.000000
0
0 1000 2000 3000 4000 5000
Travel Costs (US$)

Frequency distribution of per person travel cost.

201
Appendix 2 Non Parametric Results

Raw Response Frequency Data

Tourism Tax
BID (t) Yes No % Yes
10 31 5 0.861111
50 26 12 0.684211
90 25 24 0.510204
490 7 31 0.184211
990 0 11 0

Environmental Tax
BID Yes No % Yes
10 41 3 0.9318182
50 23 13 0.6388889
90 33 11 0.75
490 9 32 0.2195122
990 1 15 0.0625

Hypothesis test (t-test)

H0: WTPTOUR =WTPENV; H1: WTPTOUR<WTPENV

165.15 130.07
4.28
41.58 25.50

Critical t value = 1.96. Test statistic is greater than the critical value therefore we can

reject the null hypothesis that WTPTOUR =WTPENV.

202
Appendix 3 Parametric Results

CONTINGENT BEHAVIOR SURVEY

Tourism Tax
Model I
Variable Coefficient Std. Error z-Statistic Prob
CONSTANT 1.075161 0.224334 4.792679 0
BID -0.0057 0.001061 -5.37106 0

Model II
Variable Coefficient Std. Error z-Statistic Prob Mean
CONSTANT 1.30092325 1.04091912 1.249783 0.2114 1
BID -0.00609899 0.00111791 -5.455694 0 211.287
NIGHTSJ -0.05123786 0.04281843 -1.196631 0.2315 8.1930
AGE -0.01569732 0.01411148 -1.112379 0.266 43.0351
EDUC 0.28438474 0.26865158 1.058563 0.2898 2.9240
GENDER 0.80791054 0.39388883 2.051113 0.0403 0.4561
HHINC0000 -0.00664278 0.00204 -0.324896 0.7453 12.5708
US_CAN -0.18063731 0.51866166 -0.348276 0.7276 0.8480

Environmental Tax
Model I
Variable Coefficient Std. Error z-Statistic Prob
CONSTANT 1.54574307 0.24205 6.386051 0
BID -0.00546248 0.000877 -6.231403 0

Model II
Variable Coefficient Std. Error z-Statistic Prob Means
CONSTANT 0.786658 1.048337 0.739617 0.4595 1
BID -0.00558 0.000897 -6.219599 0 232.762
NIGHTSJ 0.010091 0.033884 0.296041 0.7672 7.9116
AGE 0.00113 0.013868 0.102202 0.9186 41.1381
EDUC 0.305714 0.266725 1.144129 0.2526 2.8674
GENDER -0.25348 0.378352 -0.667797 0.5043 0.4917
INC0000 -0.00786 0.00000184 -0.427605 0.6689 12.1586
US_CAN 0.00609 0.605144 0.012724 0.9898 0.8729

203
Likelihood Ratio tests (2)

H0: 3=4=5= 6=7=8 = 0

LR = 2[ln Lr ln Lu] Chi Squared Critical Value = 12.59

LR Tourism6df = 10.62; Cannot Reject Null Hypothesis

LR Environmental6df = 2.28; Cannot Reject Null Hypothesis

Hypothesis test (t-test)

H0: WTPTOUR =WTPENV; H1: WTPTOUR<WTPENV

t test for Model I

282.97 188.60
10.87
39.78 35.66

t test for Model II

282.27 191.67
10.63
38.92 33.69

Critical t value = 1.96. For both tests the t statistic is greater than the critical value

therefore we can reject the null hypothesis that WTPTOUR = WTPENV.

Direct Graphical WTP Estimation Canonical Response Model

1 exp
Pr
1 exp

Mean WTP is calculated by using estimated probabilities for each bid amount to calculate

the area under the curve bounded by $0 and $1000.

204
 That is: Mean WTP, Pr

Median WTP (C*) is taken directly from the response probability curve

That is: C* = (Prob (Yes) = 0.5

Tourism Survey
1+ exp() = 4.21869
1+ exp()/
BID 1+ exp( +A) 1+ exp( +A) %Yes %No
1 4.238381 0.995354 100% 0.5%
2 4.258192 0.990723 99% 0.9%
10 4.421108 0.954216 95% 4.6%
50 5.366335 0.78614 79% 21.4%
1000 1434.629 0.002941 0% 99.7%
Median 137.3251 8.437384 0.50 50% 50.0%
Mean 151.3218 9.100167 0.463584 46% 53.6%

Prob(YesTourismTax)
1.0
0.9
0.8
0.7
Prob(YesTax)

0.6
0.5
0.4
0.3
0.2
0.1
0.0
1 5 10 50 100 200 300 400 500 600 700 800 900 1000
TourismSurcharge($US)

Linear probability response curve, Tourism Tax

205
Environmental Survey
1+ exp() = 5.829270159
1+ exp()/
BID 1+ exp( +A) 1+ exp( +A) %Yes %No
1 5.856286 0.995387 100% 0.5%
2 5.88345 0.99079 99% 0.9%
10 6.106335 0.95463 95% 4.5%
50 7.382937 0.78956 79% 21.0%
1000 1279.3882 0.004556 0% 99.5%
Median 141.9109 11.65854083 0.499999978 50% 50.0%
Mean 160.2611 12.80753372 0.455143846 46% 54.5%

Prob(YesTax)
1.0
0.9
0.8
0.7
Prob(YesTax)

0.6
0.5
0.4
0.3
0.2
0.1
0.0
1 5 10 50 100 200 300 400 500 600 700 800 900 1000
EnvironmentalTax(US$)

Linear probability response curve, Environmental Tax

206
Combined Data

Model I
+--------+--------------+----------------+--------+--------+----------+
Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]| Mean of X|
+--------+--------------+----------------+--------+--------+----------+
---------+Characteristics in numerator of Prob[Y = 1]
Constant| .96925138 .72688958 1.333 .1824
ENVTAX | .48536463 .26250642 1.849 .0645 .51420455
NIGHTSJ | -.02105504 .02113149 -.996 .3191 8.04829545
AGE | -.00750950 .00960890 -.782 .4345 42.0596591
EDUC | .23587571 .18983271 1.243 .2140 2.90340909
FEMALE | .26786291 .26485426 1.011 .3118 .47443182
INC0000 | -.01009641 .01366428 -.739 .4600 12.3588375
US_CAN | -.08689966 .38058407 -.228 .8194 .86079545
BID | -.00561781 .00067872 -8.277 .0000 222.329545

Model II
+--------+--------------+----------------+--------+--------+----------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]| Mean of X|
+--------+--------------+----------------+--------+--------+----------+
---------+Characteristics in numerator of Prob[Y = 1]
Constant| 1.29422224 1.58376084 .817 .4138
ENVTAX | .53422534 .26610141 2.008 .0447 .51420455
NIGHTSJ | -.01914727 .02263919 -.846 .3977 8.04829545
AGE | -.01156092 .00991350 -1.166 .2435 42.0596591
EDUC | .23517835 .19760160 1.190 .2340 2.90340909
FEMALE | .28658176 .27053254 1.059 .2895 .47443182
INC0000 | -.00737338 .04653294 -.158 .8741 12.3588375
US_CAN | -.04940140 .39098724 -.126 .8995 .86079545
BID | -.00578508 .00069931 -8.273 .0000 222.329545
W50 | -.45320995 .58852071 -.770 .4413 .20170455
W100 | .24802661 .97183270 .255 .7986 .32670455
W_200 | .27145852 .77977636 .348 .7277 .55113636
W300 | -.42413734 .54483412 -.778 .4363 .75852273
W301 | -.71896562 1.00331275 -.717 .4736 .07954545

207
Model III

+--------+--------------+----------------+--------+--------+----------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]| Mean of X|
+--------+--------------+----------------+--------+--------+----------+
---------+Characteristics in numerator of Prob[Y = 1]
Constant| .90602626 .84576897 1.071 .2841
ENVTAX | .50600753 .28234052 1.792 .0731 .51420455
NIGHTSJ | -.02318250 .02168217 -1.069 .2850 8.04829545
AGE | -.00688282 .00979852 -.702 .4824 42.0596591
EDUC | .22533603 .19378498 1.163 .2449 2.90340909
FEMALE | .27554130 .26587311 1.036 .3000 .47443182
INC0000 | -.00933169 .01381716 -.675 .4994 12.3588375
US_CAN | -.08065277 .38156249 -.211 .8326 .86079545
BID | -.00559557 .00068261 -8.197 .0000 222.329545
MONTEGO | .01912083 .43993323 .043 .9653 .24431818
OCHORIOS| -.01691854 .45240124 -.037 .9702 .23295455
NEGRIL | .16286331 .45263019 .360 .7190 .23863636
RUNAWAYB| -.06190905 .60496115 -.102 .9185 .06818182
OTHERLOC| .11272394 .48258189 .234 .8153 .13636364

208
Log Linear Probit Models

Tourism Tax
+--------+--------------+----------------+--------+--------+----------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]| Mean of X|
+--------+--------------+----------------+--------+--------+----------+
---------+Index function for probability
Constant| 2.67374829 .72322585 3.697 .0002
NIGHTSJ | -.02172561 .02250973 -.965 .3345 8.19298246
AGE | -.00650419 .00819639 -.794 .4275 43.0350877
EDUC | .11278855 .15806777 .714 .4755 2.92397661
GENDER | .36643282 .22457924 1.632 .1028 .45614035
INC0000 | -.00384808 .01258550 -.306 .7598 12.5708234
US_CAN | -.02530216 .32016444 -.079 .9370 .84795322
LOGBID | -.58062932 .08737855 -6.645 .0000 4.44088157

Median WTPTOUR = $94.81, SE = 18.9336232 ($37.11)

Mean WTPTOUR = $417.79, SE = 87.6131621 ($171.72)

Environmental Tax
+--------+--------------+----------------+--------+--------+----------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]| Mean of X|
+--------+--------------+----------------+--------+--------+----------+
---------+Index function for probability
Constant| 2.56458595 .74724693 3.432 .0006
NIGHTSJ | .00845054 .02044733 .413 .6794 7.91160221
AGE | .00440309 .00814656 .540 .5889 41.1381215
EDUC | .17397988 .15753436 1.104 .2694 2.86740331
GENDER | -.14002224 .22070645 -.634 .5258 .49171271
INC0000 | -.01134487 .01168563 -.971 .3316 12.1585635
US_CAN | -.03812395 .36381797 -.105 .9165 .87292818
LOGBID | -.61883124 .08423926 -7.346 .0000 4.44459723

Median WTPENV = $143.05, SE = 27.8018570 ($54.49)

Mean WTPENV = $527.87, SE = 106.713030 ($209.16)

209
STATED CHOICE SURVEY

Model I - Basic
+---------------------------------------------+
| Start values obtained using MNL model |
+--------+--------------+----------------+--------+--------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]|
+--------+--------------+----------------+--------+--------+
FEE | -.01336514 .00200730 -6.658 .0000
ASCLOW | -.31263568 .16133172 -1.938 .0526
ASCBASIC| 1.67247443 .20893544 8.005 .0000
ASCADV | 2.16196966 .28192413 7.669 .0000

+---------------------------------------------+
| Random Parameters Logit Model |
| Replications for simulated probs. = 500 |
| Halton sequences used for simulations |
| Number of obs.= 522, skipped 0 bad obs. |
+---------------------------------------------+
+--------+--------------+----------------+--------+--------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]|
+--------+--------------+----------------+--------+--------+
---------+Random parameters in utility functions
ASCLOW | -.69890865 1.09299962 -.639 .5225
ASCBASIC| 1.03924879 1.28077793 .811 .4171
ASCADV | 2.36201172 .35682579 6.620 .0000
---------+Nonrandom parameters in utility functions
FEE | -.01535037 .00268566 -5.716 .0000
---------+Derived standard deviations of parameter distributions
NsASCLOW| 1.21654136 1.78585123 .681 .4957
NsASCBAS| 3.41354933 3.87600323 .881 .3785
NsASCADV| .07973216 1.53917073 .052 .9587

Simulated Welfare Estimates

===============================================================================
Variable Mean Std.Dev. Minimum Maximum Cases Missing
===============================================================================
------------------------------------------------------------------------------
All observations in current sample
------------------------------------------------------------------------------
DECLINE | 96.6149 110.786 -3.63580 599.633 3456 0
IMPROVE | -22.3901 134.734 -142.836 650.248 3456 0
ACCESS | 128.211 138.946 .253262E-02 817.094 3456 0

210
Model I Socioeconomic Interactions
+---------------------------------------------+
| Discrete choice (multinomial logit) model |
+--------+--------------+----------------+--------+--------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]|
+--------+--------------+----------------+--------+--------+
ASCLOW | -.67001417 .86733058 -.773 .4398
ASCBASIC| 2.56570098 .71127732 3.607 .0003
ASCADV | 3.34716647 .75844111 4.413 .0000
FEE | -.01441551 .00213418 -6.755 .0000
INCLOW | .00422378 .00117255 3.602 .0003
INCBASIC| -.00224325 .00084967 -2.640 .0083
INCADV | -.00596536 .00290774 -2.052 .0402
AGELOW | -.00320706 .00630708 -.508 .6111
AGEBASIC| -.01008245 .00622299 -1.620 .1052
AGEADV | -.02306665 .02173099 -1.061 .2885
EDUCLOW | -.20423347 .12844012 -1.590 .1118
EDUCBASC| .12066704 .11902218 1.014 .3107
EDUCADV | .01001820 .41473110 .024 .9807

+---------------------------------------------+
| Random Parameters Logit Model |
| Replications for simulated probs. = 500 |
| Halton sequences used for simulations |
| Number of obs.= 522, skipped 0 bad obs. |
+---------------------------------------------+
+--------+--------------+----------------+--------+--------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]|
+--------+--------------+----------------+--------+--------+
---------+Random parameters in utility functions
ASCLOW | .16263369 .89071311 .183 .8551
ASCBASIC| 1.71190390 1.11888244 1.530 .1260
ASCADV | 2.80227971 .44157947 6.346 .0000
---------+Nonrandom parameters in utility functions
INCLOW | .02452359 .01071592 2.289 .0221
INCBASIC| -.00355683 .00252508 -1.409 .1590
INCADV | -.00140090 .00065643 -2.134 .0328
FEE | -.01558860 .00289159 -5.391 .0000
---------+Derived standard deviations of parameter distributions
NsASCLOW| 1.33364992 1.73187131 .770 .4413
NsASCBAS| 3.60610697 4.24740603 .849 .3959
NsASCADV| .11816074 1.81147146 .065 .9480

211
Model II Environmental Attributes
+---------------------------------------------+
| Discrete choice (multinomial logit) model |
+--------+--------------+----------------+--------+--------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]|
+--------+--------------+----------------+--------+--------+
BEACH | .41517289 .13832926 3.001 .0027
WATER | .41982022 .14287554 2.938 .0033
REEF | -.11172894 .11585594 -.964 .3349
FISH | -.00396942 .13453134 -.030 .9765
FEE | -.00922228 .00206321 -4.470 .0000

+---------------------------------------------+
| Random Parameters Logit Model |
| Replications for simulated probs. = 500 |
| Halton sequences used for simulations |
| Number of obs.= 522, skipped 0 bad obs. |
+---------------------------------------------+
+--------+--------------+----------------+--------+--------+
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]|
+--------+--------------+----------------+--------+--------+
---------+Random parameters in utility functions
BEACH | .73770038 .20498825 3.599 .0003
WATER | .57536289 .18538588 3.104 .0019
REEF | -.29535361 .16015568 -1.844 .0652
FISH | -.08475066 .17600110 -.482 .6301
---------+Nonrandom parameters in utility functions
FEE | -.01228622 .00312019 -3.938 .0001
---------+Derived standard deviations of parameter distributions
NsBEACH | .00933970 .34338507 .027 .9783
NsWATER | .72855047 .23678259 3.077 .0021
NsREEF | .00902643 .23673369 .038 .9696
NsFISH | .83330818 .31547051 2.641 .0083

Simulated Implicit Prices

===============================================================================
Variable Mean Std.Dev. Minimum Maximum Cases Missing
===============================================================================
------------------------------------------------------------------------------
All observations in current sample
------------------------------------------------------------------------------
BEACH_IP| 60.0218 .745323 57.5842 62.6388 3000 0
WATER_IP| 45.1796 58.1395 -144.965 249.328 3000 0
REEF_IP | -24.0599 .720324 -26.4157 -21.5306 3000 0
FISH_IP | -8.78562 66.4993 -226.271 224.717 3000 0

212
Appendix 4 Sample Copies of Surveys

TOURISM TAX

CONTINGENT BEHAVIOR SURVEY

213
 Tourism and Beach Recreation Survey,
Jamaica

Please place your completed questionnaire in the envelope provided

and return it to the person or address indicated.

Environmental Foundation of Jamaica

214
Hello and Good Day,

I am a Jamaican PhD student asking for your help in a study of tourism and beach recreation. This study is
part of a research project that aims to better understand how visitors to Jamaica value their vacation
experience.
You were randomly chosen to participate in this PhD research project. Your help is voluntary and your
answers are completely confidential. This survey was approved by the University of Delawares Human
Subjects Review Board so your anonymity is assured. This means, none of the results of this survey can be
linked to particular individuals such as yourself.

Please note this is not a government funded survey, however your thoughts and opinions are important in
helping us understand how we can assist tourism officials better manage the tourism industry in Jamaica.

The survey will take approximately 15 minutes to complete. Your time and cooperation in completing this
questionnaire are greatly appreciated.

After completing the survey please place the booklet in the envelope provided. Be sure to seal the envelope
to ensure confidentiality.

If for some reason you choose not to respond, please let us know by returning the blank questionnaire in the
enclosed self-addressed envelope.

Thank you very much for your participation.

Sincerely

..
PETER EDWARDS
PhD Student,

University of Delaware
College of Marine and Earth Studies
Newark, DE 19716, USA

215
START HERE

1. Including your most recent trip, how many times have you visited Jamaica? (Check one)

1
2
3
4
5 or more visits
I am a Jamaican living abroad

2. Please indicate the number of trips for pleasure (vacations) including this one you have taken
outside of your country of residence in the last five years.

1
23
45
6 10
11 or more

3. Where did you stay on your most recent visit to Jamaica?

(Check one)
Hotel/Resort
Small Hotel/Inn
Bed and Breakfast
With Friends and Family

Other _______________________________

216
4. Where was your hotel or accommodation located?

Ocho Rios
Runaway Bay
Montego Bay
Negril
Whitehouse

Other _______________________________

I am not sure

5. Please check all the activities you engaged in on your most recent trip to Jamaica.

Sun Bathing
Swimming
Scuba Diving
Glass Bottom Boat Tours
Reef Snorkelling
Para Sailing
Pleasure Cruises
Tours (Nature etc.)
Golfing
Shopping
Business Meetings
Conference

Other ____________________________

Other ____________________________

217
6. How satisfied were you with your overall vacation experience? (Check one)

Very satisfied
Somewhat satisfied
Neither satisfied nor dissatisfied
Somewhat dissatisfied
Very dissatisfied

7. The factors listed below may have influenced your level of satisfaction on your vacation. Please
rate all the factors shown below. (Check one box per factor).

FACTORS Very Good Good Fair Poor Very Poor Not Sure
Hotel Amenities
Customer Service
Food
Night Life
Safety and Security
Shopping
Tours and Attractions
Friendliness of Locals

What did you like most about your trip to Jamaica?

___________________________________________________________

What did you like the least about your trip to Jamaica?

____________________________________________________________

218
 Now we would like you to give us your opinion on the quality of the
beaches, marine and underwater life you experienced on your
vacation to Jamaica.

8. How would you rate the overall quality of the beaches and swimming waters you experienced in
Jamaica?
(Check one box in each row).

Very Good Good Fair Poor Very Poor Not Sure

Beach Sand Quality
Beach Cleanliness
Water Clarity

If you participated in any underwater viewing activities such as snorkelling, glass bottom boating or
SCUBA diving while on your vacation please answer the question below. If you did not please proceed to
the next page.

9. How would you rate the overall (or average) quality of the underwater life that you observed
during these activities?

Very Good Good Fair Poor Very Poor Not Sure

Coral Reef Health*
Fish Abundance
Other Sea Creatures**
Underwater Visibility

*Coral reef health did you see mostly live coral (Good)? Was it overgrown with green plants (Poor)?
**Other sea creatures Did you see many other animals such as lobsters, crabs, eels and starfish?

219
10. What would you say is your general level of awareness about the following issues affecting the
planet as a whole?
(Check one box per issue)

ISSUES Very Aware Somewhat Very Little Not at All

Air Pollution
Water Pollution
Solid Waste Pollution
Nature Conservation
Climate Change
Deforestation
Coral Reef Protection
Population Growth

Now we ask you for some details about your most recent trip to Jamaica.

11. How many nights did you spend in Jamaica? _________ nights

12. With whom did you travel on this vacation?

(check all that apply)

Alone
Spouse/Partner
Family
Friends
Business Associates

Other______________________________

Please write below

Total Number of Persons (including yourself)__________

220
13. Persons can either pay one price for airfare, hotel, meals, etc, or they can pay for these items
separately. Please indicate where applicable the cost(s) to you and your immediate group (eg partner
or family) for your trip to Jamaica. (All prices are quoted in US$)

If you paid for a package deal please indicate below

If NOT check Not Applicable and Skip to the next table

Package Deal (Airfare, Accommodation and Meals)

Not Applicable $2001 $2500
$0 $500 $2501 $3000
$501 $1000 $3001 $4000
$1001 $1500 $4001 $5000
$1501 $2000 More than $5000

If you did NOT purchase a package deal (ie, you paid for your airfare, hotel, meals, etc. separately)
please indicate the costs in the table below.

Airfare Accommodation & Meals

Not Applicable Not Applicable
$0 $500 $0 $500
$501 $750 $501 $750
$751 $1000 $751 $1000
$1001 $1500 $1001 $1500
$1501 $2000 $1501 $2000
$2001 $3000 $2001 $3000
$3001 $4000 $3001 $4000
More than $3000 More than $3000

Please indicate the cost of any other items you may have paid for

Rental Car Diving Fees Golfing Fees Tours & Excursions

Not Applicable Not Applicable Not Applicable Not Applicable
$0 $250 $0 $30 $0 $50 $0 $50
$251 $500 $31 $50 $51 $100 $51 $100
$501 $750 $51 $80 $101 $150 $101 $150
$751 $1000 $81 $100 $151 $200 $151 $200
$1001 $1500 $101 $150 $201 $250 $201 $250
More than $1500 More than $150 More than $250 More than $250

Other Expenses : Type __________________ Cost ______________

14. How many persons did the costs above pay for? ________________
Person(s)

221
 Please note the following

The Jamaican government currently charges a tourism surcharge (tax) of US$10 per person to all overnight
visitors to Jamaica. This surcharge is typically included in your overall travel costs (airfare).

The proceeds from this tourism tax go towards a general tourism development fund and are to be used to
support management of the local tourist municipalities with activities such as:

General beautification,
Human resource training.

The purpose of the tourism development fund is to help maintain the tourism industry and assist in
Jamaicas social and economic development.

Please continue

222
 Please consider the following hypothetical scenario

Suppose that prior to your most recent trip to Jamaica, the Jamaican government decided to increase this
tourism tax. This new tax would result in an increase in your overall travel costs. The extra revenue
from this tax would go to the Government of Jamaica to be used to support necessary government
programs.

15. If, because of the increased tax, you now had to pay a per person surcharge of US$60 (in other
words an additional $50 on top of the existing $10) as part of your overall travel expenses, would
you still have decided to visit Jamaica?

Before you proceed, I want to talk to you about a problem that we have in studies like this one. Because this
is a hypothetical situation, people tend to behave differently when they know they wont have to dig into
their pocket and pay money. We often find, if the decision they are being asked to make involves something
that is good like contributing to a worthy cause the typical reaction is to agree to pay. But if it were a
real situation they would be faced with the option of spending money on this or something else. So, I am
asking you to consider what decision you would really make if you had to spend the extra money, given
your current budget.

(Check one)

Definitely yes
Probably yes
Probably not
Please go to the next page
Definitely not
If Yes, please indicate below your reasons for paying the tax.

____________________________________________________________

____________________________________________________________
Thanks for your comment, skip the next page and proceed to page 10

223
A Reminder
If you answered Definitely or Probably Yes to the previous question skip this page
and proceed to the next page.

16. If you answered Probably not or Definitely not, to the question on the previous page, please state
your reason: (Check all that apply)

This increased tax would be too expensive

Instead the most I would be willing to pay is $______extra
I am against paying any extra taxes
I dont think the government will use the money effectively
Other reasons (please specify below)

____________________________________________________________

____________________________________________________________

____________________________________________________________

Please proceed to the next page

224
 FINALLY here are a few questions about yourself that will help us to interpret our results.
As a reminder the information you provide is completely confidential and is needed for our
statistical analyses.

17. Please indicate your gender Female Male

18. Please state your year of birth Year______________

19. What is your current marital status?

Now married
Widowed
Divorced
Separated
Never Married

20. Where do you reside permanently? (Check one)

USA _______________ Canada_______________

zip code postal code

Japan __________________
City
European Union ______________________________________
Country City
Latin America ______________________________________
Country City
Caribbean ______________________________________
Country City
Other Country ______________________________________
Country City

225
21. What city is your final (airport) destination? 23. What is your current employment status?

________________________ Employed for wages

City Self employed
________________________
Out of work
Country Homemaker
Student
Retired

22. What is the highest level of education you have 24. Are you an active member of an
completed? (Check one) environmental organization, club or society?

Yes
Elementary school No
High school or GED
Associate degree
Bachelors degree
Masters degree
Professional degree (MD, PE)
Doctorate degree (Ph.D./Ed.D.)

226
25. Are you a certified SCUBA diver? 28. How many years have you been diving?

Yes ________ years

No Proceed to the next
page (Question 30).

26. What is your highest level of dive certification? 29. Approximately how many dives have
(Check one) you made in your lifetime?

Resort Diver ____________ dives

Open Water
Rescue Diver (Sports Diver)
Advanced Open Water
Dive Master (Dive Leader)
Master Scuba Diver

Other _________________

27. What diving organization are you affiliated with?

(Check all that apply)

PADI
NAUI
BSAC
CMAS
SSI
Other _______________________

227
 And finally for statistical purposes only....

32. Is the house, apartment or mobile home you live in:

30. Which category best describes your annual Owned by you/someone (with a mortgage)
household income before taxes in 2006? Owned by you/someone (with no mortgage)
(Check one). Rented for cash rent
Less than $20,000 Occupied without paying cash rent
$20,000 $50,000
$50,001 $80,000 33. If owned what is the approximate market value of
$80,001 $100,000 your home?
$100,001 $150,000 Not Applicable
$150,001 $200,000 $50,000 $100,000
$200,001 $300,000 $100,001 $150,000
$300,001 $400,000 $150,001 $200,000
More than $400,000 $200,001 $300,000
$300,001 $400,000
Please indicate Currency $400,001 $500,000
US Dollars $500,001 $600,000
Euros (treat $ amounts as )
More than $600,000
Other _____________
34. If rented what is your monthly rent?
Not Applicable
31. In your home country, do you consider your $0 $500
household income $501 $1,000
$1,001 $1,500
Below average $1,500 $2,000
About average $2,001 $2,500
Above average $2,500 $3,000
More than $3,000

END OF SURVEY

228
Thank you for your participation in this survey!

Your time and contribution are greatly appreciated. If there is anything else you would like to tell us about
your most recent trip to Jamaica please let us know. You may use the space provided below for that
purpose. If you have any comments that you think may help us better understand how you felt about the
quality of your vacation experience they are most welcome.

Additional Comments

Should you have any further questions or concerns about this survey please contact Peter Edwards at
pedwards@udel.edu.

Please place your completed survey in the envelope provided and return it to the address indicated on the
envelope.
Peter Edwards
University of Delaware
College of Marine and Earth Studies
Robinson Hall
Newark, DE 19716, USA
BCH4

229
 ENVIRONMENTAL TAX

CONTINGENT BEHAVIOR SURVEY

230
 Coral Reef and Beach Recreational Survey,
Jamaica

Please place your completed questionnaire in the envelope provided

and return it to the address indicated.

Environmental Foundation of Jamaica

231
Hello and Good Day,
I am a Jamaican PhD student asking for your help in study of coral reef and beach recreation. This study is
part of a research project that aims to better understand how visitors to Jamaica value the ocean and
beaches.
You were randomly chosen to participate in this PhD research project. Your help is voluntary and your
answers are completely confidential. This survey was approved by the University of Delawares Human
Subjects Review Board so your anonymity is assured. This means, none of the results of this survey can be
linked to particular individuals such as yourself.
Please note this is not a government funded project, however your thoughts and opinions are important in
helping us understand how we can assist environment and tourism officials better manage the beaches and
coral reefs in Jamaica.
The survey will take approximately 15 minutes to complete. Your time and cooperation in completing this
questionnaire are greatly appreciated.
After completing the survey please place the booklet in the envelope provided. Be sure to seal the envelope
to ensure confidentiality.
If for some reason you choose not to respond, please let us know by returning the blank questionnaire in the
enclosed self-addressed envelope.
Thank you very much for your participation.

Sincerely

..
PETER EDWARDS
PhD Student,

University of Delaware
College of Marine and Earth Studies
Newark, DE 19716, USA

232
START HERE
1. Including your most recent trip, how many times have you visited Jamaica? (Check one)

1 4
2 5 or more visits
3 I am a Jamaican living abroad

2. Please indicate the number of trips for pleasure (vacations) including this one you have taken
outside of your country of residence in the last five years.
1
23
45
6 10
11 or more

3. Please check all the activities you engaged in on your most recent trip to Jamaica.

Swimming Sun Bathing

Scuba Diving Reef Snorkelling
Glass Bottom Boat Tours Pleasure Cruises
Para Sailing Tours (Nature etc.)
Golfing Shopping
Business Meetings Conference

Other ____________________________
4. Where did you stay on your most recent visit to Jamaica?
Hotel/Resort (If you can please write the location)
Small Hotel/Inn
Bed and Breakfast ________________________
With Friends and Family Town or City (eg. Negril)
Other _______________________________

233
5. How satisfied were you with your overall vacation experience?
(Check one)

Very satisfied
Somewhat satisfied
Neither satisfied nor dissatisfied
Somewhat dissatisfied
Very dissatisfied

6. The factors listed below may have influenced your level of satisfaction on your vacation. Please
rate all the factors shown below. (Check one box per factor).

FACTORS Very Good Good Fair Poor Very Poor Not Sure
Hotel Amenities
Customer Service
Food
Night Life
Safety and Security
Shopping
Tours and Attractions
Friendliness of Locals

What did you like most about your trip to Jamaica?

_____________________________________________________________

What did you like the least about your trip to Jamaica?

_____________________________________________________________

234
 Now we would like you to give us your opinion on the quality of the
beaches, marine and underwater life you experienced on your
vacation to Jamaica.

7. How would you rate the overall quality of the beaches and swimming waters you experienced in
Jamaica?
(Check one box in each row).

Very Good Fair Poor Very Not

Good Poor Sure
Beach Sand Quality
Beach Cleanliness
Water Clarity

If you participated in any underwater viewing activities such as snorkelling, glass bottom boating or
SCUBA diving while on your vacation please answer the question below. If you did not please proceed to
the next page.

8. How would you rate the overall (or average) quality of the underwater life that you observed
during these activities?

Very Good Fair Poor Very Not

Good Poor Sure
Coral Reef Health*
Fish Abundance
Other Sea Creatures**
Underwater Visibility

*Coral reef health did you see mostly live coral (Good)? Was it overgrown with green plants (Poor)?
**Other sea creatures Did you see many other animals such as lobsters, crabs, eels and starfish?

9. What would you say is your general level of awareness about the following issues affecting the
planet as a whole?
(Check one box per issue)
ISSUES Very Aware Somewhat Very Little Not at All
Air Pollution
Water Pollution
Solid Waste Pollution
Nature Conservation
Climate Change
Deforestation
Coral Reef Protection
Population Growth

235
 Now we ask you for some details about your most recent trip to
Jamaica.

10. How many nights did you spend in Jamaica? _________ nights

11. With whom did you travel on this vacation?

(check all that apply)

Alone
Spouse/Partner
Family
Friends
Business Associates

Other______________________________

Please write below

Total Number of Persons (including yourself)__________

236
12. Persons can either pay one price for airfare, hotel, meals, etc, or they can pay for these items
separately. Please indicate where applicable the cost(s) to you and your immediate group (eg
partner or family) for your trip to Jamaica. (All prices quoted in US$)

If you paid for a Package Deal please indicate below

If NOT check Not Applicable and Skip to the next table

Package Deal (Airfare, Accommodation and Meals)

Not Applicable $2001 $2500
$0 $500 $2501 $3000
$501 $1000 $3001 $4000
$1001 $1500 $4001 $5000
$1501 $2000 More than $5000

If you did NOT purchase a package deal (ie, you paid for your airfare, hotel, meals, etc. separately)
please indicate the costs in the table below.

Airfare Accommodation & Meals

Not Applicable Not Applicable
$0 $500 $0 $500
$501 $750 $501 $750
$751 $1000 $751 $1000
$1001 $1500 $1001 $1500
$1501 $2000 $1501 $2000
$2001 $3000 $2001 $3000
$3001 $4000 $3001 $4000
More than $3000 More than $3000

Please indicate the cost of any other items you may have paid for
Rental Car Diving Fees Golfing Fees Tours & Excursions
Not Applicable Not Applicable Not Applicable Not Applicable
$0 $250 $0 $30 $0 $50 $0 $50
$251 $500 $31 $50 $51 $100 $51 $100
$501 $750 $51 $80 $101 $150 $101 $150
$751 $1000 $81 $100 $151 $200 $151 $200
$1001 $1500 $101 $150 $201 $250 $201 $250
More than $1500 More than $150 More than $250 More than $250

Other Expenses: Type __________________ Cost ______________

13. How many persons did the costs above pay for? ________________
Person(s)

237
 In order to answer the next questions, please consider the
following environmental issues facing Jamaica

Coral reefs, seagrass beds and beaches provide a number of benefits and environmental services such
as:
Coastal protection
Habitat and breeding areas for fish and other marine life
White sandy beaches
Food and jobs
Recreational activities (swimming, scuba diving etc).

Preserving the health of the coral reef ecosystem and beaches in Jamaica requires active management
to address threats such as:

Overfishing
Damage from: boats, divers and snorkelers,
Water pollution from; sewage, garbage and stormwater
Beach erosion
Overgrowth of coral reefs with algae (sea plants)

Environmental management is the responsibility of local Marine Park Management Organisations

along with the National Environment and Planning Agency. These organisations require funding to
sustain the following activities:

Marine Patrols - To: monitor anchor damage, beach erosion and coastal water quality; work
with fishermen to reduce overfishing; monitor diving and snorkeling behaviour
Public Education - with the public, schools and businesses; beach clean up programs
Joint Programs with other agencies - TO: reduce deforestation; address river pollution;
improve agricultural practices; improve garbage and sewage collection.

14. What would you say was your general level of awareness of the environmental issues described
on the previous page?

Very aware
Somewhat aware
A little aware
Not at all aware

15. Do you think that the presence of the coral reefs and beaches added to the enjoyment of your
most recent vacation experience in Jamaica?

Yes
No
I am not sure

238
 Please note the following

The Jamaican government currently charges a tourism surcharge of US$10 per person to all overnight
visitors to Jamaica. This surcharge is typically included in your overall travel costs (airfare). The proceeds
from this tourism tax go towards a general tourism development fund and are to be used to support
management of the local tourist municipalities with activities such as:

General beautification,
Human resource training.

The purpose of the tourism development fund is to help maintain the tourism industry and assist in
Jamaicas social and economic development.

Please continue

239
 Please carefully consider the following hypothetical plan to
protect Jamaicas beaches and reefs

Suppose that prior to your most recent trip to Jamaica, the Jamaican government decided to add an
environmental tax to the existing US$10 surcharge, as part of its efforts to provide funding for the
management of the coastal environment. These funds would go directly to the relevant environmental
management agencies for activities such as; marine patrols, public education and joint environmental
programs and therefore preserve the existing conditions and prevent a decline in environmental quality.

16. If, because of this mandatory environmental tax, you now had to pay a per person surcharge of
US$60 (in other words an additional $50 on top of the existing $10) as part of your overall travel
expenses, would you still have decided to visit Jamaica?

Before you proceed, I want to talk to you about a problem that we have in studies like this one. Because this
is a hypothetical situation, people tend to behave differently when they know they wont have to dig into
their pocket and pay money. We often find if the decision they are being asked to make involves something
that is good like protecting the environment the typical reaction is to agree to pay. But if it were a real
situation they would be faced with the option of spending money on this or something else. So, I am asking
you to consider what decision you would really make if you had to spend the extra money, given your
current budget.

(Check one)

Definitely yes
Probably yes
Probably not
Please go to the next page
Definitely not
If Yes, please indicate below your reasons for paying the tax.

_____________________________________________________________

_____________________________________________________________
Thanks for your comments skip the next page proceed to page 10

240
A Reminder
If you answered Definitely or Probably Yes to the previous question skip this page
and proceed to the next page.

17. If you answered Probably not or Definitely not, to the question on the previous page, please state
your reason: (Check all that apply)

This increased tax would be too expensive

Instead the most I would be willing to pay is $______extra
I am against paying any extra taxes
I dont think the government will use the money effectively
Other reasons (please specify below)

_____________________________________________________________

_____________________________________________________________

_____________________________________________________________

Please proceed to the next page

241
 FINALLY here are a few questions about yourself that will help us to interpret
our results. As a reminder the information you provide is completely
confidential and is needed for our statistical analyses.

18. Please indicate your gender Female Male

19. Please state your year of birth Year______________

20. What is your current marital status?

Now married
Widowed
Divorced
Separated
Never Married

21. Where do you reside permanently? (Check one)

USA _______________ Canada_______________

zip code postal code
Japan __________________
City

European Union ______________________________________

Country City
Latin America ______________________________________
Country City
Caribbean ______________________________________
Country City
Other Country ______________________________________
Country City

242
22. What city is your final (airport) 24. What is your current employment
destination? status?

Employed for wages

________________________ Self employed
City
Out of work
_________________________ Homemaker
Country Student
Retired

23. What is the highest level of education you 25. Are you an active member of an
have completed? environmental organization, club or
society?

Elementary school Yes

High school or GED No
Associate degree
Bachelors degree
Masters degree
Professional degree (MD, PE)
Doctorate degree (Ph.D./Ed.D.)

243
26. Are you a certified SCUBA diver? 29. How many years have you been
diving?
Yes
No Proceed to the next ________ years
page (Question 31).

27. What is your highest level of dive certification? (Check

one) 30. Approximately how many dives
have you made in your lifetime?
Resort Diver
Open Water ____________ Dives
Rescue Diver (Sports Diver)
Advanced Open Water
Dive Master (Dive Leader)
Master Scuba Diver

Other _________________
28. What diving organization are you affiliated with?
(Check all that apply)

PADI
NAUI
BSAC
CMAS
SSI

Other _______________________

244
 And finally for statistical purposes only....

31. Which category best describes your 33. Is the house, apartment or mobile home you
annual household income before taxes in live in:
2006? (Check one). Owned by you/someone (with a mortgage)
Owned by you/someone (with no mortgage)
Less than $20,000 Rented for cash rent
$20,000 $50,000 Occupied without paying cash rent
$50,001 $80,000
$80,001 $100,000 34. If owned what is the approximate market
$100,001 $150,000 value of your home?
$150,001 $200,000 Not Applicable
$200,001 $300,000 $50,000 $100,000
$300,001 $400,000 $100,001 $150,000
More than $400,000 $150,001 $200,000
Please indicate Currency $200,001 $300,000
US Dollars $300,001 $400,000
Euros (treat $ amounts as ) $400,001 $500,000
Other _____________
$500,001 $600,000
More than $600,000
32. In your home country, do you consider
your household income 35. If rented what is your monthly rent?
Below average Not Applicable
About average $0 $500
Above average $501 $1,000
$1,001 $1,500
$1,500 $2,000
$2,001 $2,500
$2,500 $3,000
More than $3,000

END OF SURVEY

245
Thank you for your participation in this survey!
Your time and contribution are greatly appreciated. If there is anything else you would like to tell us about
your most recent trip to Jamaica please let us know. You may use the space provided below for that
purpose. If you have any comments that you think may help us better understand how you felt about the
quality of your vacation experience they are most welcome.

Additional Comments

Should you have any further questions or concerns about this survey please contact Peter Edwards at
pedwards@udel.edu.
Please place your completed survey in the envelope provided and return it to the address indicated on the
envelope.
Peter Edwards
University of Delaware
College of Marine and Earth Studies
Robinson Hall
Newark, DE 19716, USA

EN4

246
STATED CHOICE SURVEY

247
 Coral Reef and Beach Recreational Survey,
Jamaica

Please place your completed questionnaire in the envelope provided

and return it to the address indicated.

Environmental Foundation of Jamaica

248
Hello and Good Day,
I am a Jamaican PhD student asking for your help in study of coral reef and beach recreation. This study is
part of a research project that aims to better understand how visitors to Jamaica value the ocean and
beaches.
You were randomly chosen to participate in this PhD research project. Your help is voluntary and your
answers are completely confidential. This survey was approved by the University of Delawares Human
Subjects Review Board so your anonymity is assured. This means, none of the results of this survey can be
linked to particular individuals such as yourself.
Please note this is not a government funded project, however your thoughts and opinions are important in
helping us understand how we can assist environment and tourism officials better manage the beaches and
coral reefs in Jamaica.
The survey will take approximately 15 minutes to complete. Your time and cooperation in completing this
questionnaire are greatly appreciated.
After completing the survey please place the booklet in the envelope provided. Be sure to seal the envelope
to ensure confidentiality.
If for some reason you choose not to respond, please let us know by returning the blank questionnaire in the
enclosed self-addressed envelope.
Thank you very much for your participation.

Sincerely

..
PETER EDWARDS
PhD Student,

University of Delaware
College of Marine and Earth Studies
Newark, DE 19716, USA

249
START HERE
1. Including your most recent trip, how many times have you visited Jamaica? (Check one)

1
2
3
4
5 or more visits
I am a Jamaican living abroad

2. Please indicate the number of trips for pleasure (vacations) including this one you have taken
outside of your country of residence in the last five years.

1
2
3
4
5
6
7 or more visits
3. Where did you stay on your most recent visit to Jamaica?
(Check one)
Hotel/Resort
Small Hotel/Inn
Bed and Breakfast
With Friends and Family
Other _______________________________
4. Where was your hotel or accommodation located?

Ocho Rios
Runaway Bay
Montego Bay
Negril
Whitehouse

Other _______________________________
I am not sure

250
5. Please check all the activities you engaged in on your most recent trip to Jamaica.
Sun Bathing
Swimming
Scuba Diving
Glass Bottom Boat Tours
Reef Snorkelling
Para Sailing
Pleasure Cruises
Tours (Nature etc.)
Golfing
Shopping
Business Meetings
Conference
Other ____________________________
Other ____________________________
6. How satisfied were you with your trip to Jamaica? (Check one)
Very satisfied
Somewhat satisfied
Neither satisfied nor dissatisfied
Somewhat dissatisfied
Very dissatisfied

Please state the reason below

___________________________________________________________

___________________________________________________________

___________________________________________________________

7. How many nights did you spend in Jamaica? _________ nights

8. How many people were in your travel party?

Number of Persons (including yourself) __________

251
9. What was the approximate out of pocket cost to you for your trip to Jamaica? Please consider
all expenses including airfare, hotel, meals, etc.
(All prices quoted in US$)

Check one
Approximate Costs
$0 $250
$251 $500
$501 $750
$751 $1000
$1001 $1500
$1501 $2000
$2001 $3000
$3001 $4000
$4001 $5000
More than $5000

10. How many persons did the costs above pay for? ________________
Person(s)

252
 In order to answer the following questions, please consider
the following environmental issues facing Jamaica.

Coral reefs, seagrass beds and beaches provide a number of benefits and
environmental services such as:

Coastal protection
Habitat and breeding areas for fish and other marine life
White sandy beaches
Food and jobs
Recreational activities (swimming, scuba diving etc).

Preserving the health of the coral reef ecosystem and beaches in Jamaica requires
active management to address threats such as:

Overfishing
Damage from boats, divers and snorkelers
Water pollution from sewage, garbage and stormwater
Beach erosion
Overgrowth of coral reefs with algae (sea plants)

11. What would you say was your general level of awareness of the environmental
issues described above? (Check one)

Very aware
Somewhat aware
A little aware
Not at all aware

253
As you may know, the Jamaican government currently charges a tourism surcharge of
US$10 per person to all overnight visitors to Jamaica. This surcharge is typically
included in your overall travel costs (airfare and hotel fees). The proceeds from this
tourism tax go towards a general tourism development fund.

The purpose of the tourism development fund is to:

support the maintenance and beautification of the local tourist municipalities,

help maintain the tourism industry
assist in Jamaicas social and economic development.

12. Were you aware that there was a per person surcharge?

Yes
No

254
 Now please carefully consider the following hypothetical plan to
protect Jamaicas beaches and reefs

Suppose the Jamaican government were considering three options for environmental
management.

1. Low Management
Engage in minimal management activities (such as sporadic beach cleaning). The US$10
surcharge would stay in place. It is expected that this option would result in an eventual
decline in environmental quality.

OR

2. Basic Management
Conduct basic management activities (eg. marine patrols, beach protection & fisheries
management). The $10 surcharge would remain but a compulsory environmental tax
would be added. These extras funds would be used to fund the basic management
activities. With this option it is expected that the current environmental quality will be
preserved.

OR

3. Advanced Management
Implement advanced management activities which would result in additional programs
such as solid waste management, forest preservation and environmental education. With
this option it is expected that the environmental quality will be improved.

Noticeable changes (in the case of an improvement or decline) would take effect in about
one year.
Please continue

255
The management options described on the previous page would affect the quality of
Jamaicas

Beaches
Swimming Waters
Coral Reefs
Fish

The table below shows a rating scheme of different levels of environmental quality that I
will refer to on the following pages.
Rating Beach Water Quality Fish Coral Reef
POOR Eroded, poor sand Poor underwater Only few small 90% of corals dead, other
quality visibility (cloudy) fish seen marine life absent
FAIR Some erosion, slightly Variable visibility, Moderate number 15% live corals (85%
improved sand quality (sometimes cloudy) of small fish live dead), other marine life
on the reefs (eg lobsters) rarely seen
GOOD No erosion, with Good underwater Moderate number 40% live corals, other
mainly white sand visibility (clear) of small and few marine life seen
large fish seen sometimes
Wide, with very white Crystal clear Many large and 75% live corals, other
EXCELLENT sand underwater visibility small fish seen on marine life seen (lobsters,
reefs octopi)

256
 Now I am going to ask you three hypothetical questions each has the
same format.

Suppose you were planning to take a trip to Jamaica and one of the previously mentioned
management options were already implemented. Suppose also that each option would vary by the
cost of the tourism surcharge as well as environmental quality.
13. First, assume you were faced with one of the following three options

Low Management Basic Management Advanced Management

Per Person
Tourism Surcharge $10 $40 $80
Beach Quality Fair Good Good
Water Clarity Fair Good Excellent
Coral Reef Fair Fair Good
Fish Fair Fair Good
(Feel free to refer to Page 8 for a review of quality ratings)

which of the options (if any) would you prefer? (Check one box below)

Before you answer, please remember that the tourism surcharge is currently US$10 per person. Although
this is a hypothetical situation I am asking you to consider what decision you would really make if you had
to spend the extra money, given your current budget.

Low Management
Basic Management
Advanced Management
I would choose none of the options, instead I would
(If none check one below)

Stay home (cancel my vacation)

Take a vacation in my home country
Visit some other island or country

Note: The environmental characteristics for Basic Management are based on current estimates of quality in Jamaica.

257
Second hypothetical scenario
14. Now assume you were faced with the following options

Low Management Basic Management Advanced Management

Per Person
Tourism Surcharge $10 $100 $200
Beach Quality Poor Good Good
Water Clarity Fair Good Good
Coral Reef Fair Fair Excellent
Fish Poor Fair Good
(Feel free to refer to Page 8 for a review of quality ratings)

which of the options (if any) would you prefer? (Check one)

Low Management
Basic Management
Advanced Management
I would choose none of the options, instead I would
(If none check one below)

Stay home (cancel my vacation)

Take a vacation in my home country
Visit some other island or country

Note: The environmental characteristics for Basic Management are based on current estimates of quality in Jamaica.

258
Third hypothetical scenario
15. Finally, assume you were faced with the following options
Low Management Basic Management Advanced Management
Per Person
Tourism Surcharge $10 $80 $100
Beach Quality Fair Good Excellent
Water Clarity Fair Good Excellent
Coral Reef Fair Fair Good
Fish Poor Fair Good
(Feel free to refer to Page 8 for a review of quality ratings)

which of the options (if any) would you prefer? (Check one)
Low Management
Basic Management
Advanced Management
I would choose none of the options, instead I would
(If none check one below)

Stay home (cancel my vacation)

Take a vacation in my home country
Visit some other island or country

Note: The environmental characteristics for Basic Management are based on current estimates of quality in Jamaica.

16. Consider the following statement. The three previous questions were extremely confusing.
Would you agree with this observation?
Strongly agree
Agree
Neither agree nor disagree
Disagree
Strongly disagree

259
 FINALLY here are a few questions about yourself that will help us to interpret
our results. As a reminder the information you provide is completely
confidential and is needed for our statistical analyses.

17. Please indicate your gender Female Male

18. Please state your year of birth? Year______________

19. What is your current marital status?

Single
Married
Widowed

20. Where do you reside permanently? (check one)

USA _______________ Canada_______________

zip code postal code
Japan __________________
City
European Union ______________________________________
Country City
Latin America ______________________________________
Country City
Caribbean ______________________________________
Country City
Other Country ______________________________________
Country City

260
21. What city is your final (airport) 25. Which category best describes your
destination? annual household income before taxes in
2006? (Check one).

______________________________ Less than $20,000

City $20,000 $50,000
______________________________
Country
$50,001 $80,000
22. What is the highest level of education you $80,001 $100,000
have completed? $100,001 $150,000
$150,001 $200,000
Elementary school $200,001 $300,000
High school or GED $300,001 $400,000
Associate degree More than $400,000
Bachelors degree
Masters degree Please indicate Currency
Professional degree (MD, PE) US Dollars
Euros (treat $ amounts as )
Doctorate degree (Ph.D./Ed.D.)
Other _____________
23. Are you an active member of an
environmental organization, club or 26. In your home country, do you consider
society? your household income
Yes
No Below average
About average
24. Are you a certified SCUBA diver?
Above average
Yes
No
END OF SURVEY

261
Thank you for your participation in this survey!

Your time and contribution are greatly appreciated. If there is anything else you would like to tell us about
your most recent trip to Jamaica please let us know. You may use the space provided below for that
purpose. If you have any comments that you think may help us better understand how you felt about the
quality of your vacation experience they are most welcome.

Additional Comments

Should you have any further questions or concerns about this survey please contact Peter Edwards at
pedwards@udel.edu.
Please place your completed survey in the envelope provided and return it to the address indicated on the
envelope.

Peter Edwards
University of Delaware
College of Marine and Earth Studies
Robinson Hall
Newark, DE 19716, USA

CM2

262
Appendix 5 Jamaican Tourism Statistics (JTB 2008)

263
Appendix 6 Respondent Feedback: Likes and dislikes of Jamaica visit
This appendix contains quotations of the written comments provided by respondents for
their answers to the question about factors influencing the level of satisfaction on the
respondents vacation.

What did you like most and what did you like least about your trip to Jamaica?
Answers from Tourism Survey (Question 7)

ID What I liked most about my trip What I liked least about my trip
terribly bumpy road from Montego Bay to
1 relaxation, food sun Whitehouse
the shuttle to and from the airport and
4 the accommodations and people we met resort
5 All natural beach Chaos at airport
6 the people the airports
7 Friendly service and good food travel distance from airport to resort
8 the beach water clarity
9 Relaxation Our resort
10 the warm of Jamaica Nightlife
11 Customer service, food weather being pushed by locals to buy things
12 A sense of relaxation and not stress other tourists
18 Hospitality and sunny warm temperatures lost luggage (North West Airlines)
19 beach
20 customer service "tips"
21 beautiful beach, ocean, shore poverty, personal safety concern
22 the peacefulness
street vendors especially at Dunn's River
23 sun and beach Falls
everyone seems to be so happy to do their people trying to sell you something
24 job, duties and so happy to serve you constantly
25 the warm weather waiting at the airport
26 getting married , atmosphere the lines in the airport
people outside resort trying to get us to
28 weather, sandals resort purchase items
29 weather leaving
30 fun in the sun, everything is alright
31 weather, location, English speaking people quality of resort, local town
32 the nice beaches, people offered drugs by locals
ID What I liked most about my trip What I liked least about my trip
33 Weather Reservation was lost
Driving through the countryside - St car accident resulting from reckless
34 Elizabeth and Manchester Jamaican driver
Dunn's River Falls market - too
aggressive very scary experience.
Sonny's markets Ocho Rios no quality
shops like those at the airport or near
35 Sailing, Beach, Sun Rosehall.
The good people of Jamaica, friendly
36 people the beach at Starfish Hotel not very clean
37 Beach and water condition of roadways = poor
38 meeting Jamaicans electricity blackouts
39 the people people wanting something from you
40 Welcoming of Sandals employees Travelling
Compared to other Caribbean places I was Everything was good except the check
41 happy with not having a language barrier into the hotel was very long
42 Weather Services at motel
43 Friendly locals, good attitude lack of entertainment at resort
Getting off the resort and interacting with
44 locals heckling of street vendors
bus trip from airport to hotel, would
prefer direct transportation no stops other
45 warm weather than at hotels to drop off passengers
46 Learning more about the culture Being harassed by street people
47 relaxed atmosphere trip to and from airport
48 people locals and laid back attitude hustlers
49 weather activities traffic
50 Climate resort was loud, busy
I got some sun; and showed them around Jealous people, referring to me as "that
51 Jamaica for their first time foreign boy"
52 the people falsely advertised tours
53 friendly people nothing
54 Weather Airport security, stupidity
(possibly perceived) sense of poor
55 Weather and local people security
56 Ocho Rios Being pestered in the street
weather, friendliness of locals, quality of
57 food and water bartering method of some shopping
58 people the hassling of locals
59 the friendliness of the people the roads
The starry nights and long walks in the
68 "bush" and farm animals long car rides to different parishes

265
ID What I liked most about my trip What I liked least about my trip
69 the weather travelling to get here
70 Weather Not sure
71 Sun and people being hassled on the street
72 Hospitality food quality at the resort
73 Everything about the resort not sure
74 Sunshine attitude of average people
Flight into Jamaica was 3 hours late,
75 Sandals Dunn's River Falls problems with original booking
76 Food and hotel the roads
77 Meeting family
78 Swimming in the sea Food
79 Weather, ocean and tours! bus ride to depart to the airport
80 Weather It was too short
the natural beauty and friendliness of the
82 resort staff and locals nothing
83 Weather long drive to resort
84 the people made to feel like you have to tip everyone
85 the people quality of roads
86 All inclusive, weather travel time- wait at airports
badgering from locals when out shopping,
87 my resort couples swept away! and when they would come on the resort
the ocean temperature was cooler than
expected. And one hotel which advertised
hot tubs had cold water in hot tub. False
advertising promises "it will work" not
sunshine, sunsets, relaxing on beach or by kept. Other problems with false
88 pools advertising eg free wireless in airport.
89 friendly resort staff airport staff
90 swimming pools and locals 10 hour flight
91 the food and Dunn's river Sandals Beach Resort
walking down the streets in Negril, West prices at our hotel/resort compared to
92 end and meeting the people outside
Staff at Sandals Royal Jamaica, very
93 accommodating Poor food quality
94 Sun
95 The people at the hotels were great! the dead animals on the roads
96 Water color cheap service
97 Sun and Sand I cannot think of one!
overly aggressive vendors on beach and
98 The friendliness of the local people Dunn's River Falls
99 The ocean and beach annoying sales people

266
ID What I liked most about my trip What I liked least about my trip
100 the warmth relaxation and friendliness the brevity
109 Sun Bathing nothing
110 Seeing friends and family, night life driving
Seeing the country and getting to see the many local men try to get to fresh with
111 pretty blue water single women. It is a little scary
112 Sunshine and food the attitude of some people
113 Great Weather resort ran out of many items
114 Sun food
115 Ripping off
116 Walking down street talking to locals On tours, severe pressure to buy things
117 swimming and snorkelling noise
118 the people observing the poverty level
The sand flies and ridiculous prices for a
119 small snack at the airport
121 Ocean, sun, rest, drinks service should be better
122 sun sea food could not open a bank account
123 relaxing leaving
124 beautiful sunshine and great food feeling scared
125 the beach and sunbathing food
126 sun and food the road was good (?)
127 Caribbean Sea bus/shuttle trips
128 Helping others the way tourists treat residents
129 We won at rugby got sun burnt
130 Black River tour Humidity
131 Shopping and sight seeing I came for a funeral
132 Sun No smile from service people
our hotel had several exotic parrots locked
133 the reef snorkelling in small cages
134 the weather transportation
135 Driving
136 Good weather police shake down for money
137 Rest/sex having to leave
not enough smiling, garbage everywhere
139 Weather, feeling of freedom on the resort on the way to the resort
140 Swimming getting hassled
141 weather/resort no major complaint
142 Relax - sun - food - our cook transportation - roads
143 Quality of the resort nothing
144 the beach being bothered by locals to buy

267
ID What I liked most about my trip What I liked least about my trip
Able to visit friends and relatives and to
145 explore the island
146 climate, nice island peoples attitude, aggressiveness
147 The resort experience long travel to tours
149 Seeing my family members
150 the culture hustlers
151 swimming the road
152 Sun - Beach being shaken down by locals
153 tours the breakfast
getting robbed at our hotel and hassled by
183 reggae music locals
184 the variety of things to do being approached by beach vendors
185 Relaxed attitude of locals the roads and traffic congestion
186 attractions
187 Sun bathing by the pool 2 hour drive from airport to hotel
188 all natural beach ride to resort
when somebody just pressure you to do
189 the people things
the taxi drive from Negril to Montego
190 the scenery and hotel accommodations Bay
the people, great weather, Rockhouse
191 hotel, where we stayed and 7 mile beach a lot of the food was fried!
Jamaica needs more professionals on
193 the country side tourism
194 weather transportation from resort to airport
195 the weather the ride to Ocho Rios
We were changed rooms 3 times in our
196 the warm weather and the Jamaican people resort
197 bonding with my travel companions the overzealous market vendors
sense of insecurity! Guards everywhere:
warnings not to leave relatively small
198 the people resort area. I felt restricted
199 the total vacation package zero
200 weather bus ride to resort
201 weather
waiting in line at the airport for entering
202 Good service and friendly people the country (approx 1 hour)
203 weather
204 Graciousness of people too short
205 warm weather - people airport wait time
206 meeting people and visiting friends leaving!

268
ID What I liked most about my trip What I liked least about my trip
I wouldn't have left my resort for fear of
207 warm weather, welcoming people crime
208 no pressure leaving so soon
(Bad?))customer experience at resort - 1
209 resort time event though
210 beach
I couldnt use my calling card to call
211 Locals welcoming us! home
Service at Sandals Dunn's River, pleasant
212 smiles
213 People, tours and attractions A room change when we first arrived
214 All inclusive amenities ground travel, primitive roadways
215 Scenery, food, people Jamaicans are beggars!
216 weather, water temperature and clarity unsafe roads
217 friendliness of people flight
220 Relaxing
221 sun bathing/beach beer
the bus trip from Montego Bay to Dunns
222 Sunshine and friendly people River. Clearing customs took 2 hours
223 too short nothing
Friendly people on the island and at the non-organization of arrival to airport
224 resort when coming through customs
225 the weather and the rum having to leave
226 the people sales people on the beach
227 the beach the food
228 Being home Customer service
230 Abundance of ganja the resort
231 the ocean, weather, getting away construction going on, food at the buffet
232 snorkelling my room location

269
Answers from the Environmental Survey (Question 6)

ID What I liked most about my trip What I liked least about my trip
1 Friendliness of natives hounding by small shop proprietors
2 people, beaches customs
3 Sun and ocean resort
4 Meeting locals Airport travel
5 weather, personnel, tours going home
Clean, clear water(ocean) and very friendly
6 local people flight difficulties
7 People, food, tours bus rides
8 the ocean leaving
9 Relaxation and people of Jamaica Curvy roads
10 Sunshine and friendly people too short
11 Dunn's River Falls
12 Food Late Flight!!!
14 Beach
15 The weather/sunshine and heat not long enough
16 Dunns River Falls Size of beaches
17 The weather and the swim up bar the construction at our hotel
18 the weather and beaches resort near noisy road
19 weather local people felt unsafe away from resort
20 Sunshine and warmth we had a lot of problems at our resort
21 Beaches lack of access to town
poverty- do not like that Jamaica has so
22 weather, people (local) culture much poverty
23 Sunshine, reef snorkelling hotel housekeeping
24 weather and hospitality pressure at craft markets
not enough beach shopping on Bloody
25 people, locals Bay
26 interacting with the Jamaican children garbage in the streets, water etc
poor service, long wait times in hotel
27 Sunshine, scenery dining rooms/restaurants
29 customer service
people drive like maniacs! Service is
31 Direct flight from Atlanta, weather slow, have to ask for checks, refills etc.
32 Cool Runnings trip to Dunn's river Falls walls in hotel too thin
snorkelling, horseback riding, friendly High prices, not enough good American
33 people friendly food
35 Perfect weather to enjoy the outdoors intermittent services like phone and water
36 friendly people rain
37 Helping the people
38 Ocean, Dunn's River Falls & Ocho Rios Hustlers
39 Meeting People Traffic
40 friendliness of staff food was ordinary

270
ID What I liked most about my trip What I liked least about my trip
41 friendliness of staff road to Negril
42 friendliness of staff delayed flight
43 warmth of climate and people
44 Resort Drive to resort
45 the friendly people leaving
46 Warm weather cost of food at restaurants
47 the culture and the locals leaving Jamaica
48 beach vendors
49 serenity
the weather, hospitality culture and locals are too pushy and solicit(ate) too
50 philosophy much
51 Sun and beach locals harassing you
52 the beauty and food nothing it was awesome
53 Our hotel
54 Very nice people long ride to resort
55 nice beach/ocean water/ all inclusive resort resort too close to airport
56 relaxing not serving coke products in the resort
57 water sports resort under renovations
58 character of the people lack of Jamaican's political awareness
59 warm weather travel agency. Expedia.com
60 weather pushy vendors
tourist harassment is less on Negril beach
61 from April 2007 nothing
62 weather and relaxation rain
63 friendly people poor sidewalks and potholes in roads
64 canopy tour food
relentless craft merchants, accuracy in
66 relaxed atmosphere and scenery service (food)
67 Dunn's River waterfall pushy vendors
68 the beauty and accessibility constant selling of goods
69 the beauty and the beaches the roads
71 my time was my own airport sitting
72 the people long flights
73 weather bus rides
hotel amenities, beaches, nightly
75 entertainment people not friendly, no bottled water
customs took 11/2 hours not a pleasant
76 friendliness of people welcome!
77 Bob Marley tour in nine miles/weather hustlers and expenses
78 Dunn's River falls not feeling safe outside of the resort
79 weather nothing
80 Dunn's river falls, the people, the scenery resort food
81 the water seeing the poor people
82 Delays
83 everything very little

271
ID What I liked most about my trip What I liked least about my trip
84 Sailing
85 Golf
86 The people aggressive selling
87 the cruise being pestered all the time
88 the beach food prices
89 laid back, weather charging tourist too much money
90 the ocean the hospital
91 the friendly people long ride to resort
92 the beauty of the island not enough time
93 weather Jamaican time, or lack of punctuality
94 weather, hotel staff cleanliness, honesty
95 beaches shops
96 cool breezes, cool evenings pot holes!
97 Beautiful country too many people selling things
98 the nice weather and people the long trip, the waiting
the cost of our tours for such a short time
I liked the people, weather, trees, ocean, 150US for a 3-4 hour tour. At Cancun it
99 beaches and the weed mon! was cheaper
100 friends made here got sick once
101 sunshine and local staff telephone inoperative
102 Weather slow speed of everything
103 the coffee! maniac drivers
104 Weather, people, food Travel airlines
people approaching on the street to sell
105 trip to Ocho Rios and Dunn's River Falls goods/services
106 the weather and relaxation customs and airport experience
personal satisfaction, rest and relaxation in
108 the sun too short
109 Relaxation, people and sunshine the departure, rudeness of Americans
relaxing atmosphere silver sands beach and
110 snorkelling nothing
Being able to relax and have fun when
111 needed nothing (wish I could say)
leaving so early for our resort to get to the
112 Tubing, sandals employees airport
113 Jamaica (scenery and weather) transportation
nights were not very good, mosquitoes.
115 Beach life Hotel not much entertainment
116 Efficient service, friendliness cant get out the resort, not safe
117 people and culture, beautiful country haggling of shop owners , driving
118 relaxation and the Jamaican people travel time
120 Good weather, nice people the security required around the resort
friendly people in Beaches Negril, weather,
121 hotel going home, travelling to hotel
122 Sun, Sea and vegetation Rude staff

272
ID What I liked most about my trip What I liked least about my trip
123 Service
124 the beach the people the black flies
125 the people the flight
lack of training and service from the
126 friendship of people employees from scuba caribe
127 watching the sunset in Negril the crime
128 weather, beach, good vibe attempting to buy something
129 local accommodations and hospitality getting burnt in the sun
130 the weather cable channels
131 snorkelling leaving
hotel nightlife was not great too many
133 Friendly people, stunning beach and sea mosquitoes
134 the friendliness of the people
concerns of personal security after dark
135 the people outside of hotel
136 people
137 weather and scenery local road conditions can be better
138 the beauty of the land the haggling of the shop keepers
139 atmosphere of the resort airport transportation
140 the sea the roads in the country
141 Sunshine view from hotel window
142 learning the local culture, recipes driving
143 saying good bye the food
did not feel comfortable to leave
144 weather - golf compound (Tryall)
146 weather
147 the beaches an the Jamaican smile when shopping Jamaican use too insistent
148 Quality of services and food
149 Relaxing Unsure
150 catamaran/snorkelling in room hotel maintenance issues
151 sun/beach street safety
152 Jerk chicken girls
153 weather excessive tipping
154 the beach the roads
155 the hotel beach and weather
156 weather and drinks nagging people at shops
157 friendliness leaving
158 weather and people flight time 10 hours
159 The hospitality the food
160 the beach and falls
161 catamaran, Dunn's River Beach
162 the ocean and friendliness seeing so much poverty
163 Mingling with locals nothing - loved it all
my knowledge that a dual class system
164 the serenity exists

273
ID What I liked most about my trip What I liked least about my trip
166 weather beautiful countryside hotel room
The hotel I stayed at and staff were
exceptional. I met some very nice locals the police could have more of a presence
where I feel lasting friendships will in downtown Montego Bay and on the
167 develop hip strip
168 relaxation, meeting locals & snorkelling aggressive vendors
169 ocean and sun beggars
the airport employees were very
170 The beach and meeting new people unfriendly and rude
171 the wonderful people and beaches bad pay for local workers
172 weather people food leaving
rest and relaxation, friendly people meeting
173 people from around the world all day trip to get here
174 Boat, catamaran beach
175 the sun and food
unable to get American $ changed in the
176 beach sun water hotel
177 arriving leaving
178 Sun heat
179 Sun Hotel
180 the beauty of Jamaica waiting at the airport
181 Local parties in the country fun holiday - lost most of our beach
182 Beaches Pushy vendors
183 the happy friendly people, weather Infrastructure
bag did not arrive with me, got it next
184 relaxing atmosphere, low key feeling evening
185 The tours the food at the hotel
have to go home the number of people
186 Great weather trying to sell marijuana
187 weather in January/Jamaican food having to leave
188 Sun Drug pushers
189 the natural beauty of the land and water the litter

274
 REFERENCES

Adamowicz, W., J. Louviere, and M. Williams, (1994). Combining Revealed and Stated

Preference Methods for Valuing Environmental Amenities. Journal of

Environmental Economics and Management 26:271-92.

Adamowicz, W., J. Swait, P. Boxhall, and M. Williams, (1997). Perception Versus

Objective Measures of Environmental Quality in Combined Revealed and Stated

Preference Models of Environmental Evaluation. Journal of Environmental

Economics and Mangement 32:65-84.

Adamowicz, W., P. Boxall, , J. Louviere, J. Swait, J. and M.Williams, (1998). Stated

preference approaches for measuring passive use values: Choice experiments and

contingent valuation. American Journal of Agricultural Economics 80(1):6475.

Adamowicz, W., J, Louviere, P Boxhall, J Swait, and M. Williams, (2001). Stated

Preference Methods for Valuing Environmental Amenities. In: Valuing

Environmental Preferences Theory and Practice of the Contingent Valuation

Method in the US, EU , and developing Countries Eds Ian Bateman, and Kenneth

G Willis, Oxford University Press, 2001; pp 460-479.

Anderson, J.E.C., (2007). The recreational cost of coral bleaching A stated and revealed

preference study of international tourists. Ecological Economics 62: (3-4):704-

715.

275
Arin, T. and R.A. Kramer, (2002). Divers willingness to pay to visit marine sanctuaries:

an exploratory study. Ocean and Coastal Management 45:171-183.

Bartlett E., (2008). Ministerial Presentation: The Caribbean Hotel Association

Marketplace, Jan 14, 2008. http://www.eturbonews.com/840/presentation-hon-

edmund-bartlett-minister-tourism-jamaica.

Bateman I.J., I.H. Langford, P.K. Turner, K.G. Willis and G.D. Garrod, (1995).

Elicitation and truncation effects in contingent valuation studies. Ecological

Economics; 12: 161-179.

Bateman, I.J., R.T. Carson, B. Day, M. Hanemann, N Hanley, T. Hett, M. Jones-Lee, G.

Loomes. S. Mourato, E. zdemiroglu, D.W. Pearce, R. Sugden, and J. Swanson,

(2002). Economic Valuation with Stated Preference Techniques, A Manual,

458pp. Edward Elgar Publishing.

Batemen, I.J. and K.J. Willis, (1999) (Eds). Valuing Environmental Preferences. Theory

and Practice of the Contingent Valuation Method in the US, EU and Developing

Countries. Oxford University Press, 1999.

Batsell, R. R. and J.J. Louviere (1991) Experimental Choice Analysis: Marketing Letters

2:199-214.

Bennett, J.W. and W.I. Adamowicz, (2001). Some Fundamentals of Environmental

Choice Modelling. In J.W. Bennett and R. Blamey (eds.), The Choice Modelling

Approach to Environmental Valuation (pp. 3769). Northampton, MA: Edward

Elgar Publishing Limited.

276
Bennett, J. and R. Blamey, (2001). The Choice Modelling Approach to Environmental

Valuation, New Horizons in Environmental Economics, Edward Elgar Publishing,

Cheltenham, UK.

Bergmann, A., N. Hanley and R. Wright, (2006). Valuing the attributes of renewable

energy investments. Energy Policy, 34: 1004-1014.

Bergstrom, J.C., J.R. Stoll, and A. Randall, (1989). Information effects in contingent

markets. American Journal Agricultural Economics 71: 685-691

Birol, E., K. Karousakis and P. Koundouri, (2006). Using a choice experiment to account

for preference heterogeneity in wetland attributes: The case of Cheimaditida

wetland in Greece. Ecological Economics 60:145-156.

Blackburn, M., G.W. Harrison and E.E. Rustrom, (1994). Statistical Bias Functions and

Informative Hypothetical Surveys. American Journal of Agricultural Economics

76(5):1084 88.

Blomquist G.C., and J.C. Whitehead, (1998). Resource quality information and validity

of willingness to pay in contingent valuation. Resource and Energy Economics;

20: 179-196

Boadway, R., and A. Shah, (1992). How Tax Incentives Affect Decisions to Invest in

Developing Countries. World Bank Policy Research Working Paper 1011,

Washington, D.C.: The World Bank

Boardman, A., D. Greenberg, A. Vining and D. Weimer, (2001). Cost-Benefit Analysis:

Concepts and Practice. 2nd ed. Upper Saddle River, NJ: Prentice Hall.

277
Bockstael, N.E., (1995). Travel Cost Models: Handbook of Environmental Economics.

Cambridge, MA: Blackwell Publishers.

Boxall, P.C. and W.L. Adamowicz, (2002). Understanding heterogeneous preferences in

random utility models: a latent class approach. Environmental and Resource

Economics, 23:421446.

Boxhill, I., (2004). Towards an alternative tourism for Jamaica. International Journal of

Contemporary Hospitality Management; 16(4/5): 269-272

Boyle, K., (2003). Introduction to Revealed Preference Methods. In P.A. Champ, K.J.

Boyle & T.C. Brown (Eds.), A Primer on Nonmarket Valuation (Vol 3, pp. 576)

Dorcdrecht: Kluwer Academic Publishers.

Brown, B. E., and J.C. Ogden, (1993). Coral Bleaching. Scientific American 268: 6470.

Bruner, A.G., J.E.N. Sweeting and A.B. Rosenfeld, (1998). The Green Host Effect: An

Integrated Approach to Sustainable Tourism and Resort Development.

Conservation International.

Bryant, D., L. Burke, J. McManus, M. Spalding, (1998). Reefs at Risk: A Map-Based

Indicator of Threats to the Worlds Coral Reefs. World Resources Institute,

Washington, DC.

Bunce L, K. Gustavson, J. Williams and M. Miller, (1999). The human side of reef

management: a case study analysis of the socioeconomic framework of Montego

Bay Marine Park. Coral Reefs; 18:369-389.

Cameron, T. A., (1992). Combining Contingent Valuation and Travel Cost Data for the

Valuation of Nonmarket Goods, Land Economics, 68:302-317.

278
Carpenter K.E., M. Abrar, G. Aeby, R.B. Aronson, S. Banks, A. Bruckner, A. Chiriboga,

J. Corts, J.C. Delbeek, L. DeVantier, G.J. Edgar, A. J. Edwards, D. Fenner, H.M.

Guzmn, B.W. Hoeksema, G. Hodgson, O. Johan, W.Y. Licuanan, S. R.

Livingstone, E. R. Lovell, J. A. Moore, D.O. Obura, D. Ochavillo, B.A. Polidoro,

W.F. Precht, M.C. Quibilan, C. Reboton, Z.T. Richards, A.D. Rogers, J.

Sanciangco, A. Sheppard, C. Sheppard, J. Smith, S. Stuart, E. Turak, J.E.N.

Veron, C. Wallace, E. Weil, and E. Wood, (2008). One-Third of Reef-Building

Corals Face Elevated Extinction Risk from Climate Change and Local Impacts.

Science 321(5888):560-563.

Carson, R.T., J.L.Wright, A. Albertini, N. Carson and N.E. Flores, (1994). A

Bibliography of Contingent Valuation Studies and Papers. Natural Resource

Damage Assessment, La Jolla, CA., March.

Carson R.T., N.E. Flores, K.M. Martin, and J.L. Wright, (1996). Contingent Valuation

and Revealed Preference Methodologies: Comparing the Estimates for Quasi-

Public Goods. Land Economics 22(1):80-99.

Carson, R.T. and W. M. Hanemann, (2005). Contingent Valuation. In K.G. Maler and

J.R. Vincent, Eds. Handbook of Environmental Economics, North Holland, 2005.

Central Intelligence Agency, (2008). CIA, The World Fact Book, Jamaica.

https://www.cia.gov/library/publications/the-world-factbook/geos/jm.html

Chase, L.C., D.R. Lee, W.D. Schulze and D. J. Anderson, (1998). Ecotourism demand

and differential pricing of national park access in Costa Rica. Land Economics

74(4):466-482.

279
Clarke, H. and Y.K. Ng, (1993). Tourism, Economic Welfare and Efficient Pricing.

Annals of Tourism Research; 20:613 632.

Cummings, R.G., D.S. Brookshire and W.D. Schulze, (1986) Eds. Valuing

Environmental Goods: An assessment of the Contingent Valuation Method,

Totowa, NJ: Rowman & Allanheld.

Cummings R.G., S. Elliot, G.W. Harrison G.W. and J. Murphy, (1997). Are

Hypothetical Referenda Incentive Compatible? The Journal of Political

Economy; 105(3):609 621.

Cummings R.G. and L.O. Taylor, (1999). Unbiased Value Estimates for Environmental

Goods: A Cheap Talk Design for the Contingent Valuation Method. The

American Economic Review; 89(3):649 665.

Dharmaratne, G.S., F. Yee Sang, L. Walling, (2000) Tourism potentials for financing

protected areas. Annals of Tourism Research 27(3):590-610.

Dillman, D. (2007). Mail and Internet Surveys: The Tailored Design Method. 3rd Ed.

Wiley: New York, NY.

Dixon, J.A., L.F. Scura, and T. vant Hof, (1993). Meeting Ecological and Economic

Goals: Marine Parks in the Caribbean. Ambio 22(23):117125.

Dixon, J.A., L.F. Scura and T. vant Hof, (2000). An economic and ecological analysis

of the Bonaire Marine Park. Cesar HSJ (ed) Collected essays on the economics of

coral reefs. Kalmar Sweden, Cordio, 242 pp.

Dixon, J., K. Hamilton, S. Pagiola and L. Segnestam, (2001). Tourism and the

Environment in the Caribbean: An Economic Framework. Environment

280
 Department Papers, No. 80. Environmental Economics Series, The World Bank,

March 2001.

Edwards, P.E.T., (2002). Mangrove, Seagrass and Coral Reef Community Interactions of

the Falmouth Coast, North West Jamaica. M.Phil. Thesis, University of the West

Indies, Mona, April 2002.

Edwards P.E.T., (2009). Sustainable Financing for Ocean and Coastal Management in

Jamaica: The Potential for Revenues from Tourist User Fees. Marine Policy

33:376-385.

Emanuel, K., 2005. Increasing destructiveness of tropical cyclones over the past 30

years. Nature 436:686-688.)

Englin, J., and T.A. Cameron, (1996). Augmenting Travel Cost Models with Contingent

Behavior Data: Poisson Regression Analyses with Individual Panel Data.

Environmental and Resource Economics. 7:133-47.

Environmental Management Unit, (2001). Socio-Economic Valuation Study of the Ocho

Rios Marine Park. Department of Geology and Geography, University of the

West Indies.

Fox, J., J. Shogren, D. Hayes and J. Kleibenstein, (1999). CVM-X: Calibrating

Contingent Values with Experimental Auction Markets. American Journal of

Agricultural Economics 80(3):455 65.

Freeman, A.M.I., (2003). Economic valuation: what and why. In P.A. Champ, K.J.

Boyle & T.C. Brown (Eds.), A Primer on Nonmarket Valuation (Vol 3, pp. 576)

Dorcdrecht: Kluwer Academic Publishers.

281
Gooroochurn, N. and M. T. Sinclair, (2005). Economics of tourism taxation: Evidence

from Mauritius. Annals of Tourism Research; 32 (2):478 498.

Greene, W.H., (1997). Econometric Analysis, Third Edition. Prentice Hall.

Grijalva. T.C., R.P. Berrens, A. K. Bohara, and W. D. Shaw, (2002). Testing the Validity

of Contingent Behavior Trip Responses. American Journal of Agricultural

Economics; 84 (2):401-414.

Gustavson, K., (1998). Values Associated with the Local Use of the Montego Bay

Marine Park. Report Prepared for the Environmentally and Socially Sustainable

Development Sector Department, Latin America and the Caribbean Region,

(LCSES). World Bank, Washington, D. C.

Haab, T.C. and K.E. McConnell, (2002). Valuing Environmental and Natural Resources:

The Econometrics of Non-Market Valuation. Edward Elgar Publishing Inc. MA,

USA.

Hanemann, W.M., (1984). Welfare Evaluations in Contingent Valuation Experiments

with Discrete Responses. American Journal of Agricultural Economics (66)

3:332-341.

Hanemann, M. and B. Kanninen, (1999). Statistical Analysis of Discrete-Response CV

Data (302-441). In: I.J. Bateman & K.G. Willis (Eds) Valuing Environmental

Preferences. Theory and Practice of the Contingent Valuation Method in the US,

EU and Developing Countries. Oxford University Press, 1999.

282
Hanley, N., R.E. Wright and G. Koop, (2002). Modelling Recreation Demand Using

Choice Experiments: Climbing in Scotland. Environmental and Resource

Economics 22:449-466.

Hensher, D.A., J.M. Rose, and W. H. Greene, (2005). Applied Choice Analysis: A

Primer. New York: Cambridge University Press.

Hiemstra, S. and J. Ismail, (1992). Analysis of Room Taxes Levied on the Lodging

Industry. Journal of Travel Research; 31(1):42 49.

Hoegh-Guldberg, O.,P.J. Mumby, A.J. Hooten, R.S. Steneck, P. Greenfield, E Gomez,

C.D. Harvell, P.F. Sale, A.J. Edwards, K. Caldeira, N. Knowlton, C.M. Eakin, R.

Iglesias-Prieto, N. Muthiga, R.H. Bradbury, A. Dubi, and M. E. Hatziolos, (2007).

Coral Reefs Under Rapid Climate Change and Ocean Acidification. Science (14)

318: 1737-1742

Hoehn J.P. and A. Randall, (1987). A satisfactory benefit-cost indicator from contingent

valuation. Journal of Environmental Economic Management; 14:226-247.

Huang, J.C., P.J. Poor and M. Q. Zhao, (2007). Economic Valuation of Beach Erosion

Control. Marine Resource Economics 22(3):221-239.

Hughes, T.P., A.H. Baird, D.R. Bellwood, M. Card, S.R. Connolly, C. Folke, R.

Grosberg, O. Hoegh-Guldberg, J.B.C. Jackson, J. Kleypas, J.M. Lough, P.

Marshall, M. Nystrm, S.R. Palumbi, J. M. Pandolfi, B. Rosen, J. Roughgarden,

(2003) Climate Change, Human Impacts, and the Resilience of Coral Reefs.

Science (301) 5635, 929 933.

283
Issa J.J. and C. Jaywardena, (2003). The all inclusive concept in the Caribbean.

International Journal of Contemporary Hospitality Management 15(3):167-171.

Jamaica Tourist Board Annual Travel Statistics, (2007). Research & Market Intelligence

Unit, Marketing Department, Jamaica Tourist Board, 64 Knutsford Boulevard,

Kingston 5, Jamaica.

Jamaica Tourist Board, (2008). Monthly Statistical Report, December 2008, Volume

XVIII, Issue 12.

http://www.jtbonline.org/statistics/Monthly%20Statistics/Monthly%20Statistical

%20Report%20December%202008%20Vol%20xviii%20No%2012.pdf.

Jamaica Trade and Invest (2007). Jamaica at a Glance: Tourism Incentives.

http://www.jamaicatradeandinvest.org/index.php?action=investment&id=6&oppa

ge=1&optyp=mm

Kolstad, C.D., (2000). Environmental Economics. Oxford University Press, New York

NY.

Krueger A., (2007). Valuing Public Preferences For Offshore Wind Power: A Choice

Experiment Approach. PhD Dissertation, College of Marine and Earth Studies,

University of Delaware, 158pp.

Lancaster, K., (1966). A new approach to consumer theory. Journal of Political Economy,

74: 132-157.

Leeworthy, V.R., (1996). Technical Appendix: Sampling Methodologies and Estimation

Methods Applied to the Florida Keys/ Key West Visitors Survey. Strategic

284
 Environmental Assessments Division, National Ocean Service, National Oceanic

and Atmospheric Administration. Silver Spring, MD.

Lindsey, G and A Holmes, (2002). Tourist support for marine protection in Nha Trang,

Viet Nam. Journal of Environmental Planning and Management 45(4):461-480.

Loomis, J.B., (1992). The Evolution of a More Rigorous Approach to Benefit Transfer:

Benefit Function Transfer, Water Resources Research; 28(3), 701-705

Loomis, J.B., (1993). An investigation into the reliability of intended visitor behavior.

Environmental and Resource Economics 3:183-91.

Loomis, J.B., (1999). Contingent Valuation Methodology and the US Institutional

Framework. (613-628). In: I.J. Bateman & K.G. Willis (Eds) Valuing

Environmental Preferences. Theory and Practice of the Contingent Valuation

Method in the US, EU and Developing Countries. Oxford University Press, 1999.

Loomis, J.B., M. Creel and T. Park, (1991). Comparing Benefit Estimates from Travel

Cost and Contingent Valuation Using Confidence intervals for Hicksian Welfare

Measures. Applied Economics 23:1725-1731.

Mathieu, L.F., I.H. Langford, and W. Kenyon, (2003). Valuing Marine Parks in a

Developing County: a Case Study of the Seychelles. Environment and

Development Economics 8:373-390.

McFadden, D., (1974). Conditional Logit Analysis of Qualitative Choice Behavior. In P.

Zarembka, (Ed) Frontiers in Econometrics. Academic Press: New York, NY.

285
McGonagle, M.P. and S. K. Swallow, (2005). Open Space and Public Access: A

Contingent Choice Application to Coastal Preservation. Land Economics

81(4):477-495.

Mercado L.Y., (2003). Economic Valuation of Environmental Resources and Sustainable

Development in the Caribbean Region: A Case Study of Tourist Preferences in

the Area of Punta Cana, Dominican Republic. PhD, Cornell University; 204pp.

Ministry of Transport and Works, Jamaica (2006). The Highway 2000 Project.

http://www.mtw.gov.jm/works/projects/ongoing/highway_2000.htm

Mitchell R.C. and R.T. Carson, (1989). Using Surveys to Value Public Goods: the

Contingent Valuation Method. Washington DC: Resources for the Future.

Mitchell R.C. and R.T. Carson, (1995). Current Issues in the Design, Administration,

and Analysis of Contingent Valuation Surveys. In Current Issues in

Environmental Economics; Johansson, P.O., Kristrom, B., Maler, K. G., Eds.;

Manchester University Press: Manchester, U.K., 1995.

National Oceanic and Atmospheric Administration, NOAA, (1993). Natural Resource

Damage Assessments under the Oil Pollution act of 1990; Federal Register

58:4601 4614.

Neter J., M.H. Kutner, C.J. Nachtsheim, and W. Wassermann, (1996). Applied Linear

Statistical Models. McGraw-Hill.

Office of the Contractor General of Jamaica, (2006). Report of Investigation conducted

into the Sandals Whitehouse Hotel Project.

http://www.ocg.gov.jm/website_files/media_releases_issued/sandals_report.pdf

286
Organization of American States, OAS Document, (2000). Citizen Management of

Public Parks in Jamaica. http://www.ispnet.org/Documents/jamaica.htm.

Parsons G., (2003). The Travel Cost Method. In P.A. Champ, K.J. Boyle and T.C.

Brown (Eds.), A Primer on Nonmarket Valuation; Vol 3, pp. 576. Dorcdrecht:

Kluwer Academic Publishers.

Parsons G.R. and S. Thur, (2008). Valuing changes in the quality of coral reef

ecosystems: A stated preference study of SCUBA diving in the Bonaire National

Marine Park. Environmental Resource Economics (40) 4: 593-608.

Pielke Jr., R. and C.W. Landsea, (1998). Normalized hurricane damage in the United

States: 1925-95. Weather and Forecasting, American Meteorological Society

13:621-631.

Ramsaran, R., (1995). Ed. The Savings/Investment Environment in the Caribbean. St.

Augustine: University of the West Indies, Caribbean Centre for Monetary Studies.

Randall A., (1986). The possibility of satisfactory benefit estimation with contingent

markets. In: Cummings, R.G., Brookshire, D.S., Schulze, W.D. Eds., Valuing

Environmental Goods: An Assessment of the Contingent Valuation Method.

Rowman and Allanheld, NJ, 1986

Randall, A., J.P Hoehn and D.S. Brookshire (1983). Contingent valuation surveys for

evaluating environmental assets. Natural Resources Journal 23: 635648.

Reid-Grant, K. and M. Baht, (2009). Financing Marine Protected Areas in Jamaica: An

Exploratory Study. Marine Policy 33:128-136.

287
Richmond, R.H., (1993). Coral reefs: Present problems and future concerns resulting

from anthropogenic disturbance. American Zoologist, 33:524536

Ruitenbeek, H.J. and C.M. Cartier, (1999). Prospecting for marine biodiversity: a case

study in Montego Bay, Jamaica (Chapter 13). In Turner K., Bateman I, Adger N

(eds) Economics of Coastal and Water resources: Valuing Environmental

Functions. Kluwer Academic Press, Dordrecht.

Smith V.K., (2006). Fifty years of contingent valuation. In Handbook on Contingent

Valuation. Alberini and Kahn eds. Edward Elgar Publishing Inc, UK, USA.

Spash, C.L., (2000). Assessing the benefits of improving coral reef biodiversity: The

contingent valuation method. In Cesar, HSJ, ed. Collected Essays on the

Economics of Coral Reefs. Cordio: Kalmar, Sweden. pp242.

Spurgeon J.P.G., (1992). The Economic Valuation of Coral Reefs. Marine Pollution

Bulletin, 24(11):529-536.

Thur, S., (2003). Valuing Recreational Benefits in Coral Reef Marine Protected Areas:

An Application to the Bonaire National Marine Park. PhD Dissertation, College

of Marine Studies, University of Delaware.; 175pp

Todd, S., (2006). A Framework To Manage Jamaicas Protected Areas; No.25. Jamaica

Biodiversity Clearing House Mechanism.

http://www.jamaicachm.org.jm/Articles/Current%20Article.asp

Tongson, E and M Dygico, (2004). User fee system for marine ecotourism: the

Tubbataha Reef experience. Coastal Management 32:17-23.

288
Train, K., (2003). Discrete Choice Methods with Simulation. Cambridge UK: Cambridge

University Press.

Train K. and M. Weeks, (2005). Discrete choice models in preference space and

willingness to pay space. In Alberini and Scarpa (eds.), Applications of

Simulation Methods in Environmental Resource Economics, Springer &

Dordrecht, Netherlands.

UVI, (2001). Threats to Coral Reefs. http://www.uvi.edu/coral.reefer/threats.htm .

Ward, F.A., (1987). Economics of water allocation to stream in uses in a fully

appropriated river basin: Evidence from a New Mexico wild river. Water

Resources Research 23: 381-92.

Whitehead, J.C., S.K. Pattanayak, G.L. Van Houtven and B.R. Gelso, (2005).

Combining Revealed and Stated Preference Data to Estimate the Nonmarket

Value of Ecological Services: An Assessment of the State of the Science.

Department of Economics, Appalachian State University. Working Paper, Series;

No. 05-19.

Whittington, D., (1998). Administering Contingent Valuation Surveys in Developing

Countries. World Development, 2 (1): 21-30.

Wielgus, J., N. Chadwick-Furman, N. Zeitouni, and M. Shechter, (2003). Effects of coral

reef attribute damage on recreational welfare. Marine Resource Economics, 18:

225-37.

289
World Travel and Tourism Council, (2004). Jamaica: Travel and Tourism Forging

Ahead. Travel & Tourism Economic Research. Oxford Economic Forecasting,

London.

Wright, M., (1995). An Economic Analysis of Coral Reef Protection in Negril, Jamaica.

Working Paper, Series No. 11, Insights in Environment and Development.

University of the West Indies, Centre for Environment and Development 1995

290