Download Economic Assessment of Rapid Land

Preferences for Timing of Wetland Loss
Prevention in Louisiana
Ross Moore, Daniel Petrolia, and Tae-goun Kim
Dept. of Agricultural Economics,
Mississippi State University
May 28, 2010
Motivation





Approximately 40 percent of the coastal wetlands of the lower 48
states is located in Louisiana
Coastal Louisiana lost 1,900 square miles from 1932 to 2000
Has lost an average of 34 square mile per year for the last fifty
years
Hurricanes Katrina and Rita destroyed more than 217 square
miles of marsh in a single season.
By the year 2050 an additional 700 square miles is projected to
be lost
Restoration Projects




Congress passed the Coastal Wetlands Planning, Protection
and Restoration Act (CWPPRA) in 1990
Designates approximately $60 million annually for work in
Louisiana.
April of 2007 the Louisiana Governor signed Louisiana’s
Comprehensive Master Plan for a Sustainable Coast
“A sustainable landscape is a prerequisite for both storm
protection and ecological restoration.” – CPRA Master Plan
Objectives

Estimate the value that the public of Louisiana
places on preventing the future loss of wetlands
within their state

Identify motivating factors that have an effect on
willingness to pay/accept: demographics, proximity to the
coast, risk preferences and risk perceptions, time
preference of money, climate change, confidence in
government, believe responses will influence decisions,
previous knowledge of coastal protection efforts.
Survey




Mail survey sent to 3,000 taxpaying households in Louisiana.
Stratified by county population
Survey design
 Willingness to Pay/Willingness to Accept
 Long Run/Short Run Proposal
 Order of the two proposals
Response Rate: 681 surveys (22.7%)
Survey: Preference
We would now like to ask about your relative preferences between the two proposals.
Please review the descriptions of each below and check the box above the one you would
prefer if given the choice.
Proposal #1
Proposal #2
Future losses prevented:
Future losses prevented:
Starting in 2015
Ending in 2050
Starting in 2035
Ending in 2185
Tax Cost:
$X per year for 10 years
Tax Cost:
$X per year for 10 years
No Action
Future losses NOT
prevented.
Tax Cost:
$0
Preference
Preference
Short run proposal
Long run proposal
No action
Total
Frequency
370
34
107
511
Percentage
72.41%
6.65%
20.94%
Potential Benefits
Potential benefits
Frequency Percentage
Storm protection
282
55.19%
Protection of recreational opportunities
16
3.13%
Protection against sea-level rise due to climate change
25
4.89%
Protection of the environment/ecosystem
102
19.96%
Protection of commercial fisheries
8
1.57%
Other
11
2.15%
No potential benefits in mind
67
13.11%
Total
511
Category 3 or Greater Hurricane Expectation
Expected Hurricane Frequency
Once a year or more
Once every 2-5 years
Once every 10 years
Once every 20 years
Once every 30 years
Once every 50 years
Once every 100 years or more
I don’t know
Total
Frequency Percentage
69
13.50%
204
39.92%
105
20.55%
29
5.68%
9
1.76%
7
1.37%
24
4.70%
64
12.52%
511
Actual Hurricane Frequency
Multinomial Logit
Model Variables and Descriptions
Variables
Income-Bid
Type
Ordered
Categorical
Gender
Race
Age
Household
Binary
Binary
Continuous
Ordered
Categorical
Ordered
Categorical
Description
( y j  t j )
ln 

 t j

1 if male; 0 if female
1 if white; 0 otherwise
Continuous between 19 - 84
Household size 1 if # is 1; 2 if # is 2; 3 if #; 4
if # is 4; 5 if # is 5 or greater
Education
Highest level of education 1 if some school
or high school; 2 if associates or bachelors; 3
if masters, professional, or doctoral
Latitude
Continuous Latitude based upon zip code of respondent
StormBenefit
Binary
1 if storm protection was most important
benefit, 0 if otherwise
EnivronmentBenefit
Binary
1 if environment protection was most
important benefit, 0 if otherwise
CCBenefit
Binary
1 if protection against sea-level rise due to
climate change was most important benefit, 0
if otherwise
Mean
5.15
0.56
0.76
54.31
2.46
1.64
30.69
0.54
0.18
0.06
Multinomial Logit
Model Variables and Descriptions
Variables
CCperception
Type
Binary
PreKnowledge
Binary
Government
Binary
Influence
Binary
RiskPref
Binary
HurrFreqHI
Binary
LongRunFirst
Binary
WTP
Binary
Description
Mean
1 if respondents do not at all believe in
0.13
climate change; 0 otherwise
1 if respondent had prior knowledge of
0.76
actions to protect wetlands; 0 otherwise
1 if no confidence that government agency
0.43
can accomplish such actions; 0 otherwise
1 if respondents believe responses will
0.20
influence policy; 0 otherwise
1 if respondents does not take a gamble; 0
0.69
otherwise
1 if respondent believes a Category 3
0.73
hurricane will affect them between 1 and 10
years; 0 otherwise
1 if long run proposal was presented first; 0 if 0.51
short sun was presented first
1 if the payment mechanism was willingness
0.47
to pay; 0 if willingness to accept
Estimated Coefficients, Standard Errors, Average Marginal Effects,
and Significance Levels for the Multinomial Logit Model.
Short Run
Long Run
SE
ME (%)b
Coef
SE
ME (%)b
0.1
0.0009c
0.33
0.1
0.0001c
0.32
0.55
0.05
0.32
0.06
0.34
8.61
0.76
0.34
0.96
0.02
0.01
0.38
-0.03
0.02
-0.23
0.1
0.13
1.07
0.1
0.13
0.12
Educationa
-0.06
0.2
-0.58
-0.06
0.2
-0.07
Latitudea
0.09
0.17
0.94
0.09
0.17
0.1
0.32
29.13
2.33
*
0.32
2.91
EnvironmentBenefita 2.8
*
0.53
CCBenefit
1.26
0.67
a
CCperception
-0.06
0.37
PreKnowledgea
0.69
*
0.33
* Significant at p = 0.05 level or greater
22.94
2.3
2.8
2.46
*
*
0.53
0.79
2.4
11.39
-0.65
-0.06
0.37
-0.07
7.81
0.69
0.33
0.87
Variable
Income-Bida
Coef
Gendera
0.05
Racea
0.76
Age
Householda
StormBenefit
a
0.33
2.33
*
*
*
*
*
*
a
Coefficient constrained to be equal across Short Run and Long Run Equations
b
Marginal effects shown for binary variables are for a discrete change from the base
c
Marginal effect for a $1,000 change in income
Estimated Coefficients, Standard Errors, Average Marginal Effects,
and Significance Levels for the Multinomial Logit Model.
Short Run
Long Run
SE
ME (%)b
Coef
0.32
-11.52
-1.06
0.11
0.41
1.17
0.23
0.34
HurrFreqHIa
0
LongRunFirst
0
Variable
Governmenta
Coef
-1.06
Influencea
RiskPrefa
WTP
-1.56
Constant
-5.87
*
*
SE
ME (%)b
0.32
-1.3
0.11
0.41
0.13
2.39
0.23
0.34
0.27
0.36
-0.04
0
0.36
0
0.29
8.79
-2.68
0.72
-11.35
0.31
-20.49
-0.84
0.5
2.24
-5.62
5.68
5.65
*
*
Observations = 511
Log Pseudoliklihood = -282.80
Wald chi2(22) = 138.51
Prob > chi2 = 0.00
Pseudo R2 = 0.28
* Significant at p = 0.05 level or greater
a
Coefficient constrained to be equal across Short Run and Long Run Equations
b
Marginal effects shown for binary variables are for a discrete change from the base
c
Marginal effect for a $1,000 change in income
Parametric and Turnbull Nominal (Annual) Willingness to Pay
and Willingness to Accept Estimates
95% Confidence Interval
Estimates Lower Bound Upper Bound
Multinomial Logit
$33,067
$1,493
$3,943
$60,430
$13,651
$53,855
Short Run:
WTP
WTA
Long Run:
$0.00
$0.78
WTP
$0.11
$10
WTA
Turnbull Distribution Free Estimate
$742
$746
Short Run
$24
$169
$749
Net Present Value of Willingness to Pay for the Short
Run Proposal
$350,000
$300,000
Median
Lower
Upper
Turnbull
NPV of WTP
$250,000
$200,000
$150,000
$100,000
$50,000
$0
0
0.1
0.2
0.3
Discount Rate
0.4
0.5
0.6
Net Present Value of Willingness to Accept for the
Short Run Proposal
$600,000
$500,000
Median
Lower
Upper
Turnbull
NPV of WTA
$400,000
$300,000
$200,000
$100,000
$0
0
0.1
0.2
0.3
Discount Rate
0.4
0.5
0.6
Present Value Estimates of Aggregate Welfare (millions of
dollars) by discount rate
Discount Rate
2%
6%
18%
26%
WTP
$76,465
$62,000 $37,542 $28,963
WTA
$1,044,282 $846,723 $512,706 $395,544
Turnbull
$14,459
$11,724
$7,099
$5,477
Assuming $0 WTP/WTA for Non-Respondents
2%
6%
18%
26%
WTP
$17,358
$14,074
$8,522
$6,575
WTA
$237,052 $192,206 $116,384 $89,788
Turnbull
$3,282
$2,661
$1,611
$1,243
50%
$16,634
$227,168
$3,145
50%
$3,776
$51,567
$713
Summary of Results

Probability of choosing short run over no action:


Increases:
 Income
 White
 Storm protection primary benefit
 Environmental benefits primary concern
 Had prior knowledge of protection efforts
Decreases:
 No confidence in government
 Received WTP payment mechanism
Summary of Results

The probability of choosing long run over no action


Increases:
 Income
 White
 Storm protection primary benefit
 Environment protection primary benefit
 Climate change primary benefits
 Prior knowledge of protection efforts
Decreases:
 No confidence in government
 Presented with long run first
Conclusions





Found respondents are highly willing to fund prevention of
wetland loss
Overwhelming support for short run proposals over long run
proposals
Protection from hurricane and storm damage is primary benefit
driving support
Other factors: Government, Payment Mechanism, and
Environmental Benefits
Problems


Low response rate
High welfare estimates
Thank you