ERF presentation 20073-2.ppt · 2013. 10. 23. · Ratio of Fully Compensated P e # of / # of sample...
Transcript of ERF presentation 20073-2.ppt · 2013. 10. 23. · Ratio of Fully Compensated P e # of / # of sample...
-
10/23/2013
1
Ami K Kang & George R. ParsonsMarine Policy Program University of DelawareEstuarine Research Federation Conference, RI
5th Nov. 2007
Non-Monetary Compensation PolicyFor Beach Closure:
An Application Using the Random Utility Model
2
Purpose
To develop travel cost random utility models for beach choice and
to estimate use values of beach closures in monetary and non-monetary terms
on the Gulf coast of Texas
-
10/23/2013
2
3
Study Area:Padre Island National Seashore line
Source: http://www.nps.gov/pais/naturescience/coasts.htm
4
• Demographic Information• Beach Site Characteristics• Beach Choice Information• Beach Trip Date-Month, Weekday/Weekend• Nested Logit Model (Random Utility Model)
Welfare Measurement
-
10/23/2013
3
5
Sampling
A map made by AMI 2007 Oct
6
Choice Set
Map made by AMI 2007 October
-
10/23/2013
4
7
0
20
40
60
80
100
120
140
160
180
200
Region (beach)
Wei
ghte
d V
isit
s
Beach Visits by Beach
SabinePass
Galveston Freeport CorpusChristi
SouthPadreIsland
PINS
Freeport CorpusChristi
PortLavaca
8
Weekend PINS Visits
0
5
10
15
20
25
30
5 6 7 8 9
PAIS North Beach PAIS Malaquite Beach PAIS South Beach`
PAIS Little Shell Beach PAIS Big Shell Beach PAIS Mansfield Cut
Month
-
10/23/2013
5
9
Nested Logit Model
Decision
Northern
Central
SouthernNo Beach
No Go Go
10
Estimated Coefficients of Household Characteristics
Variable Estimate Std.Err Variabe Estimate Std.Err
Weekend -1.20 0.06 Owning a beach property -0.25 0.11
Child -0.10 0.06 Owning a boat -0.36 0.07
Log(Age) 0.33 0.10 Owning a pool 0.30 0.08
Retired -0.16 0.14 Owning a surfing equipment -0.07 0.06
Spanish -0.20 0.10 June 0.15 0.08
HighSchool 0.17 0.08 July 0.27 0.09
College -0.10 0.08 August 0.20 0.09
Graduate -0.30 0.11 September 0.94 0.12
Full tile job -0.15 0.07 Constant 5.12 0.48
Woman 0.09 0.06
-
10/23/2013
6
11
Estimated Coefficients of Site Characteristics
variable Estimate Std.Err Variable Estimate Std.Err
Travel cost -0.03 0.00 Concession -0.48 0.10
Time -0.15 0.06 Machine Cleaning 0.88 0.11
Time* Income 0.00 0.00 Manual Cleaning 0.49 0.13
Gulf coast 0.42 0.14 No-fishing -0.14 0.12
Restroom 0.46 0.11 No-swimming -0.58 0.24
Lifeguard 0.27 0.12 Remote -0.28 0.12
Statepark -0.01 0.23 Closing -0.60 0.17
May_Padre 1.36 0.32 Vehicle free 0.42 0.13
June_Padre 1.75 0.32 Vehicle free area 0.48 0.14
July_Padre 2.49 0.30 Red tide -0.97 0.22
August_Padre 1.13 0.40 Length 0.23 0.04
September_Padre 2.57 0.37
12
Criteria to Measure the Efficiency of Non-Monetary Compensation Policy
The potential Kaldor-Hicks efficiency criterion: A winner can potentially compensate a loser.
The ratio of population who are fully compensated
BeforeLosing
TheBeach
AfterLosing
TheBeach
AfterNew
Services/Goods
< Person A>
BeforeLosing
TheBeach
AfterLosing
TheBeach
AfterNew
Services/Goods
< Person B>
-
10/23/2013
7
13
Policy Analysis: Machine CleaningKaldor Hick Index / Machine Cleaning/ Weekend/
each 26 beach
0
0.20.4
0.60.8
1
1.21.4
1.6
Texa
s Po
int
Palm
Bea
ch
Gal
vest
on
Jam
aica
Surf
side
Mag
nolia
Port
Mat
agor
da
Nor
th B
each
Cole
Par
k
Pack
ery
Kauf
er
Fred
Sto
ne
Beach Name
KH in
dex
May July
14
Fully Compensated People/ Machine Cleaning/ Weekend/each 26 beach
0
0.2
0.4
0.6
0.8
1
Texa
s Po
int
Palm
Bea
ch
Gal
vest
on
Jam
aica
Surf
side
Mag
noli
a
Port
Mat
agor
da
Nor
th
Col
e Pa
rk
Pack
ery
Kauf
er
Fred
Sto
ne
Beach
May July
Policy Analysis – Machine Cleaning (Cont.)
-
10/23/2013
8
15
KH Index/ May Weekend
0
2
4
6
8
10
12
14
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
# of Beach with NEW Machine Cleaning Service
Kald
or H
icks
Inde
x
Policy Analysis – Machine Cleaning (Cont.)
Percentage of People with Full Compensation/ May Weekend
0
20
40
60
80
100
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
# of Beach with NEW Machine Cleaning Service
%
16
Equality Issues
Damage & Net Gain Comparison
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
Corpus Christi Gulf Adjacent Inland Houston
Region
Mea
n W
elfa
re
damage Fort Crockett
-
10/23/2013
9
17
Future work
Estimating a Mixed Logit Model
Estimating an Individual_Level Parameter Mixed Logit Model
Coding a searching algorithm to suggest more efficient policy
18
-
10/23/2013
10
19
20
-
10/23/2013
11
21
22
Policy Analysis – Machine Cleaning
Kaldor Hick Index / Machine Cleaning/ Weekend/ each 26 beach
0
0.5
1
1.5
2
2.5
Texa
s Poi
nt
Palm
Bea
ch
Gal
vest
on
Jam
aica
Surf
side
Mag
nolia
Port
Mat
agor
da
Nor
th B
each
Cole
Par
k
Pack
ery
Kau
fer
Fred
Sto
ne
Beach Name
KH
inde
x
May June July August September
-
10/23/2013
12
23
Fully Compensated People/July
Fully Compensated People/ Machine Cleaning/ Weekend/each 26 beach
0
0.2
0.4
0.6
0.8
1
Texa
s Po
int
Palm
Bea
ch
Gal
vest
on
Jam
aica
Surf
side
Mag
noli
a
Port
Mat
agor
da
Nor
th
Col
e Pa
rk
Pack
ery
Kauf
er
Fred
Sto
ne
Beach
July August
24
Status Quo
AfterLosing
PDR
NewPolicy
Applied
Newlygain
Net LackStatus
QuoDamage
AfterLosing
PDR
NewPolicy
Applied
Status Quo
AfterLosing
PDR
NewPolicy
Applied
Newlygain
Net Surplus
Status Quo
Damage
AfterLosing
PDR
NewPolicy
Applied
∑(NetLack)∑(NetSurplus)
Potential Kaldor Hicks Crite
Ratio of Fully Compensated Pe
# of / # of sample
Criteria to Measure the Efficiency of Non-Monetary Compensation Policy
-
10/23/2013
13
25
Fully Compensated People/ Machine Cleaning/ Weekend/each 26 beach
00.10.20.30.40.50.60.70.80.9
1
Texa
s P
oint
McF
adde
nP
alm
Bea
chFo
rtG
alve
ston
Trea
sure
Jam
aica
San
Lui
sS
urfs
ide
Por
t Alto
Mag
nolia
Indi
anol
aP
ort
Mat
agor
daM
atag
orda
San
Jos
eN
orth
Bea
chM
cGee
Col
e P
ark
Mus
tang
Pac
kery
Kle
berg
Kau
fer
Dru
m P
oint
Fr
ed S
tone
Boc
a C
hica
Beach
May June July August September
Policy Analysis – Machine Cleaning (Cont.)
26
Compensatory Policy Analysis -Lifeguard
0.00
0.05
0.10
0.15
0.20
bc36 bc37 bc38 bc41 bc43 bc44 bc46 bc49 bc56 bc57
Beach service applied
Hic
ks K
alde
r Ind
ex
May June July August September
10 beaches in the central areaWeekend / May
-
10/23/2013
14
27
Compensatory Policy Analysis (Cont.)- Lifeguard
0
0.05
0.1
0.15
0.2
0.25
1 2 3 4 5 6 7 8 9 10Number of the central region beach with new servic
Hic
k-K
alda
r In
dex
May/Weekend
10 aggregated beaches in the central areaWeekend / May
28
Compensatory Policy Analysis –Lifeguard (cont.)
42 applicable beachMay/ Weekend
0.000.100.200.300.400.500.600.700.800.901.00
Beach ID
HK
May June July August September
-
10/23/2013
15
29
0
0.5
1
1.5
2
2.5
3
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Number of beach with new service
Hic
ks-K
arda
r In
dex
Compensatory Policy Analysis –Lifeguard (cont.)
42 aggregated beachMay/ Weekend
30
0
0.2
0.4
0.6
0.8
1
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Applied Beach number
May/Weekend
Compensatory Policy Analysis –Lifeguard (cont.)
Perc
enta
ge o
f peo
ple
who
is fu
lly co
mpe
nsat
ed 42 aggregated beachMay/ Weekend
-
10/23/2013
16
31
More?
32
Compensatory Policy Analysis-Machine Clean
-0.10
0.10
0.30
0.50
0.70
0.90
1.10
1.30
bc37 bc41 bc42 bc43 bc44 bc46 bc49 bc56 bc57Month
HK
eff
icie
ncy
MayJunJulAugSep
9 beach in the central areaWeekend May
-
10/23/2013
17
33
How to measure the Welfare Loss?ΔWelfare = Expected Utility BEFORE the event
-Expected Utility AFTER the event
Expected Utility
Base Case After the Accident
?
34
Random Utility Model
ijijijij XU
ijU Utility of a beach j to an individual i
ij Vector of coefficients of beach j orindividual i characteristics varibles
ijX Vector of beach j and individual i characteristics
ij Error components
-
10/23/2013
18
35
Compensatory Policy Analysis (cont.)-Machine Clean
0.00
0.50
1.00
1.50
2.00
2.50
beach id
HK
May June July August September
26 beach Weekend May
36
Calibrated Recreational Value of 6 Public Beaches in PNS
-
10/23/2013
19
37
Recreational Value of 6 Public Beaches in PNS
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
5 6 7 8 9
Month
$ (2
001)
WeekDay Weekend
?
38
Possibly Over-stated sample• Graph between sample and NPS
-
10/23/2013
20
39
Random Utility ModelNested Logit
• Merits:1.Taste Heterogeneity among population2. 3.
ijijjij XU ),(~ sdm
40
Nested Logit Model
Decision
Not Go Go
No Beach Northern Area Central Area Southern Area
-
10/23/2013
21
41
Oil Spill/Red Bloom and Beach Recreation
• http://celebrating200years.noaa.gov/magazine/recreation_restoration/welcome.html#comp
42
Beach closing in Texas• http://www.texasep.org/html/wql/wql_5cst_gulf.html
By oil spillBy Red tide
-
10/23/2013
22
43
• Insert Texas Beach map:high lighted nesting structure
44
Data
• Research focusing area:6 public beaches in Padre Island National
Shoreline Park///Galveston?• Revealed preference data:2001/5 waves/884 full participants and 565
after data cleaning/2704 tripsDiscrete choice modelUsing demographic and site characteristic
variable
-
10/23/2013
23
45
Mixed Logit Estimates
• Triangular…100 halton draws• Insert a table
46
Candidates of Non-Monetary Compensatory Equivalents
From Model
|Variable| Coefficient | Standard Error |b/St.Er.|P[|Z|>z]|+--------+--------------+----------------+--------+--------+RSTR | .45512310 .11287491 4.032 .0001LFGRD | .26994272 .11698975 2.307 .0210MCHCLN | .88386254 .10994835 8.039 .0000MANCLN | .49402739 .12523103 3.945 .0001NFS | -.14450127 .12062248 -1.198 .2309NSWIM | -.57962945 .24182955 -2.397 .0165VFR | .41896868 .12941636 3.237 .0012VFRA | .48434288 .14494130 3.342 .0008
-
10/23/2013
24
47
Non-Monetary Public Service Applied Beaches in the Region5
Insert a map which highlighted the region5
Go[NBch(s1,s2,s3,s4,s5,s6,s7,s8,s9,s10,s11,s12,s13,s14,s15,s16,s17,s18,s65,s19,s20,s21,s22,s23,s24,s25,s26,s27,s28,s29,s30,s31,s32,s33,s34,s35),
CBch(s50,s51,s52,s53,s54,s55,s36,s37,s38,s39,s40,s41,s42,s43,s44,s45,s46,s47,s48,s49,s56,s57),SBch(s58,s59,s60,s61,s62,s63,s64)]
48
Non-monetary compensatory policy
Scenario 1: Expanding Manual Cleaning in the Region3 based on
distance proximity
insert a graph
-
10/23/2013
25
49
Non-monetary compensatory policy
Scenario 1: Expanding Machine Cleaning Service in the Region3
based on distance proximity
insert a graph
50
Non-monetary compensatory policy
• Scenario 2:• insert a graph
-
10/23/2013
26
51
Or…
• Result of basic searching algorithm
52
dfdfdf
• Good
-
10/23/2013
27
53
Jsijfdd