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


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


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


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



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


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


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


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


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


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

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


Scenario 1: Expanding Machine Cleaning Service in the Region3

based on distance proximity

insert a graph

50


• 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

erf presentation 20073-2.ppt · 2013. 10. 23. · ratio of fully compensated p e # of / # of sample...

Documents