Valuing the Visual Disamenity of Offshore Wind … Wind...state of Delaware for offshore wind turbines located at different distances from the coast. The annual costs to inland residents

Valuing the Visual Disamenity of Offshore WindPower Projects at Varying Distances from the Shore:An Application on the Delaware ShorelineAndrew D. Krueger, George R. Parsons, and Jeremy Firestone

ABSTRACT. Several offshore wind power projectsare under consideration in the United States. A con-cern with any such project is the visual disamenity itmay create. Using a stated preference choice model,we estimated the external costs to residents of thestate of Delaware for offshore wind turbines locatedat different distances from the coast. The annual coststo inland residents was $19, $9, $1, and $0 (2006dollars) per household for turbines located at 0.9,3.6, 6, and 9 miles offshore. The cost to residents liv-ing near the ocean was $80, $69, $35, and $27 perhousehold for the same increments. (JEL Q42, Q51 )

I. INTRODUCTION

The United States derives about 70% of itsenergy for electricity from fossil fuel sources. 'As regulators look to address climate changeconcerns and reduce dependence on foreignsources of oil, alternative energy sources ap-pear increasingly attractive as a way to reduceglobal carbon dioxide (CO2) emissions andincrease the domestic supply of energy. Cur-rently, wind power is the only utility-scale, re-newable, IOW-CO2 energy resource that islarge enough to become a significant fractionof electric supply (Kempton et al. 2005), Ap-proximately 24,000 megawatts (MW) of newwind power capacity was installed over thepast three years, breaking all previous recordsand increasing the nation's total wind powergenerating capacity to over 35,000 MW in 36states and making the United States the worldleader in installed capacity. The American

' Nuclear accounts for 20%, hydro 6.5%, and renewa-bles 3.5%. See the U.S. Energy Infortnation Administrationat www.eia.doe.gov/.

Land Economics • May 2011 • 87 (2): 268-283ISSN 0023-7639; E-ISSN 1543-8325© 2011 by the Board of Regents of theUniversity of Wisconsin System

Wind Energy Association (AWEA 2009) es-timated that wind power generated about1.5% of the U.S. electricity supply in 2008,powering the equivalent of 5,7 millionhomes,^ The Department of Energy has set agoal of 20% wind generation by 2030, includ-ing 54 gigawatts (GW) of offshore windpower (DOE 2008),

Although no wind turbines have been in-stalled offshore in the United States, there area number of proposals under consideration,^States such as Texas, Rhode Island, Delaware,Massachusetts, Michigan, Ohio, Wisconsin,Maine, Maryland, New York, New Jersey,Virginia, and North Carolina are all consid-ering wind power development off their coast-lines, with some issuing requests for proposalsand sponsoring competitions to select pre-ferred developers. The Bureau of OceanEnergy Management, Regulation, and En-

The authors are, respectively, researcher. College ofEarth, Ocean, and Environment, University of Dela-ware; program director and professor. Marine Policy,College of Earth, Ocean, and Environment, Univer-sity of Delaware; and associate professor. Marine Pol-icy and Legal Studies, College of Earth, Ocean, andEnvironment, University of Delaware,

^ See the American Wind Energy Association (AWEA)reports for data on the tJnited States at www.awea.org/pub-lications/reports. An additional 10,000 MW was added in2009, enough generation capacity for an additional 2.4 mil-lion homes, bringing tJ.S. generation to 35,000 MW (AWEA2010). By comparison, in 2008, wind power in the EuropeanUnion accounted for about 4.2% of ail electricity, and therewas approximately 65,000 MW of capacity installed. Ger-many at 24,000 MW and Spain at 17,000 MW accountedfor most of this capacity. See the European Wind and EnergyAssociation (EWEA) at www.ewea.org/.

3 It is a different story in Europe, where as of the end of2009 there are 30 offshore wind projects accounting formore than 2,000 MW of installed capacity, with an addi-tional 1,000 MW of capacity expected to be installed in 2010(EWEA 2010).

87(2) Krueger, Parsons, and Firestone: Diasmenity Values from Wind Power 269

forcement (formerly Minerals ManagementService), the federal agency in charge of reg-ulating offshore wind power, finished its finalenvironmental impact statement for the CapeWind Energy Project off of Cape Cod, Mas-sachusetts, and recently published its Recordof Decision to issue a commercial lease forthe United States' first offshore wind facility.Additionally, the first power purchase agree-ment in the United States has been reachedfor a proposed wind power project off theDelaware coastline.

The visual disamenity of and possible en-vironmental impacts associated with windpower projects often raise concerns in localcommunities considering such developments.The well-publicized public opposition to theCape Wind Energy Project off of Cape Cod,Massachusetts, is perhaps the most recog-nized example of this concem (Eirestone andKempton 2007).

From an economic standpoint, as the bene-fits and costs of offshore wind power are con-sidered, one should account for the potentialvisual disamenity of wind turbines located inseascapes valued for their natural beauty. Likepollutants from the burning of fossil fuels, orfear of accidents from the storage of nuclearwastes, visual disamenities associated withenergy projects, including wind power, are ex-temalities missed in the calculus of markets.Erom an efficiency standpoint, these extemaleffects should be brought into the social ac-counting. Interestingly, the economics of off-shore wind power is such that disamenitycosts are almost certain to decline with in-creased distance from the coast in the near-shore environment, while transmission,construction, and maintenance costs typicallyrise with distance.

In this paper we present the results of achoice experiment designed to value the vi-sual disamenity associated with wind turbinesin the waters off the Delaware coast. This areaof study is particularly interesting because itis favorable for wind power from a purelyphysical standpoint and has recently wit-nessed the first power purchase agreement foroffshore wind power in the Americas. Weconducted a mail survey over a stratified ran-dom sample of Delaware residents in the fallof 2006. We analyzed the choice data using

random utility theory and found that disamen-ity costs decline with distance from the coast,level off at approximately nine miles, and aresignificantly higher for people living nearerthe coast. A broader discussion of the surveyresults has been presented by Eirestone,Kempton, and Krueger (2008) and Krueger(2008). We begin with a brief review of valu-ation studies related to offshore wind powerprojects before presenting our model andresults.

II. VALUATION STUDIES RELATED TOLOCATION OF OFFSHORE WIND

POWER PROJECTS

Ladenburg (2009) provides a nice reviewof the valuation literature, limited as it is, onthe location of offshore wind power projects.This section draws heavily from his review.Table 1 is a list of the relevant studies. Threeof these, by Aravena, Martinsson, and Scarpa(2006) in Chile, Ek (2006) in Sweden, andMcCartney (2006) in Australia, are concernedonly with the value of the location of windpower projects in the broad sense of whetherthey are located offshore or onshore. Eor ex-ample, in the context of choice experiments,Aravena, Martinsson, and Scarpa (2006) andEk (2006) ask households to consider windpower projects in mountain versus inland ver-sus offshore locations. Neither give respon-dents specific geographic areas or specificdistances at which the wind power projectswould be located offshore. Both find a pref-erence, all else constant, for locating windpower projects offshore compared to onshorelocations. McCartney's (2006) study is differ-ent in that she gives respondents a specificgeographic area, which happens to be a ma-rine park. Given that setting, she finds a pref-erence for onshore versus offshore locationsof wind power projects. So, as one might ex-pect, whether households prefer wind powerprojects onshore or offshore is likely to de-pend on where onshore or offshore they arelocated. While useful for broader policy de-liberations, none of these studies helps alongthe lines of establishing the size of the disa-menity effect at different distances offshoreand hence the optimal siting for offshore windpower projects. The studies by Ladenberg and

270 Land Economics May 2011

TABLE 1Stated Preference Studies Addressing Location of Wind Power Projects Offshore

Author(s)

Aravena,Martinsson,and Scarpa2006

Ek 2006

McCartney 2006

Ladenburg andDubgaard2007, 2009

YearConducted

2005

2002

2004

2003-2005

Resource Studied

Wind power projectsin Chile

Wind power projectsin Sweden

Wind power projectin Jurien BayMarine Park,Australia

Wind power projectsin Denmark

Number ofRespondents

^=300Random draw of

residents frommetropolitan areaof Concepción

^ = 547Random draw of

Swedishpopulation

Af = 96Local residents and

tourists on-siteduring holidayweekend

N=^62Stratified random

sample of Danishpopulation (withtargeting of areaswith offshore windpower projects)

Method

Choice experiment.pictures notshown, paymentwith cost ofelectricity

Choice experiment.mail survey.pictures shown.payment withelectricity bill

Contingent valuation.in-person survey.pictures shown.payment withelectricity bill

Choice experiment.mail survey.pictures shown.payment withelectricity bill

How OffshoreDistance Gradient

Was Valued

Compared windpower projects inmountains, alongcoast, inland, andoffshore

Compared windpower projects inmountains, inland.and offshore

Compared windpower projectsinland, on beach.and offshore

Compared windpower projectslocated at 8, 12,18, and 50 kmoffshore

Dubgaard (2007, 2009) are the only ones todate that provided preference data on thevalue of that disamenity gradient.

Ladenburg and Dubgaard (2007, 2009)conducted a choice experiment in a mail sur-vey of 362 Danish residents. They consideredoffshore wind projects only. Their experimentposed different-sized wind projects at varyingdistances from shore (8, 12, 18, or 50 km) andat varying annual costs per-household (€0,€12.5, €23, €40, €80, or €175). They foundthat residents were willing to pay approxi-mately $58, $121, and $153 per household peryear (2006 dollars) to have a wind project lo-cated at 12, 18, and 50 km from the coast ver-sus a baseline of 8 km. Respondents livingnear the coast, having a summer home on thecoast, or engaging in recreational activities onthe coast reported significantly higher values.Our analysis builds directly on this work. Weconsider a location in the United States, con-sider distances closer to the coast (as near as0.9 miles [1.5 km]), use specific geographicareas, and use a more ñexible econometricmodel.

Before moving on to our analysis it isworth noting that there are some wind powervaluation studies of onshore locations andsome examining wind power in the broadercontext of its acceptability versus other re-newable energy sources. Some of these con-sider distance to residential areas and are ofinterest to our application. Meyerhoff, Ohl,and Hartje (2010) in Germany, for example,consider distances of 750, 1,000, and 1,500 mfrom residential areas and find that people in-deed prefer that wind power projects be lo-cated further from away from their homes.Fimereli, Mourato, and Pearson (2008) havea similar finding in the United Kingdom. Arecent hedonic study examining the effect ofwind projects on property values found no sta-tistically significant difference in sales priceamong homes located less than 3,000 feet,3,000 to 5,000 feet, 1 to 3 miles, or 3 to 5miles from the project, as compared to the ref-erence case of greater than 5 miles (Hoen etal. 2009).

Other studies that address preferences forwind power in various contexts, but do not


address the distance-value gradient or evenonshore versus offshore value, include thoseby Hanley and Nevin (1999), Alvarez-Farizoand Hanley (2002), Bergmann, Hanley, andWright (2006), Groothuis, Groothuis, andWhitehead (2008), Borchers, Duke and Par-sons (2007), Navmd and Braten (2007),Bergmann, Colombo, and Hanley (2008),Dimitropoulos and Kontoleon (2009), andKoundouri, Kountouris, and Remoundou(2009). As noted, most of these examine pref-erences for wind versus other sources of re-newable energy.

i n . SURVEY

Our survey began with 12 semistmcturedinterviews of Delaware residents to help usunderstand knowledge and perception of, andattitudes toward, energy issues. A pilot ver-sion of the mail survey was tested in personat the Department of Motor Vehicles in Wil-mington, Delaware, on June 29-30,2006. Thepilot test was used to refine wording and sur-vey format, to ensure that respondents under-stood the questions, to test the layout andusefulness of a map and photo prop, to gatherinformation on whether respondents per-ceived the survey instmment to be biased, andto see if the survey was appropriate in length.In particular, the test was critical to ensure thatthe choice experiment section was under-standable and realistic, that the attributes cho-sen and the range of their corresponding levelswere appropriate, that the respondents couldunderstand and properly complete the choiceexperiment questions, and that the questionswere producing usable data.

The final version of the survey has foursections. The first covers attitudes and opin-ions concerning wind power and the possibil-ity of having offshore wind power inDelaware. The second contains the choice ex-periment in which respondents are asked tochoose among two different offshore windpower scenarios and an opt-out fossil fuelpower scenario. The choice experiment is de-scribed in detail in the next section. The thirdand fourth sections cover beach use and dem-ographics. The protocol for survey constmc-tion, testing, and administration followedDillman's tailored design method (Dillman

2000) as closely as possible, given time andbudget constraints.

From September 9 to September 20, 2006,we mailed 2,000 surveys to a stratified ran-dom sample of Delaware households. Thethree strata are (1) households living in censusblock groups bordering the Atlantic Ocean,(2) households living in census block groupsbordering Delaware Bay, and (3) all otherhouseholds in the state. The initial samplefrom each of these strata was 400, 400, and1,200, respectively. Each mailing included (1)a cover letter describing the survey and whythe addressee's participation was important,(2) the survey booklet, (3) a map and photosimulations prop, and (4) a stamped return en-velope. Three weeks after the initial mailing,reminder postcards were sent out to thank allrespondents for their participation and to re-mind those respondents who had not yet com-pleted their survey to promptly do so.Following the postcard reminder, a secondmailing of 1,250 surveys was sent betweenOctober 28 and October 30, 2006, to thoseindividuals who had not yet returned theiroriginal completed surveys. These packetscontained a modified cover letter reaffirmingthe importance of the study, reminding re-spondents of the confidentiality of their an-swers, and asking respondents to take a fewminutes to complete and mail back the survey.

A total of 949 retumed surveys were usedin our final analysis. After accounting for badaddresses and for deceased and otherwise in-capacitated respondents (based on statementsmade by relatives), the response rate was52%. Table 2 provides some descriptive dataon the sample population. Survey respondentswere more likely to be male, older, andwealthier than the overall Delaware popula-tion. For additional detail on survey devel-opment see Firestone, Kempton, and Krueger(2008, 2009) or Krueger (2008)."

IV. CHOICE EXPERIMENT

Each respondent faced three hypotheticalreferenda in our choice experiment. Figures 1

"* The entire survey including the wind turbine visual isavailable at www.ceoe.udel.edu/windpower/docs/FinalDNRECOpinionReport.pdf.


TABLE 2Sample Means over Key Respondent Characteristics

Inland Bay Ocean

Sample SizeAgePercent retiredHousehold income^Gender (percent male)Mean distance from nearest beachAverage number of days per year spent at the beachPercent who have seen a wind turbine

56457

33.7$50,000-$75,000

68.135 miles

1454.3

20361

42.4$50,000-$75,000

69.54 miles

7659.4

18261

48.4$100,000-$ 150,000

69.70.6 miles

10472.9

" Median values.

Offshore Wind Power in Deiaware

Assume Delaware needs to increase its energy supply by 20% and that you bavethe opportunity to vote on energy deveiopment options. One way to meet the energydemand would be to place a 500-turbine wind farm offshore. Another option wouid be tocontinue Delaware's current energy policy and obtain additional energy from a new plantthat burns coal or natural gas.

Because wind energy uses no fuel, your eiectric bill wouid not increase over timedue to higher fuel cost; however, like other sources of electricity, it could increase forother reasons. To offset the initiai costs of providing wind energy to Delaware residents,assume that there wouid be a "Renewable Energy Fee" added each month to yourelectric bill, for the first three years only.

We are now going to ask you three questions where you get to vote on windpower development. For each question, assume the option that receives the most voteswiii be carried out.

Continue on and voie—>-

FIGURE 1Preamble to Choice Experiment Section

and 2 show the preamble and an examplequestion. Respondents were asked to considera scenario in which Delaware would be ex-panding its power capacity to meet future en-ergy needs. Respondents were then asked toconsider three development scenarios: twooffshore wind options and a status quo optionof expanding natural gas or coal power. Theattributes for the two wind options and theirlevels are shown in Table 3. These were se-lected based on a review of current regulatorypolicy and pertinent literature and from in-sight gained during semistructured interviews.People were told in the preamble leading upto the choice questions that 500 turbineswould be placed offshore. This was held con-stant throughout the experiment and was con-

sistent with the size of the wind power projectactually being considered. The project size is450 MW.

The first attribute is the location of thewind power project. In Delaware, there arethree logical areas: Delaware Bay, the north-em Atlantic coast (adjacent to the town of Re-hoboth Beach), or the southern Atlantic coast(adjacent to the town of Fenwick Island). Weused these three areas for candidate locationsand provided respondents with the map shownin Figure 3.

The second attribute is distance from shore.We provided a range of realistic distancesbased on current technological limitations andon existing proposals for offshore wind powerprojects off other parts of the U.S. Atlantic


18. Now, for which option wouid you vote?

Refer to the Delaware tnap itisert for the "wind farm location." Refer to theocean photo insert for simulated views of the wind farm at different distances.

Wind farmlocation

Distancefrom shore

Annualrent/royaity

Renewabieenergy fee

on yourmonthiy

eiectricitybiii for 3

years

Option A

Ocean (South)

0.9 miles

$1 million toBeach

NourishmentFund

$1

Option B

Ocean (North)

6 miles

$8 million toBeach

NourishmentFund

$20

Option C

No wind power

Expansionof coal

cr natural gaspower

I would vote for...

D Option AD Option BG Option C

EIGURE 2

Sample Choice Experiment Question

coast. A page with photo simulations was in-cluded to help respondents visualize changesthat would occur to the seascape if wind tur-bines were to be placed at different distancesfrom shore. The unusual increments (e.g,, 0,9instead of 1,0 miles) correspond to actual dis-tances under consideration. In all cases thewind turbines shown were 440 feet high. Thenacelle was 258 feet high and the blades ex-tended that to 440 feet. These were also theturbines actually being considered for theDelaware project.

The third and fourth attributes are the typeof royalty fund and the amount of payment inthe fund. We used Delaware's existing GreenEnergy Fund (which subsidizes home solar

TABLE 3

Attributes and Levels for the Choice Experiment

Attribute Levels

Location of wind farm

Distance from shore(miles)

Royalty fund

Renewable payment

Renewable energy fee

Delaware Bay; RehobothBeach; Fenwick Island

0.9; 3.6; 6; 9; too far out to see

Beach nourishment fund;Delaware green energy fund;Delaware general fund

$1 million; $2 million; $8million

$0;$l;$5;$10;$20;$30


DELAWAREBAY

Cape May Point

ATLANTICOCEAN

FIGURE 3Delaware Map Depicting Hypothetical Offshore Wind Development Areas

and wind power, among other things), theState General Eund, and a hypothetical BeachNourishment Eund. We included the latter totest whether individuals living near the coast-line would be more willing to accept a visualdisamenity from wind turbines if a nearbybeach benefits from the collected revenues.The range of royalty revenues to be collectedby the state was based on low and high esti-mates of other onshore and offshore wind pro-ject royalty payments.

The final attribute is a renewable energyfee, which was used as the payment vehicle.It was chosen because it is related to the de-livery of the electricity, was believable in ourpretests, and is easily understood. We used amonthly payment period of 3 years. Theranges were determined based on the pretest.

The five attributes and their correspondinglevels, presented in Table 3, result in 810 pos-sible treatment combinations. We used an or-thogonal main effects only design to generate


TABLE 4Twenty-Five Choice Combinations Used in the "Tailored" Choice Question

SurveyVersion

12345678910II1213141516171819202122232425

Option A

Distance

0.9 miles0.9 miles0.9 miles0.9 miles0.9 miles3.6 miles3.6 miles3.6 miles3.6 miles3.6 miles6 miles6 miles6 miles6 miles6 miles9 miles9 miles9 miles9 miles9 miles

0.9 miles3.6 miles6 miles

0.9 miles3.6 miles

Fee

$0$1$5

$10$20$0$1$5

$10$20$0$1$5$10$20$0$1$5

$10$20$5$10$5

$10$1

Option B

Distance

TFTS9 miles6 miles

3.6 milesTFTS

6 miles9 milesTFTS

6 miles9 miles9 milesTFTS

9 milesTFTS

9 milesTFTSTFTSTFTSTFTSTFTS

3.6 miles6 miles9 milesTFTS

9 miles

Fee

$5$10$20$30$30$10$20$10$20$30$1

$30$30$30$30$20$5$20$20$30$10$30$20$20$10

Options

Royalty

GreenBeachGreenBeach

GeneralGreenBeach

GeneralGeneralGreenBeachGreenGreenBeachBeach

GeneralGeneralGeneralBeachBeachGreenBeachGreen

GeneralGreen

AandB

Fund

$2 M$8 M$8 M$2 M$1 M$1 M$1 M$8 M$2 M$2 M$8 M$1 M

' $8 M$1 M$2 M$1 M$2 M$8 M$2 M$8 M$8 M$1 M$1 M$2 M$2 M

Note: M, million; TFTS, too far out to see.

25 choice experiments for two of the threechoice questions using standard SAS macrosand following Hensher, Rose, and Greene(2005). The other choice question was de-signed to focus specifically on the trade-offbetween distance from shore and willingnessto pay. In this question, the respondent waspresented with two offshore wind power de-velopment options that were identical in everyrespect except that they were at two differentdistances from shore, with the further distancealways having a higher fee. The levels of theother attributes varied across respondents butwere held constant in each choice experiment.We tailored one choice question in this waybecause of our focus on estimating the valueof the distance gradient. This question gaverespondents a choice that focused on thattrade-off. The question also did well in dis-cussions with respondents following the pre-tests. The pairings used for our tailoredquestion are shown in Table 4. While there isno doubt room for added statistical efficiency

in the design, our format introduces no biasand, given the standard errors on our esti-mates, which we discuss shortly, there appearsto be adequate statistical efficiency.^

V. RUM MODEL

We modeled the choices made by respon-dents using a conventional random utilitymaximization (RUM) model. In the modeleach individual n faces a decision among a setof three altematives: two wind projects (/ =1,2) and the status quo expansion of fossilfuel power (i = 0). Random utilities are givenas

ßx„,- + ay„£«0

(¿=1,2)[1]

' See Ferrini and Scarpa (2007) for more on design strat-egies that might be used in further research along these lines.


where the first expression is the utility for in-dividual n for one of the two wind projects /,and the second is the utility for the status quooption (coal or natural gas), ß and a are pa-rameter vectors to be estimated, x„,- is a vectorof project attributes, y„ is a vector of individ-ual attributes, and £„,- is a random error term.An individual is assumed to choose the alter-native i that gives the highest utility. In aRUM model that choice can be explained onlyup to the probability of altemative i beingchosen. We used mixed logit in estimation toallow for a fairly general pattem of correlationamong the error terms in equation [1]. Weconsidered normal and triangular mixing dis-tributions over most parameters, but ended upemploying normal distributions in all cases.The results were not sensitive to this choice.We used Halton draws in our simulation andfound that the parameters stabilized around500 draws. We allowed for correlation amongeach respondent's error terms (one for eachquestion) by making the Halton draws personspecific instead of choice specific. Separatemodels were estimated for the inland, bay, andocean populations. The theory and approachare well known and fully developed by Train(2003) and Louviere, Hensher, and Swait(2000).

In our application, the wind power projectdisamenity is measured using a set of fourdummy variables representing the differentdistances offered in the choice experiment:3.6 miles, 6 miles, 9 miles, and too far awayto see under any lighting conditions (esti-mated at 20 miles). Because the nearest dis-tance offered (0.9 miles) is the excludedvariable, these coefficients capture a utility in-crease for having the wind power project ateach distance relative to that nearshorelocation.

The extemal cost of wind turbines at anylocation d versus having them out of sight is

[2]

where /3TFTS is the coefficient on distance at"too far to see" and )3¿ is a coefficient on adistance dummy variable {d = 0.9, 3.6, 6, 9)in the Xni vector; ßö.9 = 0 since it is the ex-

cluded variable in estimation. /3fee is the es-timated coefficient on the fee variable, whichis fixed in estimation and is a measure of themarginal utility of income. Because the dis-tance coefficients are estimated as random pa-rameters, the ratios also are random and arethus calculated as simulated means. The su-perscript r on ßxFTS and ß'^ signifies that thecoefficient is one of R draws from the esti-mated distribution of ßjFTS ^nd /3¿. We useÄ = 5,000 in our application.

VI. ESTIMATION RESULTS

Estimation results are shown in Table 5. Asnoted earlier, separate models were estimatedfor the inland, bay, and ocean populations.^The results are more or less as expected. Thecoefficient on fee is negative and statisticallysignificant in all three models, so the likeli-hood of choosing one of the wind altemativesdeclines as the fee increases. There also is aclear preference for the wind power projectbeing located further from the coast. The co-efficients on all of the distance variables arepositive and statistically significant. Gener-ally, the sizes of the coefficients rise at a de-clining rate as you move away from the coastin all three models, although the distance co-efficient in the inland model peaks at 9 milesand then drops somewhat at "too far to see."The relative size of these coefficients is sig-nificantly larger in the bay and ocean models,implying, as expected, greater disutility forthose living near the coast. We discuss im-plicit values from these coefficients in the nextsection.

Model results indicate no clear preferencefor the location of a wind power project at thetwo ocean and one bay locations offered. Thecoefficients on Rehoboth Beach, Fenwick Is-land, and Delaware Bay are not statisticallysignificantly different from one another in anyof the models. In one section of the survey,respondents were asked directly to choosewhether they preferred a certain location foroffshore wind development. Descriptive sta-

* When tested against a pooled tnodel that constrainedparameters to be constant across these three groups, thepooled tnodel was easily rejected in favor of the splitmodels.


TABLE 5Mixed Logit Estimation Results

Variable

Random parameters (means)BayRehoboth BeachFenwick IslandDistance 3.6Distance 6Distance 9Distance too far to seeGreen Energy FundBeach Nourishment FundRoyalty $2 millionRoyalty $8 million

Fixed parametersFeeIncomeSome college educationFour-year college degreePost grad degreeAgeMaleRetiredDelmarvaSee oceanDistance from beachBeach daysSeen a turbineBeach house

Random parameters (dispersion)BayRehoboth BeachFenwick IslandDistance 3.6Distance 6Distance 9Distance Too Far To SeeGreen Energy FundBeach Nourishment FundRoyalty $2 millionRoyalty $8 million

Number of observations (respondents)Log likelihood value

Inland

Parameter /;-Value

6.25**** 0.0006.22***6.37***1.05***1 90***2.39***1.93***0.53**0.78***

-0.54***-0.57**

-0 .11*** '0.0020.59

-0 .60- 1.93****-0.07***'-0 .14

0.52-0.45

3.39***0.01**0.0030.46

-0.47

1.85**0.761.14**1.66**2.19***1.49**0.011.25*1.65***0.781.58***

* 0.000* 0.000" 0.000" 0.000" 0.000" 0.000

0.0240.0010.0100.030

" 0.0000.4920.1930.139

• 0.000" 0.000

0.6400.2460.1720.0070.0460.6500.1300.443

0.0100.3730.0460.0440.0020.0240.9940.0980.0010.2020.004

1,692(564)- 1,401.152

Bay

Parameter

5.606.57*7.17*2.58***3.28**3.60***400***1.350.22

-0.21-0.12

-0.12***-0.01- 1.31-5 .51**-2.79-0.04

0.260.924.96**

-0.24- 0 . 1 1 *

0.0013.05**NA

0.256.89**4.45**0.265.261.265.41**3.442.080.170.78

/7-Value

0.1120.0800.0660.0060.0100.0040.0090.1660.7290.7540.856

0.0050.1100.5030.0210.1800.4230.7990.5200.0190.8840.0690.8330.034NA

0.8720.0110.0150.8400.1010.4610.0430.2260.1760.9210.626

609 (203)-454.8983

Ocean

Parameter

-3 .91**-4.64**-4.78**

0.622.29****2.72****4.14****0.120.760.26

-0 .56

-0.05***0.0020.36

-0 .19-0 .44

0.03-0 .78*

0.72NA

0.562.49***0.0031.48***NA

5.14*3.85***0.04

10.350.330.811.991.130.010.090.05

/;-Value

0.0500.0240.0190.7800.0000.0000.0000.8350.1060.5600.295

0.0050.4950.7650.8580.6780.2580.0890.200

NA0.4590.0010.1500.007

NA

0.0570.00100.9750.2830.8990.5540.1890.3400.9850.9470.953

546(182)-442.3946

Note: NA, not applicable.* Signifieant at the a =0.1 level of eonfidence; ** significant at the a = 0.05 level of confidence; *** significant at the a = 0.01 level of

confidence; *•** significant at the a = 0.001 level of confidence.

tistics show that significantly more respon-dents prefer wind power development in theocean (40%) than in Delaware Bay (16%); al-though a plurality (45%) expresses no pref-erence (Firestone, Kempton, and Krueger2008).

Respondents prefer that royalty funds fromwind power go to targeted funds rather thanto general revenues: the coefficients on the

beach nourishment and green energy fundsare positive in all three models (the generalstate fund being the excluded category). Thecoefficients show significance only in the in-land model. The implied willingness to payby inland residents to have funds distributedto the beach nourishment and green energyfunds versus the state's general fund is about$7 and $5 per month for 3 years. At the same


time, the variables for the amount of the roy-alties give counterintuitive results: negativeand significant coefficients in the inlandmodel and negative but insignificant coeffi-cients in the other models in three of the pos-sible four cases. Because a $1 million royaltyis the excluded category, the $2 million and$8 million royalty variables represent higherpayments and would presumably be utility en-hancing. The insignificant results in the bayand ocean models suggest that the amount ofroyalties collected is not important to indi-viduals. For the inland model, the negativecoefficients suggest a dislike for royalty pay-ments to the state. Perhaps respondents per-ceive that royalty payments might increase theactual cost of delivered energy, or that suchpayments might hamper development of off-shore wind power, or they distrust the statewith such funds.

We estimated all of the parameters on theindividual characteristics as fixed. Initially weexperimented by interacting individual char-acteristics with the distance variables to see ifthere might be preference variation alongthese lines, but detected none. Even as simpleshifters of preferences for wind over fossilfuel power, which the parameters in Table 5show, the demographic variables have onlymodest explanatory power.

Consider the variables that show some sta-tistical significance. The distance one livesfrom the coast is the only variable that is sig-nificant or borderline significant in all threemodels, but the results across models are in-consistent. The probability of choosing off-shore wind over fossil fuel power sourcesincreases with the distance one lives from thecoast for the inland and ocean samples, butdecreases for the bay sample. Distance awayfrom the beach in the ocean sample is a largepredictor of voting in favor of wind power.With each additional quarter mile away fromthe coast that a household is, we find that will-ingness to pay increases by about $12, Al-though significant for the inland sample aswell, that value is only about 90 cents per 10miles. For these models, then, value increasesby either moving the turbines away from peo-ple or moving people away from turbines. Thebay sample willingness to pay decreases with

distance from the coast by about $1 per milewith statistical significance.

We also wanted to test whether preferencesdiffered between households that had anocean or bay view and those that did not. Onemight expect the parameter {See ocean) to benegative and statistically significant, but suchwas not the case in our results. Indeed, for thesmall population of residents living inland buthaving a view (7%) of the ocean or bay, theyshowed a preference in favor of wind.

Having seen a wind turbine at some timein one's life increases the probability of votingfor an offshore wind option in all three modelsand with significance in the bay and oceanmodels, Braunholtz (2003) found that thosewho most frequently saw wind power projectson their day-to-day routine were most favor-able toward them. Our results may imply thatthose who have not seen turbines before mayenvision a far more objectionable visual im-pact than is actually the case. Or, they mayimply that those who favor this type of tech-nology are more likely to have made an effortto view turbines in some setting.

Finally, over the inland sample, whosenumbers dominate Delaware's population, theprobability of favoring wind over fossil fuelsources of power decreases with educationand age. The education result holds weakly inthe bay and ocean models, and the age resultdoes not hold. For the most part, then, ourresults show only modest observed heteroge-neity along the lines noted above and suggestthat people's preferences for offshore windversus fossil fuel sources are not easily clas-sified demographically. Perhaps most inter-esting among the variables with littlesignificance is the coefficient on number ofdays an individual spent at the beach over thelast year. It is positive but insignificant in allmodels, suggesting the impact on in-state vis-itors to the shore may not be significantlylarger than on the general population.

While there is limited observed heteroge-neity, there is a reasonable degree of unob-served heterogeneity realized through thedispersion measures on the random parame-ters. This is most evident in the inland model,as one might expect, since this is the mostdiverse of the sampled populations. Eight ofthe 11 estimated standard errors are signifi-


TABLE 6External Cost per Household for Wind Turbines Located a Different Distances

Offshore

Distance

0.93.669

Extemal

Inland

$18.868.740.780"

Costs AnnuallyPerpetuity

Bay

$34.3911.175.832.06

in

Ocean

$80.0368.7935.1026.65

Extemal Costs Monthly3 Years

Inland

$17.998.340.750"

Bay

$32.7810.645.551.96

for

Ocean

$76.3065.5933.4725.41

" Value is slightly negative in estimation.

cant for this model. There is strong evidenceof variation in the preferences of inland resi-dents for the view disamenity, the allocationof royalty funds, and the size of the royaltyfunds. The sizes of the standard errors, rela-tive to their means, on the royalty funds andthe royalty amounts suggest particularly widevariation in preferences for these attributes,with some residents favoring, and some op-posing, targeted funds. And finally, the size ofthe dispersion terms for the location variablesin the bay and ocean models stand out, show-ing strong variability in the wind power pref-erence and for the location of turbines amongthe coastal samples.

VII. THE EXTERNAL COST OFTURBINES AT DIFFERENT DISTANCES

OFFSHORE

Table 6 shows the estimated extemal costof wind turbines located at different distancesfrom the coastline. The values are shown sep-arately for the inland, bay, and ocean samples.All estimates are per household in 2006 dol-lars and are shown in annual values in per-petuity and monthly values for 3 years. Theestimates were calculated using equation [2].The annual values in perpetuity were con-verted from the 3-year monthly values usinga discount rate of 3%.

The external cost for each of the threepopulation groups decreases with distancefrom the coast. The further a wind project islocated from the shore, the lower the disa-menity effect. The values for the ocean sam-ple are largest, followed by the bay and inlandsamples. The extemal costs per household peryear for the inland population are $19, $9, $1,

and $0 at 0.9, 3.6, 6, and 9 miles. The externalcosts for the bay residents for the same dis-tances are $34, $11, $6, and $2, and the valuesfor the ocean residents are highest at $80, $69,$35, and $27. The extemal costs captured inour estimates may include more than visualdisamenity. In some cases people may be ex-pressing, for example, a concern over the con-fiict of turbines with recreational boaters andfishers in the nearshore area.

As we noted earlier, Ladenburg and Du-bgaard (2007, 2009) provide the only otherpublished estimate of willingness to pay tomove wind turbines further offshore. Eigure 4overlays their estimated values with ours.^Their values are higher and persist at greaterdistances from shore. One explanation mightbe the size of the wind turbines. We use sim-ulations of wind turbines with the nacelle at258 feet, and blades extending to 440 feetabove the ocean; they use larger, next-gener-ation wind turbines, with the nacelle at 328feet, and blades extending to 520 feet.* At the

'' Note: The value for Inland sample at 9 miles actuallygoes slightly negative in estimation. The distance "too farto see" is shown at 20 miles, which seems to be the greatestdistances at which a modem offshore wind turbine could bevisible from shore under even the clearest of conditions. Ourproject has 500 turbines each 440 feet high and is locatedin Delaware. Ladenberg and Dubgaard's project has approx-imately 700 turbines each 520 feet high and is located inDenmark. Ladenburg and Dubgaard's estimates are for arandom sample of all residents (after adjustment for strati-fication), including some with and some without a view ofthe coast. Also, the closest view in our experimental designwas 0.9 miles; Ladenburg and Dubgaard's was 8 km (about5 miles).

* They also use a larger wind project—700 versus 500wind turbines—but each simulated project is very large, solarge in fact that they are each significantly larger than any


/ear

»use

ho

ld/

$/h

c

180 1

160 -

140 -

100 -

40 -

20 -X-._

0

0

\

\

\ X

4 8

Miles

-A--Ocean —x—Bay

—înnnn.....T...-••..

12

Offshore

—•— Inland —

16 20

••—L&D

FIGURE 4Extemal Cost per Household for Wind Turbines Located at Different Distances Offshore with Ladenburg and

Dubgaard (L&D) Results Overlaid, in 2006 Dollars

same time, the differences may be due simplyto differences in the Danish and U.S. popu-lations, differences in the timing of the studiesrelative to development (their study occurredafter the installation of a large offshore windpower project of approximately 70 wind tur-bines), or differences in study design. On thelast point, for example, we use a monthly pay-ment period over 3 years and they use an an-nual payment that continues indefinitely.Also, we include an opt-out altemative thatallows a respondent to choose neither windproject; their design includes no opt-outaltemative.

One caveat worth noting in our findings isthat the maximum monthly fee offered in ourchoice experiment ($30) was well below ourpredicted values in monthly terms for theocean population. While our parameter esti-mates on the distance variable are significantin the ocean model and the stepwise values

project now in existence. We do not feel this difference ex-plains the disparity in results.

have the order one would expect, we are pre-dicting values outside the range of our dataand hence have somewhat lower confidencein these numbers. All of our pretests weredone on inland populations, which led us to arange of prices that worked in the bay andinland samples but were on the low side forthe ocean population. This was simply a mis-take in our pretest design.

Aggregate values for the state of Delawareat 0.9, 3.6, 6, and 9 miles are $7.6 million,$4.2 million, $1.1 million, and $870,000 an-nually in perpetuity. This is the extemal costto all Delaware residents at each distance. Thenumber of households for each area (inland282,691, ocean 22,579, and bay 12,369) wasobtained using 2007 U.S Census Bureaustatistics.

The estimated values, of course, ignore theeffects on visitors to the shore, many of whomare out-of-state visitors from New Jersey,Maryland, Virginia, Pennsylvania, and theWashington, D.C., area (Lilley, Firestone, andKempton 2010). Because there are about 1.3


million out-of-state visitors each year to theDelaware shore (DEDO 2007), this value ispotentially large. Ladenberg and Dubgaard(^2009), for example, find that recreationalboaters and anglers have higher extemal coststhan other populations. But, there is also evi-dence that wind turbines may attract moretourists than they dissuade, so there may bepositive amenity effects to account for as well(Lilley, Firestone, and Kempton 2010).

VIII. CONCLUSIONS

Offshore wind power is a promising alter-native energy source that has gained consid-erable attention recently given the size of theresource, concem over climate change, healthimpacts of air pollutants from conventionalenergy sources, and degree of domestic con-trol over global fossil fuel stocks. Whilenearly free of the textbook extemal effects ofpollution (NAS 2010), offshore wind power,like any other energy facility or transmissionline, has extemal effects in the form of a vi-sual disamenity.

Using stated preference data over house-holds in the state of Delaware, we estimatedthe economic value associated with the visualdisamenity from wind turbines at various dis-tances off the coast. Our results pertain to awind power project with 500 turbines each440 feet high (about 450 MW). The resultsare shown in Figure 4. We estimate that inlandresidents have an extemal cost of $19, $9, $1,and $0 annually in perpetuity (2006 U.S. dol-lars) for wind turbines located at 0.9, 3.6, 6,and 9 miles offshore. Respondents living onthe ocean have extemal costs of $80, $69,$35, and $27 at the same distances. Aggre-gated statewide, the extemal costs at these dis-tances are $7.6, $4.2, $L1, and $0.9 millionper year. Keep in mind that these costs ex-clude the disamenity (and amenity) effects onout-of-state visitors to the Delaware coast andresidents from nearby states such asMaryland.

Looking back at Figure 4, perhaps the moststriking finding of this study is the descent ofdisamenity values after 6 to 9 miles offshore.The conventional wisdom, without much in-formation on extemal costs, has been that tur-bines should be located outside the viewshed.

Given the sizable cost savings associated withmoving turbines closer to shore, our resultsmay call this conventional wisdom into ques-tion. For example, based on rough estimatesfor other projects in the United States, the costsavings of moving the wind project currentlyunder consideration in Delaware from outsidethe viewshed to 9 miles is likely to range be-tween $7 to $20 million per mile. By com-parison the external costs of moving theproject over the same range is about $8 to $10million per mile. So, the numbers are close.The extemal costs exclude tourists and resi-dents in nearby states, so these numbers willbe larger. At the same time, our estimates forthe cost of moving turbines closer are basedon projections for past projects and may de-cline with technological advancement. Takentogether these adjustments would lower trans-mission cost and raise visibility costs, sug-gesting that moving turbines outside theviewshed or perhaps nearer the proposed dis-tance of 13 miles may make sense. But again,these numbers are rough, and getting sharperestimates on both sides would be useful forpolicy.^

Acknowledgments

We acknowledge the contributions of Jesse Fernan-des, Amardeep Dhanju, Meredith Blaydes Lilley,Philip Whitaker, Jacqueline Piero, and April Rouleaufor their assistance in survey development, pretesting,survey assembling and mailing, and/or data entry. Theauthors also thank Patrick Devine-Wright, BonnieRam, and Martin Pasqualetti for reviewing a 2008report on the Delaware survey on which this article isin part based. The research was funded by the Dela-ware Green Energy Fund and the University ofDelaware.

References

Alvarez-Farizo, Begoña, and Nick Hanley. 2002. "Us-ing Conjoint Analysis to Quantify Public Prefer-ences over the Environmental Impacts of Wind

' The cost estimates are derived from Wright et al.(2002), who reported costs for projects in Rhode Island andMaine. Their numbers are per megawatt. Our estimates fordisamenity values convert annual losses to discounted pres-ent values (to be compared with Wright et al.'s numbers)using a 3% discount rate.


Facilities: An Example from Spain." Energy Pol-icy 30 (2): 107-16.

American Wind Energy Association (AWEA). 2009.American Wind Energy Association Annual WindEnergy Report, Year Ending 2008. Available atwww.awea.org/publications/reports/AWEA-Annual-Wind-Report-2009.pdf.

. 2010. Year End 2009 Market Report. Avail-able at www.awea.org/publications/reports/4Q09.pdf.

Aravena, Claudia, Peter Martinsson, and RiccardoScarpa. 2006. "The Effect of a Monetary Attributeon the Marginal Rate of Substitution in a ChoiceExperiment." Paper presented at the Environmen-tal and Resource Economists 3rd World Congress,Kyoto, Japan, July 3-7.

Bergmann, Ariel, Sergio Colombo, and Nick Hanley.2008. "Rural versus Urban Preferences for Renew-able Energy Developments." Ecological Econom-ics 65 (3): 6\6-25.

Bergmann, Ariel, Nick Hanley, and Robert Wright.2006. "Valuing the Attributes of Renewable En-ergy Investments." Energy Policy 34 (9): 1004-14.

Borchers, Allison M., Joshua M. Duke, and GeorgeR. Parsons. 2007. "Does Willingness to Pay forGreen Energy Differ by Source?" Energy Policy35 (6): 3327-34.

Braunholtz, Simon. 2003. Public Attitudes to Windfarms: A Survey of Local Residents in Scotland.Edinburgh: Scottish Executive Social Research.

Delaware Economic Development Office (DEDO).2007. 2006 Delaware Visitor Profile. Available athttp://dedo.delaware.gov/Delaware06.pdf.

Department of Energy (DOE). 2008. 20% Wind En-ergy by 2030: Increasing Wind Energy's Contri-bution to U.S. Electricity Supply. DOE/GO-102008-2567. Available at www Leere.energy.gov/windandhydro/pdfs/41869.pdf.

Dillman, Don A. 2000. Mail and Internet Surveys:The Tailored Design Method. 2nd ed. New York:Wiley.

Dimitropoulos, Alexandros, and Andreas Kontoleon.2009. "Assessing the Determinants of Local Ac-ceptability of Wind-Farm Investments: A ChoiceExperiment in the Greek Aegean Islands." EnergyPolicy 37 (5): 1842-54.

Ek, Kristina. 2006. "Quantifying the EnvironmentalImpacts of Renewable Energy: The Case of Swed-ish Wind Power." In Environmental Valuation inDeveloped Countries: Case Studies, ed. David W.Pearce, 181-210. Northampton, MA: Edward El-gar Publishing.

European Wind Energy Association (EWEA). 2010.The European Offshore Wind Industry: Key Trendsand Statistics 2009. Available at http://ewea.org/

fileadmin/emag/statistics/2009offshore/pdf/offshore%20stats%2020092.pdf.

Fenini, Silvia, and Riccardo Scarpa. 2007. "Designswith A Priori Information for Nonmarket Valu-ation with Choice Experiments: A Monte CarloStudy." Journal of Environmental Economics andManagement 53 (3): 342-63.

Fimereli, Eleni, Susana Mourato, and Peter Pearson.2008. "Measuring Preferences for Low-CarbonEnergy Technologies in South-East England: TheCase of Electricity Generation." Paper presentedat ENVECON 2008: Applied Environmental Eco-nomics Conference, London, March 14.

Firestone, Jeremy, and Willett Kempton. 2007. "Pub-lic Opinion about Large Offshore Wind Power:Underlying Factors." Energy Policy 35 (3): 1584-98.

Firestone, Jeremy, Willett Kempton, and Andrew D.Krueger. 2008. Delaware Opinion on OffshoreWind Power. Final Report. Report submitted to theDelaware Department of Natural Resources(DNREC), Dover, Delaware. Available at www.ceoe.udel.edu/windpower/docs/FinalDNRECOpinionReport.pdf.

. 2009. "Public Acceptance of Offshore WindPower Projects in the United States." Wind Energy12(2): 183-202.

Groothuis, Peter A., Jana D. Groothuis, and John C.Whitebead. 2008. "Green vs. Green: Measuringthe Compensation Required to Site Electrical Gen-eration Windmills in a Viewshed." Energy Policy36 (4): 1545-50.

Hanley, Nick, and Ceara Nevin. 1999. "AppraisingRenewable Energy Developments in RemoteCommunities: The Case of the North Assynt Es-tate, Scotland." Energy Policy 27 (9): 527^7.

Hensher, David A., John M. Rose, and William H.Greene. 2005. Applied Choice Analysis: A PrimerNew York: Cambridge University Press.

Hoen, Ben, Ryan Wiser, Peter Cappers, Mark Tbayer,and Gautam Sethi. 2009. The Impact of WindPower Projects on Residential Property Values inthe United States: A Multi-Site Hedonic Analysis.LBNL-2829E. Berkeley, CA: Lawrence BerkeleyNational Laboratory.

Kempton, Willett, Jeremy Firestone, Jon Lilley, TracyRouleau, and Philip Whitaker. 2005. "The Off-shore Wind Power Debate: Views from CapeCod." Coastal Management 33 (2): 121-51.

Koundouri, Phoebe, Yiannis Kountouris, and KyriakiRemoundou. 2009. "Valuing a Wind Farm Con-struction: A Contingent Valuation Study inGreece." Energy Policy 37 (5): 1939-44.

Krueger, Andrew D. 2008. "Valuing Preferences forOffshore Wind Power: A Choice Experiment Ap-proach." Ph.D. dissertation. University of Dela-ware, Newark.


Ladenburg, Jacob. 2009. "Stated Public Preferencesfor On-Land and Offshore Wind Power Genera-tion: A Review." Wind Energy 12 (2): 171-81.

Ladenburg, Jacob, and Alex Dubgaard. 2007. "Will-ingness to Pay for Reduced Visual Disamenitiesfrom Offshore Wind Earms in Denmark." EnergyPolicy 35 (S): 4059-11.

Ladenburg, Jacob, and Alex Dubgaard. 2009. "Pref-erences of Coastal Zone User Groups Regardingthe Siting of Offshore Wind Earms. Ocean andCoastal Management 52 (5): 233-42.

Lilley, Meredith B., Jeremy Eirestone, and WillettKempton. 2010. "Offshore Wind Energy Devel-opment and Coastal Tourism in Delaware: AnExamination of Potential Impacts and Opportuni-ties." Energ/ei 3 (1): 1-22.

Louviere, Jordan J., David A. Hensher, and Joffre D.Swait. 2000. Stated Choice Methods: Analysis andApplication. Cambridge, UK: Cambridge Univer-sity Press.

McCartney, Abbie. 2006. "The Social Value of Sea-scapes in the Jurien Bay Marine Park: An Assess-

ment of Positive and Negative Preferences forChange." Journal of Agricultural Economics 57(3): 577-94.

Meyerhoff, Jürgen, Cornelia OhI, and VolkmarHartje.2010. "Landscape Externalities from OnshoreWind Power." Energy Policy 38 (1): 82-92.

Navrud, Stale, and Kirsten Gronvik Braten. 2007."Consumers' Preferences for Green and BrownElectricity: A Choice Modeling Approach." RevueD'économie Politique 117 (5): 795-811.

National Academy of Sciences (NAS). 2010. Elec-tricity from Renewable Resources: Status, Pros-pects, and Impediments. Washington, DC:National Academies Press.

Train, Kenneth E. 2003. Discrete Choice ModelsMethods with Simulation. Cambridge, UK: Cam-bridge University Press.

Wright, Sally D., Anthony L. Rogers, James E. Man-well, and Anthony Ellis. 2002. "Transmission Op-tions for Offshore Wind Earms in the UnitedStates." Paper presented at the AEWA AnnualConference, Portland, Oregon, June.

Copyright of Land Economics is the property of University of Wisconsin Press and its content may not be

copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

Documents

Valuing the Visual Disamenity of Offshore Wind … Wind...state of Delaware for offshore wind turbines located at different distances from the coast. The annual costs to inland residents