PipelinesandPropertyValues:AnEclec5cReviewoftheLiterature

LouisWilde,ChristopherLoos,andJackWilliamson1

February15,2012

[ForthcominginJournalofRealEstateLiterature,Vol.20,No.2,Fall2012]

1

1 LouisWilde is a Director at GnarusAdvisors LLCandVisi5ngAssociate in Economicsat theCalifornia Ins5tute ofTechnology.Christopher Loosis aDirector atGnarusAdvisorsLLC.JackWilliamsonisPresidentofAlmostConvexEconomics,Inc. Por5onsofthisreviewareadaptedfromaprojectfundedbytheKernRiverGasTransmissionCompany.Theauthorsthanktheeditorsandrefereesforhelpfulcommentsanddiscussions.

PipelinesandPropertyValues:AnEclec5cReviewoftheLiterature

I. Introduction

Theincidents involvingnaturalgaspipelinesinSeptember2010inSanBruno,CaliforniaandinFebruary

2011inAllentown,Pennsylvania,havebroughtwidespreadmediaaVen5ontotherisksassociatedwith

naturalgaspipelinesystems.2Thispaperconcernswhetherproximitytopipelines,especiallynaturalgas

pipelines,hasanyeffectsonresiden5alpropertyvalues.Assuch,itispartofabroadliteratureonthe

effectsofproximitytodisameni5esonpropertyvalues,par5cularlyhazardousorpoten5allyhazardous

sites.Generallythatliteraturecanbedividedintostudiesdealingwithmereproximitytoa hazardousor

poten5allyhazardoussite3andstudiesdealingwiththeactualpresenceofahazardorpoten5alhazard.4

A comprehensivebibliographyof studiesoftheeffectsonresiden5alpropertyofproximitytoairports,

high‐voltagetransmission lines (“HVTLs”), transporta5on corridors, and landfillsandhazardouswaste

facili5esasof themid‐1990scan befound in Bell (1999).5 That book also includesa comprehensive

bibliographyof studiesof theeffectsofcontamina5ononbothresiden5alandnon‐residen5alproperty

valuesasofthemid‐1990s.6Reviewar5clesincludeFarber(1998),BoyleandKiel(2001),Jackson(2001),

PalmquistandSmith(2002),andKiel(2006).7

This paper falls in the former of the two categories described above, that is, studies dealing with

proximity to ahazardousor poten5ally hazardoussite.Within that category is asub‐class of studies

dealingwithu5litycorridors,par5cularlyHVTLs,which includesa rela5velysmallsetofstudiesdealing

2

See, for example, “Transporta5onChief toUnveil Pipeline SafetyEffort,”The New York Times, April 4, 2011; “U.S.MakesPipeline‐SafetyPush,”TheWallStreetJournal,April4,2011.

3See,e.g.,KielandWilliams(2007).

4See,e.g.,Dotzour(1997)orRogers(2000).

5Bell(1999),pages89‐97.

6Bell(1999),pages142‐151.

7Someofthepapersreferencedinthesear5clescanbefoundinRoddewig(2002).SeealsothebibliographiesinSimons(2005),especiallychapters4and5.

with the effects of proximity to pipelines as well as the effects of pipeline releases or ruptures on

propertyvalues.8

This literaturereviewis eclec5cforatleasttworeasons.Itfocusesprimarilyonpipelinesandonstudies

usingactualsales data.9Asnotedabove,therearealimitednumberofsuchstudies,andmanyarenot

publishedinpeerreviewedjournals.Theinclusionofnon‐peerreviewedstudiesinareviewsuchas this

is not uncommon; infact,many of thestudies, reviewar5cles,andbibliographiescitedin this paper

include references to such studies. Among these are so‐called “working papers” by academicians,

“whitepapers”by researchersatpublicandprivateins5tu5ons, andpaperspresentedatmee5ngsof

professional socie5es or conferences. Some of the studies we discuss fall into these categories. In

addi5on,they also includework producedby researchersac5ng asexpertwitnessesinthecontextof

li5ga5on,whichwealsohereindiscuss.Suchworkisworthyofconsidera5onforseveral reasons.Most

obviously,totheextentthatitaddstothebodyofknowledgeregardingapar5cularissue,itshouldnot

be ignored a priori. Furthermore, while not always peer reviewed in the strict sense of the term,10

studiesrelieduponbyexpertwitnessestypicallyaresubjecttoscru5nybyexpertsretainedbyopposing

par5esandmustpassmusterwithrespecttoadmissibility.11This said,itisimportanttorememberwhen

discussing the results of a study produced by an expert witness in a lawsuit, that there oken exist

an5podalresults producedby anotherexpertwitnessretainedbyanopposing party in thelawsuit.12

3

8 Early workon the effects of high voltage transmission lines on property values incl8udes Colwell (1990) andDelaneyandTimmons(1992).ForlaterstudiesseeDesRosiers(2002)andWolvertonandBoVemiller(2003).

9Weassumethatpropertyvalueequalsmarketprice.Adiscussionofwhen,ifever, thisfailstobetrueisbeyondthe scope of this paper. See, however, sec5on III.D below which provides references related to the ques5onwhetherrealestatemarketsareinforma5onallyefficient.

10 The peer review process variessignificantlyacross publica5ons, but even for themost demanding academicjournals it isfar from perfect.This isnot the place to debate themeaningofpeer review;wesimplynote thatbecauseastudyhasbeen“peer reviewed”doesnot implythatit iserror freeor thatanyremainingerrorsdonothaveasignificant impact upon the conclusionsdrawn fromthestudy.Furthermore,in theauthors’view,whenaresearcher retainedasanexpertwitness for onepartyin alawsuitpublishesastudybasedon thatworkbut theresearcherretainedasanexpertwitnessbytheopposingpartydoesnot,thisshouldnotsubsequentlybeviewedas a reflec5on on the rela5ve validity of the two studies or the conclusions based upon them by the tworesearchers.

11 Recent court rulingson theadmissibilityofexpertwitnesstes5monyemphasizethat itmust relyon thesametechniquesandmeetthesamestandardsasnon‐expertwitnesswork. Foradiscussionoftheissuesandexamplesof excluded tes5monyregarding claims of damages from property value diminu5on due to contamina5on seeWilde(2009).

12Allof thematerialcitedinthispaperhasbeenprovidedto theeditor.It isalsoavailableuponrequestfromtheauthors.

The eclec5c nature of this literature review also is reflected in thediscussion of “related literature”

offeredinsec5on III.D. Thisdiscussion,whilenotintendedtobecomprehensive, coversahandfulof

papersrelatedtovariousthemesintroducedby thestudiesreviewedinsec5onsIII.AthroughIII.C,and

offersseveralavenuesoffutureresearchsuggestedbythem.

Ourreviewofstudiesof theeffectsofpipelinesonpropertyvaluesindicatesthatthereisnosystema5c

evidence,basedonactualsalesdata,thatproximitytopipelinesreduces propertyvalues. Inaddi5on,

thereis nosystema5cevidence, basedonactual salesdata, thatpipelineruptures, evencatastrophic

ones, reducepropertyvalues.However, theextant literatureislimited, so it shouldnotbepresumed

without futureanalysis of actual salesdata, that such incidents as in San Bruno and Allentown, for

example, have no effect on property values. Furthermore, the reviewed research suggests that

hypothe5cal surveys of actual or poten5al market par5cipants may not be a good subs5tute for

systema5canalysisofmarketdata,as theymayoverstatetheeffects,ifany,ofproximitytodisameni5es,

includingpipelines,onpropertyvalues.Whetherthisis,infact,trueremainsanopenques5onpending

furtherresearch.

The remainderof this paper isorganized as follows: Sec5on II overviews the natureof naturalgas

pipeline risks, Sec5on III provides an extensive review of the available literature, published and

unpublished,andSec5onIVoffersabriefsummaryandconclusions.

II. NaturalGasPipelineRisksAsnotedabove,thetopicofpipelinesafetyhassurfaceddrama5callyonthena5onalsceneasaresultof

the incidents in September 2010 in San Bruno, California and in February 2011 in Allentown,

Pennsylvania.Thus,abriefdiscussionofnaturalgaspipelinesafetyisinordereventhoughtheliterature

reviewed below deals with other kinds of pipelines as well. Overall, while the public concern is

understandable,aninformalreviewofthedatatendstosuggestthatsucheventsarerela5velyrare.

Naturalgassupplies25percentof thetotalenergyconsumedintheUS.13Toservicethatdemand,the

countryhasdevelopedsignificantpipelineinfrastructure.Thereareover2.1millionmilesofnaturalgas

pipeline in the US, consis5ng of gathering lines, transmission lines, and distribu5on lines.14 A

4

13SeePHMSAGeneralPipelineFAQs,hVp://www.phmsa.dot.gov.

14SeePHMSAGeneralPipelineFAQs,hVp://www.phmsa.dot.gov.

transmission linewasinvolvedin therecent San Bruno incident,whileasmallerdistribu5onlinewas

involvedintheAllentownincident.

Given their broad geographical reach and the vola5le nature of the product conveyed, there is

understandable interest in the safety of natural gas pipelines. The Department of Transporta5on

(“DOT”)PipelineandHazardousMaterialSafetyAdministra5on(“PHMSA”),ac5ngthroughtheOfficeof

Pipeline Safety (“OPS”), administers the na5onal regulatory program to assure safe natural gas

transporta5onbypipelines.15 TheOPSis responsibleforregula5ng thesafetyof design,construc5on,

opera5on,maintenance,tes5ng,andemergencyresponseofthesesystems. AccordingtothePHMSA,

“pipelines are the safest and most cost‐effec5ve means to transport the extraordinary volumes of

naturalgas”and“[r]ela5vetothevolumesofproductstransported,pipelinesareextremely safewhen

comparedtoothermodesofenergytransporta5on.”16

Safeguards notwithstanding,therearenumerouscontributorstopipelinefailures.ThePHMSAtracksall

incidents by pipeline type and collects extensive informa5on on each incident and the physical

characteris5csof the line involved. Data compiled by the agency generally indicate that corrosion,

material/weld flaws, and excava5on damage are the threemain causes of pipeline failure.17 Other

causes of pipeline incidents include construc5on error, operator error, and malfunc5on of control

systemsor relief equipment.18 Themedia, in commen5ng on the recentSanBrunoandAllentown

accidents, has noted in par5cular that a pipeline’s age, configura5on, and type of material play

importantrolesinthelikelihoodoffailure.

Whileempiricalanalysis of thedataregardingpipelineincidentsmightbeof independentinterest,itis

not the focus of this paper. Nevertheless, OPS sta5s5cs indicate a declining trend in the failure

frequencyfromallcauses.19 Forexample,thenumberof overall leaksduetocorrosion from2002to

2009decreased from1,179to608,with significant incidents– definedasthoseresul5ng in property

5

15 Other agencies like the Department of HomelandSecurity (“DHS”)Transporta5on Security Administra5on (“TSA”), theDepartmentofEnergy(“DOE”),andtheFederalEnergyRegulatoryCommission(“FERC”)alsoregulatethesesystems.

16SeePHMSAGeneralPipelineFAQs,hVp://www.phmsa.dot.gov.

17 See Pipeline Safety brochure prepared by the Interstate Natural Gas Associa5on of America (“INGAA”), hVp://www.ingaa.org/cms/6211/11402.aspx.

18 See, for example, the Arkansas Public Service Commission, Pipeline Safety Office, hVp://www.apscservices.info/PSOIndex.asp.

19SeePipelineSafetybrochurepreparedbytheINGAA,hVp://www.ingaa.org/cms/6211/11402.aspx.

damagegreaterthan$50,000–averagingless than10ayear,whileseriousincidents–thoseresul5ngin

deathandinjuries–averagingaround1perannum.20

III. LiteratureReview

A. Overview

Theempiricalliteratureon theeffects of proximitytooilandnaturalgaspipelinesonproperty values

basedonsales dataissmallintermsof thenumberofpapers(publishedorunpublished)andnarrowin

termsof subjectmaVer. As far aswehavebeen abletodetermine, it consistsof (1) threepublished

papersbyRobertSimons,oneofwhichisco‐authored,dealingwithoilpipelineleaksorruptures;21 (2)

threeinter‐relatedpapersbyWilliamKinnard,withoutandwithSueDickeyandMaryGeckler,onlyone

ofwhichhasbeenpublished,dealingwiththeeffectsofproximitytonaturalgaspipelinesonproperty

values;22 and(3)miscellaneousotherpapersincludingapaperbyBoxall,Chan,andMcMillan(2005)on

the effectsof proximity tosour gaswellsand flaring oilbaVeriesonpropertyvalues.Thenext three

subsec5onsofthispaperdealwitheachofthesesetsofpapers.

B. OilPipelineLeaksandRuptures

Webeginwith thepapersdealing with theeffectsoil pipeline leaksor ruptures onproperty values.

Chronologically,thefirstof theseisSimons(July1999). Thispaperdeals withtheeffectson property

values of an oil pipeline rupture in Fairfax County, Virginia in 1993 that caused the release of

approximately 430,000 gallons of petroleum product. The release caused a surface spill, “with no

evidenceof groundwater contamina5on,”whichflowed down acreek, “eventually ending up in the

PotomacRiverseveralmilesaway.”23

As described in Simons (1999), “Very few proper5es along the pipeline right‐of‐way were directly

contaminatedbytherupture,butanumberofmostlyresiden5aldownstreamproper5esalongthecreek

6

20 Seebrochure5tled“PipelineSafety”preparedbythe INGAA,hVp://www.ingaa.org/cms/6211/11402.aspx. Inthiscontext,“propertydamage”referstoinjurytophysicalproperty, notpropertyvaluediminu5onassociatedwithraw landor improvedproper5es.

21Simons(1999),Simons(Summer1999),andSimons,Winson‐Geideman,andMilkelbank(2001).

22 Kinnard(1993),Kinnard, Dickey,andGeckler, (1994a); andKinnard, Dickey, andGeckler (1994b). These three papers arecloselyrelatedandarecitedinBell(1999)atpage90.

23Simons(1999),page256.

were impactedas thepetroleumpassed by theirhomes.”24 Thus, thestudy dealswith single family

residences and townhouses which were encumbered by pipeline easements but were not actually

contaminatedby the1993release. Themethodologywas a “beforeand aker”analysisusing hedonic

regressiontechniques.25

Forsingle‐family residences,Dr. Simonsbeganfirstwithananalysisof 19subjectproper5es inNorth

FairfaxCountyencumberedbypipelineeasementsthatsoldbetween1990and1997.His controlgroup

for theseproper5es consisted of 768 sales of other homes in the same subdivisions as thesubject

proper5esthatwere notencumberedbypipelineeasements.26 Theresultsof his hedonic regression

modelforthesesalesareshowninTable2onpage260ofSimons(1999).Asshownthere,thecoefficient

onthevariable“saleonpipelinepost‐1993rupture(0‐no,1‐yes)”is‐0.055witha t‐sta5s5cof‐1.68.Dr.

Simonsconcludesthat this “indicatesa5.5%reduc5on inthesalesprice,holdingallothervariablesin

themodelconstant.”27Leavingasidethefactthattherelevantcoefficientisnotsta5s5callysignificantat

the95%level,acri5calproblemwiththisconclusionisthatthehedonicregressionmodelunderlying it

omitsa keyvariable.28 In par5cular,Dr.Simons’modelincludesvariablesforwhetherasalewasfora

property encumbered by apipeline easement and forwhether such a sale occurred aker the1993

rupture, but it does not include a variable for whether a sale, encumbered or not by a pipeline

easement,occurredakerthe1993rupture.Asaresult,thecoefficientonthevariable“saleonpipeline

post‐1993rupture(0‐no,1‐yes)”comparessuchsalestoall sales notencumberedbyapipeline,notjust

thosethatoccurredakerthe1993rupture.Inotherwords,atbestonecansaythatsalesofproper5es

encumbered by a pipeline easement aker 1993 were lower, other things equal, than sales of all

proper5es not encumbered by a pipeline easement, both before and aker the 1993 rupture. The

relevantques5on,however, iswhether salesof proper5esencumbered by apipelineeasementaker

7

24Simons(1999),page256.

25 Hedonicregressionanalysisisusedtoexplain the value ofagoodbydecomposingitsmarketprice intoindividualeffectsassociatedwiththecharacteris5csofthegood. Inthecontextofrealestate, forexample, thesecharacteris5cstypicallyincludesuchitems asthenumberofbedrooms,thenumberofbathrooms,presenceofaswimmingpool,and,proximitytoameni5es ordisamen5tes.See,e.g.,thevariousbooks,reviews,andar5clescitedabove.

26 Simons(1999),page258.Noinforma5onisgivenonhowmanyofthesubjecthomessoldbefore therupture (March1993)versusaker.

27Simons(1999),page260.

28 The typicalcriterionforsta5s5cal significance isbasedonwhetherthe 95% confidence intervalfor thees5matedcoefficientincludeszero.Atechnicaldiscussionofthisstandardisbeyondthescopeofthisreview.

1993werelower,otherthingsequal,thansalesofproper5esnotencumberedbyapipelineeasement

aker1993.Itisimpossibletoanswertheques5onusingDr.Simons’model.29

Besides analyzingsingle‐familyresidencesinNorthFairfaxCounty,Simons(1999)alsoconductsseparate

hedonicregressionanalysesoftheeffectsoftheruptureonsingle‐familyresidencesandtownhousesin

all of Fairfax County. With respect to single‐family residences, he reports no sta5s5cally significant

results.However,theactualhedonicregressionmodelisnotshown,makingitimpossibletoascertainif

the omiVed variableproblemwhich greatly limits the generalizability of his results for single‐family

residencesinNorthFairfaxCountyalsoappliestohisresultsforsingle‐familyresidencesinall ofFairfax

County.Ofcourse,thisleavesopenthepossibilitythatthemodeldoesomitthekeyvariableandthatifit

wereincluded,sta5s5callysignificantresultswouldbeobtained. Ontheotherhand,theresultsofthe

hedonicregressionmodelfor townhousesinallofFairfax CountyareshowninTable4onpage262of

Simons(1999),andinthis casethereis noomiVedvariableproblem.30Theresultsshownthereindicate

adecreaseinthevalueof townhousesencumberedby pipelineeasementsaker the1993ruptureas

compared tobeforeof 2.6%,controlling for changes in thevalueof townhousesnotencumbered by

pipelineeasements.Whilethiseffectisrela5velysmall,itappearstobebasedonsoundmethodology.

Chronologically,thesecondSimonspaper isSimons(Summer1999).Thispaperanalyzesthetermsand

results of a property buy‐out and seVlement agreement resul5ng from a lawsuit alleging long‐term

leakagefromapipelinecarryingpetroleumproductsinnortheastOhio.Becausethetransac5onsdonot

representarms‐lengthtransac5onsbetweentypicallymo5vatedbuyersandsellers,thepaperprovides

onlyverylimitedinsightintotheactualimpactsoftheleakage,ifany,onthemarketvalueofproper5es

adjacenttothepipeline.Assuch,weoffernofurtherdiscussionofitinthispaper.31

The finalpaper dealingwith theeffectsof oilpipeline ruptureson property valuesisSimons, et al.

(2001).ThispaperdealswiththeeffectsofanoilpipelineruptureinPrinceGeorgeCounty,Maryland,in

April 2000. The released oil ini5ally flowed in a creek and surrounding marsh area. It subsequently

8

29ThisisanexampleofanomiVedvariablemisspecifica5onandisdiscussedat lengthinRogers (2000).The issueis notsimplyone ofinterpreta5on.Aproper“beforeandaker”hedonicregressionanalysisoftheeffectsofa contamina5onevent requirespropermodelspecifica5onasthereinoutlined.

30 The addi5onal variable that eliminates the problem is labeled “TH sale post‐1993 rupture (0‐no, 1‐yes)” in Table 4. Asdiscussedabove,noanalogousvariableispresentinTable2.Bothspecifica5onscannotbecorrect.

31 Simonsiden5fies21proper5esthatwere bought andthenresoldby the responsibleparty.Of the21,11wereresoldatorverycloseto20%lessthanthepurchaseprice.Noevidenceisofferedthateithertheini5alpurchasepricesorthere‐salepricesreflectmarketvalue.

spreadtothePatuxentRiverwhereithada poten5alimpactonupto10milesofshoreline.32Simons,et

al.(2001) reportstheresults ofahedonicregressionmodelanalyzingtheeffects ofthereleaseonboth

impactedwaterfrontproper5esandthose“interior”proper5eswith“ownershiprightstoanimpacted

beach area.”Therewerethreeof the formerand 32of the laVer.33 The reportedresultsindicateno

sta5s5cally significant effectsonwaterfrontproper5esbutdid showasta5s5cally significant average

decreaseinthevalueof“interior”proper5esofapproximately11%.34

Evenwithintheconfinesof Simons,etal.(2001),itisclearfromthespecifica5onshowninTable2on

page415 that thekeyhedonic regressionmodel in thepaper omitskey variables.35 Butthere is an

addi5onal, significant problem with the reported analysis which goes beyond the paper itself. In

par5cular,althoughtheauthors neveracknowledgeit inthepaper, theanalysisreportedwaspartofa

consul5ngassignmentassociatedwithaclassac5onlawsuit.Thiscaseul5matelywas cap5onedAnthony

Williams,etal.v.PotomacElectricPowerCompany,etal. (“PEPCo”).36 Inthecourseof thatli5ga5on,

certain problemswith theanalysisdescribed in Simonsetal. (2001) wereiden5fied,and inresponse

addi5onal analyses were conducted by Dr. Simons that were not subsequently published. As next

described,Dr.Simons’revisedresultsfailtosupport,andul5matelycontradict,hisini5alconclusions.37

Specifically,betweenthe5meDr.Simons conductedtheanalysisreported inSimons,etal. (2001) and

the5mehetes5fiedinPEPCo,heupdatedhisanalysis withaddi5onalsales data.38 Intheprocessof

doing so, among other things, Dr. Simons also eliminated certain non‐arms length sales that were

9

32Simons,etal,(2001),page410.

33Simons,etal.(2001),page413.

34SeeSimons,etal.(2001).Table2anddiscussionpage415.

35 A number of neededarea dummy variables are omiVedas per Rogers (2000). The authorsanalyze two subject areas,waterfront proper5eswith beachesthatwere impactedand interior proper5eswith ownership rights inbeaches that wereimpacted.Simple dummyvariablesforeachofthese typesofproper5esaswellasdummyvariableswhichare interactedwiththesewithapost‐spill5medummyvariableareneeded.

36 AnthonyWilliams, et al. v. PotomacElectricPowerCompany, et al., Inthe United StatesDistrictCourt for the District ofMarylandSouthernDivision,CivilNo.PJM‐00‐1429.

37IssuesrelatedtotheconclusionsreachedinSimons,etal.(2001)werefirstreportedinWilson(2004).

38“[W]edidtworoundsofresearch.Onewasfromthefirstyear’ssalesseason,whichincludedsalesfromMay15,2000upun5lOctober4,2000.We foundthere there [sic]were 32oiledinteriorproper5eswith access rights inwaterfrontproperty andthreewaterfrontsalesoverthat5meperiod.Thenwerecentlyredidouranalysisandnowwehaveatotalof65sales,ofwhichsevenwaterfrontsaleshave occurredupfromthree andatotalnowofabout58interiorproper5eswithaccessinwaterfrontproperty.”Mo5onsHearing before the Honorable Peter J.MessiVe, UnitedSatedDistrict Judge, Anthony Williams, et al. v.PotomacElectricPowerCompany,etal., Inthe UnitedStatesDistrictCourt for theDistrictofMarylandSouthernDivision,CivilNo.PJM‐00‐1429.October9,2001,page169.[“PEPCoHearing”].

previouslyincludedinhisanalysis.39Asaresultofthesechanges,atleastonespecifica5onofDr.Simons’

hedonic regression model showed a sta5s5cally significant decrease in the value of waterfront

proper5esof12%–13%butnosta5s5callysignificanteffectsoninteriorproper5es,essen5allyflipping

the publishedresults.40 Thedetailsof this revisedmodelarenotpublishedandarenot availablefor

review.

C. ProximitytoNaturalGasPipelines

In thefinal analysis,giventhe limita5onsondataand thetechnicalproblems, liVlecanbedefinitely

concluded from the published papersdealing with the impacts of oil pipeline ruptureson property

values. The situa5on issomewhat beVer, but not ideal, regarding the threeKinnard papersdealing

directly with theeffects of proximityto natural gaspipelineson property values:Kinnard (1993) and

Kinnard,etal.(1994a,1994b).Kinnard(1994b)is merelyapublishedsummaryoftheresultsobtainedin

Kinnard (1993), so our discussion isrestricted to Kinnard (1993) andKinnardet al. (1994a). Kinnard

(1993) reportstheresultof two“marketresearchassignments”conductedbyDr.Kinnard’sfirm, Real

EstateConsul5ng Group of Connec5cut, Inc. (“RECGC”), to “test claims that proposedhigh‐pressure

naturalgastransmissionpipelineswould resultinsubstan5aldecreasesin themarketvalueof single‐

family residen5al proper5es ‘near’ the proposed pipelines and their rights of way.” One of these

assignmentsinvolveda“1440‐poundspersquareinch(psi)naturalgastransmissionpipelineproposedto

transverse” Southwestern Connec5cut. The other involved a “1000/1200 psi transmission pipeline

circlingtheNorthernandWesternboundariesofLasVegas,Nevada.”41

In both studies, proper5es were categorized according to their distance from the pipeline under

considera5on.TheseDistanceZoneswereasfollows:

T:Lotistraversedbypipelinerightofway

O:Lotabutspipelinerightofway

A:Lotis200feetorlessfromrightofway

10

39 PEPCoHearing, pages 174‐177, 184. Non‐arms length sales typically are excluded, to the extent possible, from hedonicregressionmodelsofpropertyvaluessincetheyruntheriskofnotreflec5ngmarketvalue.

40 PEPCoHearing,pages180,214. These resultsmakemuchmoreeconomicsensethantheauthors’ ini5alresults– the effectonwaterfrontproper5eswithimpactedbeachesshouldbe greaterthantheeffectoninteriorproper5eswithaccessrightstoimpactedbeaches.Butsincetherewereonlythreeoftheformer,eventheseresultsshouldbeviewedwithcau5on.

41Kinnard(1993),page3.

B:Lotis201‐400feetfromrightofway

C:Lotis401‐800feetfromrightofway

D:Lotis801‐1300feetfromrightofway

E:Lotis1301‐2600feetfromrightofway

F:Lotis2601‐5280feetfromrightofway

For theConnec5cutstudy, DistanceZoneEwasused asa control area,andfor theLasVegasstudy,

DistanceZoneFwasusedasacontrolarea.

IntheConnec5cutstudy,twonaturalgastransmissionlinerightof wayswereanalyzed, jointlyand in

separatemodels.One,theAlgonquinrightofway,containstwonaturalgas pipelinesandtraversesfour

townsfromwhichsalesdatawasusedinthestudy.Theother,theTennesseerightof way,containsa

singlenaturalgaspipelineandtraverses fivetownsfromwhichsalesdatawasusedinthestudy.Allthree

naturalgas pipelineswerebuiltinthe1960s.IntheLasVegasstudy,onenaturalgaspipelinerightofway,

whichwasnewasof the5meofthestudy(late1991),wasanalyzed.Salesdata fortheLas Vegasstudy

werefromamasterplannedcommunitycalledTheLakes.

Thecompletehedonicregressionmodelses5matedarenotgiveninKinnard(1993).Rather,theresults

shownfocusonthedummyvariablesrelatedtotheDistanceZones.Assummarizedinthepaper,twoof

theprimaryconclusionsreachedare:

1. No paVern of measurable and significant nega5ve impacts on sales prices of residen5al

proper5esclosetoanexis5ngorproposedhigh‐pressurenaturalgastransmissionpipelinewas

observed.Noneshouldbeexpectedinfuture,similarsitua5ons.

2. Nosystema5cpaVernofvaria5onsinsalespriceeffectswasobserved.Noneshouldbeexpected

in future, similar situa5ons. No correla5on was observed between price levels (total or per

squarefoot)anddistancefromthepipeline.42

Thesecondof theKinnardpapershereindiscussed,Kinnard,etal. (1994a),waspresentedatthe1994

AnnualConferenceof theAmericanRealEstateSociety,andwas insomewaysaheadof its 5me.The

focus of the paper is on the divergence between es5mates of themarket impacts of proximity to

11

42Kinnard(1993),page16.

allegedlyhazardouslandusesbasedonsurveys,especiallyso‐calledCon5ngentValua5onMethodology

surveys,ascomparedtoactualmarketdata,inpar5cularsalesprices.

Con5ngent Valua5onMethodology (“CVM”) surveys are dis5nct from informal canvasses of market

par5cipants okenrelieduponbyappraisers.Ini5ally,CVM surveysweredesignedtobeformalsurveys

thataVempt toelicit valua5onsfornon‐marketgoodsorservicesin responsetoahypothe5cal injury

scenario. Theyhavebeenadopted foruseinproperty valuediminu5oncasesfromNaturalResource

DamageAssessment(“NRDA”),buteveninthecontextofNRDAtheyarecontroversial.43

ThecoreofKinnard,etal.(1994a)consistsof twonewcasestudies,buttheauthors men5ontwoprior

papersinasec5onen5tled“EarlyProximityImpactStudyResults:MarketFactv.ExpertOpinion.”

ThefirstofthesewasanunpublishedreportpreparedbyRECGCdealingwiththeeffectsofproximityto

HVTLsonpropertyvalues.44Thekeyresultsofthatreportaredescribedasfollows:

Inthecourseofconduc5nginterviewswithlocalrealestatebrokers,realestateappraisers and

mortgagelendersac5veinthelocalresiden5alrealestatemarket,wediscoveredtheintriguing

factthatthepriceorvalueimpactopinionsof locallyac5veresiden5alrealestateprofessionals

weresubstan5allymorenega5vethanthoseof residents livingonproper5esthatabuVedthe

HVTL. In turn, the awtudes and opinions about value impacts expressed by the abuwng

propertyresidentweremarkedlymorenega5vethanthefacts ofthemarketplaceindicated.This

par5cular study was based on an analysis of over 700 sales of single‐family residences in

proximitytotheHVTLbeingstudied.BothPairedSalesandMul5pleRegressionAnalyses(MRA)

intheHedonicPricingModelformatwereappliedtothisdataset.45

12

43 Foradiscussionofissues relatedtotheuse of theCVMinthe contextofNRDA, see,forexample,DiamondandHausman(1994).Foradefense oftheuse ofsuchsurveysinpropertyvaluediminu5oncases,seeMundyandMcClean(1998)andforacri5que ofsame, seeWilson(2006).Fora recentcomprehensive reviewofthe literatureontheuse ofthe CVMinrealestatedamagees5ma5on,seeLipscomb,etal. (2011).Wethanka refereeforpoin5ngustoastudypreparedbyIndustrialEconomics,Incorporatedfor theOPSthatincludesdiscussions oftheuse ofthe CVMinthe costbenefitanalysis ofOPS’sregula5onsandprograms. See The Joint OPS Stakeholder Workgroup (1999) at hVp://www.phmsa.dot.gov/sta5cfiles/PHMSA/DownloadableFiles/Files/cba_rpt.pdf.

44Kinnard,Geckler,Geckler,andMitchell(1984).

45Kinnard,etal.(1984).

ThesecondpriorpapersummarizedbyDr.Kinnard andhisco‐authors,DelaneyandTimmons(1992),

alsosurveys realestateprofessionalsregardingtheimpactofproximitytoHVTLsonpropertyvalues.46

Theyconcludethat “themean declinein value notedby respondentswho had appraised residen5al

propertysubjecttotheinfluenceof[HVTLs],anddependingongeographicregion,rangedfrom7.77%to

15.5%, …”47 However, they also observe that “[t]his finding is in contrast to much of the research

conducted to date thatfindsliVleorno impact fromhighvoltagepower linesonresiden5alproperty

values.”48 Furthermore,as notedbyDr.Kinnard andhisco‐authors, “[a]n interes5ng footnoteto that

study is that thoserespondentswho had no experiencein appraising HVTL‐impactedproper5eshad

much stronger nega5ve feelings about the price impacts of such proximity than did designated

residen5alappraisersexperiencedinvaluingsuchproper5es.49

ThetwonewcasestudiesanalyzedinKinnard,etal.(1994a) stemfromcondemna5onac5onsinvolving

atransporta5oncorridor fornuclearwasteandfuelandahigh‐pressurenaturalgastransmissionline.

Thefirstof thesewasapar5altaking relatedtoaby‐passhighway intended,inpart, forshipmentsof

nuclearfuelandnuclearwastetoandfromLos AlamosinNewMexico.Atelephonesurveywasusedby

experts retained on behalf of the plain5ffs to es5mateseverance damagesfor certain undeveloped

residen5alpropertyadjacenttotheactualtaking.50Dr.Kinnardandhisco‐authorsconductedapost‐trial

face‐to‐facesurveyandananalysisofmarketdatausingvarioustechniquesincludinghedonicregression

analysis.Like thehedonicregressionanalysisinKinnard (1993),described above,aseries ofDistance

Zonesweredefined,rangingfromlessthanone‐quartermiletomorethantwomilesfromthehighway

rightof way.ThelaVerwas usedas thecontrolareaforpurposesof evalua5ng theimpacts,ifany,of

proximitytothehighwayrightofway.

AssummarizedinKinnard,etal.(1994a),“[t]hees5matednega5veimpactin[thesurveyconductedby

theplain5ffs] ranged fromslightly lessthan10%up to40%.Within thisrangeof impact, 71%of the

respondents…forecastanega5veimpact.”Theiranalysisofmarketdata, “ontheotherhand,showed

13

46DelaneyandTimmons(1992),pages315‐329.

47DelaneyandTimmons(1992),page324.

48DelaneyandTimmons(1992),page323.

49Kinnard,etal.(1994a),page2.

50”WhentheUnitedStatesacquiresonlypartofaunitaryholding,federallawrequiresthatcompensa5onbemadenotonlyforthe property interest acquired, but also for the diminu5on, if any, in the value of the remainder directly caused by theacquisi5onand/orbytheuse towhichthepartacquiredwillbe put.This diminu5oninthevalueoftheremainderisokenand‘somewhat loosely’ referred toasseverance damage.”The Appraisal Ins5tute,2000, Uniform Appraisal StandardsforFederalLandAcquisi5ons,page47,emphasisintheoriginal,ci5ngUnitedStatesv.Miller,317U.S.369,376(1943).

liVleeffectfromproximity to thehighwayrightof way,whether the focusof theanalysiswassingle‐

familyresiden5alproper5esorresiden5allots.”51

ThesecondcasestudyanalyzedinKinnard,etal.(1994a),thehigh‐pressurenaturalgas transmissionline

study,isen5relydis5nctfromtheNewMexicotransporta5oncorridorstudy,butis basedonthesame

condemna5onac5onanalyzedinKinnard(1993).As withtheNewMexicotransporta5oncorridorstudy,

the focusof thehigh‐pressurenaturalgastransmission line study inKinnard, et al. (1994a) ison the

comparison between survey results, which are driven by responses to hypothe5cal scenarios, and

hedonicregressionresults,whicharedrivenbyactualsalesdata.52

AccordingtotheresultssummarizedinKinnard,etal.(1994a),57%of thesurveyrespondentssaidthey

wouldnotbuypropertywithin240feetofthesubjectnaturalgastransmissionpipelinerightofway.This

percentagefellmonotonicallywithdistanceto33%forpropertywithin1240feetof thesubjectnatural

gas transmission pipeline right of way. Themedian discount expected by thosewilling to buy was

10.5%.53 On theotherhand, thehedonicregressionmodels showednosystema5cnega5veeffecton

propertyvaluesofproximitytothepipeline.54

ThekeyoverallfindingsofKinnard,etal.(1994a)areasfollows:

1. Exanteopinionsofinterviewees notnecessarilyinvolvedinbuying(orconsideringbuying)in

areasclaimedtobeaffectedbyproximitytoasourceof fearorhazardtohumanhealthand

safety are not a subs5tute of proxy for market sales transac5ons in iden5fying and

measuringtheimpactofsalespricesofresiden5alproper5es.Theopinionsaremuchmore

nega5vethanthereflec5onsofactualmarketbehaviorofindividuals[sic]buyersexpost.

14

51Kinnard,etal.(1994a),page10.

52 The authorsknow ofnootherpublishedorunpublishedstudiesusingCVMtypesurveystoes5matethe effectsofpipelinesonpropertyvalues.Forstudiesrelatedtoleakingundergroundstoragetanks(“LUSTS”),seeSimons(2005),chapter5.

53Kinnard,etal.(1994a),Table2,page15.

54 AsinKinnard(1993),theactualhedonicmodels arenotreportedinKinnard,etal.(1994a).The laVercasestudymodelsaredis5nctfromtheformerinthattheyinvolvetwopoten5allyaffectedcommuni5esandabroader5meperiod.

2. Actualmarket behavior isanunequivocal fact, especially when it is possibleto construct

priceindexesofstandardizedhousesorlotsforcompara5vepurposes.55

3. There isnoconsistent, systema5cmarketpaVernof lowersalespricespersquare footof

livingareaasproper5esbecomeclosertothesourceofthefearofhazardstohumanhealth

andsafety.56

D. RelatedResearch

Anumberofpapersrelatedtotheissuesraisedabovehaveappeared,bothpublishedandunpublished.

Someofthesearediscussedinthissubsec5on.

Boxall,atal.(2005)reportsuponatechnicallysophis5catedanalysisoftheeffects ofoilandnaturalgas

facili5es, including wells and pipelines, on rural residen5al property values in Alberta, Canada. A

par5cular focusof thestudy is on thedifferen5al impacts of “sweet”wells versus “sour” wells, the

dis5nc5onbeingwellsthatcontaingaswithoutorwiththepresenceofhydrogensulfide,whichimposes

apoten5alhealthrisk.Forpurposesofthispaper,thekeyresultisthatthe“presenceofwells,especially

sour gaswells, was found to depressproperty values but thenumber of pipelines carrying sour gas

variabledidnothaveasignificantcoefficient.”57

In another paper dealing with theeffectsof proximity to oil produc5on facili5es, Flower and Ragas

(1994)usehedonicregressiontechniquestoanalyzetheeffectsof tworefineries inSt.BernardParish,

Louisiana,on nearby residen5alproperty values. Theyfind examplesof no effects,short‐termeffects

(based,e.g.,oncatastrophicevents),andpermanenteffectsassociatedwithproximitytothepetroleum

refineries.58

15

55Arefereehaspointedoutthatthisfindingisnotuniformlyacceptedbyallrealestateprofessionals.Theissueiswhetherrealestatemarketsgenerallyaresemi‐strongformefficient.Seefootnote9aboveandtextsupra,aswellasthediscussionandreferencesbelowinthissec5on.

56Kinnard,etal.(1994a),page12.

57Boxall(2005),page266.

58Whileonemightpresume thatrefineriesareassociatedwithpipelines,thereisnoexplicitmen5onofpipelinesinFlowerandRagas(1994).

Next,Simons(1999)referencesanunpublishedstudyof theimpacts apetroleumreleasethatoccurred

duringa floodeventinHarrisCounty,Texas.AccordingtoDr.Simons,theauthorofthestudy(Dr.Barton

Smith)“concludedthattheneteffectofthereleasewasassociatedwitha10%reduc5oninvalueforthe

affectedresiden5alsubdivision.”NofurtherdetailsareprovidedinSimons(1999),includingwhetherDr.

Smith’sstudywaspreparedinthecontextof li5ga5onresul5ngfromthereleaseand,ifso,whetherany

otherstudieswereperformedinthecontextofthatli5ga5on.59

Regarding the issue of hypothe5cal surveys versus the analysis of actual sales data, Frankel (1991)

reports on the resultsof a survey of realtorsand appraisers in 35communi5es surrounding O’Hare

AirportinChicago,Illinois.Whilenotrelatedtopipelines perse,theauthorfindsa significantdisparity

betweenes5matesof reduc5ons in property values dueto noisefrom O’Hare Airport based on the

surveysversus es5matesof reduc5onsinproperty valuesdue to noisefromotherairportsbasedon

hedonicregressionmodels,theformerbeingsubstan5allyhigherthanthelaVer.

Finally,Skantz andStrickland(1987)findthattherewasnomarketreac5onintermsofpropertyvalues

to a major flood event, allegedly due to the availability of low‐cost flood insurance, but that a

subsequent increase in insurance rates was followed by an immediate and significant decrease in

property values. These results, according to Skantz andStrickland, areconsistent witha semi‐strong

formefficientrealestatemarket;thatis,thatpricesintherealestatemarketreflectallpubliclyavailable

informa5on.60 This raises the obvious ques5on whether similar results hold for catastrophic events

relatedtopipelines,suchastheincidentsinSanBrunoandAllentown;thatis,whetherthemarkethas

alreadydiscountedforthelikelihoodofacatastrophicevent.61

IV. Summary&ConclusionsAt least one conclusion and one research hypothesis follow from a review of the exis5ng, limited

literatureontheeffectsofproximitytopipelinesonpropertyvalues.Theconclusionis thatthereisno

credibleevidencebased onactualsalesdatathatproximity to pipelinesreducespropertyvalues.The

researchhypothesisis thathypothe5calsurveysofactualorpoten5almarketpar5cipants shouldnotbe

16

59Simons(1999),footnote4andtextsupra.

60Forclassicdiscussionsofmarketefficiency,seeFama(1970,1991).

61Thepointiswhetherthemarkethadalreadydiscountedforthepossibilityofaneventthatwouldhaveanega5ve impactonproper5es. For a comprehensive review of the literature on the efficiencyofrealestate markets asof the early 1990s, seeGatzlaffandTir5roglu(1995).

usedasasubs5tuteforthesystema5canalysisofmarketdataas theymayoverstatetheeffects,ifany,

ofproximitytodisameni5es,includingpipelines,onpropertyvalues.

Infact,bothof theseobserva5onssuggestareas forfurtherresearch.Oneproblemwithmuchof the

exis5ng literature on the effects of disameni5es such as poten5al sources of contamina5on or

catastrophicincidents onpropertyvaluesisafailuretocontrolforothernega5vefactors.Thisissuewas

putsuccinctlyasfollowsbytwoleadingresearchersinthefield:

Whereenvironmental degrada5oniscausedbyemissionsfrompolluters,thespa5aldistribu5on

ofenvironmentalqualitymaybehighlycorrelatedwiththespa5aldistribu5onofemiVers.Ifthe

emiVers themselves are undesirable, then it will be difficult to separate the effects of

environmentalqualityvaria5onfromtheeffectsofproximitytotheseemiVers.62

Studies of the effects of proximity to pipelines should be rela5vely immune to such methodological

issues since most of the geographical loca5on of pipelines in not in proximity to other poten5al

disameni5es.Aswell,in situa5onssuchastheSanBrunoincident,issuesofwhetherand/orwhenthe

relevantinforma5onaboutthepoten5alrisktoownersofpropertyinproximitytothepipelinesbecame

publiclargelyaremootsincetheincidentsaresignificantandwell‐publicized.Whatareneedednoware

before‐and‐aker studies of the effects of these incidents on directly affected proper5es as well as

indirectlyaffectedproper5es.

Withrespecttovalua5onsbasedonhypothe5calsurveysversus actualmarketdata,theissueisnotnew,

nor is it confined to the effects of pipelines on property values.What is needed at this point is a

comprehensivereviewofexis5ng resultsaswell asnewresearchcomparingpastpredic5onsofeffects

onvaluebasedonsurveystoactualmarketoutcomesasreflectedinsalesprices.63

17

62LeggeVandBockstael(1998),pages6‐7.

63See,e.g.,Roddewig(2006).

References

Bell,Randall,RealEstateDamages:AnAnalysisofDetrimentalCondi5ons,TheAppraisalIns5tute,1999.

Boxall,Peter C.,WingH.Chan,andMelville L. McMillan,“The Impact of Oil andNatural Gas Facili5es on Rural

Residen5alPropertyValues:ASpa5alHedonicAnalysis,”ResourceandEnergyEconomics,2005,pages248‐269.

Boyle, Melissa and Katherine Kiel, “A Survey of House Price Hedonic Studies of the Impact of Environmental

Externali5es.”JournalofRealEstateLiterature,2001,vol.9,No.2,pages116‐144.

Colwell,Peter,"PowerLinesandLandValue,"JournalofRealEstateResearch,(Spring1990),pages117‐127.

Delaney,C.and D.Timmons,"High Voltage Power Lines:DoTheyAffectResiden5alPropertyValue?," Journal of

RealEstateResearch(Summer1992),pages315‐330.

Des Rosiers, F., “Power Lines, Visual Encumbrance, and House Values: A Microspa5al Approach to Impact

Measurement,”TheJournalofRealEstateResearch,2002,pages23‐?.

Diamond,P.A.,andJerryHausman,“Con5ngent Valua5on:IsSomeNumberBeVer ThanNoNumber?”Journal of

EconomicPerspec5ves,1994,vol.8,No.4,pages45‐64.

Dotzour,Mark,“GroundwaterContamina5on and Residen5al PropertyValues,”TheAppraisal Journal, July1997,

pp.279‐85.

Fama,Eugene,F.,“EfficientCapitalMarkets:AReviewofTheoryandEmpiricalWork,”TheJournalofFinance,May

1970,pages383‐417.

Fama,EugeneF.,“EfficientCapitalMarkets:II,”TheJournalofFinance,December,1991,pages1575‐1617.

18

Farber, Stephen, "Undesirable Facili5es and Property Values: A Summary of Empirical Studies," Ecological

Economics,24(1998),pages1‐14.

Flower,Patrick,andWadeRagas,“TheEffectofRefineriesonNeighborhoodPropertyValues,”TheJournal ofReal

EstateResearch,Summer1994,pages319‐338.

Frankel,Marvin,“AircrakNoiseandResiden5alPropertyValues:ResultsofaSurveyStudy,”TheAppraisalJournal,

January,1991,pages96‐110.

Gatzlaff, Dean and Doğan Tirtiroğlu, “Real Estate Market Efficiency: Issues and Evidence,” JournalofRealEstate

Literature,1995,pages157‐189.

Jackson,ThomasO., “TheEffectsof EnvironmentalContamina5on onRealEstate:ALiteratureReview,”Journal of

RealEstateLiterature,9:2(2001),pp.93‐116.

Kiel, Katherine, “Environmental Contamina5on and House Values,” in J.I. Carruthers and B. Mundy (eds),

EnvironmentalValua5on:InterregionalandIntraregionalPerspec5ves,Ashgate:Aldershot,2006,pp.121‐146.

Kiel,Katherine,andMichaelWilliams,“TheImpact of SuperfundSiteson Local PropertyValues:AreAllSites the

Same?”JournalofUrbanEconomics,2007,pages170‐192.

Kinnard, William N., Jr., “The Impact of Proximity to High‐Pressure Natural Gas Pipelines on Single‐Family

Residen5alPropertyValues,”RealEstateConsul5ngGroupofConnec5cut,Inc.,April1993(presentedat the1993

AnnualMee5ngoftheAmericanRealEstateSociety,KeyWest,Fla.,April1993).

Kinnard,WilliamN.,Jr.,SueAnnDickey, andMaryBethGeckler,“Fear(AsaMeasureofDamages)StrikesOut:Two

Case Studies, Comparisons of Actual Market Behavior with Opinion SurveyResearch,” (presented at the 1993

AnnualMee5ngoftheAmericanRealEstateSociety,SantaBarbara,Ca.,April1994a).

19

Kinnard,William N., Jr., Sue Ann Dickey, andMaryBeth Geckler, “Natural Gas Pipeline Impact on Residen5al

PropertyValues:AnEmpiricalStudyofTwoMarketAreas,”RightofWay,June/July1994b,pages26‐29.

Kinnard,WilliamN.,Jr.,MaryBethGeckler,JohnK. Geckler,andPhillip S.Mitchell,“AnAnalysisof the Impact of

High VoltageOverheadElectric Transmission Lines on Residen5al PropertyValuesin OrangeCounty,NewYork,”

Storrs,CT:RealEstateCounselingGroupofConnec5cut,Inc.,May1984.

LeggeV,ChristopherandNancyE.Bockstael,“EvidenceoftheEffectsofWaterQualityonResiden5alLandPrices,”

PresentedattheAnnualMee5ngoftheAmericanAgriculturalEconomicsAssocia5on,August4,1998.

Lipscomb,CliffordA.,MaxKummerow,WillSpiess,Sarah Kilpatrick,and JohnA.Kilpatrick.“Con5ngent Valua5on

andRealEstateDamageEs5ma5on,”JournalofRealEstateLiterature,19:2(2011),pp.283‐305.

Mundy,William,andDaveMcLean,“TheAddi5onofCon5ngent Valua5onand ConjointAnalysisto theRequired

BodyofKnowledgefortheEs5ma5onofEnvironmentalDamagetoRealEstate,”TheJournalofRealEstatePrac5ce

andEduca5on,1998.

Palmquist, Raymond B., and V. Kerry Smith, "The Use of Hedonic Property Value Techniques for Policy and

Li5ga5on," in Tom Tietenberg and Henk Folmer, eds., Interna5onal Yearbook of Environmental and Resource

Economics,EdwardElgar,2002,pp.115‐164.

Roddewig, Richard (editor), Valuing Contaminated Property: An Appraisal Ins5tute Anthology. Chicago: The

AppraisalIns5tute(2002).

Roddewig,Richard,andJamesFrey,“Tes5ngtheReliabilityofCon5ngentValua5onintheRealEstateMarketplace,”

TheAppraisalJournal,Summer2006,pages267‐280.

Rogers,Warren,“Errors inHedonicModelingRegressions:CompoundIndicatorVariablesandOmiVedVariables,”

TheAppraisalJournal,April2000,pp.208‐13.

20

Simons,RobertA.,“TheEffectofPipelineRupturesonNoncontaminatedResiden5alEasement‐HoldingPropertyin

FairfaxCounty,”TheAppraisalJournal,July1999,pages255‐263.

Simons,RobertA.,“SeVlementofanOilPipelineLeakwithContaminatedResiden5alProperty:ACaseStudy,”Real

EstateIssues,Summer1999,pages46‐52.

Simons,RobertA.,WhenBadThingsHappentoGoodProperty,WashingtonDC:ELIPress(2005).

Simons,Robert A.,KimberlyWinson‐Geideman,andBrianMilkelbank,“The EffectsofanOil PipelineRuptureon

Single‐FamilyHousePrices,”TheAppraisalJournal,October2001,pages410‐418.

Skantz,Terrance,and ThomasStrickland, “House Pricesand aFloodEvent:An Empirical Inves5ga5on ofMarket

Efficiency,”TheJournalofRealEstateResearch,Winter1987,pages75‐83.

TheJointOPSStakeholderWorkgroup(1999),”ACollabora5veFrameworkforOfficeofPipelineSafetyCost‐Benefit

Analyses,”September2,1999.

Wilde, Louis, “Keeping the Gate: Damages Tes5mony in Cases Alleging Property Value Diminution Due to

Contamina5on,”BNA,Inc.ExpertEvidenceReport,volume9,No.5(2009),pages129‐143.

Wilson,Albert, “TheQues5onableReliabilityof‘Peer Reviewed’RealEstateLiterature,”Analysis andPerspec5ve,

BureauofNa5onalAffairs,Inc.,Vol.4,No.1,2004,pages16‐24.

Wilson,Albert,“Con5ngentValua5on:NotanAppropriatevalua5onTool,”TheAppraisalJournal,Winter,2006.

Wolverton,Marvin,andStevenBoVemiller, “Further AnalysisofTransmissionLineImpact onResiden5al Property

Values,"TheAppraisalJournal,July2003,pages244‐252.

21