AttorneyandJudgeExperienceinTortsLitigation:AnEmpiricalStudy†

Yun‐chienChang*

Kong‐PinChen**

Chang‐ChingLin***

†AdraftofthispaperhasbeenpresentedattheTheodoreEisenbergEmpiricalLegalStudiesConferenceheldatTelAvivUniversityFacultyofLawonJune2–3,2015;HamburgLecturesonLawandEconomicsonMay27,2015;BACTseminaratRotterdamInstituteofLawandEconomicsonJune5,2015;StaffSeminaratSingaporeManagementUniversityFacultyofLawonJuly29,2015;the2015EastAsianLawandSocietyAnnualMeetingheldatWasedaUniversity,TokyoonAugust3–6,2015;UniversityofPennsylvaniaLawSchoolAdHocFacultyWorkshoponOct.27,2015;LawandEconomicsColloquiumatNorthwesternLawSchoolonOct.29,2015;2015ConferenceonEmpiricalLegalStudiesheldatWashingtonUniversityatSaintLouisonOct.31,2015;RCHSSEmpiricalLegalStudiesWorkshoponNov.6,2015;QuantitativeEmpiricalLegalStudiesConferenceheldatInternationalSchoolofFinanceLaw,EastChinaUniversityofPoliticalScienceandLawinShanghai,ChinaonNov.30,2015;FacultyWorkshopatUniversityofIowaCollegeofLawonJan.25,2016;LawandEconomicsColloquiumatNYUSchoolofLawonFeb.9,2016;Work‐in‐ProgressWorkshopattheUniversityofChicagoLawSchoolonFeb.25,2016.WethankDavidAbrams,GailAgrawal,JanetAinsworth,ShilpiBhattacharya,BernieBlack,ArthurBonnfield,GaryChan,DawnChutkow,JacquesdeLisle,ThomasEger,MichaelFaure,TaliaFisher,ChrisReindersFolmer,EzraFriedman,HualingFu,ThomasGallanis,JeanGalbraith,JonahGelbach,JergGutmann,PaulHeaton,MichaelHeise,Han‐weiHo,DavidHoffman,WilliamHubbard,DavidHyman,NicholasJohnson,MichaelKnoll,Hon.JamesLeach,PeyWoanLee,KKLim,KateLitvak,JingLiu,KeeYangLow,PeterMascini,SandraGMayson,AlanMiller,KonstantinosPilpilides,VictorD.Quintanilla,JasonRantanen,JohnReitz,IssiRosen‐Zvi,MaxSchanzenbach,MargoSchlanger,JeanetteShao,MatthewSpitzer,AgnesStrauß,HangWuTang,LeaVanderVelde,StefanVoigt,FranziskaWeber,KerenWeinshall‐Margel,YixinXu,EyalZamir,andWeiZhangforhelpfulcomments.Withduerespects,wethanktheseveraljudgesandattorneyswhoattendthetworoundtablediscussionsonjudgeandattorneyexperienceforkindlysharingtheirinsightswithus.WealsothankAliceKuo,Jian‐HuaLai,Chieh‐YuLiu,HilaryTsai,Yu‐JuneTseng,andTien‐hsinWangforresearchassistance.Da‐weiDavidJuang(theformerCEOofpingluweb.com)provideduswiththejudgeandattorneyexperiencedata,whichareinvaluableforthisresearch.*AssociateResearchProfessor&DirectorofCenterforEmpiricalLegalStudies,InstitutumIurisprudentiae,AcademiaSinica,Taiwan.kleiber@sinica.edu.tw. **DistinguishedResearchFellowandDirector,ResearchCenterforHumanitiesandSocialSciences,AcademiaSinica;ExecutiveDirector,CenterofInstitutionandBehaviorStudies,AcademiaSinica;Professor,DepartmentofEconomics,NationalTaiwanUniversity.kongpin@gate.sinica.edu.tw***AssociateProfessor,DepartmentofEconomics,NationalChengKungUniversity.ever@mail.ncku.edu.tw

Chang, Chen & Lin

ii

Abstract

Theeffectofattorneyandjudgeexperienceinlitigationhasnotbeenfully

explored,asdetailedmeasuresofjuristexperiencearenotavailable.Usinga

uniquedatasetfromTaiwan,wemeasureattorney(judge)experiencebythe

totalnumberofcivilcasestheyhaverepresented(rendered)from2000to2014.

Combiningthesedatawithanotheruniquedatasetwecompiledfrompainand

sufferingdamageslawsuitsregardingpersonalinjuryinastructuralequation

model,wefindthatbothdistrictcourtjudgesandplaintiffattorneysfollowthe

recentchangesindamagesassessmentpracticeinhighcourts.Courtawards

weresubjecttotheanchoringeffectcastbyplaintiffs’claims.Inaddition,the

moreexperiencedtheplaintiffattorneyswere,themoretheplaintiffs’claims

deviatedfromthehistoricaltrendofcourt‐adjudicatedpainandsuffering

damages.Thus,throughtheindirecteffectofhigherclaims,experiencedplaintiff

attorneysearnedtheirclientshigheramountsofpainandsufferingdamages.

Whetherdefendantshireattorneysandwhetherdefendantattorneysandjudges

areexperienceddonotaffectwhethercourt‐adjudicatedpainandsuffering

damagesdeviatedfromthehistorictrend.

Keywords

Painandsufferingdamages,compensationpercentage,filingfee,deviationfrom

historictrend,structuralequationmodel(SEM)

Chang, Chen & Lin

iii

TableofContents

I.  Introduction ............................................................................................................ 1 

II.  Pain and Suffering Damages Law and Legal Practice in Taiwan .......................... 3 

III.  Research Questions and Prior Literature ............................................................... 7 

A.  Judges ............................................................................................................. 7 

B.  Attorneys ........................................................................................................ 8 

1.  Assignment of Cases to Attorneys ......................................................... 9 

2.  Plaintiff Attorneys ................................................................................ 14 

3.  Defendant Attorneys ............................................................................ 17 

IV.  Models ................................................................................................................. 18 

A.  OLS Models to Predict District Court and High Court Awards ................... 20 

B.  Structural Equation Model on Deviation of Court Award from Trend ........ 21 

V.  Data ...................................................................................................................... 26 

VI.  Findings and Implications .................................................................................... 34 

A.  Judges Pay Close Attention to High Court Decisions .................................. 34 

B.  Attorney Influence on Plaintiff Claim ......................................................... 37 

C.  Plaintiff Attorney Influence on Adjudication ............................................... 38 

D.  Judge Experience Does Not Affect Deviation Pattern Error! Bookmark not

defined. 

E.  Defendant Attorney Has No Role ................................................................ 39 

VII. Conclusion ........................................................................................................... 40 

Appendix A: Predicting Court-Adjudicated Pain and Suffering Damages .................. 42 

Appendix B: Attorney Hiring Decisions ...................................................................... 44 

Chang, Chen & Lin

1

I. INTRODUCTION

Whenyougetinvolvedinalegaldisputeandthecaseisgoingtocourt,

wouldyouratherhireanexperiencedattorneythananinexperiencedone,

assumingyoucanaffordtopaythefeechargedbytheformer?Ifyes,why?Senior

lawyerschargemore,butwhatistheaddedvalueoftheirexperience?Inthelong

debateontheissueof“haveversushave‐not”inlitigation,thepremiseisthatthe

haves(i.e.therich)arebetterabletomobilizelitigation‐usefulresources,

includinghiringmoreexpensiveattorneys(Kuo‐ChangHuang,Lin,andChen

2014;Boyd2015b:296;Chen,Huang,andLin2015).1 Fromintuitionsof

ordinarypeopletothelegalliterature,theoftenimplicitassumptionisthat

experiencedattorneysarebetter.Nonetheless,veryfewempiricalworkshave

beenabletodemonstratethedifferencesbetweenexperiencedand

inexperiencedattorneys.

Astudyonattorneyexperiencewouldbeincompleteiftheroleofjudges(in

ajurisdictionwithoutjuries)isignored.Theeffectofattorneyexperiencecould

varyaccordingtothesittingjudges’experience.Theeffectofjudgeexperience

alsohasindependentinterests.Judgesfollowprecedentsthatsettlequestionsof

law,butwhethertheyfollowprecedentsthatdealwithquestionsoffacts,suchas

howtoassessdamages,hasnotbeenrigorouslyexamined.Towhatextentjudges’

experienceaffectstheirdecisionstofollowhighercourtsregardingassessmentof

damageshasalsonotbeenstudied,either.Poweredbyauniquedatasetthat

containsdetailedmeasuresofmultiplefacetsoftheexperienceofallattorneys

andjudgesinTaiwanasof2014,thisarticlesetsouttofillthegapinthe

empiricalliterature.

Thelegalissueusedtoexaminetheimportofjuristexperienceisthe

assessmentofpainandsufferingdamagesforpersonalinjuriesincurredincar

accidents.Theeffectofjuristexperiencecanbebetterisolatedwhenthejudicial

decisionsareneitherformulaicnorrule‐based.Painandsufferingdamagesare

1 Thisarticledoesnotdirectlyfitinthe“haveversushavenot”literatureitself.Mostnatural‐personplaintiffsanddefendantsincaraccidentcasesare“have‐not”—ourdataonparties’incomecansupportthisclaim.Asaresultofthelackofvariance,partycapabilityandcourts’ideologicalpreferencewouldbeaminorconcern,ifatall,inthisstudy,sothattheeffectofjuristexperiencecanbeidentifiedbetter.

Chang, Chen & Lin

2

highlydiscretionaryinTaiwan,2 andthusjudgesmaybeaffectedbyextra‐factual

factors,includingattorney“manipulation.”Albeitdiscretionary,theamountof

painandsufferingdamagesis,toacertainextent,predictable.Twoprior

empiricalstudies(Changetal.2014;Changetal.2015)havedemonstratedthat

painandsufferingdamagescanbecapturedbysimpleregressionmodels3—this

isimportant,asattorneyexperiencewouldnotmatteriftheamountof

non‐pecuniarydamagesisarbitrarilydecided.

Assessmentsofpainandsufferingdamagesprovideastreamlinedsettingto

examinejudicialbehaviors.Avastempiricalliteratureisdevotedtoshowthat

judgesareideological(forarecentreview,seeEpstein,Landes,andPosner2013:

77–85).Ontheotherhand,asPosner(2008)pointsout,easycasescanbe

determinedbystatutorytexts.Assessingtheamountofpainandsuffering

damagesisneitheralegalistnoranideologicalexercise.Followingthepragmatic

labor‐modelofjudges(Posner2008;Epstein,Landes,andPosner2013),this

articleempiricallyexamineswhetherthecareerjudgesinTaiwandemonstrated

reversalaversionand“audition(desireforpromotion),”withouttheconfounding

influenceofideologyandlegalisticpressure,andwhetherthejudges’experience

onthebenchaffectstheirbehaviors.

Thisstreamlinedsettingisalsosuitablefortestingtheeffectofattorney

experience,asattorneys’ideologycanbeignored.Granted,lawyerskillsarenot

singular.Threadingthestatutorytextsandjudicialcasestogethertomakean

innovativeandconvincingargumentisahallmarkofgoodlawyers.Nevertheless,

thisskillislargelyuselessintheassessmentofpainandsufferingdamages.Thus,

ourfindingregardingtheeffectofattorneyexperienceisnotreadily

generalizabletootherlegalcontextswherestatutoryinterpretationiscritical.Yet,

intermsofidentifyingtheeffectofattorneyexperience,perhapsitisbetterthat

thelegalissuewestudyenablesustoignoreotheraspectsofattorneyskillsand

focusontheaddedvalueofexperiencewhenattorneysmakesnumericclaims

basedonfacts.

2 Nostatuteorsupremecourtprecedenthasprovidedclearguidanceonhowtoassessthistypeofnon‐pecuniarydamages.3 TheR‐squaresofthemultipleregressionmodelsusedtopredictthesetwotypesofpainandsufferingdamagesrenderedbycourtsinTaiwanbetween2008and2012arebetween0.50and0.85.

Chang, Chen & Lin

3

Usingstructuralequationmodelsthatcancontrolforendogeneityproblem,

wefindthatdistrictcourtjudgesdetectthechangingtrendinhighcourtsand

dutifullyfollowthelatestwayhighcourtsassesspainandsufferingdamages.

ConsistentwithChang,Chen,andLin(2016),weagainfindthatjudgeswere

influencedbytheamountofplaintiffs’claims.Judicialexperiencedoesnotaffect

whetherjudgesdeviatefromthehistoricaltrendornot.Judgesaresuspiciousof

seniorattorneys.Seniorplaintiffattorneystendtoover‐claim,andthroughthe

anchoringeffect,winmorepainandsufferingdamagesfortheirclients.Whether

thedefendanthiredanattorneyandhowexperienceddefendantattorneysaredo

notmatter.

Thestructureofthisarticleisasfollows:PartIIprovidesanoverviewof

Taiwanlaw.PartIIIexplainstheresearchquestionsandsummarizestheprior

literature.PartIVlaysoutourOLSandSEMmodels.PartVdescribesthe

pertinentdata.PartVIdiscussestheimplicationofourfindings.PartVII

concludes.

II. PAINANDSUFFERINGDAMAGESLAWANDLEGALPRACTICEINTAIWAN4

Pursuant to Articles 193 and 195 of the Taiwan Civil Code, victims of a

tortiousact canrequest the tortfeasor topaypecuniarydamagesandpainand

suffering damages.5 No formula exists for courts to determine the amount of

painandsufferingdamages.Thecivilcodeprovidesnoguidance.Afewleading

casesrenderedbytheSupremeCourtofTaiwaninthe1950sand1960sdeclared

thatthefollowingfactorsshouldbeconsidered:thesocio‐economicstatus,total

asset, annual income, age, educational background, etc. of both sides, the

plaintiff ’s level of pain and harm, the victim’s negligence, the defendant’s

repentance,andsoon.Otherthanthis,todate,noconventionalwisdomorrules

of thumbexists forquantifyingpainandsuffering. Inpractice,plaintiffs simply

4 PartofthissectionisadaptedfromChangetal.(2014).5 TaiwanCivilCodeart.193Ipromulgates:“Ifapersonhaswrongfullydamagedtothebodyorhealthofanother,andcausedtheinjuredpersontoloseordecreasehislaboringcapacity,ortoincreasetheneedinliving,thetortfeasorsshallbeboundtomakecompensationtotheinjuredpersonforanyinjuryarisingtherefrom.”TaiwanCivilCodeart.195Iprescribes:“Ifapersonhaswrongfullydamagedtothebody,health,reputation,liberty,credit,privacyorchastityofanother,ortoanother'spersonalityinasevereway,theinjuredpersonmayclaimareasonablecompensationinmoneyevenifsuchinjuryisnotapurelypecuniaryloss.”

Chang, Chen & Lin

4

claimanamountandcontendthatitisjust,withlittlesupportingevidence.Court

decisionstypicallystartwithatemplatediscussionthatcarbon‐copiesthelistof

factorsemphasizedbytheleadingcases,6 thensummarizethefactsofthecaseat

hand,andconcludebyawardinganamount.Asjudgeshaveneverelaboratedon

theirformulasandrarelyprovidedconcreteinformationregardingthefactors,it

isdoubtful towhatextent those factors listed in the templateargumentsaffect

thefinalamountofpainandsufferingdamages.7

Courts inTaiwanwill review the receipts of all pecuniary expenses and

only grant plaintiffswith reasonable expenses. Due to themandatory national

health care system that covers most medical treatments andmedication, only

medical expenses that are not covered by the health care plans (such as

co‐payment, certain special medicines and operations, and domestic nursery

cares)canberecoveredbythevictimfromthetortfeasor.

Plaintiffsdonothaveanincentivetoclaimunrealisticallyhighamountsof

painandsufferingdamages.First, filing feesareproportional to theamountof

claimed total damages—roughly, around 1%of the total claimed damages; see

(Kuo‐ChangHuang2008:208fn.17).8 Second,the losingpartyhastopay filing

feesandothercourt fees. Ina tort lawsuit,aplaintiffusuallyhastopaypartof

thecourtfees(includingthefilingfee)ifthecourtdoesnotgrantallherclaims.

Theplaintiffgenerallyhastopay[1–(courtaward/plaintiff ’sclaim)]×courtfee.

So claiming a high amount of pain and suffering damages increases both the

amountoffilingfeesandtheprobabilityofbearingahigherpercentageofcourt

fees.One important caveat:whenaplaintiffmakespainandsufferingdamages

claimsaspartofthecriminalproceedingsagainstadefendant,andthedefendant

wasfoundguilty,theplaintiffdoesnothavetopayfilingfeesforhercivillawsuit

in the courtof first instance. Suchplaintiffs still have to shareother court fees

6 Notallcourtsusethesametemplate.Thefactorsthatacourtexplicitlyclaimstotakeintoaccountslightlydiffer. 7 Inunreportedtables,weexploredthefactorsthatTaiwanesecourtspurporttohaveconsideredindeterminingpainandsufferingdamages.Thetables,however,suggestnoclearpattern.8 PursuanttoArticle77‐13ofCivilProcedureCodeofTaiwan,thefilingfeeisassessedinthefollowingway:“1,000NTDonthefirstNTD100,000ofthepriceorclaim'svalue,andanadditionalamountshallbetaxedforeachNTD10,000thereafterinaccordancewiththefollowingrates:NTD110ontheportionbetweenNTD100,001andNTD1,000,000inclusive;NTD99ontheportionbetweenNTD1,000,001andNTD10,000,000inclusive;NTD88ontheportionbetweenNTD10,000,001andNTD100,000,000inclusive;NTD77ontheportionbetweenNTD100,000,001andNTD1,000,000,000inclusive;andNTD66ontheportionoverNTD1,000,000,000.AfractionofNTD10,000shallberoundeduptoNTD10,000forpurposesoftaxingcourtcosts.”

Chang, Chen & Lin

5

(suchasperdiemgiventotestifyingwitnesses).

Taiwancangenerallybeconsideredacivil‐lawcountry.Almostalljudgesare

career judgeswhomay ormay not have (most have not) practiced lawbefore

servingonthebench.Atoplawgraduatecanbecomeajudgeat25yearsoldor

so(theaverageinrecentyearsis28yearsold).MostjuristsinTaiwanmajorin

lawasanundergraduate, andonlyaminorityof juristsare trained ina JD‐like

graduate program. Jurists who pass the bar exam and finish six months of

practical training are qualified to practice law. Jurists who pursue a career as

judges or prosecutors have to take the “court officer” examination. Thosewho

pass the examination receive training in theAcademy for the Judiciary for two

years. At the end of their training, based on their grades, preferences, and

openings,theywillbecomejudgesorprosecutors. Judgesaretenured,andthus

presumably less influenced by external political influences. For civil matters,

there are three levels of courts: district courts, high courts, and the supreme

court.Theformertwocandeterminebothquestionsoffactandquestionsoflaw,

whilethesupremecourtonlydealswithquestionsoflaw.Appealingtothecourt

of second instance (for non‐small claim cases, the high courts) is as of right,

whereas large‐stake cases represented by attorneys can be appealed to the

supremecourt,subjecttoitsdiscretion(EisenbergandHuang2012;Chen,Huang,

andLin2015).

Asforattorneyfeescollectedbyplaintiffs’anddefendants’lawyers,

contingentfeesandhourlyfeesinthistypeoftortlitigationareveryrare,though

notprohibited.9 Flatfeesaremainstream.Thus,attorneysdonotnecessarily

haveincentivestoselectcasesthataremorelikelytowin,astheycouldreceive

similaramountsoffees.Also,accordingtotheattorneysweinterviewed,small

9 ZamirandRitov(2010)provideaconvincingbehavioralaccountofwhyplaintiffsintortlitigationoptforthecontingentfeestructurewhiledefendantsdonot.AccordingtotheTaiwaneseattorneysweinterviewed,however,fixedfeeshavebeendominantinTaiwan.Thecontingentfeearrangementwasusedinthefewcaseswhereclientsrequestit,anditwasoftenusedwhentheprobabilityofwinningisnothigh.Thecontingentfeepercentagesare20%–30%.Also,theattorneyfeearrangementisseldompurelycontingent‐based;usuallyattorneyswillstillchargeasmallamountofflatfee;smallfirms,ratherthanbigfirms,aremorewillingtoconsenttoacontingentfeearrangement.Moreimportantly,andstrikinglyoppositetothepracticeintheU.S.market,ourinterviewedattorneysbelievethatitisunethicaltotakecontingentfeesincaraccidentcases,asattorneyswouldbetakingawaymoneytocompensateforthepainandinjuryoftheirclients.Iftortvictimscannotaffordtheusualattorneyfees,attorneysmaychargealowerfeeratherthanenteringintoacontingentfeearrangement.Legalaidisanotherwayforpoorplaintiffstoseeklegalrepresentationwithoutresortingtocontingentfeearrangements.

Chang, Chen & Lin

6

lawfirmsandsolopractitioners,whohandletortcases,donothavestableand

wideclientele;thus,theyaregenerallyunwillingtoturndownclientswhoare

willingtopayaflatfee.10 Attorneysoftenadjustthefeesexanteaccordingtothe

complexityofthecases.Seniorattorneysgenerallychargeahigherfee.Case

selectionmaybeintheformofpoorpartiesinsearchofattorneyswhochargea

lowerflatfee.Charginglowfeesisnotagoodindicatoroftheeffectivenessofthe

attorneys.Someattorneysmaybecheapbecausetheyareincapable,whileother

attorneysmaychargealowerfeebecausetheirmarginalcase‐handlingcostis

lower.

Thefollowingfactsarealsoworthnoting.Insurancecompanieshavea

minorroleintortslitigation.Insurancecompaniesarerarely,ifever,apartyin

suchlitigation(noobservationinourdatasetfeaturesaninsurancecompanyas

aparty).11 Evenwhentortfeasorshaveboughtliabilityinsurancepolicies,in

additiontothemandatorymotorvehicleinsurance,tortvictimssuethe

tortfeasors,notthelatter’sinsurancecompany.Also,thereisnodistinction

betweentheplaintiffbarandthedefendantbar(Kuo‐ChangHuang2008:227).

Finally,AbramsandYoon(2007)haveraisedtheissueofthedifficultyto

evaluatethecontributionofanindividualattorney.Thisisnotaprobleminour

study.Whilethepriorliteraturemostlyexaminestheeffectoflawfirms,this

researchfocusesonthatofindividualattorneys.BiglawfirmsinTaiwanmostly

practiceincorporatelawandIPlawandusuallychargebythehour.Seldomdo

theytakecaraccidentcases,mostlybecauseplaintiffswhocannotaffordthefee

chargedbybiglawfirmswillnotseektheirrepresentationinthefirstplace.After

interviewingwithseveralexperiencedattorneys,weidentified12biglawfirms

inTaiwanandcodedtheattorneyswhoworkthereasofSeptember2014.Only5

observationsinourdatabasecontainattorneysaffiliatedwiththesebiglawfirms.

Outsidebiglawfirms,mostattorneyspracticesolo—sometimeshiringafew

associatesandsharingofficespaceandsecretarieswithotherattorneys.Inother

words,litigationoutcomescanbeattributedtoindividualeffortsbyattorneys,

10 Twoveryseniorattorneysweinterviewed(oneofthemrecentlybecameajusticeinTaiwan’sConstitutionalCourt)toldusthattheywouldrefusetorepresentmeritlesscases.Theyadmitthattheyenjoythisprivilegebecausetheyarenotshortofclients.11 Healthinsurancecompaniescansubrogatevictimstosuethetortfeasors,butitisveryrare.Insurancecompaniesmaybesuediftheyrefusetoindemnifyaninsuredperson.Thesearenotthetortslitigationwestudyhere,asjudgesinthosecasesdonothavetoassesspainandsufferingdamages.

Chang, Chen & Lin

7

notlawfirms.

III. RESEARCHQUESTIONSANDPRIORLITERATURE

A. Judges

Ourexaminationofjudicialbehaviorsfocusesonwhetherjudges’experience

affectsthewaytheyfollowprecedents.Aspointedoutabove,inassessingpain

andsufferingdamages,neithertextnorideologyisthejudges’guidingposts.The

roleofprecedentspresumablyloomslarge.Itisnotbreakingnewsthatlower

courtjudgeslargelysticktolegaldoctrinesexpoundedbyhighercourts.The

precedentsunderdiscussionhere,however,refertothepatternofhowhigher

courtsordistrictcourtsinpriorcaseshaveassesseddamagesaccordingtocase

facts.Morespecifically,lowercourtjudgesmayallrefertothesameprovisionin

thecivilcodeandallcitethesameTaiwanSupremeCourtcasesthatlaidoutthe

keyfactorstobeconsidered.Nonetheless,judgesmaygivedifferentweightsto

thesefactors.Followingprecedentsinourempiricalprojectmeansthatjudges

givesimilarweightstofactorssuchasinjurylevel,medicalexpenses,andthe

lengthofdeclaredincarceration.

Wehypothesizethatdistrictcourtjudgeswillfollowtheassessmentpattern

ofappellatecourts.Ingeneral,judgesareaversetotheirdecisions’being

reversed.Overrulingmaytarnishthejudges’reputationandaffecttheirchanceof

beingpromotedtohighercourts.12 InTaiwan,accordingtoourinterviewwith

judges,theaforementionedconcernsdoexistanddistrictcourtjudgesthusstrive

tofollowtheimmediatelysupervisingcourts.13 Wearenotthefirstto

empiricallytestwhetherlowercourtsfollowhighercourtprecedents.14

Nonetheless,tothebestofourknowledge,wearethefirsttoempiricallyexamine

whetherlowercourtsmimicthewayhighercourtsevaluatefactstoassess

12 OurhypothesisisconsistentwithChoi,Gulati,andPosner(2012:518)’sempiricalworkthatfindsthatjudgesinFederalDistrictCourtsintheU.S.“adjusttheiropinion‐writingpracticestominimizetheirworkloadwhilemaximizingtheirreputationandchanceforelevationtoahighercourt.”13 TaiwanHighCourtevenliterallygradedtheDistrictCourtdecisionsappealedtotheHighCourt.Ineachjudge’spersonalwebaccount,shecanseethegradesofallhercasesthatwereappealedaswellastheaveragegradeofherfellowjudgesinthesamedistrictcourt.ThisrulewasabolishedaslateasJuly2015. 14 Forinstance,Boyd(2015a)’srecentempiricalstudyidentifiesthecriticalfactorsinenhancingthehierarchicalinfluenceoffederalcourtsofappealsonfederaldistrictcourtsintheU.S.

Chang, Chen & Lin

8

damages.

Whetherjudges’experienceaffectsthepatternoffollowingprecedentshas

yettobestudiedextensively.Chang,Chen,andLin(2016)reportthat

experiencedjudgesinreal‐worldcaseswerenotsubjecttotheanchoringeffect,

whileinexperiencedjudgeswerestronglyinfluencedbyanchors.Nonetheless,by

counter‐claiming,defendantscanweaken,evenfullyerase,theanchoringeffect

createdbyplaintiffs’claims.Judges’experience,therefore,isusefulindebiasing.

Asjudgeshaveincentives(reversalaversionanddesireforpromotion)tofollow

highercourtprecedentsthroughouttheirtenure,weexpectthatexperiencedoes

notmatter.Still,arigorousempiricaltestingiswarranted.

B. Attorneys

Theeffectofattorneyexperiencehasnotbeenthoroughlystudied,perhaps

forlackofgooddataandtheproblemofselectioneffect.Severalexistentstudies

havefoundthatseniorattorneysperformbetterinlitigationthanunseasoned

colleagues.AbramsandYoon(2007)studyalmost12,000felonycasesinwhich

publicdefenderswererandomlyassignedtoclients,andfindthatexperienced

attorneys,measuredbytheirtenureinthepublicdefenderoffice,achieve

substantiallymorefavorableoutcomesfortheirclientsthanlessexperienced

attorneys.Harris,Peeples,andMetzloff(2005:235–36,241)measureexperience

ofmedicalmalpracticeattorneysbytheyearofpracticeandthenumberofcases

anattorneyhashandledwithinthe348sampledcases,andfindthatattorneys

whohadhandledmoremedicalmalpracticecases,hadmoretrialexperience,and

wenttobetterlawschoolsperformedbetterthanattorneyswithoutthese

attributes.Harris,Peeples,andMetzloff(2008:267,280),using“yearssince

admittedtopractice”and“numberofmed‐malcaseshandledduringthestudy

period”asameasurementofattorneyexperience,findthatplaintiffs’attorneys

whohavehandledatleastfourcasesweremorelikelytoobtainmoneyforthe

plaintiffs.Krishnan,Davidoff,andThomas(2014)andHymanetal.(2015)find

thattoplawfirmsarebetteratwinninglawsuits.Sloan(1993:196–201)finds

that“specialist”plaintiffattorneysinmedicalmalpracticefaredbetterthan

“non‐specialist”plaintiffattorneysinreceivingcourt‐awardedmonetaryrecovery.

Chang, Chen & Lin

9

Specialistattorneysweredefinedasattorneyswhohavehandledfourormore

casesregardingmedicalmalpractice(basedoncourtrecordsavailableincounty

courthousesinFlorida),attorneyswhodesignatedthemselvesasexperts,and

thosewhowerelistedbyothersasexperts(Sloan1993:170).Ledermanand

Hrung(2006)useyearsofexperiencetomeasureattorneyexpertiseandfind

thatintaxlitigationagainsttheIRS,plaintiffswithattorneyrepresentationfare

betterthanthoseprose,andplaintiffswithmoreexperienceattorneyshave

higherrecoveryratios.Feldman(2015)talliesthenumberofcasesanattorney

representsbeforetheU.S.SupremeCourtandfindsthatbriefsofmore

experiencedattorneysaremorewidelyadoptedbySCOTUSinitsopinions.

Otherstudiesdisputetheresultsoftheaforementionedstudies.Greinerand

Pattanayak(2011:2125),usingdatafromreal‐worldrandomizedexperimenton

legalrepresentation,findthatthey“couldcometonofirmconclusionregardinga

use‐of‐representationeffectonthewinrate.”Goodman‐Delahuntyetal.(2010)

survey481sampledlitigatingattorneysintheU.S.andfindthattheyare

overconfidentinpredictingtheoutcomeoftheirowncasesandcalibrationdoes

notincreasewithyearsoflegalexperience.Kuo‐ChangHuang(2008),basedon

crudeyetcomprehensiveofficialdataoncivilcourtcases,arguesthatpartiesin

Taiwansoughtlegalrepresentationonlywhentheybelievedtheyhadsome

chancestowin.Thatis,civillitigationresultsweredrivenmainlybymeritsofthe

cases,andthedifferencesinwinningpercentagesshouldbeattributedtothe

selectioneffectofpartiesseekingrepresentationbasedonperceivedwinning

chances,notattorneyrepresentationitself.

1. AssignmentofCasestoAttorneys

RandomassignmentsofpublicdefendersinAbramsandYoon(2007)and

thosebetweenpublicdefendersandappointedcounselsinAndersonandHeaton

(2012)andKuo‐ChangHuang,Chen,andLin(2010)avoidtheselection

effect—attorneysaimtorepresentstrongercases.Otherpriorempiricalworks

oncivillawsuits,however,sufferfromtheproblemofselectioneffect,asplaintiffs’

anddefendants’attorneysarerarely,ifever,randomlyassignedtoclientsincivil

cases.15 Studiesofplaintiffs’attorneysintheU.S.foundthattheyroutinelyturn

15 Seethelonglistofliteraturecitedandcritiquedas“unworthyofcredence”(forlackofrandom

Chang, Chen & Lin

10

awaymorethanhalfofthepotentialclientsseekingrepresentation,aslawyers

whotakecontingencyfeesriskalotinhandlingmeritlesscases(Sloan1993:77;

Kritzer2004:73;Harris,Peeples,andMetzloff2008:257).

AttorneysinTaiwanwerenotrandomlyassignedtotortscasesaswell.Thus,

attorneyrepresentationandseniorityofattorneysmaybeassociatedwithcase

characteristics.Casecharacteristicsobservabletousdonotconfoundourresults,

astheireffectscanbecontrolledforintheregressionmodels.Manyimportant

casecharacteristics,includingproxiesforseverityofinjuries,werealreadycoded

andusedinourmodel.Unobservablefactorsmaycauseproblems.Yetthenature

ofcaraccidentlawsuitsandfixedattorneyfeestructureinTaiwanwouldgreatly

reducethefrequencyofcaseselection.Belowweadvanceacaseselectionstory

forfixed‐feeattorneys(inparticularplaintiffattorneysinTaiwan)basedon

rationalchoicetheoryandqualitativeinterviews.

Thefirststageofselectiontakesplacewhenatortvictimsearchforan

attorney.Thesolopractitionersweinterviewedtoldusthattheyreceivednew

clientsmostlyviareferralbynon‐attorneyfriends.Thepreconditionforstronger

casessystematicallyflowingtoseniorattorneysisthatfriendsofsenior

attorneysreferredonaveragestrongercasesorthatoldclientsofsenior

attorneyswhocamebackwithsystematicallystrongercases.Bothscenarios

strikeusasimplausible.Victimswhosearchforlawyersby,say,comparing

lawyerpracticestatistics(seebelow)on‐linemaytrytofindbetterattorneys,but

whileitisplausiblethatvictimsbelievedthatseniorattorneysaregenerally

better,thesevictimswhosearchbythemselvesdonotnecessarilyhavestronger

cases.Hence,thereappearstobenogoodreasontobelievethatseniorattorneys

systematicallyencounterclientswithstrongercases.

Thesecondstageofselectioniswhenattorneysdeterminewhethertotake

cases,thefocusoftheU.S.literature.IntheU.S.context,astrongcar‐accidenttort

caseforaplaintiffattorneyisoneinwhichthetortfeasor’snegligencecanbe

easilyestablishedandthevictimisseriouslyinjured,asthiskindofcaseismore

likelytoleadtohighamountofdamages,athirdorsoofwhichgoestothe

attorney.Contingent‐feeplaintiffattorneysareinclinedtotakeonlystrongcases.

Bycontrast,plaintiffanddefendantattorneysinTaiwanmostoftenreceivedflat

assignmentofattorneys)byGreinerandPattanayak(2011:2175–2184).

Chang, Chen & Lin

11

retainerfeesfortheirserviceinordinarytortslitigation(Kuo‐ChangHuang2008:

216).ThisfeestructureshouldnotmotivateTaiwaneseattorneystoturndown

casesasaggressivelyastheirAmericancolleaguesdo(seePartII).Manylawyers

weintervieweddidinformusthattheytendtoturndownmeritlesscases,even

thoughtheycancollecttheirfeesanyway.Amoralistreasonforlawyerstodoso

isthattheselawyersbelievethatthecomplainantshavenolegalcaseandshould

notsueorthattheyshouldsimplyusetheattorneyfeestopaybacktheirdebt,

notdefendinglegitimateclaims.Apracticalconcernisthatclientswhopayfixed

feesandgetnothingfromthelitigationwillkeep“bugging”theattorneys

afterwards.

Onemightconjecturethatseniorattorneysmaycherry‐pickthebestcases

thatgotothem.Whilethismightbetrue(wedonotyethavedatatoexaminethis

conjecture),thisislargelyinapplicableincar‐accidenttortcases.Our

intervieweesthinkthatthistypeofcasesmightbetheeasiestcasesamongcivil

matters.Noclassaction,complexcontracts,ormulti‐nationalconglomeratesare

involved.Thelawsuitisusuallyone‐on‐one,andjudgmentsofcausationand

negligencecanbeaidedbythewide‐spreadcar‐camsandCCTVsaswellasother

moderntechnologies.Theitemsthatvictimscanclaimareclear,theonly

questionishowmuch.Thetortfeasorsknowthattheyhavetocompensateand

occasionallycontendthatvictimswerecomparativelynegligent.Thequestionis

againhowmuch.Therefore,attorneysdonotseemtohavestrongincentivesto

turndownrepresentingtheplaintiffsorthedefendantsinagenuinecaraccident

case.Theycouldadjustthefixedfeesaccordingtotheamountatstakes.Wesee

nostrongreasonthatseniorattorneyswouldselecttorepresentonlyclientswith

moreseriousinjuriesormoreclearlynegligent.

Thethirdstageofselectionissettlement.Thereareseveralexistent

empiricalstudiesonsettlementofcivilcasesinTaiwan.Kuo‐ChangHuang(2008:

217)findsthatattorneyrepresentationisassociatedwithsettlementratesand

contendsthatitisduetothe“representationselection”—“aparty’sinitial

decisiontolitigatethecase[leads]tohimorherretainingcounsel.”Kuo‐Chang

Huang(2009)findsthatsettlementratesafteracivilcaseenteredcourtsin

Taiwanisabout30%.Kuo‐ChangHuang(2016)findsthatsettlementratesbefore

acivildisputeenteredcourtsinTaiwanisonaverage60%,andthatthe

Chang, Chen & Lin

12

scatterplotofsettlementrate(Yaxis)againstamountofstakes(Xaxis)revealsan

inverted‐Ushaperelation.Fromourinterviews,welearnedthatseniorattorneys

arebetteratsettlement.Thereis,however,noempiricalresearchonwhether

settlementratesvarybystrengthofplaintiffclaims.

Forourpurposes,aslongasproseparties,attorneysrepresentedbyjunior

attorneys,andthoserepresentedbyseniorattorneysdonotsystematicallysettle

differenttypesofcases,settlementshouldnotbiasourempiricalfindings.Thatis,

wecontendthatinthefirsttwostages,nostrongselectiontakesplace.Ifdisputes

withdifferentrepresentationstatusesweresettledindifferentpatterns,a

statisticallysignificantresultregardingrepresentationstatusesmaybespurious.

Again,thereseemstobegoodreasontobelievethatsettlementratesamong

differentrepresentationstatuseswillcorrelatewithafactorthatisnotalready

controlledbyourregressionmodels.

Finally,weconducttwo‐samplet‐testsandFisher’sexacttests16 toexamine

whethercasecharacteristicsarebalancedacrossthefollowingfourcomparison

groups:

1) caseswithproseplaintiffsversusthosewithrepresentedplaintiffs;

2) caseswithprosedefendantsversusthosewithrepresenteddefendants;

3) casesinwhichplaintiffswererepresentedbyseniorattorneysversus

thoseinwhichplaintiffswererepresentedbyjuniorattorneys;and

4) casesinwhichdefendantswererepresentedbyseniorattorneysversus

thoseinwhichdefendantswererepresentedbyjuniorattorneys.

AsshowninTable1,casecharacteristicsontheplaintiffsidearebalanced.Only

twoassociationsarestatisticallysignificantatthe5%levelandtheydonot

correlatewithcasestrength.Casecharacteristicsonthedefendantsideareless

balanced,asseveralcasecharacteristicsarestatisticallysignificantatthe5%

levelandsomeofthemareproxiesforseverityofinjury(alsoproxiesforcase

strength).

Asawhole,althoughourdataarenotproducedbya(natural)experimentin

which attorneys are randomly assigned to cases, thanks to the attorney fee

structure,theselectioneffectappearstobeminor,especiallyregardingselections

16 Inthesetests,weused,inrows,thevariablesusedinthehedonicregressionmodels;thecolumnvariablesareeither“hiringattorneysversusnothiringattorneys”or“hiringattorneyswithabove‐medianexperienceversushiringattorneyswithbelow‐medianexperience.”

Chang, Chen & Lin

13

ontheplaintiffside.Totheextentthatattorneyrepresentationstatusesarenot

associatedwithunobservablevariables,ourregressionmodelsshouldbeableto

tease out the effect of attorney representation and the value of attorney

experience.17

17 Forrobustnesscheck,wealsotriedpropensityscorematchingmethodbasedonwhetherattorneyshaveabove‐orbelow‐medianexperience.Duetothebalancedcasecharacteristicsmentionedinthetext,theregressionresultsareverysimilartothosereportedinTable3.

Chang, Chen & Lin

14

Table1HomogeneityofCaseCharacteristicsAcrossLegalRepresentation

Statuses

PanelA:Continuousvariables

Prose

plaintiffsor

not

Prosedefendants

ornot

Plaintiff

attorneysenior

orjunior

Defendantattorney

seniororjunior

Judgeexperience 0.584 0.745 0.745 0.887

Declaredincarcerationtime 0.617 0.912 0.149 0.931

Incurredmedicalexpenses 0.212 0.069+ 0.402 0.049*

Plaintiffage 0.626 0.003** 0.423 0.003**

Plaintiffincome 0.739 0.061+ 0.304 0.139

Defendantincome 0.551 0.394 0.739 0.365

Cellscontainp‐valuesfortwo‐samplet‐tests.Valuesinrowvariablesare

transformedtonaturallogbeforethet‐tests.

***p<0.001,**p<0.01,*p<0.05,+p<0.1.

PanelB:Categoricalvariable

Prose

plaintiffsor

not

Prose

defendantsor

not

Plaintiff

attorneysenior

orjunior

Defendant

attorneysenior

orjunior

Defendantsincludecorporations 0.003** 0.152 0.189 0.095+

Drivingundertheinfluenceofalcohol 0.594 1.000 0.843 1.000

Hitandrun 0.114 0.563 0.567 0.569

Plaintiffpaysfilingfees 0.020* 0.808 0.465 0.808

7injurylevels 0.576 0.020* 0.167 0.019*

2injurylevels 1.000 0.002** 0.360 0.002**

Cellscontainp‐valuesfortwo‐tailedFisher’sexacttests.

***p<0.001,**p<0.01,*p<0.05,+p<0.1.

2. PlaintiffAttorneys

PlaintiffattorneysinTaiwanmostlychargefixedfeesfortortdamagescases,

andtheirincentivesarethusdifferentfromthosein,forinstance,theU.S.and

Israelwhochargecontingentfees.Theincentivesofcontingent‐feeattorneysare

alignedwiththeinterestsoftheirclients.Aslongasitiscost‐justified,attorneys

Chang, Chen & Lin

15

workhardertowinlargeramountofdamages,astheytakehomeathirdorsoof

thedamages.Bycontrast,theprinciple‐agentproblemismoreacuteamong

fixed‐feeattorneys.Thefinancialincomeoftheseattorneysisindependentof

caseoutcomes.Attorneys,thus,totheextentthattheycannotexanteadjustthe

amountoffees,wouldprefereasytocomplexcases.Attorneysalsohave

incentivestosettleearly,evenatalowfigure.Otherthingsbeingequal,fixed‐fee

attorneysmaynotworkashardascontingent‐feeattorneysoncases.

Attorneyscareaboutmaintainingreputationsandgettingmoreclientsin

thefuture.Injurisdictionswhereplaintiffshavetospelloutaddamnumclauses

andblockbustercasesarerare,theclaim‐grantratiocanroughlymeasurethe

extenttowhichplaintiffsanddefendantseach“win.”CourtsinTaiwan

summarizetheclaim‐grantratiointhecourtfeepercentage(seePartII).

Assumingthatattorneyscontrolhowmuchtoclaim,thecourtfeepercentagecan

serveasaproxyforattorneyskills.PingLuWeb,18 aleadinglegalserviceprovider

inTaiwan,hasusedcourtfeepercentagesasthecentralstatisticsincomparing

attorneyskills.Inthelongrun,ahighaveragecourtfeepercentageindicatesthat

anattorneylosesalotofpleadingordefenses.Usingaveragecourtfee

percentagestoevaluatehowgoodsattorneysareatpersuadingjudgesand

predictingcaseoutcomesexanteandmakingclaimsaccordinglyisparticularly

aptincar‐accidenttortcases,asthesecasesaremostlyabouttheamountof

compensation.

Fixed‐feeattorneyswhocareabouttheircourtfeepercentagesquawinning

recordswouldtendtomakeconservativeclaims,sothatcourtswouldgrantmost

oftheirclaims,keepingtheircourt‐fee‐percentagestatisticslowandtheclients’

feeslow.Conservativeclaims,however,donotnecessarilymaximizeclient

interests.Courtscannotawardmorethantheamountsplaintiffsclaim,andfiling

feesareabout1percentoftheclaimedamount.Whenclaiming100dollarsmore

wouldleadtoanincreaseinexpectedawardsofmorethan1dollarorso,

attorneyswhocaresolelyabouttheinterestoftheirclientsshouldandwilldoso.

Iftheexpectedincreaseinawardsis,say,5dollars,however,theattorney’sfee

percentageisquitelikelytoincrease,unlesslessthan5%oftheotherpartofthe

claimisexpectedtobegranted.Attorneysthusmaynothaveincentivestoadvise

18 PingLuWebliterallymeansthewebsiteforevaluatingattorneysinMandarinChinese.

Chang, Chen & Lin

16

theirclientstoclaimmore.

Providedthatjudgeswereunbiased,claimingmorethantheattorneys’best

estimatesofcourtawardsbasedonwhathighercourtshaverecentlyallowedis

unlikelytoswayjudges.Asaresult,over‐claimingisirrationalfortheattorneys

andtheirclients.Nonetheless,ifjudgesweresubjecttotheanchoringeffect,

over‐claimingmakeseconomicsensefortheclients,butattorneyswhoare

sensitivetotheirwinningrecordsmaybeinclinedtobeconservative.

Ourhypothesisisthatexperiencedattorneys,ascomparedtoinexperienced

attorneys,tendtoover‐claim.Theeconomicreasoningbehindthishypothesisis

thatseniorandjuniorattorneyshavedifferentbusinessmodels.Experienced

attorneysaremorelikelytohaveafirmerclientbaseandretainnewclients

throughwordsofmouth(referralsnotjustbypersonalfriendsbutalsoby

formerclients).Withalongertrackrecordandreputation,theirbusinessistoa

lesserextentinfluencedbyanincreaseinaveragecourtfeepercentage.Senior

attorneysthusaremorewillingtomakebolderclaimsthanjuniorcolleagues.As

experiencedattorneysgenerallychargeahigherfee,theywouldneedtoshow

theirclientsthattheyareworthit,andbeatingtheaverage(claimingmoreand

gettingmore)issuchasignal.Sometimes,accordingtoattorneysweinterview,

winninganimprobableclaimwouldleadtoabonus(“redenvelope”)givenbythe

clienttotheattorney.AsFigure1shows,thefilingfeepercentagesandthe

numberofcivilcasesplaintiffattorneyshaverepresenteddoappeartohavea

positiverelationship.

Chang, Chen & Lin

17

Figure1Plaintiffattorneyexperienceandcourtfeepercentage

N=280.Districtcourtcasesinourdatasetinwhichatleastoneattorneywas

retainedbytheplaintiffareincluded.

Therearealsopsychologicalreasonsfordifferentclaimingpatterns.

Whetherattorneysareawareoftheanchoringeffectornot,seniorattorneysare

morelikelytoover‐claim,astheyaremoreconfidentintheirownpersuasion

skills(Goodman‐Delahuntyetal.2010).Oneexperiencedattorneywe

interviewedadvancedaninterestingtheory:forplaintiffattorneys,eachcasehas

areasonablerangeforpainandsufferingdamages.Experiencedattorneystend

toclaimanamountthatisattheupperechelon,becausetheirhigherincomeand

otheraspectsoftheirlifeexperiencemakethehigherfiguremore“natural.”

3. DefendantAttorneys

Defendantattorneysareexpectedtoprioritizedismissingcasesagainsttheir

clients.Ascasesinourdatasetarelimitedtothoseinwhichdefendantswere

foundliable.Wearenotabletotestwhetherhiring(experienced)attorneys

increasethechanceofcasedismissal.

Defendantattorneysarealsoexpectedtoreducetheamountofdamages.

0.2

.4.6

.81

Per

cent

age

of fi

ling

fee

pla

intif

f has

to p

ay

1 1.5 2 2.5 3Number of civil cases plaintiff attorneys represented in 2000-2014 (log10)

Lowess line 1 observation

Chang, Chen & Lin

18

Pecuniarydamagesaretypicallyformulaic.Defenseattorneyscangreatlyreduce

theamountofcompensationusuallybyprovingthatplaintiffsarecomparatively

negligent.Aspainandsufferingdamagesarediscretionary,weconjecturethat

defenseattorneyswouldchallengethisclaim.Inpractice,somedefendants

counteredaspecificamountofdamages,somesimplyobjectedthattheclaimed

damagesaretoohigh,andsomefailedtomakeanyobjection.Wehypothesize

thatthecounter‐claimingpatternisaffectedbyseveralfactorsincludinghow

muchplaintiffs’claimedamountdeviatesfromthehistorictrend,whether

defendantsorplaintiffshireattorneysandhowexperiencedtheyare,whether

defendantsincludecorporations,anddefendants’income.

ItisalsoworthnotingthatintheTaiwancaseswestudy,nodefendantisan

insurancecompany,andcorporatedefendantsareonlyvicariouslyliable.19 In

otherwords,defendantsaretherealtortfeasorswhodroveamobilevehicle

whileinjuringtheplaintiffs/victims.

IV. MODELS

Asophisticatedstructuralequationmodelisspecifiedtoclosely

approximateareal‐worlddecision‐makingprocess.Formallyandinformally,we

interviewedmanyattorneysandjudgesofallexperiencelevelsinTaiwanin

privateandinfocusgroups.Thecoreinsightisthatbeforetheplaintiffs’

attorneysformulatetherequestedamountofpainandsufferingdamages,and

beforedistrictcourtjudgesmaketheirdecisions,bothhavesearchedinthe

officialcourtcasedatabase(http://jirs.judicial.gov.tw/Index.htm)forsimilar

casespreviouslydecidedindistrictcourtsandhighcourts,tobeconsistentwith

precedents.Therefore,thebasicset‐upofourregressionmodelistotestwhether

districtcourtjudgesattimeTwoulddeviatefromtheassessmentpatternatthe

districtcourtlevelattimeT‐1iftheassessmentpatternatthehighcourtlevelat

timeT‐1isdifferentfromthatatthedistrictcourtlevelattimeT‐1.20 Casefacts

19 Insharpcontrast,defendantsinpersonalinjurycaseselsewhereareoftenlargefirmsorinsurancecompanies(ZamirandRitov2010:276).20 Inthisarticle,wefocusontheinteractionofdistrictcourtsandhighcourts(whichreviewthecasedenovo),leavingouttheTaiwanSupremeCourt.Whileafew“selectedprecedents”renderedbytheTaiwanSupremeCourtinthe1950sand1960swerestillfrequentlycitedindistrict‐courtandhigh‐courtdecisions.Thoseprecedentsprovidehardlyanyguidanceforlowercourtjudges.Nojudgesweinterviewedsuggestthattheyoranycolleaguederivedtheamountofpainand

Chang, Chen & Lin

19

(includingvictims’severityofinjury),judges’andattorneys’experience,etc.are

addedtotheregressionmodels.

Morespecifically,underourmodel,districtcourtjudgeswouldask,giventhe

factsofthecasesathand,howmuchpainandsufferingdamagesthehighcourts

andthedistrictcourtswouldaward.Topredictwhatjudgeswouldestimateas

thetwohypotheticaldamagesawards,weusehedonicregressionmodelsto

predicttheamountofpainandsufferingdamagesforeachdistrictcourtcasein

ourmainresearchperiod(Sep.5,2013–Sep.2,2014),basedonthehigh‐court

ordistrict‐courtcasesdecidedwithinthepast3months21 ofeachcase.Thatis,

thehedonicregressionmodelwasrun102(=51*2)times,twoforeachweek.22

Therefore,eachdistrictcourtcaseintheresearchperiodispairedwithtwo

estimatesofpainandsufferingdamages:oneiswhatthehighcourtwouldhave

rendered,andtheotheriswhatthedistrictcourtwouldhaveawarded.These

estimatesarehereinafterreferredtoasthetrend(ofpainandsufferingawards).

Moreconcretely,toestimatethepainandsufferingdamagesforadistrictcourt

caserenderedinthefirstweekofApril,weusedthecasesdecidedinthefirst

threemonthsoftheyearrenderedinthehighcourtinahedonicregression

model.Thecoefficientofthemodelandthefactsofthesaidcasecanprovide

estimatedpainandsufferingdamages—thebestestimateofwhatthehighcourt

wouldhavedonehadthiscasecamebeforeitinJanuary,February,orMarch.We

thenrepeatedthesameprocedureondistrictcourtcases,togetholdofthebest

estimateofwhatthedistrictcourtwouldhaverenderedhadthiscasecame

beforeitafewweeksearlier.

Afterfurthercomputations,werunstructuralequationmodels(SEM)totest

whethertherulingonpainandsufferingdamagesbydistrictcourtjudgesand

sufferingdamagesfromtheselectedprecedents.Thus,theissuewestudyhereisdifferentfromthatinpriorliteraturethatfocusesontheinteractionbetweenthehighestcourtofajurisdictionanditslowercourts. 21 Weassume(basedoninterviews)thatjudgeswouldsearchthemostrecentcasesasreferences.Choosingthreemonthsasthescopeisasomewhatarbitrary,exantedecision,though.(Wecoded4monthsofcasespriortothemainresearchperiod,becausetheplaintiffswereassumedtoclaim1monthbeforeadjudication,andwethusneededthreemoremonthspriortothefirstplaintiffclaimtoestimatewhatthisplaintiffwouldhaveconsideredtobethetrend.)Wehavetriedtouse2.5and3.5monthsasthecut‐off,andtheresultsregardingtheattorneyandjudgeexperiencevariablesareessentiallythesame. 22 Thereareintotal51weeksinourmainresearchperiod(definedinPartV).51modelsusedistrictcourtcasesasobservations,whereastheother51modelsusehighcourtcases.

Chang, Chen & Lin

20

thoseclaimedbyplaintiffswereaffectedbythetrend,andwhetherthedeviation

fromthetrendcanbeattributedtojuristexperienceandcasefacts.We

expatriatetheempiricalstrategyinthefollowing:

A. OLSModelstoPredictDistrictCourtandHighCourtAwards

OurOLShedonicregressionmodelstakethefollowingform23:

PSD=β0+β1INJURY+β2MED+β3JAIL+β4CRIME+β5DCORP+β6CHAR+ε

wherePSD is thenatural logof the judge’spainandsufferingdamagesaward;

INJURYisadummyvariableindicatingwhetherthevictimsufferedfromminor

injury or serious injury (defined according to Article 10 of Taiwan’s Criminal

Code);24 MED are a variable presenting the natural log of court‐adjudicated

medical‐relatedexpensesalready incurredandexpectedto incurandadummy

variablethatequals0ifmedicalexpensesare0;JAIListhenaturallogofmonths

criminalcourtshavesentencedthedefendantstobe incarcerated;CRIMEarea

dummy variable that equals 1 if the defendant drove under the influence of

alcoholandadummyvariablethatequals1ifthedefendanthittheplaintiffand

ranawaywithoutassistingtheinjuredplaintiff(bothareanindependenttypeof

crime under Taiwan’s Criminal Code);DCORP is a dummy variable indicating

whetheranyofthedefendantsisacorporation(whichisvicariouslyliableforits

employees);CHARincludeplaintiffs’ages,plaintiffs’income,defendants’income,

and three dummy variables that equal 1 when age or income information is

missing.Thecoefficientstobeestimatedareβn;εisanerrorterm.

Morespecifically,utilizingthecomprehensive,officialcourtcasewebsite,

23 Oneofushascollaboratedinjointresearchprojectsonpainandsufferingdamagesforpersonalinjury(Changetal.2014),wrongfuldeath(Changetal.2015),anddefamation(Chang,Ho,andHsu2016).Thoseworksexplorethedeterminantsofcourt‐adjudicatedpainandsufferingdamagesinTaiwan.Changetal.(2014)inparticularfindsthatthelevelofinjuryandmedicalexpensesalonecanexplainmorethanhalfofthevariationfromtheaverageamount.Thus,inthisarticle,wealsousethelevelofinjuryandmedicalexpensesasthemajordeterminantsinthehedonicregressionmodels(SectionA).24 Aseriousinjuryisoneofthefollowingconditions:1.Destructionoforseriousdamagetothesightofoneorbotheyes;2.Destructionoforseriousdamagetothehearingofoneorbothears;3.Destructionoforseriousdamagetothefunctionsofspeech,taste,orsmell;4.Destructionoforseriousdamagetothefunctionofoneormorelimbs;5.Destructionoforseriousdamagetothepowerofreproduction;and6.Otherseriousinjurytobodyortohealththatiseitherimpossibleordifficulttocure.

Chang, Chen & Lin

21

weareabletofindwhetherthecivildefendantshavebeenconvictedbeforethe

civil court rendered its decisions (most of them were). We chronicled the

declaredsentences(JAIL),amongothers,fromthecriminalcourtdecisions.Our

conjectureisthatthelengthofthedeclaredsentencemightaffecttheamountof

painandsufferingdamagesascivilcourt judgescouldhavetaken itasanother

measureoftheseverityofthetortfeasors’acts.Thedeclaredsentences,however,

are not a good measure of how long the tortfeasors have suffered in prison.

Tortfeasorswhosedeclaredsentencesaresixmonthsorshortercanavoidbeing

jailed by paying criminal fines instead, and 91% of our cases fall into this

category.Manyoftherestofthecasescouldstillbeappealed.

DCORP tries to capture the effect of the deep pockets of corporate

defendants. Prior literature has shown that the deep pocket effect has its

presenceinTaiwanesecourt(Changetal.2014;Chang,Ho,andHsu2016).

WhenalldistrictcourtcaseswereputintothisOLSmodel,theR‐squareis

0.63.WhenallhighcourtcaseswereputintothisOLSmodel,theR‐squareis0.60.

SeeAppendixAforregressionresults.

B. StructuralEquationModelonDeviationofCourtAwardfromTrend

Judgesindistrictcourtssurveyedrecentsimilarcases(thehistorictrend)

andthenevaluatedthecasesathandbasedonthehistorictrendintheprevious

period.Theirmaindecisionsare thereforehowmuch theadjudicatedpainand

suffering damages (R) for the current case should deviate from the trend. The

deviation (R–bl) from the trend in district court (bl) is called “intra‐court

deviation.” The intra‐court deviation, as hypothesized, might be based on the

“inter‐court deviation” (bh–bl)—that is, the historic trend in high courtsminus

that in district courts. For example, assume that the district court at time T‐1

awarded1milliondollarstoavictiminacertaintypeofcaseandthehighcourt

attimeT‐1awarded1.5milliondollars,weconjecturethatdistrictcourtattime

T would deviate from 1 million dollars and move toward 1.5 million dollars.

Simultaneously,thedeviationoftheplaintiffs’claim(P)fromblmightserveasan

anchortothejudges’decisions.Notethatinthispaper,R,blandbharemeasured

asthenaturallogarithmoftheoriginalvaluestopromotenormality.Forthesake

Chang, Chen & Lin

22

ofbrevity,thewords“naturallog”or“ln”willbeomitted.

Toascertainthevaluesoftheaforementionedvariables,wefirstconstruct

hedonic estimates of thehistoric trend set in thepreviousperiodbyusing the

cases adjudicated 1 to 12 weeks before the verdict of the case at hand, as

describedinSectionIV.A.Wethenpluginthecharacteristicsofthecurrentcase

toassessthebestestimateofpainandsufferingdamagesinthatcase.Finally,we

calculate the intra‐court deviation (R–bl), inter‐court deviation (bh–bl), and

plaintiff‐claimdeviation(theplaintiffs’claimsminusthehistorictrendindistrict

courts;P–bl).

A structural model is warranted to handle the endogeneity problem

inherent in the nature of our inquiry.25 First, the plaintiffs’ claims (or, for that

matter,plaintiff‐claimdeviation)wereaffectedbycasefactsbutinfluencedcourt

adjudication aswell. The problem is accounted for in the structuralmodel by

using the plaintiff‐claim deviation as one independent variable in the first

equation and the dependent variable in the second equation. Second, some

characteristics considered by judges and plaintiffs are not observed by

researchers. Our structuralmodel takes into account the correlations between

theerrortermsofthetwoequations, thuscontrollingtheendogeneityproblem

(Wooldridge2010:681).26 Specifically,werunthefollowingstructuralequation

model:

)1(,)()()( 1215,*,432

*,1,,, iiiilil

ajiliilihili eWXbbEEbPbbbR

.)( 2437*,

*,6

*, iiiilih

aili eZXbbEbP (2)

In equations (1), R–bl is the intra‐court deviation; bh–bl is the inter‐court

25 Asourlaterresultwouldshow,therho(Table3)thatcaptureswhetherthecorrelationbetweenthetwoerrortermsinthetwoequationsinthestructuralequationmodelisnotstatisticallysignificant.Inotherwords,thestructuralmodelinformsusthatnoendogeneityproblemwasdetected.Theexantedecisiontouseastructureequationmodelisstilljustified,asoneshouldworrythatomittedvariablesmayexist.26 Whileitistraditionallydifficulttoanalyticallyderivetheconditionaldensityandnumericallymaximizethelikelihoodinamulti‐equationsystem,Roodman(2011:681–685)providesausefulSTATAprocedureCMP(ConditionalMixedProcessestimatorwithrandomeffectsandcoefficients),whichcanbeappliedtoinstrumentalsystemproblemstoestimatethesystemunderthejointnormalityassumption.Inthestructuralequationmodelweusedandreportlater,Shapiro‐WilkWtestsfornormalitysuggestthatwecannotrejectthenullhypothesisthattheresidualsofequation1andequation2arenormal.

Applyingthemaximumlikelihoodapproachinlinearmodelstocontroltheendogenousproblem,seeGreene(2003:402);DavidsonandMacKinnon(2004:537–538);DavidsonandMacKinnon(1993:644–651);CameronandTrivedi(2005:191).

Chang, Chen & Lin

23

deviation; P–b*l is plaintiff‐claim deviation;27 Ea are a dummy variable that

indicates whether plaintiffs retain attorneys, the natural logarithm of plaintiff

attorneys’ civil experience, and adummyvariable that equals 1 if the attorney

startedpracticingbeforeyear2000,thushisorherexperiencesunder‐estimated

byourdata;Ejrepresentsthenaturallogarithmofjudges’civilexperienceanda

dummyvariablethatequals1ifthejudgestartedhisorhercareeronthebench

before year 2000; b*l–bl is the historic change in district courts between

estimatedpain and suffering awards 1 to 12weeks before the verdict (bl) and

estimatedpainandsufferingawards5to16weeksbeforetheverdict(b*l).28

The common explanatory variables in equations (1) and (2), X, mainly

capturetheinformationorfactorsthatcouldexplainthedecisionsofthejudges

and plaintiffs. They consist of 6 dummy variables on levels of injury.29 X also

includes 18 dummy variables that control for which district courts made the

27 Tobemoreexact,P–b*listhedifferencebetweenplaintiffs’claimsandestimatedpainandsufferingawards5to16weeksbeforetheverdict.Wedeductfromtheformerthelatter,ratherthanestimatedpainandsufferingawards1to12weeksbeforetheverdict,becauseplaintiffs’claimswereonaveragemadeaboutonemonthbeforejudges’ruling.Thenewtrendthathappens1to4weeksbeforetheverdictwouldbeunbeknownsttotheplaintiffswhentheymadetheclaims.Seealsofootnote28formoreexplanation. 28 Thisvariableitselfisnotofinterest.Weaddedthistoadjustthispotentialtime‐inconsistencyofthehistorictrend.Morespecifically,P–b*lisusedasanindependentvariableinequation(1)andasthedependentvariableinequation(2)toaccountfortheendogeneityproblem.Nonetheless,theanchoringeffectcreatedbytheplaintiffs’claimsmaynotderivefromthedifferencebetweentheclaimandtheoldtrend(5–16weeksbeforeverdict),butbetweentheclaimandthenewtrend(1–12weeksbeforeverdict).Tobeabletoascertainthemagnitudeoftheanchoringeffect,b*l–blisaddedasadjustment.Morespecifically,equation(1)canbere‐writtenasfollows:

,))(()()(

)()()(

12125,*,432,1,,

1215,*,432

*,1,,,

iiiililaj

iliilih

iiiililaj

iliilihili

eWXbbEEbPbb

eWXbbEEbPbbbR

wherethevariableinthesecondterm(p–b)measurestheanchoringeffectcreatedbythedeviationofplaintiffs’claimsfromthenewtrend.Astheaboveequationshows,(p–b)and(p–b*)bothhaveγ2asitscoefficient.Thatis,theregressioncoefficientγ2canberegardedasthemeasurementoftheanchoringeffect.ConductingaregressionofR–blonP–bl,b*l–blandothervariablewillresultinthesamecoefficients(exceptthatofb*l–bl)asconductingaregressionbasedonequation(1).29 Hereweclassifiedthevictims’injuriesinto9levelsbasedonNAIC(NationalAssociationofInsuranceCommissioners)scale,asitisamoredetailedclassificationofinjuries.TheNAICscalehasbeenusedinpriorempiricalstudies.See,e.g.,Vidmar,Gross,andRose(1998:283);Sloan(1993:23).The9levelsare:1.Emotionalonly(fright,nophysicaldamage);2.Temporaryinsignificant(lacerations,contusions,minorscars,rash;norecoverydelay);3.Temporaryminor(infections,fracture,fallinhospital;recoverydelayed);4.Temporarymajor(burns,surgicalmaterialleft,drugsideeffect,braindamage;recoverydelayed);5.Permanentminor(lossoffingers,lossordamagetoorgans;includesnondisablinginjuries);6.Permanentsignificant(deafness,lossoflimb,lossofeye,lossofonekidneyorlung);7.Permanentmajor(paraplegia,blindness,lossoftwolimbs,braindamage);8.Permanentgrave(quadriplegia,severebraindamage,lifelongcareorfatalprognosis);9.Death.Novictiminourcasessufferedmerelylevel‐oneinjury.Weexcludedeathcasesastheyarecategoricallyanddoctrinallydifferent.

Chang, Chen & Lin

24

decisions. Inaddition,Xcontainsplaintiffs’age innatural log;plaintiffs’annual

income in natural log; 30 two dummy variables that equal one when an

observation contains missing values in age and income; whether defendants

droveundertheinfluenceofalcohol;whetherdefendantshitandran;thelength

of defendants’ declared criminal sentences; andwhether defendants include a

corporationvicariouslyliableforthenaturalpersonwhocausedtheaccident.

W represents variables that are only used in the first equation, including

whetherdefendantshiredattorneys,thecivilexperienceofdefendantattorneys,

adummyvariablethatequals1ifthedefenseattorneystartedpracticingbefore

year 2000, natural log of court‐adjudicated medical expenses, and a dummy

variablethatequals0whencourt‐adjudicatedmedicalexpensesare0.

The specification in equation (1) examines the driving force of judicial

decision‐making, whereas that in equation (2) teases out what drove the

plaintiffs’ decisions. As the plaintiffs’ claims were on averagemade about one

month before judges’ ruling,31 we again construct hedonic estimates of the

historictrendsinhighcourtsanddistrictcourtsbyusingcasesrendered5to16

weeks before court verdict (that is, 1–12 weeks before plaintiffs made their

claimstocourts).Then,wepluginthecharacteristicsofthecasesinquestionto

estimatetheadjudicateddamages.Thus,thedependentvariableinequation(2)

is thedeviationof theplaintiffs’ claimsaway from thehistoric trend indistrict

courts 5 to 16 weeks before court verdict. The difference of the estimated

damagesatdistrictandhighcourts5to16weeksbeforecourtverdict(bh*–bl*)is

oneofthemajorindependentvariablesinequation(2).32

30 TheannualincomeinformationwasacquiredbycourtsviatheMinistryofFinanceandreportedinthecourtdecisions. 31 Accordingtoofficialsummarystatistics,theaveragehandlingtimeforadistrictcourtcivilcaseis30days.Asanapproximation,weused4weeksasthedifferencesintimebetweenplaintiffs’claimingandjudges’adjudicating.Eachcase,ofcourse,isdifferent.Somecasessurelytookmorethan30daystoreachtheconclusionofatrial,butwehavenoinformationastohowlong.Inaddition,plaintiffscanchangetheiramountofclaimsbeforemakingtheclosingstatements.Whilesomecourtdecisionsdocontaininformationastowhetherplaintiffsincreasedordecreasedtheirclaimedamount,thosedecisionsdidnotspelloutwhethertheclaimofpecuniarydamagesorthatofnon‐pecuniarydamageswaschanged. Inshort,theremustbedifferencesintimebetweenplaintiffs’claimingandjudges’adjudicating;thatis,certainnewcasescouldbetakenintoconsiderationbyjudgesbutnotbyattorneys.Butwearenotentirelysurehowlongthetimegapis.Ourmodelassumesthatplaintiffstookintoaccountdistrictcourtandhighcourtcasesrenderedinthepreviousthreemonthswhentheyformulatedtheirrequestedamountofpainandsufferingdamages4weeksbeforecourtadjudication.Seealsofootnotes27and28.32 Inunreportedmodels,adummyvariablecapturingwhetherdefendantsexplicitlycounter

Chang, Chen & Lin

25

Due to the nature of our data and for identification purposes, specific

explanatory variables, in addition to X, are added for equation (2). These

potentialdeterminantsofreactions(Z) includethethreefollowingindependent

variables:whetherplaintiffshavetopayfilingfee,plaintiffs’incurredamountof

medical expenses in nature log, and a dummyvariable that equals 0when the

incurredmedicalexpensesare0.Whetherplaintiffshavetopayfilingfeesshould

notaffecttheawardofpainandsufferingdamages,asithasnothingtodowith

thelossesofplaintiffs,whileplaintiffswhodidnothavetopaytheproratafiling

fees have incentives to claim above the trend. Plaintiffs’ incurred medical

expensesarelargerthanorequaltocourts’adjudication,becausecourtsevaluate

medical expenses and usually award only part of them. We conjecture that

plaintiffs and judges each use their own incurred and adjudicated medical

expenses, respectively, as proxies for the level of pain; thus, they are put to

equation(2)andequation(1),respectively.Inaddition,thevariable(bh*–bl*)also

servesthepurposeofidentification,asitisonlyincludedinequation(2).(bh*–bl*)

should not affect judges, because judges should not care about the changing

trendasofonemonthbeforetheadjudication;rather, judgesshouldcareabout

the changing trend as of adjudication. Hence, (bh–bl) is put in equation (1)

instead.

Finally, we assume that the error terms eki (k = 1, 2) are jointly normally

distributedwithameanofzero.33

Ourmaininterestisinthevaluesoftheestimatedcoefficientsγ1toγ7,except

γ5. Ifγ1 is statisticallysignificantandpositive, thedistrictcourt tends to follow

thelatesttrendinhighcourts.Ifγ2 isstatisticallysignificantandpositive, there

exists an anchoring effect created by plaintiffs’ claims. The statistical

significances of γ3and γ4 demonstrate whether the increases in attorney and

judgeexperience lead todistrict courts’deviating from theirhistoric trends.γ6

informs whether attorney experience affects plaintiff claims. γ7 teases out

whether plaintiffs take into account changes in trend when formulating their

plaintiffs’claimedpainandsufferingdamagesisincludedinthefirstequation.Theresultisnotrobust.Sometimesitisstatisticallysignificant(withtheexpectednegativesign),whilesometimesitisnotstatisticallysignificant.Thisdummyvariableisultimatelyomittedduetopotentialendogeneityproblem.33 Wesetthevarianceofe2ito1toidentifytheparameters.Moreover,weallowforresidualcorrelationbetweenequations1and2.

Chang, Chen & Lin

26

claims.

Finally, several technical and substantive checks have been done. Robust

standard errors are used. Judge random effects are controlled in the first

equation.Thestructuralequationmodel is clusteredbycases.Severaldifferent

sets of variables have been added to unreported models to explore the

relationship among case outcomes, judge experience, and attorney experience:

thesupremecourtrepresentationexperienceofplaintiffanddefendantattorneys;

age differences between attorneys; the number of times the plaintiff attorneys

hasappearedbeforethejudges;thenumberoftimestheplaintiffattorneysand

thedefendantattorneyshaverepresentedoppositepartiesinlitigation;andthe

gender combination of the judges and the plaintiffs. None of these unreported

variablesarestatisticallysignificant.

V. DATA

A legal service provider, Ping‐LuWeb (www.pingluweb.com), provides us

with a rich data set on the experience of all Taiwanese attorneys and all

TaiwanesejudgesasofJune30,2014.Ping‐LuWebdownloadedmillionsofcases,

freely available on the official court case database

(http://jirs.judicial.gov.tw/Index.htm), and tallied the number of times any

lawyer (judge) appears as anattorney (judge) of a case. Theonlydrawbackof

this data set is that, as most cases before year 2000 are not available in the

officialcourtcasedatabase,theexperienceofjudgesandattorneyswhostarted

theircareersbefore2000willbeunder‐estimated.34

AssummarizedinPartIII,nopriorwork,tothebestofourknowledge,hasa

comprehensive,continuousandaccuratemeasureofjuristexperience.Mostdata

measure the number of practicing years or the number of cases an attorney

representedwithinthesample.Ourdatasetallowsustogaugejudges’experience

in terms of number of years on the bench and the total number of signed

opinionsregardingcivil,criminal,andadministrativematters.Italsoenablesus

34 Intheregressionmodels,wehavetriedaddingdummyvariablesthatindicatewhetherplaintiffs’ordefendants’attorneysstartedtopracticebeforeyear2000,inordertocapturethepotentialeffectofunder‐estimatingtheexperienceoftheseseniorattorneys.Thedummyvariablesarenotstatisticallysignificant.

Chang, Chen & Lin

27

tomeasureattorneys’experienceby thenumberofyearssinceanattorneygot

herlicense;thenumberofyearssinceanattorneyfirstrepresentedaclientina

lawsuit; and the number of court cases in which an attorney is listed as a

representativeofeithertheplaintifforthedefendant.Thepercentagesofcasesin

whichanattorneyrepresentsanindividual,acorporation,orthegovernmentis

alsoknown.SeeFigure2,Figure3,andFigure4forthedistributionofattorneys’

andjudges’experience,asmeasuredbythenumberofcivilcaseshandled.

To capture the civil‐litigation experience of plaintiffs’ attorneys,

defendants’ attorneys, and judges,weuse the number of civil cases (attorneys

represent and judges adjudicate) as the measure of their legal experience.35

Whentherearemultipleattorneyshiredbyeitherparty(sometimesonelitigant

hired multiple attorneys, and sometimes there were multiple plaintiffs or

defendantsinonecaseandtheyhireddifferentattorneys),theexperienceofthe

most seniorattorney isused.Usually,one judgesitson thebench foradistrict

courttrial.Nevertheless,whenajuniorjudgewithlessthan2yearsofexperience

onthebenchisrandomlyassignedacase,twomoreseniorcolleagueswouldjoin

him or her to form a panel. (There are 40 such cases in our data set.) In this

scenario,weusetheexperienceofthemostsenior judge(usuallythepresiding

judge)tomeasurethejudgeexperienceofthosecases.

35 Numberofyearsofpracticehasbeenusedinourregressionmodels,withorwithoutthevariableonthenumberofhandledcases.Wefound,however,thattheformerisnotstatisticallysignificantinanymodel.Thisisaninterestingcontrasttothepriorstudiesthatusedthisasthemeasuringrodofattorneyexperienceandfoundstatisticallysignificantresults.

Chang, Chen & Lin

28

Figure2DistributionofPlaintiffAttorneyExperience

N=324.In63observations,plaintiffsdidnothireattorneys.

05

1015

Per

cent

age

1 10 100 1,000Number of civil cases plaintiff attorneys have represented (log10)

Chang, Chen & Lin

29

Figure3DistributionofDefendantAttorneyExperience

N=201.In186observations,defendantsdidnothireattorneys.

05

1015

Per

cent

age

1 10 100 1,000Number of civil cases defendant attorneys have represented (log10)

Chang, Chen & Lin

30

Figure4DistributionofJudgeExperience

N=387.

Tomatchwiththisuniquedataset,thisprojecthascodedpainandsuffering

damages cases between September 5, 2013 and September 2, 2014 (during

which no judge was transferred to another court).36 To better control other

facets of the cases and to isolate the effects of the experience of judges and

attorneys, only cases involving pain and suffering damages for personal injury

incurred in caraccidentsare included. Inaddition,only cases inwhichat least

one party has hired at least one attorney‐at‐law are included in the database.

Court decisions rendered in a previous period (T‐1) were used to predict the

currentperiod(T); thus,painandsufferingdamagescasesbetweenMay1and

September 4, 2013 were also coded. Both district court and high court cases

duringthis16‐monthresearchperiodwerecoded.Thesameselectioncriterion

applies.Theattorneys’andjudges’experienceasofJune30,2014willbeusedas

theirexperiencethroughoutthemainresearchperiodwithoutadjustment.

Intotal,wecoded484civildistrictcourtcases(producing546observations)

36 Everyyear,judgetransfersinTaiwantakeplaceinonepre‐specifieddateinthefirstweekofSeptember.

05

1015

Per

cent

age

10 100 1,000 10,000Number of civil cases judges have rendered (log10)

Chang, Chen & Lin

31

andallthecriminalcasesagainstthedefendantsofthosecivilcases.Amongthem,

387 observations were 1) rendered in the main research period; 2) without

missing information in key variables; 3)not extremeoutliers (weonly exclude

foursuchcases).Theseobservationswereusedinthestructuralequationmodel

toteaseouttheeffectof juristexperience.164appellantcourtcaseshavebeen

coded in the sameway aswell.Major variables used in regressionmodels are

summarizedbelowinTable2,andthedistributionofcourt‐adjudicatedpainand

sufferingdamagesandplaintiff‐claimedpainandsufferingdamagesareshownin

Figure5andFigure6.

Thisarticlefocusesondistrictcourtcasesratherthanappellatecourtcases

forthefollowingreasons:first,almostallcurrentstudiesfocusontheappellate

orsupremecourts.Thenumberofcasesinappellatecourtsislimited;asaresult,

researchers have to include caseswith different factpatterns to gain sufficient

degreesof freedom.Theadvantageof this approach is to includeplaintiffs and

defendantswithandwithoutlegalresources(thehavesandthehavenots).The

downsideof this approach is that the effect of attorneys’ experience cannotbe

isolatedifthecontrolsforthenatureofthecasesareimperfect.

Second,amajorityofcases (about60%of the tortscases inTaiwan)were

notappealed.Researchesonappellatecasesmaysufferfromtheselectionbias.37

Third, to control and compare the effect of judges’ experience, examining

district court judges is preferable, as appellate judgesmay be similarly senior,

whiledistrictcourtjudgesaremorediverseintermsofexperience.Inaddition,

underTaiwan’shierarchicaljudicialsystem,districtcourtsaremorelikelystaffed

with inexperienced judges (in the absolute sense), who may suffer from the

anchoringbias(Chang,Chen,andLin2016)andothertypesofbiases.Similarly,

attorneysrepresentingappellatecases,particularlysupremecourtcases,would

not be novices. Had this project chosen to study appellate decisions, the

variancesofjudges’andattorneys’experiencewouldnotbelarge.

37 Adjudicatedcasesarebiasedinthesensethatmostdisputesaresettled.Nonetheless,weareinterestedinstudyingjudicialbehaviorsandtheinfluenceofattorneysonjudges.Thus,lackinginformationonsettleddisputeswillnotbiasourresults.

Chang, Chen & Lin

32

Figure5Thedistributionofcourt‐adjudicatedpainandsufferingdamages

N=387.DamagesinNewTaiwanDollars.USDollars:NewTaiwanDollars=1:30.

Figure6Thedistributionofplaintiff‐claimedpainandsufferingdamages

N=387.DamagesinNewTaiwanDollars.USDollars:NewTaiwanDollars=1:30.

05

1015

Per

cent

age

10,000 100,000 1,000,000 10,000,000Court-adjudicated pain and suffering damages (log10)

05

1015

20P

erce

ntag

e

10,000 100,000 1,000,000 10,000,000Plaintiff's claimed pain and suffering damages (log10)

Chang, Chen & Lin

33

Table2SummaryStatisticsofVariables

ContinuousVariables Obs Mean Std.Dev. Min Max

intra‐courtdeviation 387 ‐0.006 0.8 ‐2.3 2.4

inter‐courtdeviation,1to12weeks

beforeverdict=A 387 0.4 0.9 ‐1.9 3.2

Inter‐courtdifference,5–16weeks

beforeverdict=B 387 0.5 0.8 ‐1.7 3.4

ThedifferencebetweenAandB 387 ‐0.003 0.3 ‐1.2 1.7

EstimatedPSdamagesinHighCourt,

5–16weeksbeforeverdict 387 12.5 0.9 9.4 15.1

EstimatedPSdamagesinDistrict

Court,5–16weeksbeforeverdict 387 12.9 1.2 9.3 16.5

Deviationofplaintiff ’sclaimfrom

districtcourtcases5–16weeks

beforeverdict

387 1.2 0.9 ‐1.9 4.2

Court‐adjudicatedPSdamages 387 444,786 519,466 10,000 3,000,000

PSdamagesclaimedbyplaintiffs 387 1,252,810 1,310,391 10,000 10,800,000

Judge’scivilexperience 387 760 500 55 2,857

Plaintiffattorney’scivilexperience 387 290 250 0 1,539

Defendantattorney’scivil

experience 387 170 245 0 1,813

Court‐adjudicatedmedical‐related

expenses 387 1,104,264 2,912,387 0 22,700,000

Medical‐relatedexpensesincurred

byplaintiffs 387 1,749,726 4,819,668 0 46,200,000

Defendant’snumberofmonthsjailed 387 3.1 2.7 0.0 24.0

plaintiff ’sage 387 27.1 25.0 0.0 88.0

plaintiff ’sincome 387 216,640 397,650 0 2,600,000

Deductionappliedtodamages 387 251,382 512,669 0 2,200,000

Plaintiff ’scomparativenegligence 387 0.2 0.2 0.0 0.8

Note:Allcontinuousvariablesexceptplaintiff ’scomparativenegligencearein

naturallog.

Chang, Chen & Lin

34

CategoricalVariables Obs =1 %

=1ifplaintiffhiresattorney(s) 387 324 83.7

=1ifdefendanthiresattorney(s) 387 201 51.9

=1ifplaintiffhastopayfilingfee 387 87 22.5

Levelofinjury(7levels) 387 100.0

2.Temporaryinsignificant 2 0.5

3.Temporaryminor 140 36.2

4.Temporarymajor 102 26.4

5.Permanentminor 74 19.1

6.Permanentsignificant 15 3.9

7.Permanentmajor 32 8.3

8.Permanentgrave 22 5.7

=1ifdefendantsincludeacorporationwhichis

vicariouslyliable

387 9223.8

=1ifdefendantdroveundertheinfluenceof

alcohol

387 2727.0

=1ifdefendanthitandran 387 12 3.1

=1ifplaintiff ’sageismissing 387 132 34.1

=1ifplaintiff ’sincomeismissing 387 119 30.8

=1ifjudgestartscareerbefore2000 387 196 50.7

=1ifplaintiffattorneystartscareerbefore2000 387 165 42.6

=1ifdefendantattorneystartscareerbefore

2000

387 96 24.8

=1ifcourt‐adjudicatedmedicalexpenseis0 387 4 1.0

=1ifplaintiff‐incurredmedicalexpenseis0 387 1 0.3

Note:DamagesinNewTaiwanDollars.USDollars:NewTaiwanDollars=1:30.

VI. FINDINGSANDIMPLICATIONS

A. JudgesPayCloseAttentiontoHighCourtDecisions

Thestructuralequationmodelshowsthatdistrictcourtjudgesfollowedthe

historictrendsetbyhighcourts.Thevariableoninter‐courtdeviationinthefirst

equationinTable3hasapositivesignandisstatisticallysignificant(p=0.002).

Thissuggeststhatwhen,inthethreemonthspriortothedistrictcourtdecisions,

highcourtshaveincreasedtheamountofpainandsufferingdamagesrelativeto

districtcourts,districtcourtjudgesinthecurrentperiodtendedtoincreasethe

Chang, Chen & Lin

35

awardedamount.Followingtheassessmentpatternssetbyhighcourtscould

reducetheprobabilityofreversalandinthelongrunincreasejudges’chancesof

beingpromotedtohighcourts.Thisisthefirstempiricalfindingthatlowercourt

judgesmimictheweightsgiventofactsbyhighercourtjudges.

Toensurethatthisfindingisnotspurious,weranandpassedaseriesof

placebotests.Wefoundthatchangingtrendinthepastaffectedjudicialdecisions,

butchangingtrendinthefutureshouldnotinanywayaffectjudicialdecisions.

Thefollowingvariablesarecreated:changingtrendinthenextthreemonths

(highcourtestimateattimeT+1minusdistrictcourtestimateattimeT+1),

changingtrendacrosstimeandcourt(highcourtestimateattimeT+1minus

districtcourtestimateattimeT‐1),andchangingtrendacrosstime(district

courtestimateattimeT+1minusdistrictcourtestimateattimeT‐1).These

variableswerethenusedtoreplaceorsupplementthechangingtrendvariablein

equation(1)inthestructuralequationmodel.Noneofthesevariablesare

statisticallysignificantatthe10%level.

Thejudges’ownexperiencedoesnotaffectintra‐courtdeviation.38

UnreportedSEMmodelsshowthatinteractiontermsofjudgeexperienceand

otherfactorsdonotyieldstatisticallysignificantresults.Thissuggeststhat

seniorjudgesindistrictcourtswerenotmoreorlessinclinedtodeviatefrom

historictrendsetbyhighcourtsthanjuniorjudges.Theinteractionterms’lackof

statisticalsignificancecouldbeinterpretedinthefollowingway:districtcourt

judgesofallexperiencehaveequallypreparedtofollowhighcourts.39 Bydoing

so,districtcourtjudgesexpecttofacelowerreversalratesandthusincreasing

theirchancesofbeingpromoted.

38 Inunreportedmodels,weusedthenumberofyearssinceajudgerendersherfirstcivil(orany)casetoreplaceorcomplementthenumber‐of‐civil‐casemeasureofexperience.Thesenewmeasuresarenotstatisticallysignificant. 39 Inunreportedmodels,wedividedthesamplesinto2equal‐sizegroupsaccordingtotheexperienceofjudges.Thesamestructuralmodelreportedinthetextwasrunonthe2groupsseparately.TheresultsareconsistentwithChang,Chen,andLin(2016)—experiencedjudgeswerenotsubjecttotheanchoringeffect,whereasinexperiencedjudgeswere.

Chang, Chen & Lin

36

Table3SEMresults

Dependentvariableintra‐courtdeviation

(ln)

Deviationofplaintiff’s

claimfromdistrictcourt

cases5–16weeksbefore

verdict(ln)

Coef. Std.Err Coef. Std.Err

Plaintiff‐claimdeviation(ln) 0.406 *** (0.090)

Inter‐courtdeviation,1–12weeksbeforeverdict(ln) 0.133 ** (0.043)

EstimatedPSdamagesindistrictcourts5–16weeks

beforeverdict(ln)minusestimatedPSdamagesin

districtcourts1–12weeksbeforeverdict(ln)

0.682 *** (0.113)

Judge’scivilexperience(ln) 0.084 (0.053)

=1ifjudgestartedbenchcareerbefore2000 ‐0.064 (0.082)

=1ifdefendanthiresattorney(s) 0.248 (0.297)

Defendantattorney’scivilexperience(ln) ‐0.040 (0.055)

=1ifdefenseattorneystartedpracticingbefore2000 0.114 (0.102)

Court‐adjudicatedmedical‐relatedexpenses(ln) 0.033 (0.023)

=1ifadjudicatedmedicalexpenseis0 0.95 * (0.410)

Defendant’smaxincome(ln) 0.005 (0.007)

=1ifdefendant’smaxismissing 0.056 (0.108)

Deductionappliedtodamages(ln) 0.016 ** (0.006)

Plaintiff’scomparativenegligence 0.099 (0.136)

Inter‐courtdeviation,5–16weeksbeforeverdict 0.142* (0.058)

Plaintiff‐incurredmedicalexpenses ‐0.076* (0.030)

=1ifincurredmedicalexpenseis0 1.462 (0.939)

=1ifplaintiffhastopayfilingfee ‐0.372** (0.136)

=1ifplaintiffhiresattorney(s) ‐0.053 (0.285) ‐0.652+ (0.371)

Plaintiffattorney’scivilexperience(ln) 0.001 (0.051) 0.128+ (0.068)

=1ifplaintiffattorneystartedpracticingbefore2000 ‐0.165 * (0.081) ‐0.086 (0.110)

6injury‐leveldummies(baseline:2.Temporary

insignificant)

3.Temporaryminor 0.876 + (0.473) ‐1.056+ (0.619)

4.Temporarymajor 1.213 * (0.473) ‐0.798 (0.621)

5.Permanentminor 1.214 * (0.478) ‐1.004 (0.624)

6.Permanentsignificant 1.136 * (0.507) ‐1.106+ (0.664)

Chang, Chen & Lin

37

7.Permanentmajor 1.173 * (0.497) ‐1.209+ (0.642)

8.Permanentgrave 0.999 + (0.513) ‐1.186+ (0.656)

Defendant’slengthofdeclaredincarceration(ln) ‐0.013 (0.058) ‐0.234* (0.098)

=1ifdefendantsincludeacorporationwhichis

vicariouslyliable0.027 (0.079) ‐0.036 (0.108)

=1ifdefendantdroveundertheinfluenceofalcohol ‐0.035 (0.127) 0.186 (0.173)

=1ifdefendanthitandran 0.190 (0.212) ‐0.043 (0.283)

plaintiff’sage(ln) ‐0.055 (0.083) ‐0.040 (0.111)

=1ifplaintiff’sageismissing ‐0.079 (0.310) ‐0.120 (0.415)

plaintiff’sincome(ln) 0.010 (0.007) ‐0.008 (0.010)

=1ifplaintiff’sincomeismissing 0.147 (0.100) ‐0.056 (0.128)

18courtdummies Yes Yes

Constant ‐2.725 *** (0.756) 3.963*** (0.852)

σ1 0.520 *** (0.028)

σ2 0.824*** (0.030)

ρ 0.063 (0.128)

Robuststandarderrorsinparentheses.Clusteredbycases.Judgerandomeffects

apply.PSdamages=painandsufferingdamages.

***p<0.001,**p<0.01,*p<0.05,+p<0.1.

Loglikelihood=‐791.25

B. AttorneyInfluenceonPlaintiffClaim

Whetherplaintiffshire(experienced)attorneyissignificantlyassociated

withhowplaintiffsclaim.Table3revealsthatplaintiffsrepresentedbyattorneys

withlittleexperienceclaimedlessthanthoseself‐representedorrepresentedby

non‐attorneys(p=0.079).Thisisintuitive,asproseplaintiffsmaybe

optimisticallybiased,whilejuniorattorneysknowbetterbutclaimconservatively.

Plaintiffs’claims,however,increasewiththeexperienceoftheirattorneys

(p=0.059).Thisisconsistentwiththeeconomicandpsychologicalaccountslaid

outinPartIII.B.2.

Asarguedabove,(senior)attorneyrepresentationisclosetorandom

assignments.TableB.1intheAppendixfurthershowsthatplaintiffincomesdo

notaffectwhetherplaintiffsseekattorneyrepresentation.Themajordriverof

Chang, Chen & Lin

38

plaintiffs’decisionstoretainattorneysiswhetherdefendantsincludea

corporation(p<0.01).Perhapsplaintiffshireattorneyswhentheyidentifyadeep

pocket.40

Allthefouridentifyingvariablesinequation(2)arestatistically

significant.41 Thedummyvariableonwhetherplaintiffspaidfilingfeesisworthy

ofmorediscussion.Thisvariablehastheexpectednegativesign,aseconomically

rationalplaintiffswhodonothavetopayproratafilingfeeswillover‐claimas

comparedtothosewhohavetopay.Assaidabove,thejudgesweinterviewed

sensethatourconjectureholdswater,andtheattorneysweinterviewed

admittedthattheytendtodoexactlyaswehypothesize.Thus,thisresultshould

notbesurprising.

Wealsofindthatinter‐courtdeviationispositivelyassociatedwiththe

deviationofplaintiffs’claimedamount(p=0.015),suggestingthatplaintiffsin

generalpaidcloseattentiontothechangingtrendinhighcourtsvis‐à‐visdistrict

courts,too.

C. PlaintiffAttorneyInfluenceonAdjudication

Experiencedplaintiffattorneysingeneralwinhigherpainandsuffering

damagesfortheirclients.Theplaintiffs’claimedamountscreatedtheanchoring

effect.42 Whenplaintiffs’claimsarehigherthanthepointestimatesbasedon

casesofthepreviousthreemonths,districtcourtjudgestendtoawardhigher

40 Duetotechnicalconstraints,wecannotuseplaintiffattorneyexperiences(orrepresentation)anddefendantattorneyexperiences(representation)asdependentvariablesinadditionalequationsinthestructuralequationmodel.

Whenweaddedoneortwomoreequationstothereportedtwo‐equationstructuralmodel,themodeldidnotconverge.Wehavetriedexcludingcaseswhereplaintiffswerenotrepresentedbyattorneystofocussolelyontheeffectofplaintiffattorneyexperience.Puttingthosecasesintothereportedmodelrevealsalargelysimilarresult.Outmainstoryholds,thoughthep‐valuesdoshiftabit,sometimescrossingthethresholdofstatisticalsignificance.Wehavealsoputthosecasesintoathree‐equationstructuralmodel—thereportedmodelplusoneequationtoaccountforthedefendants’decisiontoretainattorneys.AgainthemainstoryreportedinthetextandAppendixB.2remainsintact.41 Thefourvariables,whenaddedintothefirstequationtotestwhethertheyarevalidinstrumentalvariables,arejointlyinsignificantinthefirstequation.Andthreeofthefourvariablesarestatisticallyinsignificantatthe10%level.Bycontrast,thefourinstrumentalvariablesarejointlysignificantinthesecondequation(p<0.001). 42 ThefindingoftheanchoringeffectisconsistentwithourpriorresearchonpropertydisputesinTaiwan(Chang,Chen,andLin2016)andexperimentalresultsconductedbyothers(ChapmanandBornstein1996;Campbelletal.2014).

Chang, Chen & Lin

39

painandsufferingdamages(p<0.001).Asdiscussedabove,plaintiffs’claimswere

furtheraffectedbyattorneyrepresentationandattorneyexperience.Hence,

throughtheanchoringeffects,plaintiffswhohireseniorattorneyswereableto

garnerhigherpainandsufferingdamages.Forexample,anincreaseofplaintiff

attorneyexperienceby1%(3.47cases)wouldonaverageleadtoanincreaseof

0.128%($53)inclaim,whichfurtherresultsanincreaseof0.05%($21)inpain

andsufferingdamagesawards.43

Thisfinding,however,doesnotimplythatplaintiffswhoseekhigherpain

andsufferingdamageshavetoretainseniorattorneystoachievethisgoal.As

longasplaintiffscanidentifythehistoricaltrendsinhighcourtsanddistrict

courtsandarewillingtogamblebyover‐claiming,theymaybeawardedwith

higherpainandsufferingdamagesaswell.AsTable3suggests,plaintiffsnot

representedbyattorneystendtoclaimhigherthanthoserepresentedby

attorneys(p=0.079).Theseunassistedplaintiffsreceivedhigherpainand

sufferingdamagesthroughtheanchoringeffects.Moreover,seniorattorneys

tendedtobedistrustedbyjudgeswhentheyover‐claim.Thedummyvariable

thatequals1whenaplaintiffattorneypassedthebarandstartedpracticebefore

year2000(whoseexperienceunder‐estimated)hasanegativecoefficientandis

statisticallysignificant(p=0.042).Thisisevidencethatjudgesmightbe

suspiciousofseniorattorneys’claims.

D. DefendantAttorneyHasNoRole

AsTable3shows,whetheradefendanthiresattorneysanddefendant

attorneyexperience44 donotaffecttheextenttowhichdistrictcourtsdeviate

fromthehistorictrend.Perhapsdefendantattorneysworkondismissingthe

casealtogetherorarguingthatplaintiffsarecomparativelynegligent.Ourdatado

notallowustotestthesehypotheses.

Defendants’decisionstoretainattorneysaremainlyinfluencedbythe 43 Thesenumbersarecalculatedbasedonthesampleaveragesofplaintiffattorneyexperience(347)andplaintiff 'sclaimofpainandsufferingdamages($41,760).0.128%×0.405=0.05%.Theexchangerateusedis$1=NTD30.44 Inunreportedmodels,weusedthenumberofyearssinceajudgerendersherfirstcivil(orany)caseorthenumberofyearssinceadmittedtothebartoreplaceorcomplementthenumber‐of‐civil‐casemeasureofexperience.Thesenewmeasuresarenotstatisticallysignificant.

Chang, Chen & Lin

40

compensationrisktheywereexposedto,asTableB.2intheAppendixshows.

Morespecifically,naturallogoftheamountoftotalclaimsmadebyplaintiffsis

positivelyandstatisticallysignificantlyassociatedwiththebinarydecisionof

defendants’hiringattorneys(p<0.01).Thisresultissensible,asdefendantscare

moreabouthowmuchtheywouldhavetopaythanthecategoriesofthe

damages.Whendefendantsaremoreexposedtorisksofhighdamages,theyare

inclinedtoseeklegalrepresentation.Besides,again,defendantincomesdonot

havestatisticallysignificantrelationwithdefendantattorneyrepresentation.

VII. CONCLUSION

Theinnovationandcontributionofthisarticleisusingthemostaccurate

measureofattorneys’andjudges’experiencetodateinnewandwell‐specified

regressionmodelstodirectlyexaminetheeffectofjurists’experience,whichhas

notbeenfullyexplored.Thestructuralequationmodelrevealsthatin

adjudicatingpainandsufferingdamages,districtcourtjudgesdeviatedfromthe

historictrendinordertofollowtherecentchangesinassessmentpatternsin

highcourts.Seniorplaintiffattorneyswintheirclientsmorepainandsuffering

damagesbyclaiminghigheramountsofdamages,whichcreatetheanchoring

effect.Judgeexperienceanddefendantattorneyexperiencedonotaffectwhether

districtcourtjudgesdeviatefromthehistorictrend.Ourfindingsofjudicialand

attorneys’behaviorsshouldbegeneralizabletootherjurisdictionswithsimilar

institutionalsettings.

Plaintiffs’addamnumcreatesananchoringeffect,andseniorplaintiff

attorneyshaveexploitedthisjudicialbiasintippingthecasetowardtheirclients’

favor—wealsofindevidencethatjudgesmightbeawareofthehigherclaims

madebyseniorattorneys.Policymakersinjurisdictionswhereaddamnumisa

pre‐requisiteshouldseriouslyconsiderreformproposalsthatcouldreducethe

unduebiascreatingbyplaintiffs’claims.

Wehavedemonstratedtheaddedvalueofplaintiffattorneys’experience.Yet,

thismaynotbethewholereasonforthehigherfeechargedbyseasoned

attorneys.Thisstudyexaminestheimpactofclaimformulationbyplaintiff

attorneys,whiletheycanchangelitigationoutcomesinotherwaysaswell.45

Therearestillalotofempiricalstudiesonattorneyexperiencetobedone!

45 Fordiscussionsonhowattorneyscanchangelitigationoutcomesviaotherways,seegenerally,forexample,Galanter(1974),Kritzer(1998),Aldisert(1992),andHaire,Lindquist,andHartley(1999).

Chang, Chen & Lin

41

Chang, Chen & Lin

42

APPENDIXA:PREDICTINGCOURT‐ADJUDICATEDPAINANDSUFFERINGDAMAGES

TableA.1HedonicRegressionModelsonAdjudicatedPainandSuffering

DamagesinDistrictCourts

DependentvariableCourt‐Adjudicatedpain

andsufferingdamages

Coef. Std.Err

Lnofcourt‐adjudicatedmedicalexpenses 0.284 *** (0.023)

=1ifcourt‐adjudicatedmedicalexpense=0 1.945 *** (0.390)

injury‐leveldummy(baseline:minorinjury) 0.651 *** (0.090)

Defendant’snumberofmonthsjailed(ln) 0.192 *** (0.050)

plaintiff’sage(ln) ‐0.140 + (0.082)

=1ifplaintiff’sageismissing ‐0.768 * (0.315)

plaintiff’sincome(ln) 0.001 (0.007)

=1ifplaintiff’sincomeismissing 0.047 (0.091)

defendant’sincome(ln) 0.010 (0.008)

=1ifdefendant’sincomeismissing 0.104 (0.105)

=1ifdefendantdroveundertheinfluenceofalcohol 0.158 (0.104)

=1ifdefendanthitandran ‐0.432 + (0.240)

=1ifdefendantsincludeacorporationwhichis

vicariouslyliable0.131 + (0.071)

Constant 9.179 *** (0.414)

N=520

R2=0.628

Robuststandarderrorsinparentheses.Clusteredbycases.

***p<0.001,**p<0.01,*p<0.05,+p<0.1

Chang, Chen & Lin

43

TableA.2HedonicRegressionModelsonAdjudicatedPainandSuffering

DamagesinHighCourts

DependentvariableCourt‐Adjudicatedpain

andsufferingdamages

Coef. Std.Err

Lnofcourt‐adjudicatedmedicalexpenses 0.341 *** (0.040)

=1ifcourt‐adjudicatedmedicalexpense=0 3.107 *** (0.843)

=1ifseriousinjury(baseline:minorinjury) 0.429 ** (0.151)

Defendant’snumberofmonthsjailed(ln) 0.064 (0.125)

plaintiff’sage(ln) ‐0.167 (0.162)

=1ifplaintiff’sageismissing ‐1.022 + (0.583)

plaintiff’sincome(ln) 0.002 (0.013)

=1ifplaintiff’sincomeismissing ‐0.055 (0.175)

defendant’sincome(ln) 0.014 (0.014)

=1ifdefendant’sincomeismissing ‐0.085 (0.200)

=1ifdefendantdroveundertheinfluenceofalcohol 0.253 (0.285)

=1ifdefendanthitandran ‐1.117 *** (0.248)

=1ifdefendantsincludeacorporationwhichis

vicariouslyliable‐0.013 (0.149)

Constant 9.132 *** (0.655)

N=164

R2=0.596

Robuststandarderrorsinparentheses.Clusteredbycases.

***p<0.001,**p<0.01,*p<0.05,+p<0.1

Chang, Chen & Lin

44

APPENDIXB:ATTORNEYHIRINGDECISIONS

TableB.1ProbitRegressionModelsonPlaintiffs’HiringAttorney(s)

Dependentvariable=1ifplaintiffhires

attorney(s)

Coef. Std.Err

plaintiff’sincome(ln) 0.021 (0.017)

=1ifplaintiff’sincomeismissing 0.311 (0.230)

plaintiff’sage(ln) ‐0.337 (0.226)

=1ifplaintiff’sageismissing ‐1.282 (0.844)

7injury‐leveldummies(baseline:2.Temporary

insignificant)

3.Temporaryminor ‐0.897 + (0.541)

4.Temporarymajor ‐1.053 + (0.548)

5.Permanentminor ‐0.847 (0.562)

6.Permanentsignificant ‐1.419 * (0.661)

7.Permanentmajor ‐0.980 (0.600)

=1ifdefendantsincludeacorporationwhichis

vicariouslyliable0.672 ** (0.230)

=1ifdefendantdroveundertheinfluenceofalcohol 0.160 (0.358)

=1ifdefendanthitandran ‐0.431 (0.429)

=1ifplaintiffisamale 0.097 (0.177)

18courtdummies Yes

Constant 2.972 ** (1.048)

N=379

PseudoR2=0.092

Robuststandarderrorsinparentheses.Clusteredbycases.PSdamages=pain

andsufferingdamages.

***p<0.001,**p<0.01,*p<0.05,+p<0.1.

Logpseudolikelihood=‐155.2775

Chang, Chen & Lin

45

TableB.2ProbitRegressionModelsonDefendants’HiringAttorney(s)

Dependentvariable=1ifdefendanthires

attorney(s)

Coef. Std.Err

Plaintiffattorney’scivilexperience(ln) ‐0.022 (0.097)

Plaintiff ’stotalclaim 0.286 ** (0.108)

Defendants’maximumincome(ln) ‐0.015 (0.019)

=1ifalldefendants’incomearemissing ‐0.334 (0.253)

7injury‐leveldummies(baseline:2.Temporary

insignificant)

3.Temporaryminor ‐0.164 (0.840)

4.Temporarymajor ‐0.165 (0.836)

5.Permanentminor 0.046 (0.843)

6.Permanentsignificant ‐0.054 (0.915)

7.Permanentmajor 0.117 (0.882)

8.Permanentgrave ‐0.114 (0.903)

=1ifdefendantsincludeacorporationwhichis

vicariouslyliable0.309 (0.191)

=1ifdefendantdroveundertheinfluenceofalcohol 0.198 (0.277)

=1ifdefendanthitandran ‐0.619 (0.501)

18courtdummies Yes

Constant ‐3.966 * (1.784)

N=318

PseudoR2=0.124

Robuststandarderrorsinparentheses.Clusteredbycases.PSdamages=pain

andsufferingdamages.Casesinwhichplaintiffsdonothireattorneysare

excludedinthisProbitregressionbecauseinoursample,defendantsalwayshire

attorneysinthosecases;thus,thereisnovariance.

***p<0.001,**p<0.01,*p<0.05,+p<0.1.

Logpseudolikelihood=‐190.1241

Chang, Chen & Lin

46

References

Abrams,DavidS.,andAlbertH.Yoon.2007.TheLuckoftheDraw:UsingRandom

CaseAssignmenttoInvestigateAttorneyAbility.TheUniversityofChicago

LawReview74(4):1145–1177.

Aldisert,RuggeroJ.1992.WinningonAppeal:BetterBriefsandOralArgument.

Boulder,CO:NatationalInstituteforTrialAdvocacy.

Anderson,JamesM.,andPaulHeaton.2012.HowMuchDifferenceDoesthe

LawyerMake:TheEffectofDefenseCounselonMurderCaseOutcomes.

TheYalelawJournal122:154–217.

Boyd,ChristinaL.2015a.TheHierarchicalInfluenceofCourtsofAppealson

DistrictCourts.TheJournalofLegalStudies44(1):113–141.

———.2015b.LitigantStatusandTrialCourtAppealMobilization.Law&Policy

37(4):294–323.

Cameron,A.Colin,andPravinK.Trivedi.2005.Microeconometrics:Methodsand

Applications.NewYork,NY:CambridgeUniversityPress.

Campbell,John,BernardChao,ChristopherRobertson,andDavidYokum.2014.

CounteringthePlaintiff 'sAnchor:JurySimulationtoEvaluateDamages

Arguments.workingpaper,http://ssrn.com/abstract=2470066.

Chang,Yun‐chien,Kong‐PinChen,andChang‐chingLin.2016.AnchoringEffectin

RealLitigations.

Chang,Yun‐chien,TheodoreEisenberg,Han‐weiHo,andMartinT.Wells.2015.

PainandSufferingDamagesinWrongfulDeathCases:AnEmpiricalStudy.

JournalofEmpiricalLegalStudies12(1):128–160.

Chang,Yun‐chien,TheodoreEisenberg,TsungHsienLi,andMartinT.Wells.2014.

PainandSufferingDamagesinPersonalInjuryCases:AnEmpiricalStudy.

workingpaper.

Chang,Yun‐chien,Han‐weiHo,andJimmyChia‐ShinHsu.2016.Non‐pecuniary

DamagesforDefamation,PersonalInjury,andWrongfulDeath:An

EmpiricalAnalysisofCourtCasesinTaiwan.ChineseJournalof

ComparativeLaw4(1).

Chapman,GretchenB.,andBrianH.Bornstein.1996.TheMoreYouAskfor,the

MoreYouGet:AnchoringinPersonalInjuryVerdicts.AppliedCognitive

Psychology10:519–540.

Chen,Kong‐Pin,Kuo‐ChangHuang,andChang‐ChingLin.2015.PartyCapabilityversusCourtPreference:Whydothe“Haves”ComeOutAhead?－An

EmpiricalLessonfromtheTaiwanSupremeCourt.JournalofLaw,

Economics&Organization31:93–126.

Choi,StephenJ.,MituGulati,andEricA.Posner.2012.WhatDoFederalDistrict

Chang, Chen & Lin

47

JudgesWant?AnAnalysisofPublications,Citations,andReversals.Journal

ofLaw,Economics,andOrganization28(3):518‐549.

Davidson,Russell,andJamesG.MacKinnon.1993.EstimationandInferencein

Econometrics.NewYork,NY:OxfordUniversityPress.

———.2004.EconometricTheoryandMethods.NewYork,NY:OxfordUniversity

Press.

Eisenberg,Theodore,andKuo‐ChangHuang.2012.TheEffectofRulesShifting

SupremeCourtJurisdictionfromMandatorytoDiscretionary—An

EmpiricalLessonfromTaiwan.InternationalReviewofLaw&Economics

32(1):3–18.

Epstein,Lee,WilliamM.Landes,andRichardA.Posner.2013.TheBehaviorof

FederalJudges:ATheoreticalandEmpiricalStudyofRationalChoice

Cambrudge,MA:HarvardUniversityPress.

Feldman,Adam.2015.WhoWinsintheSupremeCourt?AnExaminationof

AttorneyandLawFirmInfluence.workingpaper.

Galanter,Marc.1974.Whythe"Haves"ComeOutAhead:Speculationsonthe

LimitsofLegalChange.LawandSocietyReview9:95–160.

Goodman‐Delahunty,Jane,PärAndersGranhag,MariaHartwig,andElizabethF.

Loftus.2010.InsightfulorWishful:Lawyers’AbilitytoPredictCase

Outcomes.Psychology,PublicPolicy,andLaw16(2):133–157.

Greene,WilliamH.2003.EconometricsAnalysis.5ed.UpperSaddleRiver,NJ:

Prentice‐Hall.

Greiner,James,andCassandraWolosPattanayak.2011.RandomizedEvaluation

inLegalAssistance:WhatDifferenceDoesRepresentation(Offerand

ActualUse)Make.YaleLawJournal121:2118‐2215.

Haire,SusanBrodie,StefanieA.Lindquist,andRogerHartley.1999.Attorney

Expertise,LitigantSuccess,andJudicialDecisionmakingintheUSCourts

ofAppeals.LawandSocietyReview33:667–685.

Harris,CatherineT.,RalphPeeples,andThomasB.Metzloff.2005.WhoAre

ThoseGuys:AnEmpiricalExaminationofMedicalMalpracticePlaintiffs'

Attorneys.SMULawReview58:225‐250.

———.2008.DoesBeingaRepeatPlayerMakeaDifference—TheImpactof

AttorneyExperienceandCase‐PickingontheOutcomeofMedical

MalpracticeLawsuits.YaleJournalofHealthPolicyLaw&Ethics8:253–

282.

Huang,Kuo‐Chang.2008.HowLegalRepresentationAffectsCaseOutcome:An

EmpiricalPerspectivefromTaiwan.JournalofEmpiricalLegalStudies5

(2):197–238.

Chang, Chen & Lin

48

———.2009.DoesDiscoveryPromoteSettlement?AnEmpiricalAnswer.Journal

ofEmpiricalLegalStudies6(2):241–278.

———.2016.TheEffectofStakesonSettlement:AnEmpiricalLessonfrom

Taiwan.InResearchHandbookonComparativeLawandEconomics,edited

byTheodoreEisenbergandGiovanniBattistaRamello.Northampton,

Mass.:EdwardElgar.

Huang,Kuo‐Chang,Kong‐PinChen,andChang‐ChingLin.2010.DoestheTypeof

CriminalDefenseCounselAffectCaseOutcomes?:ANaturalExperimentin

Taiwan.InternationalReviewofLawandEconomics30(2):113–127.

Huang,Kuo‐Chang,Chang‐ChingLin,andKong‐PinChen.2014.DoRichandPoor

BehaveSimilarlyinSeekingLegalAdvice?LessonsfromTaiwanin

ComparativePerspective.Law&SocietyReview48(1):193–223.

Hyman,DavidA.,MohammadRamhmati,BernardBlack,andCharlesSilver.2015.

Plaintiff‐SideRepresentationinMedicalMalpractice,Part1:Market

StructureandtheWagesofRisk.workingpaper.

Krishnan,C.N.V.,SolomonStevenDavidoff,andRandallS.Thomas.2014.

ZealousAdvocatesorSelf‐InterestedActors?AssessingtheValueof

Plaintiffs’LawFirmsinMergerLitigation.SSRN:

http://ssrn.com/abstract=2490098

Kritzer,HerbertM.1998.LegalAdvocacy:LawyersandNonlawyersatWork.Ann

Arbor,MI:UniversityofMichiganPress.

———.2004.Risks,ReputationsandRewards:ContingencyFeeLegalPracticein

the

UnitedStates.PaloAlto,CA:StanfordUniversityPress.

Lederman,Leandra,andWarrenB.Hrung.2006.DoAttorneysDoTheirClients

Justice?AnEmpiricalStudyofLawyers'EffectsonTaxCourtLitigation

Outcomes.WakeForestLawReview41:1235–1295.

Posner,RichardA.2008.HowJudgesThink.Cambridge,MA:HarvardUniversity

Press.

Roodman,David.2011.FittingFullyObservedRecursiveMixed‐ProcessModels

withCMP.TheStataJournal11(2):159–206.

Sloan,FrankA.1993.SuingforMedicalMalpractice.Chicago,IL:Universityof

ChicagoPress.

Vidmar,Neil,FeliciaGross,andMaryR.Rose.1998.JuryAwardsforMedical

MalpracticeandPost‐VerdictAdjustmentsofThoseAwards.DePaulLaw

Review48:265–299.

Wooldridge,JeffreyM.2010.EconometricAnalysisofCrossSectionandPanelData.

2ed.Cambridge,MA:theMITPress.

Chang, Chen & Lin

49

Zamir,Eyal,andIlanaRitov.2010.RevisitingtheDebateOverAttorneys’

ContingentFees:ABehavioralAnalysis.TheJournalofLegalStudies39

(1):245–288.