1

OpiniononCongestionRevenueRightsAuctionEfficiency

by

JamesBushnell,MemberScottM.Harvey,MemberBenjaminF.Hobbs,Chair

MembersoftheMarketSurveillanceCommitteeoftheCaliforniaISO

Final,March15,2018

1. Introduction TheCaliforniaIndependentSystemOperator(CAISO)hasproposedmajorrevisionstoitsprocessforauctioningCongestionRevenueRights(CRRs)precedingitsauctionofannualCRRstobeheldinJuly2018.1Theproposalfollowsayear‐longexplorationofseveralcon‐cernswiththeCaliforniaCRRsystemasitiscurrentlyconstructed.TheCAISOanditsDe‐partmentofMarketMonitoring(DMM)havehighlighted,inparticular,thefactthatCRRshave,onaverage,soldatauctionpricessubstantiallybelowthevalueoftherevenuestreamsassociatedwiththem.From2009through2017,payoutstoauctionedCRRshaveexceeded$1.4billionwhileauctionrevenuesforthoseCRRswasjustover$740million,adifferenceofcloseto$700million.2This“auctionrevenueshortfall”hasbeendecliningovertime.TheCAISOreportsimilarlycalculatesthatthepayouttoannualCRRsexceededtheirauctionpriceby$48millionovertheperiodJanuary2015throughMay2017,whilethepayouttomonthlyCRRsexceededtheirauctionpricebyabout$92millionoverthesameperiod.3TheDMMandsomeload‐servingentities(LSEs),whoaretheresidualclaimantsonconges‐tionrevenuesiftheywerenotsoldatauction,havecharacterizedtheauctionsasunwillingsalesoffuturerevenuestreamsthatarefatedtobesoldbelowvalueduetofundamental

1CaliforniaISO,CongestionRevenueRightsAuctionEfficiency,Track1DraftFinalProposal,February8,2018,www.caiso.com/Documents/DraftFinalProposal‐CongestionRevenueRightsAuctionEfficiency‐Track1.pdf;CaliforniaISO,CongestionRevenueRightsAuctionEfficiency,Track1DraftFinalProposalAdden‐dum,March8,2018.www.caiso.com/Documents/DraftFinalProposalAddendum‐CongestionRevenu‐eRightsAuctionEfficiency‐Track1.pdf.

2Problemsintheperformanceanddesignofthecongestionrevenuerightauction,CAISODepartmentofMarketMonitoring,November27,2017.Thus,auctionrevenueshavebeenabouthalfofthevalueofthecongestionrevenuepayouts.If,however,expressedasapercentageoftotalcongestionrevenue,thepayouthasbeenasignificantlysmallerfraction,sinceday‐aheadcongestionrentsareontheorderof$200Mto$500M/year(from2012‐2015,accordingtoDMMstateofthemarketreports).

3CaliforniaISO,CRRAuctionAnalysisReport,November21,2017pp.51‐52.Notethe$92millionfigurecitedabovecorrectsforatypointhisreportwherenetpaymentsfor2017January–Maywerereportedas‐$11millionbutwereactually‐$19million.

2

flawsintheCRRprocess.Atthesametime,CRRshavelongbeenheldtobeuseful,ifnotcritical,instrumentsforhedgingtheriskofcongestionpricesintransmissionnetworks.Thedevelopmentoftheparadigmoffinancialtransmissionrights(orCRRs)wasafunda‐mentalstepinshiftingUSpowermarketsawayfrominefficientphysicaltransmissionrightsasameansofprovidingopenaccesstotransmissionsystems.IfsignificantvalueisplaceduponCRRsasahedgingtool,especiallybypartieswhodonotreceiveallocatedCRRs,thenmarketdesignchangesthateliminateorsubstantiallyreduceaccesstothemraisepotentialconcernsaboutmarketefficiencyandcompetitiveness.TheCAISOTrack1proposalattemptstobalancetheseconcernswiththoseofDMMandthein‐vestor‐ownedutilitieswhohavearguedthatCRRssalesarecostingtheirratepayersanav‐erageofabout$75millionperyear,althoughthisnumberislowerinrecentyears.Inaddi‐tiontochangesinoutagereportingandotherprocesschanges,theproposalwouldrestrictthetypesofCRRsavailableforauctionbylimitingthesourcesandsinksoftheCRRsthatwouldbeeligibleforsaleintheauction.TheCAISOMarketSurveillanceCommittee(MSC)hasbeenaskedbytheCAISOtoprovideanOpinionontheTrack1proposal.ThecausesofshortfallsintheCAISO’sCRRauctionsalongwithpossibleremedieshavebeenpreviousdiscussedattwoMSCmeetings.TheseincludemeetingsonFeb.3,2017,whentheMSCdiscussedpossibleanalysestounderstandthereasonsfortherevenueshortfallsandtoquantifytheusesofauctionedCRRsforhedg‐ingandtradingpurposes;4andFeb.2,2018,whentheCAISO’sTrack1proposal(thepre‐sentproposal)wasfirstpubliclydiscussed.TheMSChasalsowrittenopinionspreviouslyonCRRauctionandallocationasapartoftheMRTUdesignprocess.5

4S.Harvey,BriefingonAnalyzingCongestionRevenueRightsAuctionValuation,MSCMeeting,February3,2017,www.caiso.com/Documents/BriefingonAnalyzingCongestionRevenueRightsAuctionValuation‐MSCHar‐vey‐Feb2017.pdf

5PriortotheimplementationoftheMarketRedesignandTechnologyUpgrade(MRTU)in2007,theMSCis‐suedthreeopinionsthataddressedCRRallocation.Onefocusedonlong‐termrightsallocation,supportingtheISO’sproposalandfull‐fundingofallCRRs(F.Wolak,J.Bushnell,andB.Hobbs,"OpiniononLong‐TermCongestionRevenueRightsProposal",Jan.18,2007,www.caiso.com/Documents/FinalOpiniononLongTerm‐CongestionRevenueRights.pdf).TheotheropinionsincludedmorecomprehensivediscussionsofCRRallocation.Themostrecent(F.Wolak,J.Bushnell,andB.Hobbs,“FinalOpiniononRecentChangestotheISOCongestionRevenueRightsPro‐posal,”www.caiso.com/Documents/070418_MSCFinalOpiniononCongestionRevenueRights.pdf)emphasizedtheimportanceofgivingsmallLSEsaccesstolong‐termCRRsattradinghubs.TheOpinionalsostatedthefol‐lowingtwoprinciples“which...willlimittheextenttowhichanallocationmechanismmightharmmarketeffi‐ciency.Oneprinciple...istoavoidadirectlinkagebetweenfuturemarkettransactionsandtheassignmentoffuturerights.SuchalinkagecoulddistorttheinvestmentdecisionsofafirmbecausethecostoflocatingaplantinacongestedareawouldbeatleastpartiallyoffsetbytheallocationofCRRsfromthatlocationtothefirm’sload.Asecondprincipleisthattheallocationoftransmissionrights,ratherthananauctionofallrights,helpstoguardagainsttheownershipofCRRsenhancingtheincentiveofmarketparticipantstoexerciselocalmarketpower.Forthisreason,wehaveconsistentlysupportedallocatingCRRstoLSEsandhaverecom‐mendedagainstauctioningthemajorityofthetotalCRRsavailable”(p.3).Theconcernwaswithhowrightsallocationcouldmagnifymarketpower,apointelaborateduponinanotheropinion(F.Wolak,B.Barber,J.Bushnell,andB.Hobbs,“OpiniononAspectsoftheCaliforniaISO’sMarketRedesignandTechnologyUpgrade(MRTU)ConceptualFiling”,September30,2005,www.caiso.com/Documents/OpiniononAspects‐

3

InthisOpinion,wefirstprovidebackgroundinSection2ontheroleofCRRsinrestruc‐turedelectricitymarkets,theprocedurespresentlyusedtoallocateCRRs,andtheDepart‐mentofMarketMonitoring(DMM)/SouthernCaliforniaEdison(SCE)proposalsfortheirreform.TheninSection3,wediscussseveralidentifiedissueswithCRRsandCRRauctions,beforesummarizingtheCAISOTrack1proposalinSection4.Inthelattersection,wede‐scribeandprovidecommentsonthetwomajorfeaturesoftheproposal,whicharechangesinoutagereportingandrestrictionsupontheallowedsource‐sinkpairsintheCRRauction.Section5presentsourconclusions.

2. Background

2.1 The Historical Role of CRRs in ISO Markets

CongestionRevenueRightsawardtotheirownerarevenuestreamassociatedwiththepricedifferencebetweentwolocationsinanISOnetwork.6Theywereenvisionedasameanstoprovidethefinancialequivalentoffirmtransmissionservice,inthesensethattheyentitletheholdertouseofthetransmissionnetworkwithoutpayingcongestioncharges.7ThetransitionfromphysicaltofinancialtransmissionsrightsallowedISOstoprovidemarketparticipantsnon‐discriminatoryphysicalaccesstothegrid,whilepreserv‐ingforindividualsuppliersthefinancialequivalentoftheabilityto“deliver”powertocus‐tomers.InanorderconditionallyapprovingmarketrulesoftheNewYorkISO,theFERCconcludedthatCRRsor“TCCssignificantlyenhancetheopenaccessrequirementsoftheproformatariffasanefficientsubstituteforthereassignmentofphysicaltransmissionrightsthatentitiesobtainundertheproformatariff.”8Acoreelementoffinancialtransmissionrightsisthattheyarefinancialinstruments,notphysicaltransmissionrights.Thisdesignisessentialinordertoavoiduse‐it‐or‐lose‐it

CaliforniaISO%E2%80%99sMarketRedesignandTechnologyUpgrade_MRTU_ConceptualFiling.pdf).ThisopinionarguedforasimpleallocationprocessthatwouldgiveallCRRstoload,andagainstuseofanauctionprimarilybecauseofconcernsthatCRRscouldbeusedtomagnifygeneratormarketpower.“Themarketpar‐ticipantabletocausethemostcongestioniswillingtopaythemostforCRRsthatrefundthesecongestioncharges…thereisthedangerthatfirms,uponpurchasingorbeingawardedafinancialCRR,wouldbehavelessefficientlythaniftheydidnotownthatCRR”(p.18).ThesemarketpowerconcernsweresubsequentlymitigatedbyCAISOrulesprohibitingresourcesfromholdingcertainCRRpositionsthatwouldmagnifymar‐ketpower(e.g.,owningrightsthatsinkintoacongestedloadpocketinwhichtheholderalsoownsgenera‐tion).InnoneoftheopinionsdidtheMSCaddressthepossibilitythatauctionrevenueswouldconsistentlyfallshortofpayoutstotheauctionrights,ortheequityimplicationsifthattookplace.

6Excludinglosscomponents.

7Hogan,WilliamW."Contractnetworksforelectricpowertransmission."JournalofRegulatoryEconomics4.3(1992):211‐242.

8FERC“OrderConditionallyAcceptingTariffandMarketRules…”DocketNosER97‐1523‐000;OA97‐470‐000andER97‐4234‐000,January27,199986FERC¶61,062,footnote13,page6.TheFERC’sprimaryconcernwithTCCsinthatorderwastheabsenceofrightslongerthansixmonths.TheNYISOsubsequentlyreviseditsproposaltoincludeauctionsofboth6monthandlonger‐termTCCs.

4

incentivesthatwoulddiscourageefficientparticipationintheeconomicdispatch.Becausetheserightsarepurelyfinancial,theycanpotentiallybeacquiredbyfinancialparticipantsthatdonotusethemforhedgingandwouldinsteadvaluethemaspurelyfinancialinstru‐ments.ThesefinancialfirmscanutilizetheirpurchasedCRRstocreatesecondaryhedginginstrumentstheymightinturnselltomarketparticipants,ortheymaybevaluedpurelyasrisky,butpotentiallyprofitabletrades.TheCRRparadigmalsoprovidesamechanismforISOstoredistributethecongestionreve‐nuesitcollectsfromitsoperationofmarketswithlocationalmarginalprices.Chargingeffi‐cientcongestionpricesproducesasurplusforthemarketoperatorthatcanbereturnedtomarketparticipantsviaCRRsoranalternativedesign.Althoughfinancialtransmissionrightsweredesignedsotheywouldbefundedbythecongestionrentscollectedbythesys‐temoperator,becauseofthewayCRRsaredefinedintheCAISO,andinotherISOmarketsaswell,paymentstoCRRscanexceedthecongestionrevenuescollected,aproblemknownasrevenueinadequacy.Revenueinadequacyisarelated,butdistinctissuefromtheauc‐tionrevenueshortfallsthathavemotivatedthecurrentCAISOproposal.Revenueinade‐quacycreatesanoperatingdeficitwherebyongoingcongestionpaymentstoCRRholdersexceedcongestionrevenues.Auctionrevenueshortfalls,incontrast,equatetosellinganassetforlessthanitsexpectedfuturepayout,adjustingforthetimevalueofmoney.Ofcourse,paymentstoCRRholdersareonlyoneofmanypossibleusesforCAISOcollectedcongestionrevenues.WhenthereisasurplusofcongestionrevenuenetofCRRpayments,thissurplusisappliedtoreducetherevenuerequirementforratepayersofthetransmis‐sionsystem.However,congestionrevenues,likelocationalprices,fluctuatewithmarketandnetworkconditions,attimesinunpredictableways.ThecongestionrevenuescollectedbyISOsthereforeconstituteanuncertain,or“riskcreating,”revenuestream.Themarketpartici‐pantspayingthosecongestionpricesfacetheoppositeriskposition.Importantly,whenaCRR,whichpaysthepricedifference(orcongestioncost)betweentwolocations,istrans‐ferredfromtheCAISOtoamarketparticipantthatwillusingthenetworkinthoseloca‐tions,bothsidesreducetheirriskexposuretocongestioncosts.Inotherwords,whendis‐tributedtofirmsusingthemashedginginstruments,CRRsreduceriskforbothsidesandconstituteanefficientallocationofrisk,particularlywhenthepartiesinvolvedareriskaverse.Thesignificanceofthisefficiencybenefitwilldependuponhowriskaversethepar‐tiesare,andthedegreetowhichannualandmonthlyCRRshelptoreducethoserisks.TheISO,orindirectlytheratepayerswhoareresidualclaimantstocongestionrevenues,arethereforeinauniquepositiontoprovideCRRstomarketparticipants.Theyarethenaturalcounter‐partiessincetheyhavetheoppositerevenuestream.TheDMMhasarguedthatfinancialfirmsorotherthirdpartiescouldprovideCRRstothosewhoneedthem,butthesefirmswouldbetakingonrisk,ratherthansheddingit,todoso.ThecoststotheCRRholderwouldbehigher,butitishardtodeterminehowmuchhigher.WearenotawareofqualitydataontheavailabilityandcostofCRRsprovidedbythirdparties,butwearemore

5

skepticalthanDMMabouttheabilityofthirdpartiestofulfillCRRneedsintheabsenceofISOdefinedinstruments.9Forfirmsandindividualsengagedinthegeneration,marketing,andpurchasingofpowerinISOmarkets,hedgesoflocationalpricedifferences(congestioncosts)canbehighlyvalua‐ble,andtheacquisitionofhedgesanimportantcostofdoingbusiness.Whenpairedwithaphysicaltransaction,CRRsprovidedasimilareconomicreturnaswouldthephysicalrightto“deliver”powerfromagenerationsourcelocationtoacustomerloadsinklocation.

2.2 The Present Distribution of CRRs in the California ISO Market

IntheCAISOmarket,policieshavedevelopedtoallowload‐servingentitiesinitialandex‐clusiveaccesstosomeCRRsthroughamulti‐stageallocationprocess.Thesepoliciesweredevelopedwithatleasttwomotivations.ThefirstwasanearlyconcernthatCRRscouldbeusedtoleveragemarketpowerifsuppliersinaconstrainedareawereabletopurchaseCRRssinkingintheconstrainedareaandtherebyincreasetheirprofitsfromwithholdingoutputanddrivinguppricesintheday‐aheadmarket.10IfanindividualsupplierwasabletoincreasethepayouttoCRRsbyexercisingmarketpower,itmightbewillingtooutbidothercompetitorsforthoseCRRs.Giventheeffectivenessoflocalmarketpowermitiga‐tion,andtheoverallcompetitivenessoftheCAISOmarketinrecentyears,itisnotclearhowsignificantthisconcernremains.Thesecondmotivationforallocationwastheconnectionmadebetweenaccesstoconges‐tionrentsandobligationstopayforthephysicaltransmissioninfrastructure.IntheCAISO,thecapitalcostsfortransmissionassetsarerecoveredthroughaTransmissionAccessCharge(TAC)appliedtoloadaswellasenergythatisexportedorwheeledthroughthesys‐tem.Inthissense,load(andotherTACpayers)arethetransmission“owners”,anditislog‐icalthattheywouldhaveclaimtothecongestionrentsproducedbytheassetstheyhelpedpayforthroughtheirelectricrates.Thesecongestionrentscouldbedistributedinseveralways:(1)adirectrefundofcongestionrentsinproportiontoconsumption(orequivalentoffsetsofotherchargestoconsumers);(2)directlytransferringCRRsthroughallocation,or(3)transferringtherevenuesraisedfromCRRauctions.Thefirstoption(proportionalrefund)wouldguaranteeabalancedrefundofcongestionrevenuestoalltransmissionratepayers,butitwouldalsoeliminatetheabilitytousecon‐gestionrevenuestosupportpaymentsofCRRs.Further,thedistributionoftherefundedcongestionrentscouldbeverydifferentfromthedistributionofcongestioncostsasare‐sultofdifferentconsumersfacingdifferentlocationalmarginalprices(LMPs).Hence,it

9EvenifqualitydataonthecostofthirdpartyCRRequivalents(sometimescalledbasisswaps)wereavaila‐ble,itisimportanttonotethattheseproductsarebeingsoldinthecurrentmarketenvironmentwherecon‐ventionalCRRsareallocatedandauctionedaccordingtocurrentrules.FirmsprovidingbasisswapscouldverywellbeusingconventionalCRRstobalancetheirriskexposure.Therefore,asharpreductionorelimina‐tionofconventionalCRRscouldspillovertothepricingofthird‐partysubstituteinstruments.

10Wolaketal.,2005and2007,op.cit.

6

wouldnotprovideamechanismfortransmissioncustomerstoobtaintheequivalentoffirmtransmissionservice.Instead,theprimarymechanismappliedinCalifornia,measuredintermsofcapacity,hasbeenthedirectallocationofCRRstoLSEs.DespitepayingTAC,exportandwheel‐throughtransmissioncustomersarenoteligibleforCRRallocations.ThisdifferentialtreatmentofexternalloadwasarguedtobeacceptableasexternalloadcouldaccessISO‐backedCRRsthroughtheauctionprocess.11ThisaccesswouldchangeunderproposalsbytheDMMde‐scribedbelow.Currently,theCAISOallocatesupto75%ofitsmodeledtransmissionca‐pacitythroughamulti‐stageannualallocationprocess,thatincludeslong‐term(10year)andseasonalyear‐aheadCRRs.Additionaltransmissioncapacity,thattheCAISOdescribesasuptoroughly82%ofmodelledcapacity,ismadeavailableinamonthlyallocationpro‐cess.Inpractice,however,theproportionofcongestionrentpaidtoCRRsawardedintheallocationprocesshasnotbeenintherangeof75‐82%buthasbeenonly54.6%overtheperiodJanuary2015throughMay2017.12Ifauctionswereeliminated,theallocationprocesscouldalsoproduceresultsconsideredinequitablebysomeLSEs.ItispossiblethatsomeLSEsmaybeabletotargetandacquireCRRsthataremorevaluable,intermsofeitherprojectedcongestionrevenuesorsecond‐arymarketvaluethanotherLSEs.NeithertheCAISOnorDMMhasexaminedtheequityofthecurrentallocationprocessintermsoftherelationshipbetweentheCRRpaymentsre‐ceivedandcongestionchargespaidbydifferentcategoriesofloadservingentities.ItisonlyaftertransmissioncapacityhasbeenassignedtoLSEsintheformofallocatedCRRsthatanyremainingcapacityisauctionedintheannualandmonthlyauctions.Thelevelofoverallnetworkcapacitymadeavailabletotheauctionsisthesameasthatmadeavailabletotheallocations(75%year‐aheadandroughly82%month‐ahead).Nopercent‐ageofnetworkcapacityisdirectlyreservedexclusivelyforsaleintheauctionsotherthan½oftheintertiecapacitythatisavailableattheendofthe2ndtierallocationround.13ThelimitationthatallocatedCRRsmustsinkatparticularlocationsmayindirectlypreventtheallocationofcertainCRRsthatcreateflowsonsomeconstraints.Atfirstglance,onemightthereforeinterprettheauctionsasthedisposingofunwantedca‐pacity“leftover”fromtheCRRallocationprocess,ratherthana“forcedsale”ofCRRs.

11“(E)xternalLSEswillreceiveatremendousbenefitfromtheCAISOanditstransmissionownersinbeingallowedtoacquireCRRsasneededintheCAISOcoordinatedCRRauctionswithoutbeingobligatedtopayem‐beddedcostchargesonaprospectivebasis.ItislikelythatmostexternalLSEswillacquiretheCRRstheyneedtohedgepotentialwheelingthroughandouttransactionsintheCRRauctionandonlypayembeddedcostusagechargestotheextenttheyactuallyusethesystem.”PreparedDirectTestimonyofScottHarveyandSusanPope,FERCDocketER06‐615‐000.February2006,pp.139‐140.

12WeusetheMay2017enddatetobeconsistentwiththetimeperiodcoveredbytheCAISONovemberre‐port.OvertheperiodJanuary2015throughDecember2017theratiois57%.

13CaliforniaISO,CongestionRevenueRightsAuctionEfficiency,Track1DraftFinalProposal,February8,2018,p.8.

7

However,therearecriticaldifferencesbetweentheallocationstageandtheauctions.Mostimportantly,LSEsarerestrictedtochoosefromalimitedsetofsource‐sinkpairsinthecur‐rentallocationprocess.Undercurrentmarketrules,LSEspay,forenergy,azonalpricethatisanaggregationofload‐nodeLMPs,orDLAPorCLAP,intheirregions.Therefore,mostLSEsdonotneedtohedgeLMPsatspecificloadnodes,andarerestrictedtoselectingCRRsthatsinkinsteadatLAPs.Asaconsequence,eachCRRthatisselectedbyanLSEisassumedtoimpactflowstoeveryloadnodeinitsrespectiveLAPregion,andthereforeimpactsalargenumberoftransmissionconstraints.Therefore,intheallocationprocess,whenonlyoneoftheseconstraintsbindsat75%(or82%)ofmodelledcapacity,nofurtherCRRssink‐ingattheLAPsarefeasible,andhenceeligibleforallocation,eventhoughtheflowsassoci‐atedwithallocatedCRRsoverotherconstraintsonthenetworkmaybewellbelowthe75%threshold.14This“lumpiness”intheCRRallocationphaseislikelyanon‐trivialsourceofextracapacitybeingavailableintheauctionprocess,wherebidderscanchoosefrommorepreciselytargetedCRRs.TheotherchannelthroughwhichadditionalCRRcapacitycanbecreatedintheauctionisthroughawillingnessbygeneratorsorfinancialparticipantstotakeonriskintheformof“counterflow”CRRs.15Totheextentthatgeneratorsorfinancialparticipantsacquirecoun‐terflowCRRsthatinturnfacilitatethesaleofadditionalhedgingCRRs,thisadvancesthepurposeoftheCRRmarket.NoloadservingentityiscompelledtopurchaseaCRRsup‐portedbythepurchaseofcounterflowCRRsbygeneratorsorfinancialparticipants.AloadservingentitycanchoosetopurchaseaCRRsupportedbycounterflowifthepriceoftheCRRisfinanciallyattractive.Inaddition,totheextentthatfinancialparticipantscompetewithloadservingentitiesforhedgingCRRs,theyhelpconvergeauctionpricesandCRRpayoutsbyraisingauctionprices,soparticularloadservingentitiesarenotabletobuyCRRsatalargediscounttotheauctionpricebecauseofalackofcompetitionforthoseCRRsfromotherhedgers.InadditiontoallowingfinancialparticipantstosellcounterflowCRRsandtherebyalsoena‐blethesaleofadditionalhedgingCRRsmadefeasiblebythecounterflowCRRs,CRRauc‐tionsservethreeotherroles.First,whiletheallocationprocessallowsloadservingentitiestoacquireCRRssourcingatgeneratorsandsinkingattheirDLAPorCLAP,notallloadserv‐ingentitieseitherowngenerationorcontractforthepowerofanindividualgenerator.Someloadservingentitiesmayinsteadchoosetobuypoweronadeliveredbasis,attheirCLAPorDLAP,ormaychoosetobuypoweratatradinghub.16Settlingforwardcontracts

14CRRsarerestrictedtosinkingatDLAPSintiers1and2oftheallocationprocessbutcanbedesignatedtosinkatthemoredisaggregatedCLAPsintier3.

15AcounterflowCRRisaCRRthatisexpectedtohaveanegativepayoutandwouldsellatanegativepriceintheCRRauction.Inotherwords,holderswouldrequireapaymenttotakeontheobligationsassociatedwiththecounterflowCRR.ThepaymenttoholdsuchaCRRwouldbefundedbyauctionparticipantswhoarewill‐ingtopayfortheadditionalCRRsmadefeasiblebythecounterflowCRR.

16Itcanbeseeninthe2018allocationprocessthatanumberofsmallerloadservingentitieswereallocatedCRRsfromtradinghubstotheirload,suggestingthattheywereseekingtohedgepurchasesatthesetradinghubs.Theseloadservingentitieswerenotthelargeinvestorownedutilitiesbutsmallerloadservingentities

8

attradinghubsisacommonpracticeinmostU.S.electricitymarketsbecausethecommontradingpointprovidesmoreliquiditywhenthebuyerorsellerwanttobuyorsellsomeofthepowercoveredbythecontract.InorderforsupplierstobeabletohedgecontractscallingforthedeliveryofpowerataCLAP,DLAPortradinghub,thesupplierneedstobeabletoacquireaCRRfromitssupplysourcetothepointofsale,aCLAP,DLAPortradinghub.TheCRRsasupplierwouldneedforthishedgingneedtobeacquiredinanauctionorfromthirdparties17(becausesuppli‐ersarenotentitledtoparticipateintheCAISOCRRallocationprocess),andwillbeunlikelytocorrespondtotheCRRsallocatedtoanyentity.Second,theallocationprocessdoesnotprovideamechanismforloadservingentitiestosellCRRsthattheynolongerneedforhedgingbutcouldbereconfiguredintheauctionintoaCRRthatcanhedgethecongestionchargesofanotherloadservingentityorsuppliersell‐ingtoaloadservingentity.18Third,theauctionallowstransmissioncustomersthatwanttosellpoweroutofCaliforniaorwheelpowerthroughCaliforniatoacquireCRRsthathedgethecongestionchargesonthosetransactions.Whilemorein‐depthanalysisoftheallocationprocessisrequiredfordevelopingacompre‐hensiveCRRpolicyinCalifornia,thereareatleasttwoqualitativeimplicationsoftheallo‐cationprocessforthecurrentproposal.First,eliminationofanauctionwouldfurtherex‐cludenon‐LSEparticipantsfromdirectlypurchasingISO‐backedCRRs.Itwouldnotelimi‐nateall,orevenamajorityof,CRRcapacity,whichwouldremaindirectlyassignedtoLSEs.ButpurchaseofthoseCRRswouldrequireatransactioninwhichanLSEputstheCRRupforsale.Second,certaintypesofISO‐backedCRRssimplycouldnotexistiftherewerenotanauction.Forexample,point‐to‐pointCRRssinkingattradinghubscanonlycurrentlybeassignedthroughtheauction,giventhelimitationsontheallocationprocessdescribedabove.Also,CRRssourcinginsidetheCAISOandsinkingatschedulingpointsforexportscanalsoonlybeacquiredintheCRRauction.

2.3 The DMM and SCE Proposals

BeforeturninginSection3toourdiscussionofissuesassociatedwithrevampingtheCAISOCRRauctionprocess,wesummarizeanalternativeapproachfavoredbyanumberofstake‐holders.suchasthecityandcountyofSanFrancisco,theCityofCorona,thePortofStockton,SonomaCleanPowerAuthorityandGoldenStateWaterCompany.

17ThiswouldrequirethataloadservingentityhavenominatedaCRRfromthegenerator’snodetotheLAPthatwasnotneededtohedgesupplycontractsofthatloadservingentity.

18Thiscannotbecarriedoutthroughbilateraltradesunlessthebuyerwantedtohedgecongestionbetweenexactlythesamesourceandsinkastheseller,whichisunlikely.Auctionsalesallowthebuyertohedgecon‐gestionbetweenadifferentsourceandsinkthatcreateflowsonsomeofthesamebindingconstraintsintheauctionastheCRRbeingsold.However,iftheauctionisdesignedtohavenosparecapacityonanyconstraintasproposedbyDMMandSCE,thenlikethebilateralsituation,anysaleofaCRRwouldhavetomatchedbypreciselythesameamountofcounterflowrightintheoppositedirectionbetweenthesametwolocations.

9

WhilenoteliminatingtheCAISOCRRauctionscompletely,aproposalbyDMM,aswellasarelatedproposalbySouthernCaliforniaEdison,19wouldmakesignificantchangestohowtheauctionisdefinedandexecuted.Asweunderstandtheseproposals,theywouldlimitthetransmissioncapacityavailabletosupporttheawardofCRRstoalevelequaltotheflowsimpliedbyCRRsassignedintheallocationprocess.Intheannualprocess,atleastone‐constraintwouldlikelybindat75%(becausethatisthemaximumavailableintheallo‐cationprocess),butothers,asdescribedabove,couldemergefromtheallocationwellbe‐lowthisthreshold.Unlikethecurrentprocess,undertheDMM/SCEdesignthetransmis‐sioncapacityavailableonconstraintsthatdidnotbindintheallocationwouldnolongerbeavailabletosupportthesaleofISO‐backedCRRstotransmissioncustomers.TheonlywayCRRscouldbeacquiredintheauctionwouldbeifanLSE(orathirdpartywillingtotakeontherisk)madeavailabletheexactcounterflownecessarytoenablethenewCRRtobeawardedwithoutincreasingtheflowsonanytransmissionelement,bindingornot,abovetheflowsduetotheCRRsawardedintheallocationphase.UnliketheCAISOproposaldescribedbelow,thisproposalwouldsubstantiallyreducethecapacityoftransmissionmadeavailabletoformISO‐backedCRRs,andwealsobelievewouldfurtherrestrictthetypes(intermsofsource‐sinkpairs)ofISO‐backedCRRsthatcouldbeawarded.Totheextentthatrisk‐seekingthirdpartiesarewillingtoprovidespec‐ulativecounter‐flow,suchinstrumentscouldbeavailable,butlikelyatahighercostthaniftheCAISOuseditsavailabletransmissioncapacity(ATC)tosupporttheawardofCRRs.AlthoughthetransmissionsystemwouldhaveATCavailabletosupporttheawardofCRRstosuppliersseekingtoselltoLSEsattradinghubs,thisdesignwouldnotallowtheCAISOtomakeanyCRRsavailableforsaleutilizingthisATC.20NorwoulditmakeATCavailabletosupportthesaleofCRRsthatsupportexportsorwheel‐throughtransactions.Further‐more,noATCwouldbemadeavailablebytheCAISOtosupportthepurchaseandsalebyloadservingentitiesofISO‐backedCRRsforwhichtheCRRpurchasedandtheCRRsoldhavedifferingflowsonanytransmissionelement.TheDMM/SCEdesignwouldverylikelybeeffectiveinpreventingfinancialmarketpartici‐pantsfromacquiringCRRsatadiscounttotheexpectedpayout.However,itwouldalsopreventphysicalmarketparticipantsfromacquiringCRRsatadiscountorapremiumtotheexpectpayout.Inotherwords,whiletheproposalwouldeffectivelyeliminatethepur‐chaseofspeculativeCRRsintheauction,itdoessoatacostofsharplyreducingaccessto

19DMMProposal,MarketAlternativestotheCongestionRevenueRightsAuction,November27,2017.www.caiso.com/Documents/DMMWhitePaper‐Market_Alternatives_CongestionRevenueRightsAuction‐Nov27_2017.pdf;SCEProposal,www.caiso.com/Documents/SCEComments‐CRRAuctionAnalysisReport.pdf,postedonDecember11,2017.

20Theproposalwouldallowmarket‐participantssuchasgeneratorstooffercounterflowCRRsatpricesthatwouldincorporatetheriskassociatedwithofferingsuchCRRs,buttheproposalwouldnotallowadditionalATCtosupporttheawardofadditionalCRRs.TheproposalcouldthereforebeinterpretedaswithdrawingATCthatsupportsCAISO‐backedCRRsandreplacingthoseCRRswithspeculativecounterflowCRRsofferedathigherprices.

10

ISO‐backedhedgingCRRs,andoflockingintherigiditiesofthecurrentallocationprocessdiscussedabove.Thisdesignissupportedbyanumberofloadservingentitiesbutthesupportersdonotap‐peartoincludemostofthesmallerloadservingentitiesthatacquiredCRRssourcingattradinghubsandsinkingatLAPsintheallocationprocess,suchasthecityandcountyofSanFrancisco,GoldenStateWaterCompany,PortofStockton,WesternPowerAuthority,orCityofCorona.Itisalsostronglyopposedbyloadservingentitiesservingdirectaccesscustomers,andsomesmallerLSEssuchastheLSE‐CRRAuctionSupportersGroup.Loadservingentitiesthatdependonmeetingtheirloadinpartbypurchasingpowerattradinghubs,inturndependonthesupplierstheycontractwithforpowerbeingabletosellthempoweratthetradinghubs.However,undertheSCE/DMMdesigntherewouldbenomech‐anismforsupplierstoacquireCRRssinkingattradinghubsandsupportedbytheavailableATCofthetransmissionsystemtohedgecongestionchargesontheirtradinghubsales.TheSCE/DMMdesignwouldallocatetheavailableATCatazeropricetotheloadservingentitiesusingtheirownedandcontractedgenerationtomeettheirload,butthedesignwouldpreventtheCAISOfrommakingATCavailabletosupporttheawardofCRRs,atanyprice,tosupplierssellingtoLSEsattradinghubs.WhiletheSCE/DMMproposalwouldal‐lowtheresaleofCRRsource‐sinkpairsheldbyanLSEtoothermarketparticipants,source‐sinkpairsthataloadservingentitydidnotacquirewouldnotbeavailableforpur‐chase.Moreover,ifthevastbulkoftheCRRswereallocatedtothelargeloadservingenti‐tiesateachLAP,theywouldhavenoobligationtomakethoseCRRsavailableforsaletoanyotherLSE,orsupplier.TheDMMhasremarkedupontheincentiveproblemsandregula‐toryconstraintsthatlimitregulatedLSEparticipationintheauction,yetthisproposalre‐liesuponsuchparticipationastheonlychannelthroughwhichCAISO‐backedrightscouldbereconfigured,orredistributedtonon‐LSEs.21Further,ifthecurrentruleslimitingtheallocationofCRRstoloadservingentitiestothosesinkingatLAPs,theredoesnotappeartobeanywaythatCRRscouldbeacquiredtohedgeexportorwheelthroughtransactions.IfoneacceptsthepremisearticulatedbyFERCthatISO‐backedCRRsareaformoffirm‐transmissionservice,thenundertheDMM/SCEproposaltherewouldbenosuchfirmtransmissionserviceofferedbytheCAISOtosupportpowersalesattradinghubswhenATCpermittedit.Theonlywaytosellpowerattradinghubswouldbetheuseof“non‐firmtransmission”thatincludesthevolatilecostsofcongestion.WhiletheDMM/SCEproposalmightappeartobenefitlargeCalifornialoadservingentitiesbyreducingtheauctionrevenueshortfalls,itcanalsobeconstruedasrestrictingforwardaccesstotheCAISOtransmissiongrid.Thiscanhaveunpredictablenegativeconsequencestowholesalemarkets.IfananalogouspolicyweretobepermittedbyFERCallowingthedenialofforwardaccesstothetransmissionsystemoutsideCaliforniatoCaliforniaLSEsandthosesellingtothoseLSEs,itwouldbelikelytoadverselyimpactCaliforniapowercon‐sumers.

21Problemsintheperformanceanddesignofthecongestionrevenuerightauction,CAISODMM,p.24.

11

3. IdentifiedIssueswithCRRsandtheCRRAuctionInthissectionwediscusstheissuesthathavebeenidentifiedwiththeCRRprocess,andalsoareasinwhichfurtheranalysiswouldbehelpfulingainingabetterunderstandingoftheunderlyingcausesoftheseoutcomes.Theprimaryissueofconcerninthisproceedingiswhatiscalledthe“auctionrevenuedeficiency.”Asdiscussedabove,thisisdistinctfrom,butrelatedtorevenueinadequacy.Asstatedabove,theCAISOandtheDMMhaveshowninanumberofreportsthattheaveragepayouttoCRRsacquiredinCAISOauctionshasmateriallyexceededtheauctionrevenuesfromthesaleoftheseCRRs.22Thisoutcomecanaberesultoftheunpredictabilityofcongestionpatternsinanyparticularmonth,anditispreciselythisvariabilitythatisthereasonloadservingentitiesandsuppliersenterintoforwardpowercontractsneedCRRstohedgecongestionrisk.However,theCAISOandDMManalyseshaveshownthatthepaymentstoCRRshaveexceededtheauctionrevenuesonaverageoveralargenumberofauctions,sotherelationshipbetweenauctionrevenuesandpayoutsintheseanalysesreflectsmorethanjustthevariabilityofcongestionpat‐terns.23TheCAISOdatashowsthatthetotalpaymentstoauctionedCRRsovertheperiodJanuary2015toMay2017was$365.3million,comparedtoauctionrevenuesofonly$220.1million,indicatingthattheCRRsoverallsoldforonly60.3%oftheexpectedpayout(around54%forthemonthlyCRRsand69%fortheseasonalCRRs).Asdiscussedabove,thecorepurposeoffinancialtransmissionrights(CRRs,FTRsandTCCs)istosupportforwardcontractingbyloadservingentitiesandothermarketpartici‐pantsbyfacilitatingtheparties’abilitytohedgepotentialfuturecongestioncharges.WeagreewiththeviewthatthepurposeofaCRRauctionistosupportthiscorepurposeoffa‐cilitatinghedgingofcongestioncostsandtherebypromoteopenaccesstothetransmissiongrid.Thegoalisnottofacilitatethesellingoffinancial(orphysical)transmissionrightsthathavelittleornovalueashedgesatalargediscounttotheirexpectedvalue.ThisholdsseveralimplicationsfortheinterpretationofCRRmarketresults.WhilesomeoftheCRRsintheauctionmaybeacquiredinordertohedgecongestionriskandmightbevaluedatapremiumtotheexpectedpayout,thisisnotthecaseoverall.Instead,theoverallpatternisoneofCRRsbeingvaluedbybuyersasriskyfinancialinstrumentsandpurchasedatadiscounttotheexpectedpayoutthatmateriallyexceedsthetimevalueofmoney.24WhiletheinstanceofanyspecificCRRsellingforlessthanitssubsequentcongestionreve‐nuevalueisnotsufficientevidencethatitwassoldforlessthanitsexpectedvalue(which22Seeforexample,CaliforniaISO,CRRAuctionAnalysisReport,November21,2017pp.49‐56,CaliforniaISODepartmentofMarketMonitoring,2016AnnualReportonMarketIssuesandPerformance,pp.200‐204.

23NeithertheDMMnorCAISOcomparisonsofauctionrevenuesandCRRpayoutsforannualCRRsaccountforthetimevalueofmoneydifferenceinthetimingofpaymentsforandtoannualCRRs.

24Becauseannual/seasonalCRRsarepaidforasmuchasayearbeforeanypayoutwouldbereceived,thereisatimevalueofmoneybenefitfromreceivingtheseadvancepayments.Theratesofreturnfortransmissionownersareintherangeof10%orless,however,whicharetoolowtoaccountformorethanasmallportionofthedifferencebetweenauctionrevenuesandCRRpayouts.

12

can’tbeobserved),thepersistentpatternofoverallauctionrevenuesfallingbelowoverallcongestionpayoutsisevidencethatthiseffectisdominatingtheaggregateauctionout‐comes.OnequestioninunderstandingauctionoutcomesistheextenttowhichtherearestructuralfactorscontributingtoarelativelylowdemandforCRRsashedginginstrumentsinCAISOCRRauctions,therebyyieldinglowauctionprices.Thiscouldinpartbebecausethefocushastodatebeenontheauction,andthemostvaluablehedgingCRRsarelikelyacquiredintheallocationprocess.ThevalueandpayoutoftheseallocatedCRRsisnotincludedinthesecomparisons.Inaddition,becausetheportionofloadservedbyregulatedutilitiesismuchhigherinCaliforniathaninregionslikeNYISO,PJMorISONewEngland,theremightbelessdemandforcongestionhedgesbecauseofelementsofCPUCregulationthatdiscour‐ageregulatedutilitiesfromacquiringCRRsinauctions25orbecauseCPUCregulationsdonotincentregulatedutilitiestohedgecongestionrisk.26Sincenostatisticshavebeencom‐piledonthevaluationofCRRsacquiredintheallocationprocess,wecannotassesswhetherCRRsingeneralarenotvaluedashedgesorwhetheritisparticularlytheCRRspurchasedinCAISOauctionsthatarenotvaluedatthemarginashedgesbutasriskyfinancialinstru‐ments.27

WhileonepossibilityisthatCRRsaretypicallysoldatadiscounttotheexpectedpayoutbe‐causethereisagenerallackofdemandforcongestionhedgesinCAISOauctions,anotherpossibilityisthatthelowvaluationofCRRsisspecifictotheCRRssoldintheauctionandreflectsthepayouttoCRRsthathavelittlevalueashedges.CRRshavebeenstructuredinCaliforniatoprovideloadservingentitiesaperfecthedgeforcongestionchargesbetweentheCRRsourceandtheCLAPorDLAPatwhichtheloadservingentitybuyspower,withoutregardtotheimpactoftransmissionoutages,loopflowsorchangesinloaddistributionfac‐torsfortheCLAPsandDLAPsontheextenttowhichthetransmissiongridisabletosup‐portthepaymentstotheCRRholder.BecauseofthisCRRdesign,loadservingentitiesandsupplierssellingtoloadservingentitieshavenohedgingincentivetoacquiregenerator‐to‐generatorCRRsthatwouldhaveapayoutinanyofthefollowingcircumstances:

(1) whenparticularoutagesoccur,andtheoutagecausesconstraintstobindthatwerenotbindingintheCRRauctionmodelorcausestheCRRtohaveanincreasedshiftfactorintheday‐aheadmarketoverconstraintsthatdidbindintheCRRauction,

(2) whenthereareloopflowsmodeledintheday‐aheadmarketthatchangetheavaila‐bletransmission,or

25Thiscouldbethecase,forexample,iftheregulatedutility’sshareholderswereatriskforlossesonCRRpurchaseswhileprofitswouldbepassedthroughtoratepayers.

26Thismightbethesituationif,e.g.,theregulatedutilitywasabletopassthroughanycongestionchargesinitsrates,sothatratepayershadtobearthecongestionrisks.

27BecauseallCRRsarepricedinauctions,allocatedCRRscouldbevaluedatbothseasonalandmonthlyauc‐tionpricesandthisvaluationcomparedtothepayout,butneithercomparisonhasbeencarriedoutbytheCAISOorDMM.

13

(3) whenthedistributionofloadintheday‐aheadmarketcausesdifferentconstraintstobindintheday‐aheadmarketthanintheallocationorauction.

IfCRRsthathavelittlevaluetoloadservingentitiesortheirsuppliersascongestionhedgesareofferedintheauction,onewouldexpecttheywouldbevaluedasriskyfinancialinstru‐mentsandsoldatadiscounttotheexpectedpayout.Itisimportanttounderstandthatthisoutcomeisnotnecessarilyaresultofalackofcompetitionamongthepurelyfinancialauc‐tionparticipants.Nomatterhowintensecompetitionisamongthefinancialauctionpartic‐ipants,theywillnotvalueCRRsatapremiumtotheexpectedpayout;purefinancialplay‐erswillonlybewillingtobuypositivelyvaluedCRRsatadiscounttotheexpectedpayout.Alackofcompetitionwouldincreasethediscounttotheexpectedpayoutbutmorecompe‐titionfromfinancialmarketparticipantswillnoteliminatethediscount.Competitionbe‐tweenfinancialmarketparticipantsandhedgers(becauseofsharedtransmissioncon‐straints)mightresultinahighervaluationofCRRsthatarevaluedashedges,butonlybyafewload‐servingentitiesorsuppliers.However,competitionfromfinancialparticipantscannoteliminatethediscountforCRRsthatarevaluedonlyasfinancialinstruments.Moreover,themorecomplexitistoanalyzeandvalueCRRs,themorecostsfinancialplay‐ersmustincurtoparticipate,andthelargerthedifferencebetweentheauctionpriceandday‐aheadmarketpayoutmustbeinequilibriumtoenablethoseparticipantstorecovertheircosts.Thesameisalsotrueforload‐servingentitiesthatmightvaluetheseCRRsasimperfecthedges.ThemorecomplexitistovaluetheseCRRsasimprovingtheirconges‐tionhedges,thelowerthepremiumwillbethattheloadservingentitywouldbewillingtopay(orthelargerthediscountitwouldneedtoreceive)fortheCRRsrelativetotheex‐pectedpayout.Finally,totheextentthattransmissionoutagescausedifferencesbetweenthetransmissionmodelusedintheauctionprocessandthatusedtodetermineCRRvaluesintheday‐aheadmarket,thereisapotentialforauctionparticipantstoacquireCRRsthatcreatelittleifanyflowsonconstraintsmodeledintheauction,andhencearevaluednearzero,butcreatelargeflowsontransmissionelementsthatwillbindontheday‐aheadmarketwhenparticu‐laroutagesoccur.28BecauseintheCAISOCRRdesignthetransmissiongridmodelusedintheauctiondiffersfromthegridmodelthatwilldetermineCRRvaluesintheday‐aheadmarket,thehedgingCRRsacquiredbyloadservingentitiesandsuppliersintheallocationandauctionmaynotcreateverylargeflowsonsuchconstraintsintheauctionmodel,andtheconstraintsmaynotevenbeenforcedintheauctionmodel.Ifthisisthecase,theseconstraintswillnotbindintheauctionsothatthepriceofCRRflowsontheseconstraintsintheauctionwillbezero.

28ThepriceofaCRRintheCRRauctionisdeterminedbytheflowtheCRRcreatesontheconstraintsthatbindintheauctionsolution,timestheshadowpriceofeachauctionconstraint.IfaCRRcreatesnoflowsonanyconstraintthatbindsintheauction,itwouldhaveazeroprice.IfaCRRcanbeacquiredbetweenanypairofnodesinthetransmissionmodel,itmightbepossibletoidentifyandselectCRRsthatturnouttocreateminimalflowsonbindingconstraintsandhencesellatalowprice.

14

ThisoutcomewasseenforavarietyofconstraintsintheCAISOanalysesoftheCRRauctionoutcomes,particularlyduringthetypicalmaintenancemonthsofOctoberthroughDecem‐ber.29Inanumberofcases,morethanamilliondollarsofCRRpaymentsweremadetoCRRssoldintheannualandmonthlyauctionsforconstraintsthatwerenotmodeledintheauctionsandhencetheconstraintcouldnotbindinauctionandwasvaluedatzeroinde‐terminingCRRprices.Indeed,inanumberofmonthsmorethanalloftheauctionrevenueshortfallappearstobeaccountedforbythepayouttoCRRscreatingflowsonconstraintsthatdidnotbindintheannualandmonthlyCRRauctions.Sometimesthisamountofpay‐outarosefromasingleconstraintthatdidnotbindintheCRRauctionsbecauseitwasnotmodeled.30TheCRRssoldintheauctionthatcreatedtheseflowsmayhavealsocreatedflowsonotherconstraintsthatdidbindintheauctionsothattheCRRswerelikelynotsoldatazeroprice.Theywere,however,potentiallysoldataverylargediscounttotheday‐aheadmarketpay‐out.31ThepotentialforthiskindofoutcomeinwhichCRRsthatcreateflowsonconstraintsthatbindintheday‐aheadmarketbutdonot,orcannot,bindonthatconstraintintheCRRauction(andhencedonothingtoimproveauctionvaluation)isincreasedbyrulesandpoli‐ciesthatrequiretheCaliforniaISOtodisclosethedetailsoftheauctionmodelanumberofdayspriortotheauction,anddonotallowtheCaliforniaISOtochangethemodeltoreflectadditionaloutages.ThisauctionstructureallowsmarketparticipantswhoarenotseekinghedgestoidentifyCRRsthatwillcreateflowsontransmissionelementsthatwillbindwhenoutagesoccur,butcreatelittleifanyflowsontransmissionconstraintsenforcedintheauc‐tionmodel.TheseCRRbidsdonothingtoimproveCRRauctionvalues,theyareinfactstructuredpreciselytoavoidhavingthatimpact.

29CaliforniaISO,CRRAuctionAnalysisReport,November21,2017,Tables13‐23.

30Forexample,seethediscussionofDecember2016inSection4.2below.Similarly,inJanuary2017thenetpaymenttoauctionCRRsinexcessofauctionrevenueswas$4.5million.TheCRRpaymentstoasinglecon‐straintthatwasnotenforcedintheauction23040CrossTripaccountedfor$5.7millionofpaymentstoauc‐tionedCRRs,andanotherconstraint,OMS4622069TL50003accountedforanother$1.2million.InFebruary2017,thenetpayouttoauctionedCRRsinexcessofauctionrevenueswas$4.1millionTheconstraint7820TL23040IVSPSNGthatwasnotenforcedonlyaccountedfor$1.6millionofthisshortfallbuttherewereanumberofconstraintsthatwerenotenforcedthatcontributedtotheoverallshortfall.Theoverallshortfallwas$3.6millionforMarch2017,withasingleconstraintthatwasnotenforced,7820TL23040SPSNG,ac‐countingfor$3.4millionofthisandoneotherunenforcedconstraint31486Caribou11530255CarbouM1XF11accountingforanother$.8million.InApril2017theCRRpayoutexceededauctionrevenuesby$7.7million,andasingleconstraintthatwasnotenforced,6410CPSNG,accountedforanetpayoutofmorethan$6.7million.Oneotherconstraintthatwasnotenforced,OMS3831815TMSDLOaccountedforroughlyan‐other$.7million.TheCAISOanalysisinSection7oftheNovemberreportalsoshowslargepayoutsonotherconstraintsthatwereenforced,butdidnotbindintheauction.Thisoutcomecanarisesimplybecausecon‐gestionpatternsaredifferentfromexpectedbutitcanalsoarisebecauseGenerator‐to‐generatorCRRsareacquiredintheauctionthatcreatesmallflowsontheconstraintontheauctiongridbutcreatelargeflowsandpayoutsontheseconstraintsontheday‐aheadmarkettransmissiongridbecauseofoutagesthatwerenotmodeledintheauction.

31ItwouldbepossibletoidentifywhichFTRscreatedflowsontheseconstraintsintheauctionandcomparetheFTRpricestothepayoutsbutthiswouldrequirererunningtheauctionforforcetheseconstraintstobindsothatshiftfactorswouldbecalculated.

15

4. TheCAISOProposalThechangesproposedbytheCAISOwouldbeappliedtothenextauctionofannualCRRswhere,barringchanges,75%ofthesystemcapacityismadeavailablethroughasequentialprocessofallocationandauctioning.TheCAISOalsoplanstoimplementchangestotheauctionsformonthly2019CRRs.Asmentionedabove,theannualauctionhasbeenrespon‐siblefor$48millioninnetauctionrevenueshortfallsovertheperiodJanuary2015throughMay2017.32ThesechangesarethereforetargetedattheannualauctionbothbecauseofitsprominenceinauctionrevenueshortfallsandbecausetheauctionwillbethefirstinthenextcycleofCRRauctiondistributions.TheCAISOwillconsiderfurthercomprehensivedesignchanges,includingthosethatmightimpactmonthlyauctions,initsTrack2processduringtherestof2018.33TherearetwocomponentsofthechangesproposedforTrack1.Theseincludechangingthetimelinefortransmissionoutagereportingandchangestotheallowablesourceandsinkpairsthatcanbesoldintheauction.Wediscusseachofthesecomponentsbelow.4.1 TimelineforTransmissionOutageReportingOnesourceofbothrevenueinsufficiency(CRRpaymentsinexcessofcongestionrentscol‐lected)andnetauctionpaymentdeficiencies(CRRauctionpricesbelowthesubsequentrevenuestreamassociatedwiththem)hasbeenthemismatchbetweenconstraintsthatarerepresentedinthenetworkmodelusedbytheauctionandthosethatareactuallyincorpo‐ratedinthenetworkmodelusedtocleartheday‐aheadmarketwhosepricesareusedtosettleCRRs.Thisisafundamentalconsiderationwhendefiningthetruecapacityofthenet‐work.Whenconstraintsarenotincludedintheauctionmodel,theycannotbindandthere‐forecannotimpacttheavailabilityandpricesofCRRs.Ifthoseconstraintsdobindinthedailymarket,thenthereisapotentialthatmoretransmissioncapacityhasbeensoldintheformofCRRs,thanactuallyexists.Moreover,ifauctionparticipantsknowthataparticularconstraintthatwillbindduringanoutageisnotenforcedintheauctionmodel,theycanstructuretheirCRRbidstocreateflowsonthetransmissionelementthatisnotmodeledbuthavelittleornoimpactonmodeledconstraints.Alternatively,iftheyknowthatapar‐ticularunmodeledoutagewouldincreasetheshiftfactorsofparticularCRRsonconstraintsthatwouldbindintheday‐aheadmarket,theycanstructuretheirbidstoacquireCRRsthathavenovalueashedgesontheauctiongridbutwouldbeentitledtolargepayoutsduringtheoutage.EitherbiddingstrategycancontributetobothrevenueinsufficiencyandlowauctionpricesrelativetopayoutsforCRRssoldintheauction.Thereareseveralpotentialoptionsformitigatingtheseproblems,someofwhichmaybe

32CaliforniaISO,CRRAuctionAnalysisReport,November21,2017,p.52.

33CAISO,CRRAuctionEfficiencyTrack1DraftFinalProposal,p.4.

16

consideredduringTrack2.ForTrack1theCAISOhashighlightedabasicproblemwithconstraintmodelinginitsannualauction;thefactthatauctionisheldbeforetransmissionownersarecurrentlyrequiredtonotifytheCAISOofplannedoutagesforthefollowingyear.Currently,theCAISOdeadlineforsubmittingcomprehensiveannualmaintenanceplansisOctober15ofthepreviousyear,34whereasitsannualCRRprocessisruninAugustofthepreviousyear.Inordertobestcapturetheknownfutureavailabilityoftransmissionfacilitiesintimefortheannualauction,theCAISOisthereforeproposinganadditionaldeadlineforoutagereportingrelevanttotheCRRauctionofJuly1.DiscussionThereisaninherenttrade‐offbetweenearlierreportingdeadlinesandthequalityofinfor‐mationthatcanbecollectedintimetomeetthosedeadlines.ThemaindownsideoftheproposalisthatthestateofinformationasofJulytheyearpriortotheoperatingyearwillbeofmuchlowerqualitythanwouldbethecaseclosertotheoperatingmonth.Withre‐specttotheannualauction,evenpartialorimperfectinformationaboutfuturemainte‐nanceplansislikelybetterthannoinformationatall.However,therearelimitstothetransmissionowners’abilitytoprovideaccurateinformationthisfarinadvanceoftheout‐age,astheneedforsomeoutagesmaynotbeknownthatfarinadvance.Whileearliernoti‐ficationofknownoutagesstrikesusasacommon‐sensestepwithregardstotheannualauction,theneedforsomeoutageswouldnotbeknownatthatpointintimeandwillhavetobescheduledclosertotheoperatingmonth.Hence,asecondaryOctoberreportingdatewillberetained,andtheCAISOauctionmodelingwillneedtoallowforthepossibilitythatotheroutageswillneedtobescheduledclosertotheoperatingmonth.Thereisalsothequestionofenforcement.TheCAISOhasreportedthatforthemonthlyprocessonly43%percentofplannedoutageswerereportedbytheirrequiredreportingdateandtheCAISOdidnothavetimetoevaluatealloftheseoutagesforadjustmentstothemonthlyauctionmodel.Whilethisisclearlyproblematic,significantpenaltiesformissingdeadlinescouldleadtospeculativeoroverlyconservativeestimates.Thiswouldinturnraisethequestionofenforcingthequalityofreportedinformation.Itisthereforenotclearifthegainsofmoreheavy‐handedenforcementofreportingwouldbesubstantial.4.2 RestrictionsonAllowableSourceandSinkPairsThemostsignificantchangetotheCRRparadigmproposedbytheCAISOistointroducelimitationsonthetypesofnodeseligibletobeeitherasourceorsinkforCRRssoldintheCRRauction.Figure1providesanillustrationofthesource‐sinkcombinationsthatareproposedtobeallowedinthisproposal.Incontrast,theCAISOcurrentlyallowsauctionparticipantstobidfromanybiddablepricingpointtoanyotherbiddablepricingpoint.

34CAISOop.cit.,p.24.

17

Theproposalwouldeliminate“non‐deliverypair”bids,atermthatreferstosource‐sinkpairsthatarenotrelatedtosupplytransactions.AmongthetypesofCRRbidsthatwouldnotbeallowedwouldbe“gen‐to‐gen”CRRs,whichbothsourceandsinkatsupplynodes.35ThislattercategoryaccountsforthelargestsinglesourceofCRRcapacityawardedduringtheauctionphase.TheCAISOreportsthatgen‐to‐gentransactions“haveresultedinanoverall$186millioncongestionrevenuerightauctionnetpaymentdeficiencysince2014.”TheCAISOarguesthatmanyCRRsbetweennon‐deliverypairswouldhavelimitedvalueinhedgingresource‐to‐loadtransactions.Thisisbecause,suchCRRscouldnot,ontheirown,matchoroffsetthecongestionriskassociatedanyphysicaltransaction.Theonlywaythatnon‐deliveryCRRscouldcontributetohedgingiswhentheyarecombinedwithotherin‐struments,eitherISO‐backedCRRsorpurelyfinancial,toformadditionalhedgingproducts.WediscussthepotentialhedgingusesoftheseCRRssourcingandsinkingattheseotherlo‐cationsbelow.DiscussionThisisthemostsubstantiveandcontroversialelementoftheCAISO’sproposedchanges.Thelogicbehindthisrestrictionisstrong:non‐deliverypairCRRsaccountforthevastma‐jorityoftheauctionrevenueshortfall.36Moreover,allowingsalesofCRRssourcingandsinkingatthesenon‐deliverylocationslikelycontributessubstantiallytotheabilityoffi‐nancialmarketparticipantstobuyCRRsthatcreateflowsontransmissionelementsthatbindduringoutagesbutarenotmodeledasconstraintsintheauction.Further,thisallow‐ancecanalsoenablepurchasesofCRRsthathadsmallshiftfactorsonbindingconstraintsintheauctionmodel,yethadlargeshiftfactorsonbindingconstraintsintheday‐aheadmarketduringtransmissionoutages.Itisthereforeanticipatedthateliminatingtheauc‐tioningofsuchCRRswouldsubstantiallyreducetheauctionrevenuedeficiency.37Ideally,

35CAISO,CRRAuctionEfficiencyTrack1DraftFinalProposal,p.32.

35Ibid.

36RecentCAISOcalculationsprovidedtousindicatethat$280million,or82%,oftheoverallauctionrevenueshortfallsince2014canbeattributedtonon‐deliverypairs.Asnotedabove,generator‐to‐generatorCRRsaloneaccountfor$186million.SeeCaliforniaISO,Congestionrevenuerightsauctionefficiency,Feb13,2018pp.8,10.

37Thesechangesarealsolikelytoimproveauctionrevenueadequacy.

LAP

Intertie

TradingHub

Gen

18

suchachangewouldfocusbiddingonasmallersub‐setofCRRsthatwillcreateflowsonconstraintsmodeledintheauctionandthereforeincreasecompetitionand,inalllikeli‐hood,auctionrevenuesfortheremainingeligiblepairs.Totalauctionrevenuewouldverylikelydeclinewiththeenforcementofthesesource‐sinkrestrictionsintheauction,butitisanticipatedthatthedeclineinauctionrevenueswillbesmallerthanthedecreaseinCRRpayouts.ItisanticipatedthatmanyfewerCRRswilllikelybesoldinannualandmonthlyauctionsasaresultofthischange.Theintentionofthepro‐posalisforthereductioninthesaleofCRRstothenleadtoalargereductionintheoverallcongestionpaymentstoCRRssoldintheauction,therebybringingtheauctionrevenuesmoreinlinewiththeCRRpayout.ThedegreetowhichtheoverallCRRpayoutwillbere‐ducedisuncertain,butthereisevidencelendingsupporttotheexpectationthatthereduc‐tioninCRRpayoutswillexceedthereductioninauctionrevenues.Thefundamentaldriv‐ersofdemandforhedgingCRRsaredifferentthanthoseforspeculativeones.Itisreasona‐bletoexpectthathedgerswouldbewillingtopayapremium(overexpectedrevenues)forhedgingCRRs,whilespeculatorswouldrequireadiscountoverexpectedvaluetoaccepttheriskassociatedwithaspeculativeCRR.InotherISOs,specificCRRsassociatedwithhedging(e.g.delivery)pairsconnectingmajorsourceandsinkhubshaveconsistentlyclearedatpricesabovetheirsubsequentpayout.38WhiletheeliminationofCRRssourcingandsinkingatGennodescouldintheabstractma‐teriallyincrease,ratherthandecrease,differencesbetweenauctionclearingpricesandday‐aheadmarketcongestionpayoutsontheCRRssoldintheauction,theCAISO’sanalysishasshownthataverylargeportionofthenetpayoutistoconstraintsthatdidnotbindintheauctionandoftenwerenotevenenforced.TheGentoGenCRRbidsthereforedidnoth‐ingtoconvergeauctionpriceswithmarketpayoutswithregardstothoseconstraints.Forexample,Table23forDecember2016showsthattherewerenetpayoutstomonthlyandannualCRRsofover$8millionduetotheconstraintOMS4379177IVALLEYBNK81NG2whichwasnotenforcedintheauction.ThisoneconstraintaccountsformorethanallofthedifferencebetweenauctionrevenuesandCRRpayoutsinDecember2016,whichwas$5.7million.39Atthesametime,sellinglessCRRs,whilereducingrevenueshortfalls,mightalsoreducethevariousmarketbenefitsprovidedbyCRRsthathavebeendiscussedabove.Severalstakeholdershavenotedthewaysthatnon‐deliverypairCRRscanbecombinedwitheach‐otherordelivery‐pairCRRstocreatenewhedges.Suchflexibilitywillbereduced.How‐ever,CAISOanalysisindicatestheperceptionthattheauctionitselfproducessubstantialcounterflowthatdirectlyunlocksISO‐backhedging(delivery‐pair)CRRsisnotaccurate.

38CRRs(orequivalentinstruments)associatedwiththeNewYorkISOZoneGtoZoneJ,andthePJMwesternhubtoPECOsinkhavesoldinauctionsatpricesonaveragehigherthanpayoutsover10to15yearperiods(S.Harvey,February2018.“CRRPricesandPayOuts:AreCRRAuctionsValuingCRRsasHedgesorasRiskyFinancialInstruments”).

39CaliforniaISO,CRRAuctionAnalysisReport,November21,2017,pp.135and145.SeealsothenoteabovediscussingsimilarpatternsintheCAISOanalysisofpayoutstonon‐bindingandnotenforcedconstraintsinothermonths.

19

AccordingtotheCAISO’scalculations,eliminatinggenerator‐to‐generatorCRRs,thedomi‐nantformofnon‐deliverypairCRRs,allowedforanincreaseinthesaleofnon‐generator‐to‐generatorpairCRRsinaresimulationofthe3rdquarterof2018.Thisprovidessupportfortheargumentthatnon‐deliverypointCRRsarenotprovidingdirecthedgingvaluethroughtheauction.Whilethereisreasontoexpectthatthedirectimpactsofthischangewillproducethede‐siredeffectofreducingauctionrevenueshortfallswhilepreserving,orevenexpanding,theavailabilityofhedgingCRRsintheauction,itismuchmoredifficulttopredicttheindirectimpactsofthischange.Anumberofconcernshavebeenexpressedwiththelimitationsonsource‐sinkpairsproposedbytheCAISO.Someoftheseconcernsappeartobevalid,somewouldrequirefurtherdiscussiontounderstand,whilesomeothersappearinconsistentwiththedatacompiledbytheCAISOinitsdetailedanalysisofauctionoutcomes.First,severalmarketparticipantshavepointedoutthatthepurchaseofGenerator‐to‐gen‐eratorCRRscanbeusedtoeffectivelychangethesourceofaGeneratortoLAPorgeneratortotradinghubCRRheldbyamarketparticipant.40Thus,aCRRfromnodeAtotheTradingHubplusaCRRfromnodeBtonodeAisfinanciallyexactlyequivalenttoanodeBtoTrad‐ingHubCRR.Thepurchaseofgenerator‐to‐generatorCRRsisthereforepotentiallyusefultophysicalmarketparticipantsinadaptingtheircongestionhedgestochangesingenera‐tioncontractsoroutagesovertheyear.However,thenumberofgenerator‐to‐generatorhedgesbeingpurchasedduringtheyearexceedstheirpossibleuseforthispurposeofad‐justingcongestionhedgesbyanorderofmagnitude.Instead,thesegenerator‐to‐generatorCRRsaccountforalargeportionofthepayoutinexcessofCRRauctionvalues.TheCAISOwouldlikelybeabletoaccommodatethesaleofsmalleramountsofgenerator‐to‐generatorCRRsinthelongrunbyimplementingsomerestrictionsonallowedpurchases,41butthesechangeswouldneedtobeimplementedinfutureauctionsaftertheshort‐runchangeshavebeenimplementedandthecurrentsubstantialimbalancebetweenauctionrevenuesandCRRpaymentscorrected.Second,therehasbeenanopinionexpressedthatgenerator‐to‐generatorCRRscanbeusedtoreduceexposuretoparticularnodesinthecurrenttradinghub.42Itisnotclearwhythisisdesirablefromahedgingperspectivefortransactionsatthetradinghub.Ifthereisaneedforthedevelopmentofnewtradinghubswithsomewhatdifferentdefinitionsthanthecurrenthubs,thatisadiscussiontheCAISOcouldhavewithstakeholdersperhaps

40See,forexample,DCEnergy,CommentsonDraftFinalProposal,February28,2018;FinancialMarketersCoalition,CommentsonDraftFinalProposal,February28,2018;andAppianWayEnergyPartnersCom‐ments,February28,2018.

41TheCAISOcouldimplementbidvalidationthatwouldonlyallowaCRRholdertosubmitagenerator‐to‐generatorCRRbidforaCRRsinkingatthesourceofageneratortoaLAP,ageneratortoatieorageneratortotradinghubCRRheldbythemarketparticipantsubmittingthebid.Suchadesignwouldallowthepurchaseofgenerator‐to‐generatorCRRstorestructureexistingCRRswithoutallowingtheuseofgenerator‐to‐genera‐torCRRstocreateflowsonconstraintsnotmodeledintheauction.

42SeeCommentsofCalpineEnergySolutionsLLC,February28,2018.

20

followedbyimplementationadditionaltradinghubsinfutureauctions.However,thatwouldbeadiscussionbestheldafterthechangesproposedbytheCAISOhavebeenimple‐mented.Third,ithasbeenobservedthatgenerator‐to‐generatorCRRscouldbeusedtocreateso‐called“dirty”hedges,i.e.asetofCRRsthatdonotperfectlyhedgecongestionbetweenare‐sourceandaLAPortradinghubbutprovideapartialhedge,ifincrementalgeneratortoLAPCRRscannotbeawarded.43Thisispossible,butthepatternsofCRRflowsonunen‐forcedandnon‐bindingconstraintsportrayedintheCAISOreportsuggeststhatthepri‐maryfocusofthesegenerator‐to‐generatorCRRsisnothedging,andthereisnoevidencethattheseCRRshavematerialvalueashedgesgiventhattheyaresoldatalargediscounttotheexpectedpayout.OncethechangesproposedbytheCAISOhavebeenimplementedandthecurrentsubstantialimbalancebetweenauctionrevenuesandCRRpaymentscorrected,considerationcouldbegiventowhetherthereisaneedtomakeadditionalchangesthatmightenableincreasedhedging.AfourthconcernwiththeCAISOproposalisthatitwouldpreventmarketparticipantsfrombiddingtoprovidecounterflowCRRs,forexamplefromatradinghubtoageneratororfromaLAPtoagenerator.44Thiswillbethecase,butitisanecessarypartoftheCAISO’sdesign.IftheCAISOweretoallowthepurchaseofsuchcounterflowCRRs,marketpartici‐pantscouldconstructgenerator‐to‐generatorCRRsbypurchasingaCRRfromageneratortoatradinghubandthenbuyingacounterflowCRRfromthetradinghubtoanothergener‐ator.Ifbothbidscleared,thetwoCRRswouldbecompletelyequivalenttoholdingagener‐ator‐to‐generatorCRR.WhiletheneedtosubmitseparateforthetwoCRRswouldcompli‐catebidding,45allowingthesebidswouldgiverisetothekindofconcernsDMMexpressedaboutbiddingstrategiesthatwouldcircumventtheintendedrestrictions.46Sincethefun‐damentalcharacteristicofgenerator‐to‐generatorCRRsisthattheauctionpriceismuchlessthantheCRRpayout,thereisnotmuchincentiveformarketparticipantstobuycoun‐terflowCRRsandtakenonapositionforwhichtheywouldbepaidlessthantheywouldpayout.TheCAISO’sresimulationofaseasonofCRRauctionsfoundthattherewere3800megawattsofcounterflowCRRsbuttheoverallimpactofnon‐deliveryCRRswastoreducethenumberofCRRsthatclearedbetweentheprimarylocationsforsupplyhedges.47AftertheCAISOimplementstheseTrack2reformsandobservesthedegreeofimprovedconvergencebetweenCRRauctionpricesandday‐aheadmarketpayouts,itwillbeableto

43SeeNRGEnergyInc,CommentsonDraftFinalProposal,February28,2018.

44See,forexample,AppianWayEnergyPartnersComments,February28,2018.

45Ifthebiddersubmittedtwoseparatepricecappedbids,onemightclearwhiletheotherbidnot,leavingthebidderwithapositionitdidnotintendtotake,whileifthebiddersubmittedpricetakingbidsitcouldenduppayingmorethanitintendedfortheCRRiftheCAISOenforcedaconstraintthatthebidderdidnotexpect.

46SeeCAISODepartmentofMarketMonitoring,CommentsontheCongestionRevenueRightsAuctionEffi‐ciencyTrack1DraftFinalProposal,February28,2018.

47“CongestionRevenueRightsAuctionEfficiency,Track1DraftFinalProposal,”February8,2018,p.35.

21

considerrefinementsinbiddingrulesthatcouldbeimplementedonalongertimeframethatcouldallowthesubmissionofsomecounterflowCRRbidwithoutunravelingtheef‐fectsofthesource‐sinkrestrictions.Othercommentershavepointedtothesource‐sinkrestrictionsproposedbytheCAISOwillmakeitmoredifficultorcostlytohedgeunspecifiedcongestionexposures.48Itisantici‐patedthatthesource‐sinkrestrictionsproposedbytheCAISOwillmakeitmoredifficultforCRRpurchaserstohedgecongestiononparticulartransmissionelements(suchasthosethatarenotmodeledintheauction).However,thatinabilitydoesnotpreventmarketpar‐ticipantsfrombiddingforCRRsthathedgethecongestionrisksthateitherphysicalmarketparticipantsorthosesellingthemfinancialhedgeswouldincur.Commentershavealsosuggestedthatgenerator‐to‐generatorCRRshelpsCRRauctionpricesreflectconstraintcostsbuttheCAISOdatainSection7showsthatmuchoftheCRRpayoutistoconstraintsthatarenotenforcedinauction,sogenerator‐to‐generatorbidscannotcausethemtobindandimpactprices,buttheydoincreasetheCRRpayout.Inaddi‐tion,theCAISOanalysisinSection7oftheNovemberCRRAuctionAnalysisReportshowsthatmanyoftheotherconstraintsthataccountedforlargepayoutsdidnotbindintheauc‐tion,sotheabilityofauctionparticipantstosubmitgenerator‐to‐generatorCRRbidsdidnotcausetheexpectedpayouttotheseconstraintstobereflectedinconstraintprices.Anumberofcommenterssuggestthatinsteadofimposingthesource‐sinkrestrictionspro‐posedbytheCAISO,theCAISOshouldlimittheauctionchangestoeliminatingthesaleofCRRsbetweenessentiallyidenticallocations.49Whilesuchchangesshouldbemade,theCAISO’sanalysishasshownthattheirimpactonthedifferencebetweenauctionrevenuesandCRRspayoutswouldhavebeenrelativelylimited.50Tosummarize,CAISOanalysisindicatesthatthemajorityoftheauctionrevenueshortfallsarepresentlyattributabletothetypesofCRRsthatwouldbeeliminatedunderthispro‐posal.SeveralstakeholdershaveraisedconcernsthatthisreductionintheallowedsourcesandsinksforCRRscouldleadtounintendedconsequencesthroughchangestobiddingbe‐haviorandtofinancialtransactionsoutsideofCAISOmarkets.Theremaybesomesecond‐aryimpactsfromthesechanges,buttheCAISOanalysisshowsthatwhateverthebenefitsthatmaybeforegonebecauseoftheseunintendedconsequences,thereisaverylargein‐tendedconsequenceoftakingafirststeptowardsreducingandeventuallyeliminatingthediscrepancybetweenauctionrevenuesandCRRpayouts.Moreover,wenotethatthischangeisalessdramaticchangetothelandscapeofauctionedCRRsthanthesolutionssup‐portedbyDMMandtheinvestor‐ownedutilities.

48See,forexample,CommentsofBostonEnergyTradingandMarketing,February28,2016.

49Seeforexample,AppianWayEnergyPartnersComments,February28,2018,NRGEnergyInc,CommentsonDraftFinalProposalFebruary28,2018;DCEnergy,CommentsonDraftFinalProposalFebruary28,2018;CommentsofFinancialMarketersCoalition,February28,2018.

50WeunderstandfromtheCAISOthatwhiletheseCRRsaccountedforaround12%oftheauctionrevenueshortfallin2015,theyaccountedforlessthan5%in2016andlessthan0.5%in2017.

22

Conversely,themagnitudeofthereductioninthedifferencebetweenauctionrevenuesandCRRpayoutsandtheincreaseincongestionrentsflowingtotransmissionratepayersthatwillbeachievedbytheCAISOdesignisalsouncertainasitdependsinpartonhowCRRauctionparticipantschangetheirbehaviorinfutureauctions.Therefore,whilethefullef‐fectsofthischangecannotbeunderstoodwithexantesimulationalone,wesupportthismeasureastargetingtheCRRsthataretheleastlikelytoprovidemarketbenefitsandthemostlikelytoinduceauctionrevenueshortfalls.Aswithallofthemeasurescurrentlypro‐posedbyCAISO,wesupportanin‐depthevaluationoftheirimpactsandpursuitoffurther,orevenreversalofthese,changesaswarrantedbytheanalysis.TheimpactoftheDMM/SCEproposalandthecongestionrentsflowingtotransmissionratepayersisalsouncertainbecausethisdependsnotonlyonauctionrevenuesandCRRpayoutsbutalsoonthecongestionrentscollectedintheday‐aheadmarket.ChangeswhichhindertheabilityofexternalsupplierstohedgeforwardsalesatCAISOtradinghubswilllikelyimpactfor‐wardsalespatternsaswellasCRRpurchasesandbyreducingimportscouldbothreducecongestionrentcollectionsandraiseday‐aheadmarketprices.

5. ConclusionsAtthisstage,therearethreebroadpolicydirectionsthattheCAISOcouldtake:

1) Continuethestatusquowithonlyminormodificationstotheauction,2) Eliminatetheauctioninitscurrentform,leavingitasonlyatradingplatformfor

“voluntary”transactions(asintheDMM/SCE)proposal,or3) Makesubstantialrevisionstotheauctionstructurethattrytomaintainitshedging

benefitswhilesignificantlyreducinglossestotransmissionratepayers.Theserevi‐sionscouldeitherbeintendedtobepermanent,orcouldinsteadbeprovisional,in‐tendedtoreducelosseswhiletheCAISOfurtheranalyzespossiblechangesandde‐velopsamorefinelytunedproposalthatmayincludefurtherrestrictionsonauctionsalesinsomerespects,butallowmoreauctionparticipationinotherrespects.

TheCAISOproposalfallsinthethirdcategory.Thechangestotheauctionarefarmorethanminoradjustments.However,theypreservetheabilityofnon‐LSEstodirectlyaccesstheISO‐backedCRRmarketinordertohedgesalestoloadservingentitiesatLAPsortrad‐inghubsbypurchasingCRRssourcingatgeneratorsorthetiesandsinkingatLAPsandtradinghubs.Theyalsomaintainsomeoftheotherflexibilityattributesprovidedbythecurrentauctionprocess,suchasexporting,wheeling,andevennon‐contractedmerchantgeneration.Thechangescanalsobeviewedasprovisional,andcouldbesupplantedorsupplementedbychangesaspartofTrack2oftheCRRauctionreformsorsubsequentpro‐posals.Wesupportthisproposalforthedetailedreasonsprovidedabove,whicharesum‐marizedbelow.Atthistime,wedonotsupporttheDMM/SCEproposalforthereasonsprovidedinSection2.3.Inparticular,itwouldbecountertotheopenaccessprinciplesthatmotivatedthecre‐ationofcongestionrevenuerightsasahedgeinthefirstplace;replacementhedgeswould

23

likelybeavailableonlyatamuchhigherpricesformarketparticipantswhodonotpartici‐pateinthefreeallocationstageofCRRallocation;andcautionshouldbetherulewhencon‐sideringmarketchangesthatwouldprofoundlyaffecttheavailabilityandcostoftransmis‐sionhedgingservices.IftheTrack1and2changesprovetobeineffectualinreducingCRRauctionlosses,thentheDMM/SCEproposalisonealternativethatcouldbeconsidered.Theauction,asitiscurrentlyimplemented,hasproducedarevenueshortfallthathasrangedfrom50–75$millionoverthelastthreeyears.Totalcongestionrevenuesduringthisperiodhavebeenseveraltimesaslarge.Atthesametime,CRRsareviewedbymany,includingourselves,asprovidingbenefitstotheoperationandefficiencyofwholesalemar‐kets.WhileitisextremelydifficulttoquantifytheexactcontributionthatauctionedCRRsprovidetothemarket,itdoesnothavetobesubstantialtomakeadifference.Thetotalwholesalecostofservingload,ascalculatedbyDMM,rangedfrom7.5to12.1$billionan‐nuallybetween2014and2016.TheCAISO’sCRRauctionsneedtocontributelessthanonepercenttowholesalemarketefficiencytooffsettheentireauctionrevenueshortfall,therebyactuallyproducinganetbenefittoratepayers,andwouldbeabletoprovidealargernetbenefitwhentheauctionrevenueshortfallisreducedbytheCAISOTracks1and2proposals.WhileitisextremelydifficulttoquantifytheimpactsofthechangesproposedbytheCAISOinTrack1,wenotethatqualitatively,theyareconsistentwiththegoalofpreservingaccesstocongestionhedginginstrumentsforallloadservingentities,includingthesmallerloadservingentitiesthatappeartobemorelikelytopurchasepowerattradinghubs,whileminimizingratepayerlosses.Mostplausiblehedgingtransactionsrequirerightsconsistentwiththephysicalflowofpowerorwithfinancialsupplydeliverabilityattradinghubs,andthoseinstrumentswillbepreservedundertheproposal.WhileitispossiblethatothertypesofCRRsplaysomeroleinsupportinghedging,aswellasspeculation,weknowofnoreliablewaytoquantifythecontributionsthesetypesofCRRsmaketothehedgingfunc‐tion.TheCAISOauctionanalysisdoesshowus,however,thattheserightshavebeenre‐sponsibleforthelargestnetrevenueshortfallsoverthelastseveralyears.Bycontrast,anearlierproposaltoreduceoverallcapacityavailableintheannualauctiontargetsalltypesofCRRs,andstrikesusasmorelikelytohavelimitedtheavailabilityofhedgingaswellasspeculativeinstrumentsintimeframesrelevanttosomeforwardcontracts.Furthermore,limitingthecombinationsofCRRsshouldtheoreticallyconcentratetheflowsoverfewerconstraintssotheycouldincreasepricesrelativetopayouts.Whiletheproposalisqualitativelyastepintherightdirection,theexactquantitativeim‐pactsareuncertain.Itisuncleartoushoweffectivethesource‐sinkrestrictionswillbeinlimitingthesaleofCRRsthathavelittlevalueashedgesandaresoldintheauctionatlowpricesrelativetotheexpectedpayout.Byhowmuchthesource‐sinkrestrictionswillim‐provethebalancebetweenauctionpricesandCRRpayoutsisuncertain.Nevertheless,thesechangesare,webelieve,likelytoreducetheauctionrevenueshortfallwithoutsub‐stantiallyharmingmarketefficiency,andtheCAISOcanundertakeadditionalanalysistoguideadditionalchangesoverthecomingweeks.Whiletheannualauctionsfor2019willbeginthissummer,thelargestdifferencesbetweenauctionrevenuesandCRRpayoutsare

24

inthemonthlyauctionsandadditionaldesignchangescouldbeimplementedinthoseauc‐tionspriortotheannualauctionsfor2020.EveniftheCAISOTrack1proposaliseffectiveinreducingpurchasesofCRRsatlargedis‐countstotheexpectedpayoutbyfinancialmarketparticipants,itstillmaynoteliminateallorevenmostofthedifferencebetweenauctionpricesandCRRpayouts.Themarketstruc‐tureofretailsupplyinCalifornia,combinedwiththeaggregationofdemandnodepricingtoloadaggregationpoints,couldbecontributingtorelativelylowdemandforhedgingin‐struments,atleastbyLSEs.Totheextentthatlowauctionprices,andauctionrevenueshortfallsareduetolowhedgingdemand,theCAISOchangeswillnotchangetheseout‐comesastheyarenotaresultoftheCAISO’sauctiondesign.WenotethatthisproposalisonlyafirststepinalongerprocessofevaluationoftheCRRauctionanditsbroadermarketimpacts.Analysisoftheeffectsoftheproposedchanges,someofwhichhavenotbeenexperiencedinUSmarketsbefore,willprovidevaluablein‐sighttowardwhetherfurtherchangesarenecessary,orifelementsoftheproposalshouldberevisited.Asaninterimstep,therefore,webelievethatthisoptionmakesconsiderablesenseasawaytoreducethesaleofundervaluedCRRsintheshort‐runandprovidesomeevidenceofhowsomekindofauctionmodificationsimpacttherelationshipbetweenauc‐tionvalueandCRRpayouts.Therefore,wesupporttheCAISOproposalasareasonablein‐crementalstep,thatstandsagoodchanceoflimitingauctionrevenueshortfallswhilestillpreservingthemainunderlyingfunctionoftheCRRauction.Thesechanges,togetherwithanyTrack2changesthatareimplemented,willneedtobegivensomereasonableperiodoftimetobeinplacebeforetheirsuccessisevaluated.AspartoftheTrack2stakeholderprocess,westronglysuggestthat,first,thatawiderrangeofalternativesforreducingauc‐tionlossesbeconsidered.OnealternativeshouldincludeestablishingaminimumpriceorperunitfeeforauctionedCRRs.DataprovidedtousbyISOstaffindicatethat7%oftheauctionlossesinthepastfouryearswerecomprisedofpaymentsmadetoCRRsthatsoldatazeroprice.51Anotherwouldbetolookatthestructureforfundingtheauctionrevenueshortfall.Wenotethattheproblemthatisthefocusinthisinitiative‐‐theauctionrevenueshortfall‐‐isfundamen‐tallyatransferpaymentissuefromonesetofmarketparticipantstoanother,incontrast,theproposalstoaddressthistransfercouldhaveadditionalnegativeimpactsontheeffi‐ciencyofthemarket.Whiletherearereasonstobelievethattheefficiencyimpactsassoci‐atedwiththeCAISOproposalwouldbeacceptable,thoseassociatedwiththeDMM/SCEproposalcouldbemoresubstantial.Ifthesourceoftheproblemconstitutesanunaccepta‐bletransferfromonegrouptoanother,thesolutioncouldinvolveanothertransferorfeethatwouldoffsettheselosseswithoutsignificantlycompromisingtheCRRmarket.OursecondsuggestionisthatcarefulanalysesbemadeofpotentialCRRauctionrevenuesrelativetopayouts,categorizedbysource‐sinkpair,underalternativeauctiondesignsforthe2014‐2017period.Thiswouldbestbedonebyrunninghistoricalbidsthroughimple‐mentationsofalternativeauctions(includingeliminationofcertainsource‐sinkpairs,

25

reductionsinannualauctionquantitiesawarded,andminimumpriceorbidfees).Thisshouldbedonefortheentirefouryearperiodinordertominimizetheimpactofsampleerrorupontheconclusions.Wenotethattheresultsofsuchanalysesshouldbeinter‐pretedcarefully,becausebiddingbehaviormaychangeasaresultofreformingtheauction,sothattheanticipatedreductionsinlossesmightnotberealized.Nevertheless,thesesim‐ulationsoftheperformanceofalternativeauctiondesignsusinghistoricalbidswillprovideabetterindicationofthepotentialreductioninlossesthantheanalysesthatarepresentlyavailable.AthirdsuggestionisthattheCAISOanalyzetheextenttowhichthereisageneralunder‐valuationofhedgingCRRsinCAISOmarkets,ratherthansimplyalowvaluationofCRRsthathavelittlevalueascongestionhedges.Fourth,astheTrack1changesareimplemented,theCAISOshouldassesstheextenttowhichthesechangeshavebeeneffectiveinreducingthepayouttoCRRswhoseshiftfactorsandday‐aheadmarketpayoutsareinflatedbyoutages,andconsiderwhetherchangesinthewayCRRsaresettledmightcontributetoimprovedoutcomes.Tosummarize,wesupporttheCAISOproposal,butanticipatethatfurtherchangeswillbenecessarytoeithersupplementorsupplantthoseproposedhere.Whateverchangesaremadeshouldcontinuetosupporttheabilityofsmallandnon‐LSEstoaccessamarketforISO‐backedCRRs.Wesuggestseveralanalysesthatcanquantifytheimpactoftheseandotherpotentialchanges,andadvisethesebeundertakenaspartoftheTrack2process.