y Ohio Energ Luncheon Keynote:
Ohio Energy Luncheon Sponsored by FirstEnergy Solutions Luncheon Keynote: Competitive Electricity Markets Spur Economic Growth & Create Jobs Jonathan A. Lesser, Ph.D., President, Continental Economics, Inc., Sandia Park, NM Wednesday, February 20, 2013 12:45 p.m. to 1:30 p.m. JONATHANA.LESSER,PHD Dr.JonathanLesseristhePresidentofContinentalEconomics,Inc.,andhas almost30yearsofexperienceworkingforregulatedutilities,government, and as an economic consultant. He has analyzed economic, policy, and regulatoryissuesaffectingtheenergyindustry,includingmarketstructure anddesign,environmentalandregulatorypolicy,renewableenergypolicy, economicimpactsofenergyinvestment,andutilityfinancingandthecost ofcapital. Dr. Lesser has extensive experience in valuation and damages analysis, from estimating the damages associated with breaking commercial leases to valuing nuclear powerplants.Dr.Lesserhasperformedduediligencestudiesforinvestmentbanks,testifiedon generating plant stranded costs, assessed damages in commercial litigation cases, and performed statistical analysis for class certification. He has also served as an arbiter in commercialdamagesproceedings. Dr.Lesserhasprovidedexperttestimonyonmanyelectricandnaturalgasregulatoryissueson thecostofcapital,costofservice,costallocation,ratedesign,anddepreciationincasesbefore state utility commissions; before the Federal Energy Regulatory Commission (FERC); before internationalregulatorsinLatinAmericaandtheCaribbean;aswellastestifiedincommercial damages cases in U.S. federal and state courts. He has also testified on energy policy and marketdesignbeforestatelegislativecommittees.Dr.Lesserhasalsoservedasanindependent arbiter in disputes involving regulatory treatment of utilities and valuation of energy generationassets. Dr.Lesseristheauthorofnumerousacademicandtradepressarticles.Heisalsothecoauthor of Environmental Economics and Policy, published in 1997 by Addison Wesley Longman, Fundamentals of Energy Regulation, published in 2007 by Public Utilities Reports, Inc., and Principles of Utility Corporate Finance, published in 2011 by Public Utilities Reports, Inc. Dr. Lesser is also a contributing columnist and Editorial Board member for Natural Gas & Electricity. 6 Real Place • Sandia Park, NM 87047 • main: 505.286.8833 • DC Office: 202.446.2062 www.continentalecon.com www.competecoalition.com STATE SUBSIDIZATION OF ELECTRIC GENERATING PLANTS AND THE THREAT TO WHOLESALE ELECTRIC COMPETITION Report prepared for: COMPETE Coalition Prepared by: Continental Economics, Inc. December 2012 Copyright © 2012, Continental Economics, Inc. The information contained in this document is the exclusive, confidential and proprietary property of Continental Economics, Inc. and is protected under the trade secret and copyright laws of the U.S. and other international laws, treaties and conventions. No part of this work may be disclosed to any third party or used, reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage or retrieval system, without prior express written permission of Continental Economics, Inc. TABLEOFCONTENTS ExecutiveSummary I. Introduction....................................................................................................................................1 II. WhyCapacityMarkets?...............................................................................................................3 A. CapacityMarketOpposition..................................................................................................................5 III. HowGovernmentSubsidizationofGenerationHarmsMarketsandConsumers..6 A. BuyerMarketPower—NoSuchThingasaFreeLunch............................................................7 B. TheBriggs‐KleitModel:EffectsofGovernmentSubsidizedEntry intoCapacityMarkets...............................................................................................................................8 IV. ConclusionsandPolicyImplications....................................................................................10 Appendix1:TheBriggs‐KleitMultipleMarketModel Appendix2:ReviewofElectricRestructuringandCreationofCompetitiveWholesale Markets Appendix3:TheNewJerseyandMarylandPrograms EXECUTIVESUMMARY “Theexpectationofgovernmentinterventioncreatesaself‐fulfillingconditionthatincreases costsanddrivessuppliersfromthemarket” Competitive wholesale electricity markets in the U.S. have shown themselves to be an economic success. Today, more than 40% of total generating capacity is owned and operatedbyindependent,competitivegenerators.1And,unlikemonopolyelectricutilities, whose ratepayers must bear the financial risks of investment decisions, competitive generatorsbearthoserisksthemselves. Because electricity cannot be stored cost‐effectively, ensuring reliable electric service requires that reserve capacity be available to meet unforeseen events, such as sudden increasesindemandorunexpectedgeneratoroutages.Toprovidethesecapacityreserves, a number of competitive market solutions for capacity have been developed. One such solution is a forward capacity market that uses an auction to secure commitments for sufficientcapacityresourcesintothefuture.ThePJMInterconnection(thegridoperatorin themid‐Atlanticstates)administersthelargestsuchmarket,whichiscalledtheReliability Pricing Model (RPM). The goal of capacity markets like RPM is to provide economic incentivesthatensurethereisenoughcapacityneededtomeetreliabilitystandardsatthe lowestpossiblecost.2 Some opponents have argued that capacity markets are not working properly because higher capacity prices in their states—typically, the result of sub‐regional transmission constraints—havenotincentedtheconstructionofnewcapacityresources.Consequently, capacity market opponents conclude that direct government intervention in capacity markets is required to develop local generating resources, despite the regional nature of theinterstatewholesalepowermarkets.3 This paper examines the impact of such intervention in the form of mandated state government subsidies for new generation resources. Using the results of recently published work by the Pennsylvania State University Electricity Markets Initiative,4 we concludethatgovernmentsubsidiesfornewgenerationresourcesbothraisecapacitycosts for the very customers whom the subsidies are supposed to benefit and jeopardize resourceadequacyandreliabilityinthelongrunforallconsumers. 1 Source:U.S.EnergyInformationAdministration,ElectricPowerAnnual2010,November2011,Table1.3. 2 ForanintroductiontoRPM,seeCOMPETECoalition,“KeepingtheLightsOntoPowerOurFuture:“RPM”, PJM’sReliabilityProgram,”March2012. http://www.competecoalition.com/files/Keeping%20the%20Lights%20On%20to%20Power%20Our% 20Future%20‐%20FINAL%2031212.pdf. 3 See,forexample,2011NewJerseyEnergyMasterPlan,pp.21‐22. 4 R.BriggsandA.Kleit,“ResourceAdequacyandtheImpactsofCapacitySubsidiesinCompetitive ElectricityMarkets,”WorkingPaper,Dept.ofEnergyandMineralEngineering,PennsylvaniaState University,October22,2012(BriggsandKleit). http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2165412 EX‐1 The new work by Professors R.J. Briggs and Andrew Kleit finds that consumers in states wheresubsidiesaremandatedreceivelessoftheinitial“benefits”fromstategovernment intervention in capacity markets than customers in surrounding states. However, these “benefits” quickly disappear, as government intervention drives out otherwise economic existinggenerationandhindersthedevelopmentofnewresourcesinallstateswithinthe market.Thus,whengovernmentintervenesonbehalfofonegeneratoritdrivesoutother generators, taking with it not only competitive generation capacity, but also the jobs and taxbaseassociatedwithgenerationthatexitsthemarket.Mostimportantly,theyfindthat the adverse long‐run impacts in all states far outweigh any short‐term “benefits” of temporary price reductions. Their work demonstrates that mandates by state governmentstosubsidizeinvestmentingenerationresourcesareneveroptimalandresult inlossestoallconsumers. MostU.S.electricmarketstypicallyextendbeyondindividualstates,operatingregionallyto achieve economies of scale and scope. As a result, individual state government intervention, like the recent government actions in Maryland and New Jersey,5 imposes long‐run costs on neighboring states. Such “beggar‐thy‐neighbor” policies are not only counterproductive, they also invite policy “retribution” that will further damage competitivemarkets.Forexample,notingthatsignificantnewgenerationinvestmenthad taken place in Pennsylvania, that state’s public utility commissioners expressed concern that [t]he ability to bid in new capacity at potentially artificially low prices can skew the capacity market leading to less investment in new and existing capacity,includinginPennsylvania.Withoutsuchinvestment,theendresult from the consumer's perspective, ultimately, could be higher rates in Pennsylvaniathanwithoutthisstate‐mandatedsubsidy.6 It is a basic economic fallacy that price distortions caused by subsidies from government interventioninafreemarketare“benefits.”Therealityissuch“freelunch”policiesnever work,becausetheyincorrectlyassumethatsupplierswhoseelowerpriceswillnotchange theirbehavior. Perversely, when a state government intervenes in a regional capacity market so as to benefitconsumers,itactuallyforcesthosesamein‐stateconsumerstosubsidizeshort‐run benefitsforneighboringstates’consumers.Oneconsequenceofsuchinterventionisthat merchantgenerationwillnotinvest,eitherinnewresourcesorexistingresources,forfear that governments will intervene and eliminate the ability to realize a needed economic return on the investment that compensates developers for the risks they take. In other words, even the expectation of government intervention createsa self‐fulfilling condition thatincreasescostsanddrivessuppliersfromthemarket.Thisharmsallconsumersand, in turn, will harm long‐run grid resource adequacy. Eventually, the only suppliers in a market will be subsidized ones, and the market will cease to exist, eliminating the real benefitsthatcompetitivemarketsprovide. 5 Appendix3providesadescriptionofthegenerationsubsidyprogramsinMarylandandNewJersey. 6 LettertothePennsylvaniaCongressionaldelegationfromthePennsylvaniaPublicUtilityCommissioners, July13,2011. EX‐2 StateSubsidizationofElectricityGeneration I. December2012 Introduction CompetitivewholesaleelectricitymarketsintheU.S.havebeenaneconomicsuccess.With passageoftheEnergyPolicyActof1992,Congressunleashedcompetitiveelectricmarkets, breaking the virtual monopoly on utility‐owned and operated generation.1 Today, more than40%oftotalgeneratingcapacityisownedandoperatedbyindependent,competitive generators.2 Competitive generators have also improved their reliability and operating efficiency more than their fully‐regulated, utility counterparts.3 And, unlike monopoly electric utilities, whose ratepayers bear the financial risks of investment decisions, competitivegeneratorsbearthoserisksthemselves. Because electricity cannot be stored cost‐effectively, the assurance of reliable electric service requires that reserve capacity be available to meet unforeseen events, such as sudden increases in demand or unexpected generator outages. To provide such capacity reserves, a number of competitive markets for capacity have been developed. Unlike competitive wholesale electric energy markets, whose prices fluctuate each hour, competitivecapacitymarketsarelonger‐terminnature.PJMInterconnection4operatesthe largest such market called the Reliability Pricing Model (RPM). The goal of capacity marketslikeRPMistoprovideeconomicincentivesthatensurethereisenoughcapacityto meetreliabilitystandardsatthelowestpossiblecost.5IndependentreviewsofRPMhave found that it has worked well, adding over 13,000 Megawatts (MW) of new generation since2007.6 Someopponentsofforwardcapacitymarketshavearguedthattheyprovideawindfallfor generatorsattheexpenseofcustomers,especiallyforbaseloadgeneratorslikenuclearand coal‐fired power plants, without providing a more reliable electric system. In essence, 1 Appendix2providesabriefhistoryofelectricrestructuring. 2 Source:U.S.EnergyInformationAdministration,ElectricPowerAnnual2010,November2011,Table1.3. 3 Seee.g.,P.Joskow,“MarketsforPowerintheUnitedStates:AnInterimAssessment,”TheEnergyJournal 27(January2006),pp.1‐36.http://econ‐www.mit.edu/faculty/download_pdf.php?id=1219;N.Rose,K. Markiewicz,andC.Wolfram,“DoesCompetitionReduceCosts?AssessingtheImpactofRegulatory RestructuringonU.S.ElectricGenerationEfficiency,”MassachusettsInstituteofTechnology,Centerfor EnergyandEnvironmentalPolicyResearch,04‐418WP,November2004. http://web.mit.edu/ceepr/www/2004‐018.pdf. 4 PJMInterconnectionisthelargestregionaltransmissionorganizationintheUS,covering13statesand theDistrictofColumbia. 5 ForanintroductiontoRPM,seeCOMPETECoalition,“KeepingtheLightsOntoPowerOurFuture:“RPM”, PJM’sReliabilityProgram,”March2012. http://www.competecoalition.com/files/Keeping%20the%20Lights%20On%20to%20Power%20Our% 20Future%20‐%20FINAL%2031212.pdf. 6 TheBrattleGroup,“SecondPerformanceAssessmentofPJM’sReliabilityPricingModel,”August26,2011. http://www.pjm.com/~/media/committees‐groups/committees/mrc/20110818/20110826‐brattle‐ report‐second‐performance‐assessment‐of‐pjm‐reliability‐pricing‐model.ashx. ‐1‐ StateSubsidizationofElectricityGeneration December2012 these opponents believe (wrongly) that they had previously been receiving capacity “for free,”andthatpayingforcapacitythroughaformalmarketmeansthattheyjustpaymore. Of course, that was never true. Before competitive electric markets existed, utilities still maintainedcapacityreserves,whichratepayerspaidforintheirmonthlybills.Thecostof capacitywasneithertransparentnorsubjecttocompetitivemarketforces,butitcertainly existed. Other opponents have argued that formal capacity markets are not working properly. Theseopponentsarguethathighercapacitypricesintheirstates—typically,theresultof transmission constraints caused by higher costs within the various regions—have not incented the construction of new capacity resources. Consequently, they conclude that directgovernmentinterventionincapacitymarketsisrequired.7 Insomestates,suchasMarylandandNewJersey,thisinterventionhastakentheformof requiring local distribution (i.e., “poles and wires”) utilities to either build generating capacity themselves, or auction off the right to build new capacity to independent developers.8 Whatever the specific form of intervention, the costs of the newcapacity in these state‐subsidized efforts are underwritten by the local “poles and wires” utility’s customers,thuseliminatingthenormalfinancialriskthatcompetitivegenerationsuppliers bear. (In fact, allocating financial risk back to generation suppliers was one of the key purposesofelectricindustryrestructuring.)Thenewgeneratingcapacityisthenoffered into the capacity andenergy markets so as to purposefully “suppress” market prices and supposedly“benefit”consumers. However, it is a basic economic fallacy that price distortions caused by government subsidies in a free market are “benefits.” The lower prices made possible by subsidized entry are short‐lived because they drive competitive suppliers from the market. For example,inanOctober2011hearingbeforetheNewJerseyBoardofPublicUtilities,Zamir Rauf, the Chief Financial Officer of Calpine Corporation—the largest developer of independentlyownedandoperatedgenerationinthecountry—testifiedthat:“[O]neofthe biggest risks we currently face is the regulatory uncertainties created by various states' interest in trying to jump‐start the process of developing new capacity via long‐term contracts. … [i]t will be exceedingly difficult for merchant projects to compete with ratepayersubsidizedprojects.”9 Inthisreport,wefocusontheeffectswhengovernmentsmandateconsumerpaidsubsidies ofgeneratingcapacitythatcanbeusedtosuppresspricesincapacitymarkets.Thereality is that such policies never work: they are a form of ‘free lunch’ economics that fails to account for market dynamics. This report provides a non‐technical introduction to the 7 See,forexample,2011NewJerseyEnergyMasterPlan,pp.21‐22. 8 Appendix3providesadescriptionofthegenerationsubsidyprogramsinMarylandandNewJersey. 9 IntheMatteroftheBoard’sInvestigationofCapacityProcurementandTransmissionPlanning,DocketNo. E011050309,NewJerseyBoardofPublicUtilities,Transcript,October14,2011,pp.135‐136. ‐2‐ StateSubsidizationofElectricityGeneration December2012 important work of Briggs and Kleit (2012),10 who demonstrate how such subsidized capacity actually imposes the greatest economic harm on intervening states’ own electricityconsumersandleadstohigherlong‐runelectricitypricesforconsumersinthese statesandneighboringones. Briggs and Kleit demonstrate that the adverse long‐run impacts in both markets far outweigh any short‐term “benefits” of temporary price reductions. They show that even theexpectationofgovernmentinterventioncreatesaself‐fulfillingconditionthatincreases costsanddrivessuppliersfromthemarket.Eventually,theonlysuppliersinamarketwill besubsidizedones,andthemarketwillceasetoexist,harmingallconsumersandlong‐run grid resource adequacy. This violates one of the founding tenets of electricity restructuring: less government intervention on supply and price, and more reliance on market signals to attract investment. As Briggs and Kleit conclude, “[i]n the context of restructured electricity markets, subsidized additions of base capacity by state governments are, at best, a costly and undue burden for taxpayers. At worst, these subsidieshavetheperverseeffectofreducingtheincentivesforresourceadequacyinthe longrun.”11 II. WhyCapacityMarkets? As discussed previously in the introduction, because electricity cannot be stored cost‐ effectively, ensuring the reliability of the power system requires there to be reserve marginstoaddresscontingencies.Inregionswhereelectricmarketswererestructuredto allowmarketincentivestoimproveoperatingefficiency,betterallocatefinancialrisk,and reduce costs, the old regulatory model of having vertically integrated utilities meet their reserve requirements by building their own generating resources no longer applies. Instead,restructureddistributionutilitiesbecame“polesandwires”companiesthatobtain generationfromcompetitivewholesaleelectricmarkets. Inoverseeingwholesaleelectricmarkets,FERChassoughttoensurethatgeneratingunits needed primarily for maintaining reserve margins could remain economically viable. Beforerestructuring,suchunitswereownedbyutilitiesandpaidforbyratepayers.Thus, post‐restructuring, a market mechanism was needed to ensure these same types of generatingresourcescouldremaineconomicallyviableeveniftheywererarelyusedand thusdidnotroutinelyearnsufficientrevenuesfromwholesaleelectricenergymarkets. However, FERC also imposed price caps in wholesale electric energy markets to address potential market‐power concerns, especially when electricity demand peaked. But these 10 R.BriggsandA.Kleit,“ResourceAdequacyandtheImpactsofCapacitySubsidiesinCompetitive ElectricityMarkets,”WorkingPaper,Dept.ofEnergyandMineralEngineering,PennsylvaniaState University,October22,2012(BriggsandKleit). http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2165412Appendix1providesamoredetailed reviewoftheBriggsandKleitmodel. 11 BriggsandKleit,p.27. ‐3‐ StateSubsidizationofElectricityGeneration December2012 two goals—sufficient electric supplies to ensure reliability12 and price caps to curb potentialmarketpowerabuses—resultingeneratorsnotrecoveringtherevenuesthatthey needtocontinueoperating.Thisissue,oftencalledthe“missingmoney”problem,hasbeen well‐documented.13 Thesecondproblemthataroseisthatelectricsystemreliabilityhascharacteristicsofwhat economists call a “public good.”14 Because reliability for one is really reliability for all, thereistoolittleincentiveforsupplierstoinvestfullyinresourceadequacy,andthereisan incentiveforend‐useelectricityconsumersto“free‐ride”bynotpurchasingthefullamount ofreliability. Atleasttwobasictypesofpoliciescanaddressthereliabilityproblem.First,regulatorscan raisethepricecaponwholesaleenergypricestothefulleconomicvalueofgenerationat the peak.15 This is an “energy only” solution. The Electricity Reliability Council of Texas (ERCOT)16 has adopted this approach. The Public Utility Commission of Texas, which overseesERCOT,recentlyvotedtoraisethecaponenergypricesto$9,000permegawatt‐ hour(MWh)byJune1,2015.17Incomparison,FERChascappedwholesaleelectricmarket offersat$1,000perMWhinPJM.18 12 Therearereallytwodifferenttypesofreliability.Thefirstis“systemsecurity,”whichdealswith ensuringtheelectricsystemcanmeetminute‐to‐minutechangesinelectricdemandandcompensatefor unforeseenevents,suchassuddenlossesoftransmissionlinesandgenerators.Thesecondis“resource adequacy,”whichaddressesthelong‐termneedforresourcestomeetfutureelectricdemand.Thefocus ofcapacitymarkets—andthisreport—isthelatter. 13 See,e.g.,R.Shanker,“CommentsonStandardMarketDesign:ResourceAdequacyRequirement,”Federal EnergyRegulatoryCommission,DocketRM01‐12‐000,January10;W.Hogan,“OnanEnergyOnly ElectricityMarketDesignforResourceAdequacy”FederalEnergyRegulatoryCommission. http://www.ferc.gov/EventCalendar/files/20060207132019‐hogan_energy_only_092305.pdf 14 SeeJ.LesserandG.Israilevich,“TheCapacityMarketEnigma,”PublicUtilitiesFortnightly,December2005, pp.38‐42.Seealso,P.CramtonandS.Stoft,“ACapacityMarketthatMakesSense,”TheElectricityJournal 18(August2005),pp.43‐54. 15 Inthelimit,thehighestvalueofelectricityisthevalueoflostload(VOLL),orwhatelectricityconsumers wouldbewillingtopaytoavoidaforcedoutage. 16 ForadiscussionofresourceadequacyinERCOT,seeS.Newell,K.Spees,J.Pfeifenberger,R.Mudge,M. DeLucia,andR.Carlton,“ERCOTInvestmentIncentivesandResourceAdequacy,”TheBrattleGroup,June 2012. http://www.ercot.com/content/news/presentations/2012/Brattle%20ERCOT%20Resource%20Adequ acy%20Review%20‐%202012‐06‐01.pdf. 17PUCRulemakingtoAmendPUCSubst.R.25.505,RelatingtoResourceAdequacyintheElectricReliability CouncilofTexasPowerRegion,OrderAdoptingAmendmentsto§25.505asApprovedattheOctober25, 2012OpenMeeting,October25,2012.ThePUCTincreasedthepricecapto$5,000perMWh,effective June1,2013;$7,000perMWh,effectiveJune1,2014;and$9,000perMWh,effectiveJune1,2015. http://www.puc.texas.gov/industry/projects/rules/40268/40268adt.pdf. 18 NewlyadoptedruleswithinPJMforscarcitypricingwilleventuallyallowthelevelofcappedpricestorise to$2,700perMWh. ‐4‐ StateSubsidizationofElectricityGeneration December2012 Second,regulatorscanrequireload‐servingentities(LSEs),whetherlocalelectricutilities orcompetitiveelectricityproviders,tocontractintothefuture(orforward)foraccessto specified amounts of generating capacity, such that each LSE will be individually reliable andthesystemwillbereliableasawhole.19Thesecontractsmaybenegotiatedbilaterally ortheirpricemaybedeterminedusingacentralizedmarket‐clearingmechanism,hence,a “capacitymarket.”20 Althoughtheyhaveevolvedovertime,theforwardcapacitymarketsthatareoverseenby ISONewEngland(ISO‐NE),theNewYorkISO(NYISO),andPJMallshareacommondesign. Specifically,theyallusemarketmechanismsthataredesignedtosendthemarketsignals needed to incent developers to construct additional generating capacity (or supply alternativeformsofcapacity,suchasdemandresponseresources,whichareagreementsto curtail electric usage when called on by the RTO), and thus ensure the “desired” level of system reliability over time.21 Retail electric utilities and competitive firms who supply electricitytoretailcustomersarerequiredtoobtainaminimumquantityofcapacity,soas toensuretherearesufficientreservestomeetpeakdemand. A. CapacityMarketOpposition To be sure, the capacity markets operated by PJM, NYISO, and ISO‐NE have vocal critics. Thesecriticismsfallintotwomainareas: 1. Capacity markets provide “windfall” profits to existing generators, especially baseloadgeneratorsforwhichratepayershavealreadypaid;and 2. Capacity markets are not incenting new generating resource development in constrained regions, despite capacity prices that are higher than the overall RTO prices. Althoughthefirstargumenthasbeenraisedinthepast,especiallywhennaturalgasprices weremuchhigherthantheyaretoday,ithasbeeneffectivelyrefutedandisrarelyraised today.22 Instead, it is the second argument that capacity market opponents now raise. 19 ForadiscussionofhowthePJMcapacitymarketisimplemented,seeJ.Chandley,“PJM’sReliability PricingMechanism:(WhyIt’sNeededandHowItWorks),”2008. http://www.pjm.com/documents/~/media/documents/reports/pjms‐rpm‐j‐chandley.ashx 20Foracomparisonofbilateralcapacitycontractsandcapacitymarkets,seeP.CramtonandS.Stoft,“Why WeNeedtoStickwithUniform‐PriceAuctionsinElectricityMarkets.”TheElectricityJournal20(January (2007),pp.26‐37(CramtonandStoft2007). 21 PJMandtheNYISOuseadministrativelydetermined,downwardslopingcapacitydemandcurvesandan auctioninwhichsuppliersofferincapacityresources.ISO‐NEusesaverticaldemandcurvesettothe installedcapacityrequirementandwhatiscalleda“descendingclockauction,”inwhichsuppliers respondtodescendingpricesuntilthecapacitysuppliedequalstheinstalledcapacityrequirement.The NYISOIndependentMarketMonitorhascalledforNYISOtoadoptasimilardownward‐slopingdesign. SeePotomacEconomics,2011AssessmentoftheISONewEnglandElectricityMarkets,June29,2012,pp. 117‐121.http://www.iso‐ne.com/markets/mktmonmit/rpts/ind_mkt_advsr/emm_mrkt_rprt.pdf. 22 Forarefutationofthe“windfallprofits”argument,seeCramtonandStoft2007,supranote20. ‐5‐ StateSubsidizationofElectricityGeneration December2012 Maryland and New Jersey officials have argued that the lack of new generating capacity being built in their states, despite higher capacity prices reflecting transmission constraints, requires government intervention to subsidize new generating capacity and reduce market prices. For example, the 2011 New Jersey Energy Master Plan stated, “Despite high capacity prices in New Jersey as a result of the RPM capacity market construct,newgenerationhasnotbeenbuiltrecentlyinNewJersey.”23 Theproblemwithsuchanargumentisthatrisingcapacitypricesarenotapersesignalof theneedfornewgeneratingcapacityinvestment.Theentirepurposeofcapacitydemand curvesistoprovideanefficientmarketsignal:onlywhenpricesrisetolevelsthatsupport newgeneratingcapacity,calledthe“costofnewentry”(CONE),24shouldnewcapacitybe developedtoensureresourceadequacy.Instead,therehasbeenanexplosionoflower‐cost supply alternatives, including capacity “uprates” of existing generating units, and a rapid increaseindemand‐response(DR)resourcesthathaveenteredthemarketatacostlower thanCONE.Forexample,sincePJMimplementeditscapacitymarketin2007,almost6,000 MWofcapacityuprateshaveclearedintheRPMauction.Similarly,inthemostrecentRPM auctionthatwasheldinMay2012,clearedDRresourcesincreasedtoalmost15,000MW.25 III. HowGovernmentSubsidizationofGenerationHarmsMarketsand Consumers FERChasexpressedconcernthatcapacitymarkets,arevulnerabletobuyermarketpower. Theissuewasfirstraisedin2006aspartofthesettlementthatestablishedthePJMRPM.26 More recently, FERC became concerned about state‐mandated intervention designed to suppresscapacitymarketprices.Thisissuewasraisedextensivelyinhearingsaboutboth theNewJerseyLongTermCapacityAgreementPilotProgram(LCAPP)andtheMaryland Request for Proposal (RFP) capacity procurements, in which opponents of these states’ 23 See2011NewJerseyEnergyMasterPlan,December6,2011(2011NJEMP),p.81. 24 CONErepresentsthe20‐yearlevelizedpriceoftheleast‐costnewpeakinggeneratingresourcethat wouldbebuilt,netoftheexpectedrevenuessuchaplantwouldearnfromenergyandancillaryservice sales.CONEistheestimatedaveragemarketpriceforapeakingunitthatwouldberequiredtomakean investmentinsuchaplanteconomicallyworthwhile.SeePJM,Updated2015/2016RPMBaseResidual AuctionPlanningPeriodParameters,April6,2012,p.8.http://www.pjm.com/markets‐and‐ operations/rpm/~/media/markets‐ops/rpm/rpm‐auction‐info/2015‐2016‐planning‐period‐ parameters‐report.ashx. 25 HistoricclearedDRresourcedatacanbefoundinthe2011StateoftheMarketReportforPJM,Volume2, p.89,Table4‐9.http://www.monitoringanalytics.com/reports/PJM_State_of_the_Market/2011/2011‐ som‐pjm‐volume2‐sec4.pdf.DataforthemostrecentRPMauctionheldinMay2012isavailablefrom PJM.See2015/2016RPMBaseResidualAuctionResults,p.9.http://www.pjm.com/markets‐and‐ operations/rpm/~/media/markets‐ops/rpm/rpm‐auction‐info/20120518‐2015‐16‐base‐residual‐ auction‐report.ashx. 26 SeePJMInterconnection,L.L.C.,117FERC¶61,331(2006).TheoriginalMOPRrulesexemptedresources undertakenpursuanttostatemandate,whichledtofurtherFERCordersacceptingmodificationstothe MOPRrules.SeePJMPowerProvidersGroupv.PJMInterconnection,L.L.C.etal.,135FERC¶61,022 (2011)(“PJMInterconnection2011”). ‐6‐ StateSubsidizationofElectricityGeneration December2012 plannedinterventionwarnedthatthestateswouldunderminethePJMRPM,aswellasin complaints brought before FERC.27 For example, the PJM Independent Market Monitor (IMM) argued that the Maryland RFP would have significant long‐term adverse consequencesonthecapacitymarketandalsodisagreedwiththecontentionthattheRPM failedtoprovideadequateincentivesforbuildingnewgeneration.28 A. BuyerMarketPower—NoSuchThingasaFreeLunch Althoughthephrase“marketpower”typicallyconnotesactionstakenbysupplierstoraise pricesartificiallyabovecompetitivelevels,withregardtostate‐mandatedinterventionin capacitymarkets,suchastheNewJerseyandMarylandprograms,capacitysupplierswere particularlyconcernedthatcapacitypurchaserswouldmanipulatethecapacitymarketto lower prices artificially below competitive levels.29 State policymaker mandates for local distributionutilitiestoacquirecapacityviatheuseofratepayer‐fundedsubsidiesarethe principalmanifestationofthisconcern. In formal capacity markets, which hold auctions to determine clearing prices, capacity suppliers raised concerns that LSEs, which are required to obtain sufficient capacity resources to meet their forecast demand plus a reserve for contingencies, could exercise buyermarketpower.Specifically,capacitysuppliersexpressedconcernthatLSEscouldbe ordered to build their own or contract for new generating resources—paid for by customers—and offer the resulting capacity into the market at a zero price so as to guarantee its clearing in the auction.30 By adding to the overall capacity supply, the subsidizedgeneratingcapacitywouldthuslowerthemarket‐clearingprice. Althoughsuchanapproachmayseemlikeareasonablestrategyforbenefitingconsumers, it is actually a textbook example of buyer market power. The reason is that, in a well‐ functioning competitive market, neither an individual supplier’s nor an individual consumer’s decisions will affect the market price. When buyers or sellers have market power,however,theirindividualactionsdoaffectthemarketprice.Theextremecaseisa monopolist (single supplier) or a monopsonist (single buyer). A sole buyer or seller, left unchecked,willclearlyaffectthemarketprice,becausetheyarethatsideofthemarket. 27 AsFERCstatedinPJMInterconnection2011,“Themountingevidenceofriskfromwhatwaspreviously onlyatheoreticalweaknessintheMOPRrulesthatcouldallowuneconomicentryhascausedusto reexamineouracceptanceoftheexistingstateexemption,whichweapprovedaspartofthe2006RPM SettlementOrder.Forthesereasons,weacceptasjustandreasonablePJM’sproposaltoeliminatethe currentstateexemption.”(Par.139,footnotesomitted).Inthatsameorder,FERCagreedwiththe PennsylvaniaPublicUtilitiesCommissionthat,“thereisnovalidstateinterestinensuringthat uneconomicofferscansubmitbelow‐costoffersintotheRPMauction.(Id.,Par.142). 28 IntheMatterofWhetherNewGeneratingFacilitiesareNeededtoMeetLong‐TermDemandforStandard OfferService,CaseNo.9214,CommentsoftheIndependentMarketMonitor,January28,2011. 29 AlloftheestablishedcapacitymarketshaveIndependentMarketMonitors(IMMs),whoalsoensurethat capacitysuppliersdonotexercisemarketpowertoraiseprices,suchasbywithholdingsuppliesfromthe market. 30 SeePJMInterconnection2011,Par.20. ‐7‐ StateSubsidizationofElectricityGeneration December2012 Because they must meet specific capacity requirements for reliability purposes, LSEs cannot manipulate capacity market prices by arbitrarily reducing the amount of capacity they purchase. However, left unchecked, LSEs can manipulate capacity market prices by artificially increasing capacity supplies. But in doing so, not only will LSEs harm the market, but their own customers will end up worse off. Therefore, states that promote subsidized capacity development to suppress market prices are really pursuing a “free lunch” economic theory; they assume (contrary to basic economics) that lower market priceswillnotchangethebehaviorofexistingandpotentialcapacitysuppliers. Artificial price reductions caused by subsidized entry will cause existing power plants to shut down prematurely or their owners to abandon plans to expand. Potential market entrants, fearing further government intervention, will not build new power plants. And investors, facing greater risks, will demand higher returns to compensate them for those risks, thus raising thecost of capital for all suppliers. In the end, the marketwill lose as many, or more, megawatts of supply as it gains through subsidies, which will result in risingcosts.Consumerswillnotpaylessforelectricity;theywillpaymore.Thisiswhythe “promise”ofbenefitsarisingfromgovernment‐subsidizedinterventionincapacitymarkets willalwaysremainunfulfilled. In their paper, Briggs and Kleit formally explore this “promise” of benefits from government‐subsidized intervention in capacity markets. Specifically, they pose the following question: if a utility’s retail customers are better off because that utility has subsidizedinvestmenttoloweritsoverallcapacitybill,thenwhydon’tallstatesintervene in the market by subsidizing capacity? As Briggs and Kleit show, the answer is market dynamics.Inotherwords,whenstatesintervenetolowercapacitymarketpricesthrough subsidized investment, that intervention harms existing and potential suppliers, who respond accordingly. Thus, Briggs and Kleit prove that, contrary to state promises, but consistentwithbasiceconomics,thereisno“freelunch.” B. TheBriggs‐KleitModel:EffectsofGovernmentSubsidizedEntryintoCapacity Markets To evaluate the economic impacts of state intervention in capacity markets, Briggs and Kleitdevelopedamathematicalmodelofacapacitymarket.Themodelassumesthereisa transmission‐constrained, “downstream” market and an unconstrained “upstream market.31 They then examine how existing and potential capacity suppliers in these two marketsrespondtosubsidizedgovernmententryinthedownstreammarketthathasthe goal of artificially suppressing the market prices in that downstream market. Their analysisshowsthefollowing: 31 Subsidized baseload capacity investments have significant potential for adverse effectsinboththedownstreamandtheupstreamelectricitymarkets. Appendix1providesamoredetailedreviewofthemodelandgraphicalanalysisoftheeconomicimpacts ofmarketintervention. ‐8‐ StateSubsidizationofElectricityGeneration December2012 Subsidizedinvestmentinbaseloadcapacityisneveroptimal.Inotherwords,such subsidiesalwaysreduceconsumerwell‐being. Government intervention reduces the incentive for private investment. If governmentsrespondtothisreducedincentivebyinterveningagain,aviciouscycle can arise and government’s perceived need to intervene becomes a self‐fulfilling prophecy.Ultimately,theyconcludethattheentirecapacitymarketcancollapse. Well‐designed Minimum Offer Price Rules (MOPR) can reduce the ability of the subsidized generating capacity to distort the market‐clearing price by preventing subsidizedgeneratorsfromsubmittingzero‐priceoffers,butcannotrestoremarket optimality. ThestartingpointfortheBriggsandKleitanalysisisthecostofbuildingnewgenerating capacity, which is called CONE.32 If CONE is higher than the expected capacity market price, no competitive supplier will build capacity. Thus, a utility would only build new capacity if, by doing so, it could exert market power to drive down the market price. Although,bydoingso,autilitymightreapsomeshort‐termbenefitsforitscustomers,these are not economic benefits. Rather, they are short‐term wealth transfers from existing competitivesuppliers. Althoughproponentsofsuchsubsidiesmaydismisscomplaintsabouttransferringwealth from existing suppliers, perhaps based on the “windfall” profits argument discussed previously,thepromisedbenefitstoutilityconsumersassumethatexisting(andpotential) suppliers do not respond to the changed circumstances in the market. In other words, proponentsofstatesubsidizedinterventionassumeanunchangingor“static”market. However, markets are never static. Instead, markets are dynamic, which is one of their great benefits. As circumstances change, market participants adjust their behavior. For example,asgasolinepriceshaveincreased,consumersdemandedmorefuel‐efficientcars and trucks. Automakers responded with technological innovations to improve fuel efficiency,whilemaintainingperformance.Asaresult,theimpactofhighergasolineprices hasbeenreduced. Notsurprisingly,marketdynamicsarealiveandwellincapacitymarkets.Therefore,when oneaccountsforthedynamicmarketresponsesofexistingandpotentialcapacitymarket suppliers in the long‐run, the wealth transfer vanishes and prices increase. Because governmentsubsidiesdistortmarketsignals,theyforcesupplierstochangetheirbehavior. Theartificiallyloweredmarketpriceforcesothercompetitivegenerationsuppliersoutof themarket.Forexample,onOctober22,2012,DominionResourcesannounceditwould shut down its 556 MW Kewaunee Power Station, which is located in Wisconsin, in early 32 Ineconomicparlance,CONEisthemarginalcostofnewentry.Ifthemarketpriceislessthanthe marginalcostofadditionalsupplies,thenitdoesnotmakeeconomicsensetoenterthemarket. ‐9‐ StateSubsidizationofElectricityGeneration December2012 2013 because of low market prices.33 As existing suppliers exit the market, capacity supplies decrease. Not only do capacity market prices increase, but the reduction of capacitycancreatereliabilityissues‐thesamereliabilityissuesthatsomeproponentshave citedasjustifyingmarketintervention. Governmentsubsidiesalsoincreasethecostofnewentryintothemarket,byincreasingthe costofcapital.Investorswillnotwanttorisktheircapitaltocompetewithplantsthathave been artificially subsidized by consumers. In other words, investors will demand additionalreturnstobeartheriskoffuturegovernmentinterventionthatwouldreducethe economic value of their investments. Alternatively, investors may simply wait for their own subsidies. Ultimately, such a system would return us to the pre‐restructuring, monopolyutilityworld.Itwouldresultinlessefficientplantsandconsumersonceagain havingtobeartherisksofutilityinvestments. Ultimately, market dynamics mean that artificially subsidizing new generation cannot “trick”themarket.Subsidiesdonotloweractualcosts,butinsteadraisethem.Subsidies donotencourageinnovation,butreducetheincentivetoinnovateandlowercoststhrough greater operating efficiency. Subsidies do not improve long‐term reliability, but instead harm it by driving out existing and potential competitors. To believe otherwise is to believein“freelunch”economics. Finally,BriggsandKleitshowthatawell‐designedMinimumOfferPriceRule(MOPR)34can limittheabilityofsubsidizedgeneratingcapacitytodistortthemarket‐clearingprice,but cannoteliminatealloftheharmcausedbysubsidizedgeneration.Inessence,aMOPRrule meansgovernment‐subsidizedgenerationcannotbeofferedintothecapacitymarketata zeroprice.Thislimitsthemagnitudeofthepotentiallossestoexistingmarketparticipants arisingfromsubsidizedintervention,butcannotrestorethemarketoptimum.Insummary they find, “Policies like PJM’s MOPR may mitigate this situation if they correctly and crediblyscreenoutnon‐economiccapacityadditions.”35 IV. ConclusionsandPolicyImplications The Briggs‐Kleit model shows that government intervention imposes costs on both consumers who are supposed to benefit and on consumers in interconnected markets. Because government intervention, or even the expectation of intervention, imposes costs on capacity suppliers, they will be less likely to invest. As a consequence, government intervention to “correct” for a perceived absence of market‐based investments in generatingcapacitysimplydrivessuppliersaway.Thus,governmentinterventionisaself‐ fulfilling prophecy that, if left unchecked, can eventually drive all suppliers from the capacitymarket. 33 Inpart,lowershort‐runmarketpricesarebeingdrivenbyrapidincreasesinsubsidizedwindgeneration. About20%ofallUSinstalledwindcapacity,approximately10,000MW,hasbeenbuiltwithinMISO. 34 SeePJMInterconnection,L.L.C.,139FERC¶61,011(2012). 35 BriggsandKleit,p.27. ‐10‐ StateSubsidizationofElectricityGeneration December2012 Because electric markets are interconnected and extend beyond individual states, governmentinterventioninindividualstates,suchasundertheMarylandandNewJersey programs,imposeslong‐runcostsonneighboringstates.Itiseconomicfallacythatprice distortionscausedbygovernmentsubsidiesinafreemarketare“benefits.”Such“beggar‐ thy‐neighbor”policiesarenotonlycounterproductive,theyinvitepolicy“retribution”that willfurtherdamagemarkets. Finally,justifyinginterventionasameans,notonlytoreducemarketprices,buttocreate jobs, is a last refuge of the interventionist scoundrel. The short‐run “jobs” benefits from buildingsubsidizedgenerationareunlikelytoaccruetolocalconsumers,whomustpayfor the subsidies. By reducing incentives for new private investment in the long run and increasingelectricprices,interventionwillcauseanoverallreductionineconomicgrowth. Thus, such policies, pursued for economic development reasons, are “penny wise and pound foolish.” At best, they are an undue burden on taxpayers. At worst, as shown by Briggs and Kleit, “these subsidies have the perverse effect of reducing the incentives for resourceadequacyinthelongrun.”BriggsandKleitconcludethat“ouranalysissuggests thatto“keepthelightson,”stateseitherneedtoletmarketsworkorfacetheprospectof continued, costly interventions over the long run.”36 Based on these findings and the consequences discussed in this paper, future state subsidization of electricity generation shouldbeavoided.Itharmstheveryconsumersitisintendedtobenefit,harmsconsumers inneighboringstates,andwreckscompetitivecapacitymarkets. 36 BriggsandKleit,p.27. ‐11‐ StateSubsidizationofElectricityGeneration December2012 Appendix1:TheBriggs‐KleitMultipleMarketModel TheBriggs‐Kleitmodelconsiderstheimpactsofsubsidizedgenerationwhenthereareboth upstreamanddownstreamcapacitymarkets.43Thedownstreammarketisassumedtobe transmissionconstrainedatleastsomeofthetime.Asaresult,thedownstreamcapacity market price, PD, is assumed to be higher than the price in the upstream market, PU, as showninFigureA1‐1.44 FigureA1‐1:Briggs‐KleitTwo‐MarketModel This price differential is assumed to underlie downstream policymakers’ desire to subsidizenewgenerationentry.BriggsandKleitexaminetheimpactsofbothsubsidized baseload capacity and subsidized peaking capacity to determine whether the type of generationsubsidizedaffectstheoverallchangeineconomicwell‐beingtoconsumersand existingsuppliers.Theirmodeladdressesfourpolicyquestions: How does the presence of transmission constraints affect the optimal design and functionofcapacitymarkets? Doesitevermakesenseforgovernmentstosubsidizecapacity,andifsohow? 43 TheBriggsandKleitmodelextendsthesingle‐marketanalysispreviouslydevelopedbyJoskowand Tirole,whichwasbasedona“Ramseyequilibrium”model.InplainEnglish,aRamseyequilibriumisone inwhichthegovernmentknowshowtheprivatesectorwillreacttogovernmentintervention.The governmentintervenesinawaydesignedtomaximizeoverallwell‐beingandtheresultingmarket conditions,aftertheprivatesectorresponse,istheRamseyequilibrium.SeeP.JoskowandJ.Tirole, “ReliabilityandCompetitiveElectricityMarkets,”TheRANDJournalofEconomics38(Spring2007),pp. 60‐84(JoskowandTirole2007). 44 Iftherewerenotransmissionconstraints,themarket‐clearingpriceswouldbethesameandthetwo marketswouldeffectivelybetreatedasasingleone.ThereasonPJMandotherRTOswithcapacity marketshaveseparatezonesistoreflecttransmissionconstraintsandprovidemoreaccurateprice signalsinconstrainedmarkets. A1‐1 StateSubsidizationofElectricityGeneration December2012 Are price floors like the MOPR sufficient for ensuring the optimal functioning of capacitymarkets? How will markets react to subsidized capacity investments in the short and long runs? The Briggs and Kleit framework also allows for price caps in the energy market. As discussed previously, such price caps create a “missing money” problem that capacity marketsaddress.ExtendingtheJoskowandTirole(2007)model,theyshowthatcapacity markets restore the appropriate market incentives for investment in peaking capacity as longas(1)allgeneratingunitsarepaidthecapacityclearingprice;and(2)thatthemarket price consumers pay incorporates the capacity market price, and not just the short‐run energymarketprice.45 Wenowaddressthequestionofwhethergovernmentinterventioneverimprovesoverall well‐being.Aswediscussedpreviously,reliabilityisapublicgoodthatmarketswilltendto under‐supply. Briggs and Kleit show that, in the short‐run and in an ideal world, the governmentcouldprovidethecorrectincentivesfornewpeakingcapacityinthemarket, andthusensuresufficientcapacitytomeetreliabilitystandards.Theyalsoshowthisresult doesnotholdforsubsidizedbaseloadcapacity.However,eveninthisidealized,short‐run setting, Briggs and Kleit show that economic well‐being is increased if the market determinesthecapacityprice.Astheystate: Reducing the capacity price affects the incentive for the marginal producer andreducescapacityinvestment.Thisintuitionessentiallyrestatesthelogic at the heart of a capacity market policy: the policy supplies the market information on the demand for reliability, and allows the market to determine the price for that capacity. When governments act to affect this price, they distort the information that the market provides, reduce incentivesat the margin for capacity suppliers, and reduce the efficiency of theoutcomeforconsumers.46 Briggs and Kleit consider both the short‐run and long‐run impacts of government intervention in the transmission‐constrained downstream market. In the short‐run, they showthat,becausedownstreaminterventionwithsubsidizedcapacityreducesexportsinto the downstream market, upstream consumers benefit from downstream government intervention. In other words, because downstream consumers must pay for the government‐subsidized capacity, those consumers may not see any benefit, even in the short‐run. 45 BriggsandKleitalsoshowthatthecombinationofanenergymarketpricecapandacapacitypayment meansthatthegovernmentcannotcreateaRamseyequilibrium,whenevertherearemorethantwo possiblestatesoftheworld. 46 BriggsandKleit,p.17(emphasisadded). A1‐2 StateSubsidizationofElectricityGeneration December2012 Short‐andLong‐RunImpactsofGovernmentSubsidizedEntryonCapacityMarkets Below, we provide a general overview of the short‐run and long‐run impacts of government‐subsidized entry based on the work by Briggs and Kleit. To understand the short‐runimpacts,supposetherearefourcapacitysuppliers—A,B,C,andD—whosubmit offers to provide capacity in the amount of QA, QB, etc., of increasing cost, as shown in Figure A1‐2 (supply curve S0). Based on the administratively determined demand curve, the market‐clearing capacity price is P*. Q* is the equilibrium quantity of capacity purchased,withallofthecapacityofferedbysuppliersAandBtaken,andonlysomeofthe capacityofferedbysupplierC.47 Next, suppose subsidized capacity supplier S is allowed to bid in a quantity of capacity, QSUB,intothemarketatazeroprice.Inthatcase,theoffersoftheothersuppliersareall pushedout,andthenewcapacitysupplycurveisS1.Thenewsupplycurveintersectsthe demand curve at Q and the resulting market price falls to PSUB. At this lower price, supplierCisknockedoutofthemarketentirely,andtheonlysuppliersofcapacityareA,B, andthenewsubsidizedsupplier. FigureA1‐2:Short‐runImpactofGovernmentSubsidizedCapacity 47 InFigureA1‐1,theamountofcapacitysuppliedbyCis:QC–(Q*–QB). A1‐3 StateSubsidizationofElectricityGeneration December2012 Intheshort‐run,consumersappeartogainfromthedecreaseincapacityprice.Thisgain, called “consumers’ surplus,” accrues primarily from a transfer of wealth from capacity suppliers48 (the large diagonally‐shaded rectangle) and a small increase because of the increase in total capacity supplied (the small gray triangle). The net benefit, if any, to consumersisthedifferencebetweenthemonetarygainfromthelowercapacitypriceand the cost of the subsidized capacity. A key finding of Briggs and Kleit’s work is that the consumerscalledontosubsidizecapacitythroughLCAPP‐typeprogramsareleast‐likelyto enjoy even short‐run benefits, whereas consumers upstream will capture the majority of theshort‐runbenefits. Although subsidizing capacity can reduce market prices in the short‐run, focusing exclusively on the short run ignores critical market dynamics, that is, effects over time. Whenevaluatinggovernmentinterventionintendedtosuppressmarketprices,accounting for these dynamiceffects is crucial, for several reasons. First, marketsare not static;the entire point of a capacity market, for example, is to send appropriate price signals over timeandincententry(andexit)asneeded.Second,marketparticipantsrespondtoprice signals. Thus, subsidized entry in the capacity market will have long‐run impacts on, unsubsidized suppliers. Understanding these impacts is crucial to evaluating the overall impactsofgovernmentintervention. Next, we consider the long‐run impacts. As Briggs and Kleit show, in the long‐run all consumers lose. To see this, consider Figure A1‐3. Because the subsidy reduces the market‐clearing price in the short‐run, Briggs and Kleit’s analysis finds that existing suppliers A and B respond by either exiting the market or forgoing investment. Thus, in our example, the capacity supplied by A and B decreases over time to QA’ and QB’, respectively. Moreover, because investors perceive greater risk to supplying capital to generators, the cost of capital increases, which further raises costs and the market price (Psub)abovewhat itotherwisewouldbewithoutthesubsidizedgeneratingcapacity.This resultsinthelongrunmarketprice(Psub)risingabovetheoriginalmarketprice(P*)asa result of government‐subsidized entry into the market without mitigation to protect consumers. 48 Thisisreflectedasareductionin“producers’surplus,”whichrepresentsthereturnstoindividual suppliers. A1‐4 StateSubsidizationofElectricityGeneration December2012 FigureA1‐3:Long‐RunImpactofGovernmentSubsidizedCapacity Long‐term market price increases Because government intervention downstream reduces exports from upstream, the net upstream capacity supply increases, causingthe marketprice todecrease.Consequently, upstream market consumers benefit from the downstream market intervention. For example, this suggests that the primary beneficiaries of the New Jersey and Maryland capacity subsidy programs are consumers in Pennsylvania and Ohio, not in the states themselves. Thus, advocates of intervention in the New Jersey and Maryland capacity zonesareeffectivelyforcingtheirconsumerstosubsidizebenefitsforconsumersupstream inPennsylvania. Iftheseupstreambenefitspersistedinthelong‐run,wewouldexpecttoseepolicymakers inupstreamstatescheeringontheinterventionisteffortsindownstreamefforts.Because we do not, it means there are adverse long‐term impacts stemming from the dynamic response of capacity market suppliers in both upstream and downstream markets. To evaluate long‐term, dynamic responses to government intervention, Briggs and Kleit assumethatmarketparticipantsbasetheirinvestmentdecisionsonexpectationsoffuture government intervention. Thus, if capacity suppliers believe there is a 50% probability that the government will intervene in the market at some later time, and that the probabilityofinterventionisaffectedbythemarketprice,thensupplierswilladjustchange their supply decisions. In essence, the expectation of market intervention—even if that intervention does not actually occur—imposes an expected cost on suppliers. The A1‐5 StateSubsidizationofElectricityGeneration December2012 expectedcostisaformofregulatorytaking.49Theresultofthisexpectationofintervention is that suppliers will reduce future investment, reducing reliability in both the upstream and downstream markets. According to Briggs and Kleit, “[A]s suppliers’ expectations of government action become reinforced, the expected return to capacity investments decreases.Thisstateofaffairscreatesaproblemforlongrungridresourceadequacy….”50 Theexpectationofintervention,whichistiedtothecapacitymarketprice,thusincreases themarketpriceneededtoinducesupplierinvestment.However,ifsuppliersseehigher marketprices,theywillalsoexpecttheprobabilityofgovernmentinterventiontoincrease, which means they require an even higher market price to induce them to enter, and so forth.Inotherwords,theexpectationofgovernmentinterventioncreatesaself‐fulfilling conditionthatincreasescostsanddrivessuppliersfromthemarket.Eventually,theonly suppliersinamarketwillbesubsidizedones,andthemarketwillceasetoexist. Thus,whengovernmentpolicymakersindownstreammarketscomplainthatmarketprices are not inducing new capacity investment and threaten to intervene with subsidized capacity, the policymakers effectively signal to suppliers not to invest. And, because downstream intervention also leads to lower short‐run prices in upstream markets, threatened(oractual)interventionrestrictsinvestmentinupstreammarkets.TheBriggs and Kleit model demonstrates that the long‐run impact of government intervention reduces well‐being in both downstream and upstream markets, compared with the non‐ interventioncase. Briggs and Kleit conclude their paper by analyzing how MOPR‐type instruments, which limit the extent to which subsidized capacity can be offered into the capacity market, affects the overall welfare impacts. They find that MOPR‐like instruments limit the magnitudeofthepotentialregulatorytakingsarisingfromsubsidizedintervention(similar towhatwasshowninFigureA1‐2),butcannotrestorethemarketoptimum. Finally,FigureA1‐4providesanexampleoftheshort‐runimpactsofaMOPR.Specifically, supposetheMOPRforthesubsidizedgeneratingresourceisbetweentheoffersofcapacity resourcesBandC.Inthiscase,thesubsidizedgenerationcapacityisnotallowedtobida pricebelowMOPRSUB.Thesubsidizedmarketpriceisstillbelowtheinitialmarket‐clearing price,P*,buttheshort‐runimpact,intermsofthewealth‐transferfromexistingcapacity supplierstoconsumersandoverallbenefittoconsumers,decreases.However,theMOPR instrument cannot restore optimality to the market, because the threat of government intervention, and the impact of that threat on investor expectations, will still distort the market. 49 Foracomprehensivediscussionofregulatorytakings,seeW.Fishel,RegulatoryTakings:Law,Economics andPolitics,(Cambridge,MA:HarvardUniversityPress1995). 50 BriggsandKleit,p.27. A1‐6 StateSubsidizationofElectricityGeneration December2012 FigureA1‐4:Short‐runImpactofSubsidizedCapacity–WithMOPR A1‐7 StateSubsidizationofElectricityGeneration December2012 Appendix2: ReviewofElectricRestructuringandCreationofCompetitive WholesaleMarkets By the late 1970s, the electric industry was in turmoil. After two OPEC oil embargoes, rapidly escalating costs for nuclear power plants that threatened to bankrupt electric utilities, new environmental regulations that were needed to reduce air and water pollution, and looming shortages of natural gas (ironically caused by preventing market pricesfromincentingnewexplorationanddrilling),theelectricindustrywasonitsknees. A new direction clearly was needed, one which would restore the industry’s financial health, ensure enough new generating capacity was built to meet increased demand, and keepelectricitypricesfromskyrocketingoutofcontrol. Thisnewdirectionentailedrestructuringtheelectricindustry,supplementingand,insome cases, replacing the old vertically integrated structure with competitive wholesale and retail markets. Electric utility restructuring was inspired by the success of introducing market competition into other regulated industries, including airlines, trucking, telecommunicationsand,perhapsmostimportantly,naturalgas.Inalloftheseindustries, competitionhasledtoinnovation,improvedefficiency,andlowerprices. In1992,CongresspassedtheEnergyPolicyAct(EPAct),whichcreatedthefoundationfor competitivewholesaleandretailelectricmarkets.Specifically,EPActcreatedanewclassof competitive generators, called Exempt Wholesale Generators (EWGs). EWGs were designed to compete directly with generation built and operated by electric utilities themselves,unlikethe“qualifiedfacilities”(QFs)createdaspartofthe1978PublicUtilities ResourcePolicyAct.Thelatterwereprimarilyrenewablegeneratorswhoseoutputelectric utilitieswererequiredtopurchaseatpricessetbystateutilityregulators—oftenathighly inflated prices stemming from wildly inaccurate forecasts of future oil and natural gas prices. Several years later, in 1996, the Federal Energy Regulatory Commission (FERC) issued Order888,a major policy order designed torestructure the electric transmission system and promote “open” competitive access for generators, thus enabling them to sell electricity in wholesale markets.51 Although so‐called “power‐pools” – multiple utilities combiningtheirgeneratingresourcestoreducecostsandimprovereliability–hadexisted since the late 1920s,52 Order 888 led to creation of independent transmission organizations, called “independent system operators” (ISOs) whose role was to better coordinatetransmissionsystemoperationsandensurethatopenaccessrequirementsdid not jeopardize the overall reliability of the bulk power system (i.e., the system of central stationgenerationandhighvoltagetransmission). 51 PromotingWholesaleCompetitionThroughOpenAccessNon‐discriminatoryTransmissionServicesby PublicUtilities;RecoveryofStrandedCostsbyPublicUtilitiesandTransmittingUtilities,OrderNo.888,75 FERC61,080(1996), 52 ThefirstsuchpowerpoolwasthePennsylvania–NewJersey–Marylandpool,theprecursortoPJM, whichbeganin1928. A2‐1 StateSubsidizationofElectricityGeneration December2012 Three years later, FERC issued Order No. 2000, which was designed to further enhance openaccesstransmissionandcompetitivewholesalemarketsthroughcreationofRegional Transmission Organizations (RTOs), which can be thought of as enhanced ISOs.53 In addition to operating wholesale spot markets for electric energy, several of these RTOs developed separate markets for installed generating capacity—essentially payments to generating firms to recover the fixed construction costs that were previously included in rate base and in return for those to provide sufficient revenues for firms to constructing additionalgeneratingcapacityforuseduringtimesofpeakdemand.Thoughthatcapacity wouldbeuneconomicalinawholesaleenergymarket,itwasnecessarytoensuretherewas sufficientgeneratingcapacitytomeetreliabilitystandards.54 53 RegionalTransmissionOrganizations,89FERC¶61,285(1999).Althoughtheyarenottechnicallythe sametypesofentities,theresponsibilitiesandoperationsofISOsandRTOsarequitesimilar.Further addingtotheconfusion,RTOscoveringNewEnglandandtheMidwestareISO‐NEandMidwestISO, respectively. 54 Therearetwoflavorsofreliability:long‐termresourceadequacyandshort‐termsystemsecurity. Capacitymarketsweredevelopedtoaddresstheformer.Thelatter,incontrast,focusesonspecific minute‐to‐minuteoperationofthebulkpowersystem.Foramoredetaileddiscussion,seeJ.Lesserand G.Israelivich,“TheCapacityMarketEnigma,”PublicUtilitiesFortnightly,December2005,pp.38‐42. A2‐2 StateSubsidizationofElectricityGeneration December2012 Appendix3:TheNewJerseyandMarylandPrograms The New Jersey and Maryland subsidized generation programs were established for a varietyofreasons,includingeconomicdevelopmentandjobcreation.Theprincipaldriver oftheprograms,however,wastoreducethecostofgeneratingcapacitythestate’selectric utilities are required to purchase to meet PJM’s reliability standards. For example, New Jersey Board of Public Utilities (NJBPU) President Lee Solomon criticized both PJM and FERC,arguingthatthePJMcapacitymarkethad“overcharged”thestatebymorethan$1 billion, without incenting new generating capacity within the state. This issue was also raisedinthe2011NewJerseyEnergyMasterPlan,whichstates: [m]any market participants argue that RPM has not brought enough generation into the markets where and when needed. A key finding of the June 2010 BPU Technical Conference was that generators proposing new projectsarenotabletoobtainfinancingatreasonableratestodevelopnew assetsduetouncertaincapacityrevenues.55 To address the perceived shortcomings of the PJM RPM, NJ Senate Bill 2381, which was signed into law in January 2011, created the “Long‐Term Capacity Agreement Pilot Program” (LCAPP). Under LCAPP, generation developers responded to an RFP to build a total of 2,000 MW of new, in‐state generating capacity. The winning generators were awarded 15‐year contracts with the state’s electric utilities and are required to bid and clear all of this new capacity into the PJM RPM auction. The winning bidders are guaranteed a fixed price for their capacity. If the market price is less than the bid price, utilityratepayersarerequiredtomakeupthedifference.Thisrevenueguaranteeisaform ofdirectsubsidy. Three generation developers—Competitive Power Ventures (CPV), Hess, and NRG Energy—were selected to build new capacity under the LCAPP solicitation. As a result, New Jersey ratepayers have guaranteed CPV and Hess over $2.1 billion in revenues over thenext15years.56TheCPVplantwillbeon‐linebeforetheJune1,2015startofthePJM 2015‐2016 energy year, and will be paid a price of $286.03/MW‐day for the first year.57 TheHessfacility,whichclearedintothe2015‐2016auction,willreceiveacontractpriceof $200/MW‐daybeginninginthe2016‐17energyyear.TheNRGfacilityfailedtoclearinthe PJMRPMauctionbecauseofPJM’sMinimumOfferPriceRule(MOPR).58 55 2011NJEMP,pp.21‐22. 56 SeeG.Thomas,“TimetoFold'em‐‐NewJersey'sBetonPowerPlantsGoesTerriblyWrong,”NJSpotlight, June13,2012.http://www.njspotlight.com/stories/12/0613/1432/. 57 Incomparison,themarket‐clearingpricewas$167.46/MW‐day.SeeJ.Kaltwasser,“StateReleasesNew LCAPPNumbers,”NJBIZ,October3,2012. http://www.njbiz.com/article/20121003/NJBIZ01/121009938/State‐releases‐new‐LCAPP‐numbers. 58 ISO‐NEusesasimilarconcept,knownas“out‐of‐market”(OOM)resources.SeeISONewEngland,Inc.,et al.,135FERC¶61,029(2011). A3‐1 StateSubsidizationofElectricityGeneration December2012 The Maryland program is smaller in scope, but is based on similar principles. The Maryland program began in 2009, when the Maryland Public Service Commission (MarylandPSC)issuedanorder“toinvestigatewhether[theCommission]shouldexercise its authority to order electric utilities to enter into long‐term contracts to anchor new generationortoconstruct,acquire,orlease,andoperate,newelectricgeneratingfacilities inMaryland.”59 Based on similar concerns that the PJM RPM was not providing adequate incentives to buildnewgeneratingcapacityinMaryland,theMarylandPSCrequiredtheutilitiestoissue a request for proposals (RFP) for up to 1,500 MW of gas‐fired generating capacity. UltimatelyCPVMarylandwontheRFP,proposingtobuilda660MWgas‐firedgenerating unitinCharlesCounty,Maryland,withacommercialoperationdateofJune1,2015.Like the New Jersey LCAPP, the Maryland program provides a revenue guarantee to CPV, in which, if the RPM market price falls below CPV’s guaranteed revenue stream, Maryland ratepayerswillmakeupthedifference.60ThetotalcostoftheMarylandrevenueguarantee isover$800millionoveraten‐yearperiod.61 59 IntheMatterofWhetherNewGeneratingFacilitiesareNeededtoMeetLong‐TermDemandforStandard OfferService,CaseNo.9214,OrderNo.82936,September29,2009,pp.2‐3. 60 Id.,OrderNo.84815,April12,2012. 61 CaseNo.9214,MemoofBostonConsultingGrouptoMarylandPublicServiceCommission,April3,2012, p.5. http://webapp.psc.state.md.us/Intranet/Casenum/NewIndex3_VOpenFile.cfm?ServerFilePath=C:\Casen um\9200‐9299\9214\\138.pdf. A3‐2 This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy Wind Generation Patterns and the Economics of Wind Subsidies Dr. Jonathan Lesser is the founder and President of Continental Economics, Inc., an economic and litigation consulting firm providing services to utilities, industry, and regulators on a broad array of market, regulatory, investment, and environmental issues affecting all segments of the energy industry in the U.S., Canada, and Latin America. Dr. Lesser has coauthored three textbooks, most recently Principles of Utility Corporate Finance, which was published in 2011 by Public Utilities Reports, Inc. He holds a B.S. in Mathematics and Economics from the University of New Mexico, and a M.A. and Ph.D. in Economics from the University of Washington. He can be reached at [email protected]. Funding for this research was provided, in part, by Exelon Corporation. However, the views expressed are solely those of the author and do not necessarily represent the views of Exelon Corporation or its subsidiaries. 8 An analysis supports the conclusion that there is no economic rationale for continued subsidization of wind generation. At the federal level, direct subsidies, such as the federal production tax credit, should not be continued. State-level subsidies, whether feed-in tariffs established by state regulators or statutory RPS mandates, further exacerbate market distortions and raise electricity prices, again to the detriment of consumers. Jonathan A. Lesser I. Introduction The United States has subsidized the wind industry for 35 years. At the federal level, subsidies began with the Public Utility Regulatory Policy Act (PURPA) of 1978. Under PURPA provided indirect subsidies for renewable generation through mandates that electric utilities purchase the output of qualifying facilities (QFs) based on forecasts of avoided costs, essentially ‘‘but for’’ cost projections made by the utilities and approved by state regulators, or made by those regulators themselves. With passage of the Energy Policy Act of 1992 (EPAct), wind subsidies were increased through a variety of programs. The most prominent was the federal production tax credit (PTC).1 Although not specifically limited to wind generation, approximately 75 percent of the 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 The Electricity Journal Author's personal copy total PTC credits claimed since its inception have been for wind generation.2 Wind generation benefits from other subsidies as well. Since 2009, for example, the wind industry has received payments under the $831 billion American Recovery and Reinvestment Act of 2009 (ARRA). Perhaps the largest subsidy has been through statelevel renewable portfolio standards (RPS), which mandate minimum levels of renewable generation that electric utilities or competitive generation suppliers must obtain as part of their overall resource mix used to serve customers. Currently, 30 states, plus the District of Columbia, have such RPS mandates. nlike market prices for other commodities, the market price of electricity varies by season, day, and hour. In part because electricity cannot be stored cost-effectively, the price is highly dependent on daily fluctuations in demand—higher demand during the day and lower demand at night—and seasonal changes. In most of the U.S., for example, electricity demand now peaks in the summer, driven by increased use of air conditioning in commercial and residential buildings. As a result of this price variation, the value of subsidized wind generation also varies by season, day, and hour. In some hours, the value of electricity can be thousands of dollars per MWh. In other hours, the value actually can be less than zero. U Jan./Feb. 2013, Vol. 26, Issue 1 The purpose of this article is to examine the economic value of subsidized wind generation. Specifically, are taxpayers and consumers who are forced to pay for subsidized wind power receiving high- or low-value electricity? Answering this question has important policy implications. First, Congress is currently considering whether or not to extend the PTC for an additional year, at an estimated Are taxpayers and consumers who are forced to pay for subsidized wind power receiving high-value or low-value electricity? cost of over $12 billion. Second, because the percentages of renewable generation required under state RPS requirements continue to increase, electricity consumers will be forced to subsidize greater amounts of wind power, which will have larger impacts on electricity costs. Third, continued subsidization of wind generation will lead to higher long-run retail prices for electricity,3 which will have adverse impacts on economic growth. Given these reasons, determining the value consumers obtain for their subsidy dollar is highly relevant to policy decisions regarding continued subsidies. II. Economic Costs of Wind Power Subsidies Renewable energy subsidies have been advocated for a variety of reasons, ranging from common arguments about protecting emerging or ‘‘infant’’ industries so they may become established, 4 to ‘‘two wrongs make a right’’ justifications, i.e., that because fossil fuel generating resources have been subsidized, it is only ‘‘fair’’ that renewable generation be subsidized, to arguments that renewable subsidies offset external environmental costs of fossil fuel generation.5 Regardless of how they are justified, subsidies distort competitive markets, drive out unsubsidized competitors, and reduce the incentives to innovate and improve operating efficiency.6 In addition to these economic costs, wind power subsidies create four other types of adverse economic spillovers because of the nature of electric markets and integrated power grids. irst, because baseload generators, e.g., nuclear and coal-fired power plants, cannot be cycled easily, these generators operate even when the market price of electricity is less than their variable operating costs.7 As a consequence, when the demand for electricity is sufficiently low, market prices can fall below zero. In such situations, baseload generation owners are then forced to pay to generate power and inject that power into the grid, F 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 9 Author's personal copy which exacerbates economic losses.8 With a current after-tax PTC of $22/MWh, it is economically rational for wind generators to sell power into the market even when prices are as low as $34/MWh.9 As a result, negative pricing periods are exacerbated, which increases the costs for baseload generators who are unable to cycle their units and may hasten their retirement. econd, the inherent intermittency of wind generation increases the costs of maintaining power system reliability. The intermittent nature of wind generation requires additional generating reserve capacity so as to ‘‘firm’’ wind supply. Moreover, rapid variations in wind output can require additional voltage support through automatic generation control (AGC) that automatically adjusts the output of flexible generating resources (e.g., gas-fired turbines) so as to maintain voltage and frequency within acceptable levels. A study published by the National Renewable Energy Laboratory (NREL) estimated these integration costs to be about $5/ MWh.10 In Texas, which has over 10,000 MW of installed wind capacity, in 2011 these integration costs added an estimated additional $140 million in power system costs. Nationally, integration costs were over $500 million in 2011.11 Third, wind generation requires additional investment in high-voltage transmission lines, because wind resources are S 10 geographically dispersed and typically located far from load centers. The costs of highvoltage transmission lines are generally socialized across all transmission system users. Texas alone spent over $6.9 billion on Competitive Renewable Energy Zone (CREZ) high-voltage transmission lines to interconnect wind power.12 Fourth, the demonstrated inaccuracy of short-term forecasts of wind generation increases the overall cost of meeting electric demand as system planners must reimburse other generators who had been scheduled to operate, but were not needed because actual wind generation was greater than forecast, or had not been scheduled, but were required to operate because actual wind generation was less than forecast. Although generators can be penalized for erroneous forecasts, most of the resulting system costs are socialized across all users. Despite claims by wind power advocates that wind generation can be predicted accurately several days in advance, allowing system operators to reduce, if not eliminate, the impacts of wind’s volatility, actual data does not bear this out.13 III. The Economic Value of Wind Generation To examine the economic value of subsidized wind generation, we analyzed wind generation in three regions where there has been extensive—and rapid— development of wind power: the PJM Interconnection, which covers the mid-Atlantic states and the Ohio Valley; MISO, which covers much of the remaining Midwestern States; and ERCOT, which oversees the electric system in almost the entire state of Texas. Together, these three regions account for over 27,000 MW of wind generating capacity, more than half of the approximately 50,000 MW of installed wind generating capacity in the U.S.14 Because of weather patterns that can change from year to year, we examined hourly wind generation and load data over a 44-month period, Jan. 1, 2009, through Aug. 31, 2012, to assess the relative economic value of wind power. We then evaluated the performance and availability of wind power in each of the four seasons, where each season was defined as including the months shown in Table 1.15 From both a system planning and customer perspective, the highest-value generating resources are those that are available when electricity demand peaks: like taxicabs that never show up when it is raining, generating resources that fail to produce when most needed have little value. Table 1: Month-Season Mapping. Season Winter Months December–February Spring March–May Summer Fall June–August September–November 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 The Electricity Journal Author's personal copy [(Figure_1)TD$IG] Figure 1: PJM Hourly Load and Wind Generation, July 1–8, 2012 Consider, for example, the pattern of hourly load and wind generation in PJM for the week of July 1–8, 2012, when much of the eastern U.S. was in the grip of a record heat wave (Figure 1). ver that week, a strong negative correlation between hourly demand and wind generation is apparent. The actual correlation coefficient is 0.40.16 As Figure 1 shows, over this week, wind generation usually peaked in the late night and early morning hours, whereas peak demand occurred in the late afternoon. Electricity demand peaked at 5 PM on July 6 of this week, when demand was over 151,000 MW. During that same hour, 201 MWh of wind power was generated by the approximately 4,700 MW of installed wind capacity in PJM, O Jan./Feb. 2013, Vol. 26, Issue 1 less than 5 percent of the potential generation. As little generation as that was, it represented an increase from earlier in the day, as only 14 MWh was generated during the hour between Noon and 1 PM. In the Northern Illinois zone, which encompasses Chicago, the demand for electricity averaged 22,000 MW over the entire day; the average amount of wind power generated was just 4 MW. From a system planning standpoint, the ‘‘gap’’ between high hourly loads and low wind output makes wind a far less valuable and far less reliable resource than conventional generating resources. This ‘‘gap’’ between peak electric demand and low wind generation is not only observable on a daily basis, but can also be observed on a seasonal basis. T o evaluate the load-wind gap, we first calculated average daily wind availability, Wd,y, during a standard 16-hour on-peak portion of each day, 7 AM–11 PM, as total wind generation relative to total potential generation based on installed wind capacity, WC,m,y.17 Thus, 16 1 X W d;y ¼ w =W C;m;y ; (1) 16 h¼1 h;d;y where wh;d;y equals hourly wind generation on day d of year y. Next, we average these daily wind availability values in each season of each year to define ¯ S;y . seasonal wind availability, W Thus, Ds X ¯ S;y ¼ 1 W : W Ds d¼1 d;y (2) Similarly, we define the annual ¯ A;y , as the wind availability, W 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 11 Author's personal copy average daily wind availability over year y, or 365 1 X ¯ W : WA;y ¼ 365 d¼1 d;y (3) The seasonal wind ratio is just equal to the ratio of the seasonal and annual wind ¯ S;y =W ¯ A;y . availability levels, or W Next, we define the seasonal load ratio, L¯S;y , as the average load during season S of year y relative to the average annual load in year y, L¯y . Thus, Ds L X d;y L¯S;y ¼ ; (4) ¯ Ly value of subsidized wind generation, the load-wind gap should be as large as possible when load and market prices are at a maximum. That is, the economic value of subsidized wind generation will be maximized if the relative wind generation is greatest when loads are greatest. Intuitively, during peak demand hours, wind d¼1 where 365 1 X L¯y ¼ L 365 d¼1 d;y (5) Finally, the load–wind ‘‘gap,’’ GS,y, equals the difference between the seasonal wind availability ratio and the seasonal load ratio: GS;y ¼ ¯ S;y W L¯S;y ¯ A;y W (6) For example, suppose the seasonal load in spring of year y equals 90 percent of annual average load, but seasonal wind generation is 120 percent of annual average wind generation. Then the spring load–wind gap, GSpring,y equals 120–90 percent, or +30 percent. A positive load-wind gap value means there is relatively more wind generation available to serve load; a negative load-wind gap value means there is relatively less wind generation available to serve load. rom the standpoint of maximizing the economic F 12 generation will displace high-cost fossil generating units; the greater the availability of wind power, the greater will be the cost savings from displacing fossil-fuel peaking units. n contrast, when load and market prices are low, wind generation will displace lower variable-cost baseload resources. Moreover, when load is especially low and baseload resources cannot be cycled, wind generation will not displace any generation. Instead, wind will simply force baseload generation owners to pay to continue operating, driving prices below zero. In such cases, the value of wind displacement is zero; subsidized wind generation I simply results in a wealth transfer from existing generation owners to wind generators and consumers. Although consumers may benefit from lower wholesale prices in the short run if load-serving entities are relying on the market, in the long run, consumers will be worse off, as demonstrated by Briggs and Kleit.8 Figures 2–4 illustrate the seasonal load-wind gaps for ERCOT, MISO, and PJM. As Figures 2–4 demonstrate, however, the economic value of subsidized wind generation does not follow this pattern. In each region, there is a strong lack of wind generation during the last four summers, when electricity demand was greatest. Instead, in all three regions, the highest relative amount of wind generation occurred when loads were lowest, and the smallest amounts of wind were available when loads were greatest in summer. In PJM, this effect has been particularly pronounced, with a summer load – wind gap of almost 70 percent in summer 2010 and 2011, and 59 percent in summer 2012. Although we did not evaluate wind generation in the Southwest Power Pool (SPP), which has about 4,800 MW of installed wind capacity, the SPP Independent Market Monitor reports similar wind output behavior during peak load hours. In 2011, for example, wind availability during all peak hours averaged just over 15 percent, whereas in the hours where loads were lowest, wind 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 The Electricity Journal Author's personal copy [(Figure_2)TD$IG] Figure 2: ERCOT Load–Wind Gap, 2009–2012 availability averaged over 40 percent.18 ext, we evaluated availability ratios each N year during the hour when demand peaked on the 10 days with the highest greatest electricity demand in each RTO. We compared the median of the availability ratios in each year with the overall median availability over the entire [(Figure_3)TD$IG] Figure 3: MISO Load–Wind Gap, 2009–2012 Jan./Feb. 2013, Vol. 26, Issue 1 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 13 Author's personal copy [(Figure_4)TD$IG] Figure 4: PJM Load–Wind Gap, 2009–2012 Table 2: Median Wind Availability, Peak Demand Days and Overall. Year ERCOT 2009 14.2% 1.8% 14.6% 2010 6.0% 2.5% 8.2% 2011 2012 15.9% 14.0% 7.6% 7.2% 14.0% 13.8% Median, All-hours, All years 30.9% 27.0% 25.9% four-year period, based on the individual daily availability ratios.19 The results are shown in Table 2. s Table 2 shows, in MISO, median wind availability ranged between 1.8 percent and 7.6 percent of total installed wind capacity at the peak hour on the 10 highest-demand days. In ERCOT, median wind availability ranged between 6.0 percent and 15.9 percent. In PJM, the range was between 8.2 percent and 14.6 percent. As shown, these availability values are, at best, half A 14 MISO PJM the median availability for the entire period and, in the case of MISO, at best less than one-fourth of the median availability. From a system planning perspective, therefore, planners must assume that little wind generation will be available on the highest-demand days. inally, we examined wind generation based on its relation to an average daily load profile, both seasonally and over the entire year. This is shown in Figure 5, which compares average wind availability by hour in F ERCOT to average hourly electric demand over the entire four-year period, both in the summer season and on an average annual basis. As Figure 5 shows, average hourly loads in summer are higher than during the year overall, whereas average wind availability is lower in summer. Thus, we see the same high-load/ low-wind generation relationship: high-load hours are associated with low wind availability.20 IV. Policy Implications Our analysis shows that continued subsidies for wind generation represent both bad economics and bad energy policy, for at least three reasons. First and foremost, wind generation’s 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 The Electricity Journal Author's personal copy [(Figure_5)TD$IG] Figure 5: Summer Season and Annual Daily Wind Generation and Load Patterns production pattern not only is volatile and unpredictable, but even more significantly, has low economic value. Rather than displacing high-variable-cost fossil generating resources used to meet peak demand, wind generation’s observed availability peaks when electricity demand is lowest. As a result, wind generation tends to displace lowvariable-cost generation or simply forces baseload generators to pay greater amounts to inject power onto the grid because the units cannot be cycled cost-effectively. The low economic value of wind power is comparable to the government paying farmers to plow under high-value crops in order to plant low-value ones, or even weeds. Second, as with all subsidies, subsidized wind generation distorts electric markets by artificially lowering electric prices Jan./Feb. 2013, Vol. 26, Issue 1 in the short run, but leads to higher prices in the long run. This imposes economic harm on competitive generators and consumers, thus reducing economic growth. hird, because geographic dispersion of wind resources does not address inaccurate forecasts of wind availability, additional fossil generating resources are required to maintain system reliability. Moreover, geographic dispersion requires billions of dollars to be spent on additional transmission lines. These costs, along with most of the system integration costs, are socialized across all grid customers, that is, borne by all generators and, ultimately, consumers. In other words, wind generation imposes external costs on other market participants. After 35 years of direct and indirect subsidies, there is no T economic rationale for continued subsidization of wind generation. At the federal level, direct subsidies, such as the federal PTC, should not be continued. Statelevel subsidies, whether feed-in tariffs established by state regulators or statutory RPS mandates, further exacerbate market distortions and raise electricity prices, again to the detriment of consumers. Ultimately, continued subsidization of wind generation simply rewards the few at the expense of the many. Given a massive federal debt and anemic economic recovery, this type of pernicious redistribution cannot be justified.& Endnotes: 1. More recently, payments to the wind industry have increased still further with billions of dollars in additional monies paid-out as part of 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 15 Author's personal copy the $831 billion American Recovery and Reinvestment Act of 2009 (ARRA). 2. M. Sherlock, CRS. ‘‘Impact of Tax Policies on the Commercial Application of Renewable Energy Technology,’’ Statement Before the House Committee on Science, Space, and Technology, Subcommittee on Investigations and Oversight & Subcommittee on Energy and Environment, April 19, 2012, p. 3. 3. The reasons why are discussed in the next section. increase. Moreover, the threat of intervention raises the expected costs of market entry, leading to higher long-run market prices than would prevail in the absence of subsidies. For a discussion of subsidies and price suppression in organized capacity markets, see Briggs, Robert, and Andrew N. Kleit, Resource Adequacy and the Impacts of Capacity Subsidies in Competitive Electricity Markets, Working Paper, Dept. of Energy and Mineral Engineering, Pennsylvania State University, 4. The ‘‘infant industry’’ argument historically was used to justify protection of domestic firms from international trade. It was first developed by Alexander Hamilton at the beginning of the nineteenth century. A classic article discussing why infant industries should not be protected is Robert Baldwin, ‘‘The Case Against Infant Industry Protection,’’ Journal of Political Economy 75 (1969), pp. 295–305. 5. Arguments that subsidies account for external costs incorrectly assume that the effects of subsidies and taxes are equivalent. They are not. See William Baumol and Wallace Oates, The Theory of Environmental Policy, 2d ed., (Cambridge: Cambridge University Press 1988). See also, Daniel Dodds and Jonathan Lesser, ‘‘Can Utility Commissions Improve on Environmental Regulations,’’ Land Economics 70 (1994), pp. 63–76. 6. There is an extensive literature on the effects of subsidies in agriculture, energy, housing, environmental quality, and so forth. General discussions on the impacts of subsidies on markets can be found in any intermediate microeconomics textbook. 7. Equivalently, the marginal cost of cycling the plant is greater than the variable operating cost. Hence, it is economically rational to continue operation. 8. Some argue that price suppression ‘‘benefits’’ consumers. While subsidies can reduce market prices in the very short-run, markets are dynamic. Thus, as competitors are driven out, prices 16 Encourage Wind Energy Predicated on a Misleading Statistic?’’ The Electricity Journal 25 (April 2012), pp. 42–54. 14. Source: SNL Financial. Data through August 31, 2012. 15. We defined the ‘‘Winter’’ season contiguously. Thus, for example, Winter 2012 is defined as the three months December 2011 through February 2012. 16. In ERCOT, the correlation coefficient between hourly wind availability and annual hourly loads is 0.83. The correlation coefficient for the Summer season is 0.74. 17. We used wind capacity data as published by SNL Financial, which provided wind capacity installed in each month of the 44-month analysis period. The capacity used to calculate wind availability in month m was the amount of reported capacity installed at the end of month m 1. Oct. 22. 2012. http://papers.ssrn. com/sol3/papers.cfm?abstract_id= 2165412. 9. This value is based on a federal corporate tax rate of 35%. 10. NREL, Eastern Wind Integration and Transmission Study, NREL/SR550-47086, Revised February 2011. http://www.nrel.gov/docs/ fy10osti/47086.pdf. 11. This value is based on total wind generation of just over 28.2 million MWh in ERCOT in 2011. According to the US Energy Information Administration, total wind generation was about 120 million MWhs in 2011. 12. Public Utilities Commission of Texas, Competitive Renewable Energy Zone Program (CREZ) Oversight, CREZ Progress Report No. 8, July 2012, p. 6. http:// www.texascrezprojects.com/ page2960039.aspx. 13. Forbes, Kevin, Marco Stampini, and Ernest Zampelli, ‘‘Are Policies to 18. SPP, Independent Market Monitor, 2011 State of the Market, July 9, 2012, pp. 59–60. The Independent Market Monitor reports that similar wind availability patterns— decreasing availability as load increased—were observed in the three previous years. 19. The median was selected as a more representative planning value for the data. Consider a simple (albeit extreme) example: suppose wind availability was 0% on nine of the ten days, but 100% on the 10th day. In that case, the median availability would be 0% and the average availability would be 10%. However, system planners who assumed 10% wind availability each day in order to schedule generating resources would have to rely on replacement generation on nine of the days, and be forced to back down generation on the 10th. In contrast, using the median availability, system planners would only have to back down generation on the 10th day. 20. The correlation coefficient between average annual hourly wind availability and average annual hourly load is 0.83. The correlation coefficient for the Summer season is 0.74. 1040-6190/$–see front matter # 2012 Elsevier Inc. All rights reserved., http://dx.doi.org/10.1016/j.tej.2012.11.015 The Electricity Journal The Economic Impacts of U.S. Shale Gas Production on Ohio Consumers Prepared by: Continental Economics, Inc. January 2012 Copyright © 2012, Continental Economics, Inc. The information contained in this document is the exclusive, confidential and proprietary property of Continental Economics, Inc. and is protected under the trade secret and copyright laws of the U.S. and other international laws, treaties and conventions. No part of this work may be disclosed to any third party or used, reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage or retrieval system, without prior express written permission of Continental Economics, Inc. The Economic Impacts of Shale Gas on Ohio Consumers January 2012 TABLE OF CONTENTS Executive Summary I. Introduction ........................................................................................................................... 1 II. Historical Overview ............................................................................................................... 2 A. The Emergence of Shale Gas ............................................................................................... 5 B. Overview of Shale Gas Production Economics ................................................................... 8 C. Natural Gas Prices and Demand ........................................................................................ 10 D. Trends in U.S. and Ohio Natural Gas Demand .................................................................. 12 E. Natural Gas Prices and Ohio Consumers’ Energy Costs. .................................................. 13 III. Estimating the Impacts of Shale Gas Production on U.S. Wellhead Natural Gas Prices and Ohio Consumers’ Energy bills .................................................................................... 14 A. The Impacts of Shale Gas on Wellhead Natural Gas Prices .............................................. 14 B. Impacts on Ohio Natural Gas Consumers .......................................................................... 17 Appendix 1: Econometric Model Specification The Economic Impacts of Shale Gas on Ohio Consumers EXECUTIVE SUMMARY January 2012 † Shale gas has fundamentally changed the U.S. energy picture, providing a boon in an otherwise moribund economy. A decade ago, shale gas and liquids production were inconsequential. As the gas supply “bubble” of the 1990s ended and crude oil prices accelerated, so did wellhead natural gas prices, because of the historic linkage between the prices of the two fuels. In 2005, the damage caused by Hurricanes Katrina and Rita to the U.S. Gulf natural gas supply infrastructure caused a further spike in wellhead prices, and concerns grew that natural gas prices would continue to escalate. As shale gas production has accelerated, U.S. natural gas prices have plummeted. Although the severe economic recession that began in late 2008 and the resulting decrease in the demand for natural gas have contributed to lower wellhead natural gas prices, much of that price decrease stems from the rapid increase in domestic shale gas supplies, which increased almost tenfold between 2005 and 2010. The rapid expansion of shale gas production in the United States has created hundreds of thousands of new jobs directly and in supporting industries. The effect of this expansion on people and communities within the geographic areas of the shale plays has received considerable attention. However, domestic shale gas developments have also been the catalyst for far broader economic benefits throughout the country. More specifically, the lower wellhead natural gas prices that have resulted from this expanding shale gas production have lowered businesses’ and consumers’ energy bills, not only for natural gas, but also for electricity, an increasing percentage of which is generated from natural gas. Without seeking to divert attention away from the important economic development and retention benefits that shale gas development has had or will have on local populations and communities, this report provides information about the broader beneficial dividends that shale development is paying to the public at large. While conventional natural gas production in the U.S. has decreased over time, shale gas has become a rapidly increasing source of U.S. gas supplies, accounting for about 20 percent of total U.S. onshore domestic natural gas production in 2010. The U.S. Energy Information Administration (“EIA”) forecasts that, by 2035, shale gas could account for over 50 percent of onshore natural gas production. Of greater interest for Ohioans is the Utica Shale, which lies beneath the better-known Marcellus Shale, and extends into the eastern half of the state. Although reserve data is based on preliminary drilling in the Utica Shale, geologists expect the Utica Shale to be relatively rich in † Funding for this report was secured through the Industrial Energy Users-Ohio (IEU-Ohio), an Ohiobased organization of customers that helps customers address issues affecting the price and availability of energy. Information on IEU-Ohio is available at http://ieu-ohio.org EX-1 The Economic Impacts of Shale Gas on Ohio Consumers January 2012 oil and natural gas liquids that are currently worth significantly more than natural gas on an energy-equivalent basis. Preliminary estimates by Ohio's Department of Natural Resources (ODNR) suggest a recoverable reserve potential of between 1.3 and 5.5 billion barrels of oil as well as 3.8 to 15.7 trillion cubic feet (“Tcf”) of natural gas. The overall economic value of the Utica Shale region in Ohio may be especially large, because it lies relatively close to the surface, which reduces exploration and development costs.1 Although overall natural gas consumption in Ohio has decreased since 1997 (in part because of reductions in the energy intensity of Ohio’s economy), expenditures on natural gas remain significant. In 2009, Ohio consumers and businesses, including electric generators, consumed 724 billion cubic feet (“Bcf”) of natural gas, at a cost of $7.46 billion. Thus, lower natural gas prices owing to shale gas production can have real benefits for Ohio energy consumers as well as the public at large. To estimate how much shale gas has contributed to the decline in wellhead natural gas prices and how those price decreases have flowed through to benefit Ohio’s natural gas consumers, Continental Economics developed a model to isolate the impacts of shale gas on wellhead prices. Then, using the results of that model, we determined the savings to different classes of Ohio consumers.2 Our analysis showed that, for each Tcf of shale gas produced, the average annual wellhead price is $0.46 per thousand cubic feet (“Mcf”) lower that it otherwise would be. Equivalently, the average wellhead price would be $0.46 per Mcf higher for each Tcf of shale gas not otherwise produced. The results of our analysis are shown in Figure EX-1 and Table EX-1 on the following page. As the table shows, the impact of shale gas production on wellhead gas prices has increased steadily as shale gas supplies have increased relative to total natural gas supplies. For example, in 2010 we estimate that shale gas production, which was over 4.7 Tcf, caused observed average wellhead natural gas prices to be $2.43 per Mcf lower than what they would have otherwise been. 1 The depth of shale gas deposits below the surface is not uniform. All other things equal, the closer to the surface, the lower are exploration and development costs. 2 We will address the impacts of lower wellhead natural gas prices on wholesale and retail electricity prices for Ohio consumers in a subsequent report. EX-2 The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Figure EX-1: Estimated Annual Wellhead Natural Gas Prices Without Shale Gas (1990–2010) Table EX-1: Estimated Annual Price Impact of Shale Gas Production (1990-2010) Year Price Reduction ($/Mcf) Year Price Reduction ($/Mcf) 1990 ($0.01) 2001 ($0.13) 1991 ($0.02) 2002 ($0.17) 1992 ($0.02) 2003 ($0.20) 1993 ($0.03) 2004 ($0.24) 1994 ($0.04) 2005 ($0.30) 1995 ($0.05) 2006 ($0.43) 1996 ($0.07) 2007 ($0.70) 1997 ($0.09) 2008 ($1.14) 1998 ($0.09) 2009 ($1.68) 1999 ($0.09) 2010 ($2.43) 2000 ($0.11) Based on the results of the analysis described above and average use per customer data for 2010, Table EX-2 provides an estimate of the resulting natural gas energy bill reductions for Ohio commercial, industrial, and residential customers. EX-3 The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Table EX-2: Estimated Annual Cost Savings for Ohio End-Use Customers Customer Class Commercial Industrial Residential Total Average Use Per Customer (Mcf) Price Reduction ($/Mcf) 2010 Estimated Cost Savings Number of Customers Estimated Savings (Millions of $) 562.1 ($2.43) $1,366 258,422 $353.0 35,266.8 ($2.43) $85,698 5,738 $491.7 88.2 ($2.43) $214 3,198,883 $685.4 $1,530.2 As this table shows, we estimate that Ohio businesses and consumers saved over $1.5 billion on their natural gas bills in 2010 because of lower wellhead natural gas prices. The average residential customer, for example, burned 88 Mcf of natural gas and saved $214 in 2010. The average commercial customer used 562 Mcf and saved $1,366, while the average industrial customer used over 35,000 Mcf and saved almost $87,000. In addition, electric generators reduced their costs because of lower wellhead gas prices. This translated into lower fuel charges levied by electric utilities with fuel cost recovery mechanisms,3 such as Columbus Southern Power and Ohio Power Company, and also contributed to lower wholesale electric prices paid by retail electric suppliers. The results of our analysis demonstrate that shale gas production has significantly reduced U.S. wellhead natural gas prices and reduced Ohio consumers’ natural gas bills. The estimated savings of $1.5 billion on natural gas bills alone in 2010 affect all sectors of the Ohio economy. As Ohio’s Utica Shale gas resource is developed, Ohio businesses and consumers are likely to benefit even more in the future. Furthermore, because of the increasing importance of natural gas used in generating electricity, Ohio consumers are reaping even more benefits from lower electric bills. (In a subsequent report, we will present the estimated savings for Ohio consumers on their electric bills.) The decreases in natural gas and electricity prices will benefit the Ohio economy, not only by creating jobs directly in the shale gas development and extraction industries as the Utica Shale is developed, but by lowering home energy bills and improving the overall competitiveness of Ohio businesses and industry. 3 The default generation supply prices of Ohio Power and Columbus Southern Power (sometimes referred to as AEP-Ohio) continue to be administratively set by the Public Utilities Commission of Ohio (“PUCO”) based on a rate structure that includes a fuel adjustment clause (FAC). Other Ohio electric distribution utilities (“EDUs”) establish default generation supply prices through a competitive bidding process (“CBP”) conducted under the PUCO’s supervision. The downward pressure that shale gas development has placed on electric prices is observable from the inputs that go into the FAC as well as the pricing results of the CBPs that have been approved by the PUCO. EX-4 The Economic Impacts of U.S. Shale Gas Production on Ohio Consumers† I. INTRODUCTION Shale gas has fundamentally changed the U.S. energy picture, providing a boon in an otherwise moribund economy. A decade ago, shale gas and liquids production were inconsequential. As the gas supply “bubble” of the 1990s ended and crude oil prices accelerated, so did wellhead natural gas prices, because of the historic linkage between the prices of the two fuels. In 2005, the damage caused by Hurricanes Katrina and Rita to the U.S. Gulf natural gas supply infrastructure caused a further spike in wellhead prices, and concerns grew that natural gas prices would continue to escalate. As shale gas production has accelerated, U.S. natural gas prices have plummeted. Although the severe economic recession that began in late 2008 and the resulting decrease in the demand for natural gas have contributed to lower wellhead natural gas prices, much of that price decrease stems from the rapid increase in domestic shale gas supplies, which increased almost tenfold between 2005 and 2010. The rapid expansion of shale gas production in the United States has created hundreds of thousands of new jobs directly and in supporting industries. The effect of this expansion on people and communities within the geographic areas of the shale plays has received considerable attention. However, domestic shale gas developments have also been the catalyst for far broader economic benefits throughout the country. More specifically, the lower wellhead natural gas prices that have resulted from this expanding shale gas production have lowered businesses’ and consumers’ energy bills, not only for natural gas, but also for electricity, an increasing percentage of which is generated from natural gas. Without seeking to divert attention away from the important economic development and retention benefits that shale gas development has had or will have on local populations and communities, this report provides information about the broader beneficial dividends that shale development is paying to the public at large. † Funding for this report was secured through the Industrial Energy Users-Ohio (IEU-Ohio), an Ohiobased organization of customers that helps customers address issues affecting the price and availability of energy. Information on IEU-Ohio is available at http://ieu-ohio.org. -1- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Much attention has been paid to the jobs created by the shale gas industry, both directly and indirectly.1 Much less has been focused on these broader economic benefits provided by shale gas stemming from lower natural gas and electricity prices.2 The purpose of this report, therefore, is two-fold. First, we estimate the magnitude of the decrease in wellhead natural gas prices that has been caused by increased shale gas production. Second, we estimate how this decrease in wellhead natural gas prices has reduced natural gas expenditures by Ohio businesses and consumers.3 II. HISTORICAL OVERVIEW In the late 1960s, the conventional wisdom was that natural gas supplies would soon be exhausted. Wellhead natural gas prices were regulated and capped. Supplies began to diminish as the incremental cost of production exceeded the revenue available from regulated prices and production from existing wells declined. Growth in the natural gas industry came to a standstill because there was little economic incentive to undertake new, more costly exploration. By 1967, estimated domestic reserves had peaked and actual production began to fall steadily. Natural gas supply shortages on peak usage days began to occur. As these gas supply shortages became more common, in states like Ohio new customer hookups were unavailable, supplies for industrial and commercial customers were interrupted and curtailed, and predictions that “the spigot would run dry” within a decade became prevalent. By 1978, proved natural gas reserves had dropped by 30%. Something had to be done and policy makers turned to market-based strategies to balance natural gas supply and demand. First, as part of comprehensive energy legislation that year, Congress passed the Natural Gas Policy Act (“NGPA”), which began to dismantle the complex historic system of natural gas price regulation that stemmed from a 1954 decision by the United States Supreme Court decision regarding the meaning of the Natural Gas Act passed in 1938.4 For example, Ohio initiated the Natural Gas 1 See, e.g., Kleinhenz and Associates, “Ohio‘s Natural Gas and Crude Oil Exploration and Production Industry and the Emerging Utica Gas Formation,” (Sept., 2011). http://www.oogeep.org/downloads/file/Economic%20Impact%20Study/Ohio%20Natural%20Gas%20and %20Crude%20Oil%20Industry%20Economic%20Impact%20Study%20September%202011.pdf. For a different viewpoint on job creation impacts, and the costs and benefits of shale gas development in Ohio, see A. Weinstein and M. Partridge, “The Economic Value of Shale Natural Gas in Ohio,” Department of Agricultural, Environmental, and Development Economics, Ohio State University, December 2011. http://go.osu.edu/shalejobs. 2 In this report, we do not analyze employment impacts, which are not benefits per se. However, businesses and consumers clearly do benefit from lower energy prices. 3 We will address the impacts of lower wellhead natural gas prices on wholesale and retail electricity prices in a subsequent report. 4 The Natural Gas Act (“NGA”) of 1938 allowed the Federal Power Commission (“FPC”), the precursor to the Federal Energy Regulatory Commission (“FERC”), to regulate the prices charged by interstate natural gas pipelines. From 1938 to 1954, the FPC did not regulate wellhead natural gas prices. Instead, independent producers sold natural gas to interstate pipelines at unregulated prices, -2- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Self-Help Program5 that allowed retail customers to develop and obtain their own natural gas supply and then use the unbundled delivery capabilities of natural gas companies to transport the gas supply from the wellhead to the point of utilization. The NGPA provided for the gradual elimination of a labyrinth of price rules, and full decontrol of prices was achieved by 1993.6 Not surprisingly, eliminating price controls and creating a truly competitive market for natural gas supplies created the economic incentives needed for renewed natural gas exploration and development. Coupled with FERC’s severing of the traditional connection between production, pipeline transportation, and distribution in 1992,7 by the early 1990 the natural gas market was vibrant; the predicted shortages had turned into a gas supply “bubble” that led to much lower prices. Those lower prices, in turn, spurred development of competitive wholesale electricity markets that were envisioned under the Energy Policy Act of 1992, as well as calls for electric industry restructuring, because of advances in gas-fired generating technologies, such as combined-cycle units that were energy efficient and could be constructed more quickly than traditional coal-fired or nuclear baseload generating plants. As a result of the increasing reliance on natural gas-fired generation, the demand for natural gas has increased since the mid-1990s. Between 1997 and 2010, for example, total natural gas consumption increased by about six percent, whereas natural gas consumption for electric power generation increased by 80%.8 __________________________ (cont.) with any subsequent sales for resale being regulated by the FPC. Advocates who sought to keep consumer prices low through price regulation went to the FPC to close what they alleged was a regulatory “loophole,” because the FPC exempted wellhead sales from price regulation as “production and gathering” activities. The FPC rejected this claim, but in 1954, the U.S. Supreme Court, in Phillips Petroleum v. Wisconsin, 347 U.S. 672 (1954), reversed the FPC, ruling that the NGA applied not only to pipelines, but also to natural gas producers. This led to the FPC regulating natural gas wellhead prices and creating the natural gas “shortages” beginning in the late 1960s. 5 Ohio’s Self-Help Program was one of the first unbundled natural gas transportation programs in the Nation and during the 70’s it sparked a surge of exploration and development activities in Ohio. It also served as a model for other unbundled open access transportation programs that evolved to provide the foundation for the natural gas industry structure that is in place today. 6 In 1989, Congress passed the Natural Gas Wellhead Decontrol Act, which fully decontrolled wellhead prices as of January 1, 1993. 7 Pipeline Service Obligations and Revisions to Regulations Governing Self-Implementing Transportation Under Part 284 of the Commission’s Regulations, Regulation of Natural Gas Pipelines After Partial Wellhead Decontrol, FERC Order No. 636, 59 FERC ¶ 61,030 (1992). 8 Source: U.S. Energy Information Administration, Natural Gas Annual, 2011. -3- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 The Many Sources of Natural Gas Natural gas is produced from a number of sources. The diagram below is a schematic of the different types of gas and their relative locations underground. Nearest the surface is coalbed methane, which is just natural gas found in coal seams.9 Associated gas is natural gas that is found on top of crude oil deposits. Often, crude oil wells produce both crude oil, natural gas, and so-called “natural gas liquids” (“NGLs”), which are valuable types of hydrocarbons, such as propane and butane. In other cases, natural gas is found in separate deposits, called non-associated gas. Further below the surface, one finds “tight-gas.” Tight gas is natural gas that has migrated upwards into sandstone formations and which, because sandstone has low permeability,10 cannot migrate further. Further below still lies shale gas. Although market forces had eliminated fears of natural gas shortages, wellhead prices remained linked to world crude oil prices. Thus, when the events of September 11, 2001, and the subsequent invasions of Afghanistan and Iraq, led to a rapid increase in crude oil prices, natural gas prices followed; the gas supply “bubble” had burst. Natural gas prices increased, spiking in 2005 because of the damage caused by Hurricanes Katrina and Rita to the Gulf Coast gas supply 9 Coal seams are often saturated with water, and the pressure of that water forces methane (natural gas) into the coal. When the water is removed, the pressure drops, and natural gas can be extracted. 10 “Low permeability” means that gas molecules do not flow easily. For shale gas, that is the reason producers use hydraulic fracturing techniques to release the natural gas and let it easily flow to their wells. -4- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 infrastructure, and causing renewed fears that U.S. natural gas production once again faced inexorable decline and that consumers would face increasing prices.11 The search was on for new sources of natural gas. One of the first was liquefied natural gas (LNG) that could be imported from the Middle East, where gas was still considered a waste byproduct from crude oil production and simply flared (burned) off. Plans for huge new facilities capable of receiving LNG were developed, but many such facilities faced intense siting opposition because of the perceived risks, such as explosions. Other “unconventional” domestic natural gas supplies also emerged. By the mid-1990s, for example, production of coal-bed methane (“CBM”) had increased to about 1 trillion cubic feet (“Tcf”) per year, or about four percent of total U.S. natural gas production. In 2008, coal-bed methane production peaked at just under 2 Tcf. The other unconventional resource—and the one that has already provided huge economic benefits—is shale gas, which now accounts for over 20 percent of all total domestic natural gas production. A. The Emergence of Shale Gas By the late 1970s, natural gas was already known to exist in deep shales, such as the Barnett in Texas and Marcellus in Pennsylvania (Figure 1). Figure 1: U.S. Shale Gas Plays 11 The same was true for natural gas supplies exported to the U.S. from western Canada, which had also increased over the previous decade. -5- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 However, the technology to retrieve gas from such “low-permeability” areas did not exist. It was not until the 1980s that improvements in hydraulic fracturing, a drilling technique that had been widely used since the 1940s to enhance production in existing oil and gas wells, began to change the economics of shale gas. The U.S Energy Information Administration (“EIA”) has tracked shale gas production for each of the major shale gas plays since 1990 (Figure 2). As can be seen in Figure 2, shale gas production began to accelerate rapidly after the year 2000, as production ramped up in the Barnett shale of Texas. By the middle of the decade, production in the Fayette and Haynesville regions began to increase. Most recently, production in the Marcellus region, which is estimated to have far larger reserves, has begun to accelerate. Figure 2: U.S. Annual Shale Gas Production, (1990 – 2010) -6- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Figure 3: U.S. Annual Natural Gas Production, by Source (1990–2010) While conventional natural gas production in the U.S. has decreased over time, shale gas has become a rapidly increasing source of U.S. gas supplies (Figure 3), and now accounts for about 20 percent of total U.S. onshore domestic natural gas production. The EIA forecasts that, by 2035, shale gas could account for over 50 percent of onshore natural gas production.12 Of greater interest for Ohioans is the Utica Shale, which lies beneath the better-known Marcellus Shale, and extends into the eastern half of the state. Although reserve data is based on preliminary drilling in the Utica Shale, geologists expect the Utica Shale to be relatively rich in oil and natural gas liquids that are currently worth significantly more than natural gas on an energy-equivalent basis. Preliminary estimates by Ohio's Department of Natural Resources (ODNR) suggest a recoverable reserve potential of between 1.3 and 5.5 billion barrels of oil as well as 3.8 to 15.7 trillion cubic feet (“Tcf”) of natural gas. The overall economic value of the Utica Shale region in Ohio may be especially large, because it lies relatively close to the surface, which reduces exploration and development costs.13 12 http://www.eia.gov/forecasts/aeo/IF_all.cfm#prospectshale. 13 The depth of shale gas deposits below the surface is not uniform. All other things equal, the closer to the surface, the lower are exploration and development costs. -7- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Figure 4: Marcellus and Utica Shale Plays in Ohio B. Overview of Shale Gas Production Economics As we discuss in the next section, shale gas production has helped reduce wellhead natural gas prices. But what factors affect shale gas production? It turns out, there are a number of factors, including not only day-to-day production costs, but also the costs of leasing land, the productivity of the wells drilled, and the mix of natural gas and NGLs produced. The economic benefits of drilling an oil or gas well—and, often, the same well produces both— depend on a number of factors. Broadly, these are the expected future revenues from what the well produces, and the fixed and variable production costs of drilling and operating the well. Expected future revenues depend on how much a typical well is likely to produce over its lifetime and future prices. For example, wells that produce both crude oil and NGLs tend to be more profitable than wells producing just natural gas, given current and expected prices. The reason is that world crude oil prices are much higher (on a Btu basis, i.e., the price per million Btus, based on the relative heat content of oil and natural gas) than the price of natural gas. Similarly, some NGLs, such as propane and butane, tend to sell at higher market prices than methane, which is the major component of what we term “natural gas.” Thus, all else equal, a developer is more likely to drill where natural gas is likely to be found with crude oil and natural gas liquids. -8- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Production costs can be broken down into fixed and variable costs. Fixed costs are those that do not change with the quantity of oil or gas produced, such as the cost of obtaining a lease. For example, recent data shows that leases for land in Athens County are costing developers $2.500 per acre.14 The costs of these leases, together with the cost of leasing the actual drilling equipment, are the largest fixed costs associated with well-drilling. Variable costs of production are those that depend on how much oil and natural gas is produced. For example, the state of Ohio levies a severance tax on crude oil and natural gas producers of$0.10 per barrel of oil and $0.0025 per thousand cubic feet of natural gas produced, regardless of the market prices.15 Landowners typically assess a royalty fee on producers that, unlike the state severance tax, is based on the value of natural gas produced. Finally, there are the direct variable production costs, such as the cost of operating the well equipment every day. Because a significant portion of the overall production costs are fixed, drillers will often continue to produce oil and natural gas from wells even when the average production costs are greater than market prices, which can tend to further decrease market prices. Estimates of the overall average production cost of shale gas wells vary widely, because shale gas plays differ in their characteristics, such as depth. Typically, drilling costs are reported on a per-foot basis. Thus, the equivalent cost per MMBtu of natural gas produced depends on how deeply a well is drilled, and the well’s average daily production. Publicly available data on the costs of shale gas wells, and production costs per MMBtu, are difficult to obtain. Moreover, because of technological advances, production costs continue to decrease. A 2010 report by the World Energy Council states that estimates of average shale gas production costs in North America range between $4 per Mcf and $8 per Mcf.16 However, in 2010, Chesapeake Energy estimated average direct production expenses, including taxes, of just over $1 per Mcf.17 It reports another $0.44 per Mcf in administrative and general costs, and $1.56 per Mcf in depreciation and amortization costs, for an overall average cost of about $3 per Mcf. Moreover, because shale gas resources tend to be located near demand centers, transportation costs on natural gas pipelines can be less than for natural gas sourced from traditional supply basins, such as the Rocky Mountains, Western Canada, and the Gulf Coast. 14 “County oil and gas leasing just goes on & on,” The Athens (OH) News, December 15, 2011. 15 http://codes.ohio.gov/orc/5749. One thousand cubic feet (“Mcf”) is approximately 1.04 million Btus (MMBtu). One barrel of oil has an average heat content of 5.6 MMBtus. Thus, on a per-Btu basis, Ohio levies a slightly higher severance tax on oil than natural gas. 16 World Energy Council, “Survey of Energy Resources: Focus on Shale Gas,” 2010, page 14. http://www.worldenergy.org/documents/shalegasreport.pdf 17 Chesapeake Energy, 2010 Annual Report, page 4. http://phx.corporateir.net/External.File?item=UGFyZW50SUQ9OTEzODB8Q2hpbGRJRD0tMXxUeXBlPTM=&t=1 -9- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 C. Natural Gas Prices and Demand The gas supply “bubble” of the 1990s caused an extended period of low natural gas prices, with prices generally less than $2 per Mcf (Figure 5). Starting in 2001, however, gas prices, which were historically linked closely with crude oil prices, began to increase rapidly, in response to the events of September 11, 2001, and the subsequent invasions of Afghanistan and Iraq, which caused crude oil prices to increase rapidly (Figure 6). Wellhead natural gas prices peaked in 2005 at an average of over $7 per Mcf, in part because of the damage to the production and gathering infrastructure along the U.S. Gulf Coast caused by Hurricanes Katrina and Rita, and continued increases in natural gas demand, especially for generating electricity. Prices then decreased to about $6 per Mcf in 2006 and 2007. However, in 2008, prices spiked to their highest annual level ever, about $8 per Mcf, caused by increased demand and surging crude oil prices.18 Figure 5: Average Annual U.S. Wellhead Natural Gas Prices (1990–2010) The rapid increase in U.S. shale gas production has more than compensated for decreases in conventional natural gas production from oil and gas wells, because advances in drilling technology have made the economics of shale gas production so favorable. In fact, according to 18 The June 2008 wellhead price was $10.79 per Mcf, the highest nominal value ever. Historically, U.S. natural gas prices and crude oil prices were closely linked, owing to the substitutability of oil and natural gas. As discussed below, because of shale gas, that historic link is much weaker. -10- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 IHS CERA, the cost of producing shale gas is now less than the cost of producing “conventional” natural gas.19 Another important benefit of the rapid increase in shale gas production has been to weaken the historical link between wellhead natural gas prices and volatile world crude oil prices. Because of the limited ability to export natural gas overseas, increased domestic production has reduced wellhead natural gas prices, even though world crude oil prices remain in the $100 per barrel (“Bbl”) range.20 For example, between January 2009 and August 2011, the price of Brent crude (one of the world’s benchmark oil prices) tripled, rising from under $40 per Bbl to almost $120 per Bbl. During that same period, U.S. wellhead natural gas prices remained relatively constant (Figure 6).21 Thus, while natural gas prices and crude oil prices in Europe and Asia continue to move in tandem, the weakening of the historic link in the U.S. is providing consumers with the economic benefits of lower and more stable natural gas prices. 19 “Fueling America’s Energy Future,” IHS Cambridge Energy Research Associates Report, 2010. A copy of the Executive Summary may be found at: http://groundwork.iogcc.org/sites/default/files/IHS%20CERA%20Executive%20Summary%20Fuelin g%20North%20America%27s%20Energy%20Future%20The%20Unconventional%20Natural%20Ga s%20Revolution%20and%20the%20Carbon%20Agenda.pdf. 20 LNG facilities were developed to import natural gas, not export it. Today, some of these facilities are being modified so that natural gas can be delivered to them, and then exported, but the cost is high. 21 A portion of the increase in the Brent crude price can be explained by decreases in the value of the U.S. dollar relative to other major world currencies. According to data published by the U.S. Federal Reserve, between January 2009 and August 2011, the dollar declined in value by 15% relative to these other currencies. -11- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Figure 6: U.S. Monthly Wellhead Natural Gas Prices and Brent Crude Price (2009–2011) D. Trends in U.S. and Ohio Natural Gas Demand The increase in the production of shale gas has also outstripped the overall growth in the demand for natural gas (Figure 7). This has also contributed to the decrease in wellhead natural gas prices since 2008 and provided greater price certainty. As shown in Figure 7, between 1997 and 2010, total natural gas delivered to domestic customers increased from 20.8 Tcf to 22.1 Tcf. Between 1997 and 2010, residential and commercial sector demand remained essentially constant, despite increases in the number of customers, because of increases in the energy efficiency of space and water heating equipment. Industrial demand during this same time period decreased by just over 22%, owing to a general decrease in manufacturing output, as well as improved energy efficiency. However, the demand for natural gas by electric generators increased over 80% between 1997 and 2010, reflecting the rapid increase in gas-fired generating capacity. -12- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Figure 7: U.S. Annual Natural Gas Deliveries, by Customer Class Unlike the U.S. as a whole, natural gas consumption in Ohio decreased during this period, especially in the industrial sector, where demand fell by about one-third between 1997 and 2009, reflecting the loss of heavy manufacturing industry in the state during this period. E. Natural Gas Prices and Ohio Consumers’ Energy Costs Although overall natural gas consumption in Ohio has decreased since 1997 (in part because of reductions in the energy intensity of Ohio’s economy), expenditures on natural gas remain significant. In 2009, Ohio consumers and businesses, including electric generators, consumed 724 billion cubic feet (“Bcf”) of natural gas, at a cost of $7.46 billion, as shown in Table 1.22 Thus, lower natural gas prices owing to shale gas production can have real benefits for Ohio energy consumers as well as the public at large. 22 Source: EIA, State Energy Data System. http://www.eia.gov/state/seds/seds-states.cfm?q_state_a=OH&q_state=Ohio. -13- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Table 1: Summary of Ohio 2009 Natural Gas Consumption and Expenditures Sector Consumption Cost Average Price (Bcf) (Millions of $) ($/Mcf) Residential 292 $3,708 $12.70 Commercial 161 $1,676 $10.41 Industrial 233 $1,908 $8.19 38 $166 $4.36 724 $7,458 $10.30 Electric Power Totals Of the total expenditure, residential customers spent $3.7 billion, about half of the total, and paid an average delivered retail price of $12.70 per Mcf. That delivered price includes the wholesale cost of gas, which reflects the wellhead price, plus the cost of transportation via pipeline and the cost of retail distribution. The reason that the average price for electric generators is so much lower than other customers is that most electric generators are directly interconnected to interstate pipelines, thus avoiding all of the costs associated with retail distribution. III. ESTIMATING THE IMPACTS OF SHALE GAS PRODUCTION ON U.S. WELLHEAD NATURAL GAS PRICES AND OHIO CONSUMERS’ ENERGY BILLS Shale gas has contributed to the decline in wellhead natural gas prices, but by how much? And, how does the wellhead price decrease caused by shale gas translate to savings for Ohio natural gas consumers? To answer these questions, we developed a model to isolate the impacts of shale gas on wellhead prices. Then, using the results of that model, determined the savings to different classes of Ohio consumers. A. The Impacts of Shale Gas on Wellhead Natural Gas Prices Natural gas supplies reflect complex relationships between expectations of future demand, market prices, and technology. Moreover, because significant quantities of natural gas are produced from oil wells, supplies are also influenced by expectations about crude oil markets. The U.S. EIA, for example, uses a complex set of interdependent models to prepare forecasts of natural gas production and prices.23 The EIA models combine engineering relationships, such as exploration costs per drilled foot, with econometric models and economic projections, to determine the economic returns from exploration and development. These factors further interact with demand projections, which are based on macroeconomic forecasts of the U.S. and world economies. 23 The EIA crude oil and natural gas supply model (“OGSM”) is part of its larger National Energy Modeling System (“NEMS”). Documentation for the OGSM can be downloaded at: http://www.eia.gov/FTPROOT/modeldoc/m063(2011).pdf -14- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 For the purposes of this report, however, it would be difficult to modify this type of modeling approach to determine what historic natural gas prices would have been without production from shale gas. Thus, we developed an econometric framework that models annual natural gas supply, demand, and average wellhead prices, and which isolates the impacts of shale gas production on wellhead prices. (A detailed description of the model can be found in the Appendix.) The advantages of an econometric approach include its relative transparency: factors that influence natural gas supply and demand, such as the price of crude oil and the delivered price of coal used by electric generators, are easily modeled and evaluated. The disadvantages of the econometric framework used here is that it cannot incorporate all of the variables that affect natural gas supply and demand.24 Once the model was estimated, we evaluated how well it predicted wellhead natural gas prices (Figure 8). As this figure shows, the model predicted natural gas prices that closely follow the actual annual prices. Figure 8: Actual v. Predicted Wellhead Natural Gas Price (1990 – 2010) To estimate the impact of shale gas production on average wellhead prices, we used the estimated relationship between wellhead prices and the quantity of natural gas produced. Specifically, our analysis showed that, for each Tcf of shale gas produced, the average annual 24 In economic terms, we have developed a “partial equilibrium” model, rather than a “general equilibrium” one. -15- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 wellhead price would be $0.46 per Mcf lower. Equivalently, the average wellhead price would be $0.46 per Mcf higher for each Tcf of shale gas not otherwise available. The results of the analysis are shown in Figure 9 and Table 2. Figure 9: Estimated Annual Wellhead Natural Gas Prices Without Shale Gas (1990 – 2010) Table 2: Estimated Annual Price Impact of Shale Gas Production (1990-2010) Year Price Reduction ($/Mcf) Year Price Reduction ($/Mcf) 1990 ($0.01) 2001 ($0.13) 1991 ($0.02) 2002 ($0.17) 1992 ($0.02) 2003 ($0.20) 1993 ($0.03) 2004 ($0.24) 1994 ($0.04) 2005 ($0.30) 1995 ($0.05) 2006 ($0.43) 1996 ($0.07) 2007 ($0.70) 1997 ($0.09) 2008 ($1.14) 1998 ($0.09) 2009 ($1.68) 1999 ($0.09) 2010 ($2.43) 2000 ($0.11) -16- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 As Table 2 shows, the impact of shale gas production on wellhead gas prices has increased steadily as shale gas supplies have increased relative to total natural gas supplies. In 2010, for example, we estimate that shale gas production, which was over 4.7 Tcf, caused observed average wellhead natural gas prices to be $2.43 per Mcf lower than what they would have otherwise been. To gauge the reasonableness of this price impact, we reviewed a 2004 analysis prepared by the EIA at the request of Representative Barbara Cubin, Chairman of the Subcommittee on Energy and Mineral Resources of the U.S. House Committee on Resources. The EIA analysis examined the projected impacts on U.S. natural gas production and wellhead prices under three different “low-supply” scenarios, and a combination of all three scenarios:25 No increased availability of Alaska natural gas; No significant increase in production of tight sands natural gas (or other nonconventional sources); and Inability to permit more than three additional average-sized liquefied natural gas offloading facilities. The EIA study estimated that, in 2010, the combination of these three restrictive supply assumptions would increase the average wellhead price of natural gas by $0.47 per Mcf (2002$) and reduce production in the lower 48 states by 0.96 Tcf. This implies an average increase of $0.49 (2002$) per Tcf of reduced production. Adjusting for inflation to 2010 dollars, this translates into an average price impact of $0.59 per Mcf for each Tcf reduction in natural gas production in the lower 48 states. As discussed above, we estimated a somewhat smaller price impact, $0.46 per Mcf for each Tcf reduction in gas supplies. B. Impacts on Ohio Natural Gas Consumers Natural gas retail distribution customers typically pay for natural gas on a pass-through basis. Thus, if the wholesale price increases by, say, 10 cents per Mcf, the retail customer will see an additional 10 cent charge on his bill. Thus, we believe it reasonable to assume that the full impacts of the wellhead price reductions stemming from increased production of shale gas would be fully reflected on customers’ bills. Based on the analysis described above, Table 3 provides an estimate of the resulting natural gas energy bill reductions for commercial, industrial, and residential customers, using average use per customer data from 2010. 25 EIA, Office of Integrated Analysis and Forecasting, “Analysis of Restricted Supply Cases,” February 2004. http://www.eia.gov/oiaf/servicerpt/ngsupply/pdf/sroiaf%282004%2903.pdf . There do not appear to be any more recent EIA analyses that have estimated wellhead prices and production related specifically to unconventional gas. -17- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Table 3: Estimated Annual Cost Savings for Ohio End-Use Customers Customer Class Commercial Industrial Residential Total Average Use Per Customer (Mcf) Price Reduction ($/Mcf) 2010 Estimated Cost Savings Number of Customers Estimated Savings (Millions of $) 562.1 ($2.43) $1,366 258,422 $353.0 35,266.8 ($2.43) $85,698 5,738 $491.7 88.2 ($2.43) $214 3,198,883 $685.4 $1,530.2 As this table shows, we estimate that Ohio businesses and consumers saved over $1.5 billion on their natural gas bills in 2010 because of lower wellhead natural gas prices. The average residential customer, for example, burned 88 Mcf of natural gas and saved $214 in 2010. The average commercial customer used 562 Mcf and saved $1,366, while the average industrial customer used over 35,000 Mcf and saved almost $87,000. In addition, electric generators reduced their costs because of lower wellhead gas prices. This translated into lower fuel charges levied by electric utilities with fuel cost recovery mechanisms,26 such as Columbus Southern Power and Ohio Power Company, and also contributed to lower wholesale electric prices paid by retail electric suppliers.27 The results of our analysis demonstrate that shale gas production has significantly reduced U.S. wellhead natural gas prices and reduced Ohio consumers’ natural gas and electric bills. The estimated savings of $1.5 billion in 2010 affect all sectors of the Ohio economy. As Ohio’s Utica Shale gas resource is developed, Ohio businesses and consumers are likely to benefit even more in the future. The decreases in natural gas and electricity prices will benefit the Ohio economy, not only by creating jobs directly in the shale gas extraction industry as the Utica Shale is developed, but by improving the overall competitiveness of Ohio businesses and industry. 26 The default generation supply prices of Ohio Power and Columbus Southern Power (sometimes referred to as AEP-Ohio) continue to be administratively set by the Public Utilities Commission of Ohio (“PUCO”) based on a rate structure that includes a fuel adjustment clause or FAC. Other Ohio electric distribution utilities (“EDUs”) establish default generation supply prices through a competitive bidding process (“CBP”) conducted under the PUCO’s supervision. The downward pressure that shale gas development has placed on electric prices is observable from the inputs that go into the FAC as well as the pricing results of the CBPs that have been approved by the PUCO. 27 A subsequent report will estimate how much Ohio consumers saved on their electric bills because of the lower wellhead natural gas prices. -18- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Appendix 1: Econometric Model Specification This appendix provides supporting detail on the econometric model used to estimate the impact of shale gas production on U.S. wellhead natural gas prices. The appendix first discusses general issues in estimating econometric models of supply and demand, and how those issues affected the specification of the model we developed. We then present the details of the model itself, the data used to estimate it, and the results of the estimation. The Identification Problem One well-recognized problem in modeling supply and demand is called the “identification” problem.28 What this means is that, when modeling supply and demand, observed data can reflect changes in both, as shown in Figure A-1. Figure A-1: Identifying Supply and Demand Curves Regression line Price D1 observed D2 D3 D4 S4 S3 S2 S1 Quantity Figure A-1 illustrates four annual supply-demand equilibrium points, each corresponding to a different supply-demand curve combination. In this example, the four observed points do not trace out a single demand or supply curve. Thus, if we performed a simple linear regression of price on quantity, the resulting regression line (shown as the bright red line in Figure A-1), would not correspond to either a supply or demand curve. 28 For a discussion, see P. Kennedy, A Guide to Econometrics, 6th Ed., (Malden, MA: Blackwell Publishing 2009), pp. 173-76. -19- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 If we graph the annual average natural gas wellhead prices and total natural gas withdrawals, we see a similar problem. This is shown in Figure A-2. Figure A-2: Average Annual Wellhead Natural Gas Prices and Gross Withdrawals (1990–2010) In Figure A-2, we have graphed the supply-demand combinations for the years 1990 through 2010. Although the trendline shown looks like an upward sloping supply curve, it is far more likely that it reflects changes in demand and supply curves over time, as in Figure A-1. Model Specification To address this simultaneity issue, we developed a 4-equation model reflecting the supply and demand for natural gas, as well as the supply and demand of coal for electric generating purposes, because natural gas is increasingly being substituted for coal to generate electricity.29 Thus, we write the general model structure as: 29 We assume the world crude oil price (measured as the published price of Brent crude) as independent of the U.S. natural gas market and the market for coal used to generate electricity. -20- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Pt G f1 (QtG , Pt O , Pt C ) QtG f 2 ( Pt G , Pt O , Pt C ) Pt C f3 (QtC , Pt G , Pt O ) (A-1) QtC f 4 ( Pt C , Pt O , Pt G ) where: Pt C = Average annual price of coal delivered to electric generating plants, year t. Pt G = Average annual wellhead price of natural gas, year t. Pt O = Average annual price of Brent crude, year t. QtG = Gross withdrawals of U.S. natural gas from all sources, year t. QtC = Receipts of coal at electric generating plants, year t. Thus, the model consists of demand and supply equations for natural gas and coal delivered to electric generating plants. The specification treats the wellhead price of natural gas, gross withdrawals of natural gas, the quantity of coal delivered to electric generators, and the price of coal delivered to electric generators as endogenous variables. The price of Brent crude is treated as exogenous. To address the data shown in Figure A-2, we evaluated a number of functional forms for the general model structure in (A-1), ultimately settling on the following specification, (t-1 subscripts indicate one-year lagged values). QtC 02 12 Pt C 22 Pt O1 32 Pt C1 43GDPt t2 (A-2) Pt C 03 13 Pt C1 23 Pt G 33 Pt O1 t3 (A-3) Pt G 04 14QtG 24 Pt O 34 Pt C1 44 D2002 54 D2005 t4 (A-4) QtG 05 15 Pt G 25 Pt G1 35 Pt C1 45 Pt O1 t5 (A-5) where: D2002 = Dummy variable for the year 2002. D2005 = Dummy variable for the year 2005. GDPt = Real U.S. gross domestic product, year t. tj = Random error term, equation j. The inclusion of lagged price variables reflects the fact that gas and coal production and consumption decisions are influenced by observed historic prices, as well as contemporaneous -21- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 prices. The inclusion of a “dummy” variable for the year 2002 reflects the economic downturn brought on by the events of September 11, 2001. The inclusion of the dummy variable for the year 2005 reflects the damages caused by Hurricanes Katrina and Rita to the natural gas drilling and gathering infrastructure off the U.S. Gulf coast. To estimate the impact of shale gas production on wellhead natural gas prices, we note that QtG QtShale QtNon shale (A-6) where QtShale and QtNon shale represent production shale gas and gas from all non-shale sources, respectively. Thus, the estimated wellhead price of natural gas without shale gas, Pt G, NS is Pt G, NS Pt G 14QtShale , (A-7) where the minus sign in equation (A-7) reflects the fact that total wellhead production is reduced by the quantity of shale gas produced. Data Sources All of the data used to estimate the model is publicly available and published by the U.S. EIA. The specific data sources are as follows: Pt C Electric Power Annual. Available at: http://www.eia.gov/electricity/data.cfm Pt G Available at: http://www.eia.gov/dnav/ng/ng_pri_sum_dcu_nus_m.htm Pt O Based on daily spot prices, as published by EIA from ThomsonReuters. QtG Natural Gas Annual. Available at: http://www.eia.gov/dnav/ng/ng_prod_sum_dcu_nus_a.htm Electric Power Annual QtC Analysis Results Equations (A-2) – (A-5) were modeled using the three-state least-squares (3SLS) method in STATA.30 The results are summarized in Table A-1. 30 See Kennedy (2009), pp. 180-81. -22- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Table A-1: Three-stage Least-squares Regression Summary31 No. of Obs Parameters RMSE "R‐sq" Chi‐ square Significance (A‐2) Qcoal 20 3 50662.89 0.7064 50.17 0.0000 (A‐3) Pcoal 20 4 0.6913167 0.9783 909.27 0.0000 (A‐4) Pgas 20 5 0.3771036 0.8931 167.70 0.0000 (A‐5) Qgas 20 5 0.3141652 0.9670 584.57 0.0000 Equation The estimated coefficients for all of the equations are shown in Table A-2. Table A-2: Three-Stage Least Squares Regression Results Equation Dep. Variable (A‐2) (A‐3) (A‐4) (A‐5) Coef. Std. Err. z P>|z| [95% Confidence Interval] Pcoal(t) Qcoal(t) ‐5.01E‐06 3.95E‐06 ‐1.27 0.205 ‐0.0000128 2.73E‐06 Pcoal(t‐1) 0.9101757 0.0439476 20.71 0 0.82404 0.9963114 Pgas(t) 0.6102044 0.1782589 3.42 0.001 0.2608233 0.9595854 Poil(t‐1) 0.0627027 0.0122623 5.11 0 0.0386691 0.0867364 Constant 2.756187 3.41147 0.81 0.419 ‐3.930172 9.442546 Pcoal(t) ‐2065.633 4251.958 ‐0.49 0.627 ‐10399.32 6268.052 Poil(t) 2004.676 1515.881 1.32 0.186 ‐966.3957 4975.748 Real_GDP 24.01612 16.67534 1.44 0.15 ‐8.666949 56.69919 Constant 625558.4 253000.7 2.47 0.013 129686.1 1121431 ‐0.4617642 0.0809128 ‐0.2108965 0.7724999 ‐1.175227 18.91655 0.0873941 0.0047686 0.0164349 0.3563397 0.3269946 2.210884 ‐5.28 16.97 ‐12.83 2.17 ‐3.59 8.56 0 0 0 0.03 0 0 ‐0.6330535 0.0715666 ‐0.2431083 0.0740869 ‐1.816125 14.5833 ‐0.290475 0.0902591 ‐0.1786848 1.470913 ‐0.5343296 23.2498 Pgas(t) ‐0.2000595 0.0905154 ‐2.21 0.027 ‐0.3774665 ‐0.0226526 Pgas(t‐1) ‐0.6352557 0.1392652 ‐4.56 0 ‐0.9082104 ‐0.362301 Pcoal(t‐1) ‐0.0856178 0.0413864 ‐2.07 0.039 ‐0.1667337 ‐0.004502 Poil(t‐1) 0.0523611 0.0178196 2.94 0.003 0.0174353 0.0872869 Year 0.2041171 0.0379631 5.38 0 0.1297108 0.2785235 Constant ‐380.1301 76.49281 ‐4.97 0 ‐530.0532 ‐230.2069 Qcoal(t) Pgas(t) Qgas(t) Poil(t) Pcoal(t‐1) Dummy_2005 Dummy_2002 Constant Qgas(t) 31 Note that the “R-sq” is not the same as the traditional “goodness-of-fit” measure in OLS regressions, because with 3SLS “R-sq” can be negative. For comparison purposes, Appendix 1 provides the results of OLS regressions of each of the four equations (A-2) – (A-5). -23- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 The regression results for the coal equations are consistent with economic theory. For example, equation (A-2) shows that the delivered price of coal to electric generators increases as the price of natural gas and oil increase, which would be expected for substitute fuels. Equation (A-3) shows that the quantity of coal deliveries increases as the price of oil increases, but decreases as the price of coal increases. It also shows that coal deliveries increase as economic growth, measured by real GDP, increases. Equation (A-4) shows that the wellhead price of gas tends to increase as the price of oil increases, as was generally the case until 2005. Equation (A-4) also includes the two dummy variables for the years 2002 and 2005, respectively. The coefficients for both dummy variables have the expected signs. Equation (A-4) also shows that the wellhead price of gas is strongly related to the previous year’s price of utility coal receipts. However, the coefficient is negative. That is, an increase in last year’s utility price of coal tends to decrease this year’s wellhead price of natural gas. We hypothesize that this initially counterintuitive result stems from the response by natural gas producers to anticipated increases in the demand for natural gas. In other words, producers respond to higher coal prices and, hence, expected increases in the demand for natural gas, by increasing production. These production increases more than offset the expected increase in demand. In fact, this phenomenon has clearly contributed to the reductions in natural gas prices generally; technological improvements in drilling technology have enabled rapidly increasing quantities of shale gas to be produced, more than compensating for the general increase in natural gas demand. Equation (A-5) yields the expected results for the coefficients on the prices of natural gas and crude oil. Thus, we expect increases in the price of crude oil to increase production of natural gas. Not only does this stem from increased demand for gas, but higher oil prices encourage additional exploration of development of domestic oil resources, and significant quantities of natural gas are produced from oil wells (so called “wet gas”). Equation (A-5) also shows the same counterintuitive result with respect to coal prices. Impacts of Shale Gas Production on Wellhead Prices Equation (A-7) and the coefficient 14 from equation (A-4) are used to estimate the impacts of shale gas production on wellhead natural gas prices. Thus, the impact of shale gas production in year t on the average wellhead natural gas price in year t is Pt G, NS Pt G (0.4617642) x QtShale , which implies that, for each Tcf of shale gas produced, the wellhead price decreases by just over $0.46 per Mcf. -24- The Economic Impacts of Shale Gas on Ohio Consumers January 2012 Limitations of the Model Specification As discussed in Section II of the report, econometric modeling has both advantages and disadvantages, with a specific disadvantage being the partial equilibrium framework that we have used. Such a framework necessarily trades off accuracy for greater simplicity. The model estimated presumes there is a linear relationship between shale gas production and wellhead prices. In fact, the relationship may be nonlinear, especially as shale gas production accounts for an increasing proportion of total U.S. natural gas production. In 2010, for example, shale gas accounted for almost 20% of total gross production in the U.S., and that percentage is expected to increase over time, barring environmental regulations that restrict or curtail shale gas production in the future. To the extent the relationship is nonlinear, the predicted impacts of shale gas production on wellhead natural gas prices may be overstated. Moreover, additional shale gas production, to the extent it reduces wellhead prices, may reduce production of conventional natural gas. The impact will be a function of the market price and the production cost of conventional gas: the higher the cost to produce conventional gas, the larger the likely reduction in conventional production, and the smaller will be the net price impact. Finally, higher natural gas prices could have other macroeconomic impacts that are not considered in the econometric model, such as reductions in economic growth that would reduce overall natural gas demand and temper such price increases. Thus, a recommended next step in researching the impacts of shale gas on wellhead prices is to use more complex econometric specifications that test for and account for these potential nonlinearities. -25-
© Copyright 2024