Redrawing the energy-climate map - World Energy Outlook Special ...

REDRAWINGTHEENERGY-CLIMATEMAPWorld Energy Outlook Special Report

REDRAWINGTHEENERGY-CLIMATEMAPWorld Energy Outlook Special Report10 June 2013

© OECD/IEA, 2013lle Brornelis Bloason E BordoEric BorremansDavid Brayshawnna BroadhurstCarey BylinBen CaldecoPatri CarlnRichenda Connellvo de Boeros DelbeePhilippe DesfossesBo Dicfalusyoic Douilletmar EdenhoferDolf Gielenudi Greenwaldeigh aceDavid awinsDavid oneEsa yvrineniay yereun iangaroon hesghians rgen ochihiro uroiSarah adislawames eatonoesung eeMar ewisSurabi Menonlan MillerStephanie Millerincent Minierinistry of inance, SwedenEcofysColumbia University, United StatesBP Paribas nvestment PartnersUniversity of Reading, United ingdoministry of oreign airs and Trade, ew ealandUS Environmental Protecon gencySmith School of Enterprise and the Environment, University offordaenfallcclimasePGDirectorate General for Climate con, European CommissionRetraite ddionnelle de la oncon Publiue, ranceinistry of Enterprise, Energy and Communicaons, SwedenlstomPotsdam nstute for Climate mpact Researchnternaonal Renewable Energy gencyCentre for Climate and Energy Soluons, United Stateslstomatural Resources Defense Council, United StatesShellortumorld BanEnergy Research nstute, ChinaExxonMobilEnergy gency, Ministry of Climate, Energy and Building,DenmarThe nstute of Energy Economics, apanCenter for Strategic and nternaonal Studies, United StatesCarbon Tracerorea University Graduate School of Energy and EnvironmentDeutsche BanClimateors oundaonnternaonal inance Corporaonnternaonal inance CorporaonSchneider Electric4 World Energy Outlook | Special Report

Giuseppe MontesanoClaude ahonaushige obutaniPatric livaSimonEri llusenri Paillereindene E PaonStephanie Pfeiferolmar Pugntonio PgerMar RadaTeresa Riberaic Robinsrgen RosenowStéphanie Saunierrisn SeybothStephan SingerPremysaw Sobasieena Srivastavandrew Steericholas Sternaren Sundosué Tanaao Tyndallras TynynenStefaan ergotePaul atinsonMelissa eitMechthild rsdrferi ouEEEDMinistry of Economy, Trade and ndustry, apanMichelinortumuclear Energy gencyurich nsurance Groupnstuonal nvestors Group on Climate ChangeSiemensederal Ministry of Economics and Technology, GermanyUnited aons Environment ProgrammeSTSBCECarbon imitsIntergovernmental Panel on Climate Change InternaonalMinistry of the Environment, PolandThe Energy and Resources Instute, Indiaorld Resources Instuteondon School of EconomicsSUD EnergyEuropean Ban for Reconstrucon and DevelopmentMinistry of oreign airs and Trade, ew ealandParliament of inlandDirectorate General for Climate con, European CommissionMinistry of Ecology and Sustainable Development, ranceEnvironmental Protecon gency, United StatesDirectorate General for Energy, European Commissionaonal Center for Climate Change Strategy and InternaonalCooperaon, China1234The individuals and organisaons that contributed to this study are not responsible for anyopinions or udgements it contains ll errors and omissions are solely the responsibilityof the IE© OECD/IEA, 2013Acknowledgements 5

Dr ah BirolChief EconomistDirector, Directorate of Global Energy EconomicsInternaonal Energy gency, rue de la édéraon3 Paris Cedex 1ranceTelephoneEmail(331) 4 weoieaorgMore informaon about the World Energy Outlook is available atwww.worldenergyoutlook.org© OECD/IEA, 20136 World Energy Outlook | Special Report

Table of ContentsAcknowledgements 3Executive Summary 91Climate and energy trends 13Introduction 14The energy sector and climate change 15Recent developments 18International climate negotiations 19National actions and policies with climate benefits 20Global status of energy-related CO 2 emissions 26Trends in energy demand and emissions in 2012 26Historical emissions trends and indicators 29Outlook for energy-related emissions and the 450 Scenario 33Sectoral trends 36Investment 38Broader benefits of the 450 Scenario 392Energy policies to keep the 2 °C target alive 43Introduction 44GDP-neutral emissions abatement to 2020 45Methodology and key assumptions 45Emissions abatement to 2020 50Abatement to 2020 by policy measure 53Implications for the global economy 70Building blocks for steeper abatement post-2020 75The long-term implications of the 4-for-2 °C Scenario 75Technology options for ambitious abatement post-2020 76Policy frameworks post-2020 80© OECD/IEA, 20133Managing climate risks to the energy sector 83Introduction 84Impacts of climate change on the energy sector 84Energy demand impacts 87Energy supply impacts 91Climate resilience 96Table of Contents 7

Economic impact of climate policies on the energy sector 97Existing carbon reserves and energy infrastructure lock-in 98Power generation 101Upstream oil and natural gas 106Coal supply 110Implications of delayed action 113Annexes 117Annex A. Units and conversion factors 117Annex B. References 119© OECD/IEA, 20138 World Energy Outlook | Special Report

The world is not on track to meet the target agreed by governments to limit the longtermrise in the average global temperature to 2 degrees Celsius (°C). Global greenhousegasemissions are increasing rapidly and, in May 2013, carbon-dioxide (CO 2 ) levels in theatmosphere exceeded 400 parts per million for the rst me in several hundred millennia.The weight of scienc analysis tells us that our climate is already changing and that weshould expect extreme weather events (such as storms, oods and heat waves) to becomemore freuent and intense, as well as increasing global temperatures and rising sea levels.Policies that have been implemented, or are now being pursued, suggest that the long-termaverage temperature increase is more likely to be between 3.6 °C and 5.3 °C (comparedwith pre-industrial levels), with most of the increase occurring this century. hile globalacon is not yet sucient to limit the global temperature rise to 2 °C, this target sllremains technically feasible, though extremely challenging. To keep open a realisc chanceof meeng the 2 °C target, intensive acon is reuired before 2020, the date by which anew internaonal climate agreement is due to come into force. Energy is at the heart of thischallenge: the energy sector accounts for around two-thirds of greenhouse-gas emissions,as more than 80 of global energy consumpon is based on fossil fuels.espite posive developments in some countries global energy-related C 2 emissionsincreased by 1.4% to reach 31.6 gigatonnes (Gt) in 212 a historic high. Non-OECDcountries now account for 60 of global emissions, up from 45 in 2000. In 2012, Chinamade the largest contribuon to the increase in global CO 2 emissions, but its growth wasone of the lowest it has seen in a decade, driven largely by the deployment of renewablesand a signicant improvement in the energy intensity of its economy. In the United States,a switch from coal to gas in power generation helped reduce emissions by 200 milliontonnes (Mt), bringing them back to the level of the mid-1990s. However, the encouragingtrends in China and the United States could well both be reversed. Despite an increase incoal use, emissions in Europe declined by 50 Mt as a result of economic contracon, growthin renewables and a cap on emissions from the industry and power sectors. Emissions inJapan increased by 70 Mt, as eorts to improve energy eciency did not fully oset theuse of fossil fuels to compensate for a reducon in nuclear power. Even aer allowing forpolicies now being pursued, global energy-related greenhouse-gas emissions in 2020 areprojected to be nearly 4 Gt CO 2 -equivalent (CO 2 -eq) higher than a level consistent withaaining the 2 °C target, highlighng the scale of the challenge sll to be tackled just inthis decade.© OECD/IEA, 2013We present our 4-for-2 °C cenario in which we propose the implementaon of fourpolicy measures that can help keep the door open to the 2 °C target through to 2020 at nonet economic cost. Relave to the level otherwise expected, these policies would reducegreenhouse-gas emissions by 3.1 Gt CO 2 -eq in 2020 80 of the emissions reduconExecutive Summary 9

equired under a 2 °C trajectory. This would buy precious me while internaonal climatenegoaons connue towards the important Conference of the Pares meeng in Parisin 2015 and the naonal policies necessary to implement an expected internaonalagreement are put in place. The policies in the 4-for-2 °C Scenario have been selectedbecause they meet key criteria: they can deliver signicant reducons in energy-sectoremissions by 2020 (as a bridge to further acon) they rely only on exisng technologiesthey have already been adopted and proven in several countries and, taken together, theirwidespread adopon would not harm economic growth in any country or region. The fourpolicies are:• Adopng specic energy eciency measures (49 of the emissions savings).• iming the construcon and use of the least-ecient coal-red power plants (21).• Minimising methane (CH 4 ) emissions from upstream oil and gas producon (18).• Accelerang the (paral) phase-out of subsidies to fossil-fuel consumpon (12).Targeted energy eciency measures would reduce global energy-related emissionsby 1. Gt in 2020 a level close to that of ussia today. These policies include: energyperformance standards in buildings for lighng, new appliances, and for new heang andcooling equipment in industry for motor systems and, in transport for road vehicles.Around 60 of the global savings in emissions are from the buildings sector. In countrieswhere these eciency policies already exist, such as the European Union, Japan, theUnited States and China, they need to be strengthened or extended. Other countries needto introduce such policies. All countries will need to take supporng acons to overcomethe barriers to eecve implementaon. The addional global investment required wouldreach $200 billion in 2020, but would be more than oset by reduced spending on fuel bills.nsuring that new subcrical coal-red plants are no longer built and liming the use ofthe least ecient eisng ones would reduce emissions by 640 t in 2020 and also helpeorts to curb local air polluon. Globally, the use of such plants would be one-quarterlower than would otherwise be expected in 2020. The share of power generaon fromrenewables increases (from around 20% today to 27% in 2020), as does that from naturalgas. Policies to reduce the role of inecient coal power plants, such as emissions and airpolluon standards and carbon prices, already exist in many countries. In our 4-for-2 °CScenario, the largest emissions savings occur in China, the United States and India, all ofwhich have a large coal-powered eet.© OECD/IEA, 2013ethane releases into the atmosphere from the upstream oil and gas industry would bealmost halved in 2020 compared with levels otherwise epected. Around 1.1 Gt CO 2 -eqof methane, a potent greenhouse-gas, was released in 2010 by the upstream oil and gasindustry. These releases, through venng and aring, are equivalent to twice the total naturalgas producon of Nigeria. Reducing such releases into the atmosphere represents an eecvecomplementary strategy to the reducon of CO 2 emissions. The necessary technologies arereadily available, at relavely low cost, and policies are being adopted in some countries,such as the performance standards in the United States. The largest reducons achieved inthe 4-for-2 °C Scenario are in Russia, the Middle East, the United States and Africa.10 World Energy Outlook | Special Report

ccelerated acon towards a paral phase-out of fossil-fuel subsidies would reduce C 2emissions by 360 t in 2020 and enable energy eciency policies. Fossil-fuel subsidiesamounted to $523 billion in 2011, around six mes the level of support to renewableenergy. Currently, 15% of global CO 2 emissions receive an incenve of $110 per tonnein the form of fossil-fuel subsidies while only 8% are subject to a carbon price. Growingbudget pressures strengthen the case for fossil-fuel subsidy reform in many imporngand exporng countries and polical support has been building in recent years. G20 andAsia-Pacic Economic Cooperaon (APEC) member countries have commied to phase outinecient fossil-fuel subsidies and many are moving ahead with implementaon.1234The energy sector is not immune from the physical impacts of climate change and mustadapt. In mapping energy system vulnerabilies, we idenfy sudden and destrucveimpacts (caused by extreme weather events) that pose risks to power plants and grids, oiland gas installaons, wind farms and other infrastructure. Other impacts are more gradual,such as changes to heang and cooling demand, sea level rise on coastal infrastructure,shiing weather paerns on hydropower and water scarcity on power plants. Disruponsto the energy system can also have signicant knock-on eects on other crical services.To improve the climate resilience of the energy system, governments need to design andimplement frameworks that encourage prudent adaptaon, while the private sector shouldassess the risks and impacts as part of its investment decisions.The nancial implicaons of stronger climate policies are not uniform across the energyindustry and corporate strategy will need to adjust accordingly. Under a 2 °C trajectory,net revenues for exisng nuclear and renewables-based power plants would be boosted by$1.8 trillion (in year-2011 dollars) through to 2035, while the revenues from exisng coal-red plants would decline by a similar level. Of new fossil-fuelled plants, 8% are reredbefore their investment is fully recovered. Almost 30% of new fossil-fuelled plants are ed(or retro-ed) with CCS, which acts as an asset protecon strategy and enables more fossilfuel to be commercialised. A delay in CCS deployment would increase the cost of powersector decarbonisaon by $1 trillion and result in lost revenues for fossil fuel producers,parcularly coal operators. Even under a 2 °C trajectory, no oil or gas eld currently inproducon would need to shut down prematurely. Some elds yet to start producon arenot developed before 2035, meaning that around 5% to 6% of proven oil and gas reservesdo not start to recover their exploraon costs in this meframe.© OECD/IEA, 2013elaying stronger climate acon to 2020 would come at a cost 1. trillion in lowcarboninvestments are avoided before 2020 but trillion in addional investmentswould be reuired thereaer to get back on track. Delaying further acon, even to theend of the current decade, would therefore result in substanal addional costs in theExecutive Summary 11

energy sector and increase the risk that the use of energy assets is halted before theend of their economic life. The strong growth in energy demand expected in developingcountries means that they stand to gain the most from invesng early in low-carbon andmore ecient infrastructure, as it reduces the risk of premature rerements or retrots ofcarbon-intensive assets later on.© OECD/IEA, 201312 World Energy Outlook | Special Report

Chapter 1Climate and energy trendsMeasuring the challengeHighlights There is a growing disconnect between the trajectory that the world is on and onethat is consistent with a 2 °C climate goal the objecve that governments haveadopted. Average global temperatures have already increased by 0.8 °C comparedwith pre-industrial levels and, without further climate action, our projecons arecompable with an addional increase in long-term temperature of 2.8 °C to 4.5 °C,with most of the increase occurring this century. Energy-related CO 2 emissions reached 31.6 Gt in 2012, an increase of 0.4 Gt (or 1.4%)over their 2011 level, conrming rising trends. The global increase masks diverseregional trends, with posive developments in the two-largest emiers, China andthe United States. US emissions declined by 200 Mt, mostly due to low gas pricesbrought about by shale gas development that triggered a switch from coal to gas inthe power sector. Chinas emissions in 2012 grew by one of the smallest amountsin a decade (300 Mt), as almost all of the 5.2% growth in electricity was generatedusing low-carbon technologies mostly hydro and declining energy intensitymoderated growth in energy demand. Despite an increase in coal use, emissionsin Europe declined (-50 Mt) due to economic contracon, growth in renewablesand a cap on emissions from the industry and power sectors. OECD countries nowaccount for around 40% of global emissions, down from 55% in 2000. Internaonal climate negoaons have resulted in a commitment to reach a newglobal agreement by 2015, to come into force by 2020. But the economic crisishas had a negave impact on the pace of clean energy deployment and on carbonmarkets. Currently, 8% of global CO 2 emissions are subject to a carbon price, while15% receive an incenve of $110 per tonne in the form of fossil-fuel subsidies.Price dynamics between gas and coal are supporng emissions reducons in someregions, but are slowing them in others, while nuclear is facing dicules and largescalecarbon capture and storage remains distant. Despite growing momentumto improve energy eciency, there remains vast potenal that could be tappedeconomically. Non-hydro renewables, supported by targeted government policies,are enjoying double-digit growth.© OECD/IEA, 2013 Despite the insuciency of global acon to date, liming the global temperaturerise to 2 °C remains sll technically feasible, though it is extremely challenging. Toachieve our 450 Scenario, which is consistent with a 50% chance of keeping to 2 °C,the growth in global energy-related CO 2 emissions needs to halt and start to reversewithin the current decade. Clear polical resoluon, backed by suitable policies andnancial frameworks, is needed to facilitate the necessary investment in low-carbonenergy supply and in energy eciency.Chapter 1 | Climate and energy trends 13

ntroduconClimate change is a dening challenge of our me. The scienc evidence of its occurrence,its derivaon from human acvies and its potenally devastang eects accumulate. Sealevels have risen by 15-20 cenmetres, on average, over the last century and this increasehas accelerated over the last decade (Meyssignac and Caenave, 2012). Oceans arewarming and becoming more acidic, and the rate of ice-sheet loss is increasing. The Arccprovides a parcularly clear illustraon, with the area of ice covering the Arcc Ocean inthe summer diminishing by half over the last 30 years to a record low level in 2012. Therehas also been an increase in the frequency and intensity of heat waves, resulng in moreof the world being aected by droughts, harming agricultural producon (Hansen, Sato andRuedy, 2012).Global awareness of the phenomenon of climate change is increasing and polical aconis underway to try and tackle the underlying causes, both at naonal and internaonallevels. Governments agreed at the United Naons Framework Convenon on ClimateChange (UNFCCC) Conference of the Pares in Cancun, Mexico in 2010 (COP-16) that theaverage global temperature increase, compared with pre-industrial levels, must be heldbelow 2 degrees Celsius (°C), and that this means greenhouse-gas emissions must bereduced. A deadline was set at COP-18 in Doha, atar in 2012 for agreeing and enacng anew global climate agreement to come into eect in 2020. But although overcoming thechallenge of climate change will be a long-term endeavour, urgent acon is also required,well before 2020, in order to keep open a realisc opportunity for an ecient and eecveinternaonal agreement from that date.There is broad internaonal acceptance that stabilising the atmospheric concentraonof greenhouse gases at below 450 parts per million (ppm) of carbon-dioxide equivalent(CO 2 -eq) is consistent with a near 50% chance of achieving the 2 °C target, and that thiswould help avoid the worst impacts of climate change. Some analysis nds, however, thatthe risks previously believed to be associated with an increase of around 4 °C in globaltemperatures are now associated with a rise of a little over 2 °C, while the risks previouslyassociated with 2 °C are now thought to occur with only a 1 °C rise (Smith, et al., 2009). Otheranalysis nds that 2 °C warming represents a threshold for some climate feedbacks thatcould signicantly add to global warming (enton, et al., 2008). The UNFCCC negoaonstook these scienc developments into account in the Cancun decisions, which include anagreement to review whether the maximum acceptable temperature increase needs tobe further reduced, including consideraon of a global average temperature rise of 1.5 °C.Global greenhouse-gas emissions connue to increase at a rapid pace. The 450 ppmthreshold is drawing ever closer (Figure 1.1). Carbon-dioxide (CO 2 ) levels in the atmospherereached 400 ppm in May 2013, having jumped by 2.7 ppm in 2012 the second-highest© OECD/IEA, 201314 World Energy Outlook | Special Report

ise since record keeping began (Tans and eeling, 2013). 1 Average global temperatureshave already increased by around 0.8 °C, compared with pre-industrial levels, and, withoutaddional acon, a further increase in long-term temperature of 2.8 °C to 4.5 °C appearsto be in prospect, with most of the increase occurring this century. 2Figure 1.1 World atmospheric concentration of CO 2 and average globaltemperature changeppm4504003500.750.500.25°CAtmosphericconcentraonof CO 2Temperaturechange (right axis)12343000250-0.25200-0.501900 1920 1940 1960 1980 2000 2010Note: The temperature refers to the NASA Global and-Ocean Temperature Index in degrees Celsius, baseperiod: 1951-1980. The resulng temperature change is lower than the one compared with pre-industriallevels.Sources: Temperature data are from NASA (2013) CO 2 concentraon data from NOAA Earth System Researchaboratory.The energy sector and climate changeThe energy sector is by far the largest source of greenhouse-gas emissions, accounng formore than two-thirds of the total in 2010 (around 90% of energy-related greenhouse-gasemissions are CO 2 and around 9% are methane CH 4 , which is generally treated, in thisanalysis, in terms of its CO 2 equivalent eect). The energy sector is the second-largestsource of CH 4 emissions aer agriculture and we have esmated total energy-related CH 4emissions to be 3.1 gigatonnes (Gt) of carbon-dioxide equivalent (CO 2 -eq) in 2010 (around40% of total CH 4 emissions). Accordingly, energy has a crucial role to play in tacklingclimate change. et global energy consumpon connues to increase, led by fossil fuels,which account for over 80% of global energy consumed, a share that has been increasinggradually since the mid-1990s.© OECD/IEA, 20131. The concentration of greenhouse gases measured under the yoto Protocol was 444 ppm CO 2 -eq in 2010 andthe concentration of all greenhouse gases, including cooling aerosols, was 403 ppm CO 2 -eq (EEA, 2013).2. The higher increase in temperature is consistent with a scenario with no further climate action and thelower with a scenario that takes cautious implementation of current climate pledges and energy policies underdiscussion, the New Policies Scenario. Greenhouse-gas concentration is calculated using MAGICC version 5.3v2(UCAR, 2008) and temperature increase is derived from Rogelj, Meinhausen and nutti (2012).Chapter 1 | Climate and energy trends 15

Carbon pricing is gradually becoming established, and yet the worlds largest carbonmarket, the EU Emissions Trading System (ETS), has seen prices remain at very low levels,and consumpon of coal, the most carbon-intensive fossil fuel, connues to increaseglobally. Some countries are reducing the role of nuclear in their energy mix and developingstrategies to compensate for it, including with increased energy eciency. Renewableshave experienced strong growth and have established themselves as a vital part of theglobal energy mix but, in many cases, they sll require economic incenves and appropriatelong-term regulatory support to compete eecvely with fossil fuels. Acon to improveenergy eciency is increasing, but two-thirds of the potenal remains untapped (IEA,2012a). It is, accordingly, evident that if the energy sector is to play an important part inaaining the internaonally adopted target to limit average global temperature increase,a transformaon will be required in the relaonship between economic development,energy consumpon and greenhouse-gas emissions. Is such a transion feasible Analysesconclude that, though extremely challenging, it is feasible (IEA 2012a OECD 2012).Our 450 Scenario, which shows what is needed to set the global energy sector on a coursecompable with a near 50% chance of liming the long-term increase in average globaltemperature to 2 °C, suggests one pathway. Achieving the target will require determinedpolical commitment to fundamental change in our approach to producing and consumingenergy. All facets of the energy sector, parcularly power generaon, will need to transformtheir carbon performance. Moreover, energy demand must be moderated throughimproved energy eciency in vehicles, appliances, homes and industry. Deployment ofnew technologies, such as carbon capture and storage, will be essenal. It shows that,to stay on an economically feasible pathway, the rise in emissions from the energysector needs to halted and reversed by 2020. Acon at naonal level needs to ancipateimplementaon of a new internaonal agreement from 2020. Achieving a 2 °C target in theabsence of such acon, while technically feasible, would entail the widespread adopon ofexpensive negave emissions technologies (Box 1.1), which extract more CO 2 from theatmosphere than they add to it. By the end of the century, energy-related CO 2 emissionsin the 450 Scenario need to decrease to around 5 Gt CO 2 per year, i.e. less than one-sixthtodays levels. 3It is the cumulave build-up of greenhouse gases, including CO 2 , in the atmosphere thatcounts, because of the long lifeme of some of those gases in the atmosphere. Analysis hasshown that, in order to have a 50% probability of keeping global warming to no more than2 °C, total emissions from fossil fuels and land-use change in the rst half of the century needto be kept below 1 440 Gt (Meinshausen, et al., 2009). Of this carbon budget, 420 Gt CO 2has already been emied between 2000 and 2011 (Oliver, Janssens-Maenhout and Peters,2012) and the World Energy Outlook 2012 (WEO-2012) esmated that another 136 Gt CO 2© OECD/IEA, 20133. The RCP2.6 Scenario in the Fifth Assessment Report of the Intergovernmental Panel on Climate Changeis based on negative emissions from 2070 (IPCC, 2013). The RCP2.6 Scenario is more ambitious than the450 Scenario in that it sets out to achieve an 80% probability of limiting the long-term (using 2200 as thereference year) global temperature increase to 2 °C, while the probability is around 50% in the 450 Scenario.16 World Energy Outlook | Special Report

will be emied from non-energy related sources in the period up to 2050. This means thatthe energy sector can emit a maximum of 884 Gt CO 2 by 2050 without exceeding its residualbudget. hen mapping potenal emissions trajectories against such a carbon budget, itbecomes clear that the longer acon to reduce global emissions is delayed, the more rapidreducons will need to be in the future to compensate (Figure 1.2). Some models esmatethat the maximum feasible rate of such emissions reducon is around 5% per year (Elen,Meinshausen and uuren, 2007) Chapter 3 explores further the implicaons of delayedacon in the energy sector.Box 1.1 What are negative emissions?1234Carbon capture and storage (CCS) technology could be used to capture emissionsfrom biomass processing or combuson processes and store them in deep geologicalformaons. The process has the potenal to achieve a net removal of CO 2 from theatmosphere (as opposed to merely avoiding CO 2 emissions to the atmosphere as is thecase for convenonal, fossil fuel-based CCS). Such negave emissions result whenthe amount of CO 2 sequestered from the atmosphere during the growth of biomass(and subsequently stored underground) is larger than the CO 2 emissions associatedwith the producon of biomass, including those resulng from land-use change andthe emissions released during the transformaon of biomass to the nal product (IEA,2011a). So-called bio-energy with carbon capture and storage (BECCS) could be usedin a wide range of applicaons, including biomass power plants, combined heat andpower plants, ue gas streams from the pulp and paper industry, fermentaon inethanol producon and biogas rening processes.From a climate change perspecve, BECCS is aracve for two reasons. First, netgains from BECCS can oset emissions from a variety of sources and sectors that aretechnically dicult and expensive to abate, such as emissions from air transportaon.Second, BECCS can migate emissions that have occurred in the past. For a given CO 2stabilisaon target, this allows some exibility in the ming of emissions higheremissions in the short term can, within limits, be compensated for by negaveemissions in the longer term. Of course, the projects have to be economically viable.© OECD/IEA, 2013To achieve net negave emissions, it is essenal that only biomass that is sustainablyproduced and harvested is used in a BECCS plant. Assuring the sustainability ofthe biomass process will require dedicated monitoring and reporng. As this willencompass acvies that are similar to those required to monitor and verify emissionsreducon from deforestaon and forest degradaon (REDD), the development ofnaonal and internaonal REDD strategies will contribute to the deployment of BECCSand vice versa. Increasing the share of sustainably managed biomass in a countrysenergy mix, in addion to decreasing CO 2 emissions, has a number of benets in termsof economic development and employment.Chapter 1 | Climate and energy trends 17

Figure 1.2 Mapping feasible world CO 2 emissions trajectories within acarbon budget constraintAnnual CO2 emissionsCumulave CO2 emissionsCO 2 budget with a 50% chance of 2 °CTimeTimeTaking as its starng point the proporonate contribuon of the energy system to theglobal greenhouse-gas emissions, this chapter focuses on the disconnect between the levelof acon that science tells us is required to meet a 2 °C climate goal and the trajectory theworld is currently on. It looks at recent developments in climate policy, both at global andnaonal levels, and at those elements of energy policy that could have a signicant posiveimpact on the migaon of climate change. It maps the current status of global greenhousegasemissions, illustrang the dominant role of energy-related CO 2 emissions in this pictureand the important underlying trends, drawing on the latest emissions esmates for 2012.It then looks at the prospecve future contribuon of the energy sector to the totalemissions up to 2035, comparing the outcome if the world pursues its present course (ourNew Policies Scenario) with a trajectory compable with liming the long-term increasein average global temperature to 2 °C (our 450 Scenario) (IEA, 2012a). This enables us tohighlight the addional eorts that would be necessary to achieve the 2 °C climate goaland to point to the short-term acons (see Chapter 2) which could contribute signicantlyto make its realisaon possible.Recent developmentsGovernment policies are crical to tackling climate change: what has been happeningAnswering this queson requires an examinaon both of policies that are directed mainlyat climate change and of policies with other primary objecves, such as energy securityand local polluon, which also have consequences for global emissions. hile key climatecommitments may be internaonal, implemenng acons will be taken primarily atnaonal and regional levels. So far, to take one indicave example, carbon pricing appliesonly to 8% of energy-related CO 2 emissions, while fossil-fuel subsidies, acng as a carbonincenve, aect almost twice that level of emissions.© OECD/IEA, 201318 World Energy Outlook | Special Report

centralised set of commitments that characterise the yoto Protocol, in order to allowfor exibility to take account of naonal circumstances. It is expected to bring togetherexisng pledges into a co-ordinated framework that builds mutual trust and condencein the total emissions abatement that they represent. It will also need to create a processthat provides for the ambion of these pledges to be adequately developed to match theevolving requirements of meeng the 2 °C goal.As discussed above, policies adopted at the naonal level which deliver emissionsreducons are central to tackling climate change whether that is their primary movaonor not. The global economic crisis has constrained policy makers scope for acon in recentyears, but there have been some encouraging developments. In parcular, many developingcountries that made voluntary emissions reducon pledges under the Cancun Agreementshave announced new strategies and policies, in many cases involving measures in theenergy sector (Figure 1.4).Figure 1.4 Linkages between climate change and other major policiesCarbon pricing is one of the most direct ways of tackling emissions. Currently some 8% ofglobal energy-related CO 2 emissions are subject to carbon pricing. This share is expectedto increase, as more countries and regions adopt this pracce (Spotlight). However, theroll-out is by no means free of concerns, notably on compeveness and carbon leakage.© OECD/IEA, 2013Power plant emissions are being regulated in a number of countries. Regulaons limingemissions from new power plants, which would have an impact parcularly on investmentin new convenonal coal generaon, have been proposed by the US EnvironmentalProtecon Agency (EPA). New standards are also expected to be promulgated for exisngplants. US EPA regulaons targeng convenonal pollutants are also expected to promotemodernisaon of the power generaon eet (though they may face legal challenge).Canada has introduced regulaons for new power plants that rule out new convenonalcoal investment. High levels of local polluon connue to be a signicant issue for some20 World Energy Outlook | Special Report

of Chinas largest cies and the government has spulated mandatory reducons insulphur dioxide (SO 2 ) and nitrogen oxide (NO x ) emissions per kilowa-hour (kh) of powergenerated by coal-red power plants and a target to cut by at least 30% the emissionsintensity of parculate maer (PM 2.5 ) coming from energy producon and use. Thesenaonal measures will all have associated climate change benets.Although WEO-2012 demonstrated that only a fracon of the available energy eciencybenets are currently being realised, fortunately, many countries are taking new stepsto tap this potenal. In early 2013, the US government announced a goal to doubleenergy producvity by 2030. WEO-2012 had already highlighted the contribuon newfuel-economy standards could make in moving the United States towards lower importneeds and the queson now is whether similar eects can be achieved in other sectorsof the economy. The US Department of Energy has put in place in recent years energyeciency standards for a wide range of products, including air condioners, refrigeratorsand washing machines. More standards are expected to come into force for eciency inbuildings and appliances. The European Union has adopted an Energy Eciency Direcve,to support its target of improving energy eciency by 20% by 2020 and pave the way forfurther improvements beyond this. In China, the 12 th Five-ear Plan (2011-2015) includesindicave caps on total energy consumpon and on power consumpon for 2015. Thereare also mandatory targets to reduce the energy intensity of the economy by 16% and toreduce CO 2 emissions per unit GDP by 17% the rst me a CO 2 target has been set. Chinahas published energy eciency plans consistent with the 12 th Five-ear Plan, includingthe Top 10 000 programme that sets energy savings targets by 2015 for the largestindustrial consumers. In India, a Naonal Mission on Enhanced Energy Eciency has beenlaunched, aimed at restraining growth in energy demand. Indias Perform Achieve andTrade mandatory trading system for energy eciency obligaons in some industries waslaunched in 2011, and is a key element in plans to deliver its pledge to reduce carbonintensity by 20-25% by 2020 (from 2005 levels).1234© OECD/IEA, 2013Many forms of intervenon to support renewable energy sources have contributed to thestrong growth of the sector in recent years. Installed wind power capacity increased by19% in 2012, to reach 282 gigawas (G), with China, the United States, Germany, Spainand India having the largest capacity (GEC, 2013). Sub-Saharan Africas rst commercialwind farm also came online, in Ethiopia. US solar installaons increased by 76%, to 3.3 Gin 2012 (Solar Energy Industries Associaon, 2013) and, while a federal target is not inplace, most US states have renewable energy porolio standards designed to increasethe share of electricity generated from renewable sources. The European Union has inplace a target contribuon from renewable energy to primary demand of 20% by 2020.Japan has also expressed strong expectaons for renewables, mainly solar photovoltaics(P), in its new energy strategy following the Fukushima Daiichi nuclear accident. Chinahas an extensive range of targets for all renewables, which are regularly upgraded. Oneexample is the recent strengthening of the target for P installaons to 10 G per year,promising to make China the world leader for P installaon from 2013 onwards. IndiaChapter 1 | Climate and energy trends 21

Nuclear policies vary by country. In 2012, Japan announced new energy eciency andrenewable energy targets, supported by feed-in taris, in light of the 2011 accident atthe Fukushima Daiichi nuclear power staon. However, plans in the major nuclear growthmarkets, such as China, India and orea, are largely being maintained. Condence in theavailability of low-carbon alternaves needs to be high in countries contemplang movingaway from nuclear power.In transport, policies to increase eciency and support new technologies go hand-in-hand.Most major markets have fuel-economy standards for cars and have scope to introducesimilar standards for freight. Sales of plug-in hybrids and electric vehicles more thandoubled, to exceed 100 000, in 2012. Nonetheless, these sales are sll well below the levelrequired to achieve the targets set by many governments. Collecvely these amount toaround 7-9 million vehicles by 2020 (IEA, 2013a).1234Figure 1.7 CCS capacity by region and project status, 2012Mt CO21008060Rest of worldOECD Asia OceaniaEuropeNorth America4020Operang2012UnderconstruconAdvancedplanningEarly planningNotes: Relates to large-scale integrated projects and, where a range is given for CO 2 capture capacity, themiddle of the range has been taken. Exisng EOR projects are not included where they are not authorisedand operated for the purpose of CCS.Sources: Global CCS Instute (2013) and IEA analysis.© OECD/IEA, 2013Carbon capture and storage (CCS) technology can, in principle, reduce full life-cycle CO 2emissions from fossil-fuel combuson at staonary sources, such as power staonsand industrial sites, by 65-85% (GEA, 2012). However, the operaonal capacity of largescaleintegrated CCS projects, excluding enhanced oil recovery (EOR), so far provides forthe capture of only 6 million tonnes (Mt) of CO 2 per year, with provisions for a further13 Mt CO 2 under construcon as of early-2013 (Figure 1.7). If all planned capacity wereto be constructed, this would take the total to around 90 Mt CO 2 , sll equivalent to lessthan 1% of power sector CO 2 emissions in 2012. hile the technology is available today,projects need to be scaled-up signicantly from exisng levels in order to demonstratecarbon capture and storage from a typical coal-red power plant. Experience gained fromChapter 1 | Climate and energy trends 25

large demonstraon projects will be essenal, both to perfecng technical soluons anddriving down costs. Ulmately, a huge scale-up in CCS capacity is required if it is to make ameaningful impact on global emissions (see Chapter 2).Global status of energy-related CO 2 emissionsTrends in energy demand and emissions in 2012Global CO 2 emissions from fossil-fuel combuson increased again in 2012, reaching a recordhigh of 31.6 Gt, according to our preliminary esmates. 5 This represents an increase of0.4 Gt on 2011, or 1.4%, a level that, if connued, would suggest a long-term temperatureincrease of 3.6 °C or more. The growth in emissions results from an increase in global fossilfuelconsumpon: 2.7% for natural gas, 1.1% for oil and 0.6% for coal. Taking into accountemissions factors that are specic to fuel, sector and region, natural gas and coal eachaccounted for 44% of the total energy-related CO 2 emissions increase in 2012, followed byoil (12%). The global trend masks important regional dierences: in 2012, a 3.1% increasein CO 2 emissions in non-OECD countries was oset, but only partly, by a 1.2% reducon inemissions in OECD countries (Figure 1.8).Figure 1.8 CO 2 emissions trends in 2012Mt4003002001000-100-20032%24%16%8%0%Change from2011Share of worldemissions(right axis)-300UnitedStatesEuropeanUnionOtherOECDIndia Middle Japan Other ChinaEast non-OECD© OECD/IEA, 2013hile China made the largest contribuon to the global increase, with its emissions risingby 300 Mt, or 3.8%, this level of growth is one of the smallest in the past decade andless than half of the emissions increase in 2011, reecng Chinas eorts in installing lowcarbongenerang capacity and achieving improvements in energy intensity. Coal demandgrew by 2.4%, most of it to supply industrial demand. hile electricity generaon in Chinaincreased 5.2%, coal input to power generaon grew by only 1.2%. Most of the addionaldemand was met by hydro, with 18 G of capacity addions coming online in 2012,complemented by a wet year in 2012. Increased wind and solar also played a role. Hydrocapacity at the end of 2012 was 249 G, on track to meet the 2015 target of 290 G. The5. Global emissions include international bunkers, which are not reflected in regional and country figures.26 World Energy Outlook | Special Report

decarbonisaon eorts in the power sector resulted in a decade long improvement of itsemissions per unit of generaon (Figure 1.9). Energy intensity improved by 3.8%, in linewith the 12 th Five-ear Plan target, indicang progress in diversifying the economy and inenergy eciency.Figure 1.9 CO 2 emissions per unit of electricity generation in Chinag CO2/kWh90085012348007507002000 2003 2006 2009 2012In the Middle East, energy-related CO 2 emissions increased by around 55 Mt CO 2 , or 3.2%,on the back of rising gas consumpon in power generaon and the persistence of subsidisedenergy consumpon. Indias emissions grew by some 45 Mt CO 2 , or 2.5%, mainly driven bycoal. This gure was much lower than the previous year due to lower GDP growth andissues related to domesc coal producon.In OECD countries, the trends are very dierent. CO 2 emissions declined in the UnitedStates year-on-year in 2012 by 200 Mt, or -3.8%, around half as a result of the ongoingswitching from coal to natural gas in power generaon (Box 1.2). Other factors contributedto the decline: increased electricity generaon from non-hydro renewables, lower demandfor transport fuels and mild winter temperatures reduced the demand for heang. CO 2emissions in the United States have now declined four of the last ve years, 2010 being theexcepon (Figure 1.10). Their 2012 level was last seen in mid-1990s.© OECD/IEA, 2013CO 2 emissions in the European Union in 2012 were lower year-on-year by some 50 Mt,or 1.4%, but trends dier markedly from country to country. ith electricity demanddeclining by 0.3% in 2012, in line with a contracon in the economy, cheap coal and carbonprices meant that many large emiers turned partly to coal to power their economies. Coaldemand grew 2.8%, compared with an average 1.3% decline over the past decade. et datashow a 0.6% decline in power sector emissions that are capped under the EU ETS, and alarger, 5.8% fall in emissions from industry sectors such as cement, glass and steel. Nonhydrorenewables generaon increased by 18%, thanks to support policies. Emissions inEuropes biggest economy, Germany, increased by 17 Mt CO 2 or 2.2% (UBA, 2013). Drivenby low coal and low CO 2 prices, consumpon of coal in power generaon increased by 6%.Chapter 1 | Climate and energy trends 27

CO 2 emissions increased also in the United ingdom by 21 Mt, or 4.5%, due to higher coaluse in power generaon and higher demand for space heang (DECC, 2013). Electricitygeneraon from coal increased by 32%, displacing gas in the electricity mix.Box 1.2 6The decline in energy-related CO 2 emissions in the United States in recent years hasbeen one of the bright spots in the global picture. One of the key reasons has beenthe increased availability of natural gas, linked to the shale gas revoluon, which hasled to lower prices and increased compeveness of natural gas versus coal in the USpower sector. Over the period 2008-2012, when total US power demand was relavelyat, the share of coal in US electricity output fell from 49% to 37%, while gas increasedfrom 21% to 30% (and renewables rose from 9% to 12%). 6 The large availability ofspare capacity facilitated this quick transformaon. In 2011, when the share of gas hadalready increased signicantly, the ulisaon rate of combined-cycle gas turbines wassll below 50% (IEA, 2013b). Gas-red combined-cycle plants produce on average halfthe emissions per kilowa hour than convenonal coal-red generaon. Part of thisgain, however, is oset on a life-cycle basis due to methane emissions from natural gasproducon and distribuon.hether the trend in emissions reducon from coal-to-gas switching in powergeneraon will connue depends on relave coal and gas prices. Preliminary signs of areversal were seen in the rst quarter of 2013: coal consumpon in power generaonincreased 14% compared with the same period in the previous year, as natural gasprices at Henry Hub increased by around 40% from $2.45 per million Brish thermalunits (MBtu) in 2012 to $3.49MBtu in the same period of 2013. In the absence ofenvironmental or other regulaons posing addional restricons on CO 2 emissionsstandards on exisng power plants, exisng coal plants could again become economicrelave to gas for natural gas prices in the range $4.5-5MBtu or higher.The resource base for unconvenonal hydrocarbons holds similar promise for othercountries heavily relying on coal in the power sector, such as China. But due to theexpected relave coal to gas prices in regions outside North America, the US story isnot expected to be replicated on a large scale in the period up to 2020. Our analysisdemonstrates increased gas use in all scenarios, including that compable with a 2 °Ctrajectory (the 450 Scenario), but on its own, natural gas cannot provide the answerto the challenge of climate change (IEA, 2011b and 2012b). In the 450 Scenario, forexample, global average emissions from the power sector need to come down to120 g CO 2 kh by the 2030s, almost one-third the level that could be delivered by themost ecient gas-red plant in the absence of CCS technology.© OECD/IEA, 20136.Based on US Energy Informaon Administraon data for 2012.28 World Energy Outlook | Special Report

Figure 1.10 Change in fuel consumption and total energy-related CO 2emissions in the United States1Mtoe60300-306.26.05.85.6GtOilCoalGasCO 2 emissions(right axis)234-605.4-905.2-1202008 2009 2010 2011 20125.0Japans emissions rose by some 70 Mt CO 2 ,or 5.8%, in 2012 a rate of growth last seen twodecades ago, as a consequence of the need to import large quanes of liqueed naturalgas and coal in order to compensate for the almost 90% reducon in electricity generaonfrom nuclear power following the Fukushima Daiichi accident. The increase in fuel importcosts was a key reason for Japans record high trade decit of 6.9 trillion ($87 billion) in2012.Historical emissions trends and indicatorsThe data for 2012 need to be seen in a longer term perspecve. Since 1900, emissionslevels and their geographical distribuon have changed signicantly, with the rst decadeof this century seeing the accumulaon in the atmosphere of eleven mes more CO 2 thanthe rst decade of the previous century. Excluding internaonal bunkers, OECD countriesaccounted for almost all of the global emissions in the 1900s, yet now non-OECD emissionsaccount for 60%. OECD countries emied 40% of global energy-related CO 2 emissionsin 2012, down from 55% in 2000 (Figure 1.11). This compares with around 40% of totalprimary energy demand and 53% of global GDP (in purchasing power parity terms). Thegrowth in Chinas emissions since 2000 is larger than the total level of emissions in 2012of the other BRICS (Brail, Russia, India, China, South Africa) countries combined. Indiasemissions increased in 2012, reinforcing its posion as the worlds third-largest emier.Developing countries tend to be net exporters of products whose producon gives rise toCO 2 emissions, opening up scope for debate as whether responsibility for the emissions lieswith the producer or the importer.© OECD/IEA, 2013Chapter 1 | Climate and energy trends 29

Figure 1.11 Energy-related CO 2 emissions by countryGt353025201510Other non-OECDMiddle EastIndiaChinaOther OECDEuropean UnionJapan161446558700600500400300200Cumulave emissions (Gt)5United States 921001900 1920 1940 1960 1980 2000 2012Sources: IEA databases and analysis Boden et al., (2013).1900-'291930-'591960-'891990-2012Trends in energy-related CO 2 emissions connue to be bound closely to those of the globaleconomy (Figure 1.12), with the few declines observed in the last 40 years being associatedwith events such as the oil price crises of the 1970s and the recent global recession. Thecarbon intensity of the economy has generally improved over me (GDP growth typicallyexceeds growth in CO 2 emissions), but the last decade has seen energy demand growthaccelerate and the rate of decarbonisaon slow mainly linked to growth in fossil-fueldemand in developing countries.Figure 1.12 Growth in global GDP and in energy-related CO 2 emissions8%6%GDP growth (MER)Change in CO 2 emissions4%2%0%-2%1972 1975 1980 1985 1990 1995 2000 2005 2010Note: MER market exchange rate.2012© OECD/IEA, 2013A simple comparison between OECD Europe or the United States, and China or Indiareveals a signicant dierence in GDP and CO 2 trends over me (Figure 1.13). In OECDEurope and the United States, GDP has more than doubled or tripled over the last 40 yearswhile CO 2 emissions have increased by 2% and 18% respecvely. In China and India, GDPand CO 2 emissions have grown at closer rates, reecng their dierent stage of economic30 World Energy Outlook | Special Report

development. This resulted in Chinas emissions overtaking those of the United States in2006, despite its economy being less than one-third the sie.Figure 1.13 GDP and energy-related CO 2 emissions in selected countriesTrillion dollars (2011)2016128420161284GtGDP (MER):United StatesOECD EuropeChinaIndiaCO 2 emissions:(right axis)United StatesOECD EuropeChinaIndia12341971 1975 1985 1995 2005Note: MER market exchange rate.2012Global CO 2 per-capita emissions, having uctuated within a range from around 3.7 to4 tonnes CO 2 from the early 1970s to the early 2000s, have now pushed strongly beyondit, to 4.5 tonnes. Developed countries typically emit far larger amounts of CO 2 percapita than the world average, but some developing economies are experiencing rapidincreases (Figure 1.14). Between 1990 and 2012, Chinas per-capita emissions tripled,rapidly converging with the level in Europe, while Indias more than doubled, thoughremaining well below the global average. Over the same period, per-capita emissionsdecreased signicantly in Russia and the United States, yet remained at relavely highlevels.Figure 1.14 Energy-related CO 2 emissions per capita and CO 2 intensity inselected regionsTonnes per thousand dollars of GDP($2011, MER)2.42.01.61.20.80.4IndiaEuropean UnionChinaRussiaJapanUnited StatesKey19902012© OECD/IEA, 20130 3 6 9 12 15 18 21Tonnes per capitaNotes: Bubble area indicates total annual energy-related CO 2 emissions in that region. MER marketexchange rate.Chapter 1 | Climate and energy trends 31

Trends by energy sectorThe power and heat sector is the largest single source of energy sector CO 2 emissions. Itproduced over 13 Gt of CO 2 in 2011 7 (Figure 1.15), more than 40% higher than in 2000.Trends in CO 2 emissions per kilowa-hour of electricity produced in a given country largelyreect the nature of the power generaon. Countries with a large share of renewables ornuclear, such as Brail, Canada, Norway and France have the lowest level. Of those regionsrelying more heavily on fossil fuels, the large natural gas consumers, such as Europe andRussia, have levels below the world average. Despite eorts in many countries to developmore renewable energy, in global terms the power sector is sll heavily reliant on coal,accounng for nearly three-quarters of its emissions. Australia, China, India, Poland andSouth Africa are examples of countries sll heavily reliant on coal to produce electricity,reecng their resource endowment. In the United States, electricity generaon from coalhas decreased 11% since 2000, coal consumpon for power generaon falling by 64 milliontonnes of oil equivalent (Mtoe) and yielding a decline in overall emissions from the powersector of 0.8% per year on average.Figure 1.15 World energy-related CO 2 emissions by fuel and sector, 2011CO 2 emissions from transport, the largest end-use sector source, were just under 7 Gtin 2011. 8 Emissions from the sector, which is dominated by oil for road transport, haveincreased by 1.7% per year on average since 2000, but with diering underlying regionaltrends. OECD transport emissions are around 3.3 Gt: having declined to around year-2000levels during the global recession, they have remained broadly at since. Market saturaonin some countries and increasing eciency and emissions standards appear to be curtailing© OECD/IEA, 20137. CO 2 emissions data by sector for the year 2012 were not available at the time of writing. Unlike previoussections, this one uses 2011 as latest data point.8. At the global level, transport includes emissions from international aviation and bunkers.32 World Energy Outlook | Special Report

emissions growth. Non-OECD transport CO 2 emissions have increased by more than 60%since 2000, reaching 2.5 Gt in 2011, with increased vehicle ownership being a key driver.Emissions in China and India have both grown strongly but, collecvely, their emissions fromtransport are sll less than half those of the United States. More than 50 countries haveso far mandated or promoted biofuel blending to diminish oil use in transport. Emissionsfrom internaonal aviaon and marine bunkers are on a steady rise. They reached 1.1 Gtin 2011, up from 0.8 Gt in 2000.Having remained broadly stable at around 4 Gt for much of the 1980s and 1990s, CO 2emissions from industry have increased by 38% since the early 2000s, to reach 5.5 Gt. Allof the net increase has arisen in non-OECD countries, with China and India accounngfor some 80% of the growth in these countries. China now accounts for 60% of globalcoal consumpon in industry. Iron and steel industries account for about 30% of total CO 2emissions from the industrial sector.1234Total energy-related CO 2 emissions in the buildings sector (which includes residenal andservices) reached 2.9 Gt in 2011, connuing the gradually increasing trend of the lastdecade. Natural gas is the largest source of emissions about 50% of the total with theOECD (mainly the United States and Europe) accounng for two-thirds of the total. Non-OECD emissions from oil overtook those of the OECD, which are in decline, in 2011. Manyother changes in the buildings sector, such as increasing electricity demand for lighng,cooking, appliances and cooling, are captured in changes in the power sector.Outlook for energy-related emissions and the 450 ScenarioThis secon analyses the disconnect between the energy path the world is on and anenergy pathway compable with a 2 °C climate goal. It does so by presenng and analysing,by fuel, region and sector, the essenal dierences between the New Policies Scenario, ascenario consistent with the policies currently being pursued, and the 450 Scenario, a 2 °Cclimate scenario, both of which were fully developed in WEO-2012 (Box 1.3). Our analysisshows that the energy projecons in the New Policies Scenario are consistent, other thingsbeing equal, with a 50% probability of an average global temperature increase of 3 °C by2100 (compared with pre-industrial levels) and of 3.6 °C in the longer term. 9 This comparesto 1.9 °C by 2100 and 2 °C in the long term in the 450 Scenario. This indicates the extent towhich the energy world is going to have to change: connuing on todays path, even withthe assumed implementaon of new policies, would lead to damaging climac change.The present energy trajectory indicates increasing energy-related CO 2 emissions throughto 2035. By contrast, to meet the requirements of the 450 Scenario, emissions need topeak by 2020 and decline to around 22 Gt in 2035 around 30% lower than in 2011, alevel last seen in the mid-1990s (Figure 1.16). The cumulave emissions gap between© OECD/IEA, 20139. The long-term temperature change is associated with a stabilisation of greenhouse-gas concentrations, whichis not expected to occur before 2200.Chapter 1 | Climate and energy trends 33

the scenarios over the projecon period is around 156 Gt, an amount greater than thatemied by the United States over the last 25 years. Such a path of declining emissionsdemands unprecedented change. The 450 Scenario shows how it could be achieved, basedon policies and technologies that are already known but, crucially, it requires urgentcommitment to strong acon, followed by robust, unwavering implementaon. If the450 Scenario trajectory is successfully followed, by 2035, non-OECD countries will haveachieved more than 70% of the total reducon (10.5 Gt) in annual CO 2 emissions in the450 Scenario, compared with the New Policies Scenario.Figure 1.16 World energy-related CO 2 emissions by scenarioGt403530New Policies Scenario156 Gt2520450 Scenario151990 1995 2000 2005 2010 2015 2020 2025 2030 2035In both scenarios considered here, GDP growth averages 3.1% per year and populaongrowth averages 0.9%, pushing total primary energy demand higher but this demandis met increasingly from low or ero-carbon sources. To be consistent with the requiredtrajectory in the 450 Scenario, energy-related CO 2 emissions must begin to decline thisdecade, even though the level of energy demand is expected to increase by 0.5% per year,on average: CO 2 emissions peak by 2020 and then decline by 2.4% per year on average unl2035. ooking across the fossil fuels, gas demand increases by 0.7% per year on average,oil decreases by 0.5% per year and coal declines by 1.8% per year. Energy eciency policiesare the most important near-term emissions migaon measure (see Chapter 2 for moreon ways to save CO 2 in the short term). By 2035, acons to improve energy eciencysuccessfully reduce global emissions in that year by 6.4 Gt equivalent to about 20%of global energy-related CO 2 emissions in 2011. The payback periods for many energyeciency investments are short, but non-technical barriers oen remain a major obstacle.It is these barriers that governments need to tackle (see WEO-2012 and Chapter 2).© OECD/IEA, 201334 World Energy Outlook | Special Report

Box 1.3 Overview of the New Policies Scenario and the 450 Scenario 101112The analysis in this chapter of the disconnect between the energy path the world iscurrently on and an energy trajectory consistent with a 50% chance of achieving the2 °C climate goal relies on two scenarios, both of which were fully developed in theWEO-2012. 10• The New Policies Scenario, though founded essenally on exisng policies andrealies, also embodies some further developments likely to improve the energytrajectory on which the world is currently embarked. To this end, it takes intoaccount not only exisng energy and climate policy commitments but alsoassumed implementaon of those recently announced, albeit in a cauousmanner. Assumpons include the phase-out of fossil-fuel subsidies in imporngcountries and connued, strengthened support to renewables. The objecve ofthis scenario is to provide a benchmark against which to measure the potenalachievements (and limitaons) of recent developments in energy policy in relaonto governments stated energy and climate objecves.• The 450 Scenario describes the implicaons for energy markets of a co-ordinatedglobal eort to achieve a trajectory of greenhouse-gas emissions consistent withthe ulmate stabilisaon of the concentraon of those gases in the atmosphere at450 ppm CO 2 -eq (through to the year 2200). This scenario overshoots the 450 ppmlevel before stabilisaon is achieved but not to the extent likely to precipitatechanges that make the ulmate objecve unaainable. The 450 Scenariooers a carefully considered, plausible energy path to the 2 °C climate target.For the period to 2020, we assume policy acon sucient to implement fullythe commitments under the Cancun Agreements. Aer 2020, OECD countriesand other major economies are assumed to set emissions targets for 2035 andbeyond that collecvely ensure an emissions trajectory consistent with ulmatestabilisaon of greenhouse-gas concentraon at 450 ppm, in line with what wasdecided at COP-17 to establish the Durban Plaorm on Enhanced Acon, tolead to a new climate agreement. e also assume that, from 2020, $100 billionin annual nancing is provided by OECD countries to non-OECD countries forabatement measures.1234The projecons for the scenarios are derived from the IEAs orld Energy Model (EM) a large-scale paral equilibrium model designed to replicate how energy marketsfuncon over the medium to long term. 11 The OECDs EN-inkages model has beenused to provide the macroeconomic context and for the projecons of greenhouse-gasemissions other than energy-related CO 2 . 12© OECD/IEA, 201310.The detailed list of policies by region, sector and scenario is available at: www.worldenergyoutlook.org/media/weowebsite/energymodel/policydatabase/WEO2012_AnnexB.pdf.11.A full descripon of the EM is available at www.worldenergyoutlook.org/weomodel.12.For more informaon on the OECD EN-inkages model see Burniaux and Chateau (2008).Chapter 1 | Climate and energy trends 35

In the 450 Scenario, CO 2 emissions per capita decline gradually prior to 2020 and then,reecng more robust policy acon, decline more rapidly, the global average reaching2.6 tonnes CO 2 per capita in 2035 (compared with 4.3 tonnes CO 2 in the New PoliciesScenario). Signicant variaons persist across regions, with the non-OECD average percapita level sll being less than half that of the OECD in 2035.In the 450 Scenario, the carbon intensity of the world economy is around one-thirdof exisng levels by 2035, with many non-OECD countries delivering the biggestimprovements (Figure 1.17) as they seie the opportunity to base their extensive investmentprogrammes in addional energy supply on low-carbon sources. OECD countries are,however, far from free of challenge. In the 450 Scenario, energy-related CO 2 emissionsin the OECD are around half current levels by 2035, reaching just over 6 Gt a decline ofnearly 3% per year on average.Figure 1.17 CO 2 intensity in selected regions in the 450 ScenarioTonnes per thousand dollars of GDP($2011, MER)1.2India1.0 China0.8 Russia0.6World0.40.2Middle EastUnited StatesEuropean Union2010 2015 2020 2025 2030 2035Note: MER market exchange rate.Sectoral trendsThe 450 Scenario requires a rapid transformaon of the power sector. In some respects itinvolves only an acceleraon of trends already underway, such as moving to more ecientgeneraon technologies and the increased deployment of renewables, but innovaon isalso required, such as the adopon of CCS technology. Overall, electricity generaon in2035 is 13% lower than in the New Policies Scenario, but CO 2 emissions from the powersector are more than 10 Gt (70%) less (Figure 1.18). Electricity demand in transport in thatyear is 85% higher in the 450 Scenario than in the New Policies Scenario, but it is 17% lowerin buildings, due to more ecient appliances, heang equipment and lighng. In industry,electricity demand is 12% lower in 2035, mainly due to more ecient motor systems.© OECD/IEA, 201336 World Energy Outlook | Special Report

Figure 1.18 World energy-related CO 2 emissions abatement by sector inthe 450 Scenario relative to the New Policies Scenario1Gt40Indirect2302010DirectPower sector*TransportIndustryBuildingsAgriculture3420112035New PoliciesScenario2035450 ScenarioIndirect electricity savings in the power sector result from demand reducon in end-use sectors, whiledirect savings are those savings made within the power sector itself (e.g. plant eciency improvements).Direct savings include heat plants and other transformaon.In the 450 Scenario, the share of electricity generaon from fossil fuels declines from morethan two-thirds in 2011 to one-third in 2035. Electricity generaon from coal declinesto half of exisng levels by 2035 and installed capacity is 1 100 G lower than in theNew Policies Scenario (see Chapter 3 on the risk of stranded assets). In the OECD, thegreatest change in coal-red capacity occurs in the United States, but the biggest changesglobally are in non-OECD countries, where the recent reliance on new fossil-fuel capacity(especially coal) to meet rising demand gives way to increased use of low-carbon sources.Natural gas is the only fossil fuel with increasing electricity generaon in the 450 Scenario,but it sll peaks before 2030 and then starts to decline, ending 18% higher in 2035 thanin 2011. CCS becomes a signicant source of migaon from 2020 and saves 2.5 Gt CO 2in 2035, equivalent to around one-and-a-half mes Indias emissions today. In severalcountries, including China and the United States, very ecient coal-red power staonsare built up to 2020 and are retroed with CCS in the following years. Installed globalnuclear capacity doubles by 2035 in the 450 Scenario, signicantly higher than in the NewPolicies Scenario, with the largest increases in China and India, and addional capacitybeing installed in the United States and Europe. Electricity generaon from renewablesincreases almost 11 000 terawa-hours (Th) to 2035, with wind, hydro and solar Pgrowing most strongly. Renewables-based electricity generaon supplies almost half theworlds electricity in 2035.© OECD/IEA, 2013In the 450 Scenario, global transport CO 2 emissions peak around 2020 but then decline,ending 5% below 2011 levels in 2035 (2.4 Gt below the level in the New Policies Scenarioin 2035). A range of migaon measures is incorporated in the 450 Scenario, with fueleciency gains and an increase in the use of biofuels being parcularly important up to2020. Such policies are already in place in the United States, which has mandated the useChapter 1 | Climate and energy trends 37

of 36 billion gallons of biofuels by 2022, and in the European Union, where the RenewableEnergy Direcve requires a mandatory share of 10% renewable energy in transport by2020. Improved eciency becomes even more important globally aer 2020, alongsidelower growth in vehicle usage in countries where subsidies are removed.In industry, global energy-related CO 2 emissions in 2035 are 5% lower than in 2011 inthe 450 Scenario, at around 5.2 Gt, 21% lower than the New Policies Scenario. Improvedenergy eciency accounts for more than half the reducon in cumulave terms, with CCSin energy-intensive industries and fuel switching also playing a role. More than 80% of theCO 2 savings in the sector in the 450 Scenario come from non-OECD countries, with China,India and the Middle East all making notable improvements. By 2035, global emissions inbuildings are 11% lower than 2011 in the 450 Scenario, at around 2.6 Gt, with the savingsrelave to the New Policies Scenario being spread relavely evenly between OECD andnon-OECD countries. Much more energy ecient buildings are adopted from around 2020onwards.InvestmentA 2 °C world as in the 450 Scenario requires increased investment in the power sectorand in end-use sectors, but reduced investment in fossil-fuel supply. In the 450 Scenario,total investment in fossil-fuel supply is $4.9 trillion lower than in the New Policies Scenariothrough to 2035, and investment in power transmission and distribuon networks isaround $1.2 trillion lower. However, this saving is more than oset by a $16.0 trillionincrease in investment in low-carbon technologies, eciency measures and other formsof intervenon. Part of the incremental investment is oset by savings in consumersexpenditure on energy. Addional investment across OECD countries reaches around$590 billion per year in 2035 and in non-OECD countries around $760 billion (Figure 1.19).Figure 1.19 World annual additional investment and CO 2 savings in the450 Scenario relative to the New Policies ScenarioTrillion dollars (2011)1.61.20.80.4-4-8-12Addionalinvestment:Non-OECDOECDCO 2 Savings:(right axis)OECDNon-OECD© OECD/IEA, 20132015 2020 2025 2030 203538 World Energy Outlook | Special Report-16Gt

Transport requires the largest cumulative additional investment in the 450 Scenario,relative to the New Policies Scenario $6.3 trillion (Figure 1.20). Most of this is directedtowards the purchase of more efficient or alternative vehicles. The buildings sectorrequires $4.4 trillion in cumulative additional investment, but this reflects investment thatboth delivers direct abatement from buildings and indirect abatement through reducedelectricity demand. The decarbonisation of the power sector requires a net additional$2.0 trillion, after accounting for the lower investment need for transmission anddistribution lines. More than 80% of the additional investment in electricity generationgoes to renewables-based technologies. Industry invests an additional $1.5 trillion,around one-quarter of it directed to CCS.1234Figure 1.20 Cumulative change in world investment by sector in the450 Scenario relative to the New Policies Scenario, 2012-2035Trillion dollars (2011)7654321Transport Buildings Power plants Industry Biofuels supplyNote: Investment in power plants increases, but investment for transmission and distribuon (not shownhere) declines by a cumulave total of around $1.2 trillion over the period.© OECD/IEA, 2013The transformaon of the global energy system in the 450 Scenario delivers signicantbenets in terms of reduced fossil-fuel import bills, enhanced energy security, beer airquality, posive health impacts and reduced risk of energy-related water stress. Fossil-fuelprices are lower in the 450 Scenario than the New Policies Scenario (Table 1.1), driven bylower demand. In real terms, the IEA crude oil import price needed to balance supply anddemand in the 450 Scenario reaches $115barrel (in year-2011 dollars) around 2015 andthen declines to $100barrel in 2035 ($25barrel lower than the New Policies Scenario).Coal and gas prices are also lower in the 450 Scenario. The steam coal import price in theOECD is almost 40% cheaper in 2035 and the natural gas price in Europe and the Pacic isaround 20% cheaper. ower internaonal fuel prices and lower demand might be expected,other things being equal, to lead to lower fuel expenditure by consumers. But we assumethat end-use fuel prices in the transport sector are kept at higher levels through taxaon,Chapter 1 | Climate and energy trends 39

The 24% reducon in the cost of local polluon controls (for SO 2 , NO X and PM 2.5 ) in 2035in the 450 Scenario, relave to the New Policies Scenario, is small when compared withenergy sector investment costs or potenal fossil-fuel import bill savings, but is associatedwith improved quality of life and health. In China, local polluon in several cies hasalready prompted increased government acon. In our New Policies Scenario, polluoncontrol costs increase by nearly 80% to 2035, and non-OECD polluon control costs as awhole overtake those of the OECD around the middle of the projecon period (Table 1.2).In the 450 Scenario, world polluon control costs sll rise, but at a much slower rate, withthe OECD level being similar to 2011 in 2035 and the non-OECD level being much lowerthan in the New Policies Scenario.1234Table 1.2 Pollution control costs by region and scenario ($2011 billion)New Policies Scenario450 Scenario2011* 2020 2035 2020 2035OECD 203 256 261 244 206United States 72 89 85 86 65Europe 81 106 112 100 94Japan 22 23 21 22 15Other OECD 29 38 42 36 32Non-OECD 124 234 325 220 237Russia 8 14 18 14 14China 49 96 124 89 81India 6 15 34 14 28Middle East 11 21 32 20 24an America 17 31 41 30 33Other non-OECD 33 57 76 53 58World 327 489 586 463 443European Union 86 108 124 100 94 Esmate.Source: IIASA (2012).© OECD/IEA, 2013Chapter 1 | Climate and energy trends 41

© OECD/IEA, 2013

Chapter 2Energy policies to keep the 2 °C target aliveShort-term actions for long-term gainHighlights© OECD/IEA, 2013 The absence of early, tangible achievement in the internaonal climate negoaonsand the sluggish global economy are threatening the viability of the 2 °C climate goalby weakening condence in the investment case for a low-carbon economy. To keepthe door to the 2 °C target open, we propose a set of pragmac policy acons that,without harming economic growth and using available technologies and policies, canresult in a global peak in energy-related GHG emissions by 2020. The four priorityareas in our 4-for-2 °C Scenario are: specic energy eciency measures limits tothe use and construcon of inecient coal power plants minimising methane (CH 4 )releases to the atmosphere in oil and gas producon and a paral phase-out offossil-fuel subsidies. In the 4-for-2 °C Scenario, energy-related CO 2 and CH 4 emissions increase from 33.3 Gtin 2010 to 34.9 Gt in 2020 (measured on a CO 2 -eq basis) and decline thereaer.Emissions in 2020 are 3.1 Gt lower than the course on which we otherwise appear tobe set, delivering 80% of the abatement needed to be on track with a 2 °C trajectory. Energy eciency accounts for 49% of the savings realised, limitaons on inecientcoal-red power plants for 21%, lower methane emissions in upstream oil and gasfor 18%, and the paral phase-out of fossil-fuel subsidies for 12%. Restricons oncoal use support the growth of renewables, which increase their share in powergeneraon to 27% in 2020, up from around 20% today. In addion to addressing climate change, the policies assumed in the 4-for-2 °CScenario reduce local air polluon and increase energy security without hamperingeconomic growth of any given region. Required addional investment to 2020 ismore than oset by reduced spending on fuel bills. A gradual reorientaon of theeconomy resulng from the implementaon of the four policies entails losses insome sectors, including oil and gas upstream and electricity, but gains in other areas. The 4-for-2 °C Scenario buys precious me to keep the 2 °C target alive, whileinternaonal negoaons connue, avoiding much carbon lock-in but it isinsucient to limit the long-term temperature increase to 2 °C. A frameworkconducive to more ambious abatement aer 2020 needs to be developed, notleast to provide clear market signals to businesses and long-term investors, notablyincluding a global carbon price and roll-out of low-carbon technologies at scale. Inthe 450 Scenario, delaying CCS deployment by ten years would increase the cost ofdecarbonisaon in the power sector by $1 trillion and result in lost revenues for coalproducers ($690 billion) and oil and gas producers ($660 billion).Chapter 2 | Energy policies to keep the 2 °C target alive 43

Introduconarious iniaves have recently been undertaken with the explicit objecve of reducinggreenhouse-gas emissions, such as new schemes to price carbon-dioxide (CO 2 ) emissions(either through cap-and-trade programmes or carbon taxes) in Australia, orea andCalifornia, along with other measures that serve implicitly to incorporate CO 2 pricing intoinvestment decisions in the energy sector (Chapter 1). Some acons taken primarily forother purposes, for example to reduce local air polluon or improve energy eciency orin response to price changes, also benet CO 2 abatement. The recent switch from coal tonatural gas in the power sector of the United States as a result of lower gas prices is oneexample of climate benets derived from changes driven by the market, rather than bydeliberate climate policy acon. Nonetheless, the chance of achieving abatement on thescale needed to follow a trajectory consistent with a global average temperature rise ofno more than 2 degrees Celsius (°C) now appears more remote than it was several yearsago, parcularly as governments grapple with economic crisis in many parts of the world.A rst eect of lower economic acvity in some regions has been to reduce the expectedlevel of emissions but the crisis has also curtailed direct government acon to limit climatechange, partly as a result of fears that more stringent climate policies could result in a lossof economic compeveness. In some cases, these concerns have been heightened bywide divergences in energy prices between dierent regions.In view of the long lifeme of capital stock in the energy sector, lack of momentum towardsconcerted global climate policy acon directly increases the scale of the challenge to meetthe 2 °C climate goal by failure to deter addional investment in emissions-intensiveinfrastructure, thereby locking in emissions for decades to come. The date at which theexisng energy infrastructure will lock in all the CO 2 emissions from the energy sectorprovided for in a global CO 2 emissions budget consistent with a 2 °C trajectory, leaving noprovision for emissions from new carbon-eming infrastructure to meet growing demand,is close. Thereaer, it becomes ever more costly and dicult to achieve the stated goal(see Chapter 3). In addion, research suggests that there is a point of no return at whichclimate feedbacks could become self-reinforcing (though there is remaining uncertainty asto exactly when this occurs), thus closing the door to 2 °C forever (enton, et al., 2008).hile the metable to which internaonal climate negoators are working provides forimplementaon of a legally-enforceable agreement to reduce emissions from 2020, ourprojecons suggest that, without earlier addional acon at naonal level, global energyrelatedCO 2 and methane (CH 4 ) emissions in 2020 will already be 3.9 gigatonnes (Gt)CO 2 -equivalent (CO 2 -eq) above the level needed to follow a 2 °C trajectory.© OECD/IEA, 2013It thus becomes essenal to consider what can be done in the short term to keep the doorto 2 °C open. It seems unlikely that naonal policy makers will implement acons that arechallenging to their naonal economy given the economic situaon in many countries. In thischapter, therefore, we set out to idenfy a set of pragmac and achievable policy measureswhich, in net terms, do no harm to naonal economic growth yet which, taken together,would reduce global greenhouse-gas emissions in the period to 2020 by substanally more44 World Energy Outlook | Special Report

than the reducon expected to be achieved by exisng and planned policies alone. Thesemeasures would take us only part of the way towards an emissions trajectory that wouldachieve the 2 °C goal but in the second part of the chapter we explore addional elementsto support ambious abatement aer 2020, which would help the overall goal to be met.GDP-neutral emissions abatement to 2020Many policies to support the growth of low-carbon technologies and to moderate thegrowth of energy demand unl 2020 and beyond are in place or already planned today:these are the policies embodied in the New Policies Scenario of the World Energy Outlook2012 (WEO-2012), a scenario consistent with the course on which governments appear atpresent to be embarked (IEA, 2012a). Exceponally, as a result of the policy focus over thelast decade, the deployment of renewables today is already broadly on track towards themore ambious level required to deliver their expected contribuon in 2020 to meenglong-term climate targets (IEA, 2013a). On the other hand, while energy eciency, too,has been high on the policy agenda in recent years, exisng and planned policies are likelyto leave two-thirds of the global economically viable energy eciency potenal untapped(IEA, 2012a). Therefore, much wider adopon of eciency measures will be necessary tofull the energy eciency expectaons of a scenario consistent with the achievement ofthe internaonal 2 °C climate target.1234The short-term measures considered in this chapter, collecvely embodied in whatwe describe as the 4-for-2 °C Scenario, go beyond policies already adopted and entailmeasures that require either signicant further strengthening and wider adopon, orthat are currently not high on the policy agenda, even though the measures required toimplement the relevant policies are known and their adopon could make a signicantaddional dierence. This is the approach adopted in the 4-for-2 °C Scenario, which isbased on two core assumpons. First and foremost, the measures that are assumed to beadopted are readily available today, meaning they do not require the idencaon andimplementaon of innovave sets of energy policies or the deployment of technologiesthat have yet to be proven in the market. Though the individual measures have notyet been adopted everywhere, they have already been proven in some countries, andtherefore just need to be tailored to naonal circumstances elsewhere. Second, the setof measures adopted, when taken together and in net terms, does not adversely aecteconomic growth in any given country or region. Although the proposed measures involvean inial deployment cost, the set of proposed policies as a whole is calculated to delivereconomic savings (such as through lower fuel bills) to the extent that the inial deploymentcosts of the proposed policies are oset within each economy, considered as a whole. 1 Asa consequence, the set of policies proposed does not harm overall economic growth up to2020. Beyond 2020, it actually improves the compeveness of the economies concerned© OECD/IEA, 20131. The judgement expressed here about economic effects apply to regions taken as a whole, using standardgroupings in the World Energy Outlook series (see www.worldenergyoutlook.org).Chapter 2 | Energy policies to keep the 2 °C target alive 45

and the ability of their energy system to make the transion towards a low-carbon basisin the long term.The emphasis of the 4-for-2 °C Scenario is on measures which can be implementedeecvely in the short term, to provide breathing space for the internaonal negoaonsaimed at policy implementaon by 2020. The measures adopted produce valuable resultsin the period to 2020, though their eect connues beyond that date. In developing the4-for-2 °C Scenario, we reviewed a wide range of measures that we assessed as being bothpraccal and implementable in a short me frame and capable of having a signicantimpact on global greenhouse-gas emissions in the period to 2020. e then analysedthe impact on global energy consumpon and emissions of the implementaon of thepackage of measures under consideraon using the IEAs orld Energy Model (EM) andthe impact on GDP at regional level using the EN-inkages model of the Organisaon forEconomic Co-operaon and Development (OECD). 2 If the package of measures as a wholewas found to reduce economic growth in the period to 2020 in any region, then the levelof ambion of the policies with the most severe negave impact on GDP was reduced or where applicable the measure was abandoned.Based on this iterave process, we have idened a package of four measures, elaboratedbelow, that meet the criteria of making a signicant contribuon to CO 2 abatement inthe period to 2020 without adversely aecng economic growth. Each of the measuresselected can be readily implemented and does not require the use of new technologieswith high upfront deployment costs that would require me to apply beyond nichemarkets (such as electric vehicles), nor major technological breakthroughs, nor radicalchanges in consumer behaviour (except those induced by changing prices or increasedavailability of capital in certain sectors). Many of the measures that were excluded from the4-for-2 °C Scenario might well be cost-eecve in the long-run, but they are judged tohave less certain potenal to make a signicant impact on global emissions by 2020.Highly successful exisng policies, like support for renewables, have not been selected forenhancement in the short term if they appear to be broadly on track to deliver in 2020 thecontribuon that they are required to make in the (more demanding) 450 Scenario, whichis consistent with achievement of the long-term climate objecve.The four policy measures adopted in the 4-for-2 °C Scenario are (Figure 2.1):• Targeted specic energy eciency improvements in the industry, buildings andtransport sectors.• iming the use and construcon of inecient coal-red power plants.• Minimising methane emissions in upstream oil and gas producon.• Further paral phase out of fossil-fuels subsidies to end-users.© OECD/IEA, 20132. EM is a partial equilibrium model. EN-inkages is a computational general equilibrium model. Thecoupling of both models allows the impact of energy policy on economic growth to be assessed.46 World Energy Outlook | Special Report

Figure 2.1 Policy pillars of the 4-for-2 °C Scenario1234Although the adopon of these measures is primarily directed at the reducon ofgreenhouse-gas emissions, they also oer important co-benets and oen complementeach other (Table 2.1, see also Box 2.3). The phase-out of fossil-fuel subsidies, for example,would incenvise energy eciency improvements, while the use of more ecient end-usetechnologies complements the limitaon on the use of inecient coal power generaon bymoderang growth in electricity demand.Table 2.1 ClimateLocal airpolluonEnergysecurityEconomicgrowthEnergypovertyImproving energy eciency iming inecient coal use in power Reducing upstream methane emissions Fossil-fuel subsidy phase-out © OECD/IEA, 2013In a special focus on energy efficiency, the WEO-2012 identified an extensive rangeof measures, by country and by sector, capable of reducing energy consumption ina cost-effective manner (IEA, 2012a). However, since implementation of some of theefficiency policies identified in WEO-2012 would depend upon the prior elimination ofserious market barriers (which in practice could take considerable time), only a selectedsub-set of the measures are adopted in 4-for-2°C Scenario, namely: (i) reducingenergy use from new space and water heating, as well as cooling equipment (ii) moreefficient lighting and new appliances (iii) improving the efficiency of new industrialmotor systems and (iv) setting standards for new vehicles in road transport. Measuresto meet the objectives are already widely deployed in many countries, using readilyavailable technologies and methods. hile there are some market barriers, steps toChapter 2 | Energy policies to keep the 2 °C target alive 47

overcome them have been identified and successfully implemented. Bilateral andormultilateral agreements could facilitate their adoption and implementation on a widerscale.In the power sector, we rst assume that a ban is introduced on the construcon of newsubcrical coal-red power plants (although it does not apply to units already underconstrucon). The means of implemenng such a policy is likely to dier by market, but avariety of opons is already available including: the adopon of stringent energy eciencyor CO 2 emissions standards for coal power plants the adopon of air polluon standardsor pricing the use of carbon, for example through an emissions trading scheme. Second, forexisng inecient coal power plants, we assume that their level of operaon is reduced tothe extent achievable, with the constraint of maintaining adequate electricity supply. Theimpact of this assumpon varies by region, reecng dierences in the power generaoneet, the quality of coal used and the level of electricity demand. Intervenon for exisngunits is likely to take a more direct regulatory form, for example assigning power produconlimits to each generator according to the make-up of its power plant eet (in liberalisedmarkets), or allocang generaon slots, renewing (or not) operaonal licences or alteringthe dispatch schedule to favour more ecient plants (in regulated markets).© OECD/IEA, 2013In the 4-for-2 °C Scenario, we also assume that policies are adopted to reduce releases ofmethane to the atmosphere in upstream oil and gas producon. This primarily aectslocaons where the incenve to reduce methane releases are currently insucient,e.g. due to a lack of domesc demand. hen producing oil and natural gas, a certainproporon of gas oen escapes into the atmosphere, either intenonally as part of normalvenng operaons, or inadvertently, for example due to the reliance on old infrastructure.In addion some natural gas can be released due to incomplete combuson either duringshort-term aring (which is somemes necessary for safety reasons or may be temporarilypermied to test the sie of newly discovered reservoirs), or when natural gas producedin associaon with oil is ared on a roune basis, as it is at a number of locaons aroundthe world, due to lack of infrastructure to ulise the gas. Most of the natural gas that isreleased into the atmosphere in these ways is methane, which is a greenhouse gas with aGlobal arming Potenal (GP) 25 mes higher than that of CO 2 over 100 years. 3 Globally,we esmate that in 2010 natural gas releases to the atmosphere during upstream oil andgas operaons resulted in 45 million tonnes (Mt) of CH 4 emissions (1 115 Mt CO 2 -eq),or around 50% of total oil- and gas-related CH 4 emissions. Other signicant sources ofmethane leakage include leakage from transmission and distribuon pipelines. Measuresto reduce greenhouse-gas emissions from such sources could have a signicant impact,but they have not been included in the 4-for-2 °C Scenario, as it is unlikely that they couldbe put in place prior to 2020, in parcular in countries with large transmission pipelinenetworks, such as Russia, or with ageing gas distribuon networks, such as the UnitedStates, Europe and Russia.3. Global arming Potential estimates the warming effect of different greenhouse gases relative to each other.48 World Energy Outlook | Special Report

Box 2.1 Mitigating short-lived climate pollutantsShort-lived climate pollutants (SCPs) are substances with a lifeme in the atmosphereranging from a few days to several decades that mainly aect the climate in the(relavely) short term. CO 2 emissions, by contrast, aect the climate system overa much longer me horion. SCPs are responsible for a substanal fracon of theradiave forcing to date. The major SCPs are black carbon, methane, troposphericoone and some hydro uorocarbons (HFCs). Black carbon is produced by theincomplete combuson of fossil fuels and biomass, and is a primary componentof parculate maer and parculate air polluon. In 2010, household air polluonand ambient outdoor parculate maer polluon were esmated to have caused,respecvely, over 3.5 and 3.2 million premature deaths (im, et al., 2012). Accordingto the United Naons Environment Programme (UNEP), black-carbon emissions areexpected to remain stable overall through 2030, decreasing in OECD countries andincreasing in non-OECD countries (UNEP, 2011).1234Although the adopon of strategies to reduce SCPs (with the excepon of methane) 4are not considered in the 4-for-2 °C Scenario, recent studies have idened sixteenmigaon measures related to SCPs which use technologies and pracces that alreadyexist (UNEPMO, 2011 UNEP, 2011). These studies esmate that the adopon ofsuch measures by 2030 would reduce the warming expected by 2050 by 0.4-0.5 °C(and, in the Arcc, by about 0.7 °C even in 2040), while each year prevenng morethan two million premature deaths and over 30 Mt of crop losses. There could beassociated reduced disrupon of rainfall paerns.Strategies that reduce emissions of SCPs complement CO 2 migaon by reducingshort-term increases in temperature, thereby minimising the risk of dangerous climatefeedbacks. However, lasng climate benets from fast acon on SCPs are conngenton stringent parallel acon on longer-lasng CO 2 emissions. In other words, whilefast acon to migate SCPs could help slow the rate of climate change and improvethe chances of staying below the 2 °C target in the near term, longer term climateprotecon depends on deep and persistent cuts in CO 2 emissions being rapidly realised.Subsidies for fossil-fuel consumpon lead to an inecient allocaon of resourcesand market distoron by encouraging excessive energy use. hile they may have wellintenonedobjecves, social ones for example, in pracce they have usually proven tobe an unsuccessful or inecient means of achieving their goals. Moreover, they invariablyhave unintended negave consequences, such as encouraging wasteful and inecientconsumpon, thereby contribung to climate change. The latest IEA esmates indicatethat fossil-fuel consumpon subsidies amounted to $523 billion in 2011, up almost© OECD/IEA, 20134.The methane emissions reducon measures discussed in the 4-for-2 °C Scenario also contribute toreducon of black carbon emissions and their eect on climate change (Stohl, et al., 2013).Chapter 2 | Energy policies to keep the 2 °C target alive 49

30% on 2010 and six mes more than the global nancial support given to renewables(IEA, 2012a). In those regions where the subsidies exist, this level of subsidy equatesto an incenve of $110 per tonne CO 2 to consume fossil fuels. Fossil-fuel subsidies areoen intended to improve access to modern energy services for the poor, but our analysisindicates that only 8% of the subsidy granted typically reaches the poorest income group(IEA, 2011a). Polical support for fossil-fuel subsidy reform has been building in recentyears, and G20 and APEC (Asia-Pacic Economic Cooperaon) member economies havemade commitments to phase out inecient fossil-fuel subsidies and many are now movingahead with implementaon. In the 4-for-2 °C Scenario, recognising the polical challengeinvolved in subsidy phase-out, we assume, as in the New Policies Scenario of WEO-2012, atotal subsidy phase-out by 2020 in fossil-fuel imporng countries but in exporng countries(where sustained reforms are likely to be even more dicult to achieve) subsidisaon ratesare only reduced by an addional 25% by 2020, relave to the New Policies Scenario, withsubsidies completely removed by 2035. 5 Average end-use prices in net-exporng regionsremain signicantly lower than in many other parts of the world as we make no assumponthat they introduce any new taxes or excise dues on energy. For example, average gasolineprices in the Middle East are around one-h of the OECD average in 2020.Emissions abatement to 2020Eecve implementaon of the proposed policy measures would have a profound impacton energy-related greenhouse-gas emissions. In the 4-for-2 °C Scenario, emissions arelower by 3.1 Gt (in CO 2 -eq terms) in 2020, compared with the New Policies Scenario, 6although they are sll higher than today (Figure 2.2). Energy eciency makes the largestcontribuon to abatement, at 1.5 Gt (or 49%) in 2020. 7 Contribuons to abatement fromrestricons on subcrical coal-red power plants are around 640 Mt (21%), the reduconof methane emissions in upstream oil and gas producon at more than 570 Mt (18%) andthe paral phase-out of subsidies to fossil fuels consumed by end-users at more than360 Mt (12%). In each case, these savings come on top of those assumed in the trajectoryresulng from policies that are already adopted or under consideraon by governments(the New Policies Scenario).© OECD/IEA, 20135. Subsidisation rate is calculated as the difference between the full cost of supply and the end-user price,expressed as a proportion of the full cost of supply. For countries that import a given product, subsidy estimatesare explicit. In contrast, for countries that export a given product, subsidy estimates represent the opportunitycost of pricing domestic energy below market levels.6. All emission reductions in this section are presented relative to the New Policies Scenario, unless indicatedotherwise.7. Energy efficiency-related savings in 2020 take account of the rebound effect, i.e. the effect of increased useof a product or facility as a result of efficiency-related operating costs savings or higher disposable income fromreduced energy expenditures. The rebound effect in the 4-for-2 °C Scenario is largely related to decreases inconsumer prices as GDP does not change in comparison to the New Policies Scenario, but is counterbalanced bythe assumed fossil-fuel subsidy phase-out that leads to increased energy conservation.50 World Energy Outlook | Special Report

The assumed policy measures go a long way toward closing the gap between expectedemissions levels in 2020 on the basis of present government intenons, as modelled in theNew Policies Scenario, and those required to achieve the 2 °C target (the 450 Scenario).They avoid 80% of the dierence in emissions levels. Nonetheless, a gap of around 770 Mtsll remains, indicang that yet more stringent measures will be required aer 2020 inorder ulmately to meet the 2 °C goal.Figure 2.2 Change in world energy-related CO 2 and CH 4 emissions bypolicy measure in the 4-for-2 °C Scenario1234Gt CO2-eq393735333129UpstreamCH4 reduconsPowergeneraonEnergyefficiencyFossil-fuelsubsidiesOther energyrelatedCH 4Upstream oil andgas CH 4 emissionsEnergy-related CO 2272010NPSPolicy measures20204-for-2 °C450S2020Notes: Methane emissions are converted to CO 2 -eq using a Global arming Potenal of 25. NPS NewPolicies Scenario 450S 450 Scenario.More than 70% of abatement occurs in non-OECD countries, where projected demand forenergy in 2020 is around 480 million tonnes of oil equivalent (Mtoe) (or 5%) lower thanin the New Policies Scenario (Figure 2.3). China alone is responsible for more than onequarterof the global emissions savings from these measures in 2020, resulng from thesignicant scope to reduce emissions that accompanies its rapidly rising energy demand,large potenal to further improve energy eciency and heavy reliance on coal-redpower generaon. The Middle East (9% share of savings in 2020) and India (9%) togetheraccount for almost one-h of the savings, driven primarily by fossil-fuel subsidy reformand reduced upstream methane emissions in the former and eciency improvements andchanges in the power generaon mix in the laer. Although energy eciency policy playsan important role in the Middle East too, it is the assumed enhanced phase-out of fossilfuelsubsidies that encourages its realisaon, as this reduces the payback period of moreecient technologies to the necessary extent to make eciency policy viable. 8 OECDcountries see a smaller share of the savings at below 30%, although the United States(13% share of savings in 2020) is the second-largest contributor to emissions reducons,© OECD/IEA, 20138. For example, given heavily subsidised low petrol prices in Saudi Arabia, the payback period for a car thatconsumes half as much fuel per 100 kilometres as todays average car is currently close to twenty years.Chapter 2 | Energy policies to keep the 2 °C target alive 51

aer China, and, together with the European Union (8%), accounts for around one-h ofthe global total. The larger share of savings in non-OECD countries is directly linked to thehigher growth in their energy demand 90% of global demand growth to 2020 in the NewPolicies Scenario. ith energy demand per capita today 70% below the OECD average,non-OECD countries have expectaons of higher growth to 2020 in both energy demandand emissions and associated scope for savings especially because populaon growth(90% of global growth to 2020) and economic growth (almost three-quarters of globalGDP growth unl 2020) are much stronger than in OECD countries.Box 2.2 The role of renewables in the 4-for-2 °C ScenarioIn many countries, renewables deployment is driven by government targets. Examplesinclude the targeted share of 20% in total energy demand by 2020 in the EuropeanUnion US state-level renewable porolio standards, covering 30 states and the Districtof Columbia exisng capacity targets by technology type in China, India and Brailand biofuels blending mandates in many countries. A wide variety of such policiesand mechanisms are in place today. All are taken into account in the New PoliciesScenario, the central scenario of WEO-2012. They include the enforcement and furtherstrengthening of these policies where governments have announced this intenon.Renewable energy accordingly plays an important role in all our scenarios, in parcularin power generaon. Though not characterised specically as one of the addionalpolicies of the 4-for-2 °C Scenario, the share of renewables in global power generaonincreases from 20% today to 27% in 2020. This is two percentage points above thelevel reached in the New Policies Scenario, due to the proposed policy to reduce theuse of inecient coal-red power generaon and lower electricity demand fromenergy eciency policies. In net terms, renewables meet about 60% of the increasein global electricity demand up to 2020 in the 4-for-2 °C Scenario, installed capacityreaching around 1 350 gigawas (G) of hydropower, 580 G of wind, 265 G of solarphotovoltaic, 135 G of biomass-red power plants and 35 G of other renewables.The 4-for-2 °C Scenario sees cumulave investment in renewables of $2.0 trillion up to2020, contribung to the reducon in renewable energy technology costs post-2020,thereby facilitang steeper emissions reducons then.© OECD/IEA, 2013Increasing deployment of renewables is supported by subsidies, which help overcomedeployment barriers. In power generaon, these subsidies are set to increase to$142 billion in 2020 in the 4-for-2 °C Scenario, up from $64 billion in 2011. This is 5%over the level reached in the New Policies Scenario in 2020 (due to lower wholesaleelectricity prices from lower internaonal fuel prices), but is oset by the widereconomic gains achieved from lower fossil-fuel prices. Biofuels (mostly supportedby blending mandates) received subsidies totalling $24 billion in 2011 these rise to$47 billion in 2020 in the 4-for-2 °C Scenario. The European Union, United States andChina account for the bulk of renewables subsidies, today (85%) and in 2020 (77%).52 World Energy Outlook | Special Report

Figure 2.3 Change in energy-related CO 2 and CH 4 emissions in selectedregions in the 4-for-2 °C Scenario relative to theNew Policies Scenario, 2020100%80%60%855 Mt 399 Mt 284 Mt 279 Mt 241 Mt 203 MtFossil-fuel subsidiesPower generaonEnergy efficiencyUpstream CH 4reducons123440%20%China United Middle India European RussiaStates EastUnionNote: Savings are allocated by enabling policy and total emissions are in CO 2 -eq.Abatement to 2020 by policy measureIn the 4-for-2 °C Scenario, energy eciency is the largest contributor to the reducon inglobal greenhouse-gas emissions, resulng in savings of 1.5 Gt CO 2 -eq in 2020, or almosthalf of the total abatement relave to the New Policies Scenario (Figure 2.4). As indicatedabove, while there is a ra of eciency policies capable of reducing energy consumponand therefore emissions, 9 we have focused on just four key measures on the basis that theycan be quickly implemented and that the mechanics of implementaon have already beendeveloped in numerous countries. The selected policies are applied to new equipment andtechnologies: they exclude the early rerement of exisng stock.ey energy eciency measures include:• More ecient heang and cooling systems in residenal and commercial buildingsthrough minimum energy performance standards (MEPS) for new equipment, andtechnology switching, such as through greater use of heat recovery and beer use ofautomaon and control systems.• More ecient appliances and lighng in residenal and commercial buildings.• Use of more ecient electric motor systems in industrial applicaons, such as pumping,compressing air, and other types of mechanical handling and processing.• Fuel-economy standards and fuel-economy labelling for new passenger light-dutyvehicles (PDs) and freight trucks in road transport.© OECD/IEA, 20139. There are already numerous energy efficiency policies in place in many countries an overview of key policiesby country and sector is available in the energy efficiency focus in WEO-2012 (IEA, 2012a). All figures hererepresent the additional gains resulting from the specified additional measures.Chapter 2 | Energy policies to keep the 2 °C target alive 53

Figure 2.4 Change in world CO 2 and CH 4 emissions in the 4-for-2 °CScenario by policy measure relative to the New PoliciesScenario, 2020Fossil-fuel subsidies12%Savings: 3.1 Gt CO 2 -eq6%Road transportUpstream CH 4reducons18%Efficiency49%14%14%Industrial motorsAppliances and lighngPower generaon21%15%Heang and coolingAmong these measures, those targeng heang and cooling, appliances, lighng andindustrial motors have a similar eect on reducing greenhouse-gas emissions, eachcontribung around 30% of the addional eciency-related savings. Policies targeng roadtransport make up a smaller share of abatement, partly because of the lead mes requiredfor more ecient vehicles to penetrate the vehicle stock and because the New PoliciesScenario takes into account the numerous policies already in place to improve eciency inroad transport (thus reducing the scope for further gains in the 4-for-2 °C Scenario).Figure 2.5 CO 2 and CH 4 policy in the 4-for-2 °C Scenario relative to the New PoliciesScenario, 2020ChinaUnited StatesEuropean UnionRoad transportIndustrial motorsAppliances andlighngHeangand coolingIndiaRussiaRest of world100 200 300 400 500 600 700Mt CO 2 -eq© OECD/IEA, 2013Almost 80% of energy eciency-related savings occurs in ve regions: China, the UnitedStates, the European Union, India and Russia (Figure 2.5). China sees by far the largestreducon in emissions through more ecient use of energy, at around 40% of the global54 World Energy Outlook | Special Report

total. Many of these savings are made in industry, at around 280 Mt CO 2 -eq in 2020, or44% of eciency-related savings, stemming from the use of more ecient industrial motorsystems. Industry in China currently accounts for about two-thirds of the countrys totalelectricity consumpon, of which it is esmated that 60-70% is used by electric motors(IEA, 2011b). hile China has already adopted MEPS for some motors, their typicaloperaonal eciency is 10-30% below the standard in internaonal best pracces(SwitchAsia, 2013). The vast majority of electricity savings from motor systems, however,comes from a combinaon of appropriate use of variable speed drives, proper motorsiing, prevenve maintenance and opmising the motor-driven equipment, as thenominal eciency of an electric motor can only be enhanced by around three percentagepoints for a medium-sied motor. In China, a combinaon of further ghtening ofMEPS, their wider adopon and, parcularly, the imposion of requirements for energymanagement systems could considerably reduce electricity demand and thereby emissionsfrom the currently carbon-intensive power generaon sector (Figure 2.6). India, too, hasconsiderable potenal for emissions reducons through the use of more ecient industrialmotors. At present, India has no MEPS for electric motors in industry and a highly carbonintensivepower mix. The adopon of such standards in India is assumed in the 4-for-2 °CScenario and lowers its emissions from the industry sector by about 55 Mt CO 2 -eq in 2020(almost 40% of the projected abatement related to energy eciency). hile MEPS are animportant instrument to encourage the use of more ecient industrial motor systems,there are barriers to their deployment, such as inadequate assessment of the actualservice required and the complexity of motor systems. However, much has already beendone to study policy opportunies and policy best pracces in this area, paving the way forthe swi and eecve introducon of this measure. 10 This results in widespread adoponof more ecient industrial motor systems in the 4-for-2 °C Scenario.1234Figure 2.6 policies in the 4-for-2 °C Scenario relative to the New PoliciesScenario, 2020ChinaUnitedStatesEuropeanUnionIndiaRussiaRest ofworldIndustrial motors-100-2%Appliances andlighngHeang andcoolingChange in electricitydemand (right axis)-200-4%-300-6%-400-8%-500-10%© OECD/IEA, 2013TWh-600-12%10. See for example IIP (2011) and IEA (2011b).Chapter 2 | Energy policies to keep the 2 °C target alive 55

Ecient use of energy in buildings, including energy used for heang, cooling, appliancesand lighng, has recently aracted considerable aenon as policies in place or underconsideraon tap only around one-h of the economic potenal (IEA, 2012a). In 2013,for example, the Major Economies Forum iniated a dialogue among its member countrieswith a view to their seng voluntary intensity targets for energy consumpon in buildings.In terms of heang and cooling, installing more ecient equipment (such as gas heangsystems, heat pumps and high eciency air-condioners) is one of the best means ofreducing emissions in the short term, although the potenal to improve the buildingenvelope is also vast (IEA, 2013b). Several countries have already adopted voluntaryprogrammes, e.g. India or Brail, or binding ones, such as the United States, to advanceuptake of more ecient equipment.In the 4-for-2 °C Scenario, China achieves 40% of the global emissions reducon relatedto more ecient heang and cooling systems. The high share reects the expected rapidincrease in projected demand in China for such services, parcularly for air-condioning,which means that the adopon of MEPS can signicantly curtail growth in energy demand.The United States and the European Union are together responsible for a further one-thirdof the global emissions reducons related to more ecient heang and cooling equipment.In both cases, the deployment of new higher eciency heang and cooling systemshas a signicant impact on emissions, bringing reducons of almost 80 Mt CO 2 -eq and65 Mt CO 2 -eq in 2020 for the United States and the European Union, respecvely.Just as for industrial motors, there are barriers to the use of more ecient heang andcooling systems. hile MEPS are an important means of achieving emissions reduconsin the 4-for-2 °C Scenario, they need to be accompanied by policies to ensure theirenforcement. Typical barriers include public acceptance and other general market risksof new technologies, and can be related to a shortage of skilled labour in some countries.Split incenves are another problem that needs careful aenon. 11 Providing informaonthrough awareness campaigns and training programmes can be helpful tools to overcomethese barriers.There is considerable scope in all regions to reduce emissions stemming from the use ofappliances and lighng. This is linked, in part, to their important share in overall electricitydemand today: lighng and appliances alone are responsible for 37% of electricity demandin OECD countries and 26% in non-OECD countries. Due to the relavely short operanglifespan of the equipment concerned, MEPS for appliances and lighng are parcularlyeecve and are already widely used in many countries. Most OECD countries haveadopted such standards for a wide range of products, as has China. Russia is phasing outincandescent light bulbs (100 was and above), while India is set to adopt mandatorystandards and labelling for room air condioners and refrigerators. At the Clean Energy© OECD/IEA, 201311. A split incentive refers to the potential difficulties in motivating one party to act in the best interests ofanother when they may have different goals andor different levels of information.56 World Energy Outlook | Special Report

Ministerial in New Delhi in April 2013, ministers highlighted the importance of the SuperecientEquipment and Appliance Deployment (SEAD) iniave as a means to progressquickly and cheaply towards a more sustainable future. 12In the 4-for-2 °C Scenario, the contribuon of appliances and lighng to addional energyeciency-related savings is parcularly large in the United States, at 44% in 2020. The bulkof these savings could be achieved by ghtening the MEPS that already exist. Appliancesand lighng are responsible for close to 40% of the eciency-related savings in India, a highshare that reects the current dearth of eciency standards. In absolute terms, the largestreducons are made in China (125 Mt CO 2 -eq), followed by the United States (around85 Mt) and the European Union (around 60 Mt), where we assume that the new EcoDesignDirecve that covers een product groups is further strengthened. Across all countries,there is sll considerable potenal to expand both the range of products that are coveredby MEPS and the stringency of the standards.1234Road transport, which is currently responsible for around 16% of CO 2 emissions fromthe energy sector, has received a lot of policy aenon in recent years, as high oil pricesand rising demand for mobility have strengthened the case for eciency improvements.Many governments have adopted fuel-economy policies in a bid to reduce the burden onconsumers and the cost of oil imports. PD standards have been adopted most widely,including in many of the major car markets in OECD countries (IEA, 2012b). OutsideOECD countries, only China has adopted such standards, though India plans to do so.Fuel-economy standards for trucks are also increasingly receiving the aenon of policymakers and have been adopted in several OECD countries. Though essenal to realisingfuel eciency in road transport, standards are and should be complemented by supporngpolicies to overcome the barriers associated with their deployment, such as informaongaps. 13In the 4-for-2 °C Scenario, the impact of ghter fuel-economy standards, i.e. beyond thoseimplemented in the New Policies Scenario, is moderate in the period to 2020, comparedwith the other energy eciency measures proposed. This reects the me it takes for thefull eect of fuel-economy standards for new vehicles to be felt across the enre eet. Bycontrast, they have a much greater impact aer 2020. The relavely limited impact alsoreects the fact that fuel-eciency regulaons are already in place in many of the majoreconomies. Nonetheless, fuel-eciency standards in road transport do play a signicantrole in the overall abatement: in Russia, they account for about 30% of the eciencyrelatedsavings, compared with the New Policies Scenario. In the 4-for-2 °C Scenario, theaverage tested fuel eciency of new PD sales in 2020 reaches around 40 miles per gallon(mpg) (or 5.9 litres per 100 kilometres l100km) in the United States 95 grammes of CO 2per kilometre (g CO 2 km) in Europe (or 3.8 l100km) 5.0 l100km in China and 4.8 l100kmin India. The global average is 5.1 l100km.© OECD/IEA, 201312. For more information, see .13. For an overview of suitable policy packages, see IEA (2012b).Chapter 2 | Energy policies to keep the 2 °C target alive 57

In the 4-for-2 °C Scenario, the use of the least ecient coal-red power plants is reduced,relave to the New Policies Scenario. e assume a ban is introduced prohibing theconstrucon of new subcrical coal-red power plants. Plants that have recently been builtor are already under construcon and have therefore yet to recover their investment cost,connue to operate, albeit at reduced levels. Those inecient plants that have alreadyrepaid their investment costs are either rered or idled. Possible levers to achieve thispolicy include:• Adopon of energy eciency or CO 2 emissions standards for coal-red power plants.• Adopon of air polluon standards.• Pricing the use of carbon, for example through an emissions trading scheme.• Assigning power producon limits for each generator to incenvise the use of themost ecient plants (typically in liberalised markets).• Allocaon of generaon slots, renewing (or not) operaonal licences or altering thedispatch schedule in favour of more ecient plants (typically in regulated markets).As a result of the proposed policy, the global installed capacity of subcrical power plantsin operaon decreases by more than one-fourth in 2020, or about 340 gigawas (G),compared with the New Policies Scenario (Figure 2.7). Exisng plants account for the vastmajority of this reducon: while 170 G of new subcrical plants are added in the NewPolicies Scenario by 2020, only about 50 G of them are not already under construconand therefore do not go ahead in the 4-for-2 °C Scenario. Of the 1 270 G of subcricalcoal-red power plants in 2020 in the New Policies Scenario, 290 G with the lowesteciencies are either rered or not used at all by 2020. A more complete phase-out of coalsubcrical plants by 2020 is unrealisc in most regions both because it would unacceptablyreduce the reliability of electricity supply and because of the costs involved.Figure 2.7 Change in subcritical coal electrical capacity in the NewPolicies and the 4-for-2 °C Scenarios, 2020GW1 4001 2001 000800600400200RerementsUnder construconOther required addionsAddions not requiredIdled plants© OECD/IEA, 20132011 2020 2020New Policies Scenario4-for-2 °C Scenario58 World Energy Outlook | Special Report

The extent of the potenal reducon in use of inecient coal plants by region isdetermined by two main factors: the extent of the reducon of electricity demand (whichis achieved through the proposed energy eciency measures in the 4-for-2 °C Scenario)and the extent of the opportunity to switch to other technologies. The switch in powergeneraon is mostly possible to gas-red power plants or more ecient coal plants upto 2020, as the addional reliance on nuclear power is constrained by long construconlead mes and the New Policies Scenario already embodies rapid growth in renewables,mainly driven by targets in many countries (Box 2.2). The decrease of electricity demandgenerally provides an opportunity to reduce the use of subcrical coal plants by at leastthe same amount.1234The possibility of switching to other, more ecient, technologies depends on severalfactors, which include the exisng capacity mix, the extent of the need for capacityaddions, the nature of the support schemes in place, the relave eciency of theplants available and the construcon periods for new plants. For example, in China andin the United States, the reducon of power generaon from inecient coal plants in the4-for-2 °C Scenario is greater than the reducon in electricity demand, due to the possibilityto switch to more ecient coal technologies and gas-red generaon (Figure 2.8). InEurope, on the other hand, the CO 2 price assumed in the New Policies Scenario alreadyprovides an incenve for higher-eciency power plants, which limits the scope foraddional producon from these plants in the 4-for-2 °C Scenario, with the result that thefall in the use of coal plants fails to keep pace with the reducon in electricity demand.Figure 2.8 the New Policies Scenario by selected regions, 2020TWh-200-400-600-800-1 000UnitedStates-12%EuropeanUnion-11%China India ASEAN-15%-13% -7%20%40%60%80%100%Lower electricitydemandLower electricitygeneraon from lessefficientcoal plantsShare of total coalgeneraon from lessefficientcoal plants(right axis)X% Percentage pointsdifference relave tothe share of lessefficientcoal in theNew Policies Scenario© OECD/IEA, 2013At a global level, the reduced use of subcrical coal plants combined with the greater useof more ecient coal plants increases the average eciency of global coal generaon by3.3 percentage points in 2020 in the 4-for-2 °C Scenario, relave to 2011. This is moreChapter 2 | Energy policies to keep the 2 °C target alive 59

than twice as high as the eciency gain achieved in the New Policies Scenario, where theaverage eciency of the coal power plant eet increases by 1.5 percentage points over thesame period.The reduced use of inecient coal-red power plants in the 4-for-2 °C Scenario cutsglobal CO 2 emissions by around 570 Mt in 2020, relave to the New Policies Scenario,as the average emissions intensity of power generaon is almost 10% lower, at about420 grammes of CO 2 per kilowa-hour (g CO 2 kh) (Figure 2.9). Methane emissions fromcoal mining, transport and use, at around 70 Mt CO 2 -eq, are also reduced as a result of thelower use of coal. In overall terms, the addional emissions savings are most pronouncedin countries which currently have low average power plant eciencies (such as India) or alarge coal power eet (such as the United States and China). They are the result of a sharpdrop in coal capacity ulisaon from 60% in 2011 to 54% in 2020, driven by a decline in theuse of subcrical coal plants from 59% in 2011 to 39% in 2020. There is no scope for furtherreducon while maintaining reliability of electricity supply.Figure 2.9 Average power generation emissions intensity andcorresponding CO 2 and CH 4 savings in the 4-for-2 °C Scenariorelative to the New Policies Scenario, 2020g CO2/kWh800600400800600400Mt CO2-eqSavings comparedwith New PoliciesScenario4-for-2 °C ScenarioCO 2 savings(right axis)200200UnitedStatesEuropeanUnionChina India World© OECD/IEA, 2013Relave to the New Policies Scenario, almost 30% of global CO 2 and CH 4 emissions savingsresulng from reduced use of inecient coal plants occurs in China. China is increasinglysuering from the impact of local air polluon, partly caused by the substanal use ofcoal in power generaon. According to recent analysis by the Chinese Academy forEnvironmental Planning (CAEP), the associated societal cost of environmental degradaon,including health-related damage, amounted to the equivalent of 3.5% of GDP in 2010. In anaempt to improve the eciency of its power sector, China phased out over 70 G of small,inecient coal-red power capacity between 2006 and 2010 as part of its 11 th Five-earPlan. China has also tested further policy opons for reducing emissions of air pollutantsfrom coal power staons, including through the Energy Saving Dispatch Policy (ESDP) that60 World Energy Outlook | Special Report

was tested in ve provinces in 2007 and 2008 (and that could help achieve the projectedreducon in CO 2 emissions seen in the 4-for-2 °C Scenario). In China, power dispatch usuallyworks according to predened quotas allocated to generators by provincial governments,with generators receiving a xed price for their power output and, in some cases, free-totradequotas to opmise the generaon paern. The ESDP sought to maximise the overalleciency of fossil fuel-based power plants by allocang higher quotas to the most ecientunits, without changing the compensaon to power generators. The pilot phase raiseda number of problems such as challenges to system reliability and was seen as onlya temporary device before the eventual transion to a fully market-based power systemas envisaged by the central government. But the scheme demonstrated how one policyto reduce the use of the least-ecient coal power staons can work in China. In addionto reducing growth in CO 2 emissions, the 4-for-2 °C Scenario also sees an improvement inlocal air quality in China: sulphur dioxide (SO 2 ) emissions from the use of coal in powergeneraon are 9% lower than in the New Policies Scenario by 2020, nitrogen oxides (NO x )emissions are 8% lower and parculate maer (PM 2.5 ) emissions are 3% lower.1234Almost one-quarter of the global reducon in CO 2 and CH 4 emissions from reducing theuse of the least-ecient coal power staons in the 4-for-2 °C Scenario occurs in the UnitedStates. Following a US Supreme Court ruling in 2007 that classied greenhouse gases aspollutants, the US Environmental Protecon Agency (EPA) determined that climate changeendangers public health and welfare, and that CO 2 and other greenhouse gases contributeto this endangerment. This nding established the authority of the US EPA to regulate CO 2emissions (including from power plants) under the Clean Air Act. The US EPA proposed acarbon polluon standard for new power plants in March 2012, which, if adopted, wouldeecvely prevent the construcon of new coal power plants without carbon capture andstorage (CCS). Addionally, the US EPA has the authority to propose performance standardsfor exisng fossil fuel-red power plants, which are responsible for about 33% of totalenergy-related CO 2 emissions in the United States, though there are no ocial plans to doso currently and would likely only follow the nalisaon of standards for new power plants.The Clean Air Act appears to allow the US EPA considerable exibility in applying standardsto exisng sources, such as allowing facilies that emit less than the standard to generatecredits that can be sold to higher-eming facilies. hile any such standard is likely to facesignicant opposion from some electric power producers, its applicaon would open theway to realising the reducons envisaged in the 4-for-2 °C Scenario and help natural gas,despite increasing prices towards 2020, to maintain the market posion that it gained inthe power sector in 2012, relave to coal, as a result of low gas prices.© OECD/IEA, 2013India sees the third-largest reducon in emissions from coal-red power generaon as aresult of the assumed coal power plant restricons in the 4-for-2 °C Scenario. Despite therecent construcon of more ecient coal capacity under the Ultra Mega Power Projects(UMPP) policy, India sll has one of the lowest average conversion eciencies in coal-redpower generaon in the world, esmated at just 28%, or eleven percentage points belowthe global average. This is linked to the average age of the coal-red power plant eet andChapter 2 | Energy policies to keep the 2 °C target alive 61

the relavely poor quality of domesc coal, which has an ash content of up to 60%. hileincreased coal washing at mining complexes is a possibility (which would also help alleviatetransportaon bolenecks by reducing the amount of coal transported), plant managersare oen reluctant to aempt to change coal quality due to concerns about operaonalproblems.The low average conversion eciency of coal in India is exacerbang local concerns that airpolluon is increasingly causing health problems and having adverse economic eects. Arecent study esmated the extent of health eects at 80 000 to 115 000 premature deathsin 20112012, at an economic cost of $3.3-4.6 billion (Goenka and Gukunda, 2013).India currently does not have strict standards for pollutants from power plants, except forparculate maer, and is suering from peak shortages which make it dicult to imposeaddional constraints on power plant operaon and dispatch. A new Naonal Missionon Clean Coal Technologies is under discussion, whose task would be to foster work onintegrated gasicaon combined-cycle and advanced ultra supercrical technologies, aswell as CCS. ith air polluon concerns growing, interest in clean coal technologies andminimum conversion standards might increase, with spin-o benets for the climate: inthe 4-for-2 °C Scenario in India, SO 2 emissions from the use of coal in power generaonare 14% lower than in the New Policies Scenario by 2020, NO x emissions are 8% lower andPM 2.5 emissions are 3% lower.Emissions savings in the European Union due to the reduced use of the least-ecientcoal power plants are the fourth-largest globally by 2020 in the 4-for-2 °C Scenario. Thereare several measures readily available with which to implement this policy. They include,parcularly, the EU Emissions Trading System (ETS), although the level of CO 2 prices underthe EU ETS is currently too low to incenvise a shi away from the least-ecient coal powerplants, parcularly given the current low price of coal relave to natural gas. Measureswould be needed to ensure a level of CO 2 prices sucient to facilitate the switch. Thearge Combuson Plants Direcve, established in 2001, limits operang hours of thermalpower plants that exceed specied emissions levels for SO 2 , NO x and dust, is another toolthat could be used to implement the policy. Demand-side measures tempering electricitydemand growth, as included in the Energy Eciency Direcve, can support the reduced useof the least-ecient coal plants.Energy-related methane emissions stem from the producon, transportaon, distribuonand use of all fossil fuels and from biomass combuson. e esmate that such emissionscurrently amount to 125 Mt CH 4 per year. Using the standard 100-years GP of 25 fromthe Intergovernmental Panel on Climate Change (IPCC), this amounts to 3.1 Gt CO 2 -eq. 14 Itshould be stressed however that there is a shortage of hard, measured, data on methane© OECD/IEA, 201314. Considering shorter time periods than 100 years, the CO 2 -equivalent emissions are even larger, given thatthe 20-years GP of the IPCC is 72, which increases the need to address CH 4 emissions in the short term. Seealso Alvare, et al. (2012) for a discussion of the choice of GP.62 World Energy Outlook | Special Report

emissions esmates rely primarily on mulplying emissions factors for various acviesby acvity levels the emissions factors themselves can be traced to studies made by theGas Research Instute and US EPA in the United States (US EPA, 2013).In the oil and gas industry, methane emissions occur across the enre value chain. 15Transmission and distribuon of natural gas releases considerable amounts of methaneinto the atmosphere due to leakage or venng (which may be voluntary or involuntary),parcularly in countries with a large and ageing distribuon network, such as Russia andthe United States. Addional methane emissions occur during incomplete combuson,both in end-use and in aring. The extent of emissions in transmission, distribuon andend-use is poorly known, as many of these emissions result from unintended leaks in ageinginfrastructure. Addressing such leakage is a challenging and potenally costly task, beyondthe short-term focus considered here. The larger potenal for reducing methane emissionsfrom oil and gas in the short term lies in opmising operaonal pracces upstream, wherethe sources of emissions are relavely well-known. Technologies to reduce them areavailable (in large part through the work of the US EPA Gas Star Program) and the necessaryacon can be implemented through the exisng sophiscated industry, dominated by largecompanies with strong technical skills and budgets. e esmate that the global oil and gasupstream industry released 45 Mt of CH 4 emissions (1 115 Mt CO 2 -eq) to the atmospherein 2010 (Spotlight).1234Both venng and aring give rise to methane emissions during oil and gas eld operaons.enng (as dened here) includes both the intenonal release of methane to theatmosphere (as part of normal operaons) and fugive emissions, which are unintended the results of leaks, incidents, or ageing or poorly maintained equipment. Someemissions from venng can be reduced at comparavely low cost by applying operaonalbest pracces, such as increased inspecon and repairs, minimising emissions duringcompleon operaons and workovers 16 , and reducing the frequency of start-ups andblow-downs. Equipment can also be converted, or designed, to reduce emissions: low-costopons include modifying dehydrators and converng gas-driven pumps and gas pneumacdevice controls to mechanical controls. Addional but more capital-intensive potenallies, for example, in replacing leaking compressors with new ones and installing vapourrecovery units on tanks. Producon of unconvenonal gas has been parcularly cricisedbecause of the large amount of methane that can be released to the atmosphere duringthe owback phase aer hydraulic fracturing. Controlling such emissions is part of the IEAGolden Rules for unconvenonal gas development, and such rules are being adoptedin a growing number of countries, for example in the US EPAs New Source PerformanceStandards for the oil and gas industry in the United States (IEA, 2012c).© OECD/IEA, 201315. Research efforts are underway, including at the University of Texas at Austin and the Environmental DefenseFund, to study methane emissions at each process step of the oil and gas value chain.16. orkover is the term used for maintenance operations requiring interventions inside an oil or gas well,requiring temporary interruption of production. Depending on the sequence of operations, small volumes of gasmay be released to the atmosphere.Chapter 2 | Energy policies to keep the 2 °C target alive 63

S P O T L I G H T© OECD/IEA, 2013How large are methane emissions from upstream oil and gas?Though data on methane emissions are generally poor, it is esmated that about550 Mt of methane emissions in total are released into the atmosphere every year(IPCC, 2007), of which around 350 Mt come from anthropogenic sources. In the WEO, we esmated total energy-relatedmethane emissions to be 125 Mt, of which 90 Mt come from the oil and gas supplyand distribuon (IEA, 2012c). The US EPA has recently published a comparable globalassessment of 129 Mt, with the contribuon of oil and gas supply and distribuon at80 Mt in 2010 (US EPA, 2012a). Much more work is needed fully to understand themagnitude of methane emissions in the absence of widespread detailed measurements.There is no global database available that distinguishes methane emissions fromupstream oil and gas field operations from those that occur during the processing,transmission and distribution of gas. For the purpose of this , wetherefore conducted a detailed assessment of methane emissions from oil and gasfield operations. During oil field operations, methane emissions occur either fromincomplete combustion in flaring (where associated gas cannot be brought to themarket due to the remoteness of the oil fields and a lack of infrastructure, such asin Russia, the Middle East, Africa and the Caspian Region) or as a result of leakageduring associated gas handling processes and (predominantly) venting at hydrocarbonstorage tanks, compressors or pneumatic devices. During field operations dedicatedto natural gas production, CH 4 emissions occur mostly from venting during normaloperations of drilling and well completion, and unloading (including flowback afterhydraulic fracturing), but also from condensate tanks, pneumatic devices, compressorsand dehydrators.For the analysis of the volume of gas flared during oil field operations, we used thesatellite data made available through the Global Gas Flaring Reduction Partnership ofthe orld Bank to estimate the amount of methane which might remain unburned,based on an assessment of regional practices. For the analysis of methane emissionsfrom venting during other oil and gas field operations, we used a detailed bottomupanalysis by the US EPA that assessed US methane emissions by process stepand equipment type as a basis for our global assessment (US EPA, 2013). Using thisanalysis as a starting point, together with production levels by region, we analysedcountry-specific field operation practices according to the type of development(unconventionalconventional and onshoreoffshore), by region and by type ofhydrocarbon, taking into account the average age of existing oil and gas fields, theregulatory environment and the availability of technology. This enabled us to derive aglobal assessment of total methane emissions from oil and gas field operations, whichare assessed as 45 Mt CH 4 (1 115 Mt CO 2 -eq) in 2010. Of this, 17 Mt CH 4 comes fromgas fields and 27 Mt CH 4 from oil field operations. Of the latter, 3.2 Mt are releasedas a consequence of incomplete combustion during gas flaring. Unsurprisingly, thelargest emitters are the regions with high oil and gas production levels, i.e. Russia(10 Mt) followed by the Middle East (9 Mt) and Africa (5 Mt).64 World Energy Outlook | Special Report

Gas that becomes available in relavely large quanes as a by-product of oil producon(associated gas) oen has no commercially viable outlet. It will not normally be vented,for safety reasons, but will be ared. Gas aring converts methane into CO 2 , i.e. sll agreenhouse gas, but with lower Global arming Potenal. Reducing aring has been along-standing goal of the internaonal community it would substanally reduce both CO 2and methane emissions but the large investment required cannot materialise quickly.On the other hand, combuson is not always fully complete, which means that unburnedmethane is inadvertently released to the atmosphere from an otherwise controlled process,the amount varying with the design of the aring equipment, and other parameters, suchas wind speed (US EPA, 2012b). To reduce aring on a large scale, infrastructure andequipment, such as compressors and pipelines, need to be built to bring the gas to marketsor to enable it to be used for local power generaon a comparavely capital-intensiveprocess. ess capital-intensive opons, such as the opmisaon of aring equipment or gasre-injecon, need to be promoted in the short term.1234All the technologies to pursue short-term opmisaon from upstream operaons inorder to reduce methane emissions from venng and aring are readily available, whichmeans that the pace of reducon can be signicant if the right policies and enforcementprocedures are adopted (Figure 2.10). In the 4-for-2 °C Scenario, such short-term policies,including reducing venng and improving aring eciency, reduce methane emissions fromoil eld operaons by about 300 Mt CO 2 -eq. in 2020 (or 40% of oil-supply related methaneemissions), relave to the New Policies Scenario, in which no addional regulaon toaddress venng and aring is assumed beyond that in place today, such as those targenggreen compleon equipment in the US EPAs New Source Performance Standards for theoil and gas industry. For gas eld operaons, the decrease is 280 Mt CO 2 -eq (or about 55% ofgas supply-related methane releases). The largest reducons are in Russia, the Middle East,Africa and the United States. They are achieved through a combinaon of rapid and broadbasedimplementaon of low-cost and technological best operaonal pracces, e.g. fewerstart-upsshutdowns, more frequent inspecons, installaon of electronic are ignion,replacement of pneumac controls by mechanical ones and upgraded dehydrators.These measures would account for about half of the reducon in emissions in 2020. Theremainder would be accounted for by the rst results from reducon endeavours thatare more complex, take more me to implement and require larger investments. Thiscategory includes modicaons like the installaon of pressurised storage tanks withvapour recovery units, replacing compressors by ones with higher emissions standardsand capturing emissions from individual wells. The impact of these measures would beeven larger beyond 2020, as methane emissions from the upstream are likely to connueto increase in line with increasing oil and gas producon.© OECD/IEA, 2013Regulaons exist in many countries to reduce venng and aring, for example, in Russia,Ukraine, Argenna and Colombia. But there is oen a lack of means of enforcement,parcularly for venng. hile the extent to which gas is ared is visible, vents are invisibleand eecve enforcement demands installaon of specic equipment (for example,infrared cameras) and carrying out specic measurements. These equipment or processesChapter 2 | Energy policies to keep the 2 °C target alive 65

are oen unavailable. Another essenal ingredient for success is raising awareness.Operators themselves, in parcular in dispersed operaons, are oen unaware of theextent of their emissions and lack appropriate detecon and measurement equipment.In relaon to the reducon of venng, at least, this points to an inial focus on large,concentrated operaons. A number of related eorts are currently underway, including theGlobal Methane Iniave and the US Natural Gas STAR Program. Supplementary oponsinclude extending carbon tax or trading schemes to methane, and imposing mandatoryrequirements to implement appropriate methane emissions control technologies andadopt best pracces.Figure 2.10 Methane emissions from upstream oil and gas by scenario, 2020Mt CO2-eq350300250Addional emissions inNew Policies Scenario4-for-2 °C Scenario201020015010050UnitedStatesOtherOECDMiddleEastRussiaAfricaOthernon-OECDEsmates from the WEO-2012 suggest that fossil-fuel consumpon subsidies worldwideamounted to $523 billion in 2011, up almost 30% on 2010 and six mes higher than thenancial support given to renewables (IEA, 2012a). 17 Fossil-fuel subsidies were mostprevalent in the Middle East, at around 40% of the global total. These esmates indicatethe extent to which end-user prices are reduced below those that would prevail in an openand compeve market. Such subsidisaon occurs when energy is imported at world pricesand sold domescally at lower, regulated prices, or, in the case of countries that are netexporters of a product, where domesc energy is priced below internaonal market levels.© OECD/IEA, 2013In recognion that subsidy reform is likely to be a challenging and slow process in manycountries because of polical obstacles, the 4-for-2 °C Scenario does not encourage highexpectaons for a universal phase-out in the short term. A total phase-out by 2020 isassumed in fossil-fuel imporng countries, as in the New Policies Scenario but in exporngcountries (where sustained reforms are likely to be more dicult), we assume a moregradual phase-out: relave to the New Policies Scenario, subsidisaon rates are reduced17. See IMF (2013) for additional discussion of subsidies.66 World Energy Outlook | Special Report

y an addional 25% by 2020, before being completely removed by 2035. 18 As a result ofthese eorts, CO 2 emissions are reduced by 360 Mt in 2020, relave to the New PoliciesScenario (Figure 2.11). Savings are greatest in the countries of the Middle East, whichaccount for 54% of all savings, followed by Africa at 15%, and an America at 11%. Besidesthe cauous approach adopted towards subsidy reform in the 4-for-2 °C Scenario, the factthat these savings come on top of those already achieved in the New Policies Scenarioexplains the relavely low share of abatement resulng from fossil-fuel subsidy reform,compared with the eect of the other policies adopted in the 4-for-2 °C Scenario.1234Figure 2.11 Change in world CO 2 emissions through fossil-fuel subsidyreform in the 4-for-2 °C Scenario relative to the New PoliciesScenario, 2020Savings: 360 MtOthernon-OECD13%Russia7%Lan America11%Middle East54%Africa15%Subsidy reform is dicult as the short-term costs imposed on certain groups of society canbe very burdensome and induce erce polical opposion. In Indonesia, for example, anaempt to increase gasoline and diesel prices by 33% in April 2012 induced strong publicprotests. Similarly, several weeks of naon-wide protests followed the complete removalof gasoline subsidies in Nigeria in January 2012. Concerns about inaon in several othercountries in Asia and polical and social unrest in parts of the Middle East and North Africahave delayed, and in some cases reversed, plans to reform energy pricing. Nonetheless,fossil-fuel subsidies represent a signicant burden on many naonal budgets and policalsupport for fossil-fuel subsidy reform has been building in recent years. In net-imporngcountries, in parcular, eorts to reform have been closely linked to the unsustainablenaonal nancial burden created by the growth of subsidies as import prices rise. Evensome net-exporng countries have taken steps to curtail the eect of arcially lowdomesc prices on export availability and foreign currency earnings (Table 2.2).© OECD/IEA, 201318. Subsidisation rate is calculated as the difference between the full cost of supply and the end-user price,expressed as a proportion of the full cost of supply.Chapter 2 | Energy policies to keep the 2 °C target alive 67

Table 2.2 Recent developments in fossil-fuel consumption subsidy policiesin selected countries© OECD/IEA, 2013CountryBoliviaChinaEgyptGhanaIndiaIndonesiaIranJordanMalaysiaMexicoRecent developmentsIn January 2012, the government returned to the issue of phasing out subsidies for gasolineand diesel, aer eorts in 2011 failed in the face of strong opposion.Implemented a ered electricity pricing system in July 2012. Announced in March 2013 thatprices of oil products would be adjusted every ten working days to beer reect changes inthe global oil market.Announced in August 2012 a commitment to gradually phase out subsidies to energyintensiveindustries. Plans to implement a smart card system to manage sales of subsidisedgasoline and liqueed petroleum gas (PG).Cut fuel subsidies in February 2013. As a result, prices of premium gasoline, diesel prices,kerosene, heavy fuel oil and PG increased.In January 2013, allowed state fuel retailers to start increasing the price of diesel on a monthlybasis unl it reaches market levels and raised the price cap on PG cylinders. The 2013-2014budget for petroleum product subsidies has been cut by more than 32%, compared to theprevious year, from Rs 969 billion to Rs 650 billion (approximately $12 billion).Announced policies to reduce subsidy expenditure in May 2012: tracking fuel use byvehicle banning state-owned and (certain) company vehicles from using subsidised fuelssubstung natural gas for gasoline and diesel and reducing electricity use in state-ownedbuildings and street lighng.Signicantly reduced energy subsidies in December 2010 as part of a ve-year programmeto gradually increase prices of oil products, natural gas and electricity to full cost prices. InJanuary 2013, ended supplies of subsidised gasoline for cars with engines of 1 800 cubiccenmetres and above, and restricted sales of subsidised gasoline near border areas.Raised the price of gasoline and electricity taris for selected industrial and services subsectorsin June 2012. Since November 2012, subsidies have been removed from all fuelsexcept PG and global oil prices have been reected via a monthly review.In April 2012, announced that subsidies for gasoline, diesel and cooking gas would connueto be provided under the current administraon.Gasoline and diesel prices are being raised slightly every month in 2013 to bring them closerto internaonal levels.Morocco In June 2012, raised the price of gasoline by 20% and diesel by 10%.NigeriaA naon-wide strike followed a complete removal of gasoline subsidies in early January2012, which doubled prices. Gasoline prices were then cut by a third, parally reinstungthe subsidy. Announced in March 2013 that there were no plans to reduce subsidies onpremium gasoline.Russia Plans to increase regulated domesc natural gas taris by 15% for all users from July 2013.SaudiArabiaSouthAfricaIn May 2013, the Economy and Planning Minister indicated that subsidy raonalisaon wassomething the country is seeking to address as they have become expensive and are causingdamage to the economy.Energy regulator granted power ulity Eskom an 8% per year average electricity priceincrease over the next ve years, which will eecvely reduce electricity subsidies.Sudan Commenced a subsidy reducon programme in June 2012, but in December 2012,announced that there were no plans to cut fuel subsidies further in 2013.ThailandIn early 2013, announced that PG prices would be increased monthly by 50 satang(approximately $0.02) monthly over the next year.68 World Energy Outlook | Special Report

Because of the social sensivity of the issue (and because every country must consider itsspecic circumstances), there is a ra of key principles to be adhered to when implemenngsuch reforms. For example, inadequate informaon about exisng subsidies is frequentlyan impediment. Before taking a decision about reform, governments must rst preciselyexamine energy subsidies, including their beneciaries, to idenfy low-income groupsthat depend on subsidies for access to basic energy services, and quanfy their costs andbenets, in order to determine which subsidies are most wasteful or inecient. Makingmore informaon available to the general public, parcularly about the budgetary burdenof subsidies, is a necessary step in building support for reform.1234hile the removal of fossil-fuel subsidies tends to improve long-term economic compevenessand scal balances, it may, nonetheless, have negave economic consequences in the shortterm, parcularly for certain groups, and any such reform must be carried out in a way thatallows both energy and other industries me to adjust. Governments may well be wise todissociate themselves from direct responsibility for price-seng, either by liberalising energymarkets, or, at least, by establishing automac mechanisms for price changes.Box 2.3 Sustainable Energy for All and the 4-for-2 °C ScenarioProviding access to modern energy oers mulple economic and social benets. et,today 1.3 billion people do not have access to electricity and 2.6 billion people relyon the tradional use of biomass for cooking (IEA, 2012a). These people mainly live inrural areas in developing Asia and sub-Saharan Africa. The United Naons SustainableEnergy for All (SE4All) iniave addresses this urgent problem, but investments undercurrent and planned policies will not be enough to achieve universal energy access by2030 (IEA, 2011a).The SE4All iniave sets specic targets for reaching the goal of universal access tomodern energy services by 2030, including reducing energy intensity at an averageannual rate of 2.6% between 2010 and 2030, and increasing the share of renewables.The policy measures proposed in the 4-for-2 °C Scenario allow important steps tobe taken towards these goals: in parcular, the energy intensity target is more thanreached as a result of the proposed policy package. The proposed ban on the leastecientcoal power plants helps to increase the share of renewables, but the levelreached in 2030 is sll short of the SE4All target.© OECD/IEA, 2013Reducing methane emissions from upstream oil and gas is not part of the SE4Alliniave, but it could also support the achievement of universal access in countrieswith considerable aring. Nigeria, for example, had the third-largest populaon in theworld without access to electricity in 2010, around 79 million people or half of thetotal populaon. The country already makes steps towards reducing aring in oil andgas producon, and full implementaon of such measures as in the 4-for-2 °C Scenariocould save natural gas at a level that, if supplied to the domesc market, would besucient to provide basic energy needs to the currently deprived.Chapter 2 | Energy policies to keep the 2 °C target alive 69

Even a commitment to subsidy reform will not be sucient in the absence of certaininstuonal and administrave capabilies, and even physical infrastructure. There mustbe instuons that are capable of accurate and mely collecon of data about exisngsubsidies, their distribuon and the need for oseng selecve relief accompanyingreform. Governments ulmately have the responsibility for gathering this far-reachinginformaon, but other organisaons may have the technical experse necessary to aid theeort. 19Implicaons for the global economyThe policy package suggested in the 4-for-2 °C Scenario does not aect global and regionalgrowth of GDP to 2020. GDP grows globally at 4.1% per year between 2012 and 2020,represenng annual average growth of 2.2% and 6.0% in OECD and non-OECD countriesrespecvely. 20 This neutral impact on GDP results from the combined implementaonof the four policies that are assumed to be adopted and from relave price adjustmentsacross all commodies, goods and services. In the period post-2020, however, the adoptedpolicy measures foster economic growth, as investments in the programme are increasinglyoutweighed by fuel bill savings and resources get allocated more eciently across theenre economy. 21Energy prices in the 4-for-2 °C Scenario are lower than in the New Policies Scenario: oilprices increase to $116 per barrel in 2020, or $4barrel lower than in the New PoliciesScenario, before declining in 2035 to $109barrel, which is $16barrel lower than in theNew Policies Scenario. 22 Natural gas prices are lower in imporng regions such as Europe orJapan. OECD steam coal prices reach $100tonne in 2035, $15tonne lower than in the NewPolicies Scenario. The acvity level of each sector in each country is boosted or reduced,depending on the specic policies to which they are exposed (Figure 2.12).© OECD/IEA, 201319. The World Energy Outlook 2013 to be released on 12 November 2013 will examine the extent of fossilfuelsubsidies globally.20. Slight deviations from New Policies Scenario GDP levels lie within the margin of error of standard mid-termeconomic forecasting, particularly in times of high uncertainty on projected economic activity (OECD, 2012 IMF,2012).21. Some of the proposed policy measures, such as energy efficiency, can foster economic growth evenbefore-2020. See IEA (2012a) for a discussion of economic benefits of energy efficiency policy.22. Each policy pillar may impact energy prices. For example, the multilateral and progressive phase out offossil-fuel subsidies tends to push international fossil-fuel prices down, but domestic end-user prices increasein countries conducting the reform. Targeted energy efficiency measures also put a downward pressure oninternational prices by lowering energy demand. In contrast, minimising upstream methane emissions andreducing power generation from inefficient coal-fired power plants tend to increase production costs, thusleading to higher end-user prices of oil, natural gas and electricity.70 World Energy Outlook | Special Report

Figure 2.12 Average annual GDP growth by scenario in selected countries,2012-202018%6%4%4-for-2 °CScenarioNew PoliciesScenario2342%2.7%UnitedStates1.5%EuropeanUnion1.4%3.7%7.7%7.4%4.0%4.9%Japan Russia China India Brazil Africa4.1%MiddleEastThe value of economic acvity in energy sub-sectors (comprising fossil-fuel extraconand processing, transport fuel producon and shipping, and power generaon) is slightlyreduced, as they bear extra costs due to the reducon in methane emissions in upstreamoil and gas operaons and lower use of subcrical coal-red power plants. By 2020, theglobal reducon in acvity in the energy sector, measured by real value-added, reachesabout $150 billion, a 3.7% decline relave to the New Policies Scenario. By contrast,other sectors of the economy benet from lower energy prices and, in some cases, fromaddional investments linked to the adopon of more energy-ecient technologies thatbring about savings in fuel costs. These variaons in sectoral acvity level oset each other,resulng in overall GDP-neutrality.Energy eciency measures adopted in the 4-for-2 °C Scenario bring about a $900 billionincrease, relave to the New Policies Scenario, in cumulave investment from 2012 to 2020(Figure 2.13). More than half of the increase is due to households purchasing more ecientenergy consuming equipment (IEA, 2012a). The increase in cumulave investment in theservice and transport sectors respecvely reaches more than $160 billion and $170 billion.Energy intensive industries are responsible for only a small share of total energy eciencyinvestments, as their potenal for energy savings is comparavely limited in the period to2020. The reducon of methane emissions from upstream oil and gas requires a cumulaveinvestment of around $20 billion up to 2020, while power generaon investment is slightlyreduced, relave to the New Policies Scenario, due to lower electricity demand, driven byenergy eciency policies. 23© OECD/IEA, 201323. The implementation of the four policies will require transfer of technology from developed to developingcountries. One such mechanism, Joint Crediting Mechanism (JCM), is being established by the JapaneseGovernment. Through this mechanism a host country receives technology and sets up measuring, reporting,and verification of a projects emissions reductions. Projects include renewable energy, highly efficient powergeneration, home electronics, etc., which facilitates low-carbon growth in developing countries.Chapter 2 | Energy policies to keep the 2 °C target alive 71

Figure 2.13 non-energy value-added in the 4-for-2 °C Scenario relative tothe New Policies ScenarioBillion dollars (2011)300200100Value-added inenergy sectorsValue-added innon-energy sectorsEnergy efficiencyinvestments0-100-2002010 2012 2014 2016 2018 2020Capital-intensive sectors, facing sieable fuel spending, benet the most from the policiesthat are assumed to be adopted in the 4-for-2 °C Scenario (Figure 2.14). Transport services(including freight and shipping of other goods) are directly smulated by targeted energyeciency investments. Cumulave value-added to 2020 increases by 7%, relave tocurrent levels, and 0.7% relave to the New Policies Scenario. Despite limited investmentsin energy eciency, energy-intensive industries are parcularly sensive to the reducedenergy prices stemming from the full policy package. This enables those industries toredirect spending to other primary factors, e.g. capital and labour, which translates into anincrease in acvity of around 1-3% through to 2020 (Chateau and Magné, 2013).Figure 2.14 Change in household spending, investment in energyScenario relative to the New Policies Scenario, 2012-2020© OECD/IEA, 2013Household spendingServicesTransport servicesTransport equipmentManufactured goodsChemicalsIron and steelCementOther2% 4% 6% 8% 10%100 200 300 400 500Billion dollars (2011)CumulavespendingCumulaveinvestmentsChange in cumulavevalue-added relaveto 2012 levels(top axis)72 World Energy Outlook | Special Report

The manufacturing sector also sees reduced producon costs and a 4% increase incumulave acvity. In relave terms, the policy impact for the services sector is limited.Given the sheer sie of services in the global economy currently around 60% of totalvalue-added the amount of capital invested in energy eciency relave to capital inplace is only a few percentage points. In addion, energy use is in the services sector is toolimited to benet signicantly from reduced energy prices in the 4-for-2 °C Scenario.The overall objecve of reducing CO 2 and methane emissions entails sectoral and regionalreallocaons of supply and demand across all commodies, goods and services. Energyeciency investments by households and rms reduce their energy bills, freeing up nancefor the purchase of other goods. Prices of non-energy goods and services are moderated,as energy costs are lower. This smulates an increase in acvity in non-energy sectors thatmore than compensates for the reducons in the energy sector. The global trade impactsof the policies remain very limited a mere 0.1% increase in 2020.1234Figure 2.15 Impact on consumption of goods and services in households inthe 4-for-2 °C relative to the New Policies Scenario, 2020OECDElectricityOil productsNatural gasServicesTransport*-3.8%-2.3%-4.2%Manufactured goodsOtherNet effect+0.1%Non-OECDElectricityOil productsNatural gasServicesTransport*Manufactured goods-3.9%-3.2%-8.1%OtherNet Effect+0.4% Includes transport equipment and transport services.-80 -60 -40 -20 0 20 40 60 80Billion dollars (2011)© OECD/IEA, 2013The reshuing of sectoral acvity is chiey triggered by the altered consumponbehaviour of households, a disnct driver of economic growth, parcularly in OECDcountries (Figure 2.15). Goods and services with relavely low energy content or whoseadopon may bring about signicant energy savings, such as in transport through thedeployment of energy-ecient vehicles, are specically targeted by the policy packageimplemented in OECD countries. In 2020, the four policies in OECD countries lead to anChapter 2 | Energy policies to keep the 2 °C target alive 73

increase in household expenses of above 1% for transport services and equipment, andalso for manufactured products relave to the New Policies Scenario. 24 Both categories ofaddional expense are of similar magnitude, around $35 billion.Energy expenses in OECD countries are between 2% and 4% lower than in the New PoliciesScenario, equivalent to a net reducon of about $40 billion. The net increase in OECDhousehold consumpon is limited to 0.1%. Similar net deviaons are observed in non-OECDcountries, though non-OECD economies are generally more industry-oriented and energyspending accounts generally for a larger share in consumpon. Energy eciency measuresredirect consumpon towards goods and services which embed less energy. Therefore,the set of policies in the 4-for-2 °C Scenario induces a more signicant boost in householdconsumpon. The services sector is further developed, as economic development proceedsin these countries. In 2020, household spending on energy goods is cut by almost 4% inthe case of oil products. The electricity bill diminishes by more than 8%, incenvised by thereform of fossil-fuel subsidies.Figure 2.16 4-for-2 °C relative to the New Policies Scenario, 2020OECDElectricityOil productsNatural gasOtherServicesTransport*-1.5%-2.1%-4.7%Energy-intensive goodsNet effect-0.2%Non-OECDElectricityOil productsNatural gasOtherServicesTransport*Energy-intensive goodsNet effect-4.2%-2.7%-0.2%-5.5%-140 -120 -100 -80 -60 -40 -20 0Billion dollars (2011) Includes transport equipment and transport services.The overall increase in household consumpon is, to some extent, counterbalanced byan overall reducon in consumpon of goods and services by rms (Figure 2.16). Energyexpenses by rms are reduced in similar proporon to those of households. But demand© OECD/IEA, 201324. elfare impacts of the 4-for-2 °C Scenario are qualitatively similar to consumption trends illustrated in thissection.74 World Energy Outlook | Special Report

y rms for other goods, notably manufactured products, is maintained, though changedin detail. The resulng net impact on consumpon by rms is a 0.2% decrease ($90 billion)in OECD countries, oseng the net increase in households. arger cuts in the energy billsof non-OECD rms also lead to a net 0.2% drop in consumpon in 2020 (-$80 billion).The assumed mullateral reform of fossil-fuel consumpon subsidies leads to moreecient resource allocaon across the enre economy and is thus welfare-enhancingin countries implemenng the reform. This measure benets Middle Eastern countriesparcularly, which results, in combinaon with other elements of the policy package, in aslight increase in their GDP.1234Building blocks for steeper abatement post-2020The implementaon of assumed policy measures in the 4-for-2 °C Scenario signicantlyreduces growth in global CO 2 emissions from the energy sector. Global energy-related CO 2emissions connue to grow in the short term, to 32.1 Gt in 2020, 25 but this is only some 2%higher than is required to put the world on track for a global temperature rise in the longterm no higher than 2 °C. Emissions stabilise aer 2020 and start falling slowly, reaching30.8 Gt in 2035, 6.2 Gt (or 17%) lower than in the New Policies Scenario (Figure 2.17).Almost 60% of the CO 2 savings in 2035 occur due to the reduced use of coal as a resultof lower electricity demand and less use of the least-ecient coal power plants. The useof oil is also reduced, contribung 25% to overall emissions reducons, largely due toeciency standards in road transport and the phase-out of fossil-fuel subsidies. Natural gascontributes another 17% to emissions reducons, due to lower electricity demand and thephase-out of fossil-fuel subsidies. However, the use of natural gas, sll grows unl 2035 inthe 4-for-2 °C Scenario, though at a reduced average annual rate, relave to 2010, of 1.1%.It is the only fossil fuel for which demand sll increases signicantly over todays levels.Figure 2.17 World energy-related CO 2 emissions by scenarioGt393633New Policies Scenario4-for-2 °C Scenario3027Coal© OECD/IEA, 201324Oil450 Scenario Gas212010 2015 2020 2025 2030 203525. Including CH 4 , energy-related emissions reach 34.9 Gt CO 2 -eq in 2020.Chapter 2 | Energy policies to keep the 2 °C target alive 75

These addional measures are not sucient alone, however, to reach the 2 °C target inthe long term, as CO 2 emissions are 8.8 Gt (or 40%) higher than the required level in 2035,a level which, if realised, would represent only a 13% probability of stabilisaon at 2 °C,and a 50% likelihood of reaching 2.9 °C. In order to change course post-2020 and put theworld rmly on track consistent with a for a 50% chance of reaching the 2 °C target, furtherreducons are required. Relave to the 4-for-2 °C Scenario, these addional reduconsamount to a cumulave 78 Gt through 2035.For the required transformaon of the energy sector post-2020 to achieve climate targets,all technology opons will be needed and their early availability is essenal to minimisethe addional costs associated with their deployment. hile deep emissions reduconsare possible if consumers were to reduce demand for energy services such as mobility orcomfort, such changes are considered unlikely and might entail lower economic acvity.The acceptable keys to the required emissions reducon are, therefore, technologicaldevelopments and ongoing improvements in eciency.Figure 2.18 World electricity generation from low-carbon technologies byscenarioTWh7 0006 0005 0004 0003 0002020:4-for-2 °C ScenarioIncremental overperiod 2020-2035:4-for-2 °C Scenario450 Scenario2 0001 000Nuclear Hydro Wind Bioenergy Solar PV Other CCSNote: Other includes geothermal, concentrated solar power and marine.© OECD/IEA, 2013In the power sector, for example, a profound change in the way electricity is generated isneeded post-2020. In the 4-for-2 °C Scenario, the share of low-carbon technologies includingrenewables, nuclear and CCS, reaches 40% in 2020, up from 32% today, but this is sll well shortof the required level of almost 80% in 2035, as reected in the 450 Scenario (see Chapter 1).Achieving this target will require the use of all low-carbon technologies, with the largestcontribuon coming from increased use of renewables, as electricity output from hydro, wind,biomass, solar and other renewables combined in 2035 is over 4 000 terawa-hours (Th)76 World Energy Outlook | Special Report

(or almost 40%) higher than in the 4-for-2 °C Scenario (Figure 2.18). Electricity generaonfrom nuclear power needs to increase by almost 1 800 Th in 2035 (or about 40%) over thelevel achieved in the 4-for-2 °C Scenario. In relave terms, the largest scale-up, post-2020,is needed for CCS, at seven mes the level achieved in the 4-for-2 °C Scenario, or around3 100 Th in 2035, with installaon in industrial facilies capturing close to 1.0 Gt CO 2 in2035. Projects in operaon today in all sectors capture only 6 Mt CO 2 , implying a very rapiddeployment of CCS in many applicaons. For all low-carbon technologies, the removalaer 2020 of market and non-market barriers towards their wider adopon will require aconsistent policy eort over the next decade.1234In the transport sector, a shi towards low-carbon fuels is required as improving theeciency of road vehicles alone will not lead to the steep reducons required aer 2020(IEA, 2012b). hile natural gas and biofuels are promising alternaves to oil, their potenalto reduce emissions relave to oil is limited, either due to their carbon content (naturalgas) or quesons with regard to their sustainability and conicts with other uses for thefeedstock (biofuels). From todays perspecve, high expectaons fall on the deploymentof electric and plug-in hybrid vehicles, with their share of all PD sales required to riseby above one-quarter by 2035 (as in the 450 Scenario). Such a dramac shi away fromcurrent sales paerns is unprecedented in global car markets. In order to aain such asteep increase in market shares, electric vehicles need to be freely available to the massmarket at compeve costs by 2020, soluons having been idened to address issuessuch as driving range (for example, fast recharging infrastructure) or other issues crucial toconsumer acceptability.© OECD/IEA, 2013The large deployment of CCS aer 2020 is required partly as a fossil-fuel assets proteconstrategy. 26 In 2020, there are almost 2 000 G of coal-red capacity and almost 1 800 Gof gas-red capacity installed worldwide in the 4-for-2 °C Scenario, together represenng58% of total electricity generaon. Deploying CCS and retrong fossil-fuel plants with CCSavoids the need to mothball large parts of this eet and improves the economic feasibilityof the climate objecve, in parcular in regions where geological formaons allow forCO 2 storage (IPCC, 2005). So far, only a handful of large-scale CCS projects in naturalgas processing are operang, together with some low-cost opportunies in industrialapplicaons. hile many projects are economically viable because CO 2 is purchased forenhanced oil recovery (EOR), there is no single commercial CCS applicaon to date in thepower sector or in energy-intensive industries. Addional to technological and economicchallenges, CCS must overcome legal challenges related to liabilies associated with theperceived possibility of the escape of the CO 2 gases that are stored underground. Exisngpolicies so far are insucient to incenvise investments in commercial-scale CCS (Box 2.4).Although progress has been made towards improving the regulatory framework, sucient26. See Chapter 3 for an analysis of the economic implications of stronger climate policies.Chapter 2 | Energy policies to keep the 2 °C target alive 77

technology and deployment support is lacking and the absence of a substanal price signalhas impeded necessary development of CCS technology.Past analysis has demonstrated that emissions migaon becomes more costly without CCS(IEA, 2011c). 27 In the power sector, delaying introducon of CCS from 2020 to 2030 wouldincrease the investment required to keep the world on track for the 2 °C target by more than$1 trillion, as the need for addional investment in other low-carbon technologies, suchas renewables and nuclear, would more than oset the reduced investment in coal powerplants and CCS (Figure 2.19). Although a reducon of electricity demand can accommodatelower CCS deployment in the power sector, there are limits to the extent to which energyeciency can reduce energy demand without reducing energy services.Figure 2.19 Change in cumulative investment in power generation if CCS isdelayed, relative to the 450 Scenario, 2012-2035Trillion dollars (2011)2.01.51.00.50-0.5Renewables:OtherSolar PVWindBioenergyHydroCoal and gas:Without CCSCCS-fied-1.0Coal Gas Nuclear Renewables Totalhile the delayed availability of CCS can be compensated in the power sector by increasinginvestment in renewables and nuclear, albeit at higher costs, the fact that alternavesare not available to compensate for a shorall of the deployment of CCS technologiesin industry is a bigger challenge. Energy-ecient equipment can go a long way (and isdeployed to its maximum in the 450 Scenario), but the potenal for renewables inindustrial applicaons is limited. A higher use of decarbonised electricity in industry hassome potenal, for example in iron and steel via secondary steelmaking, but this wouldnot allow the producon of certain product qualies. ithout the deployment of CCS or analternave low-carbon technological breakthrough in industrial processes, industry wouldstruggle to reach the levels of decarbonisaon necessary to achieve the 450 Scenario,so pung further pressure and imposing greater costs on sectors with more opons todecarbonise, such as transport and power generaon.© OECD/IEA, 201327. The analysis of the cost of delaying CCS in this section is based on a comparison of the cost of reachingthe 450 Scenario from WEO-2012 with those of the Delayed CCS Case that was presented in WEO-2011 and thatassumes that CCS is introduced in 2030, i.e. ten years later than in the 450 Scenario.78 World Energy Outlook | Special Report

Box 2.4 Policies to support CCSCCS deployment requires strong policy acon, as present market condions areinsucient and current CO 2 pricing mechanisms have failed to provide adequateincenves to drive it. Governments need to put in place incenve policies that supportnot only demonstraon projects but also wider deployment. The opmal porolio ofincenve policies needs to evolve as the technology develops from being relavelyuntested at a large scale to being well-established. The incenve policy porolioshould inially be weighted towards technology-specic support, explicitly targengthe development of CCS into a commercial acvity through the provision of capitalgrants, investment tax credits, credit guarantees andor insurance (Figure 2.20). Atthe early stage, measures are needed to enable projects to move ahead in order togenerate replicable knowledge and experience. Targeted sector-specic industrialstrategies are then needed to move CCS from the pilot project phase to demonstraonand then deployment phases. In the long term, a technology-neutral form of support,e.g. in the form of a CO 2 price, allows the deployment of CCS to be considered inrelaon to other cost-eecve abatement opons.1234Figure 2.20 Policy framework for the development of CCSSource: Adapted from IEA (2012d).© OECD/IEA, 2013In addion to the tailored incenve policies needed to drive CCS forward, governmentsneed to ensure that the terms of regulatory frameworks (or their absence) do notimpede demonstraon and deployment of CCS. In this context, a regulatory frameworkis the collecon of laws (and rules or regulaons, where applicable) that removesunnecessary barriers to CCS and facilitates its implementaon, while ensuring it isundertaken in a way that is safe and eecve. Jurisdicons in the European Union, theUnited States, Canada and Australia have established legal and regulatory frameworksfor CCS over the past few years (IEA, 2011c). hile developing a legal and regulatoryframework for a novel technology is a daunng challenge, a regulatory frameworkcan, within limits, be developed in phases, with regulaons tailored to the stageof technology deployment, as has been done in some jurisdicons, so long as theregulatory process stays suciently ahead of the game. But, in any case, frameworkdevelopment must begin as soon as possible to ensure that a lack of appropriateregulaon does not slow deployment.Chapter 2 | Energy policies to keep the 2 °C target alive 79

CCS can not only safeguard otherwise stranded assets in power generaon and industry,but also has a value for fossil-fuel producers. To achieve climate targets, CCS would mainlybe applied to coal- and gas-red power plants and in the iron and steel, and cementindustries, which largely consume coal. If the introducon of CCS in power generaonand industry is signicantly delayed, coal consumpon must decrease correspondinglyif climate targets were to be met: while coal consumpon would decline from around5 200 million tonnes of coal equivalent (Mtce) today to 3 300 Mtce in 2035 if CCS wasintroduced on a large scale by 2020, it would be reduced by another 900 Mtce if theintroducon of CCS was delayed by ten years.Oil and gas producers would also be aected by delayed introducon of CCS. Although gasconsumpon would sll increase over todays level, growth would be slower without theapplicaon of CCS to gas-red power plants. For oil producers, the eect of delaying CCSwould be indirect: in order to keep cumulave CO 2 emissions the same in the absence ofCCS, the transport sector would need to compensate by reducing emissions further throughwider deployment of electric vehicles. This could reduce oil consumpon by around1.3 million barrels per day in 2035, compared with introducon of CCS by 2020. Overall, ifthe introducon of CCS was delayed unl 2030, then coal producing countries would loserevenues of $690 billion, gas producers would lose $430 billion and oil producers about$230 billion (Figure 2.21). The combined loss of revenue for oil and gas producers is roughlyequal to that of coal producers.Figure 2.21 Change in fossil fuel cumulative gross revenues by type andregion if CCS is delayed, relative to the 450 ScenarioOECDNon-OECDOilGasCoal-700 -600 -500 -400 -300 -200 -100Billion dollars (2011)© OECD/IEA, 2013The analysis in this chapter has shown that it is possible in 2020 to be within reach ofa 2 °C trajectory through the adopon in the short term of a number of well-targeted,decisive policy intervenons that will not damage economic growth. Aer 2020, the energy80 World Energy Outlook | Special Report

transion must move from being incremental to transformaonal, i.e. an energy sectorrevoluon, is required, which will be aained only by very strong policy acon. The pivotalchallenge is to move the abatement of climate policy to the very core of economic systems,inuencing in parcular, all investment decisions in energy supply, demand and use. Everyfeasible abatement opportunity will need to be seied. An important way to achieve this isby pricing carbon emissions.By reecng in energy prices the hidden cost of climate damage, well-judged carbonpricing gives all producers and consumers the necessary incenve to reduce greenhousegasemissions, while allowing exibility in the technical and business soluons adopted tomake these reducons. Carbon pricing provides an incenve for innovaon, and dependingon the policy design, could help the scal posion of governments.1234Carbon pricing can be implemented in a multude of ways, matching naonal circumstancesand climate objecves. Carbon taxes provide simplicity and investment certainty, whileemissions trading can be used if exibility and internaonal linkages are a higher priority.The revenues raised can be used to maximise overall economic welfare (for example,by reducing other distoronary taxes) and, in this case, the net benet can exceed theeconomic slow-down resulng from energy price rises (Parry and illiams, 2011). Therevenues raised can also be used in a targeted way to oset the impact of increasingprices for low-income consumers or vulnerable industries and, if designed well, can sllmaintain the appropriate incenves for cleaner energy choices. Targets in an emissionstrading scheme can be based on an absolute emissions cap, which gives certainty overthe abatement outcome, or an emissions intensity, which provides greater exibility forrapidly-developing economies and can have a lesser impact on energy prices.A key advantage of carbon pricing is its potenal to opmise acon internaonally, eitherthrough internaonal credit mechanisms or the linking of domesc emissions tradingschemes. Internaonal linking allows abatement to occur rst where it is cheapest, drivinginvestment ows and technology to regions with abatement opportunies. In theory, thisshould appeal both to buying and selling naons buyers benet from cheaper compliancewith emissions targets and sellers prot from higher unit sales. However, other policalconsideraons mean that, in the real world, linking decisions is complex. There may beconcerns about oulows of capital from buying countries, and loss of cheap abatementopons in selling countries. There may also be concern that internaonal linking willraise domesc carbon prices in regions with ample abatement opportunies, owingthrough to energy prices. Even if technical design elements enable linking, these policalconsideraons may mean that carbon pricing policies remain mostly domesc or regionalfor some me.© OECD/IEA, 2013Given the important role that carbon pricing must play from 2020, it is essenal to usethe few years ahead to design and test carbon pricing systems in order to gain experience.Experience in the EU ETS and other systems, such as the US-based Regional GreenhouseGas Iniave, has shown that it can take a number of years to put a carbon pricingsystem in place, and several more to sele on robust and sustainable policy parametersChapter 2 | Energy policies to keep the 2 °C target alive 81

(see Chapter 1). Countries that have worked through these design issues and have carbonpricing as an available tool will be at a disnct advantage in managing the ambiousemissions reducons that will be required under any new internaonal agreementconsistent with keeping the long-term average global temperature rise to below 2 °C.Box 2.5 Clean energy standardsEmissions trading does not necessarily mean a cap and trade system: there are manydesign opons for integrang a exible carbon price into investment decision-making.For example, a clean energy standard (CES) for the electricity sector could work in avery similar way to an intensity-based ETS, embedding a clear incenve for a cleanenergy transion across electricity sector decision-making and retaining many of thebenets of carbon pricing. Under a CES, permits are awarded for each unit of cleanelectricity generated, on the basis (approximately) of avoided emissions compared to abaseline, which requires careful denion. Electricity suppliers must surrender permitscorresponding to a required share of clean energy. The tradability of permits creates aprice for a unit of low-carbon electricity generaon, rather than a price for emissions. Aswith any emissions trading or creding system, seng the emissions cap (in this case therequired share of clean energy) at an ambious level is crical to ensure that there is afunconing market that results in real, addional emissions reducon.CES systems do not necessarily raise energy prices by as much as an equivalent capand trade system because the cost passed through to consumers is only that of therequired investment in clean energy, rather than a charge on all fossil-fuel generaon.Conversely, they do not raise revenue for governments to use in an economicallybenecial way and do not provide a clear signal for demand reducon.Despite its importance, carbon pricing alone will not be sucient to drive necessarychanges. Direct targeted policies will be needed to unlock the energy eciency potenal,such as those proposed in the 4-for-2 °C Scenario to 2020. Energy eciency is oen blockedby non-market barriers which, le in place, could distort the response to carbon prices.The development of new technologies also requires targeted support, both to bring downcosts and to allow for scaling-up to the level required for the long term. Typical examplesare CCS, some renewable energy technologies, smart grids and electric vehicles (whichalso require supporng infrastructure). It may also be necessary directly to discourageinvestment in long-lived energy infrastructure that might otherwise be beyond the reachof a carbon price signal. Although some sectors are less responsive to a carbon price,such a market signal may sll play a supporng role. For example, fuel-economy standardsin transport are much more eecve if complemented by fuel-excise charges to preventrebound, as is provided for in the 450 Scenario.© OECD/IEA, 201382 World Energy Outlook | Special Report

Chapter 3Managing climate risks to the energy sectorHighlightsBuilding resilience now The energy sector must ensure that its assets are resilient to the physical impactsof the climate change that is already occurring and in prospect, and also that itscorporate strategy is resilient to the possibility of stronger climate change policiesbeing adopted in the future. If these implicaons of climate change are not factoredinto investment decisions, carbon-intensive assets could need either to be reredbefore the end of their economic lifeme, idled or undergo retrong. The energy sector is not immune from the physical impacts of climate change andmust adapt. In mapping energy system vulnerabilies, we idenfy some impactsthat are sudden and destrucve, with extreme weather events posing risks to powerplants and grids, oil and gas installaons, wind farms and other infrastructure.Other impacts are more gradual, such as sea level rise on coastal infrastructure,shiing weather paerns on hydropower and water scarcity on power plants. Urbanareas, home to more than half the worlds populaon, experience annual maximumtemperatures that increase much faster than the global average. Our analysis,which takes account of the changing climate, shows that global energy demand forresidenal cooling is 16% higher in 2050 than when this eect is not factored in.Developing countries cooling needs increase the most, parcularly in China. Even under a 2 °C trajectory, upstream oil and gas generates gross revenues of$107 trillion through to 2035. Though 15% lower than they might otherwise be,these revenues are nearly three mes higher than in the last two decades. Strongerclimate policies do not cause any currently producing oil and gas elds to shut downearly. Some elds yet to start producon are not developed before 2035, but thisrisk of stranded assets aects only 5% of proven oil reserves and 6% of gas reserves.ower coal demand impacts on coal supply but, as investment costs are a smallshare of overall mining costs, the value of stranded assets is relavely low. In the power sector overall, gross revenues are $57 trillion through to 2035 under a2 °C trajectory, 2% higher than those expected on the trajectory now being followed,as higher electricity prices outweigh lower demand. Net revenues from exisngnuclear and renewables capacity are boosted by $1.8 trillion collecvely, oseng asimilar decline from coal plants. Of the power plants that are rered early, idled orretroed with CCS, only 8% (165 G) fail to fully recover their investment costs.© OECD/IEA, 2013 Delaying stronger climate acon unl 2020 would avoid $1.5 trillion in low-carboninvestments up to that point, but an addional $5 trillion would then need to beinvested through to 2035 to get back on track. Developing countries have the mostto gain from invesng early in low-carbon infrastructure in order to reduce the risk ofneeding to prematurely rere or retrot carbon-intensive assets later on.Chapter 3 | Managing climate risks to the energy sector 83

IntroduconAs the largest source of greenhouse-gas emissions, a signicant burden lies with the energysector to deliver the 2 degrees Celsius (°C) climate goal commied to by governments.Chapter 2 examined the need and the potenal for policy makers to take short-termclimate acons while negoang long-term deals. This chapter shis focus to the need andscope for self-interested acon by the energy sector. The global climate agreement thatis expected to come into force in 2020 is both within the lifeme of many energy assetsoperang today and within the current planning horions of the industry: it is, accordingly,one of the present uncertaines that the industry has to manage as it connues to invest.Many of the investment decisions being taken today do not appear to be either consistentwith a 2 °C climate goal or suciently resilient to the increased physical risks that areexpected to result from future climate change. Is the energy sector beginning to factorthese issues into its planning and investment decisions, or is there a risk that it will needto write-o some of its assets before the end of their economic life and before they havegenerated the nancial returns expected of themClimate change is, and will connue to be, an important issue for the energy sector. Theindustry can rise to the challenges brought about by climate change, but this will requirethe reorientaon of a system valued at trillions of dollars and expected to receive trillionsmore in new investment over the coming decades. This chapter analyses some keyissues that the energy sector must confront. It begins by examining the range of physicalimpacts that the changing climate might have on our energy system, highlighng thoseparts of our energy infrastructure that may be most vulnerable and need to become moreclimate resilient. It then analyses the potenal economic impact on the energy sectorof the stronger climate policies, necessary to meet the 2 °C climate goal, which may beadopted, measuring this in terms of the impact on the sectors future revenues, and onthe lifeme and protability of exisng assets (in terms of fossil-fuel reserves and energyinfrastructure) and those yet to be discovered, developed or built. Aer all, in a worldwhere condence in the conclusions of climate science is hardening and the availablecarbon budget is shrinking, those companies deriving their revenues most closely fromfossil fuels are at the highest risk from changes in energy and climate policies, potenallyundermining the business models that have historically served them well. The chapter alsolooks more broadly at the implicaons if stronger acons on climate change were delayed,considering the simple queson of whether it is beer to act now or act later.© OECD/IEA, 2013Impacts of climate change on the energy sectorThe energy sector is not immune to the impacts of climate change and here we examinethe exposure of dierent parts of the energy system to the associated physical risks. Evenstringent acon to contain the extent of climate change, such as realisaon of the 2 °Cclimate goal, will not eliminate the impacts of climate change and the need to adaptto it (Box 3.1). ithout such adaptaon, climate change will increase the physical risksto energy supply, pushing up capital and maintenance costs, impairing energy supply84 World Energy Outlook | Special Report

eliability and accelerang the deterioraon (and therefore the pace of depreciaon) ofassets. The industry must judge the extent to which it will be impacted by future climatechange and how it will need to adapt to the new physical risks, whether, for example,through changes in the locaon and the resilience of new infrastructure, through thedecentralisaon of the energy network, or by insuring against loss. Given the naonalimportance of some energy infrastructure, government will also have an important role toplay. Overall, developing an eecve strategy will involve the interplay between a broadrange of stakeholders, including governments, energy companies, climate sciensts andinsurers.1234Box 3.1 Climate change adaptationhile climate change migaon describes acons to reduce greenhouse-gas emissionsin the atmosphere, climate change adaptaon relates to adjustments that are made inresponse to actual or expected climate events (or their eects), which either moderatethe harm caused or exploit benecial opportunies (UNFCCC, 2013). Signicant eortsto migate climate change can reduce the need for adaptaon, but not dismiss itenrely because of the global warming that will result from the accumulaon of pastlong-lived greenhouse-gas emissions. Climate migaon and adaptaon are thereforenot mutually exclusive strategies and there are synergies that can be exploited toenhance their cost-eecveness.© OECD/IEA, 2013Climate change impacts, and the adaptaon needs, will vary by region and sector. Someof the impacts will be gradual, as a long-term increase in global temperature brings abouta rise in sea level, greater water scarcity in some regions and changes in precipitaonpaerns. Energy demand paerns will change (such as for heang and cooling), powerplant cooling and eciency will be aected, as will hydropower output, and coastalinfrastructure (including reneries, liqueed natural gas NG plants and power plants)will be threatened (Figure 3.1). For example, in the United States alone nearly 300 energyfacilies are located within 1.2 metres of high de (Strauss and iemlinski, 2012). Otherimpacts of climate change are likely to be more sudden and destrucve, with extremeweather events, such as tropical cyclones, heat waves and oods, expected to increasein intensity and frequency (Box 3.2). Cyclones can damage electricity grids and threatenor severely disrupt oshore oil and gas plaorms, wind farms and coastal reneries. Heatwaves and cold spells will impact upon peak load energy demand, pung greater stresson grid infrastructure and undermining the ability of power plants to operate at opmumeciency. Gradual and sudden climate impacts can also interact, such as a sea level riseand more powerful storms combining to increase storm surges. Furthermore, disruponsto the energy system caused by climate events can have signicant knock-on eects onother crical services, such as communicaons, transport and health.Chapter 3 | Managing climate risks to the energy sector 85

© OECD/IEA, 201386 World Energy Outlook | Special ReportFigure 3.1 ⊳ Selected climate change impacts on the energy sectorSources: Based on ©Munich RE (2011), with information from Acclimatise (2009), Foster and Brayshaw (2013), Schaeffer, et al. (2012) and IEA analysis.

Box 3.2 Extreme weather – a new factor in energy sector decisions?Climate change aects the frequency, intensity and duraon of many extreme weatherevents and evidence shows that this is already happening (IPCC, 2007 Peterson,Sto and Herring, 2012). A recent report nds that the distribuon of seasonal meantemperature anomalies has shied toward higher temperatures and that the range ofanomalies has increased, with an important change being the emergence of a categoryof extremely hot summerme outliers, relave to a 1951 to 1980 base period (Hansen,Sato and Ruedy, 2012). It also nds that, in the 1960s, less than 1% of the global landarea (around the sie of Iran) was aected by summerme extremes 1 , but observaonsnow point to this gure having increased to 10% (an area larger than India andChina combined). A separate study links the 35-year warming trend in ocean surfacetemperature to more intense and larger tropical cyclones (Trenberth and Fasullo, 2008).In 2012, Tropical Storm Irene made landfall much further north than was typical andHurricane Sandy was the largest hurricane ever observed in the Atlanc (NOAA Naonaleather Service, 2012). The 2003 heat wave in Europe is esmated to have caused upto 70 000 deaths (Robine, et al., 2008) and, while the summer average was only 2.3 °Cabove the long-term average in Europe, August temperatures in several cies were upto 10 °C higher than normal. Several densely populated urban areas are already at highrisk from natural haards. For example, Tokyo and New ork are at risk from cyclonesand oods, while New Delhi and Mexico City are at risk from oods (UNPD, 2012).The IPCC (2012) concluded that it is virtually certain that an increase in the frequencyand magnitude of warm daily temperature extremes will occur over the course of thiscentury. In a world where the average global temperature increases by 4 °C, relaveto pre-industrial levels, several studies nd that the most marked warming will beover land and actually be between 4 °C and 10 °C. A global temperature increase of4 °C means that a 1-in-20 year extreme temperature event today is likely to become a1-in-2 year event and, around the 2040s, about every second European summer couldbe as warm (or warmer) than the extreme summer of 2003 (Sto, Stone and Allen, 2004).1234Energy demand impactsComprehensive global studies covering the impact of climate change on the energy sectorare sll lacking, though some regional and sector-specic analysis exists. The buildingssector has been examined in more depth than most, with studies nding that temperatureincreases are expected to boost demand for air condioning, while fuel consumpon forspace heang will be reduced. The eects in the transport sector (such as higher use of aircondioners) and in the industry sector (changed heang and air condioning needs) areexpected to be on a smaller scale (ilbanks, et al., 2007). In agriculture, a warmer climateis likely to increase demand for irrigaon resulng in a higher energy demand for waterpumps.© OECD/IEA, 20131.Defined as being more than three standard deviations warmer than the average temperature.Chapter 3 | Managing climate risks to the energy sector 87

Around one-quarter of global nal energy consumpon is in the residenal sector nearly2 100 Mtoe in 2010 with space heang accounng for around 30% of this and spacecooling making up around 3%. At present, countries in cold climates, such as Canada andRussia, have high heang demand (heang degree days HDD of 4 000 or above), buta comparably low demand for space cooling (cooling degree days CDD below 300). 2Countries in hot climates, such as India, Indonesia and those in Africa, have virtually nodemand for space heang, but high cooling needs (CDD above 3 000). Other regions aresituated in a more temperate climate or extend over dierent climate ones, such as theEuropean Union and the Middle East where, for example, Iran has around 1 000 CDDs peryear, while Saudi Arabia has more than 3 000 CDDs.Urban areas, home to more than half the worlds populaon, are at the forefront of thechallenge of climate change. Annual maximum temperatures in cies increase muchfaster than the global average, fostered by the urban heat island eect. For example, anaverage global warming of 4.6 °C above pre-industrial levels by 2100 (as in the IPCCsRCP 8.5 Scenario) is projected to result in maximum summer temperatures in New orkincreasing by 8.2 °C (Figure 3.2) (Hempel, et al., 2013). In such a case, the extreme summerexperienced in Moscow in 2010 may be closer to the norm experienced in 2100, while theEuropean summer of 2003 could be cooler than the average by that me. In a case wherethe average global warming is 1.5 °C above pre-industrial levels by 2100 (as in the IPCCsRCP 2.6 Scenario), the maximum summer temperature in New ork is projected to increaseby only 1.6 °C.For this report, we have extended our orld Energy Model to allow for the impact ofclimate change on the projecons for heang and cooling energy demand in the residenalsector. 3 Given the relavely long mescales over which climate impacts occur, the mehorion for this purpose is from 2010 to 2050, though it is recognised that the largestimpacts will be felt aer this date: our New Policies Scenario is consistent with an averageglobal temperature increase of around 2 °C by 2050 (3 °C by 2100 and 3.6 °C by 2200),compared with pre-industrial levels. In addion to changes in average energy demand forheang and cooling, climate change may also increase peak-load demand for cooling.© OECD/IEA, 20132. HDD and CDD are measurements designed to reflect the demand for energy needed to heat or cool abuilding. HDD and CDD are defined relative to a base temperature the outside temperature abovebelow whicha building needs no heating or no cooling. For example, if 18 °C were the baseline temperature, a summer daywith an average temperature of 25 °C would result in a CDD of 7.3. The orld Energy Model has been extended to 2050 for the climate impact analysis, where the effect onspace cooling is based on the methodology proposed in McNeil and etschert (2007). Demand for space heatingis based on energy services demand, which is driven by factors including change in floor space per capita,price elasticity and a change in HDD. Data on population weighted degree days comes from the PASIM-ENTSmodel (Holden, et al., 2013).88 World Energy Outlook | Special Report

Figure 3.2 Projected annual maximum temperatures in selected citiesunder different global warming trends1°C46New York44.5 °C423837.9 °C3436.3 °C302000 2020 2040 2060 2080 2100Global warmingin 2100 of:4.6 °C1.5 °C234°C46423834.6 °CParis42.8 ° °CGlobal warmingin 2100 of:4.6 °C1.5 °C3436.2 ° °C302000 2020 2040 2060 2080 2100°C4238Moscow37.9 ° °CGlobal warmingin 2100 of:4.6 °C341.5 °C3033.3 ° °C31.4 °C262000 2020 2040 2060 2080 2100°C5450New Delhi51.6 °CGlobal warmingin 2100 of:4.6 °C464245.8 °C46.8 °C1.5 °C382000 2020 2040 2060 2080 2100© OECD/IEA, 2013Note: The temperature trend is the average of ve climate models.Sources: NOAA (2013a) Hempel, et al. (2013) and IEA analysis.Chapter 3 | Managing climate risks to the energy sector 89

Energy demand for cooling in the OECD was higher than in non-OECD countries in 2010.In our New Policies Scenario, not accounng for climate change, global energy demandfor space cooling already grows by nearly 145% to 2035 and is around 175% higher by2050, pushed largely by demand from emerging markets in Asia, primarily China. However,climate trends are going to change over the coming decades and, once these are takeninto account, the results show that global energy demand for space cooling increases byaround 170% to 2035 and 220% by 2050, compared with 2010 levels. In 2050, energydemand for cooling is signicantly higher in non-OECD countries than in the OECD. In non-OECD countries demand increases by nearly 400% (105 million tonnes of oil equivalentMtoe) compared to around 60% (20 Mtoe) in the OECD by 2050. The largest absolutechange in energy demand for cooling is in China, where the eect of increasing incomes(boosng ownership of air condioners) is complemented by increased cooling needs as aresult of rising temperatures. In relave terms, the biggest change in cooling demand thatoccurs as a result of climate change (i.e. a comparison between our New Policies Scenarioresults with and without climate change) is in China, followed by the United States, theMiddle East and India (Figure 3.3). All of the increase in energy demand for cooling in theresidenal sector is in the form of electricity, which can be challenging for power systemstability during extreme heat waves.Figure 3.3 Change in energy demand for space cooling by region in theNew Policies Scenario after accounting for climate change20%15%10%5%4 0003 0002 0001 000Number of cooling degree daysAddional 2035to 20502035CDD in 2005(right axis)ChinaUnitedStatesMiddleEastNote: These regions cover almost three-quarters of the global energy consumpon for space cooling.India© OECD/IEA, 2013In 2010, global energy demand for heang was ten mes the level for cooling, withOECD countries accounng for around 60% of the total. In the case of energy demand forheang, a New Policies Scenario that does not take account of climate change projectsa 20% increase to 2035 and 28% by 2050. Once climate change eects are taken intoaccount, energy demand for space heating increases by only 11% to 2035 and 12% to2050, compared with 2010 levels. In the OECD, energy demand for heang increases onlymarginally to 2035 and then declines to 2050, ending at a similar level (385 Mtoe) to 2010.Non-OECD countries drive almost all of the global increase, reaching around 260 Mtoe90 World Energy Outlook | Special Report

in 2050, with most of the increase occurring by 2035. ooking at the absolute changein energy demand before and aer climate eects are taken into account, the largestreducons are in Europe, China and the United States. The largest relave reducon inspace heang needs occurs in China (Figure 3.4). ess than 10% of global energy demandfor heang is in the form of electricity, with the rest split among fuels, mainly gas.Figure 3.4 Change in energy demand for space heating by region in theNew Policies Scenario after accounting for climate changeChinaEuropeanUnionUnitedStatesRussiaAddional2035 to 20501234-4%-8%-12%-16%1 5003 0004 5006 000Number of heang degree days2035HDD in 2005(right axis)Note: These regions cover almost three-quarters of global energy consumpon for space heang.Energy supply impacts© OECD/IEA, 2013Oil and gas exploraon and producon already takes place in a number of challengingclimates, and the industry has innovated over me to open up new froners, from desertsto deepwater to the Arcc Circle. Climate impacts on this sector will include those thatare relavely gradual, such as a rising sea level and changing levels of water stress, andsudden impacts, including extreme wave heights, higher storm intensies and changing iceoes (Table 3.1). For example, ten Chinese provinces already suer from water scarcity inper capita terms and, if this were to become more acute, exisng coal operaons (the coalsector has the largest share of industrial water use in China) and future plans to develop itshuge shale gas resources could be aected (IEA, 2012a). Infrastructure will need to adaptto boost resilience to a changing climate, and this is likely to entail addional costs. Iraq,which also suers from water scarcity, provides a current example of this, as it is invesngaround $10 billion to construct a Common Seawater Supply Facility that would treat10-12 million barrels per day (mbd) of seawater and transport it 100 kilometres to oilelds where it will help maintain reservoir pressure. This investment will migate futurepressures on Iraqs valuable freshwater sources (see our for moredetail IEA, 2012b).Chapter 3 | Managing climate risks to the energy sector 91

© OECD/IEA, 201392 World Energy Outlook | Special ReportTable 3.1 Selected climate impacts on the oil and gas sector by regionRegionMiddle EastOECDAmericasRussiaAfricaLanAmericaChinaOECD EuropeOECD AsiaOceaniaShare of world oilproducon, 2011*33%17%13%11%9%5%4%1%Climate impact ater stress Increase in air and sea surface temperature Increase in intensity of tropical cyclones(Gulf of Mexico) Sea level rise ater stress Permafrost thaw (Alaska, north Canada) Permafrost thaw (Siberia) ater stress (North Africa) Sea level rise (est Africa) Sea level rise Increase in storm acvity (Brail) Increase in air and sea surface temperature(South China Sea) ater stress Increase in intensity of storms Extreme wave heights (North Sea) Increase in intensity and frequency oftropical cyclones (Australia) Increase in air temperatureImpact on the oil and gas sector Increase producon costs Reduce cooling capacity in certain processes, liming the capacity of agiven facility, i.e. NG Increase costs of oshore plaorms, increase plaorm height, andmore frequent producon interrupons Increase the shut down me of coastal reneries Reduce the availability of ice road transportaonincrease pipelinemaintenance Reduce the availability of ice road transportaonincrease pipelinemaintenance Increase producon costs Increase the shut down me of coastal reneries Increase the shut down me of coastal reneries Increase in oshore plaorm costs Reduce cooling capacity in certain processes, liming the capacity of agiven facility Render some unconvenonal producon unfeasible or very costly (i.e. CT) Increase costs of oshore plaorms and increase producon interrupons Increase costs of oshore plaorms and increase producon interrupons Increase costs for coolingNote: Regional oil producon includes crude oil, natural gas liquids and unconvenonal oil but excludes processing gains and biofuels supply.

Growing producon of unconvenonal gas is expected to result in increased water demandfor hydraulic fracturing, as highlighted in the (IEA, 2012c). Shale gas or ght gas development can requireanything between a few thousand and 20 000 cubic metres (between 1 million and 5 milliongallons) per well. In areas of water scarcity, either now or due to climate change, theextracon of water for drilling and hydraulic fracturing may encounter serious constraints.The Tarim Basin in China holds some of the countrys largest shale gas deposits, but islocated in an area that suers from severe water scarcity. In the United States, the industryis taking steps to minimise water use and increase recycling.1234A major part 45% of the remaining recoverable convenonal oil resources (excludinglight ght oil) is located in oshore elds (1 200 billion barrels) and a quarter of theseare in deep water (a depth in excess of 400 metres). Ice melt can have both posiveand negave eects when looking at oshore oil and gas producon. For example, theregion north of the Arcc Circle is esmated to contain 90 billion barrels of undiscoveredtechnically-recoverable crude oil resources, 47 trillion cubic metres of gas resources (morethan a quarter of the global total) and 44 billion barrels of natural gas liquids (USGS, 2008).onger ice-free summers in the Arcc are expected to result in longer drilling seasons(and new shipping routes), increasing the rate at which new elds can be developed inthe future (though, in our projecons, we do not expect a signicant share of global oiland gas producon to come from the Arcc oshore before 2035). On the other hand,the technical and environmental challenges are already signicant and a number ofprojects have either been held back by the complexity of operaons and by environmentalconcerns, or suspended due to escalang costs. More prolic ice oes and polar storms arelikely to increase the risk of disrupon during Arcc drilling, producon and transportaon(Harsem, Eide and Heen, 2011). Increased ice melt also reduces the availability of ice-basedtransportaon (such as ice roads), adversely aecng oil and gas producon at higherlatudes, such as in Alaska and Siberia. In the case of Alaska, the period that ice roads areopen has halved since 1970 (NOAA, 2013b). Thawing permafrost can also shi pipelinesand cause leaks, which will necessitate more robust (and expensive) design measures.© OECD/IEA, 2013Extreme weather events can cause extensive damage that takes considerable me andmoney to repair. Employee evacuaons and downmes are increasing, as the designthresholds for oshore plaorms are breached more frequently by extreme wave heights(Acclimase, 2009). Oshore oil and gas rigs, such as those northwest of Australia andin the Gulf of Mexico, are already at risk from extreme weather events and the risks areexpected to increase with climate change, with more severe events resulng in moreproducon interrupons (IPCC, 2012). In 2005, Hurricane atrina caused damage valuedat $108 billion in the Gulf of Mexico, which included, together with Hurricane Rita, damageto 109 oil plaorms and ve drilling rigs (nabb, Rhome and Brown, 2005). arge-scalemidstream infrastructure, such as oil reneries and NG facilies, is oen located onthe coast and somemes in locaons prone to extreme weather events and could besimilarly exposed. In addion, reneries are large water consumers and may becomemore vulnerable to water stress, parcularly in those countries where water is already aChapter 3 | Managing climate risks to the energy sector 93

elavely scarce resource. NG plants are typically either water-cooled or air-cooled andtheir eciency is related directly to the temperature of the water or air available for cooling,a 1 °C temperature rise reducing eciency by around 0.7%. A temperature rise in line withour New Policies Scenario could see NG plant eciency decline by 2%-3% on average, andmore in hoer regions. Parcular care needs to be taken over the implicaons of climatechange for the locaon, design and maintenance regime for such long-life infrastructure,which is oen regarded as being of strategic naonal importance.In the case of coal producon, which requires large amounts of water for coal miningacvies (coal cung, dust suppression and washing), increasing water stress may rendercertain operaons more costly. An increase in the frequency and intensity of rainfall couldcause ooding in coal mines and coal handling facilies. In addion, extreme weather canaect transport networks, disrupng the route to market, as observed when ueensland,Australia, was hit by tropical cyclones in 2011 and 2013.e have shown the extent to which a warming climate may boost energy demand forcooling (mainly electricity). At the same me, rising air and water temperatures can havea direct impact on the eciency of thermal power plants, either decreasing electricityoutput or increasing fuel consumpon. High humidity also decreases the eciency ofthermal power plants equipped with cooling towers. ater stress, and an increase inwater temperature, can have a profound impact on power generaon. In China, waterscarcity has meant that some power plants have turned to dry cooling systems, whichcut water consumpon sharply but also reduce plant eciency. ater temperature notonly impacts directly on power plant eciency 4 but, in many countries, may also constrainoperaon because the temperature of cooling water discharged into rivers exceeds anauthorised level. During the summer heat wave in 2003, for example, out of a total of59 nuclear reactors in France, thirteen had to lower their output or shut down in orderto comply with regulaons on river temperatures leading to a total loss of 5.4 terawahours(Th) (EDF, 2013). Constraints due to these eects are expected to increase in thefuture (Table 3.2). Retrong exisng thermal power plants with closed-loop coolingsystems can signicantly reduce water withdrawal, but it involves costs from $100 perkilowa (k) up to $1 000k (BNEF, 2012).The eciency of transmission and distribuon networks is also compromised by a rise inambient temperature. Taking into account the eects of temperature changes on thermalpower plant eciency, transmission line capacity, substaon capacity and peak demand,a higher temperature scenario will either require addional peak generaon capacity andaddional transmission capacity, or a greater demand-side response at peak mes. Evenconsidering the gradual impacts of climate change, the accumulaon of relavely small© OECD/IEA, 20134. In the case of nuclear plants, a 1 °C increase in the temperature of the cooling water yields a decrease of0.12-0.45% in the power output (Durmaya and Sogul, 2006).94 World Energy Outlook | Special Report

changes in performance will have a signicant impact on the availability of generangcapacity and change the cost of electricity. 5Electricity generaon and transmission and distribuon networks are also at signicantrisk from extreme weather events, which can result in infrastructure being damaged ordestroyed and consumers losing their supply, potenally for long periods. eather-relateddisturbances to the electricity network in the United States have increased ten-fold since1992 and, while weather events accounted for about 20% of all disrupons in the early1990s, they now account for 65% (arl, Melillo and Peterson, 2009).1234Table 3.2 Review of the regional impact of water temperature and waterscarcity on thermal power generationEuropeUnitedStatesIndiaChinaWater temperatureNearly 20% of coal-red powergeneraon will need added coolingcapacity.About 1 °C of warming will reduceavailable electric capacity by up to19% in summer in the 2040s.About 1 °C of warming will reduceavailable electric capacity by up to16% in summer in the 2040s.Water scarcity60% of exisng coal-red power plants (347 plants)are vulnerable to water demand and supplyconcerns.Severe water scarcity will amplify competitionfor water and determine thermal plantscompetitiveness and location. Around 70% ofplanned power capacity is in locations consideredeither water stressed or water scarce.ater constraints could make the expected increasein thermal power output unachievable, in parcular,as 60% of thermal power capacity is in northernChina, which has only 20% of freshwater supply.Sources: Jochem and Schade (2009) liet, et al. (2012) Elcock and uiper (2010) BNEF (2013) Sauer, lopand Agrawal (2010) and IEA analysis.Renewable energy can also be aected by climate change. Hydropower currentlyaccounts for 16% of electricity generaon globally, but climate change will aect the sieand reliability of this resource in the future. ater discharge regimes will change, withrun-o from rivers in areas dominated by snow melt potenally occurring earlier in theyear, at levels temporarily higher than previously and with an amplification of seasonalprecipitation cycles. At the global level, output from hydropower is likely to change little,© OECD/IEA, 20135. For gas-fired power plants, a 10 °C change in ambient temperature can lead to a decrease of over 6% in netelectric power for a combined-cycle plant (Ponce Arrieta and Silva ora, 2005). According to Jochem and Schade(2009), transmission losses could increase by 0.7% up to 2050 in Europe, while Sathaye, et al. (2013) find thatelectricity losses in substations could increase by up to 3.6% in California by 2100.Chapter 3 | Managing climate risks to the energy sector 95

ut there will be significant regional variations, with increased generation potentialin some regions and reducons in others. The impact of climate change is parcularlyimportant in countries that rely heavily on hydropower for electricity generaon, such asBrail and Canada. Hamadudu and illingtveit (2012) nd that, for a warming of about2 °C by 2050 compared with pre-industrial levels, hydropower output increases in Russia,the Nordic countries and Canada by up to 25%, while it decreases in southern Europeand Turkey. More northerly parts of an America, including northern Brail, are expectedto see hydropower output decrease, but it is expected to increase in southern Brail andParaguay. In the northwest United States, some analysis points to a 20% reducon inhydropower generaon by the 2080s (Marko and Cullen, 2008). Hydropower plants canadapt to climate change with structural measures, such as increasing the sie of reservoirs,modifying spillway capacies and adapng the number and types of turbines.How climate change will impact other renewable energy sources is less well understoodand subject to strong regional dierences. The eect on wind power resources is dicult toassess due to the complexity of represenng near-surface wind condions in global climatemodels (Pryor and Barthelmie, 2010). Also, it is not only the overall electricity generaonpotenal from wind which is impacted by climate change, but also seasonal paerns.For example, electricity generaon from wind power could decrease by up to 40% in thenorthwest of the United States in the summer (Sailor, Smith and Hart, 2008). The eectof climate change on the operaon and maintenance of wind turbines will depend on thefrequency of extreme wind speeds and the possible reduced occurrence of icing. Futureelectricity generaon from solar photovoltaics (P) depends not only on solar radiaonbut also on ambient temperature and, for regions at higher latudes, on snow cover. Fora level of warming similar to our New Policies Scenario, research suggests that electricitygeneraon from solar P could decrease by 6% in Nordic countries in 2100, relave to ascenario without climate change (Fenger, 2007).Biomass producon, including biofuels, is aected not only by an increase in averagetemperature, but also by changing rainfall paerns, the increase in the carbon-dioxide(CO 2 ) concentraon and extreme weather events, such as storms and drought. HigherCO 2 levels and a limited temperature increase can extend the growing season, but morefrequent extreme weather events or changes in precipitaon paerns can more than osetthese posive impacts. The impacts will vary by region and type of biomass. For example, itis expected that the producon of many biomass crops will increase in northern Europe butdecrease in southern Europe, with Spain parcularly vulnerable due to increased drought(Tuck, et al., 2006).Climate resilience© OECD/IEA, 2013Climate change and weather extremes directly aect energy supply in a number of waysand illustrate how migaon and adaptaon become inextricably linked. Strong acon toreduce greenhouse-gas emissions will reduce the need to invest in climate adaptaon butwill not eliminate it. Even a global average temperature rise of 2 °C is going to demand some96 World Energy Outlook | Special Report

adaptaon. Unless the resilience of our energy system to climate change is consideredmore explicitly, energy supply and transformaon will be exposed to greater physicalrisks, which will increase capital, maintenance and insurance costs, impair energy supplyreliability and accelerate the depreciaon and deterioraon of assets.The climate resilience of the energy system could be enhanced in a number of ways. Interms of overall preparedness and planning, emergency response and co-ordinaon planscan be developed that cover crical energy infrastructure that is vulnerable to the impactsof climate events, such as in response to storms, oods and droughts (NCADAC, 2013).Energy facilies can be relocated or hardened to such events, with addional redundancymeasures also built into their design. Addional peak power generaon capacity, back-upgeneraon capacity or distributed generaon can improve power sector resilience. Also,grid systems can be physically reinforced (strengthening overhead transmission lines orusing underground cables), intelligent controls can be introduced and networks can bedecentralised (liming the impact of system failures). In those regions vulnerable to waterscarcity, power plants can move towards recirculang, dry (air-cooled) or hybrid coolingsystems for power plants, as is happening in South Africa, or using non-freshwater supplies,as is the case in oil producon in parts of the Middle East. On the demand side, ero-energybuildings, demand-response capabilies, such as smart grids and generally improved levelsof energy eciency can help either reduce the likelihood of system failures caused bypower demand spikes or reduce the impact of a supply failure.1234Governments need to design and implement policy and regulatory frameworks thatencourage prudent adaptaon to the impacts of climate change and help to overcomebarriers across dierent sectors of the economy. There have recently been some encouragingdevelopments in this respect, with the European Commission publishing a strategyintended to make adaptaon a central consideraon in European Union sector policies(European Commission, 2013), and the US Presidents Council of Advisors on Science andTechnology stang that a primary goal of a naonal climate strategy should be to help theNaon prepare for impacts from climate change in ways that decrease the damage fromextreme weatherand ways that speed recovery from damage that nonetheless occurs(US PCAST, 2013). As governments encourage acon, the private sector needs to reecton how best to bring the risks and impacts of climate change into its investment decisionmaking,especially for crical or long-lived energy assets.© OECD/IEA, 2013Economic impact of climate policies on the energy sectorMoving on from the issue of how climate change itself will aect the energy sector, thissecon analyses the related queson of the impact on energy sector assets and revenuesof the adopon by governments of more or less stringent policies to avoid or limit climatechange. In considering the capital allocaon and revenue implicaons of the transion toa low-carbon energy system, it draws on the New Policies Scenario and the 450 Scenarioof the , which reect diering levels of climateacon (see Chapter 1, Box 1.3, for informaon on these scenarios or refer to forChapter 3 | Managing climate risks to the energy sector 97

full details). The policy changes in the New Policies Scenario, which include both climateand other policies, are those that the energy sector can already expect to deal with aspart of normal business: they have already been announced and many are already beingimplemented. The 450 Scenario goes much further, tracing a plausible trajectory to theinternaonal objecve of achieving a 2 °C climate goal and idenfying the associatedaddional policies. e focus on the 450 Scenario, rather than the 4-for-2 °C Scenariodiscussed in Chapter 2, primarily because the emphasis is on the longer-term outlook(whereas there is parcular emphasis in the 4-for-2 °C Scenario on the period to 2020) andthe longer me horion is more relevant to the measurement of the economic impact ofclimate policy on energy sector investment decisions.e start by examining, rst, the extent to which exisng proven fossil-fuel reserves areconsumed under dierent climate policy paths and, second, the extent to which ourcurrent energy infrastructure has already locked-in future carbon-dioxide emissions. ethen analyse, by sector, the impact of dierent climate policies on the future gross and netrevenues of power generaon, upstream oil and gas, and coal mining. hile recognisingthat many new developments can result in energy sector assets becoming stranded, suchas the impact of the shale gas boom on NG import terminals in the United States, weseek to analyse the parcular risks associated with stronger climate change policies. Forthis analysis, we dene stranded assets as those investments which have already beenmade but which, at some me prior to the end of their economic life (as assumed at theinvestment decision point), are no longer able to earn an economic return, as a result ofchanges in the market and regulatory environment brought about by climate policy. Thismight, for example, include power plants that are rered early because of new emissionsregulaons, or oil and gas elds that, though discovered, are not developed becauseclimate policies serve to suppress demand. In measuring the scale of the loss associatedwith these stranded assets, we do not include the energy producon or capital recovery upto the point the asset becomes uneconomic, but only the lost element aer this point. edo not seek to esmate here the impact that changes in assets or revenues could have onthe nancial valuaon of energy companies, which can be aected by a very broad rangeof factors.The energy sector has always devoted considerable resources to nding, and then provingup, fossil-fuel reserves in the expectaon that they will one day be commercialised. Theextent to which these reserves which can be regarded as carbon reserves, that is fossilfuelreserves expressed as CO 2 emissions when combusted are actually consumed andthe CO 2 emissions released diers by fuel and scenario, according to the nature andintensity of the climate policies adopted. In our 450 Scenario, more than two-thirds ofcurrent proven fossil-fuel reserves are not commercialised before 2050, unless carbon© OECD/IEA, 201398 World Energy Outlook | Special Report

capture and storage (CCS) is widely deployed. 6 More than 50% of the oil and gas reservesare developed and consumed, but only 20% of todays coal reserves, which are muchlarger (Figure 3.5). Of the total coal- and gas-related carbon reserves, 3% are consumedin CCS applicaons where the CO 2 emissions are stored underground. In our less stringentNew Policies Scenario, there is higher consumpon of fossil fuels but at the price of failingto achieve the 2 °C trajectory. Even in the absence of any further acon on climate change,not even those allowed for in the New Policies Scenario, around 60% of world coal reserveswould remain underground in 2050.Figure 3.5 Potential CO 2 emissions from fossil-fuel reserves and cumulativeemissions by scenario to 20501234Gt2 0001 6001 200800Addional emissions:If remaining reserves combustedNew Policies Scenario450 Scenario:Emissions captured and storedNet emissions4000Coal Oil GasThe prole of the exisng global energy infrastructure (including facilies under construcon)means that four-hs (550 gigatonnes Gt CO 2 ) of the total volume of CO 2 emissions that theenergy sector is allowed to emit under a 2 °C trajectory up to 2035 are already locked-in simplyby the assumpon that it will connue to operate over its normal economic life. Assuming nolarge shis in relave fuel prices or technological breakthroughs, the emissions expected tocome from this infrastructure could only be avoided if policies were introduced which had theeect of causing its premature rerement or costly refurbishment. Around half of the lockedinemissions originate from the power sector and 22% from industry, as the facilies in thesesectors typically have a long life. The share of power generaon in total locked-in emissionsis highest in India, at 60%, closely followed by China, Russia and the United States. In Indiaand China, this is because the electricity sector relies to a relavely large extent on recentlyinstalled coal-red power plants, which are set to remain in operaon for decades, while inthe United States, large (relavely old) coal power plants currently lock-in a considerable© OECD/IEA, 20136. Proven reserves are usually defined as discovered volumes having a 90% probability that they can beextracted profitably. Ultimately recoverable resources (not discussed here) are much larger and comprisecumulative production to date, proven reserves, reserves growth (the projected increase in reserves in knownfields) and undiscovered resources that are judged likely to be produced using current technology.Chapter 3 | Managing climate risks to the energy sector 99

volume of emissions. The share of locked-in emissions for industry in China is around 30%,twice the level of that in the European Union: Chinas industry is dominated by the iron andsteel and cement sub-sectors, which have a relavely young age prole, indicang connuedoperaon well into the future.The share of locked-in emissions from transport (9%) and buildings (6%) is lower, as thebulk of the energy-consuming infrastructure in these sectors typically does not remainoperaonal for more than around een years. In the United States, the transport sectorhas a relavely high share (18%) of total locked-in emissions, since transport is responsiblefor a relavely high proporon of overall energy-related CO 2 emissions. Buildings accountfor 15% of locked-in emissions in the European Union, the highest share of all regions, dueto the importance of space heang in Europes energy systems. Another 6% of locked-inemissions results from other forms of energy transformaon (mainly reneries, and oil andgas extracon), 4% from non-energy use (mainly petrochemical feedstock and lubricants)and 1% from agriculture (including eld machinery).In non-OECD countries, infrastructure that exists or is under construcon locks-in360 Gt CO 2 from 2011 to 2035, led by China, India, the Middle East and Russia, while,in OECD countries, the gure is 195 Gt CO 2, led by the United States and the EuropeanUnion (Figure 3.6). The outlook in non-OECD countries is mainly a consequence of theinfrastructure expansion that has taken place over the past decade and the amount thatis currently under construcon. However, the extent of the connuing rapid expansion ofenergy infrastructure in non-OECD countries presents an important window of opportunityto avoid further lock-in of emissions by adopng ecient, low-carbon installaons. Thechallenge and opportunity for OECD countries lies, rather, with the replacement strategyadopted for the large amount of ageing fossil-fuel based infrastructure that could berered, or its use lowered, over the next few decades.Figure 3.6 CO 2 emissions locked-in by energy infrastructure in place andunder construction in 2011 by region and sector through to 2035100%80%60%40%177 Gt 84 Gt 51 Gt 39 Gt 30 Gt29 Gt 18 GtOtherBuildingsTransportIndustryPower generaon20%© OECD/IEA, 2013ChinaUnitedStatesEuropeanUnionIndiaMiddleEastRussiaJapanNote: Other includes energy transformaon, non-energy-use and agriculture.100 World Energy Outlook | Special Report

In the power sector, gross revenues are made up of a combinaon of wholesale electricityrevenues (volume of electricity generaon mulplied by the wholesale electricity price)and support received from governments for renewables (volume of supported renewablesgeneraon mulplied by the level of support). holesale electricity revenues account forthe vast majority of gross revenues, with government support for renewables accounngfor only a small share in our scenarios (just over 6% of gross revenues in the 450 Scenarioand slightly less than this in the New Policies Scenario). Dierent market structures directlyaect gross revenues by establishing the way in which wholesale electricity prices areformed. In a liberalised market, the price is based on the short-run marginal costs of powergeneraon. In an integrated monopoly, wholesale prices largely reect the average costsof generaon (Box 3.3).1234Figure 3.7 Power sector gross revenues and operating costs by scenario,2012-2035Trillion dollars (2011)605040302010Fuel costsO&M costsCO2 costsFuel costsO&M costsCO2 costsGrossrevenuesNew Policies ScenarioNote: OM operaon and maintenance.Net revenuesplus depreciaonGrossrevenues450 ScenarioNet revenuesplus depreciaonOn a consistent basis across scenarios, gross revenues in the power sector (from 2012 to2035) are $1.3 trillion (in year-2011 dollars) higher in the 450 Scenario than in the NewPolicies Scenario (Figure 3.7). 7 The higher gross revenues result from a combinaon oflower electricity demand and higher electricity prices, with the laer eect proving slightlylarger. 8 Over the projecon period, total electricity generaon is nearly 60 000 Th, or8%, lower in the 450 Scenario than in the New Policies Scenario, a reducon equivalent to© OECD/IEA, 20137. In the calculations made in this section, aggregate gross revenues for existing and new power plants arecomparable across scenarios. However, net revenues are discussed separately for existing and new plants. This isbecause of data deficiencies. For existing plants, investment data is not available for all plants so we present netrevenues before accounting for depreciation. For new plants, investment costs are based on known assumptionsso we are able to present net revenues after accounting for depreciation costs.8. holesale electricity prices are calculated endogenously within our orld Energy Model. For moreinformation, see orld Energy Model Documentation: 2012 version at .Chapter 3 | Managing climate risks to the energy sector 101

almost three mes annual world generaon in 2010, but wholesale electricity prices are16% higher, on average, in 2035. The change in electricity prices results from a combinaonof lower fossil-fuel prices, higher overall CO 2 costs (higher and more widespread CO 2 pricesbut lower levels of CO 2 emissions in the 450 Scenario) and capacity addions that aremore capital intensive. Operaon and maintenance (OM) costs are similar across the twoscenarios, as the reducon in costs that comes from phasing out some fossil-fuel plantsare oset by increased reliance on technologies with higher maintenance costs per unit ofcapacity, such as CCS and nuclear.For all power generaon capacity, net revenues before accounng for depreciaon 9(net revenues plus depreciaon in Figure 3.7) are $4.3 trillion higher in the450 Scenario than in the New Policies Scenario. These revenues essenally provide forthe recovery of investment costs and a nancial return on investment. 10 Depreciaoncosts for new capacity are $1.4 trillion higher in the 450 Scenario than the New PoliciesScenario, as this scenario requires more generang capacity to be built to oset thelower ulisaon factor of many renewables compared to the fossil alternave andgenerally more capital-intensive technologies. This addional cost is more than osetby lower fuel and OM costs, which are $4.5 trillion (19%) lower than in the NewPolicies Scenario through to 2035. This is due to lower electricity demand, lower fossilfuelprices and a more marked transion to renewables and nuclear with low or nofuel costs. CO 2 costs are $1.5 trillion higher in the 450 Scenario, with prices reaching$95-120 per tonne in many regions in 2035. hile higher CO 2 prices increase wholesaleelectricity prices (and therefore consumer bills), this revenue can potenally be recycledback to consumers in ways that parally oset the economic impact of electricity pricerises, without compromising climate policy outcomes.For exisng power generaon capacity, net revenues before accounng for depreciaon 11 areat similar levels in the 450 Scenario and the New Policies Scenario, at around $15.6 trillion.In the 450 Scenario, net revenues increase by around $900 billion each for both exisngnuclear and renewables capacity (that receive the market price in liberalised markets),compared with the New Policies Scenario (Figure 3.8). This gain osets a similar loss innet revenues by fossil-fuel plants of $1.9 trillion. Coal power plants without CCS bear theburden of the relave revenue reducons in the 450 Scenario, as rising CO 2 costs andreduced operang hours outweigh the impact of lower fossil-fuel prices, and power plantswith higher emissions are more aected than those with lower emissions. Net revenuesfrom gas-red power plants increase slightly overall in the 450 Scenario, compared with theNew Policies Scenario, with higher revenues from more ecient power plants, and somecoal to gas substuon, more than oseng lower revenues from less ecient gas plants.© OECD/IEA, 20139. This equals gross revenues minus operation and maintenance costs, the cost of fuel inputs and payments forCO 2 emissions in markets with a carbon price.10. The weighted average cost of capital (ACC) is assumed to be 8% in OECD countries and 7% in non-OECDcountries for all technologies.11. Investment data relating to all existing power plants are not available, preventing a robust estimation ofdepreciation costs. The omission of these costs means that our calculation of “net” revenues is artificially highbut, as the same approach is adopted in all scenarios, it is reasonable to compare across them.102 World Energy Outlook | Special Report

Box 3.3 Implications of decarbonisation on power marketsMarket design reforms are currently envisaged in several countries where there is aconcern that liberalised markets might not be able to smulate sucient investmentin new capacity. In liberalised markets, the spot price typically reects the short-runmarginal costs of the last power plant called to respond to demand. The hourly orhalf-hourly price received by all power generators is typically set by the costs of themost expensive plant required to be dispatched during that period. This ensures thatthe price is sucient to cover the operang costs of all generators, but it may beinsucient to also cover investment costs (or encourage new investment).1234In some instances, generators earn addional revenues via government supportmechanisms intended to encourage the deployment of selected generaontechnologies. This is typically the case for some renewables technologies but may alsobe considered for other low-carbon technologies, such as fossil-fuel plants using CCS.As these technologies become more compeve, the level of support is expected tobe reduced (and ulmately phased out) for new capacity. The capacity addions thatresult from such support mechanisms tend to lower the wholesale market price andreduce ancipated operang hours of convenonal plants, exacerbang the problem ofrecovering investment in new plants, making it harder for pre-exisng plants to recovertheir investment costs and harder to aract investment in new capacity that does notbenet from such support. Moreover, the introducon of variable renewables requiressignicant amounts of exible capacity available to be dispatched to guarantee thereliability of supply. In order to ensure adequate generaon capacity, several countriesare considering the introducon of either a capacity payment mechanism or lookingat the possibility of allowing very high wholesale prices during mes of scarcity, forexample, periods when variable renewables are not operang. Capacity addionsthat benet from support mechanisms generally receive a signicant amount of theirrevenues from outside of the wholesale market, a guaranteed feed-in tari, andtherefore do not suer from this problem.© OECD/IEA, 2013hile the exisng frameworks of many liberalised markets will be able to encouragesignicant decarbonisaon of the power mix, they will struggle to deliver a majortransion towards a decarbonised world. Further changes to market designs are likelyto be needed. This is parcularly true for those markets that are expected to relyon high levels of variable renewables. This is because, in the absence of signicantamounts of storage capacity or smart grid measures (to shi demand away from peakmes), the variable nature of their supply means they may be unable to sell into themarket when prices are highest, liming their ability to recover their investment costsfrom the wholesale market. On the other hand, the low variable costs of these sourcesof generaon mean that, under exisng market structures, wholesale prices could bereduced to very low levels possibly below the levels needed to recover their owninvestment costs unless there is some form of addional compensaon. Improvingexisng market designs and developing new ones for compeve power systems willtherefore be an essenal feature of the transion towards a decarbonised world.Chapter 3 | Managing climate risks to the energy sector 103

Figure 3.8 Net revenues before accounting for depreciation for existingpower plants by scenario, 2012-2035Trillion dollars (2011)87654321New PoliciesScenario450 ScenarioNuclear Fossil fuels RenewablesFor new power generaon capacity, net revenues aer accounng for depreciaon are$3 trillion higher in the 450 Scenario than in the New Policies Scenario (Figure 3.9). Thegeneral shi towards a 2 °C goal is reflected in the relative change in net revenues from theNew Policies Scenario to the 450 Scenario, with renewables, nuclear and fossil-fuel plantsed with CCS enjoying higher revenues as climate policies strengthen. Net revenues fromnew renewables capacity are 55% higher in the 450 Scenario than in the New PoliciesScenario while the capacity addions over the projecon period are 46% higher. In the450 Scenario, new renewables capacity provides nearly two-thirds of all net revenues fromnew capacity in the power sector.Figure 3.9 Net revenues after accounting for depreciation and investmentsfor new power plants by scenario, 2012-2035Trillion dollars (2011)1412108642Net revenuesInvestments($2011, billion)NewPoliciesScenario450ScenarioRenewables 6 020 8 778Nuclear 942 1 513Fossil-fuel plants fied with CCS 216 950Fossil-fuel plants without CCS 2 507 1 803New PoliciesScenario450 Scenario© OECD/IEA, 2013The 450 Scenario sees nearly 1 400 G of addional renewables capacity in 2035, comparedwith the New Policies Scenario. Fossil-fuel power plants without CCS play a signicantlysmaller role in the power sector in the 450 Scenario, with two-hs less new capacity104 World Energy Outlook | Special Report

uilt than in the New Policies Scenario. Furthermore, higher carbon prices reduce the netrevenues of new fossil-fuel plants without CCS in the 450 Scenario. In contrast, new plantswith CCS see a marked increase in capacity and net revenues, reaching 570 gigawas (G)of installed capacity in 2035. Nuclear capacity addions increase by 60% in the 450 Scenarioand the net revenues for each unit of capacity are higher, on average, due to elevatedwholesale electricity prices.Combining the economic prospects for exisng and new power plants, net revenues (aeraccounng for depreciaon) for the power sector are $3 trillion higher in the 450 Scenariothan in the New Policies Scenario. Stated simply, nancial opportunies could improvein the 450 Scenario for power producers with a porolio of low-carbon technologies,including harnessing the benets of CCS as a form of asset protecon strategy.1234An important eect of the decarbonisaon of the power sector in the 450 Scenario is tocause many older, inecient, fossil-fuel plants to be either idled or rered before the end oftheir ancipated technical lifeme, 12 and some power generaon capacity addions underconstrucon to become uneconomic and be rered early, despite originally appearing tobe economically sound investments. An addional 2 300 G of fossil-fuel plants are eitherrered before the end of their technical lifeme (37%), idled (47%) or retroed with CCS(16%) in the 450 Scenario, compared with the New Policies Scenario (Figure 3.10). Most ofthe rered or idled plants do recover their investment cost, but they are in operaon forfewer years than in the New Policies Scenario. Older, inecient plants are rered early as CO 2costs render their operaons uneconomic, but their investment costs have been recovered.Idled power plants remain available and may occasionally run in periods of strong demand,when the economics allow. Some exisng, but relavely new, plants require addionalinvestment to retrot them with CCS, so that they can remain in operaon. Almost 50% ofthe plants rered or idled are inecient subcrical coal-red power plants, as rising CO 2prices make them uneconomic, squeeing them out of the market in many countries. Thisshare is signicantly higher (75%) in non-OECD Asia, where a large number of subcricalcoal plants have been installed in recent years. These plants alone account for more thanone-third of the 1 940 G of global capacity that is rered early or idled.In the 450 Scenario, around 2 000 G of new fossil-fuel plants are built globally tomeet rising demand and, in some cases, to replace old inecient plants that becomeuneconomic. Almost 30% of these new plants are ed with CCS, two-thirds of these asa retrong operaon, as the technology becomes more compeve at scale. Just underone-quarter of the ancipated new fossil-fuel plants are currently under construconand they may face dicules recovering their investment costs if they have not takenthe costs of decarbonisaon fully into account. A smaller, but sll signicant, number of© OECD/IEA, 201312. A power plant may, for example, have an economic lifetime of 30 years (the period over which it recoversits capital investment) but be capable technically of operating for longer, perhaps 50 years.Chapter 3 | Managing climate risks to the energy sector 105

new plants are idled: of this 260 G of capacity, 165 G are idled before repaying theirinvestment costs, resulng in an unrecovered sunk cost of around $120 billion, or about40% of the inial investment. The remaining 90 G of new power plants that are idledrecover their investment costs. Idled plants can sll be given new economic life, reducingeconomic losses, if, at some point, they are retroed with CCS. Such retrots would beexpected to apply to the most ecient plants where the investment case is strongest (CCSreduces power plant eciency in the order of 8-10%). The availability of CCS technology,not only for the construcon of new power plants but also for the retrong of exisngpower plants, is a key assumpon in our assessment of sunk costs, as the deployment ofCCS technology has yet to be fully commercialised, making this a key challenge for therealisaon of the 450 Scenario (see Chapter 2 secon on the relevance of CCS).Figure 3.10 World installed fossil-fuel power generation capacity in the450 Scenario relative to the New Policies ScenarioGW5 0004 0003 000Reduced capacity needsfrom the New PoliciesScenario to the450 ScenarioTotal idled capacityin the 450 Scenario2 0001 000Total fossil-fuelcapacity running inthe 450 Scenario2010 2015 2020 2025 2030 2035Installed capacity:New Policies Scenario totalReduced generaon needs450 Scenario totalAddional early rerementsExisng plants built unl 2011 sll in operaonNew fossil-fuel plants installed and in operaon: Fossil-fuel plants installed idled:Without CCSNew plants, with sunk costsRetrofied with CCSNew plants, investments repaidFied with CCSExisng plants© OECD/IEA, 2013The gross revenues of oil and gas companies are determined by two key factors: the levelof producon and the prevailing prices. In the 450 Scenario, oil and gas gross revenuesare more than $105 trillion from 2012 to 2035 (in year-2011 dollars), nearly three mes106 World Energy Outlook | Special Report

higher than the level of the last two decades, but lower than in the New Policies Scenariowhich is around $125 trillion (Figure 3.11). Oil accounts for 70% of the gross revenues andnatural gas for 30% in the 450 Scenario. Oil demand peaks before 2020 and then declines,while gas demand connues to increase through to 2035, ending 17% higher than 2011.Oil prices average $109 per barrel (in year-2011 dollars) in the 450 Scenario ($120 perbarrel in the New Policies Scenario), while the course of natural gas prices varies regionally:gas prices decline in Japan, remain broadly stable in Europe and increase in the UnitedStates (see Chapter 1, Table 1.1 for our price assumpons).1234Figure 3.11 Cumulative world oil and gas gross upstream revenues bycomponent and scenarioTrillion dollars (2011)140120100806040Net revenuesTaxes and royalesGas transportOperang costsDepreciaonexpenses20Gross revenues1988-2011New PoliciesScenario2012-2035450 Scenario2012-2035Notes: Tax and royalty rates can vary between scenarios but are kept constant for this comparison. In caseswhere producon is dominated by naonal oil companies, the denion of taxes is somewhat arbitraryhere we assume tax rates comparable to internaonal averages.Cumulave net revenues, i.e. gross revenues, minus operang costs, gas transport,taxes and royales and depreciaon expenses, are projected to be $47 trillion in the450 Scenario, their level in 2035 being lower than in the New Policies Scenario, but higherthan in 2011 (Figure 3.12). Net revenues from gas grow throughout the projecon period,mainly driven by increasing demand, while net revenues from oil increase inially but peakbefore 2020 and then start to decline, as demand and prices decrease. Net revenues overthe period are esmated to correspond to around a 25% return on capital. 13© OECD/IEA, 201313. Assuming international oil companies typically operate with 10% risk-free rates of return, and up to20% in regions carrying a risk premium, with the average return on capital number being boosted by thecontribution of national oil companies operating in low production cost areas (based on our conservativedefinition of tax rates).Chapter 3 | Managing climate risks to the energy sector 107

Figure 3.12 World upstream oil and gas net revenues in the 450 ScenarioTrillion dollars (2011)2.11.81.51.20.90.60.31990 2000 2005 2011 2015 2020 2025 2030 2035Note: The absence of comprehensive historical data means that net revenues for previous years areesmated by applying the same gross-to-net revenues rao that is used for future years.Upstream oil and gas assets can become stranded if exisng elds do not operate at as higha level as originally planned, if they need to be rered before the end of their economiclife, or if a eld at which exploraon costs have been incurred does not go into produconby 2035 (the end date of our calculaons). There is an important disncon between thoseoil and gas elds that are in producon today and exisng or new elds that, depending ondemand, might be developed (and therefore start producing) at some point before 2035.Over the period to 2035, the level of producon from oil and gas elds that are producingtoday is the same in the 450 Scenario and the New Policies Scenario, as their produconremains economically viable in both cases. The investment in many of these elds hasalready been recovered and their level of operaon largely depends on the opmaldepleon rate and the addional costs associated with connuing producon. Thus, thepolicies in the 450 Scenario do not introduce signicant new risks that currently producingoil or gas elds will be forced out of operaon.© OECD/IEA, 2013In the case of oil and gas elds that have yet to start producon, or have yet to be found, thelower level of demand in the 450 Scenario means that fewer of them jusfy the investmentto bring them into producon (or to nd them) before 2035 (Figure 3.13). This meansthat some elds those that have been found but are not brought into producon by2035 do not start to recover their exploraon costs in this meframe. Relave to the levelin the New Policies Scenario, the addional risk of stranding assets in the 450 Scenarioaects 5% of proven oil reserves and 6% of proven gas reserves, all of which have yet tobe developed. The economic burden of this is relavely limited as, in the case of elds yetto be developed, the main impact relates to exploraon costs (typically around 15% ofinvestment in a new eld) which are not recovered by 2035, at least some of which could108 World Energy Outlook | Special Report

e recovered in the longer term. In the case of elds yet to be found, avoided exploraonand development costs oset the lost potenal future revenue opportunity.Figure 3.13 Development of proven oil and gas reserves by scenario12100%80%60%OilDeveloped inNew PoliciesScenarioGasDeveloped inNew PoliciesScenarioUndevelopedreserves to 2035Reserves developed:Addional in NewPolicies Scenario450 ScenarioCurrently producing3440%20%Developed in450 ScenarioDeveloped in450 ScenarioUpstream oil and gas sector assets can become stranded for a range of reasons, of whichnew climate policies is just one, but our analysis suggests that a companies or countriesvulnerability to this specic risk may be greater if their asset base is more heavily weightedtowards those that are not yet developed and towards those that have the highest marginalproducon cost (unless its development is driven by broader factors, such as energy security).Over the lifeme of upstream oil and gas assets, their nancial value and economic viabilitymay be appraised oen, including when: a company compiles its accounts a company takesinvestment decisions related to the asset (such as whether to develop a eld, which is oentested under a range of cost-benet assumpons) and, ownership of the asset changes.These, and other, reasons may mean that the nancial impact of stranded assets is realisedrelavely gradually over me and across several pares.© OECD/IEA, 2013hile not analysed here in detail, there is also the possibility that assets furtherdownstream will become stranded, such as in rening, NG plants and transportaonnetworks. In the case of rening, over-capacity is already a familiar issue in some regionsand could worsen under a range of scenarios. As oil demand grows in the Middle East andAsia, these regions have started extensive programmes to expand their rening capacityboth to meet internal needs and supply the export market. ower ulisaon rates, orpermanent shut down, of rening capacity could result in other regions, such as inEurope and North America, where domesc demand is declining. ithin the next decadesome 2 mbd of rening capacity is expected to be idled due to lack of demand, largelyirrespecve of climate change policies. This will aect not only old and inecient plants,but also relavely complex facilies that are bypassed by the changing crude and producttrade ows (see the focus on reneries in the forthcoming ). Inthe case of transportaon systems, some regions have already built addional pipelinesChapter 3 | Managing climate risks to the energy sector 109

to establish new trade corridors and, given the very long lifeme of such infrastructure,it is possible that ulisaon rates would decrease in some areas in the 450 Scenario,increasing the risk of stranded assets.Revenues generated by the global coal industry are a funcon of the volumes sold to themarket and the prices received for the product. Coal prices vary not only by type (steam,coking and brown coal) but also by region, reecng coal quality, transport costs andinfrastructure constraints. Typically, prices for coking coal are markedly higher than thosefor steam coal, due to the relave scarcity of coking coal and the lack of substutes insteelmaking, jusfying higher mining costs. Brown coal is rarely traded internaonally andconsequently does not have an internaonal market price. Instead, brown coal is usuallycombusted in power staons close to the mine, with the cost of mining determining theprice. In 2010, global coal producon stood at around 5 125 million tonnes of coal equivalent(Mtce), generang for the coal industry gross revenues of around $430 billion (in year-2011dollars). In the 450 Scenario, which assumes intensied climate policy measures, globalcoal demand falls by 1.6% per year on average through to 2035 (compared to an averageincrease of 0.8% per year the New Policies Scenario), with the change in demand leadingto a pronounced price drop.Figure 3.14 Cumulative world coal gross revenues by component andscenario, 2012-2035DepreciaonNew PoliciesScenarioMining costsTransport and otherNet revenues450 Scenario2 4 6 8 10 12 14Trillion dollars (2011)© OECD/IEA, 2013Cumulave gross revenues from coal sales are projected to amount to $8.9 trillion in the450 Scenario and $13.1 trillion in the New Policies Scenario (Figure 3.14). Coal supply ischaracterised by its relavely high share of variable mining costs, such as labour, energy,mining materials and spare parts (around 60% of total costs). In the 450 Scenario, thevariable mining cost component amounts to $4.6 trillion over the projecon period,compared with $6.9 trillion in the New Policies Scenario. Coal is also oen hauled long110 World Energy Outlook | Special Report

distances using trucks, railways, river barges and ocean-going vessels. Cumulavetransport costs stand at $2.3 trillion in the 450 Scenario, compared with $3.2 trillion in theNew Policies Scenario. Coal mining is far less capital intensive than oil and gas producon,and therefore depreciaon is a relavely minor cost component, amounng to around$0.86 trillion in the 450 Scenario and $1.1 trillion in the New Policies Scenario.Net revenues dier substanally between coal variees, around two-thirds of the totalcoming from steam and brown coal (around 85% of global producon), with coking coalcontribung the remainder. This means around 15% of global coal producon earns arounda third of the industrys total net revenues. Cumulave net revenues are $0.87 trillionlower in the 450 Scenario compared with the New Policies Scenario. Nearly 55% of thisdierence can be aributed to a change in price, whereas slightly more than 45% resultsfrom volume change. hile the price eect is almost enrely borne by coking coal, thedemand eect aects mainly steam coal. The level of coking coal producon, a key inputin the steel industry, declines by a relavely small amount across the scenarios due tothe lack of a large-scale substute for it in this sector. In contrast, there are substutesfor coal in power generaon and industry, which see greater take-up of nuclear powerand renewables in the 450 Scenario. Although coking coal demand diers by a relavelysmall amount between the scenarios, prices for coking coal drop sharply for two mainreasons: rst, coking coal prices are much more sensive to demand changes than steamcoal prices second, low demand for steam coal allows high quality steam coal to be usedfor metallurgical purposes, which further depresses the price for coking coal.1234Due to the relavely low capital costs involved in coal mining, coal prices need be onlyslightly above variable costs in order to provide an adequate return on investment. Hence,the risk of incurring large-scale losses on sunk investments is low. Moreover, exploraoncosts, a classic stranded investment risk, are relavely minor in the coal industry. Reduceddemand and lower prices in the 450 Scenario do lead to the closure of the highest-costmines, for which decreasing market prices do not cover the variable costs of producon.These are usually old mines whose compeveness suers from deteriorang geologicalcondions, depleon of the lowest-cost resources and low producvity due to the scale ofthe operaon and inecient equipment. Such mines typically have already recovered theirinvestment expenditure. Although the danger of stranded assets is, accordingly, limitedfor the industry as a whole, individual players can sll incur substanal losses on sunkinvestment. This is parcularly true for recent investments in elds which also require thelarge-scale development of railway and handling infrastructure. In the 450 Scenario, coaloperators will generally be able to cover their variable costs, but sub-opmal ulisaonand depressed prices might result in losses on the underlying investment, highlighng thebenets of early acon to idenfy and migate such risks (Spotlight).© OECD/IEA, 2013Chapter 3 | Managing climate risks to the energy sector 111

S P O T L I G H TCan corporate strategies help mitigate climate policy risk?Our 450 Scenario projects an increase in global energy demand relave to today,emphasising that a low-carbon transion is likely to represent a shi in the nature ofopportunies within a growing energy market. Corporate strategies that successfullytake account of climate policy risk could represent a source of compeve advantage,while failure to do so could result in a companys business model being undermined.Broad, non-exclusive approaches to migang climate policy risk might include:Decarbonise: Invest in technologies that reduce the carbon reserves associated with anexisng asset porolio. Coal companies could invest in underground coal gasicaon,coal-to-gas, increased washing of coal (to improve eciency) or develop coalbedmethane assets. Oil companies could focus exploraon eorts more towards natural gasor invest in enhanced oil recovery ulising CO 2 (or use depleted reservoirs to store CO 2 ).Power companies could invest in CCS and evolve their porolio of generaon assetstowards low-carbon opons.Diversify: Invest in new assets to develop a more diversied porolio dilung the risksassociated with those that are carbon intensive. Many coal companies are acve inother forms of mining (the largest private sector mining companies generate 10%-30%of their revenues from coal). At mes, some oil and gas companies have also owneda porolio of renewable energy assets, such as wind power and biofuels. Geographicdiversicaon of assets can also migate the policy risk of a parcular market.Delegate: Take acons to transfer the risk onto other pares willing to accept it,potenally through price hedging instruments or long-term take-or-pay contracts. Pricehedging can ensure a fossil-fuel producer receives a parcular price for all or part ofits supply. A take-or-pay contract can provide a degree of certainty over the volume offossil fuels to be sold and the revenues to be received.Divest: Dispose of carbon-intensive assets, parcularly those that have higher costs ofproducon, as they are at greater risk of becoming uneconomic.Disregard: The alternave to the migaon opons above is to accept the risk as itis, together with the associated impacts should it occur. The nancial impact will,ulmately, fall upon shareholders. It is therefore notable that esmatedthat nearly three-quarters of global carbon reserves are held by government-ownedcompanies, i.e. owned by taxpayers.© OECD/IEA, 2013112 World Energy Outlook | Special Report

Implicaons of delayed aconOur 450 Scenario, which is consistent with a 50% chance of liming global temperatureincrease to 2 °C, assumes a growing intensity of co-ordinated acon against climate changefrom 2014 onwards. Our 4-for-2 °C Scenario (see Chapter 2) takes a slightly dierentapproach, focusing on naonal short-term acons which can keep the door to 2 °C openwithout adversely aecng economic growth in any given country, prior to new coordinatedinternaonal acon from 2020. Both scenarios depend upon early addionalacon to tackle climate change, in one form or another. But what if this early acon is notforthcoming e analyse here some of the implicaons if governments and the energysector were to delay taking stronger acon on climate change, connuing on the path ofour New Policies Scenario 14 unl 2019 and then having to take sharp correcve acon toget back onto a trajectory compable with a long-term global temperature increase of nomore than 2 °C. It is an illustrave case, essenally a “delayed” 450 Scenario, based on thehypothesis that, for a variety of possible reasons, a number of years could pass before asignicant new boost is given to naonal policies and low-carbon investment.Delaying acon on climate change inevitably makes the 2 °C ever more challenging toachieve. In a scenario where there is such a delay, energy-related CO 2 emissions would reach34.4 Gt in 2019 (as in our New Policies Scenario) but then need, to meet the 2 °C target, todecline even more rapidly aer this date, ending at 20.6 Gt in 2035 (Figure 3.15). In essence,the addional emissions in the period to 2020 result in an emissions reducon trajectorythereaer which is even more challenging than our 450 Scenario. The emissions reduconaer 2020 is driven by improvements in energy eciency (parcularly in the industry andservices sectors), even more rapid deployment of renewable energy technologies in thepower sector and widespread adopon of CCS. Energy eciency is rapidly increased inindustry by phasing out old and inecient facilies in energy-intensive industries, as wellas by introducing new ecient motor systems. Energy eciency in buildings is stepped upby replacing oil- and gas-red boilers for space and water heang by more ecient ones.In the power sector, addional ecient coal and gas power plants are introduced, withless-ecient plants being operated less or completely rered. The increase in electricitygeneraon from renewables comes mainly from wind power, but also from hydro, bioenergyand solar P. The key regions aected are China, the United States and India. As well, CCSis very rapidly deployed, with the power sector accounng for nearly 70% of all CCS-relatedemissions savings, industry for more than 25% and the transformaon sector for 5%.Delaying climate acon takes the world beyond the date, esmated to be 2017 in, at which then exisng energy infrastructure locks-in the enre remainingcarbon emissions budget to 2035. The result is that much more costly acons are requiredsubsequently to undo the lock-in eect, including the early rerement of assets, lowerulisaon or idling of carbon-intensive capacity and increased investment in CCS retrong.In short, delayed acon creates more stranded assets in the energy sector. In the power1234© OECD/IEA, 201314. The New Policies Scenario does include cautious implementation of national targets to reduce greenhousegasemissions communicated under the 2010 Cancun Agreements.Chapter 3 | Managing climate risks to the energy sector 113

sector, the delay results in the construcon of a greater number of new fossil-fuelled plantsup to 2020, around 185 G of capacity. As a result, 164 G of power capacity must beeither rered or idled (101 G collecvely), or retroed with CCS (63 G), between 2020and 2035. Developing countries are most exposed to these lock-in eects, as they buildtwo-thirds of the addional fossil-fuel plants constructed up to 2020, many of which areinecient coal plants. To compensate for emissions from this capacity, an extra 130 G ofplants in developing countries must be rered, idled or retroed with CCS aer 2020. Itfollows that, if governments are to stand by their commitment to limit the average rise inthe global temperature to no more than 2 °C, developing countries have the most to gainfrom moving towards clean energy investment more quickly and the most tolose from carbon lock-in. A swi move away from subcrical coal-red power plants, ashighlighted in the 4-for-2 °C Scenario in Chapter 2, is a step in this regard and will help tomeet subsequent goals at a lower cost.Figure 3.15 World energy-related CO 2 emissions abatement in a“delayed” 450 Scenario relative to the New Policies ScenarioGt39New Policies Scenario35312723“Delayed” 450 Scenario192010 2015 2020 2025 2030 2035CO 2 Abatement 2025 2035Demand 5% 5%End-use efficiency 27% 31%Power plant efficiency 11% 3%Fuel and technology 2% 2%switchRenewables 25% 26%Biofuels 5% 5%Nuclear 9% 9%CCS 15% 20%Total (Gt CO 2 ) 6.2 16.4Analysis of the enre energy system shows that delaying acon on climate change is a falseeconomy. Investments of around $1.5 trillion are avoided in the period to 2020, but anaddional $5 trillion of investments are required between 2020 and 2035 (Figure 3.16). 15Prior to 2020, investments are notably lower in buildings (around $0.55 trillion) andindustry (around $0.45 trillion). In buildings, the amount of retrot in exisng buildingsis signicantly scaled back, while in transport the sales of hybrid and electric cars arelower in the period before 2020. The industry sector avoids investments before 2020 byallowing inecient old infrastructure to connue to operate for a few more years, reducinginvestments in more ecient equipment. Aer 2020, $1.4 trillion of addional investment© OECD/IEA, 201315. Using a 5% discount rate, investment costs avoided prior to 2020 are $1.2 trillion, while additionalinvestments required after 2020 are $2.3 trillion. If a 10% discount rate is used, investment costs avoided before2020 are $0.95 trillion, while additional investments required after 2020 are $1.15 trillion.114 World Energy Outlook | Special Report

is required to retrot buildings across both OECD and non-OECD countries. In industry,addional investment of $1.3 trillion is required to nance large-scale replacement bynew equipment, including in furnaces, motors, kilns, steam crackers and boilers. From atechnology perspecve, early acon can increase the potenal for accelerated learning andreduced costs. However, delaying acon could leave open the possibility of breakthroughsthat surpass current technologies.Figure 3.16 Change in world cumulative energy investment by sector in a“delayed” 450 Scenario relative to the 450 Scenario1234Trillion dollars (2011)1.81.20.62020-20352012-2019Net change2012-2035642Trillion dollars (2011)00-0.6Transport Buildings Industry Power BiofuelsgeneraonTotal-2This analysis shows that, if the internaonal community is serious about acng to limit therise in global temperature to 2 °C, delaying further acon, even to the end of the currentdecade, would result in substanal addional costs in the energy sector. As reectedthroughout this report, it highlights the importance of addional migaon acon in theperiod prior to a new global climate agreement coming into eect, to avoid the waste ofcreang stranded assets.© OECD/IEA, 2013Chapter 3 | Managing climate risks to the energy sector 115

© OECD/IEA, 2013

Annex AUnits and conversion factorsThis annex provides general informaon on units, and conversion factors for energy unitsand currencies.UnitsCoal Mtce million tonnes of coal equivalentEmissions ppm parts per million (by volume)Gt CO 2 -eq gigatonnes of carbon-dioxide equivalent (using100-year global warming potenals GP fordierent greenhouse gases)kg CO 2 -eq kilogrammes of carbon-dioxide equivalentg CO 2 /km grammes of carbon dioxide per kilometreg CO 2 /kh grammes of carbon dioxide per kilowa-hourEnergy Mtoe million tonnes of oil equivalentMBtumillion Brish thermal unitsGcal gigacalorie (1 calorie x 10 9 )TJ terajoule (1 joule x 10 12 )khkilowa-hourMhmegawa-hourGhgigawa-hourThterawa-hourGas mcm million cubic metresbcmbillion cubic metrestcmtrillion cubic metresMass kg kilogramme (1 000 kg 1 tonne)kt kilotonnes (1 tonne x 10 3 )Mt million tonnes (1 tonne x 10 6 )Gt gigatonnes (1 tonne x 10 9 )© OECD/IEA, 2013Monetary $ million 1 US dollar x 10 6$ billion 1 US dollar x 10 9$ trillion 1 US dollar x 10 12Annex A | Units and conversion factors 117

Oil b/d barrels per daykb/dthousand barrels per daymb/dmillion barrels per daympgmiles per gallonPower wa (1 joule per second)k kilowa (1 a x 10 3 )M megawa (1 a x 10 6 )G gigawa (1 a x 10 9 )T terawa (1 a x 10 12 )Energy conversionsConvert to: TJ Gcal Mtoe MBtu GWhFrom: mulply by:TJ 1 238.8 2.388 x 10 -5 947.8 0.2778Gcal 4.1868 x 10 -3 1 10 -7 3.968 1.163 x 10 -3Mtoe 4.1868 x 10 4 10 7 1 3.968 x 10 7 11 630MBtu 1.0551 x 10 -3 0.252 2.52 x 10 -8 1 2.931 x 10 -4GWh 3.6 860 8.6 x 10 -5 3 412 1Currency conversionsExchange rates (2011) 1 US Dollar equals:Australian Dollar 0.97Brish Pound 0.62Canadian Dollar 0.99Chinese uan 6.47Euro 0.72Indian Rupee 46.26Japanese en 79.84orean on 1 107.81Russian Ruble 29.42© OECD/IEA, 2013118 World Energy Outlook | Special Report

Annex BReferencesChapter 1: Climate and energy trendsBNEF (Bloomberg New Energy Finance) (2013), Clean Energy Investment Trends, BNEF,London.Boden, T., G. Marland and R. Andres (2012), 2Emissions, Carbon Dioxide Informaon Analysis Center, Oak Ridge Naonal Laboratory, USDepartment of Energy, Oak Ridge, Tennessee.Burniaux, J. and J. Chateau (2008), “An Overview of the OECD EN-Linkages Model”, , No. 653, OECD, Paris.DECC (U Department of Energy and Climate Change) (2013), , DECC, London.Elen, M. den, M. Meinshausen and D. van uuren (2007), Mul-Gas Emission Envelopesto Meet Greenhouse Gas Concentraon Targets: Costs ersus Certainty of LimingTemperature Increase, Global Environmental Change, ol. 17, No. 2, Elsevier, pp. 260-280.European Commission (2012a), “The State of the European Carbon Market in 2012”,, COM(2012) 652nal, European Commission, Brussels. (2012b), “Informaon Provided on the Funconing of the EU Emissions Trading System,the olumes of Greenhouse Gas Emission Allowances Auconed and Freely Allocatedand the Impact on the Surplus of Allowances in the Period up to 2020”, , SD(2012) 234 nal, European Commission, Brussels.EEA (European Environment Agency) (2013), , , accessed 11 April 2013.Frankfurt School UNEP Collaborang Centre and Bloomberg New Energy Finance (2012),, Frankfurt School of Finance andManagement, Frankfurt, Germany.GEA (Global Energy Assessment) (2012), Global Energy Assessment – Toward a, Cambridge University Press and the Internaonal Instute for AppliedSystems Analysis, Cambridge, United ingdom and New ork, and Laxenburg, Austria.Global CCS Instute (2013), , , accessed 11 April 2013.© OECD/IEA, 2013GEC (Global ind Energy Council) (2013), , GEC, Brussels.Annex B | References 119

Hansen, J., M. Sato and R. Ruedy (2012), “Percepon of Climate Change”, , ol. 109, No. 37, pp. 2415-2423.IEA (Internaonal Energy Agency) (2011a), OECD/IEA,Paris. (2011b), ,OECD/IEA, Paris. (2012a), , OECD/IEA, Paris. (2012b), ,OECD/IEA, Paris. (2013a), , OECD/IEA, Paris. (2013b), “Gas to Coal Compeon in the US Power Sector”, ,OECD/IEA, Paris.IIASA (Internaonal Instute for Applied Systems Analysis) (2012), , IIASA, Laxenburg, Austria.IPCC (Intergovernmental Panel on Climate Change) (2013), , forthcoming.Lenton, T., et al. (2008), “Tipping Elements in the Earths Climate System”, , ol. 105, No. 6, ashington, DC,pp. 17861793.Meinshausen, M., et al. (2009), “Greenhouse Gas Emission Targets for Liming Globalarming to 2 °C”, , ol. 458, pp. 1158-1163.Meyssignac, B. and A. Caenave (2012), “Sea level: A Review of Present-Day and Recent-Past Changes and ariability”, , ol. 58, pp. 96-109.NASA (2013), , , accessed15 May 2013.OECD (Organisaon for Economic Co-operaon and Development) (2012), , OECD, Paris.Oliver, J., G. Janssens-Maenhout and J. Peters (2012), 2 Emissions, PBLNetherlands Environmental Assessment Agency and Joint Research Centre, The Hague,Netherlands and Ispra, Italy.© OECD/IEA, 2013Pakistan Ministry of Climate Change (2012), , Governmentof Pakistan, Islamabad.120 World Energy Outlook | Special Report

Point Carbon (2013), ,, accessed 21 March2013.Rogelj, J., M. Meinshausen and R. nu (2012), “Global arming Under Old and NewScenarios using IPCC Climate Sensivity Range Esmates”, , ol. 2,No. 6, Nature Publishing Group, London, pp. 248-253.Smith, J., et al. (2009), “Assessing Dangerous Climate Change through an Update of theIntergovernmental Panel on Climate Change Reasons for Concern”, , ol. 106, No. 11, pp. 41334137.1234Solar Energy Industries Associaon (2013), ,, accessed 11 April 2013.Tans, P. and R. eeling (2013), , Earth SystemResearch Laboratory, , accessed 11 April 2013.UBA (Umweltbundesamt/German Federal Environment Agency) (2013), , UBA, Dessau, Germany.UCAR (University Corporaon for Atmospheric Research) (2008), MAGICC/SCHENGENersion 5.3.2, UCAR, Boulder, CO, United States.UNEP (United Naons Environment Programme) (2012), ,UNEP, Nairobi.Chapter 2: Energy policies to keep the 2 °C target aliveAlvare, R., et al. (2012), “Greater Focus Needed on Methane Leakage from Natural GasInfrastructure”, , ol.109, No. 17, ashington, DC, pp. 6435-6440.Chateau, J. and B. Magné (2013), “Economic Implicaons of the IEA Ecient orldScenario”, , OECD, Paris.Goenka, D. and S. Gukunda (2013), , Conservaon Acon Trust, Urban Emissions andGreenpeace, Mumbai.IEA (Internaonal Energy Agency) (2011a), , OECD/IEA, Paris. (2011b), ,OECD/IEA, Paris. (2011c), , OECD/IEA, Paris.© OECD/IEA, 2013 (2012a), , OECD/IEA, Paris. (2012b), , OECD/IEA, Paris.Annex B | References 121B

(2012c), ,OECD/IEA, Paris. (2012d), , OECD/IEA, Paris. (2013a), , OECD/IEA, Paris. (2013b), , OECD/IEA, Paris.IIP (Instute for Industrial Producvity) (2011), , IIP, ashington, DC.IMF (Internaonal Monetary Fund) (2012), , IMF, ashington, DC. (2013), , IMF, ashington, DC.IPCC (Intergovernmental Panel on Climate Change) (2005), , Contribuon of orking Group III of the IPCC, Cambridge University Press,Cambridge, United ingdom. (2007), “Climate Change 2007: The Physical Science Basis”, Contribution of orkingGroup I to the , Cambridge University Press, Nework.Lenton, T., et al. (2008), “Tipping Elements in the Earths Climate System”, , ol. 105, No. 6, ashington, DC,pp. 17861793.Lim, S., et al. (2012), “A Comparave Risk Assessment of Burden of Disease and InjuryAributable to 67 Risk Factors and Risk Factor Clusters in 21 Regions, 1990-2010: ASystemac Analysis for the Global Burden of Disease Study 2010”, The Lancet, ol. 380,No. 9859, pp. 2224-2260.OECD (Organisaon for Economic Co-operaon and Development) (2012), , OECD, Paris.Parry, I. and R. illiams (2011), “Moving US Climate Policy Forward: Are Carbon Taxes theOnly Good Alternave”, , No. 11-02, Resourcesfor the Future, ashington, DC.Stohl, A., et al. (2013), “hy Models Struggle to Capture Arcc Hae: The UnderesmatedRole of Gas Flaring and Domesc Combuson Emissions”, , ol. 13, European Geosciences Union, pp. 9567-9613.© OECD/IEA, 2013SwitchAsia (2013), , system.html, accessed 24 March 2013.122 World Energy Outlook | Special Report

UNEP (United Naons Environment Programme) (2011), , UNEP, Nairobi.UNEP and MO (orld Meteorological Organiaon) (2011), , UNEP and MO, Nairobi.US EPA (US Environmental Protecon Agency) (2012a), 2, US EPA, ashington, DC. (2012b), , US EPA, ashington, DC.1234 (2013), , US EPA,ashington, DC.Chapter 3: Managing climate risks to the energy sectorAcclimase (2009), “Building Business Resilience to Inevitable Climate Change”, Carbon, FTSE 350, Oxford, United ingdom.BNEF (Bloomberg New Energy Finance) (2012), BNEF, London. (2013), , BNEF, London.Durmaya, A. and O. Sogul (2006), “Inuence of Cooling ater Temperature on theEciency of a Pressuried ater Reactor Nuclear Power Plant”, Energy Research, ol. 30, No. 10, iley, pp. 799-810.EDF (2013), “Electricity Generation Lost in French Nuclear Power Plants due toEnvironmental Regulation in the Summer of 2003”, private communication with theInternational Energy Agency, 18 April.Elcock, D. and J. uiper, J. (2010), , Naonal Energy Technology Laboratory, Pisburgh, PA, United States.European Commission (2013), COM(2013) 216 final, European Commission, Brussels.Fenger, J. (2007), , Nordic Council of Ministers, Copenhagen.Foster, S. and D. Brayshaw (2013), ,Imperial College, London.© OECD/IEA, 2013Hamadudu, B. and A. illingtveit (2012), “Assessing Climate Change Impacts on GlobalHydropower”, Energies, ol. 5, MDPI AG, pp. 302-322.Annex B | References 123B

Hansen, J., M. Sato and R. Ruedy (2012), “Percepon of Climate Change”, , ol. 109, No. 37, pp. 14726-14727.Harsem, O., A. Eide and . Heen (2011), “Factors Inuencing Future Oil and Gas Prospectsin the Arcc”, , ol. 39, No. 12, Elsevier, pp. 8037-8045.Hempel, S., et al. (2013), “A Trend-Preserving Bias Correcon the ISI-MIP Approach”,, ol. 4, European Geosciences Union, pp. 49-92.Holden, P., et al. (2013), st, Open University, Milton eynes, Unitedingdom, forthcoming.IEA (Internaonal Energy Agency) (2012a), (2012a), , OECD/IEA,Paris. (2012b), , OECD/IEA, Paris. (2012c), ,OECD/IEA, Paris.IPCC (Intergovernmental Panel on Climate Change) (2007), “Climate Change 2007: ThePhysical Science Basis”, Contribuon of orking Group I to the , Cambridge University Press, Cambridge, United ingdom and New ork. (2012), “Managing the Risks of Extreme Events and Disasters to Advance ClimateChange Adaptaon”, , CambridgeUniversity Press, Cambridge, United ingdom and New ork.Jochem, E., and . Schade (2009), , Fraunhofer Instute for Systems and Innovaon Research,arlsruhe, Germany.arl, T., J. Melillo and T. Peterson (2009), , Cambridge University Press, New ork.nabb, R., J. Rhome and D. Brown (2005), , Miami, FL, United States.Marko, M. and A. Cullen (2008), “Impact of Climate Change on Pacic NorthwestHydropower”, Climac Change, ol. 87, Nos. 3-4, Springer, pp. 451-469.McNeil, M. and . Letschert (2007), “Future Air Condioning Energy Consumpon inDeveloping Countries and hat Can be Done about it: The Potenal of Eciency in theResidenal Sector”, , ECEEE, pp. 1311-1322.© OECD/IEA, 2013Munich RE (Mnchener Rckversicherungs-Gesellscha Akengesellscha) (2011),“orld Map of Natural Haards”, , Munich RE, Munich.124 World Energy Outlook | Special Report

NCADAC (US Naonal Climate Assessment and Development Advisory Commiee) (2013), US Global Change Research Program, ashington, DC.NOAA (US Naonal Oceanic and Atmospheric Administraon) (2013a), , , accessed 12 April 2013. (2013b), “Arcc Change Land Roads”, , accessed 22 March 2013.NOAA Naonal eather Service (2012), , NOAA,, accessed 11 April 2013.1234Peterson, T., P. Sto and S. Herring (2012), “”, , ol. 93, pp. 1041-1067.Ponce Arrieta, F. and E. Silva Lora (2005), “Influence of Ambient Temperature onCombined-Cycle Power Plant Performance”, , ol. 80, No. 3, Elsevier, pp.261-272.Pryor, S. and R. Barthelmie (2010). “Climate Change Impacts on ind Energy: A Review”,ol. 14, No. 1, Elsevier, pp. 430-437.Robine, J., et al. (2008), “Death Toll Exceeded 70,000 in Europe during the Summer of2003”, , ol. 331, No. 2, pp. 171-178.Sailor, D., M. Smith and M. Hart (2008), “Climate Change Implicaons for ind PowerResources in the Northwest United States”, Renewable Energy, ol. 33, No. 11, Elsevier,pp. 2393-2406.Sathaye, J., et al. (2013), “Esmang Impacts of arming Temperatures on CaliforniasElectricity System”, Global Environmental Change, ol. 23, No. 2, Elsevier, pp. 499-511.Sauer, A., P. lop and S. Agrawal (2010), , HSBC and orld Resources Instute (RI), RI,ashington, DC.Schaeer, R., et al. (2012), “Energy Sector ulnerability to Climate Change: A Review”,Energy, ol. 38, No. 1, Elsevier, pp. 1-12.Sto, P., D. Stone and M. Allen (2004), “Human Contribuon to the European Heatwaveof 2003”, , ol. 432, Nature Publishing Group, pp. 610-614.Strauss, B. and R. iemlinski (2012), Surging Seas: Sea Level Rise, Storms and Globalarmings Threat to the US Coast, Climate Central, Princeton, NJ, United States.© OECD/IEA, 2013Trenberth, . and J. Fasullo (2008), Energy Budgets of Atlanc Hurricanes and Changes from1970, , ol. 9, No. 9, American Geophysical Union andthe Geochemical Society, pp. 1-12.Annex B | References 125B

Tuck, G., et al. (2006), “The Potenal Distribuon of Bioenergy Crops in Europe UnderPresent and Future Climate”, Biomass and Bioenergy, ol. 30, No. 3, Elsevier, pp. 183-197.UNFCCC (United Naons Framework Convenon on Climate Change) (2013), Climate Change Acronyms, ,accessed 24 March 2013.UNPD (United Naons Populaon Division) (2012), , United Naons, New ork.USGS (United States Geological Survey) (2008), “Circum-Arcc Resources Appraisal:Esmates of Undiscovered Oil and Gas North of the Arcc Circle”, ,USGS, Boulder, CO, United States.US PCAST (US Presidents Council of Advisors on Science and Technology) (2013),Communicaon to the hite House, www.whitehouse.gov/sites/default/les/microsites/ostp/PCAST/pcastenergyandclimate3-22-13nal.pdf, accessed 3 May 2013.liet, M. van, et al. (2012), “ulnerability of US and European Electricity Supply to ClimateChange”, , ol. 2, Nature Publishing Group, pp. 676-681.ilbanks, T., et al. (2007), , US Department of Energy, ashington, DC.© OECD/IEA, 2013126 World Energy Outlook | Special Report

© OECD/IEA, 2013

WORLD ENERGY OUTLOOK 2013www.worldenergyoutlook.orgRELEASE: 12 NOVEMBERclass analysis, World Energy Outlook 2013 presentsThis year World Energy Outlook• Redrawing the energy-climate map: the short-within reach, and the extent to which low-carbon• Energy in Brazil: • Oil supply, demand and trade: • he implicans r ecnmic cmpeeness the changing energy map: what• he glal spread uncnennal gas supply• Energy trends in utheast sia© OECD/IEA, 2013The World Energy Outlookwww.worldenergyoutlook.org

TABLE OF CONTENTS1UNDERSTANDING THE SCENARIOS2ENERGY TRENDS TO 2035PART AGLOBALENERGYTRENDS345678NATURAL GAS MARKET OUTLOOKCOAL MARKET OUTLOOKENERGY EFFICIENCY OUTLOOKPOWER SECTOR OUTLOOKRENEWABLE ENERGY OUTLOOKENERGY AND GLOBAL COMPETITIVENESSPART BBRAZILENERGYOUTLOOK9101112THE BRAZILIAN ENERGY SECTOR TODAYBRAZIL DOMESTIC ENERGY PROSPECTSBRAZIL RESOURCES AND SUPPLY POTENTIALIMPLICATIONS OF BRAZIL’S ENERGY DEVELOPMENTPART COUTLOOKFOR OILMARKETS13141516FROM OIL RESOURCES TO RESERVESPROSPECTS FOR OIL SUPPLYPROSPECTS FOR OIL DEMANDOIL REFINING AND TRADE© OECD/IEA, 201317ANNEXESIMPLICATIONS FOR OIL MARKETS AND INVESTMENT

© OECD/IEA, 2013

OnlinebookshopBuy IEA publicationsonline:www.iea.org/booksInternational Energy Agency 9 rue de la Fédération 75739 Paris Cedex 15, FrancePDF versions availableat 20% discountA number of books printed before January 2012are now available free of charge in pdf formaton our websiteTel: +33 (0)1 40 57 66 90E-mail:books@iea.org© OECD/IEA, 2013

ulon reet te e o te E eretrt ut doe notneerly reet toe o nddul E eer ountre e Eke no rereenton or rrnty ere or led n reet o teulon ontent nludng t oletene or ury nd llnot e reonle or ny ue o or relne on te ulonThis document and any map included herein are without prejudice to the statuso or soereinty oer any territory to the delimitaon o internaonal ronersand boundaries and to the name of any territory, city or area. © OECD/IEA, 2013