Initiative No Rhyme or Reason arbon Tracker
Unreasonable projections in a world confronting climate change
Carbon Tracker Initiative, July 2016
About Carbon Tracker
The Carbon Tracker Initiative is a team of financial specialists making climate risk real in today’s financial markets. Our research to date on unburnable carbon and stranded assets has started a new debate on how to align the financial system with the energy transition to a low carbon future.
This report was authored by:
Rob Schuwerk and Luke Sussams.
Acknowledgements
This report draws heavily on internal memoranda prepared and commissioned by John Wunderlin on behalf of Carbon Tracker. The authors would also like to acknowledge the contributions of James Leaton, Mark Fulton, Tom Drew and Stefano Ambrogi for review and edits; Margherita Gagliardi for the design.
Disclaimer
Carbon Tracker is a non-profit company set-up to produce new thinking on climate risk. The organisation is funded by a range of European and American foundations. Carbon Tracker is not an investment adviser, and makes no representation regarding the advisability of investing in any particular company or investment fund or other vehicle. A decision to invest in any such investment fund or other entity should not be made in reliance on any of the statements set forth in this publication. While the organisations have obtained information believed to be reliable, they shall not be liable for any claims or losses of any nature in connection with information contained in this document, including but not limited to, lost profits or punitive or consequential damages. The information used to compile this report has been collected from a number of sources in the public domain and from Carbon Tracker licensors. Some of its content may be proprietary and belong to Carbon Tracker or its licensors. The information contained in this research report does not constitute an offer to sell securities or the solicitation of an offer to buy, or recommendation for investment in, any securities within any jurisdiction. The information is not intended as financial advice. This esearchr report provides general information only. The information and opinions constitute a judgment as at the date indicated and are subject to change without notice. The information may therefore not be accurate or current. The information and opinions contained in this report have been compiled or arrived at from sources believed to be reliable in good faith, but no representation or warranty, express or implied, is made by Carbon Tracker as to their accuracy, completeness or correctness and Carbon Tracker does also not warrant that the information is up to date. Contents
Executive Summary 5 Introduction 8 Part I: The Regulatory Landscape: The existing SEC regime for forward-looking disclosure and its shortcomings 14 Part II: Carbon Budgets 22 Part III: The US coal crash, revisited 31 Part IV: EIA’s Energy Model 38 Part V: Coal Company Disclosures in Context 50 Part VI: Conclusions 62
4 | www.carbontracker.org Executive Summary
EIA Reference Case is not a forecast This paper discusses why the EIA Reference Case is not an appropriate basis for coal sector financial planning and strategy, The EIA produces its Annual Energy Outlook (AEO) to provide and reviews how it has been utilized in industry SEC filings. With a range of scenarios, from one reflecting existing policies (the the 2016 Reference Case now reflecting the Clean Power Plan, Reference Case) to one that assumes economic conditions we ask whether the Paris emissions constraints will prove to be which drive accelerated coal plant retirements (Accelerated Coal the correct lead indicator? Retirements). The EIA is very clear that these are projections reflecting certain factors, NOT predictions or forecasts. Known trends The Reference Case is by definition a scenario in which no announced or new policies are enacted, existing policies with Companies must analyze “known trends, events, demands, sunset provisions do become ineffective, and technology commitments and uncertainties that are reasonably likely evolves at a steady pace. This is therefore a high case for coal to have a material effect on financial condition or operating demand, and there is every likelihood of a downside from this performance.” We would argue that increasing measures to demand level. Indeed, actual results over the past several years limit emissions from power generation, and the reducing costs have departed dramatically from Reference Case projections. of alternative power generation technologies are known trends. EIA modelling of emissions reduction targets more accurate The direction of travel of both these vectors is clear. On occasion, the EIA has built projections based on specific policy proposals, including the American Power Act and Two-part test for disclosure the American Clean Energy & Security Act. These models The SEC requires issuers to apply a two-part test to determine employed emissions constraints consistent with the laws’ whether trends and uncertainties must be disclosed. First, targets (which, in turn, were largely consistent with Obama’s it must determine whether the uncertainty is “reasonably Copenhagen pledges). These projections have proven far more likely” to occur. Second, if the uncertainty cannot affirmatively accurate over the past five years than those that assumed no be deemed “unlikely,” the corporation must then determine constraints even though the laws were never enacted. This whether the uncertainty would have a material effect on the strongly suggests that policy and technology are following in registrant’s financial condition, assuming it occurred. We would the wake of the overall emissions reductions targets and that argue that since 2010, a scenario of increased policy measures those targets should be a leading indicator to be considered by on pollution, lower costs of renewables and lower natural gas issuers in providing forward-looking information. That said, coal prices was not unlikely and would have a material impact on companies have consistently emphasized the Reference Case.
The Heart of the Matter | 5 the financial condition of US coal producers. Indeed with the Buying spree benefit of hindsight, it is clear that such a scenario came to A number of US coal mining companies undertook acquisitions pass and has had a material impact on the value and financial during the last peak of the market in 2011, taking on significant performance of US coal mining companies. Much of this decline debt in order to do so. This business strategy was predicated has been captured in EIA projections that incorporated climate on demand continuing to increase, and a strong market for targets. coal products. This misread of demand, and the failure of management to acknowledge that the reality was deviating Forward projections materially from its chosen projections saw an industry in denial. The SEC encourages the disclosure of “management’s The implications, however, were clear from sources, such as the projections of future economic performance that have a EIA, that were being cited in corporate reports. reasonable basis and are presented in an appropriate format.” It is also expected that the assumptions underlying these Too late projections should be transparent. This calls into question The 2015 10Ks1 filed in 2016 started to bring in references whether presenting a scenario where nothing changes as the to some of this information – but by this time it was too late most likely forecast to inform investors is reasonable. If it is not as many of the companies were already filing for bankruptcy made clear that this scenario also goes against the direction of protection. Arch Coal included a wider range of IEA and EIA travel indicated by government policy and market trends, then scenarios. Peabody has acknowledged examining the impact of this failure to disclose assumptions is significant. A business a range of policies that might impact coal and found that the strategy for a growth market with rising prices is very different financial impact could be material. However, in Peabody’s view, to one that is shrinking with falling prices. such analyses were too fraught with assumptions and therefore could not reasonably quantify potential future impact. All Range of scenarios scenarios contain assumptions to address future uncertainties, When selecting projections of future financial performance, including those that assume nothing changes. In the context of management is also advised it “should take care to assure that risk disclosures, the fact that there is uncertainty strengthens the choice of items projected is not susceptible to misleading the case for using assumptions that convey the extent of the inferences through selective projection of only favourable risk. items.” This would suggest that presenting a range of scenarios, rather than just the one that justified a BAU strategy would be an appropriate approach to disclosure. Reviews of the accuracy of previous projections would also inform whether an informative basis for the projections is being used. 1 A Form 10-K is an annual report required by the U.S. Securities and Exchange Commission (SEC), that gives a comprehensive summary of a company’s financial performance.
6 | www.carbontracker.org Dealing with uncertainty While companies often expressed uncertainty 1800.0 in forecasting the impact of climate policy, they readily provided third-party scenario 1600.0 information suggesting decades of continued demand for US coal—even though those 1400.0 projections discounted the aforementioned policy uncertainties to zero. From the 1200.0 point of view of a company managing risk, this seems a flawed approach; if an impact 1000.0 cannot be predicted precisely, it should not be assumed there is no impact at all. A more 800.0 reasonable approach would be to conduct a sensitivity analysis of a range of impacts 600.0 to demonstrate resilience under a number of scenarios. This does not require the 400.0 probability of the outcome to be determined. Coal demand (Million short tons) Notably, such scenarios were available from the same third-party sources cited by the 200.0 companies. A question for capital markets regulators is whether these concerns can be 0.0 1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 addressed through existing regulations or require new standards that ensure that the Actual 2014 AEO 2010 AEO key assumptions underlying projections are EIA Clean Power Plan EIA CPP Extended 2006 AEO reasonable and made available alongside the forecasts themselves. CPP = Clean Power Plan AEO = EIA Annual Energy Outlook
The Heart of the Matter | 7 Introduction
Desperately seeking transparency As the TCFD acknowledges, many existing disclosure regimes, including US SEC regulations, already require companies to disclose forward-looking information on known risks, trends and As the world turns to addressing climate change, markets uncertainties. These, too, are being revisited. In April, the SEC are turning to companies to better understand how they are issued a “Concept Release,” asking for comment on whether addressing the risks from an energy transition. the current disclosure regime is achieving its stated goals of improving efficiency, competition and capital formation. The Paris Agreement’s objective of limiting warming to “well- below” two degrees Celsius has fueled on-going efforts to better understand the implications of a transition to a low- Assessing and reporting risk carbon economy, including the impact on fossil fuel companies.
In parallel, the Financial Stability Board (FSB) has noted The foregoing efforts on transparency of climate risk dovetail increased demand for “decision-useful” information on how with efforts at the corporate level to better manage company- companies will be affected by climate risks (including transition wide risks and detail that management for investors. Risk risks) and sought to fill the gap. The FSB has launched a task assessment is central to business and oversight and risk force (TCFD) to consider such potential disclosures. The TCFD transparency is essential to proper capital allocation by notes that many companies have identified the risks, but goes securities markets. Proper risk assessment considers downside beyond risk identification to quantify and qualify those risks in scenarios—even those that may be seen as unlikely. a way that is useful to investment and engagement decision- making. As a consequence, the TCFD is considering whether In the wake of the global financial crisis, additional attention has enhanced disclosure of forward-looking information, stress been paid to risk oversight practices—the “governance” aspect tests, sensitivity analyses and industry-specific guidelines of risk—to ensure that risk mitigation is incorporated into would be appropriate. The TCFD will issue recommendations business planning and strategy. While most risk assessment in December and, while only voluntary, may prompt and mitigation happens at the management level, to varying reconsideration of whether existing disclosure regimes are degrees boards are engaged in the management of risks to the forcing decision-useful information on climate risks to market. enterprise as a whole.2 Increasingly, governance practices are
2 See, e.g., ANR 2013 Proxy Statement, at 24.
8 | www.carbontracker.org turning to enterprise risk management in response. Climate targets are carbon budgets For example, many companies have adopted the COSO framework for enterprise risk management to set firm-wide risk For many, the Paris Agreement marks the beginning of a appetite and tolerance to guide the company’s capital allocation potential sea change in international action on climate. While decisions accordingly.3 the level of governmental consensus may be new, the climate target is not. The two-degree climate target encapsulated Ideally, risk assessment, risk mitigation and risk oversight are in the Paris Agreement has been a mainstay in international conveyed to markets through corporate reporting in annual climate policy-making since the 2009 Copenhagen Accord. reports, proxy statements, and other channels. Especially important are material risks that may not be indicated by past Moreover, that climate policy target has been the basis for state results. Such forward-looking disclosure of principal risks and federal climate action targets, including President Obama’s and material trends—good or bad—is typically required by Copenhagen pledge to reduce emissions 17% from 2005 regulation.4 levels by 2020 and 83% from 2005 levels by 2050.5 In China, emissions intensity targets have been part of the country’s However, existing disclosure regimes focus primarily on risk five-year plans (although the relationship with the 2°C goal is identification, with less attention paid to how companies less clear).6 Both the climate targets and related emissions have integrated identified risks and trends into their business reductions pledges are important because they offer a means planning. The increasing focus on risk oversight further of quantifying a downside case for emissions and, therefore, suggests that companies should be discussing not just the fossil fuel demand.7 While these climate targets have long potential risks but how they are managing them. been in public view, very few companies have presented their consideration of those targets, or the demand implications, to Given these inquiries, we believe lessons might be learned from shareholders. examining past disclosures in the light of the purposes ascribed to existing disclosure requirements. Here, there is perhaps 5 http://unfccc.int/files/meetings/cop_15/copenhagen_accord/applica- no better starting point than the disclosures made by US coal tion/pdf/unitedstatescphaccord_app.1.pdf companies over the 2010-2015 period—a period of radical 6 China’s 12th Five-Year Plan was issued in March 2011, covering the pe- transformation in the industry, which went from assertions of a riod 2011-2015, and called for a 16% reduction in energy intensity per unit of GDP, a 17% reduction in carbon intensity per unit of GDP, and an increase of non-fossil coal super-cycle to market peak mergers to bankruptcy court energy use to 11.4% of total energy use. See http://www.c2es.org/international/ over the course of a few years. key-country-policies/china/energy-climate-goals-twelfth-five-year-plan 7 Central to Carbon Tracker’s research is the recognition that climate target imply climate budgets—greenhouse gases are long-lived so emissions accumulate in the atmosphere over time. Climate models analyse the probability and warming impact of a given level of accumulated emissions. The accumulated 3 http://www.coso.org/default.htm level of emissions associated with a given climate target is the effective “budget” for 4 See, e.g., 17 CFR 229.303. emissions.
No Rhyme or Reason | 9 Existential climate threat—role for not share those interests. Forcing that information to market is regulators? therefore a role for regulators.
Climate change—and actions to mitigate it—pose fundamental US coal a cautionary tale challenges to the fossil fuel sector. To even have a 50% chance of achieving the agreed upon limit of no more than two-degrees, fossil fuel use has to be substantially curtailed— The precipitous rise and fall in the valuations of US coal Carbon Tracker’s research suggests that roughly 80% of existing companies between 2009-2015 provides a natural experiment— coal reserves with listed companies need to remain unused.8 and cautionary tale—for companies, investors, and regulators Researchers as University College London have reached similar for how the assessment and mitigation of risk has been conclusions with respect to total global coal reserves.9 conveyed by companies to the markets. That period begins with buoyed optimism of a “coal-supercycle” led by robust This presents an existential risk to undiversified, pure-play demand from China and India and resilient demand for coal coal producers due to a dearth of commercially available and in the US. As of today, the industry is undergoing a wave environmentally sustainable ways of combusting coal. In theory, of bankruptcies amidst pronouncements by analysts that a pure-play coal company could seek to diversify, however the supercycle may never arrive—or, at least not in time to difficult operationally. save equity investors. Increasingly, the downturn is seen by mainstream analysts as structural, not cyclical.10 This is in However, a commonly held belief in capital markets is that stark contrast to a period of top of the market consolidations, diversification at the corporate level is less efficient than ensuing write-downs, and bankruptcies. what investors can achieve in their portfolios in liquid equities markets. Therefore, where trends such as climate change emerge that pose an existential threat to the business, a Where’s the assessment? company may be faced with the difficult choice of recognizing those trends and managing the decline of the business or This report takes a look-back at selected coal company resisting them aggressively. In either case, the investing disclosures in the 2010-2015 period, examining company usage community is best served by clarity on the extent of the risks of the US Energy Information Agency (EIA) Reference Case and the company’s assessment of the best path forward for the projections to forecast future coal demand in the context of business. A company electing to push against those trends may other company disclosures related to climate risks.
8 Carbon Tracker, Unburnable Carbon 2013: Wasted Capital and Stranded Assets, (2013). 10 See, e.g., Citi (2014) Global thermal coal: When cyclical supply met 9 See http://www.nature.com/nature/journal/v517/n7533/full/na- structural demand; Bernstein (2015) Asia strategy: Shouldn’t we all be dead by ture14016.html now?
10 | www.carbontracker.org The examination reveals that coal company corporate reports frequently treated the Reference Case as a forecast, notwithstanding the fact that the EIA merely projects past coal consumption to provide a policy baseline.11 Capital markets regulators should consider how such projections have been used, given their roles in ensuring transparency in corporate reporting.
As the EIA itself states, “EIA’s Annual Energy Outlook is a projection, not a prediction.”12
Box 1: EIA Projections Defined
EIA Projections As Defined by the EIA
The US Energy Information Administration provides a long-term outlook for energy supply, demand, and prices in its Annual Energy Outlook (AEO). This outlook is centered on the Reference case, which is not a prediction of what will happen, but rather a modeled projection of what might happen given certain assumptions and methodologies. In May 2016, EIA released an annotated summary of the AEO2016 Reference Case — which includes the Clean Power Plan — and a side case without the Clean Power Plan.
The AEO is developed using EIA’s National Modleing System (NEMS), an integrated model that seeks to reflect how various drivers to the US economy and the different compnents of energy production and use will interact to determine energy supply, demand, and prices.
The Reference case, which incorporates only existing laws and policies, is not intended to be a most likely prediction of the future. EIA’s approach to addressing the inherent uncertainty surrounding the country’s energy future is to develop multiple cases that reflect different sets of internally consistent assumptions about key sources of uncertainty such as future world oil prices, macroeconomic growth, energy resources, technology costs, and policies.
The alternative cases in the report show the model’s sensitivity to diffeent assumptions, which EIA updates and publishes each year. The assumptions that inform the modeling play an important role, and any results should be interpreted with the underlying assumptions in mind. These assumptions and results are discussed in the Annual Energy Outlook and EIA’s policy analysis reports such as the studies of removing restrictions associated with US crude oil exports or the then-proposed version of the Clean Power Plan.
11 http://www.eia.gov/todayinenergy/detail.cfm?id=26272 12 http://www.eia.gov/todayinenergy/detail.cfm?id=26272
No Rhyme or Reason | 11 Policy assumptions tend to be especially complicated. In the US, policies at the federal, state, and local levels can all affect energy supply, demand, and prices. Often these policies have timelines or other attributes that are revised by subsequent legislation or interpreted by executive departments. Some policies can interact in ways that are difficult to foresee. The AEO Reference case generally reflects current laws and regulations within the simplified but often expanding modeling parameters of NEMS.
Tecnological advances are also inherently difficult to anticipate. The Reference case typically projects technological evolutions rather than technological revolutions. For instance, energy production methods may become more effective and energy-consuming equipment may become more efficient in the Reference case, but EIA does not attempt to identify disruptive technologies on the timing of their availability and widespread adoption.
The energy sector has always been dynamic and undoubtedly will continue to change in the future. EIA has tried to make its projections as objective, reliable, and useful as possible. However, the projections in the AEO should be interpreted with a clear understanding of the assumptions that inform them and the limitations inherent to any modeling effort. Source: EIA
The future is by definition uncertain and therefore entails policy that might provide insight into the extent of the risk, assumptions and judgments that may subsequently be proven much less the implications of those targets for their business. wrong. Therefore, even if companies are required to take an objective look at the future in corporate reports, a failure to Without a discussion of such targets, it is not clear that offer accurate forecasts does not necessarily indicate a flawed companies were incorporating emissions reductions targets process. That said, many US coal companies have either of governments into their downside risk assessments or, if ignored climate policy targets completely or discounted them they were, what their conclusions were. Moreover, pairing due to an alleged lack of regulatory and policy commitment risk identification with third-party coal demand projections, to deliver them. Such a view offers little insight into what including those by the EIA, which assume that none of those might happen until it has occurred and may not meet investor risks will materialize, only adds confusion to such an assessment. expectations that companies manage capital in light of the risks. To be sure, forward-looking disclosure serves an important role In general terms, companies have discussed the potential for of conveying management’s views to the market—assuming future climate action and technology, but have professed an those demand projections (or forecasts, as the case may be) are inability to estimate the impact of those developments because reflective of management’s views, it is reasonable to provide of uncertainties with how specific policies and technology will them to investors. But it is equally clear that companies develop. However, few companies have identified key climate must sufficiently convey the downside risks and, increasingly, targets encapsulated in international agreements and national investors are asking not just how companies identify risks, but
12 | www.carbontracker.org also quantify them and identify what processes are in place to Even though most US coal companies were primarily exposed manage them. This leads to the question of whether the board to US markets, they have discounted concerns about those (ultimately responsible for corporate reporting) and company markets with the promise of a growing seaborne trade as a management were confident in the robustness of those justification for leveraging the company towards exposure to third party reports and if so, whether the management’s risk Asian markets, particularly China. There too, better disclosure assessment and the boards’ risk oversight responsibilities have of emissions reductions targets would have suggested functioned as intended. caution. Were these policy targets ignored, or considered and dismissed? And were investors fully aware of the extent of the risks? View to the future? The rapid decline of US coal equities has implications for the regulatory role in disclosures. Whether robust assessment One way of considering this issue is examining company of these risks can be developed within existing disclosure discussion of coal projections and forecasts in the period regulations or new regulations are required is somewhat of an leading up to the coal crash. Many companies referred open question; less uncertain are the benefits that delivering investors to EIA Reference Case scenarios as relevant indicators such disclosure would have for the markets in allocating capital of future demand. Such projections were never intended to be efficiently. predictions or the most likely outcome by the EIA, though they appear to be presented that way in the reports. This raises the In Part I, we review the US regulatory landscape as it related question of how accurately those scenarios were represented. to forward-looking disclosure of trends and risks. In Part II, we examine previously announced emissions reduction targets in In addition, the EIA produced other modelling which did the US and China and their carbon budget implications. Part consider potential changes in policy and which suggested a far III examines the precipitous decline of US coal companies in different trajectory for US coal. This included EIA modelling the aftermath of top-of-the market mergers. In Part IV, we from 2010 and 2015 that incorporated the emissions reductions compare the EIA’s Reference Case to other policy analysis cases targets of, respectively, the American Power Act (APA), the developed by the EIA that deploy carbon budget restrains American Clean Energy and Security Act (ACESA) (Waxman- within the NEMS model that undergirds EIA’s analysis. Even Markey) and the Clean Power Plan (CPP). Those demand though some of the policies identified in those scenarios have scenarios that have imposed demand constraints based on not been deployed, they have in fact been better predictors emissions reduction targets have proven to be far more of coal demand and consumption. In Part V, we focus on the prescient over the relevant period, yet were largely un-noted evolution of actual US coal company disclosures during the by companies until after the downside was priced in the equity 2010-2015 period. Finally, in Part VI, we make some concluding markets. remarks on how these risks can be better conveyed.
No Rhyme or Reason | 13 Part I
The Regulatory Landscape: The SEC regime for forward-looking disclosure
14 | www.carbontracker.org Already required? Risk factor disclosure is forward-looking and focused on the “most significant” factors that pose a threat to the business. The SEC requires companies to disclose forward-looking The discussion of those factors must be “concise,”18 requiring a risks13 and trends and uncertainties14 that might materially plain language statement of the relevant risk. Given the focus impact results. For all its merits, the SEC’s existing forward- on concision and clarity, SEC staff have frequently admonished looking disclosure regime may not have delivered on the most issuers for including language that mitigates or minimizes the salient information needed in the climate change context— stated risks.19 This means that companies should not identify notwithstanding the unusual but salutary decision by the SEC risks, on the one hand, while dismissing them, on the other. to publish guidance on how its existing regulations pertained to Few if any companies have downplayed climate risks directly, climate risks.15 Correcting this will require a renewed focus on but many have supplied investors with robust projections of Management Discussion and Analysis (MD&A) or consideration coal demand that do not account for climate-related risks. of new regulations aimed at forcing companies to analyse the implications of policy targets for their business models. “Risk Factor” disclosures may provide an overview of risks but not the type of information that can provide decision-useful Below, we describe the core elements of forward-looking information to investors. Indeed, many commentators20 and disclosures and their shortcomings: SEC Chair Mary Jo White21 have asked whether such disclosure is actually useful to investors or merely serves the purposes of shielding companies from litigation. More useful would Item 503(c) risk factor disclosure be information providing an estimate of the downside risk— information that, under certain circumstances, is required by the “Management, Discussion and Analysis” disclosure element, The purpose of Item 503(c) risk factor disclosure is to discuss discussed below. “the most significant factors that make the offering speculative or risky”.16 The aim of the disclosure is to make clear how a particular risk might, in the future, impact a given issuer. In the climate change context, SEC guidance states that “risk factor disclosure should clearly state the risk and specify how 18 Id. the particular risk affects the particular registrant; registrants 19 SEC Staff Comment, Mar. 4, 2009. Available at: https://www.sec.gov/ should not present risks that could apply to any issuer or any divisions/corpfin/guidance/cfsmallcompanyregistration.htm 17 20 See, e.g., Pillsbury, Winthrop, Shaw, Pittman LLP has identified risk offering.” (Emphasis added). factor disclosure as the “cheapest form of insurance” against potential future 13 17 CFR 229.503 litigation. Available at: http://www.pillsburylaw.com/siteFiles/Publica- 14 17 CFR 229.303. tions/77EA643CE089DDA568EFF79F0A35F681.pdf 15 75 Fed. Reg. 6294 Feb. 8, 2010 (“2010 Guidance”). 21 Speech, “The Path Forward on Disclosure,” SEC Chair Mary Jo 16 17 CFR 229.503. White, Oct. 15, 2013. Available at: https://www.sec.gov/News/Speech/Detail/ 17 2010 Guidance. Speech/1370539878806; IRRC report.
The Heart of the Matter | 15 Item 303: Management, discussion and Two-part test analysis The SEC requires issuers to apply a two-part test to determine whether a trends or uncertainty must be disclosed. First, it Item 303 requires disclosure of the trends and uncertainties must determine whether the uncertainty is “reasonably likely” reasonably likely to make operational results vary, giving to occur. The Commission has explained that “reasonably insight as to how past performance may not be repeated in likely” is a lower disclosure threshold than “more likely than 22 the future. Companies must analyze “known trends, events, not” but more likely than “remote.”28 As “reasonably likely” demands, commitments and uncertainties that are reasonably scenarios may have a less than 50% probability of occurring, likely to have a material effect on financial condition or issuers should be disclosing the impact of even some outcomes 23 operating performance.” Trends and uncertainties include that they believe likely will likely not occur. “legal, technological, political and scientific” developments that might indirectly impact a company by, for example, Figure 1: Evaluating likelihood lowering demand for its products.24 SEC guidance indicates that management’s review of potential material trends should extend to non-financial information and even external developments necessary to assess the impact on the company.25 remote reasonably more likely The SEC indicated in its 2010 Guidance that physical and likely than not financial impacts of climate change and mitigation efforts may < < be material to companies and, like any other material trend or uncertainty, should be disclosed.26 Issuers are clearly required to assess enacted legislation and regulation in determining whether to issue MD&A on it; the same standard applies to international accords relating to climate change.27
22 17 CFR Parts 211, 231 and 241, SEC Release Nos. 33-8350; 34-48960 (Dec. 29, 2003). 23 75 Fed.Reg. 6294 (Feb. 8, 2010). 24 75 Fed.Reg. 6294 (Feb. 8, 2010). 28 Commission Statement About Management’s Discussion and Analysis of 25 75 Fed. Reg. 6294-5 (Feb. 8, 2010). Financial Condition and Results of Operations, Release Nos. 33-8056 and 34-45321 26 75 Fed.Reg. 6294 (Feb. 8, 2010). (January 25, 2002); See also In the Matter of Bank of Am. Corp., File No 3-16028 27 75 Fed.Reg. at 6296 (Feb. 8, 2010). Release No. 72888 (Aug. 21, 2014).
16 | www.carbontracker.org Second, if the uncertainty cannot affirmatively be deemed “unlikely,” the issuer must then determine whether the uncertainty would have a material effect on the registrant’s financial condition, assuming it occurred.29 The SEC Guidance states that the materiality of possible future developments “will depend at any given time upon a balancing of both the indicated probability that the event will occur and the anticipated magnitude of the event in light of the totality of the company activity.’’30
Figure 2: MD&A Analysis
Remotely No disclosure likely obligation
Is the Uncertain trend or uncertainty No disclosure reasonably No obligation likely to Assuming the trend occur? Reasonably or uncertainty likely occurred, would it be material? Quantify the impact Obligation to Yes if reasonably disclose available More likely than not
29 Id., at 6295. 30 75 Fed.Reg. 6294 (Feb. 8, 2010) (quoting Texas Gulf Sulfur Co., 401 F. 2d 833 (2d Cir. 1968) at 849.
No Rhyme or Reason | 17 In practice, this means that if management believes that How material is the risk? there is a reasonable likelihood that an adverse event might materially impact the company’s finances, that event must As discussed in greater detail below, climate targets typically be disclosed. Pending legislation, regulations and treaties identify (or imply) emissions reductions levels that, based upon related to climate change are, under the statute, considered certain assumptions, further imply reductions in demand for “known uncertainties” and require management to employ the particular fossil fuels. As a first order approach, the salient aforementioned two-part test.31 A similar requirement applies question for investors is not the degree to which those targets to legal, technological, political and scientific developments are achieved by technology or policy, but what the implications related to climate change, though precise content and timing of achieving them are for company results — they are, at the of the disclosure is left to management. What constitutes a very least, a fair proxy of the downside case. ‘trend’ is open to some debate. However, with respect to the time frame for the impact, the SEC has rejected the notion that The foregoing standards suggest that companies should be it should be limited to a fixed number of years in the future, considering government emissions reductions pledges as a providing instead that the “necessary time period will depend potential risk to the business—certainly from a risk management on a registrant’s particular circumstances and the particular perspective, investors would expect management to consider trend, event of uncertainty under consideration.”32 that governments will do what they say. Perhaps even more important than identifying the trends and In the context of the top of the market mergers that took uncertainties themselves, companies must analyze how these place in the coal industry, this raises the question of how will impact the company’s financial condition. This may then well informed investors were of the extent of the risks that require disclosure. Such disclosure may often be qualitative, companies were facing. However, few companies have but quantitative disclosure may be required if “reasonably identified climate targets much less their implications, and the available.”33 The Commission has made clear that disclosure of SEC has not sought to compel such disclosure. known trends and uncertainties is not optional.34 This may be for several reasons. First, while management must objectively assess known trends and uncertainties, it 31 Id. may render a judgment on the “likelihood” of those trends 32 75 Fed.Reg. at 6296 (Feb. 8, 2010). coming to fruition. Certainly such a judgment cannot be made 33 17 CFR Parts 211, 231 and 241, SEC Release Nos. 33-8350; 34-48960 (Dec. 29, 2003). for self-serving reasons, but this still leaves some room for 34 “Not all forward-looking information falls within the realm of optional companies to determine that it is unlikely, including up until disclosure. In particular, material forward-looking information regarding known such disclosure is of little informational value to investors. material trends and uncertainties is required to be disclosed as part of the required discussion of those matters and the analysis of their effects.” 17 CFR Parts 211, 231 and 241, SEC Release Nos. 33-8350; 34-48960 (Dec. 29, 2003).
18 | www.carbontracker.org Yet, the two-degree target, now being reconsidered in light of Reasonable projections encouraged the more stringent goal of limiting warming to 1.5 degrees, has been a centerpiece of international climate policy since 2009. Companies are not always required to supply the market with The significant threat posed by climate change, the amount of their internal forecasts, but many deem it useful to provide attention it has engendered at all levels of government, and forward-looking information in other sections of their corporate perhaps most importantly, the ability to imply potential demand reports. The SEC encourages the disclosure of “management’s destruction from emissions reduction targets all suggest that projections of future economic performance that have a it would be within the range of known trends that companies reasonable basis and are presented in an appropriate format,”35 should consider and disclose to investors. (emphasis added) and has issued guidelines for companies offering such projections. In light of many company risk factor Second, companies may contend that while they believe such a disclosures, one would question whether the use of Reference transition likely, they cannot reasonably predict how the future Case projections, which assume a substantial build out of coal may unfold, particularly given the long time-frame considered. infrastructure under a “business-as-usual” trajectory, would be But testing against an emissions reduction trajectory need not reasonable. require a prediction of all policy and technology uncertainties; the two-degree climate target offers sufficient clarity of Although the guidelines appear to apply only to projections objective to frame disclosure of the extent of the risks in formulated by management, the purpose of the guidelines is to a general but meaningful way. But very few companies provide a framework for transparent presentation of forward- endeavored to provide this analysis to investors. looking information, recognizing that such information may be useful for investors considering future performance of the And while it is the case that no company can be asked to stock. Consistent with that purpose, the guidance should apply predict the future, it is clear that many companies offered with equal force to third party projections, such as the EIA forward-looking projections to investors that relied upon the projections, whenever management chooses to include them in critical assumption of no climate policy over the relevant period. annual reports. Several elements of the guidance are relevant A key question is whether equivalent disclosure of the risks here: should also be required. First, while projections typically focus on revenues, net income and earnings, the SEC notes that the guidelines are not limited to such projections. This suggests that demand projections for a company’s core market might similarly be subject to the guidelines.
35 17 CFR 229.10(b).
No Rhyme or Reason | 19 Second, the inferences drawn from forecasts are important, Figure 3: SEC Guidance on management’s projections and the SEC cautions that management “should take care to assure that the choice of items projected is not susceptible to misleading inferences through selective projection of only 36 favourable items,” and that management should caution Guidance applies to any projections. investors against putting undue emphasis on management’s 1 assessment. 37
Third, the SEC notes that assumptions are a critical part of forecasts and that investors’ understanding is “enhanced by Should avoid misleading inferences from disclosure of the assumptions which in management’s opinion 2 selective use of favourable data. are most significant to the projections or are the key factors upon which the financial results of the enterprise depend….”38 It therefore “encourages disclosure of assumptions in a manner 39 that will provide a framework for analysis of the projection.” 3 Critical assumptions should be disclosed. Fourth, the guidelines encourage management to provide information on the accuracy of past projections when compared to actual results in order to “highlight for investors the most significant risk and profit-sensitive areas in a business Consider comparing past and present operation.”40 4 projections to reveal forecasting accuracy. Fifth, the guidelines remind registrants of their obligation to notify investors of changes, favourable and unfavourable, which give management reason to know that “previously disclosed projections no longer have a reasonable basis.”41 Notify investors when prior projections no 5 longer have a reasonable basis.
36 17 CFR 229.10(b)(2). 37 17 CFR 229.10(b)(3). 38 Id. 39 Id. 40 17 CFR 229.10(b)(3)(ii). 41 17 CFR 229.10(b)(3)(iii).
20 | www.carbontracker.org Climate risk material, but business as usual? Applying the SEC guidelines on the use of EIA projections in company reports suggests that few companies considered the use of EIA projections in light of these guidelines. As discussed Many coal companies have included EIA Reference Case below, only one company indicated that the projections were projections in the Item 101 “Business Description” portion contingent upon a substantial build out of new coal-fired power of the corporate report. Such disclosures are substantively plants—an assumption that proved increasingly untenable over related to the purposes of the risk factor and MD&A but lie the period reviewed and, in any event, presented a one-sided outside those sections (and therefore are governed by the assumption about the future development (or lack thereof) of “Business Description” regulations). Although disclosures of climate policy and regulation. Moreover, the reports generally such forecasts are not required, there is still a duty to ensure did not offer management’s views on the accuracy of the EIA that such information is truthful, accurate and complete if it projection anywhere in the corporate reports. Finally, even is made.42 As noted above, the Reference Case is a modelled though the EIA Reference Case failed to even reflect the projection that assumes that only existing laws and policies direction of travel for US coal markets, companies continued are maintained over the projection period. Presumably the to utilize those numbers without identifying the significant gap incorporation of the Reference Case into corporate reports between actuals and projections. means they are generally reflective of company beliefs, though that is not entirely clear.43 This raises the dual questions of how those disclosures interrelate with company recognition of the material risks of climate change (discussed in Part III) and, more generally, how the company was assessing the risk from efforts to mitigate climate change.
42 17 CFR 240.12b-20. 43 In analyst presentations companies do often cite sources such as Wood- Mackenzie, Doyle Trading Consultants and their own internal analyses, disclosures to investors are largely dominated by reference to the EIA.
No Rhyme or Reason | 21 Part II
Carbon Budgets
22 | www.carbontracker.org The two-degree carbon budget In 2014, the Fifth Assessment Report (AR5) of the The 2009 Copenhagen Accord included a non-binding pledge Intergovernmental Panel on Climate Change (IPCC) concluded by governments to limit warming to no more than two-degrees that to limit climate change to 2°C, a total of 1,000 GtCO can above pre-industrial levels.44 Under the Copenhagen Accord, 2 be emitted by 2100.48 “Annex I Parties commit to implement individually or jointly According to the International Energy Agency (IEA), the total the quantified economy-wide emissions targets for 2020.”45 CO embedded in just the world’s coal, oil and gas reserves The Cancun Agreements formalized these targets by deciding, 2 is equal to 2,860 GtCO .49 This simple comparison gives an “Developed countries should submit annual greenhouse gas 2 indication of the extent to which fossil fuels will have to be left inventories and inventory reports and biennial reports on their in the ground to achieve acceptable levels of climate change. progress in achieving emission reductions, including information on mitigation actions to achieve their quantified economy-wide This 2°C climate target has subsequently been incorporated emission targets and emission reductions achieved.”46 In 2011 into a chain of UNFCCC international agreements and accords at Durban the parties established a formal reporting and review culminating in the recent Paris Agreement in 2015, which saw structure for reporting emissions.47 each participating nation submit national targets to keep global average temperature to “well below 2˚C” above pre-industrial The salience of Copenhagen was the incorporation of a climate levels.50 target that, in turn, implied a significant reduction in emissions when compared to business-as-usual. The implied emissions reductions provide a signpost for potential impact on fossil fuel demand and therefore a means of quantifying the risks — however they might materialize.
44 http://unfccc.int/resource/docs/2009/cop15/eng/11a01.pdf 45 Copenhagen Accord, ¶4. Available at http://unfccc.int/resource/ 48 https://www.ipcc.ch/news_and_events/docs/COP20/LCAHLD.pdf docs/2009/cop15/eng/l07.pdf. 49 IEA World Energy Outlook 2012. 46 Cancun Agreements, ¶40. The targets are also discussed in ¶¶36, 44, 50 The current Nationally Determined Contributions (NDCs) do not yet and 46. provide a clear regulatory pathway to the 2°C goal, but the Paris Agreement con- 47 FCCC/CP/2011/9/Add.1 ¶¶ 23-31, Annex II: Modalities and procedures tains a ratchet mechanism for increasing ambition over time. https://unfccc.int/ for international assessment and review. resource/docs/2015/cop21/eng/l09.pdf
No Rhyme or Reason | 23 The sum total of those targets currently exceed a two-degree compliant level of emissions, but countries have agreed to regularly update those targets:
Figure 4: Current INDCs exceed the emissions trajectory for 2°C 51 Annual CO2 emissions (GtCO2)
Source: IEA WEO 2015, IEA Energy and Climate Change 2015, CTI analysis 2016
51 The IEA INDC trajectory has been adjusted to the 2013 emissions level published in the 2015 World Energy Outlook. This represents the latest, most up to date data for this year.
24 | www.carbontracker.org Many organizations have analysed how these climate targets might affect demand for particular fossil fuels. Figure 5 shows how a 2°C trajectory (the 2014 IEA 450 Scenario) and a pathway that models implemented and announced policies (the 2014 IEA’s New Policies Scenario) would impact each fuel type differently. Evidently, demand for coal declines most dramatically, highlighting that, of all fossil fuels, coal is most at risk of being “stranded”.
Figure 5: Implications of Climate Targets for Fossil Fuels
Mtoe
Source: IEA World Energy Outlook, 2014
No Rhyme or Reason | 25 National carbon budgets The 2009 Waxman-Markey bill that passed the US House of Representatives but stalled in the Senate aimed at similar emissions reductions of 17% by 2020 and 42% by 2030, At the national level countries have sought to develop plans compared to a 2005 baseline. The recently proposed Clean compliant with these climate targets. The structure of these Power Plan, which only covers emissions from the power sector, plans varies—some are incorporated into national legislation, mandates reductions of 32% through 2030, compared to a others are coterminous with a country’s INDC pledge. 2005 baseline. Pursuant to the Paris Agreement, the US further For OECD countries, these plans are often articulated as specified a target of 26-28% emissions reductions by 2025, percentage emissions reductions levels from a baseline year. compared to a 2005 baseline. As such, they imply a fixed budget of available emissions going forward. For developing countries, these are often expressed Some commentators have noted that existing policy is so far as reductions in energy and carbon intensity of GDP in which insufficient to meet US climate targets.54 However, it is also case absolute emissions reductions are in part a function of GDP unlikely that these are the last regulatory or policy efforts growth. enacted to mitigate warming. As the Clean Air Act already vests the EPA with substantial authority to regulate GHG emissions, further action would likely not require the support US carbon budgets of the US Congress and could therefore be more easily implemented. Through the previous half decade, the US has been increasingly These climate targets imply meaningful reductions in coal proactive in its ambition to mitigate climate change, providing demand. For example, in 2015 the EIA modelled the impact clear policy signals for future emissions reductions. As of the CPP on coal demand. In Figure 6, we have compared part of the 2009 Copenhagen Accord, President Obama’s that modelling to selected Reference Case projections over administration pledged a reduction in US greenhouse gas the years. In each case, the EIA used the same underlying emissions of 17% by 2020, 42% by 2030 and 83% by 2050, economic model (NEMS), but in the CPP case it added the compared to a 2005 baseline.52 For the US, this is broadly emissions reductions constraints: consistent with a 2°C trajectory over that time period.53 This equates to 3.8% compound annual reductions per year.
52 http://www.washingtonpost.com/wp-dyn/content/arti- cle/2010/01/28/AR2010012803632.html 53 If you consider CO2 emissions as a proxy for GHG emissions more 54 Rhodium Group, “Taking Stock: Progress Toward Meeting US Climate broadly, the IEA 450 scenario determines that annual CO2 will need to fall by 65% Goals,” (Jan. 2016) (available at: http://rhg.com/wp-content/uploads/2016/01/ by 2040 on 2013 levels. IEA World Energy Outlook 2015. RHG_Taking_Stock_of_US_Climate_Goals_Jan28_2016.pdf)
26 | www.carbontracker.org Figure 6: Comparing across EIA scenarios to what actually transpired
1800.0 1.6% 1600.0
1400.0 0.4%
1200.0
1000.0 1.6%
800.0 -1.0%
600.0 -1.8% Coal demand (Million short tons)
400.0
200.0
0.0 1995 2000 2005 2010 2015 2020 2025 2030 2035 2040
Actual 2014 AEO 2010 AEO
EIA Clean Power Plan EIA CPP Extended 2006 AEO
Source: EIA Annual Energy Outlooks 2006-2014
No Rhyme or Reason | 27 Figure 6 demonstrates the drastically different results when CPP These targets inherently required China to meet the target policy action is considered—shifting the demand curve from to increase non-fossil energy supply by 11.4% over the same stable or modest growth to decline. period. Considering that China’s political environment typically ensures that the FYPs are met, the 12th FYP made clear that China carbon budgets China intended to move away from coal consumption that had grown so rapidly since the turn of the century.
The Chinese government has also established emissions Over the 2011-2015 period, coal consumption in China grew reduction targets. In its 12th Five-Year Plan (FYP) released in by 0.8% on average each year. However, this masks a more March 2011, the Chinese government announced goals for relevant story that total coal consumption inverted over this 5 the 2011-2015 period of reducing the energy intensity of GDP years, from 7.7% growth in 2011, to declines of 2.9% and 5% by 16%, carbon intensity of GDP by 17%, and 7% annual GDP in 2014 and 2015.56 China’s FYP gave US coal companies a growth. Table 1 shows that China actually exceeded its 12th clear indication that demand was likely to slow, if not decline, FYP targets through 2015: and it did, consistent with the governments’ reduction targets. It does not appear that companies disclosed these targets or incorporated them into their decision-making or forecasts. Table 1: Chinese FYP Targets and Actuals55
12th FYP target Actual Results GDP growth 7% pa 8.4% pa Energy intensity -16% -18.2% Carbon intensity -17% -20%
Sources: World Bank, WRI
55 http://www.c2es.org/international/key-country-policies/china/ener- 56 http://seekingalpha.com/article/3334735-new-china-coal-data-has-se- gy-climate-goals-twelfth-five-year-plan rious-implications-for-peabody-energy
28 | www.carbontracker.org China’s 13th FYP, released in March 2016, provided a pathway for further slowing of CO2 emissions growth, for the 2016-2020 period:
57 Figure 7: CO2 emissions implied by China’s 13th FYP, 2016-2020 (annual changes)
Source: CTI analysis 2016
57 Percentage change figures are annual compound growth estimates. Diagram sourced from: http://www.carbontracker.org/china-five-year-plan-coal-co2- emissions-renewables/
The energy-related targets in the 13th FYP are no weaker than those in the 12th FYP. As such, China’s transition away from coal Moreover, China’s Central Government are expected to consumption will not slow to 2020. The continuing restriction continue protecting domestic producers, as illustrated by the on CO2 emissions has led the IEA to state that 2013 might quality tariffs imposed on foreign coal imports. Exports to China have been peak coal in China.58 This would have catastrophic are already falling as a result. This is set to continue towards the implications for US coal companies that were hoping for a eventuality that China hits net-zero imports. robust seaborne market built around Chinese coal demand.
58 http://www.carbonbrief.org/iea-china-might-have-passed-peak-coal- in-2013
No Rhyme or Reason | 29 Box 2: The technology/policy nexus
The technology/policy nexus
The foregoing does not mean that long-term climate policy targets were a direct driver of the recent emissions reductions. But not considering those policy targets because of a lack of existing regulation or binding law to meet them is perhaps more perilous; it ignores the important ways in which policy targets influence both subsequent regulations and technology.
A top-down view—translating climate targets to emission budgets to implied demand reduction—may provide a clearer picture of trends than an assumption that business will continue as usual. This is because the policy/technology nexus is complex and self-reinforcing. Policy support has played a role in incentivizing the technology and driving initial cost reductions that has fuelled the coal-to-gas/ renewables switch in power. Regulation has further facilitated the switch by increasing compliance costs for dirtier fuels and increasing uncertainties for coal-fired generation. This has a feedback effect: as cleaner technologies become less expensive and dirtier fuels more expensive, the cost-benefit calculus for policymakers shift, allowing more significant policy action while also making the need for policy less relevant.
Precisely how policy and technology will evolve is difficult to predict. However, less difficult are the implications that a given emissions reduction target would have on fossil fuel demand (even though certain assumptions would still have to be made). This has implications for how companies consider and report the risks, suggesting that it may be useful to understand how company business plans compare to standardized demand scenarios consistent with the two-degree target.
30 | www.carbontracker.org Part III
The US coal crash, revisited
title title title title title title | 31 The last decade has seen drastic changes in the US coal sector. Global coal production and prices boomed in the immediate years after the 2007-2008 global financial crisis (GFC). Most coal companies characterized this as a resumption of a coal ‘supercycle’ and borrowed to finance acquisitions. For the most part, investors did not punish companies for taking on that additional risk.
Both regulatory and technology drivers have been influential in the shift from coal consumption to gas and renewables in the US. Figure 8 shows that environmental and pollution regulations emerged with great regularity in the US to suppress coal consumption. Simultaneously, huge technological breakthroughs were achieved that instigated the ‘dash for gas’ that caused gas prices to fall and undercut the costs of coal supply. The result has been particularly significant for investors, who have seen dramatic declines in coal equities.
Figure 8: US Coal Crash
Source: CTI, US Coal Crash, 2015
32 | www.carbontracker.org Companies leveraged up for the supercycle Investors are therefore set to lose out, with about $4bn of the outstanding $6.3bn debt held by unsecured bond investors and Companies appeared to believe that growth in developing the remaining controlled by a syndicate of banks.62 markets presented a great opportunity for US coal companies. Consequently, a number of the larger companies took on • Walter Energy: On April 1, 2011, Walter Energy substantial debt to acquire smaller competitors to gain market completed the acquisition of Western Coal Corporation, a share, focusing on seaborne thermal and metallurgical markets: Canadian company with mines located close to Canadian ports on the West coast. The acquisition cost Walter Energy $2.2bn • Alpha Natural Resources: Alpha Natural Resources in cash and $1.2bn in equities.63 The intention was that this acquired Massey Energy in the early 2010s, making Alpha the additional production and proximity to export terminals would second largest producer in the US and increasing its exposure position Walter Energy to compete for coal sales to Asia. When to metallurgical coal. The takeover cost $6.7bn, of which the additional demand failed to materialise, the capital structure of cash outlay was just $1bn. This was a major driver in Alpha the company came under stress. Prior to this takeover, Walter Natural Resource’s net debt level increasing from under $200m Energy’s net debt was zero, but the end of 3Q 2014, this had to $2.4bn at the end of 2011.59 Alpha Natural Resources filed increased to $3.2bn, largely as a result of the Western Coal for bankruptcy protection on 3rd August 2015. Corp. acquisition.
• Peabody Energy: In 2011, Peabody Energy acquired • Arch Coal: In May 2011, Arch Coal agreed to the Macarthur Coal, a metallurgical coal producer in Australia, to takeover of International Coal Group (ICG) for $3.4bn. Like increase its exposure to the Asian market. This was a very highly many of the other takeovers, this acquisition aimed to increase leveraged takeover, with $4.1bn of the $5.1bn total purchase Arch’s exposure to the metallurgical coal market, which it being funded through new loans and senior notes.60 By the expected to grow from 11mt to 14mt between 2011 and 2014.64 time of Peabody’s bankruptcy protection filing in April 2016, it The takeover was an all-cash transaction, which put Arch’s was servicing liabilities of $10.1 bn.61 balance sheet under strain when expected demand growth did not transpire.
62 http://www.afr.com/business/australian-takeover-drags-pea- body-near-bankruptcy-20160318-gnm7z1 59 http://marketrealist.com/2014/12/alpha-natural-resources-mas- 63 http://services.corporate-ir.net/SEC/Document.Service?id=P3Vyb- sey-energy-acquisition-got-messy/ D1hSFIwY0RvdkwyRndhUzUwWlc1cmQybDZZWEprTG1OdmJTOWtiM2R1Yk- 60 http://app.quotemedia.com/data/downloadFiling?webmas- c5aFpDNXdhSEEvWVdOMGFXOXVQVkJFUmlacGNHRm5aVDA0TnpZM- terId=101533&ref=7931317&type=PDF&symbol=BTU&companyName=Pea- k1UazNKbk4xWW5OcFpEMDFOdz09JnR5cGU9MiZmbj1XYWx0ZXJFbmVyZ3l- body+Energy+Corp.&formType=8-K&dateFiled=2011-11-30 JbmMucGRm 61 http://www.reuters.com/article/us-peabody-energy-bankrupt- 64 http://www.prnewswire.com/news-releases/arch-coal-to-acquire-icg- cy-idUSKCN0XA0E7 for-34-billion-121077269.html
No Rhyme or Reason | 33 Leveraged balance sheets spelled trouble
The leverage taken on at the beginning of a stunted supercycle ultimately weighed the industry down. Over the past 5 years, US coal company share prices have collapsed. Table 2 shows many lost in the region of 90% of their share value between 2010 and 2014. This resulted in approximately 30 companies going bankrupt. Those remaining are doing so by the skin of their teeth. Three of the four largest US coal producers are now in bankruptcy. The greatest losses have been borne by the companies that made market-peak acquisitions.
Table 2: US Coal Company Performance
Company Last 18m Last 5y Net debt EBITDA Peak Market Acquisitions Bankrupt? share price share price (USDm) (USDm) perf. (as of perf. (3 Jun. '16) (3 Jun. 2 Jun. '16) '16) Westmoreland Coal -73% -49% 1278.5 150.9 - No Hallador Energy -61% -56% 249.9 84.9 - No Alliance Resource Part. -64% -60% - No Consol Energy Inc. -55% -69% 3615.8 600.8 - No Cloud Peak Energy Inc. -77% -90% 491.5 71.9 Youngs Creek Mining Co. - $241m (Jun 2012) No Peabody Energy Corp. -99% -100% 7187.2 303.2 MacArthur Coal - $5.2bn (Nov 2011) Yes Arch Coal Inc -98% -100% 5487.3 158.3 Int’l Coal Grp. Inc - $3.4bn (Jun 2011) Yes Alpha Natural Resources Inc. -92% -100% 1342.5 -84.9 Massey Energy Co - $8.4bn (Jun 2011) Yes Rhino Resource Partners Lp -91% -99% 46.4 11 Elk Horn Coal Co. - $128m (Jun 2011) Yes Walter Energy Inc. -93% -100% 0.2 -178.3 Western Coal Corp. - $3.4bn (April 2011) Yes James River Coal Co. DELISTED Int’l Resource Part. LP - $519m (April 2011) Yes Foresight Energy Lp -90% N/A 1404.9 269.7 - No
34 | www.carbontracker.org Major market misread On the climate side, the overall direction of travel was signalled by President Obama through the Copenhagen pledges to cut US greenhouse gas emissions by 17% by 2020 against 2005 Each of the mergers above implied a commitment to expanding levels. this was followed by subsequent regulations.65 coal production in the 2010-2020 decade and beyond. Most Generally speaking the key factors were: companies sought exposure to seaborne markets. For thermal coal, companies made moves to lower cost basins (i.e. PRB) or • The emergence of cheap shale gas: Advances in areas closer to export markets (i.e., Australia) or sought greater horizontal drilling and slickwater hydraulic fracking techniques exposure to met coal supplies on the hope of selling those facilitated a large increase in natural gas production in the reserves in the seaborne market. These markets have failed to US. This boom resulted in the price of share gas falling by deliver. approximately 82% between the 2008 high and end of 2015. Consequently, the share of the US power market supplied by Supercyle, or boom and bust? gas increased by 11% over the same period, while that met by coal fell by 15%. Renewables captured the bulk of the remaining coal decline - refer Figure 9. It is evident that these acquisitions were a major misread of the markets, justified by a belief in high future demand in the • Environmental regulation: Since 2005, approximately metallurgical and thermal seaborne trade and resilient demand 7 key regulations have been implemented or are in the process in the US for legacy assets. Neither of these conditions held. of being implemented by the US Environmental Protection Agency (EPA), which have served to suppress coal demand. As Citi put it in mid-2013, the last five years has seen ‘the These policies are broadly aimed to mitigate the detrimental death bells ring for the commodity supercycle’. The upturn environmental and human health consequences of coal burning. that management teams were betting on has not materialised, The more influential policies include the Mercury and Air Toxics particularly in the US where structural headwinds have been Standards (MATS), air quality standards (NAAQS), the carbon building. The ensuing value destruction is the result of a number source performance standards (NSPS) and the Clean Power of structural drivers that were misread by the companies, many Plan. While harder to clearly attribute, these regulations may of which sought expansion at the wrong time. Those drivers have played a role in generating uncertainty about the long- were consistent with but not the exclusive product of climate- term future of coal and thus hampered the build out of new coal proofing the US energy supply — the economics of shale gas generation (an assumption built into the EIA Reference Case). played a central role.
65 http://www.washingtonpost.com/wp-dyn/content/arti- cle/2010/01/28/AR2010012803632.html
No Rhyme or Reason | 35 Figure 9: US Energy Mix by Energy Source
100% 2% 2% 3% 3% 4% 4% 5% 5% 6% 7% 7%
90%
80%
70%
19% 20% 22% 60% 22% 24% 24% 25% 31% 28% 28% 50% 33%
40%
30% US annual electricty demand ((TWh) 50% 49% 49% 48% 44% 45% 20% 42% 37% 39% 39% 33%
10%
0% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Other renewables Hydro Nuclear Natural gas (and other gases) Petroleum Coal
Source: CTI, US Coal Crash, 2015
36 | www.carbontracker.org Seaborne coal fails to arrive That the market collapsed so quickly raises the question of whether there was a clear understanding of the policy risks in targeted export markets whether those risks were conveyed Between 2008 and 2013, China and India alone accounted for with sufficient clarity and force. 98% of the increase in world seaborne coal trade. In this period, China’s coal import growth slowed and then declined by 10% Notwithstanding this collapse, coal companies seem to have in 2014 and 30% in 2015. This resulted in overall global coal pointed to long-term forecasts that offered light at the end of imports declining in 2014 and 2015, resulting in a collapse in the downturn. In many cases, those disclosures acknowledged prices. This was a very important driver in the collapse of US potential climate risks, but focused on scenarios, especially coal companies. those from EIA, that estimated resilient US coal demand and robust growth in the seaborne trade. These changes largely appear to have caught the coal industry by surprise. For example, in 2013, Peabody stated in their annual report that ‘coal demand is projected to expand by some 700 million tonnes, with China and India driving nearly 80% of this growth’.66 Still in 2016 Peabody expected to see ‘100mt of import growth to China and India’.67 Instead, 2013 transpired to be the year of peak demand for domestic and imported coal in China as a result of domestic air pollution measures, coal quality legislation on imports and incentives towards non-fossil energy sources.
66 Peabody Energy 2013 Annual Report, at 3. 67 https://mscusppegrs01.blob.core.windows.net/mmfiles/files/investors/ annual-reports/pe-ar2013.pdf
No Rhyme or Reason | 37 Part IV
EIA’s Energy Model
Credit: Kimon Berlin/flickr
38 | www.carbontracker.org Credit: Kimon Berlin/flickr The Energy Information Administration produces estimates of future energy demand using the National Energy Modelling System (NEMS), which takes a market-based approach based on existing regulations and standards, to model developments in energy markets and usage. NEMS includes a number of constituent modules that “represent each of the fuel supply markets, conversion sectors, and end- use consumption sectors of the energy system:”68
Figure 10: EIA National Energy Modelling System69
68 EIA, “Assumptions to the Annual Energy Outlook 2015” (Sept. 2015), at 1. 69 Id., at 3.
No Rhyme or Reason | 39 As with any model, NEMS relies on data, assumptions and key The ‘no new policies’ assumptions used in the Reference Case inputs which are updated annually. It produces alternative help inform policy makers of how markets might evolve absent scenarios by modifying key variables, such as oil prices and policy changes, and many policy makers use the EIA Reference supply and economic growth, to isolate the impact of these Case to demonstrate the need for additional policy action. The variables on energy markets. Reference Case is not intended to be a predictor of the likely actual development of energy and policy and, unsurprisingly, The EIA Reference Case assumes business as has proved to be a poor “forecast” of actual events. usual… …which is playing catch-up with actual outcomes The EIA’s Annual Energy Outlook focuses on a business-as-usual policy scenario—the EIA Reference Case—that “assumes that current laws and regulations affecting the energy sector are As we discuss in Part V, many US coal companies referred to largely unchanged throughout the projection period (including the Reference Case as if it were a “forecast.” Figure 11 shows the implication that laws that include sunset dates do, in fact, the EIA’s Reference Case projections for coal’s share of the become ineffective at the time of those sunset dates).”70 US power market from 2007 onwards as compared to actual results. It illustrates that although projections of coal’s demise in The Reference Case projection is a business-as-usual trend the US became gradually more significant, the Reference Case estimate, given known technology and technological and projections overestimated coal’s role. Over the past several demographic trends.” The EIA cautions that it should not years, the recent Reference Case scenarios have projected be considered a forecast of how things will unfold, since “[p] coal’s share of the US power market to decline, but only rojections by EIA are not statements of what will happen but of gradually. These projections were far more optimistic than the what might happen, given the assumptions and methodologies severe decline that has actually transpired. For example, the used for any particular scenario. Consequently, the Reference latest EIA data put coal’s share of power generation at 33% in Case results ‘should not be viewed in isolation’, but against 2015, but the 2015 EIA AEO Reference Case only projects coal’s 71 the other scenarios to gain perspective and a truer sense of the share of power generation fall just below 40% in 2040. This inherent uncertainties. The Reference Case does not analyse calls into question the coal company usage of the Reference the policy implications of climate targets at the national or local Case as a credible forecast of future demand. level and therefore does not incorporate emissions reductions targets. 71 In May of this year, the EIA released an early version of the AEO 2016. That version incorporates the CPP into the Reference Case and, significantly, pro- jects that coal will power just 28% of electricity generation in the US in 2040. This assumes, of course, that no additional measures after the CPP are implemented to 70 EIA AEO 2012, at 6. further reduce coal’s share.
40 | www.carbontracker.org Figure 11: Coal’s share of electricity generation As it assumes no new policy developments, the Reference Case does not incorporate the rapidly changing conditions for the US coal market from 60% +7 % change increased regulation and efforts to prevent the lock-in of coal intensive power generation. Of 7% course, the point of the Reference Case is not to 55% forecast the future, but to provide a projection -1 % change of existing policies. But this raises the question 50% -1% of the value of such a scenario in the context of a -1 % change corporate report, and specifically a discussion of -1% the potential risks to the business. 45% -5 % change Interestingly, when the EIA has modeled the -5% 40% evolution of energy consumption and generation -3% trends by embedding emissions reductions targets from stated policies—including policies 35% that were ultimately not enacted—it has produced coal demand scenarios much closer to Coal’s share of power gen. (%) share Coal’s 30% the actual trajectory than the Reference Case.
25% Other EIA scenarios In each Annual Energy Outlook the EIA also runs 20% dozens of scenarios aside from the Reference 2000 2005 2010 2015 2020 2025 2030 Case. In its most recent Outlook, this includes high and low variants on economic growth, oil Actual 2007 AEO 2009 AEO 2011 AEO 2013 AEO price, technology, renewables and fossil fuel costs and climate policy stringency. Though Source: EIA Annual Energy Outlook 2007-2016 the alternative cases were typically not referred to in company reports, some demonstrated a very different direction of travel. Figure 12 identifies the Climate Policies case, produced in 2009 and the Accelerated Coal Retirements Case, produced in 2014:
No Rhyme or Reason | 41 Figure 12: Comparison EIA Scenarios and Actuals, Coal Power Even though LW110 was still bullish compared to actual results Generation through 2013, it at least reflected the actual direction of travel.
Figure 12 also shows how other cases later 0.7% in the period examined would have shown a downward trajectory in coal generation as opposed to the stable demand projection in the Reference Case. In 2014, when 0.4% companies were able to inform their outlook -1% with the on-going crash, projections such as ‘accelerated coal retirements’ case were -1% overlooked in favour of the business-as-usual Reference Case. Indeed, by ignoring climate policy and technology risk, the Reference Case scenario represents something more
Coal power generation (Bn KwH) like a “best case scenario” for US coal rather than one built upon reasonable assumptions of future developments.
EIA’s top-down modelling constraint On occasion, policy makers ask the EIA to identify the energy market implications of proposed legislation and regulations. Source: EIA Annual Energy Outlook 2009 and 2014 Three recent examples include the August 2009 analysis of the American Clean Energy and Security Act (ACESA/ Waxman-Markey), the July 2010 analysis of the As seen above, the EIA’s strong climate policies scenario, 72 referred to as ‘LW110’, was most bearish on coal demand American Power Act (APA) , and an analysis of the Clean Power projecting -1.0% growth per annum in coal power generation Plan (CPP) in 2015. While the APA/ACESA bills were not (compared to 0.7% CAGR increase in coal power generation 72 The APA bill was very similar to Waxman-Markey, a bill that passed under the Reference Case). the US House of Representatives. The EIA analyzed both and the results were similar. Below, we have focused on Waxman-Markey.
42 | www.carbontracker.org enacted and the Clean Power Plan has yet to take effect, the Table 3: Comparison of Emissions Reduction Targets from 2005 75 projections from each are instructive since they required the Baseline EIA to integrate proposed emissions reductions targets into the NEMS model and thus represent an interesting, if temporally Year Copenhagen Waxman- Clean Power limited, comparison of the Reference Case to a NEMS-based Pledge Markey Plan analysis of emissions constraints and actual results. 2020 17% 17% -- The Waxman-Markey bill proposed an “economy-wide, GHG 2025 30% -- -- cap-and-trade system and critical complementary measures to help address climate change.”73 In line with President Obama’s 2030 42% 42% 32% Copenhagen pledge, the cap required a 17- percent reduction 2050 83% 83% -- in covered emissions by 2020, 42% by 2030 and an 83-percent reduction by 2050, all relative to a 2005 baseline, with targets By overlaying those emissions reductions targets and CPP policy that declined steadily for intermediate years.74 measures on the NEMS model, the EIA estimated the impact of 76 The CPP would establish standards to limit carbon emissions the CPP for coal demand. from existing power plants for the first time. The key emissions reduction target in the CPP is a 32% reduction by 2030 from 2005 levels—lower than the Copenhagen pledge but also only one of several policies directed at reducing emissions.
75 The CPP proposed rule contained a 30% reduction by 2030 and this level of emissions reduction but the Final Rule revised the goal upwards to 32%. See, https://www.whitehouse.gov/the-press-office/2015/08/03/fact-sheet-presi- 73 http://www.c2es.org/federal/congress/111/acesa dent-obama-announce-historic-carbon-pollution-standards. The EIA modeled the 74 See http://www.eia.gov/oiaf/servicerpt/hr2454/execsummary.html; proposed rule http://www.c2es.org/federal/congress/111/comparison-waxman-markey-ker- 76 With the exception of Arch, no companies we examined identified either ry-lieberman this analysis or any other estimate of reductions in coal demand driven by the CPP.
No Rhyme or Reason | 43 Imposing a top-down emissions target is more consistent with actual results
Figures 13 and 14 demonstrate that the EIA’s Reference Case was not an appropriate gauge for what transpired in the US coal market. Figure 13 demonstrates that the Reference Case has had to shift downwards year after year to reflect on-going reductions in coal demand and even still, suggests a flat or slightly rising increase in coal demand going forward:
Figure 13: EIA Reference Cases vs. Actual
1800.0 1.6% 1600.0
1400.0 0.4% 1200.0
1000.0 0.4%
800.0
600.0
Coal demand (Million short tons) 400.0
200.0
0.0 1995 2000 2005 2010 2015 2020 2025 2030 2035 2040
Actual 2014 AEO 2010 AEO 2006 AEO
Source: EIA Annual Energy Outlook 2006-2014
44 | www.carbontracker.org By contrast, Figure 14 shows that the EIA scenarios that modelled emissions reductions targets from Waxman-Markey—targets that were similar to what President Obama pledged to effectuate—have been far more accurate than the Reference Case, even though the bill was not enacted. Figure 14 demonstrates that EIA’s multi-scenario analysis of that bill more accurately reflected coal consumption since 2007 than the EIA Reference Case.
Figure 14: EIA ACESA Modelling 77 Coal consumption (QBtu)
Source: EIA, 2009, Energy Market and Economic Impacts of H.R. 2454, the American Clean Energy and Security Act of 2009
77 http://www.eia.gov/oiaf/servicerpt/hr2454/pdf/sroiaf(2009)05.pdf
No Rhyme or Reason | 45 The two scenarios representing the most drastic decline in US coal consumption - ‘no international offsets’ and ‘no international/limited’ - both reflect an environment in which the use of carbon offsets through credit trading is ‘severely limited by cost, regulation and/or slow progress in reaching 78 international agreements’. This means consumption of CO2- intensive fuels like coal must decline rather than be transferred to another company or region. This is tantamount to what happened in the market even absent the passage of detailed climate regulation.
This suggests that President Obama’s 2009 Copenhagen pledge to reduce US GHG emissions by 17% by 2020 (on 2005 levels) was a critical signal of future policy action. Nonetheless, most companies eschewed such projections for the Reference Case scenario. Importantly, the EIA’s analysis of the APA which reached similar conclusions to Waxman-Markey was published in August 2009, and therefore would have been available to all companies filing their 2009 10-Ks in early 2010 and thereafter.
Companies did not need to wait until COP21 in Paris to recognize the shift. Research from as early as 2012 was suggesting that the US was already on track to meet this emissions reduction target.79 Companies that accurately assessed the risks of these intentions to the coal market might have identified the alternative Waxman-Markey scenarios as being more reflective of what might transpire in the short-term than simply assuming business-as-usual.
78 http://www.eia.gov/oiaf/servicerpt/hr2454/pdf/sroiaf(2009)05.pdf 79 http://www.rff.org/files/sharepoint/WorkImages/Download/RFF- DP-12-48.pdf
46 | www.carbontracker.org The Clean Power Plan In 2015, the EIA published a similar analysis of the Clean Power Plan (CPP), which has been promulgated but stayed by the Supreme Court until the courts resolve legal challenges. EIA’s analysis of the CPP quantifies the impact it might have on coal consumption in the US:
Figure 15: Change in Electricity Generation by Fuel in Clean Power Plan Cases Relative to Baseline80
80 EIA CPP Analysis, at 32 (May 2015).
No Rhyme or Reason | 47 In Figure 16, we focus on three key EIA CPP scenarios compared to actual trends: (1) a base policy scenario (CPP), (2) a CPP policy extension scenario (CPPEXT), including extended emissions reductions of 45% below a 2005 baseline by 2040, and (3) a CPP scenario with new nuclear capacity benefiting from as favourable scoring under the CPP as other zero-carbon technologies (CPPNUC).81 These demonstrate that the CPP is largely an extension of current trends in declining coal demand:
Figure 16: Changes in CPP Scenarios82
Figure 16 shows a consensus that under each of these scenarios, coal consumption is substantially reduced compared to the EIA’s 2015 Reference Case.83 Figure 16 0.5% also demonstrates that restrictions imposed by the Clean Power Plan will likely carry through the existing -1.8% trend line of decreasing coal -2.1% demand. Coal consumption (bn kWh)
Source: EIA 2015 Analysis of the Impacts of the Clean Power Plan
81 https://www.eia.gov/analysis/requests/powerplants/cleanplan/pdf/powerplant.pdf 82 https://www.eia.gov/analysis/requests/powerplants/cleanplan/pdf/powerplant.pdf 83 The 2016 EIA Reference Case incorporates the CPP.
48 | www.carbontracker.org Mistaking the forest for the trees
The CPP was not announced until 2014 and its particularities could not have been easily foreseen. However, very few companies sought to inform investors of the magnitude of the risk it posed once that information was released. In 2015 it was run as an alternative scenario by the EIA and in 2016 forma part of the Reference Case. This shows how it takes time for laws to be put in place, but forward-looking companies would want to consider their implications before its too late. Furthermore, it is by no means clear that the CPP is the last word on climate policy; many commentators have noted that it does little more than codify the new “business-as-usual” trend of declining coal consumption. Indeed, between proposing and finalizing the rule, the Obama Administration strengthened the 2030 target by 9%. Other emissions reductions efforts to “ratchet” up ambition to meet the Paris Agreement’s two-degree goals may also be looming on the horizon. Given that the ACESA emissions forecasts led the AEO back in 2010, it is quite possible the new targets set in Paris could lead again.
The foregoing suggests that deploying a top-down view of emissions reductions targets on forward-looking models may be more reflective of policy trends than models that assume no action going forward. Here, with the many indicia of policy action on climate change already evident, the default for forward looking information in company reports should be to analyse the implications of climate targets rather than assume business-as-usual trends that continue to point away from actual results.
No Rhyme or Reason | 49 Part V
Coal Company Disclosures in Context
50 | www.carbontracker.org In Part V, we take a closer look at company disclosures in the Risks identified… annual corporate reports for the years 2010-2015. Annual filings generally acknowledged the materiality of several climate Our review of the risk factor disclosures by the coal companies related risks even as companies took on a substantial debt that pursued top of the market mergers show that that they to fund expansions into metallurgical and seaborne thermal included a discussion of climate risks that remained substantially markets. the same over the 2010-2015 period. The disclosures included: the potential for international treaties, national legislation, Certainly, such expansions suggest an expectation of growing climate and environmental regulation, renewable portfolio global demand (even whilst the Chinese government was standards, changes in financial standards for funding new pledging to disappoint), but did the focus on metallurgical and coal generating plants and cost competitive technologies to seaborne trade also signal a lack of belief in the US coal markets adversely impact coal markets. In general, companies further going forward? If so, this raises the question of how companies recognized that weakened demand for their products might saw trends in the US coal markets, especially emerging result in lower prices with material impact on results: emissions reduction targets.
The Heart of the Matter | 51 Table 5: Risk Factor Disclosures Made by Selected US Coal Companies in 2010-2015 10-Ks84
Peabody Arch ANR Cloud Peak Substitution Risk from: Renewables Cost Declines 2010-15 2010-15 2012-14 2010-2015 Natural Gas Competition 2010-15 2010-15 2010-14 2010-2015 Regulatory Risk from: UNFCCC 2010-15 2010-15 2010-14 2010-2015 CAA/EPA Regulation 2010-15 2010-15 2010-14 2010-2015
Federal Cap and Trade or Carbon Taxes 2010-15 2010-15 2010-14 2010-2015 State Utility Regulation/RPS 2010-15 2010-15 2010-14 2010-2015 Other Env’l Regulation and Enforcement 2010-15 2010-15 2010-14 2010-2015 Financing Risk to New Coal Plants 2014-15 2010-14 2015 Litigation Risk 2015 2010-2015
84 See table headlines: Peabody: Discussion of climate risks detailed in the “Business Description” of the 10-Ks section as well as “Risk Factors” section. Arch: Discussion of potential for additional regulation detailed in the “Business Description” of the 10-Ks section rather than “Risk Factors” section. ANR did not file a 2015 10-K as it had filed for bankruptcy. Noted litigation risk related to permitting. Company 10-K filings.
52 | www.carbontracker.org …But climate targets largely absent Arch’s disclosures are notable because they identify both the emissions target and the implied coal generation shrinkage— While companies identified the litany of risks from action on suggesting companies could provide valuable quantification climate change, very few identified the relevant international of the impact that climate policy would have on demand. and national emissions reductions targets themselves, nor their Unfortunately, the disclosure came long after Arch’s expansion potential impact. Some companies mentioned Copenhagen85, with the acquisition of ICG, and no similar discussion of but did not identify the key development of the two-degree President Obama’s pre-ICG merger Copenhagen pledge is target. Similarly, there was virtually no discussion of President contained in earlier disclosures.88 In any case, there is little Obama’s related goal of emissions reduction targets of up discussion of how the company intended to address such an to 83% by 2050 from a 2005 baseline, or the carbon budget impact. implications of those targets.86 While these long-term targets may not be immediately responsible for the recent declines in Arch’s 2015 10-K was submitted after the company filed for US coal markets, they more clearly reflect a long-term structural bankruptcy in January 2016.89 That filing contains an analysis problem rather than a cyclical ebb and flow of regulation. of future coal markets that considers a range of IEA and EIA scenarios for coal demand, including the IEA 450 Scenario and The exceptions suggest that a more robust discussion of the the EIA CPP Scenario, focusing on the key emissions reduction implications of emissions reductions targets on coal demand metric.90 There, Arch reveals that under the 450 Scenario, would have been useful. In its 2014 10-K, Arch identified the coal’s share of power generation would drop from 41% in 2013 Clean Power Plan goal of reducing emissions 30% of 2005 to just 12% in 2040—a substantial decline that would require levels by 2030, implying a 25% reduction in coal generation — a significant market rebalancing. The IEA has been publishing significant change: this scenario since 2010, and its inclusion here, but not in earlier reports, raises the question of how seriously the company “If the Clean Power Plan is passed as proposed, and withstands was considering this risk in the lead up to the June 2011 ICG legal challenges, it is projected to have an adverse impact on merger. the demand for coal nationally. Some studies estimate that the Clean Power Plan will reduce coal generation in the US by 25%.”87
85 See, e.g., ANR’s 2010 10-K, which notes the Copenhagen Agreement but emphasizes that it was considered a “modest” outcome by most observers. Id., 88 Arch’s 2012 10-K identifies the US emissions reductions target under the at 61-62; see also Cloud Peak 2010 10-K, noting agreement in principle with US Kyoto protocol—93% below 1990 levels—though it also notes that the US had not pledge constituting an emissions reduction of “at least 17% by 2020.” Id., at 26-27. ratified the Kyoto agreement. Arch 2012 10-K, at 27. 86 See, e.g., ANR 2010-2015 10-Ks. 89 Arch 2015 10-K (Mar. 15, 2016). 87 Arch 2014 10-K, at 28. 90 Arch 2015 10-K at 9-11.
No Rhyme or Reason | 53 A second example is Cloud Peak, which identified the 30% uncertainty surrounding each of these potential laws, emissions reduction target in its 2015 10-K filing. However, regulations and policies. We do not believe that such analyses Cloud Peak indicated that it could not quantify the implications reasonably predict the quantitative impact that future laws, for the company, noting that, “we are not currently in a position regulations or other policies may have on our results of to make any meaningful determination about the extent of operations, financial condition or cash flows.”94 the impacts to our operations.”91 This is in contrast to Arch’s 2014 disclosure, which suggests that companies can analyse Aside from these examples, most other coal companies did not the potential that given climate targets may have for reducing discuss emission targets, nor were efforts made to quantify the demand for their commodities. level of demand destruction implied by them. Instead, many emphasized that uncertainties were too great to estimate the Third, Peabody Energy identified the emissions reductions impact: targets identified by the Clean Power Plan and noted that they could have a significant impact on coal demand in the US.92 • “The content of new treaties or legislation is not yet However, no quantification of the potential impact on coal determined and many of the new regulatory initiatives remain demand is given. Separately, and as part of a settlement with subject to review by the agencies or the courts.”95 the New York Attorney General’s office,93 Peabody’s 2015 10-K • “We cannot predict the financial impact of future noted that it did examine the implications of various climate greenhouse gas or clean energy legislation on our operations or policies from time to time but that doing so required numerous our customers at this time.”96 assumptions, making those analyses an unreliable predictor of • “[I]t is not possible for us to reasonably predict the future results: impact that any such laws or regulations may have on our results of operations, financial condition or cash flows.”97 “From time to time, we attempt to analyze the potential • “We routinely attempt to evaluate the potential impact impact on the Company of as-yet-unadopted, potential on us of any proposed laws, regulations or policies, which laws, regulations and policies. Such analyses require that we requires that we make several material assumptions. make significant assumptions as to the specific provisions of such potential laws, regulations and policies. These analyses sometimes show that certain potential laws, regulations and policies, if implemented in the manner assumed by the analyses, could result in material adverse impacts on our operations, financial condition or cash flow, in view of the significant
94 Peabody 2015 10-K, at 19. 91 Cloud Peak 2015 10-K, at 18. 95 ANR 2010 10-K, at 22. 92 See, e.g., 2014 10-K at 14; 2015 10-K at 13. 96 ANR 2010 10-K, at 21-22. 93 http://ag.ny.gov/pdfs/Peabody-Energy-Assurance-signed.pdf 97 Cloud Peak 2010 10-K, at 79.
54 | www.carbontracker.org From time to time, we determine that the impact of one or more While companies often expressed uncertainty in forecasting such laws, regulations or policies, if adopted and ultimately the impact of climate policy, they readily provided third-party implemented as proposed, may result in materially adverse scenario information suggesting decades of continued demand impacts on our operations, financial condition or cash flow; for US coal—even though those projections discounted the however, we often are not able to reasonably quantify such aforementioned policy uncertainties to zero. From the point impacts.”98 of view of a company managing risk, this seems a flawed • “Enactment of laws or passage of regulations regarding approach; if an impact cannot be predicted precisely, it should emissions from the combustion of coal by the US or some of not be assumed there is no impact at all. A more reasonable its states or by other countries, or other actions to limit such approach would be to conduct a sensitivity analysis of a range emissions, could result in electricity generators switching from of impacts to demonstrate resilience under a number of coal to other fuel sources. The potential financial impact on us scenarios. This does not require the probability of the outcome of future laws or regulations will depend upon the degree to to be determined. Notably, such scenarios were available from which any such laws or regulations force electricity generators the same third-party sources cited by the companies. to diminish their reliance on coal as a fuel source. That, in turn, will depend on a number of factors, including the specific Companies use EIA numbers on long-term requirements imposed by any such laws or regulations, the time coal demand… periods over which those laws or regulations would be phased in and the state of commercial development and deployment of CCS technologies. In view of the significant uncertainty During the period reviewed, companies referenced EIA, World surrounding each of these factors, it is not possible for us to Coal Association (WCA), and IEA scenario analysis projecting reasonably predict the impact that any such laws or regulations coal demand through 2030-2040. The EIA Reference Scenario may have on our results of operations, financial condition or cash was commonly cited. That scenario assumed no new policies flows.”99 to mitigate climate change over the relevant time period, providing a baseline view assuming decades of future inaction on climate change. Typically, these forward-looking projections Each of the foregoing statements failed to acknowledge were included in the “business description” section of the the emissions reductions targets advanced by the UNFCCC reports. process, the US President, and the US Congress (Waxman- Markey) and therefore ignored a potential top-down approach. For example, this figure was included in ANR’s 2010 10-K, with updated numbers provided annually through its most recent (2014 10-K) filing. The numbers are from the EIA Reference Case, one of several scenarios that form a part of EIA’s Annual
98 Arch 2015 10-K, at 48. Energy Outlook (AEO). 99 Peabody 2010 10-K, at 48.
No Rhyme or Reason | 55 Although the Reference Case is a baseline policy case built for comparison purposes, many companies appeared to treat it as a forecast. Here, ANR simply reference EIA “data” with respect to this forward-looking information100:
100 ANR 2010 10-K, at 9. The chart includes “annual growth” calculations for the increase in coal consumption. Rather than providing a compound annual growth rates (CAGR), ANR appears to have provided an overall growth number for the referenced period. In subsequent reports, ANR used CAGR.
Cloud Peak similarly included EIA projections (likely the However, those figures appear to represent EIA’s view of what Reference Case, though not referred to as such), for coal would happen assuming no new regulations, implying that these consumption through 2035 in its 2011, 2012 and 2013 10- potential declines are, in fact, a “best case” scenario for coal Ks. Cloud Peak’s reference to EIA projections is included in demand under EIA’s analysis. Cloud Peak notes that they are its “Risk Factors” discussion. It frames them as a potential accordingly seeking to increase the export of their PRB coal downside risk from impending environmental regulation and reserves in order to grow the company.102 By characterizing natural gas competition, stating that, “[a]ccording to the EIA, the Reference Case as a downside scenario when, in fact, it the market share of coal used in electric generation is expected assumes no new climate regulations, seems to mischaracterize to decrease from 42% to 39% from 2012 to 2035 as a result of the Reference Case projections. various factors, including low natural gas prices, regulatory and environmental pressures on coal-fired electricity generation Cloud Peak no longer relied on EIA numbers in its 2014 or and domestic and foreign economic conditions and associated 2015 10-Ks, potentially suggesting that it no longer found them electricity demand.”101 useful predictors of the future.103
102 Id. 101 Cloud Peak 2011 10-K, at 24. 103 See Cloud Peak 2014 10-K, 2015 10-K.
56 | www.carbontracker.org Of all the 10-K filings reviewed for years 2010-2015, only Arch Coal identified the important assumption on which the EIA statistics relied, noting in its 2010 10-K that, “[t]hese amounts assume no future federal or state carbon emissions legislation is enacted and do not take into account subsequent market conditions.”104 Arch cited the EIA Reference case in its filings for years 2010-2014, noting in its 2013 and 2014 10-K105 “risk factor” disclosures that EIA demand projections assumed that some number of additional coal plants would be built and such build-out might be adversely affected by environmental regulations.106 However, no equivalent quantification of the downside risk using EIA analysis is explored.
In later years, some companies appeared to waiver in their use of the EIA Reference Case scenario. In its 2015 10-K, Arch identified the demand implications that the EIA modelled under its CPP scenario, noting that the CPP modelling had not yet been incorporated into the EIA AEO but that Arch expected that it would be107:
104 See, e.g., Arch 2010 10-K, at 5. 105 Arch 2013 10-K, at 7. 106 See, e.g., Arch 2014 10-K, at 46 (“The EIA’s expectations for the coal industry assume there will be a significant number of as yet unplanned coal-fired plants built in the future which may not occur.”); see also Arch 2012 10-K, at 42. 107 2015 10-K Arch, at 11.
No Rhyme or Reason | 57 …But do not consider alternative forecasts In 2011, Arch expressed optimism as to seaborne metallurgical based on emissions reduction targets markets but noted continued weakness in US steam coal consumption, due, in its estimation, to “abnormally mild 109 Coal company MD&A discussion of the implications of climate weather and muted economic growth,” ascribing power targets has varied. Several companies did not mention generation switching away from coal to “decade-low prices for 110 climate targets at all and consequently did not discuss the natural gas and abnormally high hydroelectric availability.” implications of those targets. As companies are required to Subsequent years attributed disappointing US steam coal disclose “material” trends should they deem them “reasonably demand to “unseasonably warm” winters and “low natural gas 111 likely” to occur, these omissions are difficult to square with prices” but did not discuss climate change and growth in the risks discussed elsewhere in the annual reports. After the renewables and efficiency as a related driver in the MD&A. announcement of the Clean Power Plan, some companies identified that it could have significant impact and identified It was not until after its bankruptcy filing on January 11, 2016, the CPP’s emissions reductions targets, but deemed it too that Arch’s MD&A identified environmental regulations as difficult to quantify the impact it might have. Only one company having been a significant driver of declines in US thermal coal endeavoured to quantify the impact that the CPP would have markets: on overall coal demand, demonstrating that such estimates could be made prior a regulatory rule becoming finalized. “The domestic thermal market was depressed by low natural gas prices and the implementation of new environmental Arch: No discussion of climate risk in MD&A—until after regulations. Abundant supply depressed natural gas pricing bankruptcy filing. to levels that made it increasingly economical to dispatch for electric generation relative to thermal coal. Implementation of Arch had comparatively detailed risk factor disclosures though, the MATS regulation resulted in the closure of some older coal- like other companies, failed to consider the carbon budget based generating facilities, further depressing domestic thermal implications of climate targets. In the earlier part of the period coal demand. Pricing in international thermal coal markets was 112 reviewed, Arch expressed optimism about the long-term uneconomic for our operations throughout 2015.” prospects for coal. 109 Arch 2011 10-K at 62. 110 Id. In 2010, Arch “expect[ed] growing global demand and 111 See, e.g., Arch 2012 10-K at 58; Arch 2013 10-K, at 58; Arch 2014 10-K, continuing supply constraints in traditional coal exporting at 63. 112 Arch 2015 10-K, at 62. In its business description, Arch further ascribed countries to continue to fuel seaborne coal markets for steam coal’s rapid decline to “a convergence of factors including record high natural 108 coal from the U.S.” gas production along with the lowest natural gas prices since 1999, coal unit retirements following the implementation of the Mercury and Air Toxics Standards regulations and growing power generation from wind and solar.” Arch 2015 10-K, 108 Arch 2010 10-K, at 36. at 9-10.
58 | www.carbontracker.org By this time, a focus on the potential for a precipitous decline Cloud Peak: Climate policy likely material, but no way to was not worth much to equity investors of the bankrupt quantify impact company. In its MD&A, Cloud Peak has typically noted that climate policy, While many of these regulations were disclosed as risk factors both regulation and subsidy for low-carbon energy, could result throughout the period in Arch’s 10-Ks, the omission of any in fuel switching away from coal, but that in view of policy discussion of management’s views in the MD&A over the uncertainties, “it is not possible for us to reasonably predict the early part of the period reviewed means investors were given impact that any such laws or regulations may have on our results little information on the likelihood of those risks, much less a of operations, financial condition, or cash flows.”114 quantification of how they might impact the company. Unlike unseasonably mild weather conditions or abnormally low natural In its 2014 10-K, Cloud Peak identified the emissions reduction gas prices, action on climate change and downward trends in targets in the Clean Power Plan and indicated that it could the cost of renewables represent long-term downward drivers significantly impact the business, but noted that it was “not in of demand and therefore long-term drivers of company value— a position to make any meaningful determination about the precisely the type of information that should be found in the extent of the impacts to our operations at this early stage in MD&A. This view is corroborated by other elements of Arch’s the rulemaking process.”115 Cloud Peak’s 2015 10-K MD&A 2015 10-K, where it includes the IEA’s modelling results of the section acknowledged that regulations had had and would 450 Scenario and the EIA’s modelling analysis of the CPP.113 continue to have a “significant adverse effect” on coal’s “competitive position relative to certain other sources of electricity generation,”116 and also indicated that the CPP could impact demand significantly.117 However, it did not endeavour to quantify that impact, presumably because of the uncertainties related to the details of those policies. Notably, in the first year that the CPP was announced, Cloud Peak ceased its reference to the EIA Reference Case.
113 “The IEA expects coal to retain its prominent presence as a fuel for the power sector worldwide under the Current Policies Scenario. Coal’s share of the power generation mix was 41% in 2013. By 2040, the IEA’s Current Policies Sce- 114 Cloud Peak 2010 10-K at 77; 2011 10-K at 80; 2012 10-K at 88; 2013 nario estimates that coal’s fuel share of global power generation will be 38% as it 10-K at 78; 2014 10-K at 70; 2015 10-K at 74. continues to have the largest share of worldwide electric power production. Under 115 Cloud Peak 2014 10-K, at 56. the New Policies and 450 scenarios, coal’s fuel share of global power generation is 116 Cloud Peak 2015 10-K, at 59. projected to be 30% and 12%, respectively.” Arch 2015 10-K at 9-10. 117 Id.
No Rhyme or Reason | 59 ANR: Initially acknowledges risks but then reduces discussion of ANR’s analysis of climate-related risks in the early period climate risk as conditions worsen reviewed suggested that it did not consider them too likely. Indeed, until the 2012 10-K, ANR suggested that the slow ANR acknowledged the risks posed by increasingly stringent rate of growth in electric vehicles and the “inability” of “wind climate policy in 2010 but, paradoxically, had very little and solar to become the base load source of electric power” discussion of the climate risks in the MD&A section of its 2014 as sources of its optimism.122 Understanding the implications 10-K MD&A. Moreover, its concerns related to climate policy of carbon targets, however, would have helped investors were mitigated by reference to the Reference Case suggestion understand the level of optimism implied. that coal demand would increase significantly. By 2012, ANR was no longer touting the long-term Shortly before finalizing its acquisition of Massey Energy, ANR fundamentals of US and international coal markets. However, acknowledged that, “the global demand for and use of coal it articulated the problems in the coal industry as cyclical rather may be limited by any global treaties which place restrictions on than structural and provided little analysis of the structural carbon dioxide emissions,”118 referencing the UNFCCC process threat posed by climate change mitigation. For example, ANR and the Copenhagen Accord but not the related emissions characterized the downturn in metallurgical coal as “a period of reduction target.119 This acknowledgement was paired with cyclical weakness” and that it expected “to be well-positioned the assertion that, according to the EIA International Energy to benefit from an eventual recovery in the global metallurgical Outlook (IEO), worldwide demand for coal was “expected” coal market.”123 Though climate risks appear to have increased to increase 56% by 2035, and that longer-term, “the delicate since 2010, ANR no longer mentioned climate-related risks as a balance of coal supply and increasing coal demand is expected long-term structural downward driver in its MD&A as it had in to result in strong, but potentially volatile fundamentals for its 2010 10-K. the US coal industry” and that internationally, “coal industry fundamentals remain intact and significant additional growth is expected worldwide.”120 Like the EIA AEO cases, the EIA cautions that the IEO is not a statement of “what will happen,” but instead “what might happen given the specific assumptions 122 See, e.g., ANR 2010 10-K at 61-62 (“Long-term demand for coal and and methodologies used…”121 It is therefore not what EIA coal-based electricity generation in the US will likely be driven by various factors such as the economy, increasing population, increasing demand to power residen- “expected” to happen. tial electronics and plug-in hybrid electric vehicles, public demands for affordable electricity, the inability of renewable energy sources such as wind and solar to become the base load source of electric power, geopolitical risks associated with im- porting large quantities of global oil and natural gas resources, increasing demand 118 ANR 2010 10-K, at 61-62. for coal outside the US resulting in increased exports and the relatively abundant 119 Id. steam coal reserves located within the United States.”). See also ANR 2011 10-K, at 120 ANR 2010 10-K, at 61-62. 64-65. 121 EIA IEO 2010, at ix. 123 ANR 2012 10-K, at 62-63.
60 | www.carbontracker.org Indeed, the only climate-related MD&A discussion from 2012 onwards shows that its discussion was focused on anything but the structural implications. ANR mentions the EPA’s MATS regulation in its 2013 10-K. Rather than focusing on the impending closures and demand impact, ANR instead notes “speculation that some power grid operators will designate additional plants, such as the Brayton Point power plant in Massachusetts, as “must run” facilities.”124 This note of optimism may have not provided a clear picture of the extent of structural change underway.
In its 2014 10-K, ANR’s outlook beyond the following year became less certain, noting that “Recent price trends suggest that the thermal coal market will remain challenging for 2015. Price trends thereafter will be significantly influenced by the price of natural gas, the pace of economic growth, ongoing regulatory pressures on the domestic coal-fired utility fleet and the weather.”125 As with other companies, there was no discussion of the demand implications of international or national emissions reduction targets.
124 ANR 2013 10-K, at 61-63. 125 ANR 2014 10-K, at 63.
No Rhyme or Reason | 61 Figure 17: EIA projections of coal’s share of power generation cited in coal company 10-Ks Coal’s share of power gen. (%) share Coal’s
Sources: CLD = Cloud Peak Energy ANR = Alpha Natural Resources ACI = Arch Coal HNRG = Hallador Energy JRCC = James River Coal BTU = Peabody Energy CNX = CONSOL Energy ARMS = Armstrong Energy AEO = EIA Annual Energy Outlook ACESA = American Clean Energy and Security Act
62 | www.carbontracker.org Part VI
Conclusions
63 | www.carbontracker.org No Rhyme or Reason | 63 The impact on US coal company valuations has already been This has implications for company disclosure. While many coal felt; the question now is what policymakers, regulators, and companies frequently relied upon the Reference Case, few other stakeholders can do to make such risks more transparent identified the drastically different results for coal that would going forward for other sectors, especially oil and gas. flow from even moderate climate policies.
With hindsight, it is clear that the EIA’s “Reference Case” has We believe a number of important steps could be taken not been aligned with the actual demand levels - although given to address these concerns. The EIA makes its modelling it is not intended to predict them, this should not in fact be a assumptions available and cautions that they are not forecasts; big surprise. Its modeling of emissions-reduction legislation very few corporate reports have done the same. As a first impacts has been more informative in indicating actual step, the SEC should ensure that the underlying reports are developments as they are consistent with the clear direction accurately represented and that the critical assumptions are of travel. Though the policies were not passed, they better disclosed. Second, it would help to understand whether such predicted actual developments over the last several years. In third-party forecasts are indeed reflective of management’s our view, these projections have been more accurate because beliefs — whereas robust risk disclosure suggests the company they incorporated emissions reduction targets, which are a may be taking steps to mitigate risks, business as usual proxy for the range of policy and technological actions that projections suggest the opposite. Clearly, companies should undermine fossil fuel demand. not be relying on overly optimistic scenarios. Implicitly, the use of a third-party forecast suggests that the company views it as Companies have often suggested that policy uncertainties make reasonable (without more being said). Did companies believe it difficult to quantify the impact of action on climate. What this there would be no meaningful climate regulation over the next suggests, however, is that it may be better to focus simply on 20 years? If so, it would be helpful for the markets to know. the emissions reduction targets, rather than the specific policy details, in identifying the scope of the risk. This would better Third, the SEC should also consider how to incorporate allow investors to see the forest for the trees. emissions reductions targets into company disclosures. Those targets may seem long-term, but the US coal crash Emissions reduction targets are public and long-term. This demonstrates how quickly those risks can develop. Moreover, suggests that markets should naturally incorporate them, such targets fall well within the 20-25 year time horizon for EIA assuming that the connection between those targets and projections that companies frequently discuss. That discussion sectoral impact is well understood. It may not be. The abrupt should go beyond best-case scenarios to consider the downside changes in US coal markets suggest that they were caught by case. Here, further attention should be given to the overall surprise. presentation — i.e., are business-as-usual projections and forecasts consonant with the numerous risks factors identified in corporate reports? This is also relevant for auditors—to
64 | www.carbontracker.org what extent are forecasts and projections in narrative reports It is possible that such disclosure would even inure to the consonant with the assumptions made in underlying financial benefit of management. Investors tend to disfavor corporate reports and to what extent are those assumptions reasonable? diversification so companies may not perceive many options when confronted with downward declines in their business. At root, the question is whether the future will replicate the However, growth does not always equate to value, and in past—precisely the data sought by MD&A. Emissions targets articulating the case for less being more, companies need a very much call that into question. Regulators should consider way to frame that case for investors. Detailing the demand whether, as a first order estimate, companies should be implications of climate targets may be one way for companies quantifying the impact of relevant emissions reduction targets to make that case. for their businesses. Such disclosures would provide insight into how companies are considering and managing the risks, not just identifying them.
The EIA is also a critical reference point. While there are sound reasons for continuing to publish reference scenarios, the EIA is right to admonish readers for considering them in isolation. It could augment its modeling with regular analysis of climate targets to identify pathways of how the future may unfold, including potentially a two-degree scenario as the IEA does. Such an analysis is at least as plausible, if not more plausible, than the assumption that business will continue as usual.
Emissions reductions targets also have implications for company planning, particularly in the current restructuring process sweeping the industry. For coal companies, bankruptcy is not the end of the road, but instead an opportunity for a “fresh start.” The foregoing suggests that stakeholders in the increasing number of coal bankruptcy administrations should consider how the projections will impact the ability to reorganize. This means considering the implications of emissions reduction targets to ensure that companies do not “double-dip” in bankruptcy court. This is especially relevant when evaluating the feasibility of reorganization plans.
No Rhyme or Reason | 65 Figure 18: Recommendations
66 | www.carbontracker.org Find out more about the Initiative Carbon Tracker Initiative: arbon Tracker www.carbontracker.org @carbonbubble
© Carbon Tracker Initiative June 2016
Readers are encouraged to reproduce material from Carbon Tracker reports for their own publications, as long as they are not being sold commercially. As copyright holder, Carbon Tracker requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the Carbon Tracker website.
No Rhyme or Reason | 67