Crimes of Emission

PAGE 2

C RIMES OF E MISSION : H OW D IRTY A RE THE HFC 23 A BATEMENT P ROJECTS OF THE C LEAN D EVELOPMENT M ECHANISM ? BY O LIVER L AURENCE P ECKHAM A T HESIS Submitted to the Divisions of Political Science and Environmental Studies New College of Florida in parti al fulfillment of the requirements for the degree Bachelor of Arts under the sponsorship of Dr. Frank Alcock Sarasota, Florida May, 2012

PAGE 3

ii A CKNOWLEDGEMENTS First, t hank s to my parents, V IVIANNE AND C HARLES P ECKHAM for their unconditional and unwavering support. Thanks to D R F RANK A LCOCK for directly inciting my initial interest in environmental issues and for providing invaluable support to my subsequent endeavors in the field. Thanks to D R B ARBARA H ICKS AND D R R ICHARD C OE for entertaining a tumble down the rabbit hole of carbon market policy and for helping me think of these issue s from different perspectives. Thanks to my friends, in particular I AN F OSTER L IZ G IBBONS N ATHAN H OWELL A NDREW S CHUSTER H ARRISON S HERWIN AND M ARTIN S TEELE for providing love and levity thro ughout a strenuous year. Finally, t hanks to C AFFEINE for rescuing me from the confines of my circadian rhythm.

PAGE 4

iii T ABLE OF C ONTENTS Acknowledgements ii List of Abbreviations iv Abstract v Introduction Carbon Offsets 1 The Clean Development M echanism 7 The HFC 2 3 Abatement Controversy 1 1 Scope of the Study 19 Existing Literature 20 Analyses Overview 3 5 Estimating Revenues 3 7 Assessing Aberrant B ehavior in HCFC 22 Production 43 Assessing Aberrant Behavior i n the HFC 23 Production Rat io 82 Conclusions Aggregate E f fects of Perverse Incentives 12 4 Sufficiency of Recent Adjustments to AM0001 12 6 Implications for the CDM 1 29 Appendices Appendix I 13 4 Appendix II 13 5 Appendix III 13 6 Bibliography 14 5

PAGE 5

iv L IST OF A BBREVIATIONS AM Approved Methodology BAT Best Available Technology CDM Clean Development Mechanism CER Certified Emission Reduction Credit DOE Designated Operational Entity EB Executive Board ETS Emission T rading Scheme GHG Greenhouse Gas GWP Global Warming Potential IPCC Intergovernmental Panel on Climate Change MP Monitoring Period MR Monitoring Report MY Monitoring Year ODS Ozone Depleting Substance pCER Primary Certified Emission Reduction Credit PDD Project Design Document PP Project Participant sCER Secondary Certified Emission Reduction Credit tCO 2 eq Metric Tonnes Carbon Dioxide Equivalent TEAP Technical and Economic Assessment Panel UNEP United Nations Environment Programme UNFCCC United Nations Framework Convention on Climate Change

PAGE 6

v C RIMES OF E MISSION : H OW D IRTY A RE THE HFC 23 A BATEMENT P ROJECTS OF THE C LEAN D EVELOPMENT M ECHANISM ? Oliver L. Peckham New College of Florida, 2012 ABSTRACT T he Clean Development Mechanism ( a carbon offset p rogram under the Kyoto Protocol ) has been accused of issuing carbon credits that do not represent real reductions to HFC 23 abatement projects at HCFC 22 production facilities. As a highly potent greenhouse gas, HFC 23 abatement creates a large amount of credits under a propor tional issuance system. This potentially creates a perverse incentive to increase HCFC 22 production or the HFC 23/HCFC 22 ratio for the sole purpose of generation carbon credits. This study uses a variety of market data, reported production data, and preexisting reports to draw detailed conclusions regarding the behavior of each of the nineteen HCFC 22 production facilities certified for credit issuance for HFC 23 abatement practices. Doing so, the study finds that all nineteen plants certified for CDM credit issuance for the abatement of HFC 23 have exhibited signs of behavioral responses to the aforementioned pervers e incentive The study concludes by recommend ing a way forward for HFC 23 ab atement crediting under the CDM in light of recent pol icy responses (or the lack thereof).

PAGE 7

1 first global, environmental investment and credit scheme of its kind, providing a mechanism stimulates sustainable development and emission reductions, while giving industrialized countries some flexibility in how they meet their emission reduction or UNFCCC W EBSITE 1 educting what you hope D AN W ELCH ( E THICAL C ONSUMER JOURNALIST ) 2 I NTRODUCTION C ARBON O FFSETS The argument for carbon offsetting begins with a simple premise: in terms of atmospheric composition, an emissi on reduction is an emission reduction, regardless of where it occurs. It follows that expenditures dedicated to the reduction of greenhouse gas emissions should be made in a way that seeks to maximize economic efficiency in terms of quantity of emission re duction over cost. However, given the vast range of actors driving climate change (and assuming non prohibitive transaction costs), it is highly unlikely that a rational consumer seeking to maximize the economic efficiency of a direct emission reduction in vestment would best serve 1 http://unfccc.int/kyoto_protocol/mechanisms/clean_development_mechanism/items/2718.php 2

PAGE 8

2 that interest by reducing his or her own emissions. In other words: theoretically, there is almost always a cheaper reduction. When the nature of emission reduction consumption and production is considered, the likelihood of an e fficiency gap between internal and external reductions increases even further. As climate change is an issue with highly dispersed, difficult to identify, long run effects climate change mitigation currently s. Wealthy and secure actors will therefore exhibit higher demand for emission reduction s while impoverished and insecure actors will exhibit lower demand. This is compounded by the tendency of wealthy actors to produce more emissions 3 : as the amount of d esired emission reduction is often derived from the amount emitted by the reducer, actors with higher emissions in this case, the wealthy will tend to desire greater reduction. suppliers ion reduction. Economies in developing countries tend to exhibit high impact 4 Small amounts of emission reductio n investment in these areas often produce excellent returns by contributing basic and affordable infrastructure in the early stages of industrial development. 5 Economies in developed nations, however, tend to be restrained by relatively thorough regulation s and less reliant on centralized environmental 3 Linked Environmental Science & Technology 43 (2009): 6414 6420. 4 Ander (University of Gothenburg, 2008). 5 World Economics & Politics 6 (2 003): 66.

PAGE 9

3 exploitation. 6 This has created a situation where the best consumers are poor producers and the best producers are poor consumers. Carbon offsetting is intended to address this inefficiency. An offset is an external emission reduction measure undertaken or paid for by an actor with the 7 Typically, o ffsets take the form of saleable credits issued to emission reducing entities as part emission cap or target. Theoretically, the increase in economic efficiency contribute d by offset mechanisms reduces the cost per unit of emission reduction, rewarding actors with inelastic demand (less money is spent on the same quantity of reduction) and leading to an increase in the volume of reductions where demand is elastic or unit el astic (more credits are purchased in response to the price decrease) Nevertheless, the adoption and acceptance of carbon offsets as legitimate emission reductions has been impeded by a succession of controversies, scandals, and objections. Many offset pr ojects have been accused of negative social impacts, such as displacing and disregarding local populations. 8 Other critics have argued against offsets from a position of moral principle, comparing the sale of carbon 6 Environment and Development ed. Teng Teng and Ding Yifan (EOLSS Publishers, 2009). 7 How to Create a Fair 8 Philosophical Transaction: Mathematical, Physical, and Engineering Sciences 360 (2002): 1875 1888.

PAGE 10

4 credits to the sale of indulgences by th e Catholic Church. 9 However, the most piercing and pervasive criticisms have been those questioning the functional legitimacy of the credits themselves In order to be considered a legitimate and creditable reduction of carbon emissions, an offset must me et several criteria. First, the reductions must be quantifiable and measurable. Second, the reductions must be verifiable by a third party. Third, the reductions must not be reversible. Finally, the reductions must not have been feasible without the additi onal support generated by carbon crediting. 10 Without this final requirement known as additionality credits could be issued to and purchased from projects that would have reduced emissions without offset revenue. This would effectively result in increas ed emissions, as actors would be able to reduce their emission reduction obligations by paying for offsets that would be generating no real increases in abatement. Non additional offsets would also be artificially cheap, as the reduction measures that crea ted them would have been financially feasible without revenues from credits. Additionality is measured against an assessment of the likeliest scenario in the hypothetical absence of offset crediting, otherwise known as a counterfactual baseline. Such anal yses take into account available funding, regulatory conditions, industry norms, common practice, and other factors affecting the feasibility of the project and attempt to determine whether it would be able to establish itself 9 George Monbiot last modified October 2006, http://www.monbiot.com/2006/10/19/selling indulgences/ 10 Environ ews 117 (2009): 64.

PAGE 11

5 without credit issuance. 11 Sti ll, who can say what choices would have been made, especially given the fluctuating nature of environmentalism and green initiatives? Would the project have received loans or grants? If barriers are present, would they have proved significant enough to be prohibitive in the absence of funding? Perhaps most subjectively, what constitutes common practice? Indeed, the very existence of credit issuance programs influences these factors some worry that offset mechanisms have disincentivized the establishment of domestic subsidies for emission reduction projects. 12 The elimination of negative externali ties ( detrimental effects of a process on product that are not reflected in its price ) is another serious concern in carbon offsetting. While emissions directly c aused by an emission reduction project within its boundaries are relatively straightforward to identify and internalize, the same is not true of emissions outside the project boundary. These externalities are known as leakage, which is more thoroughly defi ned by the Intergovernmental Panel on 13 Leakage can range from the direct and d iscrete impacts of a project to issued to a reduced deforestation project, but the affected loggers simply relocated to another forest, the emissions caused by the reloca ted logging would be 11 Analysis (Tyndall Centre for Climate Change Research, 2009): 6 7. 12 Development 13 US EPA, Projects, a ccessed February 2012, http://www.epa.gov/sequestration/leakage.html

PAGE 12

6 considered leakage. 14 On the other hand, even successfully reduced deforestation could cause leakage by lowering the supply of timber, thereby raising timber prices and incentivizing unsustainable logging practices. Given the plethora of difficulties facing the legitimacy of carbon offset projects, carbon trading schemes face an uphill battle. Trading mechanisms are expected to establish accurate baselines, engage in active and thorough verification, satisfactorily determine additionali ty and permanence, and internalize as many negative extern alities as possible. The issue this study discusses relates to one of the most prominent systems attempting to overcome these difficulties: the Clean Development Mechanism (CDM) of the Kyoto Protoco l. 14 Ibid.

PAGE 13

7 T HE C LEAN D EVELOPMENT M ECHANISM Developing countries, concerned that the costs of mitigation would hinder domestic industrialization, refused to support the inclusion of a universal emission reduction commitment in the Kyoto Protocol. The resulting d ichotomy between industrialized countries making reduction commitments (later known as Annex I parties) and developing countries not making reduction commitments (later known as non imp rovements in sustainability. In response to this dilemma, the parties designed an offset program called the Clean Development Mechanism (CDM), which has been operational since 2006. 15 t Parties not included in Annex I in achieving sustainable developing and in contributing to the ultimate objective of the Convention, and to assist Parties included in Annex I in achieving compliance with their quantified emission limitation and reduction 16 Essentially, the Mechanism allows Annex I countries to partially meet their reduction commitments through the purchase of credits from certified projects in non Annex I countries. CDM credit issuance may be requested for either a ten yea r period with no potential for renewal or a seven year period with the potential to be renewed 15 Robert Repet itutional Breakthrough or Insti t u tional Policy Sciences 34 (2001): 303. 16

PAGE 14

8 twice. 17 To apply, applicants first create a Project Design Document (PDD) that includes a project baseline, a methodology for the reduction of emissions against that baseline, and plans for the measurement of all relevant variables. M ethodologies for monitoring and baseline establishment are limited to the various Approved Methodologies (AMs) issued by the Executive Board (EB). 18 Once the PDD has been drafted, a De signated Operational Entity (DOE) that has been certified by the EB is hired by the applicant to review the proposed project. The DOE is responsible for assessing the additionality of a proposal through the application of four tests: the Regulatory Test, t he Barrier Test, the Investment Test, and the Common Practice Test. 19 The Regulatory Test examines any and all alternatives to the project which retain its primary function and comply with all legal and regulatory standards. When considering a hydropower p lant, for instance, a DOE would likely examine the regulatory viability of other power plant technologies. If the hydropower plant was legal and regulatory frameworks, the project would be considered non additional. 20 The project must then satisfy either a Barrier Test or an Investment Test. To satisfy a Barrier Test, the proposed technology must be significantly impeded by well documented barriers that are inconsequential to alternative technologies in the sector. To satisfy an Investment Test, the proposed technology must be significantly 17 Stephen O. Andersen and K. Madhava Sa Together to Curb HFC 2010), 9. 18 UNFCCC ast modified May 2012 http://cdm.unfccc.int/ 19 7. 20 Ibid.

PAGE 15

9 less likely to attain sufficient investment than alternative technologies in the sector. If the project satisfies neither test, it is co nsidered non additional. 21 Finally, the project must pass the Common Practice Test, which gauges the prevalence of the proposed technology. If the technology is deemed to be business as usual in the sector or the region, the project is considered non additi onal. 22 If the DOE determines that the technology is not common practice (and all other necessary tests have been satisfied), it recommends approval of the project to the EB, which almost invariably accepts that recommendation. 23 Once approved by the EB and cleared for credit issuance, the newly certified Project Participant (PP) begins to file Monitoring Reports (MRs) documenting its emission reductions during each specified Monitoring Period (MP). Reductions are calculated using the methodologies outlined terms of metric tonnes of carbon dioxide equivalent (tCO 2 eq). Where emission or mitigation of a gas other than CO 2 is involved, the IPCC approved Global Warming Potential (GWP) of the gas, which is itself a measure of CO 2 equivalency, is used as a conversion factor. 24 Monitoring Reports are also vetted by DOEs. 25 Assuming no fault is found with the MR, the EB issues a quantity of Certified Emission Reduction credits (CERs) equal to the net tCO 2 eq reduced by the project du 21 Ibid. 22 Ibid. 23 Is the CDM fulfilling its environmental and sustainable development objectives? An evaluation of the CDM and options for improvement ko Institut e.V., 2007 ), 19. 24 UNFCCC ast modified May 2012 http://cdm.unfccc.int/ 25 Is the CDM fulfilling its environmental and sustainable development objectives? An evaluation of the CDM and options for improvement

PAGE 16

10 Typically, however, actors seeking to reduce emissions do not purchase pCERs; instead, carbon market investment firms, which pool risk by investing in a portfolio of projects, sell stabilized CERs to consumers at a higher price. These transactions comprise the secondary market for CERs, which are then known as sCERs. 26 End consumers of CERs usually fall under Annex I of the Kyoto Protocol or participate in another carbon market scheme that permits the use of CERs (e.g. the Emission Trading Scheme [ETS] of the European Union); the voluntary market is very small. 27 Theoretically, this system should allow the Kyoto Protocol to capitalize on the gains in economic efficiency expected from a carbon offset program. First, t he industrialized Annex I countries presumably have, on average, a higher domestic emission reduction cost than the dev eloping non Annex I nations. Second, g iven the s 28 it is also likely that the bulk of emission reduction demand by Annex I parties is elastic or unit elastic. However, criticisms of the CDM have grown steadily louder since its inception, and faith in CERs is plummeting along with their price. 29 This ana lysis will discuss the most serious accusation of illegitimacy: the allegedly perverse influence of CDM credit issuance to HFC 23 abatement projects at HCFC 22 production facilities. 26 Green Evolution ccessed March 2012 http://www.gre en evolution.eu/default.asp?pid=138&la=1 27 28 European Environment Agency, 29 http://data.bluenext.fr/downloads/20120510_BNS_STATS.xls

PAGE 17

11 T HE HFC 23 A BATEMENT C ONTROVERSY HCFC 22 (HCF 2 Cl) is a coolant gas with applications as a feedstock for certain synthetic polymers, primarily polytetrafluoroethylene (PTFE). 30 Though HCFC 22 is an ozone depleting substance (ODS) with a GWP of 1,810 and an approximate atmospheric lifetime of ten years 31 it is typically released in very small quantities due to the closed loop nature of refrigeration systems. 32 HCFC 22 is manufactured through the controlled reaction of chloroform (CHCl 3 ) and anhydrous hydrogen fluoride (HF) as demonstrated below 33 : CHCl 3 + 2HF HCF 2 Cl + 2HCl Howeve r, the overstoichiometric reaction of HCF 2 Cl and HF, known as overfluorination, can result in the creation of a third product 34 : HCF 2 Cl + HF CHF 3 + HCL CHF 3 more commonly known as HFC 23, is a non ODS with an estimated GWP of 14,800 (recently adjusted from a previous estimate of 11, 7 00) and an atmospheric lifetime ranging from one to two hundred years. 35 While other waste products from the initial reaction may be converted back into desirable reactants, 30 23 and Other 31 Ibid. 32 Ibid. 33 A. 23 Waste Streams for Abatement of Emissions from HCFC 22 ( UNFCCC, 2004 ), 7. 34 http://cdm.unfccc.int/UserManagement/FileStorage/50KH2J9V6O1IQNBSPALXYUGRCZFED7 35 23 and Other

PAGE 18

12 this is not feasible for HFC 23. 36 As a result, the gas is generally released into the atmosphere in the absence of regulatory barriers or economic incentives. 37 HFC 23 output can be reduced in three ways: (1) reduction of HCFC 22 output, (2) optimization of the HCFC 22 to HFC w (3) dest ruction of the produced HFC 23. 38 Of these three methods, destruction is the only one to offer near complete mitigation of the gas while retaining HCFC 22 production HFC 23 is most commonly destroyed by way of thermal oxidation in a natural gas furnace, yi elding CO 2 and HF 39 : 2CHF 3 + 2H 2 O + O 2 2CO 2 + 6HF Tests measuring the mitigation of HFC 23 through thermal oxidation have demonstrated a destruction rate of over 99.996%; however, incinerator downtime frequently reduces that rate in practice. 40 As abateme nt of high GWP GHGs is a priority for the Kyoto Protocol, the CDM adopted approved methodology AM0001, regulating the mitigation of HFC 23 waste streams under the CDM, at its 10 th meeting in September 2003. 41 Initially, the limit on creditable emissions con sisted of a cap on the creditable w ratio set at the lowest w demonstrated by the facility for any year between 2000 and 2004; if the lowest 36 23 Waste Streams for Abatement of Emissions from HCFC 22 Production: A Review of Scientific, 37 Ibid. 38 23 a nd Other 39 UNFCCC http://cdm.unfccc.int/ Project 0115 MR 2. 40 23 Waste Streams for Abatement of E missions from HCFC 22 Production: A Review of Scientific, 41 UNEP Technolog y and Economic Assessment Panel, (with particular focus on the impact of the Clean Development Mechanis m) and Emissions Reduction Benefits Arising From Earlier HCFC Phase O ut and Other Practical Measures ( UNEP, 2007 ), 51.

PAGE 19

13 value was above three percent, w was capped at three percent. 42 This cap was checked against each reported MP and ap plied pro rata such that an excess over the cap during any MP resulted in a proportionate reduction in creditable HFC 23 abatement for that MP. 43 Otherwise, in accordance with HFC of HFC d produce 11,700 CERs per ton abated. 44 15 th meeting in September 2004 (before the start of CDM crediting) it had received substantial and ominous feedback on the new methodology. Due to the extreme quantity of CERs involved, critic s had reasoned that production of HCFC 22 for the sole purpose of generating HFC 23 to be abated might be economically viable, thereby creating a perverse incentive through crediting. 45 In essence, at some production and price levels, HFC 23 could displace HCFC 22 as the effective product of the facility. The potential for the existence of this incentive posed a clear threat to the legitimacy of the CDM, as c redits obtained through otherwise undesirable production of HCFC 22 constitute an excess over the iss uance ethically expected of the project. Excessive credit issuance increases the 42 23 and Other 43 UNFCCC CDM http://cdm.unfccc.int/UserManagement/FileStorage/50KH2J9V6O1IQNBSPALXYUGRCZFED7 44 te Change and Ozone Treaties Work Together to Curb HFC 23 and Other 45 UNEP Technolog y and Economic Assessment Panel, (with particular focus on the impact of the Clean Development Mechan ism) and Emissions Reduction Benefits Arising From Earlier HCFC Phase O

PAGE 20

14 respons e to this possibility, the EB put AM0001 on hold and requested a review of the methodology by the Meth Panel. 46 At its 17 th meeting (December 2004), the EB considered the Meth Panel recommendations and decided to revise the methodology. 47 In these revisions the CDM implemented two prerequisites for HFC 23 abatement project applicants. First, it mandated that the project must have had an operating history of at least three years between January 2000 and December 2004. Second, it mandated that the project mus t have been in continuous operation from 2005 to the start of the abatement activity. 48 This effectively banned new or intermittently operational HCFC 22 plants from crediting under the CDM, severely limiting the pool of potential applicants: as of April 20 12, only nineteen such plants have been certified under the CDM. 49 The revised AM0001 also implemented a second cap on the amount of creditable HFC 23 abatement. This cap, which regulated creditable HCFC 22 production, was set at the highest HCFC 22 product ion demonstrated by the facility during any year between 2000 and 2004 and applied pro rata to each Monitoring Year (MY). 50 However, these measures did not silence the criticisms. Under the new structure, critics argued, the perverse incentive to generate HFC 23 for the sole purpose of abatement could still exert its influence in two major ways: 46 Ibid. 47 Ibid. 48 Ibid, 52 53. 49 UNFCCC http://cdm. unfccc.int/ 50 http://cdm.unfccc.int/UserManagement/FileStorage/50KH2J9V6O1IQNBSPALXYUGRCZFED7

PAGE 21

15 1) Projects could act to maximize HFC 23 production through manipulation of w to its maximum value. This would be contrary to common practice, which is to optimize w t o maximize the efficiency of HCFC 22 production (thus reducing production costs). 51 2) If HCFC 22 production wa s sufficiently unprofitable and CER prices we re HCFC 22 production ( u p to the capped amount ) These concerns were amplified by the regulatory context of HCFC 22, which is primarily restricted by the Montreal Protocol on Substances That Deplete the Ozone Layer. Though use of HCFC 22 as a transitional refrigerant was initial ly encouraged, the Protocol began the global phase out of HCFC 22 consumption in 1992. 52 Initially, the phase out required non Article 5 parties (developed nations) to engage in an incremental reduction of HCFC 22, beginning in 1996 and culminating in compl ete phase out in 2030; for Article 5 parties (developing nations), the phase out prescribed a freeze in 2016 with a non incremental complete reduction deadline of 2040. 53 However, amendments to this schedule were passed in 2007 which accelerated the rate of non Article 5 party reduction and shifted Article 5 parties to a graduated system with a 2013 freeze (maintaining the 2040 complete reduction deadline). 54 Non emissive uses primarily chemical feedstock applications are not 51 23 Waste Streams for Abatement of Emissions from HCFC 22 Production: A Review of Scientific, 52 23 and Oth er 53 (Institute for Governance and Sustainable Development, 2007). 54 Ibid.

PAGE 22

16 regulated by this phase out s chedule. 55 With this phase out slowly beginning to take place, concerns emerged that CER revenues from HFC long 56 Theoretically, this could result in the displacement of more recently constru cted and likely more efficient HCFC 22 plants by CDM credited HCFC 22 plants, which would be able to offer lower prices due to CER revenues. 57 In January 2012 the European Union (EU) enacted a ban on the use of HFC 23 CERs in its Emissions Trading Sche me (ETS) effective May 2013 58 Soon after in a long awaite d response to the consistent claims that perverse production incentives were influencing the behavior of projects utilizing AM0001, the CDM EB put a nu mber of projects under review as a way of halt ing issuance until the methodology had been thoroughly reviewed. 59 A t 65 th meeting in November 2011, the CDM EB once again dramatically revised AM0001. In this most recent version of AM0001, the hard cap on w has been reduced from 3.0% to 1.0%, and the cap on creditable HCFC 22 production has been set to average historical production rather than the highest historical production from 2002 2004. 60 55 UNEP Technolog y and Economic Assessment Panel, ask Force on HCFC Issues (with particular focus on the impact of the Clean Development Mechanism) and Emissions Reduction Benefits Arising From Earlier HCFC Phase O 56 (UNFCCC, 2010), 3. 57 UNEP Technolog y and Economic Assessment Panel, (with particular focus on the impact of the Clean Development Mechanism) and Emissions Reduction Benefits Arisin g From Earlier HCFC Phase O 58 (Environmental Investigation Agency, 2011). 59 Dinakar ydro Fluorocarbon Emission Credits by 60 http://cdm.unfccc.int/UserMan agement/FileStorage/50KH2J9V6O1IQNBSPALXYUGRCZFED7

PAGE 23

17 This revised methodology will not take effect during the current approved crediting periods: as such, the revision s will not affect any of the projects that opted for ten year crediting periods, nor will it affect the first seven years of crediting under any of the renewable projects. 61 Nevertheless, in terms of the renewable projects, subsequent crediting periods will likely experience a reduction in real crediting of approximately two thirds, as most of the certified projects currently experience caps at or near 3.0%. 62 However, this latest move still has not dramatically swayed critics of HFC 23 crediting under the CD M: according to the Environmental Investigation Agency (EIA), an NGO with long term objections to encourages and subsidises increased HCFC 22 production in order to produce the HFC 2 3 by 63 The generation of illegitimate HFC 23 abatement credits has grave implications for the CDM and other carbon markets that accept the CER. As demonstrated in Appendix I, HFC 23 CERs have consistently constituted approximately half of all CERs issued by the CDM. 64 If some or all of those HFC 23 credits have been distributed based on intentional HFC 23 generation, then the real price of carbon emission reduction has been severely undercut. In addition, the loss of faith in the legitimacy of a pur chased credit itself will drive down demand, thus driving down price further. All such price reductions reduce the feasibility of legitimate carbon emission reduction projects under the CDM, as the expected 61 UNFCCC http://cdm.unfccc.int/ 62 Ibid. 63 for HFCs industry to continue." 64 UNFCCC http://cdm.unfccc.int/

PAGE 24

18 revenues from CERs will be less sufficient to jus tify project inception or operation. This possibility is made all the more upsetting by the fact that HFC 23 abatement projects constitute less than one half of one percent of all registered projects under the CDM. 65 Accordingly it is absolutely integral t o the survival of the CDM and the perceived legitimacy of carbon offsetting as a concept that any illegitimacy in crediting of HFC 23 abatement facilities is identified and rectified. 65 Ibid.

PAGE 25

19 S COPE OF THE S TUDY The purpose of this study is to answer the followi ng questions : first, in terms of revenues, is it likely that perverse incentives to increase generation of HFC 23 exist? Second, do the certified projects exhibit signs of influence by those perverse incentives? Third, is it likely that the recent adjustme nts to AM0001 will correct any current perverse incentives or non additionality in crediting ? Finally, in light of any remaining problems, what steps should be taken in terms of policy ? The bulk of the study is comprised of a thorough assessment of the pre sence and effects of perverse economic incentives resulting from CER issuance to HFC 23 abatement projects under the CDM. To accomplish this, the study primarily uses the documents submitted by certified projects to the Clean Development Mechanism (primari ly PDDs and MRs) up to December 2011 This data is complemented by a variety of studies (which are primarily covered in the following literature review), HCFC 22 price data, CER price data, and credit issuance data. T he study uses this data to establish es timates of CER to HCFC 22 revenue ratios experienced by certified HFC 23 projects and conclusions regarding the responses of w values and HCFC 22 production to the crediting caps at each credited plant Following these analyses, the study draws conclusions regarding the scale of 23 projects and the likely efficacy (or lack thereof) of the recently implemented changes to AM0001. Building on these conclusions the study recommends a way forward for HFC 23 policy under the CDM.

PAGE 26

20 E XISTING L ITERATURE A number of analyses have examined the potential for perverse incentivization in HFC 23 crediting by the CDM, and several have attempted to oduction process. The following literature review will focus on a set of prominent studies and reports that examine the comparative revenues of HFC 23 production, w value fluctuations, and HCFC 22 production rates. In April 2005, Lambert Schneider Jakob Graichen, and Nele Matz published a HFC 23 abatement versus HCFC 22 production. The authors assume an HFC 23 abatement cost of $0.20 1.00/tCO 2 2 eq, $0.73/tCO 2 eq, $4 6/kg HFC 23, and <$0.20/tCO 2 eq) 66 a CER price range of $5 $15/CER (citing market data), an HCFC 22 price range of $1.1 2.4/kg (citing a CDM estimate from November 2004), a w range of 1.5 3.0%, and a rate of CER allocation to PPs of 50 98%. 67 From these assumptions, the authors project a low impact scenario, a reference scenario, and a high impact scenario. In the low impact scenario, HFC 23 revenues eq ual 11% of HCFC 22 revenues; in the reference scenario, they equal 103%; in the high impact scenario, they equal 458%. 68 The perverse incentivization influences in HFC 23 crediting under th e CDM. First, the 66 Lambert Schneider Jakob Graichen, and Nele Matz, conventions. The Case of HFC 23 destructi on ( ko Institut e.V., 2005 ), 8. 67 Ibid, 9 10. 68 Ibid.

PAGE 27

21 authors suggest the development of a new version of AM0001 incorporating a CER discount rate, which could be accomplished either as a direct adjustment factor (e.g. 20% of earned CERs are issued) or an adjustment in the maximum w r atio cap. 69 However, they adjustment factor or the HFC 23/HCFC 23 revenues to HCF C 22 revenues and on the fluctuating market price of CERs. 70 The authors do propose that the problem of determining such a rate ex ante might be addressed in part through an ex post calculation based on real prices. 71 Second, the authors suggest that the fin ancial resources necessary for abatement of HFC 23 be provided outside of the CDM. Specifically, t he y mention preexisting funds such as the Stockholm Convention, the Montreal Protocol, and the Global Environmental Facility. This, they reason, would complet ely eliminate any perverse incentives, as HFC 23 would no longer be abated under the Clean Development Mechanism. 72 However, t he y also note that this strategy, if unaccompanied by a rate of return for the plants themselves, would eliminate the incentive for plants to seek out abatement measures. 73 and revenues in a manner similar to Schneider, et al. (2005) assuming a CER value 69 Ibid, 17 18. 70 Ibid, 18. 71 Ibid. 72 Ibid. 73 Ibid, 19.

PAGE 28

22 w and an abatement cost of $0.09/kg HCFC 22. 74 He concludes 22 producer is 75 The paper with the value of HCFC quarter 2005. 76 In response to this, Wara argues, some HCFC 22 facilities appear to have increased their production at a higher rate than the expected rate of growth of 15% in the developing world HCFC 22 sector. 77 In August 2007, the Technology and Economic Assessment Panel (TEAP) of Issues (with particular focus on the impact of the Clean Development Mechanism) and Emissions Reduction Benefits Arising From Earlier HCFC Phase Out and Other a HFC 23/HCFC 22 revenue comparison that has been commonly referenced in the argument against HFC 23 crediting The comparison assum es a CER value range of $3 $10 and an HCFC 22 value range of $1 $2/kg and proceeds to compare revenues at w levels of 1.5%, 3.0%, and 4.0%. 78 23 destruction could easily exceed the revenue from H CFC 79 Assuming thermal oxidation startup costs at $2 8 million and annual operating costs of oxidation at $189,000 74 Michael Wara, ( Stanford Unive rsity, 2006 ) 75 Ibid. 76 Ibid. 77 Ibid. 78 UNEP Technolog y and Economic Assessment Panel, (with particular focus on the impact of the Clean Development Mechanism) and Emissions Reduction Benefits Arising From Earlier H CFC Phase O 58. 79 Ibid, 57.

PAGE 29

23 on under the 80 The TEAP report also notes several trends in the production of HCFC 22, primarily a shift in production from developed to developing nations and an increased emphasis on p roduction of HCFC 22 for feedstock rather than emissive uses (both likely in response to the impending phase out of HCFC 22 in developed nations). 81 Even under the (then hypothetical) accelerated phase out schedule, the paper postulates: d require a plan to identify which HCFCs are to be phased out first and which production facilities for HCFC 22 are to close first. In the case of growth scenarios during 2007 2015, it will be very likely that the majority of HCFC 22 facilities that generate CERs can be kept in operation until 2027 2028 (thus generating CERs), even if they would have completely shifted to HCFC 82 e ways of solving the potential market distortion created by the CDM, since commitments are already in 83 phase out of HCFC 22 under the Montreal Protocol the report concedes 84 : 22 phase out is not expected to have any significant bearing on HFC 23 emissions in the first contracted period of the CDM and, in the absence of any measures to control HCFC 22 production for feedstock production, the CDM itself is t he only reliable mechanism available to prevent HFC may lie in the development of an intergovernmental agreement of all developing 80 Ibid, 58. 81 Ibid, 33. 82 Ibid, 53. 83 Ibid, 111. 84 Ibid, 7.

PAGE 30

24 countries host ing or planning to host HCFC 22 production facilities, in which 85 worst of all cases would be for HFC ts 86 Panel asserts that twelve plants stopped production when the HCFC 22 c ap was reached, two plants produced less than their HCFC 22 cap, and six plants produced more than their HCFC 22 cap, though it does not elaborate on the methods behind these observations. 87 The implication, of course, is that the plants stopped production of HCFC 22 in those periods due to the CER crediting cap. The Panel notes the behavior of Project 1105, which exhibited a w of 1.1% during one MP, as a further optimized and w [ Best Available Technology (BAT)] levels may have further 88 22 plants that produced up to the maximum level allowed by the methodology would have produced less in the absence of the CDM, becaus 89 Further, it posits: 22 could fall below the sum of the production caps of CDM HCFC project activities. In this situation, CDM HCFC 22 plants could 85 Ibid, 111. 86 Ibid, 14. 87 (UNFCCC, 2010), 2 88 Ibid, 4. 89 Ibid.

PAGE 31

25 have an incentive to produce HCFC 22 up to their c ap, because of the CER 90 HCFC 22 CDM plants beyond their normal lifetime or not improve the efficiency of the plant in the w ratio during any refurbishment because o 91 In November 2010, the Natural Resources Defense Council (NRDC) published Curb HFC 23 an by Stephen O. Andersen and K. M adhava Sarma. The authors assert: projects currently supported under the CDM are receiving vastly greater compensation than required to pay for installing and operating the technology for inci nerating HFC 23 [based on a DuPont estimate of $0.28 0.37/tCO 2 eq], and t his over compensation is perversely encouraging excess production of both HFC 23 and HCFC 92 More specifically, they asserts to maintain or even increase their ratio of HFC 23 byproduct to HCFC 22 production there is evidence that perverse CDM incentives encourage project operators to overproduce HCFC 22, above levels that otherwise would be produced just in response to underlying HCFC 22 demand, in order to maximize HFC 23 fo 93 Their graph from the UNEP TEAP report discussed earlier. Th e y also cite the Meth Panel report observation regarding plants moderating production based on HCF C 22 crediting caps; however, unlike the Meth Panel report, the authors specifically cite a project at fault: Zhonghao Chenguang Research Institute of Chemical Industry, China 90 Ibid, 2. 91 Ibid, 3. 92 23 and Other 93 Ibid, 12.

PAGE 32

26 reduced a nd then completely stopped its HCFC 22 production once it had reached the cap. The following year, the same plant once again reduced its production of HCFC 22 by 85 percent once it had reached its CDM cap. In both cases, the facility immediately resumed HC FC 22 production once the 94 They compensation for HFC 23 destruction is driving decisions whether or not to produce HCFC lants are producing HCFC 22 well above the amount warranted by market demand and pricing for HCFC from HCFC 95 Similarly, the authors M plants operate at much higher HFC 23/HCFC 22 ratios than non operate at higher HFC 23/HCFC 22 ratios at times of the year when they are earning 96 As evidence, they cite the case of Pr oject 23/HCFC 22 ratio decreased from an average of 2.9 percent during the crediting 97 They also note lower rates demonstrated by other CDM certified plants, primarily the Arkema percent in 2003, and 1.84 percent in 2004 and is capped for credit generation purposes at a maximum ratio 98 94 Ibid, 13. 95 Ibid. 96 Ibid. 97 Ibid. 98 Ibid.

PAGE 33

27 The authors also claim that HFC 23 CERs are non additional on the grounds that, despite CER issuance incentivizing incineration of HFC HFC 23 product would never have been created 99 In essence, they suggest that part of the abatement is additional, but the presence of abatement incentives causes the generation and abatement of emissions outside the baseline (which would be non additional). They also posit that if the abatement occurs as a result of waste HCFC 22 production from shifted developed countries, those reductions would also be non additional. 100 The paper recommend s the following measures be undertaken through the CDM: suspending issuance of credits to HFC 23 abatement projects u ntil AM0001 is revised; suspending renewal of HFC 23 abatement projects until AM0001 is revised; of HCFC 22 plants. 101 In terms of the Montreal Protocol, it recommends: controlling HFC of replacements for ozone 22 phase out to feeds and using the Montreal Protocol Multilateral Fund to support the phase out of HCFC 22 and destruction of HFC 23. 102 In August 2010, Lambert Schneider released another study on HFC 23 99 Ibid. 100 Ibid, 14. 101 Ibid, 17 18. 102 Ibid, 18 19.

PAGE 34

28 (CDM): an evaluation of HFC the CDM in April 2010 in a preliminary form as a Methodology Revision Request. 103 This is perh aps the most comprehensive publicly available study of the perverse incentivization controversy. In the study, Schneider analyzes the possibility that operators to produce more HCFC 22 and/or generate more HFC 23 than they would 104 To accomplish this, the study analyzed a total of 163 MRs from the nineteen certified HFC 23 projects, ranging from the beginning of crediting to February 2010. Schneider finds that w va lues span a wide range acro ss and within the projects, not HFC 23/HCFC 22 ratio significantly without any incentives offered by the CDM or any form of regulations while other plants had a relatively 105 In scatter plot form, Schneider demonstrates that reported w ratios from 2000 to 2004 exhibited a wider range than reported w ratios during crediting, HFC 23/HCFC 22 ratios that just met or slightly exceeded the cut off value 106 He highlights the case of Project 0306: 23/HCFC 22 ratio varied between 3.17% and 3.65% in the 2002 to 2004 period and stayed in a rathe r narrow range between 2.88% and 2.91% after the implementation of the CDM project just above the cut 107 103 Lambert Schneider, evaluation of HFC ( ko Institut e.V., 2010 ) 104 Ibid. 105 Ibid. 106 Ibid. 107 Ibid.

PAGE 35

29 23/HCFC 22 ratio that tends to just meet or slightly exceed the cut off value est ablished in the PDD can also be observed for projects 11, 108 Project 449, he observes, initially achieved a lower than cap w but has subsequently demonstrate d a consistently higher than cap w. 109 A operate at a HFC 23/HCFC 22 ratio just above the cut Schneider suggests that this behavior raises the question as to whether the plants were intentionally manipulating the w rat io to ensure an excess over the cap. 110 The article does note that Projects 0001, 0003, and 0115 exhibited occasionally lower than cap w values; Schneider hypothesizes that these projects might be unaware of the economic incentives associated with manipulati on of the w ratio. 111 Schneider deems the most interest ing result of the evaluation to be an analysis of the two MPs for which no CERs were issued due to production in excess of the HCFC 22 cap. For these two MPs (belonging to Projects 0151 and 1105), Schneide r demonstrates that the w ratio noticeably decreases before increasing once again at the beginning of the next MY. Schneider posits that in the absence of economic incentives for the maintenance of a high w the plants adjusted to a lower level to increase HCFC 22 production efficiency. 112 In summary, Schneider writes that the current approach to capping w 113 108 Ibid. 109 Ibid. 110 Ibid. 111 Ibid. 112 Ibid. 113 Ibid.

PAGE 36

30 The study follows with an analysis of HCFC 22 production trends As in his o bservations on w trends, Schneider demonstrates that several plants consistently produced HCFC 22 in quantities almost exactly equal to their creditable caps, again 114 Perhaps most damningly, Schneider scrutinizes Project 0767: 22 production of about 15 20 tons per day until the HCFC 22 amount eligible for crediting was reached. From that point onwards, the production was reduced or the plan t was even shut down (in April 2008). Production then resumed at the start of the next crediting year (1 May). Plant operators apparently had no incentives to produce HCFC 22 during times when no CERs could be gained from generating and destroying HFC 115 However, Schneider also notes that Projects 0115 and 0807 produced less than the capped HCFC 22 amount for several years, indicating that either the revenues from additional creditable HFC 23 abatement did not exceed the costs of excess HCFC 22 productio n or the projects were unaware of the economic incentives associated with further production. 116 Still, Schneider once again concludes that the behavior of the 117 Schneider also posits that HFC 23 CERs could be driving down the price of HCFC 22. To explain this, the study cites a price decrease from 15,000 to 8,000 seasonal variation 22 reactants such as chloroform and hydrogen fluoride had increased at the time, indicating a dissonance between 114 Ibid. 115 Ibid. 116 Ibid. 117 Ibid.

PAGE 37

31 production costs and pricing trends. 118 If this was indeed the case, Schneider concludes, the CDM could be generat ing a substitution effect by creating favorable conditions for HCFC 22 consumption versus consumption of (perhaps more sustainable) alternatives. 119 Schneider outlines four possible options for preventing perverse incentivization in HFC 23 abatement: impleme nting a discount rate on CER issuance to HFC 23 abatements projects; seizing revenues in excess of the incremental costs of abatement for use in a sustainability fund; financing HFC 23 abatement projects through an independent intermediary institution; or using a multilateral institution fund HFC 23 destruction independently of the CDM. 120 options pursue the same generic approach: they take the large profits from CER 121 Schneider explicitly f ocuses on the first option a discount rate on CER issuance 122 In particular, he favors a discount rat e achieved through the use of a default rather than dynamic emission benchmark for HCFC the profits from CER revenues are signific antly lower than the production costs for HCFC 123 118 Ibid. 119 Ibid. 120 Ibid. 121 Ibid. 122 Ibid. 123 Ibid.

PAGE 38

32 survive such a policy, he argues that a sufficient baseline would ensure that actual abatement would exceed CER issuance. 124 As in his 2005 125 Schneider takes a slight leap in his claim that any benchmark should not exceed 1%, which he bases entirely on the low w ratio exhibited by Project 1105 which had the lowest w cap of any project during its third MP. 126 Somewhat arbitrarily, Schneider suggests a 0.2% w benchmark, which he substitutes into his revenue comparison (discussed in the literature review) to show a (relatively unsurprising) drastic improvement in the potential for perverse incentivization under his model for price ranges from $10 $30. 127 As this report was submitted as a Methodology Revision Request and supported by a number of prominent NG Os, this is perhaps the most prominent external proposal for revision of AM0001. 128 Also in August 2010, the Environmental Investigation Agency (EIA) published a scathing criticism of HFC 23 pCER purchases by the World Bank titled Bank Defense of the HFC of 1.0 1.4% from 2005 and comparing them to the average w cap of 2.75% 129 the thirds of the HFC 23 production occurring at CDM plan ts is unnecessary, and therefore half to two thirds of the credits generated are not legitimate offsets and of [sic] no real 124 Ibid. 125 Ibid. 126 Ibid. 127 Ibid. 128 Ibid. 129 k Defense of the HFC 23 Scandal ( Environmental Investigation Agency 2010 ), 6

PAGE 39

33 130 w rate adjustment and interestingly notes a statement from the PDD of Project 0306: the value of w it is likely that the project will decompose more HFC 23 than the quantity of HFC 23 131 The report observes th at despite its own prediction, Project 0306 produced w ratios almost exactly equal to the cap from 2007 to 2010. 132 In a brief HFC 23 cost revenue analysis, EIA assumes w at 2.88%, HCFC 22 market value at the time of writing), and CER value at 133 22 revenues. 134 t HCFC 22 reactant prices may have risen in conjunction with HCFC 22 value, widening the apparent revenue gap. 135 The report does acknowledge that these estimates reflect sCER prices rather than pCER prices, which are expected to be significantly lower based 136 Nevertheless, it notes: revenues to corporations and the contributory effect of these revenues on the lifetime of HCFC 22 plants and pr oduction levels, SRF Chemicals in India reported that CER revenues made up to 63% and 66% of their entire Gujarat Fluoro Chemical project reported that CERs accounted for 88% of corpor 137 130 Ibid, 7. 131 Ibid. 132 Ibid. 133 Ibid, 10. 134 Ibid. 135 Ibid. 136 Ibid, 10 11. 137 Ibid, 11.

PAGE 40

34 23 Destruction Methodology is 138 138 Ibid, 13.

PAGE 41

35 A NALY SES O VERVIEW OF P ROJECTS Beginning with the submission of the PDD for Project 0001 in November 2003, and culminating in the submission of the PDD for Project 1947 in March 2009, nineteen projects have been certified for credit issuance under AM0001 (see Figure 1 on the following page) 139 Of t hese nineteen plants, just five the Indian facilities registered for a single ten year crediting period, with the remaining fourteen opting for the twice renewable seven year crediting period. The plants are highly concentrated: eleven are located in C hina, five in India, one in South Korea, one in Mexico, and one in Argentina. 140 HCFC 22 production caps and w ratio caps vary widely across projects, as do the production capacities of the plants in question. In some instances, the PP has opted to keep cer tain information confidential: eight plants did not provide production capacity information in their PDDs, six plants did not provide historical HCFC 22 production levels, and one plant did not provide historical w values. 141 For the most part, plants did pr ovide HCFC 22 production levels, HFC 23 generation levels, and credited HFC 23 during each MP. The nineteen plants often differ widely in the conditions and patterns of production, among other factors. For these reasons, the study will examine each plant o n an individual level rather than considering the aggregate data of the nineteen plants as a whole. 139 UNFCCC http://cdm.unfccc.int/ 140 Ibid. 141 Ibid.

PAGE 42

36 Project Location Project Registration Start of Crediting Type of Crediting Capacity Cap 1 Cap 2 0001 India 03/2005 02/2006 One time 27,375TPA 27,375TPA 2. 90% 0003 South Korea 03/2005 01/2003 Renewable 7,500TPA 6,914TPA 2.90%, 2.77% 0011 China 06/2006 12/2006 Renewable 30,000TPA 25,149TPA 2.86% 0115 India 12/2005 07/2004 One time 11,145TPA 2.94% 0151 Mexico 06/2006 06/2006 Renewable 7,570TPA 2.44% 0 193 China 03/2006 08/2006 Renewable 16,000TPA 12,866TPA 3.00% 0232 China 03/2006 01/2007 Renewable 36,476TPA 2.37% 0306 China 08/2006 12/2006 Renewable 40,000TPA 30,979TPA 2.88% 0499 India 02/2007 02/2007 One time 1,695TPA 2.72% 0549 China 10/2006 11/2006 Renewable 18,000TPA 10,420TPA 3.00% 0550 China 10/2006 01/2007 Renewable 17,000TPA 13,634TPA 3.00% 0767 China 05/2007 05/2007 Renewable 6,000TPA 5,887TPA 3.00% 0807 Argentina 03/2007 10/2007 Renewable 4,087TPA 3.00% 0838 India 03/2007 05/2007 One time 7,992TPA 3.00% 0868 China 04/2007 04/2007 Renewable 13,709TPA 3.00% 1105 China 02/2008 05/2008 Renewable 18,107TPA 1.64% 1194 China 09/2007 09/2007 Renewable 12,285TPA 2.96% 1867 India 11/2008 11/2008 One time 1,265TPA 1,265TPA 3.00% 1947 China 04/2009 04/2009 Renewable 25,000TPA 23,269TPA 2.89% F IGURE 1 HCFC 2 2 PRODUCTION FACILITIE S CREDITED UNDER AM0001

PAGE 43

37 E STIMATING R EVENUES As a first step, it is valuable to estimate revenues from HCFC 22 production versus CER issuance under the currently operational methodology This is a key determinant of perverse incentivization: if CER prices do not constitute a significant portion of total production line revenues (the sum of total HCFC 22 value and total CER value), then the likelihood of perverse incentivization in terms of CER issuance driving HCF C 22 production is relatively low. If, however, total CER value compares favorably to total HCFC 22 value, then the likelihood of such perverse incentivization is correspondingly higher, as decisions are more likely to be driven by the desire for HFC 23 ge neration rather than HCFC 22 production. The HCFC 22 to CER value comparison is less relevant to perverse incentivization in terms of manipulation of the w ratio: in order for a perverse incentive in terms of w to exist, CER revenues need only excee d the costs of abatement and any additional reactants expended. Based on the various estimates of abatement costs provided in the literature review, CER revenues unarguably exceed the cost of abatement even at the lowest CER price levels. Assuming relative ly low amounts of additional reactants are expended in the production of additional HFC 23, it is then safe to assume that there is a potential for perverse economic incentivization in the manipulation of the w ratio. Accordingly, analys e s should primarily focus on comparing HCFC 22 value to CER value. As is evident from the literature review previous analyses have primarily focused on theoretical value comparisons, in which the author assumes either a static value or broad range for w HCFC 22 value, and CER value. These comparisons

PAGE 44

38 are well established, and it is clear that at most (if not all) points during crediting thus far, an HFC 23 abatement proj ect being credited immediately for its reductions could theoretically expect CER revenues in excess of HC FC 22 revenues. This study will attempt a slightly di fferent analysis by estimating real revenues experienced by each of the nineteen plants. Real value estimates must take into account a wider and more complex set of factors than theoretical value estimat es. First and foremost, an estimate of real value must account for price changes in both HCFC 22 and CER sales. Unfortunately, real sales data is not readily available for either factor. In this study, both total HCFC 22 value and total HFC 23 value are ap proximated and compared on a MY to MY basis: due to the annual nature of the HCFC 22 cap, it is likely that any perverse incentivization in HCFC 22 production would be driven by expected revenues during a MY rather than by expected revenues during a month or calendar year. With this in mind, the total estimated value of HCFC 22 production during a MY must be found. First, the cumulative production of HCFC 22 by the plant during e is then multiplied by the average market price of HCFC 22 during that year. This second value has been approximated from market data on sales of HCFC 22 by Chinese firms; as Chinese HCFC 22 production facilities comprise the majority of HCFC 22 productio n in developing countries 142 it is likely that the Chinese sale price of HCFC 22 is a relatively sound indicator of HCFC 22 value at the non Chinese facilities. The data utilized ranges from January 2008 to May 2011; however, in 142 Ib id.

PAGE 45

39 certain cases, production sl ightly outside the range of the data was associated with data from the closest available date. As the data was found in CNY, the historical EUR/CNY conversion rate was applied to monthly HCFC 22 prices to standardize the price comparison. This value is th en compared to the value of CERs generated by way of HCFC 22 production during the same MY. This is a much more tumultuous variable: pCER in long term, high volume agreem ents with investment firms. 143 The most commonly used substitute sCER price (typically given as either the price at the time of writing or as a price range) is, therefore, much more representative of cost to end consumers rather than value to producers. Unfortunately, providing a more accurate estimate is extremely difficult without direct data from investment firms or the CER producers themselves: as such, sCER price has been used to estimate revenue to the CER producers. It should be noted that this est imation makes three major assumptions: first, that pCERs are sold immediately upon issuance; second, that pCERs are sold for their corresponding sCER market price; and third, that the price of HFC 23 sourced CERs does not vary significantly from the price of non HFC 23 sourced CERs. Accordingly, this analysis though more realistic, nevertheless represents an upper estimate of CER revenues. The quantities of i ssued CERs are multiplied by the sCER price corresponding to their date of issuance ( not their date of generation or request). This accounts for distributional effects, as the CDM does not issue credits upon receipt or approval of 143 Envi ronmental Investigation Agency, 23 ( Environmental Investigation Agency, 2010 ), 2.

PAGE 46

40 a MR, but rather times its credit issuances based on project external factors. Assuming that any purchasing contracts have c ost adjustment mechanisms, this is highly relevant to the real revenue experienced by any given project. Use of the real date of issuance should also slightly offset the overestimation effect created by use of the sCER price, as CER price has declined stro ngly over time. The total value of CERs issued for HFC 23 abatement during the MY in question is then calculated. If the project is based in China, a tax of sixty five percent is applied to this value in accordance with the Chinese tax on CER revenues. 144 Th ese final values are then compared with the previously established total value of HCFC 22 production during the MY. This creates a more realistic picture of the percentage of total production line revenue that is comprised by CER sales. The results of this analysis are illustrated below in Figure 2 where the listed percentages represent estimated CER revenues as a percentage of estimated HCFC 22 revenues Project Location Year Ratio Year Ratio Year Ratio 0001 India 02/2008 02/2009 413% 02/2009 02/2010 400% 02/2010 02 /2011 126% 0003 South Korea 01/2008 12/2008 384% 01/2009 12/2009 403% 0011 China 12/2007 11/2008 149% 12/2008 11/2009 148% 12/2009 11/2010 61% 0115 India 07/2008 06/2009 379% 07/2009 06/2010 209% 07/2010 06/2011 95% 0151 Mexico 06/2008 06/2009 179% 06/2009 06/2010 70% 0193 China 08/2008 07/2009 138% 08/2009 07/2010 95% 0232 China 01/2008 12/2008 123% 01/2009 12/2009 126% 01/2010 12/2010 51% 0306 China 12/2007 12/2008 117% 12/2008 12/2009 151% 12/2009 12/2010 35% 0499 India 02/2008 02/2009 3 34% 02/2009 02/2010 367% 02/2010 02/2011 155% 144 UNEP Technolog y and Economic Assessment Panel, (with particul ar focus on the impact of the Clean Development Mechanism) and Emissions Reduction Benefits Arising From Earlier HCFC Phase O

PAGE 47

41 Project Location Year Ratio Year Ratio Year Ratio 0549 China 11/2007 10/2008 103% 11/2008 10/2009 109% 11/2009 10/2010 48% 0550 China 01/2008 12/2008 108% 01/2009 12/2009 126% 01/2010 12/2010 58% 0767 China 05/2008 04/2009 134% 05/2009 04/2010 124% 05/2010 04/2010 48% 0807 Argentina 10/2007 10/2008 308% 10/2008 10/2009 507% 10/2009 10/2010 223% 0838 India 05/2008 04/2009 326% 05/2009 04/2010 295% 05/2010 04/2011 130% 0868 China 04/2008 04/2009 119% 04/200 9 04/2010 113% 04/2010 04/2011 49% 1105 China 05/2008 04/2009 34% 05/2009 04/2010 57% 05/2010 04/2011 18% 1194 China 09/2008 09/2009 143% 09/2009 09/2010 88% 1867 India 1947 China 04/2009 04/2010 114% 04/2010 04/2011 76% F IGURE 2 HCFC 2 2 PRODUCTION FACILITIE S CREDITED UNDER AM0001 Apa rt from on e plant in India with CER revenues estimated at 95% of HCFC 22 revenues during its third available year, all non Chinese plants consistently exhibit CER revenues greater than HCFC 22 revenues, with the highest estimate placing CER revenues at five times HC FC 22 revenues for one plant. The Chinese plants, on the other hand, show a much greater tendency for CER revenues to be estimated at less than HCFC 22 revenues, with the propensity increasing over time (in accordance with the increasing price of HCFC 22 a nd plummeting price of CERs). This demonstrates that the Chinese plants do show a greater tendency to produce HCFC 22 revenues in excess of C ER revenues. Given the high tax on CER revenues in China, this is unsurprising. It should also be noted that this represents a near worst case scenario for perverse incentivization: as illustrated in Appendix I I HCFC 22 prices have experienced a tremendous increase over the last two years and CER prices have

PAGE 48

42 plummeted. Since both feedstock and emissive uses account f or the increased HCFC 22 demand from developing countries, as emissive uses begin to truly phase out in subsequent crediting periods, HCFC 22 prices are likely to fall. If, as the CDM desires, faith in the CDM increases and the supply of CERs decreases, CE R prices are likely to significantly increase in price. These factors would return the ratio at Chinese facilities to a CER heavy state. However, comparative revenues do not fully answer the question of whether the potential for perverse incentivization ex ists Even if HCFC 22 revenues dramatically exceeded CER revenues on the whole, it could still be possible for CER revenues to drive HCFC 22 production up to the cap if HCFC 22 revenues diminished. As such, it is important to compare revenues between years as well. Doing so, it is evident that the decreasing percentage of revenues comprised by CER sales are primarily due to increasing HCFC 22 revenues rather than decreasing CER revenues, and CER revenues even at the Chinese plants appear to remain som ew hat steady from year to year (t hese results are illustrated in Appendix II I ) If this is, in fact, the case, then prospective revenues from CERs could remain sufficient to justify CER production if HCFC 22 prices fell.

PAGE 49

43 A SSESSING A BERRANT B EHAVIOR IN HCFC 22 P RODUCTION Previous analyses of HCFC 22 production rates have been surprisingly few and far between; most conclusions have been drawn either from the general proximity of cumulative MY HCFC 22 production to the MY cap or from isolated examples of decr eased production in response to the HCFC 22 cap being reached (most notably in the case of Project 0767). These methods are not sufficient to draw conclusions regarding the HCFC 22 production r esponse to CER incentivization across all projects. This analy sis provides a more comprehensive analysis of HCFC 22 production rates. For each firm, historical per day HCFC 22 production rates (graphed as HCFC22_d) have been charted on an MP by MP basis and as average values for each MY (graphed as HCFC22_cp). These lines are also used to illustrate pre crediting period production rates, which are easily differentiated by their lack of a corresponding HCFC 22 production cap. HCFC22_d and HCFC22_cp are juxtaposed with the HCFC 22 cap in per day form (graphed as HCFC22_ mp): while HCFC22 is not capped on a per MP basis, it is valuable to assess how production during any given period compares to a level of consistent production that would directly align with the HCFC 22 production cap. Perhaps more novel is the inclusion o available), again on a per day level (graphed as HCFC22_pc). production during each MY (graphed as HCFC22_c) against the real producti on cap (graphed as HCFC22_m) and the real reported production capacity (graphed as

PAGE 50

44 HCFC22_p). This graph better illustrates the points (if any) where production exceeded the creditable amount for the MY. At points of excess over the cap, month to month dat a from the MR(s) in question has been assessed (if available). The above data and graphs have been analyzed with the following expectations in mind: 1) If CER generation allows the Project to turn a profit from HCFC 22 production by way of HFC 23, but HCFC 22 production itself is unprofitable at the capped level HCFC 22 production for any given year to precisely meet the creditable production cap and demonstrate relatively stable production levels. 2) If, in the above sce nario, the Project did exceed its creditable production cap, one would expect HCFC 22 production to subsequently decrease until the beginning of the next MY, as HCFC 22 generation would no longer provide the revenues necessary to incentivize production. 3) If both CER generation and HCFC 22 sa les are sufficient to justify at least the capped level of HCFC HCFC 22 production to adjust upward of the production cap in response to price and demand increases, but never to fall beneath the cap. 4) If HCFC 22 sales are sufficient to justify production, but CER sales are not, one would expect HCFC 22 production to adjust entirely in response to profit and demand fluctuations without regard for the production cap. As the caps are based on historical levels of production, it is likely that such

PAGE 51

45 projects would exhibit years with aggregate production below the creditable production cap. The analysis makes two major assumptions. First, it is assumed that reactant and operation cost s are static, leaving HCFC 22 prices as the major determinant of profit per ton produced. Second, it is assumed that demand for HCFC 22 produced in developing countries is steadily increasing. This is based on the TEAP Report conclusion that production is steadily shifting from developed countries to developing countries and that production on the whole is not significantly decreasing as of yet. 145 In general, this leads to the assumption that the price of HCFC 22 has increased over time, with strong increase s in recent years. The reported production capacity of the plant, if provided, may hinder the applicability of the analysis if it is close to the creditable production cap, as proximity between the two values limits the amount to which production can respo nd in an upward direction to price fluctuations. As such, trends that may, at first, appear to exhibit strong proximity to the creditable production cap may in fact simply show the Project operating close to its full production capacity. 145 UNEP Technolog y and Economic Assessment Panel, (wi th particular focus on the impact of the Clean Development Mechanism) and Emissions Reduction Benefits Arising From Earlier HCFC Phase O

PAGE 52

46 P ROJECT 0001 Proje ct 0001 (see Figures 3 and 4) exhibit ed relatively steady increases in HCFC 22 production from MY 1 to MY 6, beginning with a sharp jump in production from the first MP of MY 1 to its second MP. HCFC 22 production rates have fluctuate d from MP to MP, most notably at points in m id year. This would seem to indicate production fluctuation with HCFC 22 price or demand. HCFC 22 production appears to have decrease d in mid 2009, when HCFC 22 experienced a significant decrease in price, and reache d very close to the cap when HCFC 22 app roached historically high price levels in 2011. P re crediting production levels are not available. During no MP or MY has ed its creditable production cap. Project 0001 is, however, the only plant to be explicitly capped base d on maximum production level rather than historic production level As noted in the introductory section, this limits the applicability of the analysis, as it is not more H CFC 22 or more CERs. Given the steady (rather than sharp) increase in HCFC 22 production over the six year period and the complete lack of aggregate yearly production values at or in excess of the cap, it seems unlikely that HCFC 22 production at Project 0001 is being driven by CER generation prospects.

PAGE 53

47 F IGURE S 3 & 4 HCFC 22 PRODUCTION AT P ROJECT 0001

PAGE 54

48 P ROJECT 0003 Project 0003 (see Figures 5 and 6) exhibited a sharp increase in production from its pre crediting values and initial MYs to its later MYs. Prior to MY 1, the production trend seemed to be downward, and the pro duction cap was set at a very early 2000 value. From MY 2 to MY 3, production more than doubled; from MY 3 to MY 4, production nearly doubled again. This rapid growth does not seem to be in line with the relatively steady increases in price of and demand f or HCFC 22. Though production wa s well below cap in MYs 1, 2, and 3, production from MY 4 onward has almost exactly met the HCFC 22 production cap (most notably in MY 7, which showed a difference of less than two tons) with the exception of MY 8 (which sho w ed an aggregate production value ten percent greater than the creditable production cap and aligns with the significant price increases in HCFC 22 in 2011). It should be noted that production did not decline suspiciously when the creditable production cap was reached. The low production values during the early years of crediting could potentially be explained by a lack of faith in or awareness of the financial returns associated with CER generation (as suggested by Schneider). Given the sharp increase in H CFC 22 production from year to year and the likely that production of HCFC 22 has, at least to an extent, been driven by CER generation prospects at Project 0003. However, it remains p robabl e that HCFC 22 production itself is still a major source of revenue, as production levels vary substantially throughout the year and production appears to have increased in response to price increases in 2011.

PAGE 55

49 F IGURE S 5 & 6 HC FC 22 PRODUCTION AT P ROJECT 000 3

PAGE 56

50 P ROJECT 0011 Project 0011 (see Figures 7 & 8) has exhibi t ed a generally upward trend in HCFC 22 production from MY to MY, with large increases in yearly production prior to crediting followed by a relatively consistent rate of production during crediting. However, aggregate yearly production noticeably decrease d in MY 5, which to MP production rate has been highly erratic, showing strong fluctuations. The Project has exceeded its creditable production cap during all MYs thus far b y a considerable amount with the exception of MY 5. While the Project did display significant production rate drops at or near the creditable production cap during MYs 2, 3, and 4, it display ed a drop of similar magnitude at the beginning of MY 2 without a corresponding drop at the end of MY 1, indicating that this may have been a seasonal shift in production rather than a response to the production cap. Given the upward trend in production prior to crediting and the consistently higher than cap production levels, it seems likely that HCFC 22 production remains profitable at the facility However, aggregate production has not align ed with the upward trend expected, and it is unclear whether the noticeable drops in production at the end of MYs 2, 3, and 4 we re seasonal or responsive to CER generation prospects. Overall, it appears unlikely that CER generation prospects have driven a 22 production.

PAGE 57

51 F IGURE S 7 & 8 H CFC 22 PRODUCTION AT P ROJECT 00 1 1

PAGE 58

52 P ROJECT 0115 Project 0115 (see Figures 9 and 10) has exhibit ed a highly consistent HCFC 22 production rate. While pre crediting production rates are not available, production increase d noticeably from MY 1 to MY 2 before stabilizing. Variation within MYs also stabilize d over time, from a state of relative fluctuation in MY 2 to a point of ve ry low fluctuation rate in later years. While production in MY 1 was noticeably below the production cap, subsequent production has almost exactly matched the creditable production cap. Though it is po ssible that the production cap approximates the production capacity of the plant (which is not provided in its PDD), it remains unlikely that natural production would so precisely meet that specific point. Given the incredibly and increasingly consiste nt production levels within and between years following lower and less stabilized levels at the beginning of crediting, it appears highly likely that HCFC 22 production at Project 0115 has, to an extent, been driven by the generation of CERs.

PAGE 59

53 F IGURE S 9 & 10 HC FC 22 PRODUCTION AT P ROJECT 0 115

PAGE 60

54 P ROJECT 0151 Project 0151 (see Figures 11 and 12) has exhibit ed fairly consistent production from its pre crediting period to its crediting period. There was a steady increase in aggregate production during the pre crediting period, culminating in the available historical prod uction/creditable production cap. Variation from MY to MY has been minimal, with slight increases in 2010 and 2011 that align with HCFC 22 price increases. Variation from MP to MP has been low to moderate, with occasional exceptions (discussed in the next p aragraph). Production capacity is unavailable. met or slightly exceeded the creditable production cap. In MY 1, MY 2, and MY 3, the Project exhibited severe production drops once th e creditable cap was reached, followed by immediate and sharp increases in production once crediting resumed. While similarly dramatic drops are not easily observed in MY 4 or MY 5, those MYs display ed signs of <1 month shutdowns near the end of the MY (in March 2010 and February 2011, respectively). Though the level of production itself is not necessarily suspect, as it has align ed adherence to the creditable production cap com bined with its highly suspect production drops after the creditable cap suggest that it is highly likely that a significant portion of HCFC 22 production is driven by CER generation prospects at Project 0151. However, given the recent increases over the ca p, it seems probable that some HCFC 2 2 production remains profitable.

PAGE 61

55 F IGURE S 11 & 1 2 HC FC 22 PRODUCTION AT P ROJECT 0 1 51

PAGE 62

56 P ROJECT 0193 Project 0193 (see Figures 13 & 14 ) exhibit ed an increase in production during its pre crediting years, culminating in its highest available historical value/creditable production cap. Duri ng the crediting period, the plant appears to have operate d consistently at or above its reported production capacity, with regular MP to MP fluctuations throughout the year. Project 0193 has exceeded its creditable production cap during every MY thus far by a notable amount. The Project has displayed no unusual drops in the per day production rate in response to said excesses over the cap; in fact, most decreases in production appear to have occurred mid year. At no point, however, has the plant drop ped be neath the per and lack of response to excesses over that cap, it seems probable that HCFC 22 production is significantly profitable and thus unlikel y that production is being driv en by CER generation prospects.

PAGE 63

57 F IGURE S 13 & 1 4 HC FC 22 PRODUCTION AT P ROJECT 0 1 93

PAGE 64

58 P ROJECT 0232 Project 0232 (see Figures 15 and 16) has exhibit ed a noticeably but not dramatically increasing production rate from MY to MY and a strongly fluctuating production rate from MP to MP. Aggregate yearly production has spiked in the most recent two MYs, which aligns with the expected response to price increases in HCFC 22. Production capacity is unavailable. The Project adhered strictly to its creditable production cap in MYs 1, 2, and 3, but showed a small increase over the cap in MY 4 and a large increase over the cap in MY 5. Though there are signs of a significant production decrease during November 2010 (after the crediting cap had been reached for MY 4, but before the end of the MP), there are no such signs for MY 5, indicating that the production over the cap was, in fact, desired. Given the proximity of real production to the cap until MYs 4 and 5, (coinciding with the sharp increase in price of HCFC 22) and the response to the excess over the cap in MY 4, it seems likely that HCFC 22 production has, to an extent, been driven by CER generation prospects. However, it appears that HCFC 22 production remains profitable.

PAGE 65

59 F IGURE S 15 & 1 6 HC FC 22 PRODUCTION AT P ROJECT 0 232

PAGE 66

60 P ROJECT 0306 Project 0306 (see Figures 17 and 18) has exhibit ed with the exception of MY 2, a c onsistently increasing yearly aggregate production rate. Production has seemed to show a trend of dropping significantly toward the end of each year, with the exception of MY 2, where production dropped before increasing at the end of the MY. Production ha s approached and exceeded the reported production capacity of the plant. Production by the Project in each MY has significantly exceed ed the creditable production cap. As noted before, production has tended to drop off after the cap has been reached, thoug h not to an extreme degree in fact, the production MP 2 of MY 1. complete lack of adherence to the creditable production cap, it seems unlikely that HCFC 22 production has been driven by CER generation prospects. Though there are drops in production after the crediting limit has been reached, the subsequent rise in production in MY 2 and the relatively moderate nature of the drops raise doubts as to whether the decrease was due to the drop in crediting. Such manipulation of the production rate would not be rational, as there is no real incentive to reaching the cap as early as possible unless the PP planned to reduce production below the production cap rate in response to reaching the cap.

PAGE 67

61 F IGURE S 17 & 1 8 HC FC 22 PRODUCTION AT P ROJECT 0 306

PAGE 68

62 P ROJECT 0499 Project 0499 (see Figures 19 and 20) d id not make available its pre crediting rates of production, but has exhibit ed an extremely consistent rate of production omparatively low rate of production (production shifts that appear small may be more significant to the plant), but the plant has exhibit ed no significant increase s or decrease s in production throughout its crediting period. Production capacity is unavaila ble. The Project has almost exactly me t its creditable production cap during each MY; during all MYs thus far, the difference between the final aggregate production total and the production cap has been less than 2% of the cap. While month to month data is not available, it appears that in MY 4, the plant significantly lowered its production rate so as to not exceed the creditable production cap. apparent adjustment of otherwise hi ghly stable production to suit production up to that level, it seems highly likely that production at Project 0499 is substantially driven by CER generation.

PAGE 69

63 F IGURE S 1 9 & 2 0 HC FC 22 PRODUCTION AT P ROJECT 0 499

PAGE 70

64 P ROJECT 0549 Project 0549 (see Figures 21 and 22) has exhibit ed a steadily increasing yearly production rate of HCFC 22 from its pre crediting years to its current to MP production rate has varied significantly throughout crediting, but the rate of variation has decreased from MY to MY and rea ched a point of relative stabilization in the most recent MYs. The Project showed aggregate production very close to the cap in MY 1, but has exhibit ed consistent and substantial growth in production over the cap in each subsequent MY. In MYs 4 and 5, prod uction capacity was met, which was likely responsible for the relative stabilization in production in those years as compared to MYs 1, 2, and 3. There have been no notable drops in the production rate following excesses over the crediting cap. Given the c onsistent growth in production over the cap, which appears to be in line with production growth trends prior to the crediting period, and the meeting of the annual production capacity in the most recent years, it appears unlikely that HCFC 22 production at Project 0549 has been significantly driven by CER generation prospects.

PAGE 71

65 F IGURE S 21 & 22 HC FC 22 PRODUCTION AT P ROJECT 0 549

PAGE 72

66 P ROJECT 0550 Project 0550 (see Figures 23 and 24) has exhibit ed steady growth in HCFC 22 production from MY to MY, although it displayed a surprising downward trend in the aggregate production of MYs 4 and 5, which does not align with growth expectations. It show ed significant variation from MP to MP, generally aligning to a trend of increasing production throughout the year. The Project barely met its creditable production cap in MY 1, but has exceeded it in all subsequent years by a notable amount. Where production has exceeded the creditable cap, there have been no corresponding significant drops in production. Production instead seems to be increasing toward the production capacity, which was nearly been met in MYs 3 and 4. Given the general lack of adherence to the creditable production cap and the lack of substantive drops in production following excesses over the cap, it seems unlikely that HCFC 22 production has been driven by CER generation prosp ects at Project 0550.

PAGE 73

67 F IGURE S 23 & 24 HC FC 22 PRODUCTION AT P ROJECT 0 550

PAGE 74

68 P ROJECT 0767 Project 0767 (see Figures 25 and 26) exhibit ed inconsistent production levels prior to the crediting period, with its latest available pre crediting year showing a relative ly low production value MP to MP production levels have shown some v ariation, generally aligning to downward trends throughout each year. Aggregate production has met the creditable production cap almost exactly during each MY to date, showing a difference from the cap of less than 2% of the amount produced in any given MY However, the difference between the amount produced and the reported production capacity was also less than 2% of the amount produced in each of those years The key determinant, then, becomes whether the rate of production dramatical ly changed when the cap wa s reached. In the final MP of MY 1 (a two mon th period), the Project exhibited a severe drop in the production rate in the first month and a complete shutdown in the second month; in the final MP of MY 2, the Project exhibit ed a severe drop in the pr oduction rate in the final month; in the final MP of MY 3, the Project exhibit ed a severe drop in the production rate in the final month; in the final MP of MY 4, the Project exhibited a severe drop in the production rate in the final month. Given the Proj total adherence to the creditable production cap and its extreme production drops in response to excesses over that cap, it seems likely that the reported production capacity of the plant is not responsible for the stability of production. Inste ad, it appears highly likely that CER generation is driving some or all HCFC 22 production at Project 0767.

PAGE 75

6 9 F IGURE S 25 & 26 HC FC 22 PRODUCTION AT P ROJECT 0 767

PAGE 76

70 P ROJECT 0807 While no pre crediting data is available, Project 0807 (see Figures 27 and 28) did exhibit a slight increase from its first MY to its subsequent MYs. Wit hin all years, production has operate d at a relatively constant level, showing few noticeable fluctuations at an MP to MP level. Production capacity is unavailable. The Project fell noticeably short of its creditable production cap in MY 1, but met it very nearly in MYs 2 ( where production fell slightly short of the cap ) and 3 ( where production slightly exceeded the cap ). Unfortunately, month to month HCFC 22 production data is not available, but the project has not display ed any obvious signs of production rate decreases following its single excess over the crediting cap (which was relatively low). Without pre crediting production data, production capacity data, month to month HCFC 22 production data, or a greater sample of MYs, it is difficult to assess th e likely driver of production at Project 0807 beyond the fallback statement that it conspicuously adhered to the HCFC 22 production cap. Nevertheless, given the available data, it appears likely that HCFC 22 production has, to an extent, been driven by CER generation prospects.

PAGE 77

71 F IGURE S 27 & 28 HC FC 22 PRODUCTION AT P ROJECT 0 807

PAGE 78

72 P ROJECT 0838 Project 0838 (see Figures 29 and 30) has exhibit ed a relatively stable HCFC 22 production rate between MYs, showing a slight trend of increased aggregate production over time, which aligns with expectations of steady growth. Within years, the Project has display ed regular flu ctuations, most notably in MY 1, which show ed a downward trend in the production rate. Production capacity is unavailable. The Project exceeded its creditable production cap during all four complete MYs, by a slight ma rgin in MYs 1 and 2 and a relatively higher margin in MYs 3 and 4. Month to month production data is unavailable; however, MP to MP (and with the exception of MY 1, which show ed a general downward trend), the Project has show n no signs of lowering producti on in response to excesses over the creditable production cap. cap (and the presence of those excesses), it seems relatively unlikely that production is being significantly dri ven by CER generation, especially in light of the general trend of production increases (which aligns with the price and demand increases expected of HCFC 22). However, again, without pre crediting period production data, production capacity data, or month to month production data, it is difficult to thoroughly assess the drivers of production.

PAGE 79

73 F IGURE S 2 9 & 3 0 HC FC 22 PRODUCTION AT P ROJECT 0 838

PAGE 80

74 P ROJECT 0868 Project 868 (see Figures 31 and 32) has exhibit ed a relatively consistent year to year production rate, showing a general trend of slight increases from its pre crediting period production to its current crediting period, w hich aligns with expectations of steady growth. The Project has displayed substantial fluctuation in the production rate from MP to MP. The Project has reliably exceeded its creditable production cap by a small amount each MY thus far; however, this level of production seems to be in line with the production levels prior to crediting. The Project did not display any significant drops in HCFC 22 generation following excesses over the crediting cap in MYs 1 and 2 ( month to month data for the following years is unavailable). crediting to crediting production levels and the lack of evidence of significant drops in production levels following crediting caps, it seems unlikely that production of HCFC 22 is being significantly driven by CER generation prospects.

PAGE 81

75 F IGURE S 31 & 32 HC FC 22 PRODUCTION AT P ROJECT 0 868

PAGE 82

76 P ROJECT 1105 Project 1105 (see Figures 33 and 34) has exhibited a relatively rapidly increasing production rate, showing sharp increases in aggregate yearly production across its pre credi ting and crediting years with the exception of MY 2, which show ed a significant decrease in production. MP to MP, the Project has show n quite a bit of fluctuation. The Project has exceeded its creditable production cap by a large amount in all MYs to date. While data is not available for MY 3, in MYs 1 and 2, the Project maintained or even increased its rate of production after the crediting cap had been exceeded with no lulls in production. cap and its undaunted rate of production following excesses over the cap, it seems unlikely that HCFC 22 production is primarily driven by CER generation at Project 1105.

PAGE 83

77 F IGURE S 33 & 34 HC FC 22 PRODUCTION AT P ROJECT 1105

PAGE 84

78 P ROJECT 1194 Project 1194 (see Figures 35 and 36) has exhibit ed a relatively stable rate of aggregate yearly production throughout its pre crediting and credit ing periods. The rate has displayed relative consistency with some infrequent, dramatic shifts in production, typically toward the end of the MY (as in MYs 2, 3, and 4). The Project precisely met its creditable production cap in all MYs to date: in MYs 1 and 3, the difference was less than 0.2% of the cap. As this production is in line with pre crediting production and production capacity is unavailable, it is tion response to excesses over the cap. In MY 1, while production does not increases toward the end of the MY, it d id noticeably increase the cap; in the final MP of MY 2, production dramatically decrease d (over 50%) in August (just when the cap had been reached) and completely stop ped in September; in the final MP of MY 3, production exhibit ed almost exactly the same trend as MY 2 ; in the final MP of MY 4, production drop ped by over 75% from the second month to third, and almost 50% from the third month to the fourth. In all cases, production resumed immediately upon the resumption of crediting. production cap and its clear and severe adjustments of production in response to the cap, it seems highly likely 22 production has been driven by CER generation prospects.

PAGE 85

79 F IGURE S 35 & 36 HC FC 22 PRODUCTION AT P ROJECT 11 94

PAGE 86

80 P ROJECT 1867 Project 1867 (see Figures 37 and 38) has very little data available: it is missing pre crediting production data and only has two reported MYs to date. Though the Project provided its designed annual production capacity as 1,265 tons of HCFC 22 (and sets its crediting cap based on this level), pre crediting production data shows a produ ction level of 1,325 tons, which is accordingly used as the Project seems to have experience d a large jump in production from MY 1 to MY 2. As there are so few MPs, it i s difficult to assess fluctuations in the production rate on an MP to MP level. The Project produced an amount noticeably less than its creditable production cap during MY 1 and substantially greater than its creditable production cap in MY 2. In MY 2, mon th to month data indicates that production remained constant following excesses over the crediting cap. substantial excess over the production cap in MY 2, and its apparent lack of a negative response to that excess over the cap, it seems unlikely HCFC 22 production has been driven by CER generation prospects. However, given the limited range of the data, it remains difficult to ascertain.

PAGE 87

81 F IGURE S 37 & 38 HC FC 22 PRODUCTION AT P R OJECT 1867

PAGE 88

82 P ROJECT 1947 Pr oject 1947 (see Figures 39 and 40) exhibit ed a wide time gap between pre crediting and crediting period data; as such, identifying a trend across periods is difficult. Generally, the Project has show n increases during its pre crediting period and a consistent production level clo se to its reported production capacity during the crediting period. It has show n fluctuation in the production rate from MP to MP. In both complete MYs thus far, the Project noticeably exceeded its creditable production cap and achieved production lev els close to its prod uction capacity. However, as the production capacity is extremely close to its production cap, it is final month of production (just after the Project had met its creditable production cap) exhibited a production rate less than half that of the other seven months in the same MP; unfortunately, month to month data is not available for the final MP of MY 2, but the average production rate does appear to h ave noticeably decreased. Given the proximity of the HCFC 22 production capacity and the creditable production cap, the lack of month to month data, and the small sample of MYs, it is difficult to assess whether HCFC 22 production is driven by CER generati on prospects at Project 1947. Given that pre crediting production levels seem to support the legitimacy of current production levels and production is over the creditable production cap, it seems likely that most HCFC 22 production is not driven by the pro duction cap. However, the Project also does appear to have a distinct response to excesses over the cap; as such, it is likely that at least some HCFC 22 production is driven by CER generation prospects.

PAGE 89

83 F IGURE S 3 9 & 4 0 HC FC 22 PRODUCTION AT P ROJECT 1 947

PAGE 90

84 A SSESSING A BERRANT B EHAVIOR IN THE HFC 23 P RODUC TION R ATIO The effects of perverse incentivization on the w ratio are somewhat easier to identify, if not to quantify. The w value has been recorded for e a ch plant under each MP, either as provided or as calculated from HFC 23 generation and HCFC 22 prod uction. The MP by MP rate (graphed as w_prd) is juxtaposed against the cumulative rate for each MY (graphed as w_cum) and the maximum rate for each MP (graphed as w_max). As with the previous graphs on HCFC 22 production, pre crediting period w rates are i dentified by their lack of a corresponding w cap. In addition, another variable, referred to as the absolute weighted variance from the cap, has been calculated. This variable is established by calculating the absolute difference between the reported w va lue of each MP and the w cap, weighing said values based on the number of days in the MP versus the number of days in the MY, and summing the results. This creates a variable that represents the in any given MY. Assumi ng the existence of a perverse incentive (that is, the value of additional HFC 23 outweighs its associated costs up to the cap), several expectations are established: 1) All firms should tend to exhibit a w rate of at least the capped rate. 2) Divergence from th e w cap is much more likely to be in an upward direction, as an excess in w is almost certainly preferable to a lower than cap w 3) As the ideal rate of w minimized), firms should seek stabilization of th e w rate with minimal fluctuations.

PAGE 91

85 4) When the HCFC 22 cap has been exceeded, firms should regain their natural incentive to minimize the w ratio, as there is no longer any possibility of crediting and thus no incentive to maintain an artificially high w rat io. If, on the other hand, the firm is unaware of or unwilling to capitalize on the perverse incentive, another set of indicators should be expected: 1) The w rate should be exhibit variations within and between years rather than holding at a steady value, as the efficiency of production will be driven by market and plant conditions rather than HFC 23 generation levels. 2) As incentives are unlikely to have experienced a paradigm shift in the interim, the w rate should not drastically differ in value or consisten cy from the historically exhibited w ratio prior to crediting. 3) The overarching trend of the w value should be downward, as the natural incentive will be to reduce reactant loss to increase HCFC 22 production efficiency; by the same logic, the w value of a non manipulative plant should generally show a greater tendency to dip beneath the capped level. adhere to the cap to a degree of relative certainty.

PAGE 92

86 P ROJECT 0001 Project 0001 (see Figure 41) exhibit ed significant variation in its pre crediting w ratios. Its w variance swiftly decrease d once crediting began however: 2, 1.6% in MY 3, 3.0% in MY 4, and 1.1% in MY 5. This shows a general trend of sharp increases in precision to the cap over time. The Project display ed significantly lower than cap w ratios once during MY 1 and once during MY 2; nevertheless, in the subsequent four MYs, it d id not achieve w rati os anywhere near those lower values. The general frequency of MPs exhibiting a lower than cap w also decrease d over the time; though the Project show ed a consistently below cap w in MY 2, subsequent years display ed primarily above cap w ratios. The Project w response to excesses over the HCFC 22 cap is not possible to assess, as there are no MYs displaying excesses over the HCFC 22 cap. w cap as compared to its pre crediting variation and its demon strated ability to achieve lower than cap w ratios, it appears highly likely that the w ratio is being manipulated to achieve a cap exact ratio at Project 0001.

PAGE 93

87 F IGURE 41 W RATIO AT P ROJECT 0001

PAGE 94

88 P ROJECT 0003 Project 0003 (see Figure 42) exhibit ed a relatively stable w ratio between its pre crediting and crediting periods. Though the necessary data for calculations is not available for MYs 1 and 2, absolute weighted variance from the cap was 3.0% in MY 3, 3.0% in MY 4, 2.9% in MY 5, and 2.0% in MY 6. This shows a general trend of strong and increasing precision to the cap across MYs. Strangely, the Project exhibit ed almost all of its variance in a downward direction. This is contrary even to pre crediting production data trends, which exhibited aggregate w values higher than the cap. Nevertheless, the t rend seems to have be en for the Project to very slightly undercut the creditable w value. In addition, the Project show ed a notable decrease in w from its first crediting period lished a lower w cap than its first, this aligns with expectations: however, the Project has continued to slightly undercut the cap. I n MY 6 (the only MY wherein the Project has notably exceeded its HCFC 22 cap thus far), the Project displayed absolutely n o signs of a decrease in the w ratio. period w ratios to its maximum w ratio, it seems likely that Project 0003 is manipulating its w ratio to a cap exact rate. However, the fact that the vast majority of variation was negative rather than positive, the w ratio seems to be generally decreasing, and exhibited w shows a clear trend in comparison with pre crediting w values, this assumption is not necessarily correct.

PAGE 95

89 F IGURE 4 2 W RATIO AT P ROJECT 0 003

PAGE 96

90 P ROJECT 0011 Project 0011 (see Figure 43) exhibit ed a relatively stable w ratio between its pre crediting and crediting periods. No clear upward or downward trend is evident weighted variance from the cap wa s 8.4% in MY 1, 5.9% in MY 2 6.8% in MY 3, 6.4% in MY 4, and 7.3% in MY 5. This shows no clear trend of increasing or decreasing precision. s MPs met or fell below cap: all of them have exhibit ed noticeably above cap ratios, with the exception of one MP in MY 2 that comes close to the lowest historical value measurement used as the cap. This indicates that the Project has the ability to achieve at least somewhat lower w ratios. Otherwise, MP to MP fluctuations have been extremely minimal. This Project is also difficult to assess. Though the Project does appear to vary relatively significantly from the cap and display values in accordance with its historical production rates, it also consistently exceeds its creditable w ratio despite a demonstrated abil ity to achieve lower rates, at least on occasion. As such, it appears likely that the Project is ensuring excesses over the w cap.

PAGE 97

91 F IGURE 4 3 W RATIO AT P ROJECT 0 011

PAGE 98

92 P ROJECT 0115 MY to MY, Project 0115 (see Figure 44) has exhibit ed fairly consistent w ratios, with yearly averages closely approximating the cap. MP to MP, Project 0115 demonstrate d highly inconsistent w ratios: absolute weighted variance from the cap was 12.5% in MY 1, 7.7% in MY 2, 8.9% in MY 3, 5.7% in MY 4, and 10.5% in MY 5. This shows a trend of general, but not strong, proximity to the cap. The MPs have also varied strongly against the cap. In MY 1, the Project achieve d a w ratio nearly a full percentage point below its w cap. Other sharp reductions can be observed in MYs 2 and 3; however, the severity and frequency of w ratios below the cap has noticeably diminished over time, and the general trend seems to be toward an increase in the w ratio. As HCFC 22 production has not noticeably exceeded the HCFC 22 creditable production cap in any MY thus far, it is not possible to assess the response in w to an excess over the cap. w ratios, its apparently decreasing propensity to do so, and its close proximity (on the aggregate) to the creditable w ratio in each MY, it seems likely tha t Project 0115 is, at the very least, working to avoid lower than cap w ratios. However, the relatively ability to actively manipulate the w ratio.

PAGE 99

93 F IGURE 4 4 W RATIO AT P ROJECT 0 115

PAGE 100

94 P ROJECT 0151 Projec t 0151 (see Figure 45) has exhibit ed a relatively consistent w rate MP to MP and MY to MY. Its w ratio appears to have remained fairly static from pre crediting to crediting years. Absolute weighted variance from the cap wa s 7.1% in MY 1, 4.9% in MY 2, 5.2% in MY 3, 4.7% in MY 4, and 3.8% in MY 5. This shows a general trend of strong and increasing precision to the cap. ed to be in a downward direction; however, by the second MP of MY 2, this trend all but di sappeared since that point, there have been only three MPs exhibiting a lower than cap w ratio (only one of them significantly lower). The Project exceeded its HCFC 22 production cap in MYs 2, 3, 4, and 5: in MY 2, the Project exhibited a drop of almost 1.5% in w following an excess over the HCFC 22 cap; in MY 3, the Project exhibited brief but noticeable drops in its final two MPs (both of which contained production beyond the HCFC 22 cap); in MY 4, the Project exhibited a noticeable decline in the w rat e when the cap was reached and a sharp decline (almost 1%) in the month when it is suspected that a shutdown occurred; in MY 5, the Project does not display a severe drop in the w ratio, but nevertheless exhibit ed a general decline following the excess ove r the cap. w cap, its demonstrated ability to achieve lower w ratios, and its clear responses to excesses over the HCFC 22 cap in terms of w it appears highly likely that Project 0151 engages in m anipulation of its w ratio to achieve maximum CER generation.

PAGE 101

95 F IGURE 4 5 W RATIO AT P ROJECT 0 151

PAGE 102

96 P ROJECT 0193 Project 0193 (see Figure 46) exhibit ed a severe shift in proximity to the cap and overall variation from its pre crediting period to its crediting period. In the pre crediting period, the Pro ject show ed wildly different w ratios, ranging between 4% and 5.5%. Once the Project was capped at 3%, the w ratio almost immediately stabilized to a range between 3% and 3.25%. Absolute weighted variance from the cap wa s 9.0% in MY 1, 5.1% in MY 2, 5.3% i n MY 3, 4.4% in MY 4, and 4.4% in MY 5, showing a trend of strong and increasing precision. Without exception, that variation has been in an upward direction: no MP has exhibited a w below the capped value. However, the Project has not display ed noticeabl e adjustments in the w ratio in response to the strong excesses over the HCFC 22 cap it has exhibited in all MYs. w ratio cap and its major shifts in production levels compared to pre crediting values indicate that it is likely that Project 0193 has, to an extent, manipulated its w ratio to adhere to the cap. However, given that the final year of the pre crediting data available indicates also a strong decrease in the w ratio and the Pr oject has not demonstrated an ability to achieve a lower w ratio, this conclusion is less certain.

PAGE 103

97 F IGURE 4 6 W RATIO AT P ROJECT 0 1 93

PAGE 104

98 P ROJECT 0232 Project 0232 (see Figure 47) has exhibit ed an extremely consistent w ratio from year to year and from MP to MP. Absolute variance from the cap wa s 18.5% i n MY 1, 2.5% in MY 2, 3.2% in MY 3, 4.3% in MY 4, and 4.0% in MY 5. At first glance this indicates an extreme adjustment in precision to the cap after MY 1. However, the Project exhibit ed only one value beneath the capped w that being the first MP of MY 1, which wa s under 1% (versus a cap of 2.37%). After examining this particular MR, it appears as though incineration began in March, while the Period itself extended to April. Taken in the context of the otherwise highly consistent w ratio, this lead s to the conclusion that the amount of HFC 23 reported was not the whole of the HFC 23 generated during the period. When data from that MP is removed from consideration, the absolute weighted variance from the cap is recalculated as 2.37% th e lowest valu e of the five MYs ability to achieve a w ratio beneath the cap, given the low pre crediting ratio in 2004 w throughout each year, it is likely t hat the Project his torically achieved lower ratios, at least on occasion Given its high level of precision to the cap, likely ability to achieve a lower w ratio, and trend of slight increases in aggregate yearly w as compared to slight decreases prior to crediting, it appears likely that Project 0232 is manipulating its w ratio to maximize CER generation.

PAGE 105

99 F IGURE 4 7 W RATIO AT P ROJECT 0 232

PAGE 106

100 P ROJECT 0306 Project 0306 (see Figure 48) exhibit ed a significantly varying w ratio prior to crediting, but swiftly adjusted to a much lower and miraculously co nsistent w ratio during the crediting period. Variation from MP to MP has been almost nonexistent, with only one MP having varied significantly from the cap: it is worth noting that this MP, technically the first of MY 2, was initially bundled with the f inal MP of MY 1, indicating that there may have been some confusion on the part of the PP. Absolute weighted variance from the cap was 0.8% in MY 1, 2.3% in MY 2, 0.5% in MY 3, 0.2% in MY 4, and 0.6% in MY 5, showing a trend of almost absolute precision to the cap. All variation from the cap has been in an upward direction: the Project has never displayed a w below the cap. I nterestingly, this trend has remained constant 22 creditable production cap in all five MYs, indicating that it may have been too difficult, time consuming, or non cost effective to readjust the w ratio post crediting. This latter point may have raised questions with other plants, as the general incentive should be to decrease w in the absence of crediting, especially in light of large excesses over the HCFC adherence to the creditable w cap versus its previously wildly divergent w ratios, it appears highly likely that the w ratio is b eing manipulated to a cap exact value at Project 0306.

PAGE 107

101 F IGURE 4 8 W RATIO AT P ROJECT 0 306

PAGE 108

102 P ROJECT 0499 Project 0 499 (see Figure 49) exhibit ed a significantly varying w across its pre crediting and crediting periods, showing no clear trend in the ratio from year to year or from MP to MP. The absolute weighted variance from the cap was 27.1% in MY 1, 15.7% in MY 2, 23.1% in MY 3, and 14.1% in MY 4. While the trend in this data seems to be in a generally downward direction, there is no clear pattern and adherence to the cap does not appear to be terribl y strong. Despite having shown average yearly values prior to the start of crediting that indicate historical production rates below the capped value, the Project has only display ed two MPs with values below the cap: the first MP ever and the final MP of MY 4. The project closely met its HCFC 22 cap in all four MYs: in the one MY where it was exceeded significantly, MY 4, the w rate dropped below the cap. fluctuation of its w ra tio, it would, at first glance, seem that the w ratio is not being 22 cap, its demonstrated ability to achieve a w ratio below the cap, and its otherwise above cap w performance, it s eems likely that Project 0499 manipulates the w ratio in such a way as to ensure that it is above cap during crediting periods.

PAGE 109

103 F IGURE 4 9 W RATIO AT P ROJECT 0 499

PAGE 110

104 P ROJECT 0549 Project 0549 (see Figure 50) exhibit ed a highly inconsistent w ratio prior to the crediting period and a highly consistent w ratio during the crediting period. MP to MP variation is minimal and shows no trends. Absolute weighted variance from the cap was 12.5% in MY 1, 11.7% in MY 2, 11.8% in MY 3, and 10.9% in MY 4, indicating stable variance from the cap with a slightly downwa rd trend. The Project has show n no signs of a w ratio below the cap during any MP. Though the Project has produced excesses over the HCFC 22 creditable production cap in all MYs thus far, there appear to have been no responses in the w ratio to said exces ses during any MY. Given the stabilization of w following previously inconsistent w ratios and the w likely that the Project is manipulating its w ratio to maintain a rate above th e cap. However, given the lack a demonstrated ability to achieve a w ratio beneath the cap and the lack of response to excesses over the HCFC 22 cap, this conclusion is less certain than with other projects.

PAGE 111

105 F IGURE 50 W RATIO AT P ROJECT 0 549

PAGE 112

106 P ROJECT 0550 Project 0550 (see Figure 51) has exhibit ed an extremely consistent w ratio from its pre crediting period to its crediting period. No remarkable trend is present in either period, nor is there any evident trend in MP to MP fluctuations. Absolute weighted variance from the cap was 3.6% in MY 1, 3.6% in MY 2, 4.9% in MY 3, AND 6.8% in MY 4, showing a trend of slight increases in variance from the cap. Without exception, variance from the cap has been in an upward direction: however, these variations have been in line with historical production, and no significantly lower w values have been demonstrated in any MP. The Project notably exceeded its creditable production cap for HCFC 22 in MYs 2, 3, and 4, but no notable drop in production was visible in any of these MPs, even on a month by month basis. G w ratio, lack of response to excess over the HCFC 22 production cap, and lack of demonstrated ability to achieve lower w ratios, it seems unlikely that the Project is actively manipulating its w ratio. However, it does not appear to have managed its w ratio in an efficient manner for HCFC 22 production, raising questions as to whether it is choosing to maintain an inefficient w ratio.

PAGE 113

107 F IGURE 5 1 W RATIO AT P ROJECT 0 550

PAGE 114

108 P ROJECT 0767 Project 0767 (see Figure 52) has exhibit ed a fairly consistent w ratio across its p re crediting and crediting periods. Generally, no clear trends are visible; however, one MY does exhibit a clear downward trend from MP to MP. Absolute weighted variance from the cap was 6.7% in MY 1, 2.1% in MY 2, 2.8% in MY 3, and 2.5% in MY 4, indicatin g narrowing and stabilization in the w ratio after MY 1. Only one MP show ed a variation below the w cap: all other variation is upward. The Project has slightly exceeded its HCFC 22 production cap in all MPs. Though a significant drop in w is v isible in MY 1 toward the end of crediting, the drop appears to have been disproportionate to the excess over production, and no such drops are visible in subsequent years. w cap and its demonstrated ability to achieve lower ratios, it appears likely that the w ratio is being manipulated to maintain a cap exact ratio at Project 0767. It is worth considering whether MY 1 represented business as usual (hence the high variance in w and otherwise rational reduction of its value) versus a more tightly manipulated ratio in subsequent years.

PAGE 115

109 F IGURE 52 W RATIO AT P ROJECT 0 767

PAGE 116

110 P ROJECT 0807 Project 0807 (see Figure 53) did not make publicly available its pre crediting period w ratio data; however, as its cap is 3.0%, it can be assumed that its pre crediting values w ere above 3.0%. From MY to MY, the Project has show n a fairly consistent w ratio with strong fluctuations from MP to MP: its absolute weighted variance from the cap was 12.4% in MY 1, 7.8% in MY 2, and 10.5% in MY 3. This indicates a trend of general proxi mity to the cap, but there is no clear evidence of a trend of increasing or decreasing precision. All variation from the cap has been in an upward direction: no MP has exhibit ed a w lower than the cap. The Project slightly exceeded its creditable HCFC 22 production cap in MY 2; however, no commensurate drop in the w ratio is observed in response to this excess. As historical data is not available and the Project has not demonstrated an ability to achieve lower w ratios, it is difficult to assess whether i t is manipulating its w ratio. However, the consistent and reasonably precise excesses over the w cap indicate that it is fairly likely that Project 0807 is ensuring an above cap w rati o, as one would assume that pre crediting values were not so closel y stabilized to the cap.

PAGE 117

111 F IGURE 53 W RATIO AT P ROJECT 0 807

PAGE 118

112 P ROJECT 0838 Project 0838 (see Figure 54) exhibited relatively high variation between its pre crediting and crediting w ratios: two of its three reported pre crediting w values we re noticeably lower than the aggregate yearly values achieved during crediting, while one wa s significantly higher. However, during crediting, the Project has exhibit ed fairly consistent w ratios and moderate MP to MP variation. Absolute weighted variance from the cap was 12.2% in MY 1, 11.3% in MY 2, 9.3% in MY 3, a nd 7.8% in MY 4. This shows a trend of increasing precision to the cap over time. The Project has shown no MPs below the w cap: all variation has been upward. However, it has demonstrated an ability to achieve lower ratios, as two of its pre crediting yea rly w average values were MP values (which were still higher than one of the two years). The Project has slightly exceeded its HCFC 22 creditable production cap in all MYs thus far: while all other MYs seem uneventfu l in this regard, MY 3 exhibit ed one month w ratios of 2.77% and 1.39% during its final MP. w cap, its increasing precision to the cap, and its consistent excess over the cap, it appears likely that the Project is manipulating its w ratio to ensure excesses over the cap when crediting is available.

PAGE 119

113 F IGURE 54 W RATIO AT P ROJECT 0 838

PAGE 120

114 P ROJECT 0868 Project 0868 (see Figure 55) has exhibit ed a downward trend in the w ratio from its pre crediting period to its crediting period. In the cr editing period, this rate stabilized tremendously and has shown almost no significant variation over time. Absolute weighted variance from the cap was 6.1% in MY 1, 6.1% in MY 2, 5.7% in MY 3, and 6.1% in MY 4. This shows an incredibly stable and slight va riance from the w cap. All variance from the cap has been in an upward direction: there is no MP in which the Project has exhibit ed a w ratio lower than the cap. In addition, as the w cap is based on the 3.0% ceiling, the pre crediting w ratios do not sug gest lower than cap values in the past. The Project has slightly exceeded its HCFC 22 creditable production cap in all MYs thus far: however, no significant adjustments in the w ratio are observed upon the generation of excess production. Given the Projec the w cap, it seems highly likely that the w ratio is being intentionally stabilized to ensure cap w response to excesses over the HCFC 22 cap and its lack of a demonstrated ability to achieve lower w ratios make this conclusion less certain than desired.

PAGE 121

115 F IGURE 55 W RATIO AT P ROJECT 0 868

PAGE 122

116 P ROJECT 1105 Project 1105 (see Figure 56) has exhibit ed a relatively varied and extremely low w ratio. While year to year w values have been fairly cons istent across the pre crediting and crediting periods, MP to MP variations have been comparatively high, with MY 1 showing a w range over 1%. Absolute weighted variance from the cap was 48.1% in MY 1, 16.4% in MY 2, and 12.9% in MY 3. This shows a strong t rend of increasing precision to the cap; however, the most recent values remain high compared to other plants. The Project show ed two crediting periods beneath the cap, which is w cap. The Project ha s significantly exceeded its creditable HCFC 22 pro duction cap in all MYs thus far. However, the two drops beneath the cap have coincide d with the highest excesses over the cap in MYs 1 and 3. ap, demonstrated ability and decreasing propensity to produce w ratios lower than the cap, and strong responses to high excesses over the HCFC 22 creditable production cap, it appears highly likely that Project 1105 is manipulating its w ratio to maximize CER generation.

PAGE 123

117 F IGURE 56 W RATIO AT P ROJECT 11 0 5

PAGE 124

118 P ROJECT 1194 Project 1194 (see Figure 5 7) has exhibit ed a relatively consistent w ratio between its pre crediting and crediting periods. Variation from MP to MP ha s been fairly strong: absolute weighted variance from the cap was 7.5% in MY 1, 3.1% in MY 2, 9.3% in MY 3, and 10.0% in MY 4. This shows decent precision to the cap, but indicates no trend of increasing or exact precision. The Project has a historically based w cap; however, only two MPs thus far have met the cap value, and none have falle n beneath it. As the Project has precisely met its HCFC 22 creditable production cap in all MYs thus far, it is impossible to 22 cap. w cap despite a demonstrated ability to achieve at least a cap exact ratio, it appears likely that Project 1194 is intentionally ensuring an above cap w ratio. However, without a demonstrated lower than cap w ratio, a trend of increasing precision to the cap, or data on t 22 cap, it is difficult to confidently make this assertion.

PAGE 125

119 F IGURE 57 W RATIO AT P ROJECT 1194

PAGE 126

120 P ROJECT 1867 Project 1867 (see Figure 58) provided very little data. Given the meager information available, aggregate w values appear to have show n a fairly consistent downward trend from year to year, including the pre crediting period. The Project has reported very few MPs, and as such, MP to MP variation and absolute weighted variance from the cap are not meaningfully applicable in this scenario. The Proje ct has exhibited no MPs with a below cap w ratio, and only one ratio the most recent meets the cap (which is based on the 3.0% ceiling rather than historical values). A month by month analysis of the second MP indicates consistency in the w ratio in re 22 production cap. Given the severe dearth of data for Project 1867, it is extremely difficult to make confident assessments of its behavior. However, it does appear that the Project is stabilizing its w ra tio to meet or slightly exceed the w cap. As such, it appears more likely than not that Project 1867 has engaged in adjustment of its w ratio to maximize CER generation.

PAGE 127

121 F IGURE 58 W RATIO AT P ROJECT 1867

PAGE 128

122 P ROJECT 1947 Project 1947 (see Figure 59) has exhibit ed a relatively consistent w ratio from its pre crediting period to its crediting period, with the data indicating a general trend of decreases in the w ratio. From MP to MP, it has exhibit ed very little variation except in service of the general downward trend. As a result of the downward trend, a bsolute weighted variance from the cap was 10.1% in MY 1 versus 7.9% in MY 2 (MY 3 appears to be on a similarly decreasing trend thus far). The Project has shown no MP with a w beneath the cap, which is (barely) non historically based. However, its exhibi ted values during crediting have generally been above its final pre crediting aggregate yearly value, indicating that the Project has the ability to generate lower w values than it has shown during crediting. The Project has slightly exceeded its creditabl e HCFC 22 production cap in both MYs thus far; however, there were no remarkable decreases in w following these excesses. crediting and crediting period w ratios and its lack of a demonstrated ability to achieve ratios lower than the cap, combined with its generally decreasing trend in w (as opposed to sharp increases upon the start of crediting), it is unclear whether Project 1947 is manipulating its w ratio: however, it has consistently achieved a higher than cap ratio and d oes appear to have increasing precision to the cap. As such, it appears more likely than not that the Project has adjusted its w ratio to ensure maximum CER generation.

PAGE 129

123 F IGURE 59 W RATIO AT P ROJECT 1 947

PAGE 130

124 C ONCLUSIONS A GGREGATE E FFECTS OF P ERVERSE I NCENTIVES Nine of the projects showed signs of HCFC 22 production driven by CER generation prospects, while all but one of the projects showed signs of manipulating the w ratio to maximize CER generation (see Figure 60) It should be noted that only one plant Project 0 001 has exhibited HCFC 22 dr iven production with aggregate yearly totals below the cap: all other non manipulative projects demonstrate d non CER dependence through excesses over the cap. This raises a question as to whether the other non manipulative projects would exhibit CER driven production i f the profitability of HCFC 22 decreased to a point where the ideal production was beneath the creditable production cap. Generally, the conclusions indicate a high occurrence of influence by perverse incentives in HFC 23 abatement. Project Loc ation HCFC 22 Manipulation Manipulation 0001 India Unlikely (under) Highly likely 0003 South Korea Likely Likely 0011 China Unlikely Likely 0115 India Highly likely Likely 0151 Mexico Highly likely Highly likely 0193 China Unlikely Likely 0232 C hina Likely Likely 0306 China Unlikely Highly likely 0499 India Highly likely Likely 0549 China Unlikely Likely 0550 China Unlikely Unlikely 0767 China Highly likely Likely 0807 Argentina Likely (low data) Likely (low data) 0838 India Unlikely (low data) Likely 0868 China Unlikely Highly likely 1105 China Unlikely Highly likely 1194 China Highly likely Likely (low data) 1867 India Unlikely (low data) Likely (low data) 1947 China Likely Likely F IGURE 60 A GGREGA TE R ESULTS

PAGE 131

125 Unfortunately, it is not feasible to reliably asse ss how the projects would have behaved differently, if at all, in the absence of crediting. Such an analysis would require knowledge of the minimum w ratio achievable by the plant and more detailed supply and demand data. Nevertheless, the results call int o question the legitimacy of the credits issued to eighteen of the nineteen plants: accounting for CER issuance until December 2011, those eighteen plants have received over 94% of the credits issued by the CDM under AM0001. These results confirm the susp icions and predictions presented in the literature review There are strong indicators of behavior that has been significantly influenced by near ly ubiquitous perverse incentives in the crediting of HFC 23 under t he Clean Development Mechanism. Specificall y, the results confirm: w ratios at Projects 0011, 0193, 0232, assertions regarding responsive HCFC 22 production rates at Project 0767; and s assertions of responsive w ratio decreases at Projects artificially precise w lants stopped production when the HCFC valuable to note that the two projects deemed non manipulative in terms of HCFC 22 by Schneider 0115 and 0807 have since increased production to a cap exact rate. Similarly, Project 0001 noted by the NRDC paper as one of the few projects to exhibit w ratios beneath the cap has since ceased doing so.

PAGE 132

126 S UFFICIENCY OF R ECENT A DJUSTMENTS TO AM0001 With these results in mind, it is now important to address whether the recent adjustme nts to AM0001 appear to be sufficient to ameliorate the effects of the observed perverse incentives. The adjustment to the HCFC 22 creditable production cap is unlikely to be of major significance, as most projects have exhibited relatively stable producti on from year to year. T he more interesting effect to observe will instead be the two thirds reduction in the creditable w ratio. In terms of manipulation of the w ratio to maximize CER generation, the adjusted 1.0% cap will likely ensure additionality. On ly two projects demonstrated an ability to achieve a ratio at or below 1.0% and BAT for HCFC 22 production has been estimated at a 1.4% ratio. 146 While lower ratios have been observed at plants in developed countries 147 1.0% is likely well below the ratio tha t would have been reached by the credited plants in the absence of crediting. When considering that many plants seem to naturally operate at a rate of HCFC 22 production at or near the capped amount, it is also likely that w ratio manipulation rather tha n HCFC 22 production rate manipulation is responsible for the bulk of non additional credits issued. While it is possible (as Schneider points out) that the non credited reductions against the baseline resulting from a lower w ratio will outweigh the ne gative effects of perverse incentives remaining in the HCFC 22 production rate, such incentives should nevertheless be avoided. Excess production of HCFC 22 not 146 147 23 and Other

PAGE 133

127 only produces non additional CERs ( assuming the new w ratio is a real expectation rather than a proxy for a discount rate) but also results in unaccounted for emissions from HCFC 22 production and potentially detrimental effects on the HCFC 22 industry. These effects create a set of especially dangerous and unpredictable negative externalities to co ntend with. When considering the elimination of the HCFC 22 production perverse incentive, t he question then becomes whether CER revenues will remain sufficient to justify excess production to the cap in light of the adjusted w ratio cap. The reduction wil l effectively result in a two thirds reduction in CERs issued to most projects. Assuming a stable price, this is equivalent to a two thirds reduction in revenues: however, as the supply of CERs will drastically decrease, their price (and perceived legitima cy) will likely increase, offsetting the reduction in revenue to an extent Still it is valuable to assess the effect of a two thirds reduction in credit production in response to perverse incentives. As the Chinese government imposes a 65% tax on CER revenues, the behavior of Chinese plants under the currently operational version of AM0001 should be comparable to the behavior of non Chinese plants under the new version of AM0001. In terms of HCFC 22 production manipulation, seven of the eleven Chinese Chinese projects. In terms of w ratio manipulation, the sole non suspect project is located in

PAGE 134

128 China. On the surface, this does seem to indicate that Chinese plants are less likely to exhibit production driven by CER generation prospects. However, the only project that consistently exhib its an aggregate yearly HCFC 22 production rate below its cap 0001 is located in India; recalling earlier conclusions, all other plants rated above the cap, which does not eliminate the possibility that CER revenues could be suffi cient to justify production. Given that at least three of the Chinese plants show indicators of production being driven by HCFC 22 production, it seems likely that such behavior remains a risk under a 65% discount rate, and thus remains a risk for non Chin ese plants under the revised methodology. With regard to Chinese plants, a further two thirds reduction will likely mitigate the incentive to produce HCFC 22 for the sole purpose of producing HFC 23 to be abated under current prices. Based on the revenue comparisons earlier in the study, a compounded reduction could lead to CER revenues rating at less than twenty percent of HCFC 22 production revenues which is almost certainly below the amount necessary to drive production. However, as noted earlier in t he study, this is assuming the continuation of the current extreme conditions: a greatly increased HCFC 22 price, an abysmal CER price, and a 65% domestic tax on CER revenues. It is also not outside the realm of possibility that the Chinese government, in response to the lowered cap, may consider lowering its tax on HFC 23 based CER revenues to ensure that its HCFC 22 facilities remain operational and competitive. Under such circumstances, the efficacy of the newly adjusted caps in preventing perverse behav ior at Chinese HCFC 22 plants cannot be guaranteed.

PAGE 135

129 I MPLICATIONS FOR THE CDM T he dramatic effects and continued existence of the HFC 23 perverse incentivization problem call into question certain gener al practices by the CDM. First and with the necessar y ca veat that accompanies twenty twenty hindsight the problem seems to have been an especially easy one to see on the horizon: the CDM composed a methodology which planned to issue millions of credits to a handful of plants without sufficiently accountin g for the possibility of perverse incentives ex ante all the while knowing that it would be extremely difficult to adjust them ex post While it is tempting to demonize the HCFC 22 plants for abusing the Mechanism one sh ould expect that businesses will act in an economically not environmentally rational man ner and plan accordingly. At the end of the day, the reactions to these perverse incentives do not indicate crimes of emission by the credited projects; rather, we are experiencing the results of a crime of omission on the part of the Clean Development Me chanism. If a problem this large can make it through the CDM EB not to mention make it six years wi thout being addressed in any real or meaningful way how far can the CDM be trusted as a source of legitimate carbon credits even if this problem w ere mitigated? Second, the results call into question proportionality in carbon offset crediti ng as a principle Currently, standard operating procedure for carbon crediting schemes is to issue credits based on the global warming potential of the aggregate greenhouse gases abated. While this practice ensures that each offset is tCO 2 eq (in theory), it also leads to wildly divergent economic incentives through imbalanced issuance. To an extent, the CDM wants to have its

PAGE 136

130 cake and it eat, too; there is a distinct tradeoff between proportionality in credit issuance and ensuring that e conomic incentives cover the costs of abatement plus a reasonable rate of return. In light of these results, it may behoove the CDM and other crediting schemes to err closer to the incentive side of that balancing act in the future. After all, it is, o f course, better for the credits to truly represent more than one tCO 2 eq than less Even if, in fact, perverse incentivization in terms of w and HCFC 22 production will both be effectively mitigated by the a djustments to AM0001, two severe problems remain. Cumulatively, many years of crediting will be conducted under the cu rrently operational version of AM0001: these unaffected years of operation and crediting, left unmitigated, will result in additional millions upon millions of illegitimate credits. By the time the new version of AM0001 begins to truly take effect that i s, when the bulk of renewable HFC 23 abatement projects are forced to reapply for crediting it may be too little, too late to truly restore legitimacy to the Clean Development Mechanism. Returning to the solutions proposed by authors in the literature re view, it is possible that alternative methods for mitigation of the ongoing problems might be found. The primary suggestion, which was mainly established by Schneider, has been the establishment of a lower w rate in AM0001, which has, to an extent, been ac complished. Nevertheless, the w ratio cap of 1.0% introduced in the most recent version of AM0001 does not appear to be sufficiently low to reliably ensure the elimination of perverse incentivization in terms of increased HCFC 22 production. As such, this report recommends an additional revision to decrease the w ratio cap

PAGE 137

131 ability to experience a reasonable rate of return beyond the costs of abatement without crediting to a le vel where a cost comparison between product and credits is necessary. Schneider also proposed managing HFC 23 abatement through an external body such as the Global Environmental Fund ; similarly, the NRDC suggested regulating HFC 23 through the Montreal Pro tocol as a highly potent greenhouse gas, HFC 23 can and should be managed under the Kyoto Protocol. It is natural that proper management will never occur if proper safeguards are not put in place: this is a problem that any institutional body will face with any major greenhouse gas. Even if responsibility were transferred, the CDM would still be responsible for the remainder of the crediting years under th e existing projects, and would be left illegitimate and with very little chance for redemption in the HFC 23 arena. As the CDM is the brainchild of the closest thing to a global climate change agreement yet devised, this seems ill advised. The TEAP suggest ed an interesting solution to the problem of perverse incentivization in current crediting periods. Though the CDM is bound to its currently approved crediting periods, the host nations for the projects in question have the ability to adjust the returns fr om crediting through taxation or to regulate the projects in other ways. Domestic regulation and taxation could remove the perverse incentives associated with high crediting levels without necessarily waiting for projects to finish their current crediting periods. This is a win win

PAGE 138

132 scenario for the CDM and host nations, as perverse incentives for unnecessary production are removed and the host nations gain revenues from taxation of CER revenues. Most dramatically, there is the option of refusing to renew th e existing HFC 23 abatement projects. The prospect is somewhat appealing: as they have likely been issued millions of credits resulting from otherwise undesirable production, it seems only fitting that crediting be made unavailable. However, such a move wo uld likely result in the cessation of thermal oxidation at the majority of the plants, causing emission of hundreds of tons of HFC 23 into the atmosphere during future production. In fact, the Chinese government recently made a veiled threat that implied t hat its HCFC 22 plants would certainly emit HFC 23 if crediting was removed from the facilities. 148 This report recommends adjustment in the opposite direction. With the introduction of a lower w cap and the presumptive elimination of perverse incentivizat ion in the revised methodology attention must be turned to the currently non creditable HCFC 22 facilities that began to consistently operate after the crediting cutoff. For the most part, these plants currently emit HFC 23 into the atmosphere, causing s evere detriment. If the CDM is confident that its methodology is sufficient to eliminate perverse incentivization, then it should allow new HFC 23 projects to apply; if it is not, then AM0001 should be revised until such confidence is established. The intr oduction of new, legitimate HFC 23 projects will also increase 148

PAGE 139

133 the average quality of CERs, which is especially valuable if the perverse incentivization in current crediting periods cannot be mitigated. Through these steps further adjustment of the w rat io under AM0001, pressure on host nations to regulate and tax HFC 23 abatement projects, and the introduction of new plants under AM0001 the CDM may yet find some small measure of redemption for the HFC 23 abatement crisis. However, the results of this s tudy do not bode well for the current or future legitimacy of the CDM, and correction will require swift and strong movement on the part of the Executive Board and participant nations.

PAGE 140

134 A PPENDIX I HFC 23 CER S VS T OTAL CER I S SUANCE

PAGE 141

135 A PPENDIX I I CER VALUE VS HCFC 22 VALUE

PAGE 142

136 A P PENDIX II I E STIMATED CER REVENUE VS ESTIMATED HCFC 22 REVENUE BY PLANT P ROJECT 0001 P ROJECT 000 3

PAGE 143

137 P ROJECT 0011 P ROJECT 0115

PAGE 144

138 P ROJECT 0151 P ROJECT 0193

PAGE 145

139 P ROJECT 0232 P ROJECT 0306

PAGE 146

140 P ROJECT 0499 P ROJECT 0549

PAGE 147

141 P ROJECT 0550 P ROJECT 0767

PAGE 148

142 P ROJECT 0807 P ROJECT 0838

PAGE 149

143 P ROJECT 0868 P ROJECT 1105

PAGE 150

144 P ROJECT 1194 P RO JECT 1947

PAGE 151

145 B IBLIOGRAPHY Together to Curb HFC 2010. modified May 2012. http://data.bluenext.fr/downloads/20120510_BNS_STATS.xls Environment and Developme nt edited by Teng Teng and Ding Yifan. EOLSS Publishers, 2009. 23 Environmental Investigation Agency, 2011. University of Gothenburg, 2008. mission trends and projections in Europe 2011: http://www.forexticket.co.uk/en/histo/EUR CNY http://www.green evolution.eu/default.asp?pid=138&la=1 Hert Linked Environmental Science & Technology 43 (2009): 6414 6420. he Clean Development Mechanism and Sustainable Development: A Panel Data for Governance and Sust ainable Development, 2007. World Economics & Politics 6 (2003): 66. Lammertjan, Dam and B

PAGE 152

146 23 Waste Streams for Abatement of Emissions from HCFC 22 Production: A Review of Scientific, Technical, and Eco http://www.monbiot.com/2006/10/19/selling indulgences/ Philosophical Transaction: Mathematical, Physical, and Engineering Sciences 360 (2002): 1875 1888. a Policy Sciences 34 (2001): 303 327. Environews 117 (2009): 63 68. ko Institut e.V., 2007. erverse incentives under the Clean Development Mechanism (CDM): an evaluation of HFC ko Institut e.V., 2010. The Case of HFC 23 de ko Institut e.V., 2005. Fluorocarbon Emission Credits by UNEP Techn particular focus on the impact of the Clean Development Mechanism) and Emissions Reduction Benefits Arising From Earlier HCFC Phase 2007. http://www.cdmpipeline.org/cers.htm UNFCCC CDM Last modified May 2012. http ://cdm.unfccc.int/ http://cdm.unfccc.int/UserManagement/FileStorage/50KH2J9V6O1IQNBSPALXYUGRC ZFED7 UNFCCC, 2010. http://www.epa.g ov/sequestration/leakage.html Accessed April 2012. http://unfccc.in t/kyoto_protocol/mechanisms/clean_development_mechanism/items/2718.php Stanford University, 2006.