76943 v2 Alternative Approaches to Addressing the Risk of Non-Permanence in Afforestation and Reforestation Projects under the Clean Development Mechanism Final Report November 20, 2012 Brian C. Murray A Christopher S. Galik A Stephen Mitchell B Phil Cottle C Prepared for The World Bank Carbon Finance Unit (BioCarbon Fund) A Nicholas Institute for Environmental Policy Solutions, Duke University B Nicholas School of the Environment, Duke University C ForestRe Ltd Acknowledgements The author team from Duke University gratefully acknowledges the financial support and content guidance from the World Bank BioCarbon Fund team, especially Rama Chandra Reddy, Marco van der Linden, Ken Andrasko, Klaus Oppermann, and Ellysar Baroudy. The report benefited greatly from the input provided at two workshops held at the World Bank in November 2011 and April 2012, as well as review comments on an earlier draft of this report by Derik Broekhoff of the Climate Action Reserve, Peter Graham of Natural Resources Canada, John Kadyszewski of Environmental Resources Trust/Winrock International, and Ruben Lubowski of the Environmental Defense Fund. We thank Tibor Vegh of Duke University for research support. All errors and omissions are those of the author team and not the sponsors or reviewers of this work. 1 Contents Executive Summary....................................................................................................................................... 4 1. Introduction: Carbon Sinks, Permanence, and Reversals in Climate Change Mitigation Policy ........... 8 1.1. Climate Change Mitigation Policies and Projects under the UNFCCC .......................................... 9 1.2. Sinks, Permanence, Reversals, and Crediting at the Project Level: Concepts and Examples ....... 9 1.2.1. Permanence Period: How Long Must Carbon be Stored? .................................................. 11 2. Reversal Risk: Types, Characteristics, and Liability ............................................................................. 15 2.1. Unintentional (Natural) Reversals............................................................................................... 16 2.2. Intentional Reversals................................................................................................................... 18 3. Risk Management Approaches ........................................................................................................... 18 3.1. Screening for Reversal Risk: Concepts and Criteria .................................................................... 19 3.1.1. Screening for Unintentional Reversals from Natural Disturbances .................................... 20 3.1.2. Screening for Intentional Reversals .................................................................................... 22 3.2. Liability Determination and Assignment..................................................................................... 22 3.2.1. Producer Liability ................................................................................................................ 22 3.2.2. Buyer Liability ...................................................................................................................... 23 3.2.3. Customized Contracts Between Buyers and Producers ...................................................... 23 3.2.4. System Liability ................................................................................................................... 23 3.3. Accounting Mechanisms for Addressing Reversals as They Occur ............................................. 24 3.3.1. Incremental Crediting Over Time (“Tonne Year� Approach) .............................................. 24 3.3.2. Full Crediting Upon Verification .......................................................................................... 27 3.3.3. Comparing Approaches ....................................................................................................... 41 4. Application of Modalities for Reversal Risks for Geological Carbon Capture and Storage (CCS) under the CDM in the Context of A/R ................................................................................................................... 45 5. Policy Decisions for Parties and Stakeholder Implications ................................................................. 46 5.1. Issues for Consideration.............................................................................................................. 46 5.1.1. Risk Screening ..................................................................................................................... 47 5.1.2. Timing of the Issuance of Credits ........................................................................................ 47 5.1.3. Issuance of Credits .............................................................................................................. 47 5.2. Implications for Countries........................................................................................................... 48 5.2.1. Implications for A/R CDM Host Countries........................................................................... 48 5.2.2. Implications for Annex I Buyer Countries ........................................................................... 48 5.3. Implications for Project Participants ........................................................................................... 49 6. References .......................................................................................................................................... 49 Appendix A. Comparison of Reversal Approaches in Existing Standards ................................................... 52 2 Appendix B: Unintentional Reversal Assessment Methodology ................................................................ 54 LANDCARB Overview .......................................................................................................................... 54 Empirical Disturbance Analysis - Chilean Historical Fire Data Overview ............................................ 56 Appendix C: Intentional Reversals Case Study: Competition from Soybean Production ........................... 58 Initial Conditions Favoring A/R ............................................................................................................... 58 Change in Market Conditions Favoring Agriculture ................................................................................ 61 Outcomes Under Price Shocks ................................................................................................................ 63 Appendix D: Insurance for Forestry Projects - Approach and Key Terms ................................................... 65 Appendix E: Project Cost Data and Assumptions........................................................................................ 67 Calculation of tCER Pricing ...................................................................................................................... 68 Appendix F: Expanded Output and Sensitivity Analysis.............................................................................. 69 Tonne Year .............................................................................................................................................. 69 Buffer Set Aside....................................................................................................................................... 69 Commercial Insurance ............................................................................................................................ 72 3 Executive Summary Afforestation and reforestation (A/R) projects can generate greenhouse gas (GHG) reduction credits by removing carbon dioxide (CO2) from the atmosphere through biophysical processes and storing it in terrestrial carbon stocks such as biomass, litter, and soils. One feature of these A/R activities is the possibility of non-permanence, whereby the stored carbon is subsequently lost though natural disturbances such as fire and wind or anthropogenic disturbances such as harvesting. These disturbances cause the stored carbon to be released back into the atmosphere as CO2, thus providing a temporary climate mitigation benefit. Adequately accounting for non-permanence under land use, land-use change, and forestry (LULUCF) activity such as A/R has been a point of ongoing discussion at the United Nations Framework Convention on Climate Change’s (UNFCCC) Conference of the Parties (COP). Specifically, the 17th Conference of the Parties serving as the meeting of the Parties to the Kyoto Protocol (CMP) in Durban, South Africa, requested the UNFCCC’s Subsidiary Body on Scientific and Technological Advice (SBSTA) to initiate a work program to consider modalities and procedures to address “the risk of non-permanence� (which this report refers to throughout as the risk of a carbon reversal) in A/R activities, starting with activities covered under the Clean Development Mechanism (CDM). Regulatory and voluntary program precedents exist for addressing reversals in LULUCF activities, including A/R, forest management, and reducing emissions from deforestation and forest degradation (REDD+). Reversal risk in A/R CDM projects is currently handled by issuing temporary credits for carbon storage which expire at some date in the future, requiring replacement at that time. The CMP decision focuses on consideration of alternative approaches to address the risk of reversals. Yet there has been limited analysis to show how different approaches perform in protecting the integrity of the offset mechanisms in which they operate and their cost-effectiveness. This report examines alternative approaches for addressing reversals to inform ongoing UNFCCC discussions on (1) the effectiveness of various approaches in handling real-world reversal scenarios in ways that ensure net carbon balance and integrity of the A/R offsets; and (2) the economic and practical feasibility of various approaches, taking into account the costs and returns of A/R projects. This report provides a conceptual basis for viewing the non-permanence issue, evaluating current approaches to address reversals and highlighting implications for policy and investment decisions. The key policy issues include:  Risk screening requirements for A/R projects.  Whether to issue credits incrementally over time (rather than all at the time of initial verification).  If credits are issued up front, whether to classify them as temporary or permanent in nature.  If permanent credits are issued, what replacement requirements should be considered when reversals occur? Should these requirements differ for unintentional reversals (caused by natural disturbances) and intentional reversals caused on purpose by project participants?  What risk management mechanisms should be put in place (if any) to ensure that projects can meet replacement requirements? 4 Toward this last issue, the report conducts analysis using forest carbon risk data and models in settings relevant to A/R projects. The analysis draws from modeling of multiple policy and accounting mechanisms for handling reversal risks using observational data on natural disturbances (e.g., fire and wind). It also explores reversals that could arise from intentional actions to clear a forest before a project is slated to end. Finally, the risk of unintentional and intentional reversals is explored together. Within each scenario, we quantitatively and qualitatively assess the impact of several different risk management mechanisms, including categorical exclusions or exceptions for risk management requirements; temporary credits, as now used for A/R projects under the CDM; “tonne year� approaches, where permanent credits are issued incrementally over time as carbon is retained; credit reserve buffers, a common method for addressing reversals in the voluntary market wherein a share of permanent credits issued at the time of verification are set aside in an account and accessed in case of a reversal; commercial insurance, in which a third party contracts to cover credit replacement risk for a fee; host country guarantees, wherein the country hosting the A/R project agrees to satisfy otherwise uncovered reversals at the subnational scale; and combinations of these approaches. The report’s analysis does not seek to recommend a specific approach for A/R mitigation projects or specific parameters for different approaches (e.g., set-aside percentage for credit buffers, insurance premium levels), which are best informed by careful examination of the risk factors affecting each country or project. Instead, the report outlines the options available and their relative strengths and shortcomings, thus providing insight to inform the UNFCCC/CMP with regard to decisions addressing reversals in A/R activities. Key messages from the analysis include:  The concept of permanence has biophysical, political, and practical foundations. Any subsequent release of stored carbon ultimately negates the original benefits of storage from an atmospheric standpoint. But practical realities dictate that policy and contract commitment are typically for finite periods of time. Policymakers may therefore opt to make “permanence� achievable within a fixed time period rather than at the elusive “end of time.�  Empirical analysis of unintentional risks from natural disturbances finds the following determinants of risk management performance: o Location matters. Ground data can reveal where projects are more (or less) likely to confront reversals. o Scale matters. Over time, large projects have less relative risk of catastrophic loss from reversal than do small projects. More area in a project means that some part of the project may experience reversal, but it is less likely that the reversal is catastrophically large. o System dynamics matter. We model a representative forest system under likely A/R project conditions, but the selection of different species or types of operation in a different disturbance regime may yield different conclusions. In the system modeled here, forest growth dynamics and disturbance characteristics combine to make longer projects more susceptible to reversals than shorter ones. o There is power in risk diversification. Building on scale effects, pooling together risks from small projects into a larger portfolio of projects can reduce the relative risk of reversal for an A/R activity.  The risk management mechanisms examined have a range of features and tradeoffs among risk conservatism, economic returns, and other factors. Some approaches deal with risk very 5 conservatively, but tend to have lower financial returns; some approaches are less conservative, thus yielding higher returns but requiring risk back-up mechanisms to ensure integrity.  The mechanisms vary in their ability to effectively address both intentional and unintentional reversals. Projects may face both kinds of risk over the life of the project, and both should be assessed at the initial risk assessment stage to inform the risk management process. Virtually all crediting mechanisms examined are designed to deal with unintentional risks from natural disturbances such as fire and wind, but some mechanisms may be less effective in addressing intentional risks from a project holder’s decisions to pursue other objectives. Key findings on this issue across the risk management mechanisms include: o Temporary crediting in the form of tCERs, such as those used in the CDM to date, may provide both intentional and unintentional reversal risk protection to the atmosphere by requiring credits to expire and be replaced periodically; this conservatism comes at a cost, however, and may not be able to adequately incentivize A/R projects. o A tonne year approach to carbon crediting, which issues credits incrementally over time as carbon storage is retained, avoids the need to reclaim credits after they are issued and reversed (either intentionally or unintentionally) to protect system integrity. Our analysis suggests that this approach can be more attractive financially than the temporary crediting approach (since the credits issued are deemed permanent, for which the market will pay more), but this depends on the specific parameters (e.g., the length of the assumed permanence period and the corresponding rate at which permanent credits are incrementally issued for carbon storage). o Permanent credit issuance up front, backed by a buffer mechanism, can provide a practical alternative to temporary crediting and can work to protect against reversals if the buffers are adequately built, and managed. A buffer is one of several mechanisms evaluated that allows permanent credits to be issued once storage is verified, which improves financial performance (assuming that permanent credits command a price premium relative to their temporary credit counterparts). In the case of buffers, effective protection against reversals requires a robust and location-specific risk assessment to determine the appropriate size of the buffer withholding requirement and other operating procedures on a case-by-case basis. Buffers may be ineffective against intentional reversals, which are inherently difficult to model at a system level. A high prevalence of intentional reversals could cause a system-wide buffer to collapse and put the entire system at risk, requiring further back-up mechanisms (such as those discussed below). o Permanent credit issuance up front, backed by commercial insurance, could be an effective and more actuarially refined mechanism than a buffer to address unintentional reversals of the issued permanent credits. Although products to insure forest carbon are still in their formative stages, insurers can draw from their experience insuring timber and other properties affected by natural risks to develop products that protect a project against extreme risks at a cost comparable to or less than alternative approaches, depending on project length. Commercial insurance, however, is not well-suited to cover against intentional actions. o Host country guarantees can provide a further backstop against reversal risk mechanisms established for projects within the country. o Modalities established for carbon capture and storage (CCS) projects under the CDM, wherein a mix of a buffer, minimum permanence period, and host country guarantees could create a workable analog for A/R, would need to be refined to capture the risk characteristics of forest carbon storage vis-à-vis the geological storage of CO2 in CCS 6 projects. For example, CCS reversal risks may diminish over time as below-ground CO2 stabilizes, whereas A/R reversal risk may increase as the biomass in above-ground pools increases. Intentional reversal factors differ between A/R and CCS projects; these should be taken into account as well. Further Policy Issues A major issue is the ease by which projects can address intentional reversals if permanent credits have been issued in advance. Screening criteria, enforceable guarantees, and opt-out provisions need to ensure that any deemed-permanent carbon credits issued are replaced, but questions remain as to the implementation of such provisions in the context of CDM activities. In the event of intentional reversal, buyers or sellers could be made liable to replace the credits issued thus far. Commercial insurance is not well-suited to cover against these intentional actions, and a system-wide buffer could put the entire system at risk if the prevalence of intentional reversals is high relative to the size of the buffer. Alternatively, temporary crediting and tonne year approaches could accommodate this form of reversal without bringing the system down; however, due to the potential lack of economic viability, the success of such projects is uncertain. Flexibility is key. It is advantageous to consider a flexible system where project investors and credit buyers have a menu of approaches for dealing with reversals, as long as safeguards are put in place to ensure environmental integrity in the most cost-efficient way. One of the advantages of a flexible menu- driven system is that it can provide incentives for innovative insurance and financing mechanisms to evolve and provide near-term and long-term options for project investors. From the perspective of project participants, choice among approaches to dealing with reversals may also be advantageous. The choice of approaches can create opportunities for project participants to pick and choose their approach. From the perspective of the project participants, it is also important that that approaches are cost-effective to apply and lead to fungible credits. From the perspective of a regulatory agency, clear guidelines need to be put in place to support the implementation of different approaches (and combinations of approaches) to ensure that each is verifiable, ensures the environmental integrity of the project, and is practicable to apply. By anticipating reversal risk and pooling such risk across projects, it remains feasible to create a mechanism that protects against net carbon loss without sacrificing the financial viability of A/R projects. Indeed, the analysis herein shows that a certain level of buffering and aggregation lowers both the chance of an offset system going negative and the extent of loss experienced should this actually occur. While the analysis focuses on A/R projects, similar conclusions may be surmised for other types of forest carbon projects (such as REDD+) and other terrestrial mitigation activities (such as wetland restoration and agriculture). The emphasis rests not on the project type but on the proper analysis of risk coupled with modeling of reversal scenarios to enhance the likelihood of the offset system remaining a net carbon sink. The analysis in this report reflects on issues that the UNFCCC Parties may wish to consider in deciding how to address non-permanence with A/R and, potentially, with other LULUCF activities under the CDM: risk screening requirements, incremental versus full issuance up front, replacement requirements, risk management options, opt-out provisions, and the like. As indicated here, Parties may want to consider allowing flexibility given the voluntary nature of the CDM; however, provisions must be established for determining which actions require credit replacement and, if so, by whom. 7 1. Introduction: Carbon Sinks, Permanence, and Reversals in Climate Change Mitigation Policy Land use, land-use change, and forestry (LULUCF) comprise about 30 percent of global greenhouse gas (GHG) emissions (IPCC 2007). A substantial part of this flow is tied to the absorption, storage, and release of carbon dioxide (CO2) in soils, biomass, and other organic pools referred to as “carbon sinks.� Sinks can accumulate carbon through both the maintenance of preexisting stocks (e.g., reduced deforestation, degradation, or other forms of land clearing) or through the creation of new stocks (afforestation, reforestation, improved management, and other forms of restoration). As such, terrestrial carbon sequestration projects are part of the GHG mitigation strategy set, typically identified as a potential “offset� for emissions from other sources.1 In principle, using a tonne of terrestrially stored carbon (or CO2 equivalent, tCO2e) as an offset is an equivalent credit against an (allowed or capped) emission if it completely negates the climatic impact of that emission.2 Recognizing the importance of terrestrial carbon sinks in climate mitigation; policies have been designed and implemented to expand carbon sinks. However, these terrestrial ecosystems are susceptible to disturbances that cause the stored carbon to be released back into the atmosphere. Problems can arise when stored carbon that has been credited as part of a climate change mitigation effort returns to the atmosphere via these disturbances, a phenomenon known as “reversal.� Reversals, when they occur, can nullify emissions reductions and undermine the permanence of these climate mitigation actions; they must be addressed through policies and accounting procedures. The distinction between reversal and non-permanence is at times a subtle one, but critically important to devising workable approaches for dealing with carbon loss. As used herein, a “reversal� is a reduction in carbon storage relative to some previously credited amount (e.g., a net loss of carbon credits), whereas “non-permanence� refers only to the inherent vulnerability of a carbon stock to reversal. See also Box 1 and Section 1.2. Box 1: Non-Permanence v. Reversal v. Non-Performance It is important to distinguish between the concepts of non-permanence, reversal, and non-performance in the context of terrestrial carbon sequestration projects. The inherent susceptibility of terrestrial carbon projects to rerelease of stored carbon is described as non-permanence; it is impossible to guarantee that a given tonne of carbon stored in a given terrestrial carbon pool will remain sequestered forever. Sequestration credits are generated during a time period if there is a net increase in carbon storage relative to the crediting baseline during that period. Should an unanticipated release of carbon subsequently occur, the loss may be termed a reversal if it causes the carbon stock to drop relative to the baseline. If prior generation of carbon gains produces a project credit, then a reversal that creates a net carbon loss can be viewed equivalently as a project debit – and some sort of accounting adjustment is necessary to balance the books. However, if the disturbance event causes a loss of carbon that is less than the total amount gained elsewhere onsite over the same time period, the end result is not a debit or reversal per se but a diminishment in the number of credits that are generated during that period. The project on balance still gains carbon, but not as much as would have been expected in the absence of the disturbance event, a phenomenon that may be referred to as non-performance or under-performance. Note that there may be other forms of non-performance unrelated to disturbances, such as the failure of an afforestation and reforestation (A/R) project to physically yield as much carbon as initially projected or the failure of certain actions to as effectively reduce emissions from deforestation. 1 Terrestrial carbon storage may also be directly regulated as part of larger emission reduction obligations, as in the case of New Zealand’s Emissions Trading Scheme. 2 The term “tonne� throughout the document will refer to a metric ton ne, or megagram (Mg), of CO2 equivalent. 8 1.1. Climate Change Mitigation Policies and Projects under the UNFCCC Carbon sink mechanisms operate at two different scales, national and project, under the UN Framework Convention on Climate Change (UNFCCC). National incentives for carbon sinks have until recently focused on the inclusion of LULUCF activity in the national accounting of Annex I (developed) countries under the 1997 Kyoto Protocol (KP). Carbon stock enhancement and emissions avoidance can help Annex I countries meet their KP emission reduction obligations. Under the KP’s Clean Development Mechanism (CDM), developing countries can host carbon sink projects that generate certified emission reduction credits. These credits can be sold to Annex I countries to help them meet their emissions reduction obligations. Joint Implementation (JI) guidelines also provide for the opportunity for LULUCF activities (JISC, 2009) at the project level within Annex I countries, coordinated with national accounting. The focus of this report is on the CDM in developing countries.3 Afforestation and reforestation (A/R) are eligible project activities under the CDM in the second commitment period (per the 2/CMP.7 decision by parties to the Protocol). Although the UNFCCC is now considering the inclusion of additional land use, land-use change, and forestry activities under the CDM, the focus of this report is A/R. Currently, CDM sinks projects address reversals by issuing expiring (temporary) credits. Upon expiration, these credits must be replaced. This replacement requirement raises the cost to the buyer of using them (relative to a full-price permanent credit), thereby reducing the monetary value of the credit and the net revenue flow to the project. As A/R projects have not been widely adopted thus far – they account for less than one percent of all CDM projects to date (UNEP, 2012) – the question is whether other approaches for dealing with reversals are needed. Toward that end, the seventh session of the Conference of the Parties serving as the meeting of the Parities to the Kyoto Protocol (CMP7) in its recent decision on LULUCF under the KP requests: the Subsidiary Body for Scientific and Technological Advice to initiate a work programme to consider and, as appropriate, develop and recommend modalities and procedures for alternative approaches to addressing the risk of non-permanence under the clean development mechanism with a view to forwarding a draft decision on this matter to the Conference of the Parties … (UNFCCC, 2011a) In light of this request, the focus of this report is on approaches to address the risk of reversal in A/R project-level activities under the CDM, with broader implications drawn for programmatic or system- level approaches beyond individual projects. 1.2. Sinks, Permanence, Reversals, and Crediting at the Project Level: Concepts and Examples LULUCF activities are subject to both natural and anthropogenic disturbances. Relevant natural disturbances include fire, wind, flood, drought, ice/snow, pest infestations, disease, landslides, earthquakes, and volcanic activity (see, Galik and Jackson, 2009, for a review). Human-induced disturbances include the legal or illegal harvesting of trees, land clearing, and incidental mortality 3 JI projects for LULUCF activity within Annex I countries can generate credits that Annex I countries can use to meet obligations under the Kyoto Protocol. However, their use has been almost non-existent because of EU ETS limitations on the use of forestry credits from either the JI or CDM. 9 occurring as a result of other activities (e.g., war). The intensity and extent of disturbance can vary for both human-caused and natural events, ranging from slight damage to complete loss and from individual trees to thousands of hectares. This section describes disturbance types for A/R activity. Afforestation and reforestation both entail the establishment of forests on land that is currently non- forested. For the purposes of the CDM, afforestation and reforestation are defined by the Marrakech Accords: “Afforestation� is the direct human-induced conversion of land that has not been forested for a period of at least 50 years to forested land through planting, seeding, and/or the human-induced promotion of natural seed sources; “reforestation� is the direct human-induced conversion of non-forested land to forested land through planting, seeding, and/or the human-induced promotion of natural seed sources on land that was forested but that has been converted to non- forested land. For the first commitment period, reforestation activities will be limited to reforestation occurring on those lands that did not contain forest on 31 December 1989 [FCCC/CP/2001/13/Add.1]. Although there are semantic and legal reasons for treating them separately, the mechanics of carbon sequestration and reversal are similar between the two. The rate of sequestration will depend on a variety of site- and project-specific factors, but the sequestration trajectory generally follows a logistic- like curve (see Figure 1). Carbon accumulates slowly as the stand is established. The sequestration rate then generally increases for a time, and then slows as the stand reaches maturity. Figure 1 depicts the carbon profile for a hypothetical 40-year project using data derived from the quantitative analysis described below. In the absence of disturbance, the area under the “Live Tree, Undisturbed� line reflects the per-hectare carbon sequestration benefits generated in the live tree carbon pool over that time. Note the general profile, where early year storage occurs slowly, building over time before finally plateauing or even declining. Under the simplifying assumption that the alternative land use to an A/R project would accumulate no carbon, these cumulative carbon stock benefits provide the starting point for project crediting. The effects of unexpected natural disturbances on the live tree pool are shown in Figure 1 by the “Live Tree, Subject to Disturbance� line. The amount of carbon lost and the rate of future carbon storage are both functions of disturbance timing, intensity, and extent. In the early years of project implementation, less carbon has been accumulated and therefore less is at risk. While the ratio of sequestered carbon to potentially lost carbon may not change over time, loss magnitude will increase; larger losses are inherently more expensive to address. Early-year disturbances are also more likely to be masked by rapid growth occurring elsewhere on the stand; this is seen in Figure 1, “Live Tree, Subject to Disturbance,� as the rare and minor early year reversals as compared to the large, recurring later-year ones. Including additional carbon pools in the project (e.g., lying dead wood, standing dead wood, and litter) can also act as an implicit hedge against disturbance. Because disturbance does not result in the instantaneous loss of carbon onsite, but rather involves a transfer between pools (e.g., “live tree� to “dead tree�), the carbon consequences of a disturbance are somewhat muted at the stand level when more pools are included (the “All pools, Subject to Disturbance� line).4 The interrelated dynamics of 4 Harvested wood products (HWP) represent another potential hedge against carbon loss. This analysis does not include planned harvests or post-disturbance salvage operations, and therefore does not assess contributions of the HWP pool to total forest carbon. 10 growth, timing, and carbon pool choice are therefore all important to consider when weighing mechanisms to address reversals in this particular activity type. Figure 1. A/R carbon stock accumulation, with and without natural disturbance. 1.2.1. Permanence Period: How Long Must Carbon be Stored? If the carbon stored in the terrestrial pool remains there forever, then it has served its offsetting function. If the stored carbon is released at any time in the future, however, a key question is whether that offset has effectively negated the emission allowed. To that end, the permanence of biologically sequestered carbon can be defined as the point in time when the stored carbon has essentially fulfilled its role in offsetting the global warming potential of the original emission that it is offsetting. Determining the equivalence of a unit emitted and sequestered in the same year is complicated by the issue of how long and at what rate CO2 and other GHGs reside in the atmosphere. The original emission that created the offset opportunity does not itself remain in the atmosphere forever. It decays over time, as would have the CO2 that was removed from the atmosphere and stored in a terrestrial carbon reservoir via carbon sequestration. The “permanence equivalence� nature of the problem stems from the relative patterns of atmospheric CO2 residency from these two events. The operative question is whether the carbon returned to the atmosphere completely negates the climate benefit of the offset or whether the timing of the subsequent release matters. In other words, is permanence absolute or relative? The answer depends on the residency time of CO2 in the atmosphere and the time horizon over which CO2 concentrations are being targeted. The relevant time horizon relies as much or more on policy judgments as on atmospheric science, as we shall now discuss. 11 1.2.1.1. Permanence in the Context of Atmospheric Chemistry5 Greenhouse gases are stock pollutants, in that it is the accumulated level in the atmosphere that matters rather than the amount introduced in any one year. Accumulated increases in GHG concentrations alter the radiative balance of the atmosphere by enhancing the absorption of outgoing long-wave radiation, which raises global temperatures. The time profile of atmospheric residency for a unit of CO2 emitted into the atmosphere is a critical consideration. From an atmospheric chemistry perspective, a pulse of “excess� CO2 released into the atmosphere decays over time (Figure 2). 1.0 Year 0.0 0 20 40 60 80 100 Figure 2. Representative decay function for CO2 in the atmosphere following emission. The horizontal axis displays time and the vertical axis represents the portion of the initial CO2 remaining in the atmosphere. One can approximate the fraction of excess CO2 that remains in the atmosphere at some point in time following a release, and some portion of it remains in the atmosphere indefinitely. Therefore, from a long-run atmospheric perspective, any reservoir created by a carbon sink to offset the excess CO2 pulse is equivalent only if the carbon it contains remains stored indefinitely.6 This is because the release of CO2 from the offset reservoir back into the atmosphere will have the same cumulative effect on the atmosphere (called the “integrated climate forcing�) as the original emission; the only effect would be a delay in when the climate-forcing effect would start (which may have some economic implications in terms of the cost of climate damages, but in the long run the climate consequences are essentially the same). This infinite horizon view of CO2 residency implicitly underlies the carbon-accounting approaches discussed below, that require any rerelease (reversal) of terrestrial stored CO2 to fully cancel any offset credits generated by the project no matter when they occur. Other approaches discussed below, meanwhile, consider the possibility of at least crediting for storage of carbon over a finite horizon. “Permanence� in a Finite Policy Horizon The warming potential created over a specific time period is often the relevant horizon for policy purposes. While the goal may be permanent reductions in atmospheric GHG concentrations, the policy itself often involves fixed emissions targets for finite periods (e.g., to 2020, to 2050), presumably adjustable by future policy decisions. 5 Much of the work in this section is based on collaboration between one of the co-authors, Brian Murray, and Duke colleague Professor Prasad Kasibhatla. That work will be released in a more extended form in a forthcoming manuscript (Murray and Kasibhatla, forthcoming). 6 Here we follow the IPCC convention of referring to a carbon sink as the flow of CO2 removed from the atmosphere and stored in a terrestrial carbon stock reservoir or pool such as biomass or soil. We clarify this point as it is not uncommon to elsewhere see the carbon sink referred to as a stock, rather than a flow. 12 1.2.1.1.1. Permanence at the End of the Policy Period The Kyoto Protocol’s CDM addresses permanence at the end of the policy period by establishing that temporary credits for A/R projects are only valid until a certain date, at which point they expire and must either be re-verified or replaced with permanent credits from another source. As a result there is, in essence, no real permanent equivalence for storage – just deferred replacement. In this case, the policy-related time horizons are more like checkpoints on the way to full replacement rather than milestones on the way to achievement of permanence. In this regard, the temporary credits are essentially a form of deferred obligation to replace A/R credits with “permanent� credits rather than an indication of cumulative progress of A/R carbon storage toward some long-term notion of permanence. The voluntary market, however, has taken a more flexible view of permanence. It issues permanent (rather than temporary) credits, typically with finite contract periods under which the landholder commits to keeping the carbon in place. Perhaps the clearest statement of the relationship (or lack thereof) between finite contract length and permanence was made by the American Carbon Registry (ACR), which in explaining its 40-year contract period, stated: AFOLU [Agriculture, Forestry, and Land Use] carbon protocols sometimes confuse permanence with the length of time for which a Project Proponent or landowner must commit to maintain, monitor, and verify the project activity. In fact, minimum project duration and the assurance of permanence are unrelated. No length of time short of perpetual is truly permanent, nor is there a sound scientific basis or accepted international standard around any particular number of years... ACR requires Project Proponents to commit to a Minimum Project Term of forty (40) years for project continuance, monitoring, and verification. ACR views forest and other AFOLU activities as a “bridge� strategy to achieve near-term reductions cost-effectively over the period from now through 2050 – the timeframe over which U.S. legislative frameworks and international negotiations propose effective de-carbonization of major emitting sectors, with reductions of around 80 [percent] below current GHG emissions. Requiring Project Proponents to commit to 40 years ensures these activities will continue over the relevant timeframe, or if they or their landowners choose to discontinue activities, that any credited [Emission Reduction Tonne] will be replaced. (American Carbon Registry, 2010, p.30) In other words, ACR sees the contract length as a means to keep sequestered carbon aligned with time commitments tied to the underlying climate policy process, at least in the context of U.S. federal policy proposals that were in place at the time the statement was written in 2010. There is not a single cap- and-trade program for carbon that establishes a cap into infinity. 1.2.1.1.2. Reversals During the Policy Period The ACR approach, and others in the voluntary market, require full replacement of credits that are reversed before the end of the time period. A possible modification of this approach is to partially credit for storage that accrues during the project and then reverses before the project is over. One such approach is the tonne year approach (Moura-Costa and Wilson, 2000; Noble et al., 2000), which is similar in some ways to the rental approach described by Sohngen (2003) in which credits accrue the longer the carbon is stored. In this approach, tonnes stored early on in a project receive small payments that progressively accumulate as the project continues and achieves storage over a longer period. Since payments are contingent on permanence, there is no “up-front� payment for permanent credits once initial storage is verified. Rather, a reversal simply reduces the basis for subsequent payments. 13 As an example, the atmospheric effects of a sink reversal under a finite time horizon of 100 years are displayed in Figure 3. The creation of the sink tonne in Year 0 produces an atmospheric credit value of - 1. At the same time, the emission that is allowed by generating a sink offset credit produces a debit value of +1. As discussed above, the emission tonne allowed by the offset decays over time (depicted by the red line). The total radiative forcing – the amount of warming potential – of the allowed emission is captured by the area A+B. The tonne of CO2 that is removed from the atmosphere during sink creation (the blue line, which is the inverse mirror image of the red line) would have the equivalent negative forcing effect (C=B+A) if the sink tonne stays intact for the full 100 years, and thus will have offset the atmospheric effects of the corresponding emitted tonne in Year 0. For the purposes of this 100-year time horizon, the sink will have met the permanence requirement. 1.00 0.75 0.50 D 0.25 Year A B 0.00 0 10 20 30 40 50 60 70 80 90 100 -0.25 C -0.50 -0.75 -1.00 Removal Benefit Emission Contribution Release Figure 3. Net radiative forcing effect of a sink (removal) created in Year 0, followed by a release in year 50. Area B+D are the cumulative effect on atmospheric CO2 concentrations. The initial emission offset by the sink has a cumulative effect of A+B. The atmospheric effect of the reversal scenario within the 100-year horizon is smaller (D RA Where RAR represents A/R project returns, including all carbon credit payments less the cost of establishment and ongoing operating costs incurred (for measurement, monitoring, and verification), and RA represents returns from alternative land use (agriculture). Both terms are further described below. RAR: Returns from an A/R Project A/R project returns include all carbon credit payments less the ongoing operating costs incurred (for measurement, monitoring, and verification). Mathematically, this can be further specified as: ∑ - EAR [A3] Where PC is the carbon price, QC is the quantity of carbon credits generated, cC is the annual operating cost, and r is the annual discount rate. T is the length of time of the project and the subscript t indicates the year of occurrence between project establishment (t=0) and project end point (t=T). EAR is project establishment cost, which include the cost of planting the trees as well as the upfront costs of planning, registering, and implementing the project. The use of the expectations operator indicates that future carbon prices are unknown at the time of the investment, a point to which we will return below.45 RA: Returns from Alternative Land Use (Agriculture) The returns from an alternative land use, such as agriculture can be specified as: ∑ [A4] Where PA is the alternate commodity (agriculture) price, QA is agricultural output, cA is the annual agricultural production cost, and all other variables and subscripts are as defined above. 45 We recognize that the other variables – yields and costs – are also uncertain, but we focus the discussion on uncertain prices as they tend to be subject to the external volatility of most concern. 58 In light of this decision rule, we consider the case of a 1,000-hectare tract of land that is arable and can be used for soybean (soya) production or can host an A/R project.46 In this example, we assume the A/R project commitment is 40 years so that we can compare the present value of A/R returns with agricultural returns over that time period. We assume for now that carbon standing at the end of the 40-year period is deemed permanent for crediting purposes. Figure AC-1 shows how sensitive this A/R investment decision is to different price assumptions. Under the efficient markets hypothesis (see, e.g., Malkiel, 1987), current prices provide the best expectation of future market prices; thus landholders will take these price levels, as well as their underlying variability and risk preferences, into account. The point where each line in the figure crosses over the horizontal axis (where the difference between A/R and soy returns is zero) represents the break-even CO2 price. The break-even price for CO2 and soybeans are positively correlated. At a low soybean price ($250/tonne), any carbon price above $2.50/tonne CO2e favors A/R. At higher soybean prices ($450/tonne), the CO2 break-even price is above about $17 for an A/R investment. Figure AC-1. Relative returns at different price combinations: A/R vs. Soybeans, marginal land (annualized) 46 As elsewhere in this report, we base our modeled scenarios on factors identified in existing A/R PDDs. Soy is mentioned as an alternative commodity in at least one PDD (“Reforestation of Grazing Lands in Santo Domingo, Argentina�). Other commodities (e.g., wheat, barley, rapeseed, rice, maize, sorghum, sugarcane, coconut, and cocoa) are likewise represented in one or more PDDs. The example presented here should therefore be seen as illustrative, not exhaustive. We used data from soya yields and costs from a South American country for illustrative purposes and do not imply that soy production is any more or less at risk of intentional reversal than any other commodity. 59 Land-Use Decisions on a Landscape of Varying Quality Not all land is of equal quality for an A/R project or agriculture, and the relative returns will reflect this. Figure AC-2 illustrates a profile of land returns for A/R and agriculture along a continuum of land quality. “Quality� here reflects arability or suitability for agriculture (see Murray 2003). Higher quality land yields higher returns for both agriculture (RA) and A/R (RAR). Given an initial set of prices, PA and PC, agricultural returns are higher than A/R up to the point that the two lines cross – the land-use margin. Under these circumstances, we would expect the highest quality land up to the land-use margin to be allocated to agriculture, and the remainder to be allocated to A/R projects. Figure AC-2 reflects the initial allocation of land after A/R project opportunities are introduced, so that LAR reflects the amount of land initially allocated to A/R; the rest of the land stays in agriculture. Any land incapable of generating positive returns for either agriculture or A/R is considered idle land. Figure AC-2. Land allocation between A/R projects and agriculture over a land quality continuum. Following the logic described above, Table AC-1 compares soybean production on an average site (about 2.7 tonnes per ha per year) with an A/R project on an average site (ranging between 5-30 tonnes CO2e per year over 40 years, following a standard S-shaped growth function), at soybean and CO2 prices in the range of recent history ($360/tonne and $10/tonne, respectively). The returns for A/R are based on the net carbon price paid to the seller after any price adjustments for reversals referenced elsewhere in the report. (i.e., the price after a buffered amount has been set aside). Table AC-1 shows that soybean production on land of average productivity out-competes A/R (it has a higher return). However, we find that an A/R investment will break even with a low yield soybean site (about 80 percent of average yield) at the indicated prices. We can think of this break-even condition as the land-use margin referenced in Figure AC-2. At these prices, we might expect land with higher soy 60 productivity than the low yield estimate to remain in soy production and land less productive than that to potentially be more profitable as an A/R project. Table AC-1. Comparison of Returns to Soybean Production and A/R Project Note Soybean A/R Soybean Avg. Yield Project Low Yield Initial Yield (tonnes/ha/yr) a 2.68 5-30 2.16 Initial Price ($/tonne) b $ 360 $ 10 $ 360 Revenue (annualized) c $ 1,029 $ 347 $ 827 Operating Cost c $ 302 $ 22 $ 302 Net Income Before Overhead $ 727 $ 324 $ 525 Overhead Costs c $ 250 $ 49 $ 250 Land Return (annualized) $ 477 $ 275 $ 275 Notes: a. Soybean yield projected to increase 0.5%/yr; forest yields follow empirical yield function for subtropical softwoods over time b. CO2 price rises at the rate of discount following standard Hotelling price assumption for storable goods; soy price is constant c. Annualized over 40-year project period Data Sources:  World Soybean Production: Area Harvested, Yield, and Long-Term Projections http://ageconsearch.umn.edu/bitstream/92573/2/20091023_Formatted.pdf.  Economic Research Service, U.S. Department of Agriculture, http://www.ers.usda.gov/data/costsandreturns/testpick.htm.  Internal project A/R data (see above). Change in Market Conditions Favoring Agriculture The underlying uncertainty and periodic discrete shifts in commodity markets may create situations in which continuation of the project may seem unprofitable relative to an alternative land use. This can also happen if carbon yields are not as large as expected.47 Just as the decision to initiate an A/R project will depend on future expectations of carbon and commodity prices, so might the decision to stay with the project. Note that in the case of an A/R project reverting to agriculture, the decision to terminate a project and switch to agriculture involves a one-time clearing cost to revert, effectively the reverse of initial establishment costs. Given the observed dynamics of commodity markets, this could be a critical factor affecting landholders’ desire to maintain an A/R project after inception. Suppose at the time an A/R project is being considered, carbon prices are $12 per tonne of CO2e and soybean prices are $360 per tonne. Under these prices, as illustrated in Figure AC-1, the expected return from an A/R project exceeds the expected return from soybean production – and we assume the landholder undertakes the A/R project.48 Suppose after 10 years, however, there is a distinct shift in the 47 As commented on by one reviewer, landholders make these land-use decisions under uncertainty and thereby hold option value; they may wait until the uncertainty resolves before committing land to a use, such as forest, which involves a long-term commitment that is costly to reverse. This “wait and see� approach may weaken the response of landholders undertaking an A/R investment in the first place, but it may also reduce the incidence of post-investment regret and desire to switch back to agriculture (see Schaatzki, 2003) . 48 C A RAR (P =$12) = $5,089 /ha > RA (P =$360) = $4,144/ha 61 carbon and commodity markets; namely the carbon price drops considerably and the soybean price rises considerably. This is not outside the realm of recent history (see Figures AC-3 and AC-4 for recent price history of CO2 prices and agricultural commodity prices). The A/R project holder may now question whether the A/R project should continue if these prices hold. The following decision rule applies: Divest A/R project if R’AR < R’A - [SARA + CR] Where R’AR is the revised value of A/R returns over the remaining years of the project under the new prices, R’A is the revised value of alternative land use (soybeans) over the same time under the new prices, SARA is the switching cost associated with clearing the trees to enable cultivation, and CR is the cost of replacing the reversed carbon (if required). Figure AC-3 and Figure AC-4 represent observed historical prices for carbon (CO2e) and a suite of relevant agricultural commodities, respectively. Carbon prices are from the EU Emissions Trading System (EU ETS); they have a relatively short history because the carbon market has only been around for less than a decade.49 But, even in that time, the prices have shown a propensity for both short-term volatility and occasional discrete shifts reflecting changes in market fundamentals (Maniloff and Murray, 2011). Three discrete shifts in the EU carbon market can be seen. One occurred in early 2006, reflecting the release of the initial national emissions data – the market had traded without this essential data in its first year. The second shift coincided with the end of the first ETS trading period in 2008 and the beginning of the global recession. The third discrete drop in price started in early 2011 as the European economy experienced its own distinct financial and fiscal crisis and global climate agreements continued to stall. Figure AC-4 shows 30-year price histories (nominal and real) for four of the most relevant agricultural commodities that may compete with A/R investment: soybeans (South America), palm oil (Southeast Asia), cocoa beans (Africa), and cattle (South and Central America). As with the carbon market, each of these commodities shows a propensity for high price volatility and periodic shifts. Figure AC-3. EU ETS emissions prices: 2005-2012. Source: Point Carbon (pointcarbon.com). Downloaded, Aug 1, 2012 49 Although forest carbon does not trade directly in the EU ETS, we use that time series as an indicator of the potential shifts and volatility in the carbon market. 62 Figure AC-4. Relevant Agricultural Commodity Prices for A/R Alternatives Outcomes Under Price Shocks Consider an example under the following price shifts in Year 10: - Consistent with our initial assumption, the starting carbon price of $12 rises at the real discount rate (6%) and at Year 10 holds the following value, . But in year 10, the price drops 50%, = $10.14.50 - The soybean price in year 10 is assumed to remain constant at its initial value of $360, but in Year 10 it shifts up 25%, %, = $450. Under these circumstances, and using the same yield and cost data referenced thus far, the returns from remaining in the A/R project for the remainder of the project period (30 more years) and switching to soybeans, respectively are: If we estimate the cost of clearing the 10-year old A/R project of trees to be $250 per hectare, then the net payoff of converting the A/R project to soybeans is $3,300 per hectare [($7,330 – $250) - $3,780] before considering whether or how to assign carbon repayment, an issue that is dealt with extensively in the main text. 50 The real rate increase for carbon prices is consistent with standard assumptions about carbon markets that allow banking and borrowing of allowances between periods. In those situations, the price rises at the real rate of interest in equilibrium, as the holder of an allowance would be indifferent between using it in the current or future period. 63 Table AC-2 considers different combinations of price shocks in the carbon and soybean market (a 50- percent drop in carbon price, a 25 percent increase in soy prices) to consider the cases where only the carbon price shock occurs and soy price remains unaffected, and vice versa. Results in Table AC-2 focus on situations where (1) all replacement liability is covered by the project holder rather than a third party, or (2) the project pays the balance of what is owed after an initial (30 percent) of credits set aside in a buffer are used (see main text for description of the buffer approach). Table AC-2. Returns to project abandonment with full credit replacement required. Values in the table are net returns to termination under different shocks to the soybean and carbon markets and different approaches to covering liability (no mechanism and buffer). For example, if the soy price rises 25 percent and the CO2 price drops 50 percent, the landholder would gain $1,877 by terminating A/R and converting to soybeans if they had to pay to replace all credits themselves. If the amount they pay into the buffer covered 30 percent of the replacement liability, then the return to conversion would be $2,304. Note however that conversion is not profitable if either the CO2 price stays the same or the soy price stays the same. 1. No Third-party Mechanism 2. 30% Buffer Soy Price Shock (%) Soy Price Shock (%) 0% +25% 0% +25% CO2 0% -- $ (3,636) CO2 0% -- $ (2,782) Price Price Shock Shock (%) (%) -50% $ (1,110) $ 1,877 -50% $ (683) $ 2,304 The results suggest that if the carbon price does not drop when the soy price goes up, the carbon liability is such that switching to soybeans is not profitable. Likewise, if the soy price does not go up when the carbon price plummets, then the return to abandoning the project may be negative. Thus it may take a swing in both directions for this type of project termination to be a threat. As would be expected, the buffer approach results in more favorable net returns in the form of lower net costs (in the presence of a carbon price drop or a soy price increase) and higher net returns (in the presence of both). This is because there are lower repayment requirements in the presence of a buffer; 30 percent of the credits were already set aside throughout the course of the project. In this respect, a buffer reduces the barrier to conversion. But if credit payback is required, there are no adverse carbon consequences to this as the atmosphere is “made whole� upon project exit. The key message here is that, under realistic conditions, commodity prices could change to favor termination of the project and conversion of the trees to agricultural production, a form of intentional reversal. Imposing the requirement that the project replace the credits reversed can make the difference between whether or not it is profitable to do so. Even with the payback requirement, however, in some cases project developers may find it optimal to opt out. The critical issue is that they do so only after making the atmosphere “whole� by replacing the reversed credits. 64 Appendix D: Insurance for Forestry Projects - Approach and Key Terms Contracts Almost universally one year insurance contracts. Any loss event occurring in that insured year is covered, even if it takes some months (or years) to measure the loss and to pay the claim. This is important in carbon projects as loss events and certification of the loss of carbon may be several years apart. Insurers would rather prefer annual verification of carbon status in order to be able to close their books on that “underwriting year.� Purpose In natural hazard insurance (fire, wind, and so forth), 90 percent of commercial losses are caused by 2 percent of events. Forestry Insurance is normally designed as catastrophic coverage. That is to say, it protects the project from the infrequent but severe event (i.e., an event that would have an adverse effect on the net present value of the project). Insurance of common and frequent losses is termed “dollar-swapping� when a premium is paid to insurers and then claimed back after losses. This is expensive and unnecessary to the survival of the project as a business proposition. Deductible Insurance contracts oblige the project to retain risk as a deductible; this is often set by the insurer after discussion with the insured. The deductible in carbon projects is designed to remove small frequent “attritional� losses (see above), often 95-99 percent of loss events. It is usually applied “each and every event� (EEL). Carbon projects need to keep back the carbon required as the deductible only for one year; after that, it may be released as the old insurance contract terminates. New deductibles are required for new contracts. They may be expressed as absolute dollar amounts, tonnes C or as a percentage of the loss with minimum and maximum values, or as a percentage of the total sum insured with minimum and maximum values. Total Sum Insured (TSI) The total value of the project (Tonnes C x Price/tonne). The premium rate is conventionally applied to the TSI. Annual Aggregate Loss Limit (AAL) this is the liability of the insurer and is the maximum value paid out by the insurer. Often described as the “loss limit, EEL, & AAL.� The AAL will be based on the 1:100 year modeled event, or the 1:250 year event on the basis that more infrequent events will not in all probability occur. It may be set arbitrarily. Once exhausted, the insurer is no longer liable to pay claims. If doubt exists about the AAL required, a reinstatement may be arranged in exchange for an additional premium (example 150 percent of the original premium). This may not be a likely tool in carbon projects, as by the time the loss is measured the insurance cover may have already expired. 65 Insurable Perils FLEXA (fire, lightning, explosion, and aircraft); wind, and tornadoes, flood, earthquakes, SRCCMD (strike riot civil commotion), ice storms, and drought. Risk Pricing Insurers price risk ideally on a pure risk basis (i.e., the price if no expenses or profits were required). Then they “gross up� for the last two items. This is done in a number of ways, according to the insurer. A simple example is that, if one needs a 50-percent margin for profit & expenses, then the pure risk price needs to be doubled. Pricing Methodology Traditionally, insurers do pricing methodology by inspecting loss data provided by the client. Where the probability of loss is very small, and this would lead to a technical pricing that is not commercial, insurers will rate on a rate-on-line basis (ROL). ROL is the percentage that the premium bears to the insurers’ liability: Liability (AAL) = $100; Premium = $4 => ROL 4%. 3% to 4% might be a typically acceptable ROL. If there is no risk, the ROL will still be at least 1 percent due to the opportunity cost of the insurer’s capital. Generally, each insured peril is rated separately. An exception is made when the rate of loss is due to all perils and it is not possible to separate out the effects of each one. Although rare in forestry, the present analysis yields only aggregate results of all perils – fire, wind, and ecological in-forest carbon fluxes. 66 Appendix E: Project Cost Data and Assumptions Afforestation/Reforestation (A/R) Costs Parameter Value Comments -1 Site Preparation $50 ha Costs will be highly variable. This is a moderate to low estimate, assuming some vegetation control or soil preparation using power equipment (power equipment cost is estimated at U.S. rates), and developing country wage rates. Inventory $30,000 Assumes developing country field technician costs of $15/day, and limited road -1 project access (e.g., relatively high amounts of time to travel to plots). Assumes enough plots to achieve +/- 10%-percent confidence interval at 95-percent statistical confidence. Assumes experienced staff compile inventory at developed country wage rates. Does not include major equipment purchases, such as multiple electronic data recorders. Occurs at project inception and again at 5-year intervals. Management Plan $30,000 Assumes a basic management plan with maps, inventory, prescriptions, and general -1 Preparation project harvest and road plans. Does not include detailed surveys of sensitive species. Occurs at project inception and again at 10-year intervals. -1 Regeneration $500 ha Assumed to be half of the cost of commercial forest regeneration in the U.S. Project Development $30,000 A low-end estimate based on observation of about 20 projects. This cost covers -1 project some map development and writing a project document. It does not include methodology development or significant payments to consultants for modeling. Pre-project Calculations $10,000 Assumes experienced staff that can quickly make calculations from inventory data. -1 project Field Verification $35,000 Slightly higher than a mid-range estimate to allow extra travel costs to remote sites. -1 project Based on observation of verification contracts of the past few years. Occurs at project inception and again at 5-year intervals. Validation $40,000 Cost is slightly higher than a mid-range estimate, based on observed validation -1 project contracts of the past few years. -1 Site Maintenance $1 ha A low “placeholder� rate. Actual costs could be lower or much higher. If higher costs occur, the higher costs should only be for the first 1-3 years after planting. Higher costs could be needed for control of competing vegetation or protection of plantings from herbivores. Field Sampling and $40,000 Includes the cost of an inventory, plus a modest amount for staff to prepare -1 Monitoring project monitoring reports for verification. Occurs at project inception and again at 5-year intervals. Annual Verification $1,000 A desk review performed in years when field verification is not performed. Although -1 Report project the time involved is low, transaction costs of contracting and liability costs of verifiers will likely cause these fees to increase. - Registry Maintenance $500 project Estimated from the APX fee schedule for a VCS account. Offset issuance and 1 -1 Fee year transfers are sometimes denominated in U.S. dollars and sometimes denominated in Euros. -1 Issuance/Registration $0.15 credit Estimated from current registry fees. Fee Carbon Price Increase 6 percent Increases at the discount rate, consistent with recent analysis of comprehensive climate policy initiatives (e.g., U.S. Environmental Protection Agency, 2009). Discount Rate 10 percent for in-country project development expenses; 6 percent for international capital. 67 Calculation of tCER Pricing The value of a temporary credit stems from the deferred compliance the credit generates. An entity that purchases a tCER offsets full compliance by the number of years the tCER stands viable. Short contracts will have heavily discounted credits, since the replacement requirement will be near at hand (Kim et al., 2008; Murray et al., 2007). Longer contracts should have lower discounts, but this depends on the expectation of future prices for replacement credits; if the price of replacement credits is expected to be much higher in the future than it is today, then temporary credits may have little value. For tCERs to maintain any value, prices of permanent credit must grow at a rate lower than the discount rate (Olschewski and Benítez 2005; Maréchal and Hecq 2006; Bird et al. 2004; Subak 2003). This assumes that tCER credits follow the same risk and cost profiles as permanent credits. Following this logic, the equation for determining tCER prices is simply (Maréchal and Hecq 2006): [A3] Where is the price of a temporary credit at the time of issuance, is the price of a permanent credit at the time of tCER issuance, is the price of a permanent credit at the time the tCER expires at year T, and is the discount rate. If the price of a permanent credit grows at a set rate (say, α) then the equation translates to: [A4] For example, if the price of a permanent carbon credit trades for $5, the discount rate is 6 percent, and carbon prices rise at 5 percent, the value of a tCER that defers compliance for five years would be: Under this equation, if the rate of growth of permanent credits is equal to or greater than the discount rate, the value of a temporary credit becomes zero or negative. The only way to alter this situation would be to include nuances in the pricing determination of tCERs for individual actors. One nuance could deal with regulatory certainty for specific industry groups. If one sector of the economy will no longer be regulated by its carbon emissions but cannot sell its purchased permanent credits, temporary credits would be a logical purchase. Further, high transaction costs for the sale of permanent credits combined with a growth in self-compliance could incentivize the purchase of temporary credits for certain individuals. For example, if an energy company plans to shut down a coal plant but finds it difficult to dispose of excess permanent credits that will result, then the purchase of temporary credits would be ideal. On the credit supply side, host countries may show favor toward temporary credits as they retain a shorter span of liability as compared to permanent credits. While some projects have traded on the delivery of future CDM A/R tCER credits, it is difficult to determine the precise price that tCERs will sell in the future. The World Bank’s BioCarbon Fund has dominated the purchase of CDM A/R credits. The price paid for these projects hovers in the $4-5 range,51 similar to the prevailing price of a permanent credit on the voluntary market (Diaz et al., 2011). 51 See the following CDM project descriptions: http://cdm.unfccc.int/Projects/DB/JACO1260322827.04/view and http://cdm.unfccc.int/Projects/DB/JACO1245724331.7/view (accessed 15 February 2012). 68 Appendix F: Expanded Output and Sensitivity Analysis Tonne Year We see that similar trends occur in the 1,000 ha projects as in the 20,000 ha, as discussed in the main body of this report. An important difference is that the smaller projects tend to perform poorer, financially, due in part to economies of scale with regard to transaction and implementation costs. The relative risk of catastrophic loss is also greater in smaller projects, as disturbance events tend to affect larger portions of the smaller project, which in turn affects income under a tonne year approach. 20 year project/100 year permanence 20 year project/40 year permanence 40 year project/100 year permanence 40 year project/40 year permanence Figure AF-1. Total credits generated in a 1,000 ha project assuming a tonne year approach, both 40- and 100-year permanence periods, and 20- and 40-year project lives. Figure AF-2. Project NPV for a 1,000 ha project assuming a tonne year approach, both 40- and 100-year permanence periods, and 20- and 40-year project lives. Buffer Set Aside As with tonne year, we find that similar trends exist in both 20,000 ha and 1,000 ha project examples operating under a buffer approach. 1,000 ha projects tend to perform poorer, financially, due in part to economies of scale with regard to transaction and implementation costs. 69 0% Withholding 10% Withholding 20% Withholding Figure AF-3. Project NPV for 0 percent, 10 percent, and 20 percent buffer for a 1,000 ha project over a 20-year period. 0% Withholding 10% Withholding 20% Withholding Figure AF-4. Project NPV for 0 percent, 10 percent, and 20 percent buffer for a 1,000 ha project over a 40-year period. 0% Withholding 10% Withholding 20% Withholding Figure AF-5. Project NPV for 0 percent, 10 percent, and 20 percent buffer for a 20,000 ha project over a 20-year period. 0% Withholding 10% Withholding 20% Withholding Figure AF-6. Project NPV for 0 percent, 10 percent, and 20 percent buffer for a 20,000 ha project over a 40-year period. 70 0% Withholding 10% Withholding 20% Withholding Figure AF-7. Ending buffer balance for 0 percent, 10 percent, and 20 percent buffer for a 1,000 ha project over a 40-year project life. Percentage above figure indicates the mean loss as compared to total credits earned by the project. 0% Withholding 10% Withholding 20% Withholding Figure AF-8. Ending buffer balance for 0 percent, 10 percent, and 20percent buffer for a 20,000 ha project over a 40-year project life. Percentage above figure indicates the mean loss as compared to total credits earned by the project. 0% Withholding 10% Withholding 20% Withholding Figure AF-9. Ending buffer balance for 0 percent, 10 percent, and 20 percent buffer for a 1,000 ha project over a 20-year project life. Percentage above figure indicates the mean loss as compared to total credits earned by the project. 0% Withholding 10% Withholding 20% Withholding Figure AF-10. Ending buffer balance for 0 percent, 10 percent, and 20 percent buffer for a 20,000 ha project over a 20- year project life. Percentage above figure indicates the mean loss as compared to total credits earned by the project. 71 Commercial Insurance Although risk of loss, and premiums and deductibles by extension, are all different, 1,000 ha projects tend to perform similar to the 20,000 ha examples. Both project lengths generate positive net present values and, as with other approaches, the longer project performs better. Figure AF-11. Project NPV for a 1,000 ha project assuming full value option insurance coverage over both 20- and 40- year project lives. 72