Text Size:   A+ A- A   •   Text Only
Site Image

CATF December Meeting Agenda
Carbon Allocation Task Force
Two World Trade Center
Mezzanine Level, Rooms 2, 3 & 4
121 SW Salmon
Portland, OR
9:00 AM – 1:00 PM
December 8, 2005
9:00     Introductions                                                           David Stewart-Smith, Chair
9:05     Climate Policy in the United Kingdom                       Hal Nelson
                                                                                            Portland State University
9:30     Emissions Accounting                                               Phil Carver
                                                                                            Oregon Department of Energy
10:30   Break
10:45   Baselines and Allowance Distribution                         Sam Sadler and Phil Carver ODOE
12:45   Planning for January Meeting                                     David Stewart-Smith
1:00     Adjourn
To participate by phone, call 877.214.0402 and enter participant code 917220. 
Please note that the sound quality may be inferior, given the size of the meeting room.
Directions to World Trade Center: www.wtcpd.com

Oregon Department of Energy staff prepared the discussion papers and the set of tables. These are drafts to help guide the discussion. They are not final policy recommendations. While staff appreciated the help they received from the members of the Emissions Subcommittee and the Baseline Subcommittee, the drafts are solely the responsibility of staff. There is no implied endorsement of the discussion papers by any members of the Task Force.

Explanation of Emissions Calculations
Explanation of Emissions Calculations [1] (See spreadsheet titled “Rough Base Emissions ”)
December 2, 2005
Phil Carver, Oregon Department of Energy
  Overview The spreadsheet is intended to provide a rough basis for policy discussions related to a load-based cap-and-trade system for the Oregon electric sector.  Emissions estimates (Tables 1 and 5 on page 1) are based on the labeling systems used for Oregon and Washington utilities.  Emissions assigned to load-serving entities (LSEs) may, in the end, be based on different estimation methods.  Transmission and distribution losses also need to be reexamined.
The values for the emissions of Oregon consumer-owned utilities are rough estimates.  The spreadsheet does not include emissions for Idaho Power Company’s Oregon loads, self-generating retail customers and independent retail electric service suppliers (ESSs under SB 1149, 1999 session).  These will have to be added later. 
The base period (2001-2004) was chosen because that is the extent of readily available emissions data.  It is not intended to prejudge the appropriate base period.
Page 2 of the spreadsheet has estimates of the aggregate cost of compliance in 2020 (Table 6), the amount of first year allowances that might be allocated to LSEs under six scenarios (Table 7) and the cost to LSEs of buying enough first year auctioned allowances to cover their base-period emissions under the six scenarios (Table 8).  These examples are only intended to be illustrative of the kinds of impacts of various policies.  These examples are neither forecasts nor recommendation of the assumptions in Tables 6-8.  They are intended to band a wide but not comprehensive range of possible alternatives.  
  The First Page of the Spreadsheet On the first page of the spreadsheet are estimates of average emissions (million metric tons [MMT] of CO2) for the four years of available data.  These are the Oregon emissions for Portland General Electric (PGE), PacifiCorp (PAC) and the 37 consumer-owned utilities (COUs).  The emissions for Idaho Power, ESSs and self-generators are not yet included.  These emissions to be added are likely about 3 to 4 percent of the total.  Emissions for PGE and PAC are based on their retail labels calculated in accordance with the Oregon PUC fuel and emissions disclosure rule (OAR 860-038-300).
Table 1 on the first page has emissions rates in pounds of CO2 per kWh.  The decline in the COU emission rate is due to increased purchases from the Bonneville Power Administration (BPA) and reduced purchases from other suppliers during the base period.  Market purchases of PGE and COUs that are not associated with specific generating plants are assigned the emission rate of the net system mix of the U.S. portion of the Northwest Power Pool.  This value is calculated by the Washington Department of Community Trade and Economic Development using data supplied by the U.S Department of Energy, PGE, PAC and Washington utilities.  BPA’s emission rate is considerable less than the net mix rate as virtually all of BPA’s generating resources are hydro or nuclear.  All of PAC’s load is covered by generation from specific power plants.  PAC’s system emissions are allocated to Oregon based on its multi-state allocation of costs. 
Table 2 is utility sales (million mega-watt hours) from Oregon Public Utility Commission statistics.  
Table 3 is transmission and distribution (T&D) losses as a percentage of electricity sold.  The PGE and PAC values are from the retail label calculations.  The 10 percent loss value for COUs is a placeholder.
Table 4 is electric load (million mega-watt hours) calculated as sales plus losses. 
Table 5 is total emission calculated as load (Table 4) multiplied by the emission rate (Table 1).  The column titled “2001-2004 Average Emiss. Rate and Loss with 1990 Sales” is the average base period emissions rate for each “utility” multiplied by an estimate of 1990 load.  The load estimates are 1990 utility sales multiplied by one plus the average loss rate for the base period.  This gives an estimate of 1990 utility emissions.  Utility specific emissions rates are not available for 1990.  The last column [“10% below ‘1990’, The 2020 Goal (the cap)”] is 90 percent of the estimate of 1990 electricity emissions.  The goal of a 10 percent reduction from 1990 greenhouse gas levels by 2020 is on page ii of the Oregon Strategy for Greenhouse Gas Reductions.   Governor Kulongoski adopted this goal for the state.  It is included in the charter for the Carbon Allocation Task Force.  The utility base-period loads and emissions and the 2020 emission goal are used in Tables 6-8 on the page 2 of the spreadsheet.  These values are in bold typeface on page 1 of the spreadsheet. 
The Second Page of the Spreadsheet  The second page of the spreadsheet has three tables.
Table 6 is intended as a rough example of possible costs of compliance in 2020.  None of the values has been adopted by the Carbon Allocation Task Force or the staff of the task force. 
The first three lines of Table 6 show the total 2020 allowances (the electric sector cap) as the sum of the allowances auctioned and distributed for free.  For simplicity, only the 5 percent value for auctions is shown.  The cap is the 2020 goal of 18.6 MMT of CO2 from Table 5.  The aggregate emission reduction needed (line 4) is the difference between the sum of 2004 utility emissions and the sector cap.  The “LSEs’ Gap” (4.77 MMT-CO2 on line 6) is the 2020 reduction needed (3.84 MMT-CO2, 80 percent of the gap) plus the amount of 2020 auctioned allowances (0.93 MMT-CO2, 20 percent of the gap) that the load-serving entities (LSEs) would buy.  The annual Net Cost Of Compliance (lines 8, 9 and 10) is the sum of the costs for CO2 reductions and auction purchases at three assumed levels of average compliance cost ($ per MT).  The cost estimates implicitly assume that the cost-effective energy efficiency and renewable resources identified by the Fifth Power Plan of the Northwest Power and Conservation Council would be sufficient to meet all load growth. 
If there is a fixed fee for non-compliance ($40 per metric ton in the straw proposal), it will put an absolute limit the price of the auctioned allowances.  The auction price and market price of the allowances should be equal if the market functions well.  The market price of allowances should be above the average cost of compliance.  Under most circumstances, the cost of virtually all CO2 reductions will be less than the market price of allowances.  CO2 reduction costs and purchase costs of auctioned allowances will major components of compliance costs. 
This system will encourage all LSEs to pursue all CO2 reduction measures that cost less than the auction price.  The costs of CO2 reduction measures are the costs of renewable resources and energy efficiency measures in excess of the expected wholesale electricity price.  The difference between the cost of low-cost reduction measures and the auction (market) price would always provide net value to the LSE, either to lower the cost of compliance or the free up allowances to sell into the market. 
If an LSE expects to need more allowances to cover its emissions, it will minimize its costs by pursuing reductions that cost less than the market price.  If a LSE has more allowances than it needs this year, low-cost measures up to the market price will free up more allowances that could sold to yield net revenues or could be banked for future use.  Whether or not a utility expects to have adequate allowances to cover its emissions, CO2 reduction measures that cost less than the market price of allowances provide net value.  To minimize net costs, an LSE would pay for up to the expected market price for CO2 reduction measures but not more.
The bottom three lines of Table 6 (lines 14-16) show the funds raised if the auction allowances are 5, 20 and 50 percent of total allowances.  Only the 5 percent row is consistent with the rest of Table 6.  Also, as discussed above, even for the 5 percent auctioned line, the auction price should be higher than the average cost of compliance.  For example, if the average cost of CO2 reduction measures were $17.50 per metric ton (MT) (not shown) and the price of auctioned allowances were $30 per MT, the weighed average cost of compliance would be $20 per MT, given the weights of 80 and 20 percent for free and auctioned allowances, respectively (as discussed above).  For this simple example, the $30 per MT auction price and the $28 million cost for purchasing auctioned allowances (row 14, highlighted) would be consistent with the $20 per MT average compliance cost rate and $95 million for annual compliance costs (row 9, highlighted), assuming that 5 percent of the 2020 allowances were auctioned.  For this example, the $95 million cost of compliances would be 3.4 percent of total 2004 Oregon electric utility retail revenue. 
Table 7 presents six scenarios of the amount of first year allowances (million metric tons [MMT] of CO2) that might be auctioned with the remainder distributed for free to PacifiCorp, PGE and the COUs.  The first three scenarios assume 5 percent of the allowances are auctioned, consistent with the top part of Table 6.  Scenarios 4, 5 and 6 assume 20 percent of the allowances are auctioned.  Scenarios 1 and 4 assume that 95 percent of the free (non-auctioned) allowances are distributed proportional to base-period emissions and 5 percent are distributed on base-period loads.  Scenarios 2 and 5 assume free allowances are distributed solely on the basis of emissions in the base period.  Scenarios 3 and 6 assume that 50 percent of free allowances are distributed on base-period emissions and 50 percent on base-period loads.
Table 8 provides the costs of if all auctioned allowances are purchased to cover base-period emissions for PacifiCorp, PGE and the COUs for the same six scenarios in Table 7.  Each “utility’s” purchase of auction allowances equals the base-period emissions minus the allowances distributed for free for that scenario.  For simplicity, a single auction price of $10 per MT of CO2 is assumed.  The actual auction price for first year allowances will depend on the expected level of emissions in first compliance year (relative to total allowances), the expected cost of CO2 reductions measures and the expected price for banked allowances in future years. 
The actual amount of first-year auctioned allowance that each utility will purchase will vary based on its first year needs or its desire to bank allowances.  Worst case, a utility that does not achieve sufficient reductions in emission would pay the fixed fee for non-compliance ($40 per MT in the straw proposal).  As the electricity sector cap declines over time, the auction price will likely rise, depending on load growth and the pace of declines in costs of renewable resources and energy efficiency measures.  In no case would it rise above the fixed fee for non-compliance.

[1]               The Oregon Department of Energy prepared this paper to aid the discussion of the Carbon Allocation Task Force in these issues.  While some members of the Task Force reviewed an earlier draft, there is no implied endorsement by the Task Force or any of its members.

Baselines and Allowance Distribution Methods
Carbon Allocation Task ForceBaselines and Allowance Distribution Methods[1]  
A.        Selecting a Baseline
The baseline allocations are the initial number of allowances that the state would distribute for free to load-serving entities (LSEs).  It is a separate concept from the cap that the state uses to determine how many allowances to issue, either for free or at auction.  For a load-based allocation standard, the baseline allocation may be determined by looking at historical emissions or the historical load, or a combination of both, for LSEs.  Data for historical loads and emissions could be for a specific year or for an average of several years.  The baseline allocation could remain fixed, or it could be adjusted over time.  Fixed free allocations based on historical data is also known as “grandfathering.”
Year Specific or Average of Years The Oregon Department of Energy has data that could be used for setting the historic baseline for LSEs beginning in 2001[2].  Historical data are available by mid-year the following year, e.g. 2005 data will be available around July 2006.  At the earliest, the first year of the allocation standard might be instituted would be 2009.  Therefore, given data that would be available before 2009, a base year period could be chosen from any year(s) between 2001 and 2007. 
For example, a single year might be the most recent year for which we shall have data.  In the example above for beginning in 2009, that would be 2007.  Or, to allow for advance planning, the base year might be 2006.  However, if such a year were a drought year or a year of high precipitation, it might be unique.  Likewise, if there were unique economic conditions that affected load, those might be considered, but it would be difficult to choose criteria that did not advantage or disadvantage a particular LSE.  That would make the choice of a particular year a highly political exercise.
Because the Northwest system is so hydro-dependent, there is a strong argument for using an average of several years for the historical baseline if the allocation is based on emissions.  Also, averaging is a way to adjust for minor changes in loads from economic conditions.  For example, one could use a data set of certain years of data.  Alternatively, the whole period of 2001-to-2007 could be used for 2009 allocations.
Choosing a method for setting the baseline will also affect the slope of the line for a decline in the emission cap to meet the 2020 and 2050 emissions targets.  For example, setting a baseline and cap to meet expected first year emissions, including load growth, would mean that the rate of the decline to reach the target would be steeper.  Likewise, holding the cap flat for several years initially might give LSEs more time to adjust to the cap, but that would also mean the rate of decline in the cap to meet a fixed target would have to be steeper.  On the other hand, the sooner the cap starts declining the lower the slope of the line to the final target.
Decision:         How will the baseline be chosen, e.g. a specific year (which year?) or an average of years (over what period?)?
Although there are different criteria and types of data that one could use to set a baseline for an LSE, there are two types of data that would be most appropriate for setting the historic baseline for a load-based allocation standard: metric tons (or tonnes) of CO2 emissions, and load, expressed as megawatt hours (MWh).
Emissions.  Setting an emissions baseline requires determining the total CO2 emissions from the resources that an LSE used to serve its customers during a specific period, including emissions associated with transmission and distribution losses.  Setting the baseline on emissions alone provides more free allowances to the LSE that has the most carbon-intense mix, i.e. it provides more allowances to the LSE whose share of total emissions is higher than its share of total load.  Concurrently, it allocates fewer allowances to those LSEs with relatively low emissions during the baseline period. 
Megawatt Hours.  Setting a baseline on megawatt hours of load would allocate allowances based on total megawatt hours that an LSE supplied during the baseline period.  Setting the baseline on megawatt hours rewards past actions to reduce carbon emissions and those LSEs with fewer or no emissions who contributed less to the increase in greenhouse gas concentrations in the atmosphere.  It also provides growth room for those LSEs with few historical emissions. 
A combination of methods could balance the benefits of each.  Table A illustrates how applying the two methods could affect allocations.  It also includes consideration of whether allowances are auctioned, which is discussed later in this paper. The table begins with the assumption that the state would issue a total of 23.94 million allowances during the first year.  That is a representative number based on an estimate of the average annual CO2 emissions from PacifiCorp, PGE, and COUs during 2001 through 2004.
All data in the table are illustrative.  The baseline might be based on different years; the methods of allocation may vary; and, not all potential LSEs are included.  Idaho Power will be included in later tables.  Self-generators have not yet been defined.  They will be included later also.  Nevertheless, the data provide a good sense of the impacts of a carbon allocation standard, given certain assumptions.
Table A shows how many allowances might be issued and how they might be distributed initially.  Part 1 of the table has six scenarios: one of two auction options apply, based on the assumption that there will be a modest auction of some allowances: Scenarios #1 - #3 show 5 percent of all allowances auctioned, and scenarios #4 ‑ #6 show 20 percent of all allowances auctioned.  Given those two options for auctions, the table then shows the distribution of the remaining free allowances under three conditions: 95 percent of free allowances distributed based on emissions and 5 percent based on MWh; 100 percent of free allowances distributed based on emissions; and 50 percent of free allowances distributed based on emissions and 50 percent based on MWh (a 50/50 distribution).
Part 1 of the table gives a reference of the baseline assumptions.  It assumes that the first year allowances equal the first year emissions, although it is unlikely that that would be the case precisely.  The differences between the emissions in Part 2 and the “distributed free allowances” in Part 1 give a sense of scale of how much PacifiCorp, PGE and the COUs might have to reduce their emissions before they had to purchase auctioned allowances.  (The funds that might be raised by an auction are shown in Tables B and C at the end of this paper.)
Table A.  Six Scenarios for Issuing Allowances the First Year

1.  Scenarios 
Total Allowances Issued (MMT CO2)
% of Total Allowances Auctioned
Allowances Auctioned (MMT CO2)
Allowances Distributed for Free (MMT CO2)
% of Free Allowances Distrib. by Emissions
% of Free Allowances Distrib. by Load
Free Allowances First Year Details 
Distributed Free to PacifiCorp
Distributed Free to PGE
Distributed Free to COUs
Total Free Allowances
2.  Assumed Base Period Emissions (First Year) 
The results show that PGE would be mostly indifferent to whether 95 or 100 percent of the free allowances were distributed based on emissions.  Even a 50/50 distribution has only a relatively small effect on PGE. 
The method of setting the baseline affects the free distribution among PacifiCorp and the consumer-owned utilities significantly.  The 50/50 distribution has the most relative impact.  Although not shown in the table, a 100 percent distribution by MWh would have the effect of decreasing PacifiCorp allowances and increasing COU allowances by another 2 to 3 MMT CO2 beyond the 50/50 distribution.  (Idaho Power, electricity service suppliers and self-generators are not included in the table, and at this time the Department does not have a breakdown for each consumer-owned utility.)
Decisions:       What data will be the basis for the initial baseline, e. g. historical emissions, historical MWh, or a combination of both? 
If there were a combination, what percentage would be assigned to emissions and what to MWh?
Adjustments[3] The baseline for an LSE could remain fixed; it could be adjusted periodically; or, it could be adjusted according to criteria for unique events.  Unique events could be extraordinary load changes or shifts in an LSE provider.  Adjustments could also be made when an LSE closes a plant that had self-generation.  Because the term “updating” has particular connotations related to a cap-and-trade system for individual generating units that are not necessarily applicable to a load-based system, this discussion uses the concept of “adjustments” to distinguish changes in a baseline within load-based system.  Updating the allocations based on changes in load or emissions on a pre-determined schedule would reward those entities that performed the worst in meeting the objectives of the allocation standard.
Because this is a load-based system, it will require each LSE to accommodate changes in load from normal economic activity.  If there is growth in customers or use per customer, an LSE will have to meet CO2 requirements with efficiency improvements or a change in the carbon intensity of its generation mix.  While adjustments could account for unique circumstances, LSEs could also buy allowances from the auction or other LSEs to help accommodate for its load growth. 
Whatever adjustments there are to the baseline of a particular LSE, the total emissions for the state would remain capped.  Total free and auctioned allowances would decline on a predetermined schedule, perhaps after being fixed for an initial period, in order to meet 2020 and 2050 goals.  Therefore, any adjustments would be a zero-sum game.  For example, if allowances were held in reserve for large new loads or for new LSEs, they would have to come from within the cap.  That would mean that fewer allowances would be distributed to other LSEs.  Adjustments would also include transfers of allowances when loads shift from one LSE to another in specific circumstances.  If adjustments result in some allowances being unallocated, such as with the closure of a self-generator, they could either be allocated to other LSEs or they could be retired.  These cases are discussed below.
Adjusting the baseline is different from the issue of the rate of decline in the cap, which is discussed elsewhere.  The rate of decline will be applied to each LSE’s allocations. 
Decision:         Would the baseline remain fixed, would it be adjusted, or would it be updated?  
Potential Adjustment Criteria If the baseline were adjusted, several criteria might apply:
LSE Closing: The situation of an LSE closing would occur only for self-generators or Electricity Service Suppliers (ESS).  A utility would not close per se because another LSE would serve the customers.  Any movement of customers to another utility, including the sale of a utility, would be dealt with as a transfer, as discussed below.  
If an LSE that was a self-generator went out of business, as opposed to closing its generator plant and putting its load onto another LSE, it would lose future allocations. The state issues allowances to serve the electrical loads in the state; therefore, there would be no grounds for continuing to give allowances to an entity that was no longer an LSE.  The free allowances would not continue to be issued to a business that had no generation and load at the site that had previously received the allocation.  Likewise, an ESS that served only a single customer that ceased operation would also not receive continued allocations.
If an LSE closed and its allowances were no longer allocated, there would need to be a policy for what happens to those allowances.  They could be retired permanently, they could be added to a reserve for new large single loads, or they could be proportionally allocated to the other LSEs.  Retiring the allocations permanently would be most consistent with the intent of the allocation standard. 
It will be necessary to determine what constitutes “closure.”  For example, there could be a minimum level of generation or there could be a grace period, e.g. one year, after closure before the allowances cut off.
Decisions:       If an LSE closes, would the allowances it would have received in the future be retired, added to a reserve for new large single loads, or allocated to other LSEs proportionally?
                        What defines closure?
Transfers of Load Between LSEs: The provisions dealing with load shifts attempt to resolve an equity problem and an efficiency problem.  The equity problem is that under a load-based cap and trade system, allowances are associated with the loads.  If the loads are transferred to a new LSE, emissions from resources to serve those loads are the responsibility of the new LSE.  Without a provision for transferring allowances, the allowances to cover these emissions would remain with the old LSE.  The efficiency problem is that a potential self-generator or ESS would be discouraged from serving a load if it had to buy all its allowances at auction.   If a customer’s existing load is transferred to a different LSE or the customer begins to self-generate to provide for its own load, the allowances associated with that load would move with the load.  The LSE transferee would receive allowances equal to its share of the former LSE’s allowances associated with the load.   
There can be three instances where a load transfers from one LSE to another:
  1. Load is shifted between a self-generator and the local utility;
  2. Load is shifted between PacifiCorp or Portland General Electric (PGE) and a retail Electricity Service Supplier (ESS) under ORS 757.600(16)); and,
  3. Load is shifted through a change in the local utility serving a territory.
These should all be treated consistently.  If less than a complete load were transferred, proportional allowances would transfer at the most recent allowance rate (metric tons of CO2 per MWh) of the entity losing load, relative to the base period load. The allowance transfer is for tonnes of CO2, whatever formula is used to determine proportionality.
It is necessary to define load reductions relative to the fixed load (presumably the base period load) because the allowance rates of the gaining and losing LSE will differ.  If a load is shifted to another LSE and then shifted back to the original LSE, the proportional allowances would transfer back to the LSE at its original allowance rate until its load was equal to the base period load.  Defining load loss or gain relative to the previous year’s load for an LSE would mean that load shifted back and forth between two LSE’s would ratchet the allowances of the two LSEs.  The LSE with the lower allowance rate would see a net gain of allowances and the LSE with the higher rate would lose allowances by virtue of transferring if there were not a provision to prevent ratcheting. 
This concern over ratcheting is most relevant if a self-generator serving its own load decides instead to sell the generation to the wholesale market and then asks the local utility to serve its load.  This could go back and forth under existing law.  This kind of back and forth could also occur between an ESS and either PGE or PacifiCorp. 
Decision:         Would allowances transfer when loads move between LSEs?
Reserve, or Set-Aside, for New Large Single Load: It may be appropriate to adjust an LSE’s allocation to account for a new large electricity load at a single site other than by transfer.  This would allow for major new industries to move into an area from out of state without placing the local LSE at a disadvantage.  However, it is not intended as a mechanism to change the baseline to account for normal growth.  On the other hand, flexibility mechanisms built into the standard, such as auctioning, trading and banking of allowances, and use of offsets, might be considered sufficient to meet any load changes.  Furthermore, it might be considered inequitable to adjust allocations for adding large single loads without adjusting for losing large single loads.  The standard is a zero-sum game, so any adjustments affect all other LSEs. 
There is a precedent under the Regional Act for defining a new large single load as least 10 average megawatts added in a single calendar year.  However, that would be relatively small for PacifiCorp and PGE and exceptionally large for a small COU.  Another approach would be to set a threshold at a percentage of an LSE’s load at a magnitude that would be truly exceptional, such as 20 percent.  The higher the threshold, the less likely the adjustment will be used. 
The allocation could be met by a withdrawal of allowances from the total allocations for the next year or there could be a set-aside to cover the need for such allowances.  In either case, the newly allocated allowances would also come from within the cap of freely distributed allowances and would reduce the free allocations to all other LSEs.  Also, reserves would have to be replenished as allowances are allocated to new large single loads.
There would have to be method for determining what would be an equitable allocation of allowances to a new large single load.  The allowance rate could be based on a best available technology standard, the statewide allowance rate, the allowance rate of the local LSE, or some other formula.  There might be consideration for a limit on the total amount free allowances that any new large single load could receive.
While “new entrants” are a concern for generation-based cap-and-trade systems, the transfer mechanism would account for new self-generators in an LSE’s territory.  Transfer would also account for new utilities serving a territory.
Decisions:       Should there be an adjustment to allocations to account for new large single loads or loss of large single loads?
What would constitute a new large single load and would it vary by size of LSE?
Should an allowance reserve be established for new large single loads or should allowances for new large single loads come from the total pool of allowances as needed? 
  • How large would a reserve be? 
  • How would allowances be distributed to meet a new large single load?
  • Would there be an annual limit of allowances for new large single loads?
B.        Distribution of Allowances (Free and Auctioned)
The state would issue serialized allowances, which means that each allowance would have a unique identification number and year.  One allowance would represent one tonne of CO2.  The policy issue is how allowances would be distributed once a baseline was set. 
There are two options for distributing allowances: distributing free allowances to LSEs (direct state allocations) and distributing allowances through an auction that the state would conduct at least once a year.  The distribution methods could be combined.  If allowances were auctioned, there would be a consequent issue of what would happen to the proceeds from the auction.  If allowances were distributed for free, it would be necessary to determine the basis for distribution.  For a load-based system, the baseline discussed above would determine the system for free distribution.  The baseline could incorporate historical emissions, historical load, or adjustments to either or both based on criteria discussed above.
Free Allowances: While the authors know of no reports that have studied the value of a free allowances compared to the cost of compliance within a load based system, there are conflicting reports on the economic effects of a fully free distribution of allowances to generators versus auctions in a generator-based cap and trade program[4] and [5].  However, some papers suggest that free distribution over-compensates the generators many fold for their cost of compliance[6] and [7].  Nevertheless, it is also broadly assumed that a free distribution of a substantial proportion of allowances is probably a necessary political concession, at least initially. 
Auction of Allowances: Studies of generator-based cap-and-trade systems that look at auctioning allowances versus a free distribution show that an auction with appropriate redistribution of the proceeds in order to lower specific types of taxes and address compliance costs are either neutral or the most economically efficient mechanism to distribute allowances[8] and [9].  However, it may be unlikely that funds would be redistributed solely based on economic principles.  If for political reasons the state distributes a large portion of allowances for free, a modest auction might be necessary for the overall system to function.  Separate from the macro-economic benefits of a pure auction, there are compelling reasons for at least a limited auction of allowances. 
An important reason for an auction is that an auction will enable a trading system to function.  With only two major LSEs in Oregon, with banking, and with a declining cap, it is likely that those LSEs would hold onto allowances they did not need in one year so they could use them in the future.  Likewise, smaller LSEs could hold onto allowances as hedges against future growth.  Without an auction, it is unlikely there would be any significant amount of trading.  Therefore, there may not be a market price to disclose the costs of compliance with the cap-and-trade system.  This would circumvent the proposed circuit breaker, which is based a high threshold of dollars per tonne of compliance costs. A second reason to auction some allowances is that a viable market would provide more flexibility for new LSEs to form, even if there were a reserve for new entrants.  A third reason is that without a market, there would be less flexibility for an LSE to purchase allowances to handle load growth. 
It will be necessary to determine what percentage of auction of allowances would be enough to establish a viable market.  The auctioned percent of allowances could increase over time as the historical basis for the free allocation of allowances becomes less relevant.  And, over time, one would expect more new entrants.  In effect, an auction provides a method to adjust the allowance distribution among LSEs through market mechanisms rather than having to rely on specific criteria, such as those discussed above.
Table B estimates the first year cost for auctioned allowances for the six scenarios described above.  The table assumes that the first year emissions equal the base year for each LSE.  The amount purchased or sold is the base-period emissions minus the allowances distributed for free.  Negative numbers in the table indicate that in some scenarios the COUS would have more free allowances than they would need for that year.  The table assumes that COUs would sell their extra allowances to PacifiCorp and PGE. 
The table assumes an auction price of $10 per tonne of CO2.  The actual auction price for first year allowances will depend on the expected level of emissions in the first compliance year relative to the total allowances, the expected cost of CO2 reductions measures, and the expected price for banked allowances in future years.  As the electricity sector cap declines over time, the auction price will likely rise, depending on load growth and the pace of declines in costs of renewable resources and energy efficiency measures. 
Table B. 
First Year Cost of Auctioned Allowances for Six Scenarios at $10 per Tonne*
PacifiCorp (million $)
PGE (million $)
COUs (million $)
Total Revenue (million $)
*      Assumes first year emissions are equal to the base period.
Looking ahead, Table C illustrates the funds that might be raised by an auction of 5, 20 and 50 percent of allowances in 2020 at prices that vary from $10 per tonne to $40 per tonne.  It assumes that the cap for 2020 is 18.6 MMT CO2, which is 10 percent below 1990 levels.  It then computes the free allowances and the auctioned allowances within that cap.  For example, if 5 percent of the 18.6 million allowances (i.e., 0.9 million allowances) were auctioned at $10 per ton, that would raise $9 million.  The remaining 17.6 million allowances would be distributed for free.  Auctioning 50 percent of the allowances would raise $93 million.  If the funds from the auction went back into energy efficiency and renewable resources, they could help reduce emissions from the LSEs and could help them stay within their caps.  That interaction is not shown. 
Table C.  Funds That Might Be Raised By An Auction In 2020  
Auction price per tonne
Auction of 5 % of allowances (million $)
Auction of 20 % of allowances (million $)
Auction of 50 % of allowances (million $)
The mid- and upper-range values for a 5 percent auction and the lower and mid-range values for a 20 percent auction are comparable to the current public purpose charges on PacifiCorp and PGE customers under SB 1149.
With an auction, there would be funds created, as illustrated in Tables B and C.  Distribution of those funds will be a key issue.  With a limited auction, it might be appropriate to designate the funds to programs that helped reduce carbon emissions.  One option would be for the proceeds to go to the Energy Trust of Oregon or a similar organization to support energy efficiency programs and renewable resources that have long-lead times or market barriers, are not commercially ready for mass deployment or are not covered by other programs.  Another option, especially for larger auctions, would be for the Legislature to apportion the revenues.
An auction could occur once a year or more often.  Two auctions a year would allow early purchasers to hedge their emissions, while the later auction could account for anticipated shortfalls.  The details of how the auction would be conducted are an appropriate topic for later discussion.
Decisions:       Should allowances be allocated for free, should they be auctioned, or should there be a combination of free distribution and auctions? 
If there were a combination of methods, what percent should be free and what percent should be auctioned? 
Should the percents change over time, and if yes, at what schedule?
If allowances were auctioned, how should the revenues be used, e.g. should they go to designated organizations or programs or should the revenues be returned to the state for the Legislature to decide?

[1]       The Oregon Department of Energy prepared this paper to aid the discussion of the Carbon Allocation Task Force in these issues.  While some members of the Task Force reviewed an earlier draft, there is no implied endorsement by the Task Force or any of its members.
[2]       See “Explanation of Emissions Calculations,” prepared for the Carbon Allocation Task Force by the Oregon Department of Energy, December 2005.
[3]       This paper is written before the Carbon Allocation Task Force has discussed in detail how self-generators would be included in the carbon allocation standard.  Later decisions about the scope of self-generators to be included might change some of the assumptions discussed in this paper.
[4]       See, for example, “Allocation of CO2 Emissions Allowances in the Regional Greenhouse Gas Cap and Trade Program,” Dallas Butraw, Karen Palmer, and Danny Kahn, Resources for the Future, Washington, D.C., June 2005.
[5]       See, for example,  “Carbon Trading Program Emission Allowances:  Practical Considerations for Allocation,” Paul J. Hibbard, Analysis Group, May 2005.
[6]       loc. cit., Butraw.
[7]       See, for example, “Allowance Allocation:  Who Wins and Loses Under a Carbon Dioxide Control Program?,” Anne E. Smith and Martin T. Ross, Charles River Associates, February 2002.
[8]       loc. cit., Butraw.
[9]       loc. cit., Hibbard.