6. Modification of Interim Avoided Cost Methodology

By far the most controversial issue in the 2006 Update phase of this proceeding revolved around modifying the interim avoided cost methodology adopted in D.05-04-024. To better understand the proposals presented by the parties, a brief summary of that methodology is presented below.

6.1. Interim Avoided Cost Methodology

The underlying theory of the interim avoided cost methodology is that long-run marginal costs (LRMC) establish proper price signals in the market to elicit the most efficient investment of new capital. The methodology uses the all-in costs of a combined cycle gas turbine (CCGT) as a proxy for this long-run price signal based on evidence from the CEC, the Western Electricity Coordinating Council and the Energy Information Association that the majority of new resources being added in the Western Interconnect are gas-fired combined cycle generators.41

In other words, the costs of a CCGT are assumed to approximate an electricity supply curve at long-run equilibrium, based on certain assumptions of free market entry and exit. Free entry means that market prices above the fully allocated cost of a CCGT cannot persist, because such high market prices would lead to the construction of new CCGTs that would tend to drive the price down. Free exit means that market prices below the fully-allocated cost of a CCGT cannot persist, because such low prices mean that existing units will be unable to earn enough margin to cover fixed costs and will exit the market (or alternatively, the construction of new resources will be delayed until growth in demand has consumed any temporary capacity surplus), driving the price back up. In the long run, therefore, these free market entry and exit assumptions dictate that the market price can neither be higher nor lower than, and must therefore be equal to, the fully-allocated cost of a CCGT.

Based on these assumptions, the interim avoided cost methodology proceeds to calculate avoided costs as follows:

6.2. Workshop Consensus and Non-consensus

The workshop discussion explored the concept of including a capacity adder in the peak hours based on the cost of a combustion turbine (CT) plant. In particular, this discussion centered around the questions of (1) whether the resource balance condition by necessity should allow entry of both a CT and CCGT and/or (2) whether the LRMC price shape should be modified to contain sufficient margin to provide recovery of the capital investment in a new CT facility. There was also discussion of how the utility proposals for rate design and evaluation of demand response incorporate CT costs in the valuation of demand reductions.

No consensus was reached at the workshop on the issue of whether to modify the current avoided cost methodology. Some parties indicated support for modifications that would add value to current avoided costs to reflect the cost of a CT (or some portion thereof), at least for certain peak hours, but agreement on the specifics of that methodological change was not reached. At the end of the workshop discussion on this topic, TURN presented a settlement proposal. TURN recommended increasing the present value of the generation avoided costs for residential and commercial a/c programs by 10% and 5%, respectively.

6.3. Positions of the Parties

In its comments, PG&E outlines its view of how short-run and long-run avoided costs should be developed, particularly in the future "when the Commission determines that there is an effective forward market for future capacity."43 In PG&E's view, when that time arrives, the forward market prices will be a more accurate measure of future avoided capacity costs than the costs associated with any particular generation resource (e.g., a CCGT or CT). More generally, PG&E argues that the best measure for determining how much capacity an energy efficiency program avoids, and the avoided cost of that capacity, would be the amount of additional capacity the resource adequacy rules require each utility to have for each time period, and the market prices that utilities pay for that capacity.

In the meantime, PG&E argues that the interim avoided cost methodology is flawed in several ways, and should be modified in this 2006 Update. For avoided cost prices once load-resource balance is reached (2008 and beyond), PG&E submits that (1) the new capacity needed to meet increases in load in those hours in which demand is highest is much more likely to be a new CT, rather than a CCGT, and (2) the new capacity needed to meet additional demand in other hours is likely to be a new CCGT. Accordingly, PG&E recommends that the interim methodology be modified so that a new CT could recover all of its annual fixed costs from selling energy in those hours in which the market prices exceeded its variable costs. In PG&E's view, this requires:

· Increasing prices in the higher price hours of the scaled up PX price shape, and

· Lowering prices in the lower price hours of that price shape by just enough to enable both the CT and the CCGT to recover amounts equal to but no greater than all of their respective fixed costs, including recovery of and return on investment.

· Making such modifications in such a way that the total area under the price shape remains unchanged, i.e., the net capacity cost of the new CT that is allocated to certain hours must be subtracted from the remaining hours.

PG&E presents three options for determining the hours in which the adjustments (increases) to avoided costs should be made, and how to allocate the appropriate amount to each hour and then subtract corresponding amounts from all the remaining hours. In addition, PG&E presents recommendations on how to modify the interim avoided cost methodology prior to resource balance, i.e., for 2006 and 2007. PG&E also outlines further methodological changes to avoided costs it recommends for Phase 3 of this proceeding.

In its opening comments, TURN also recommends modifications to the interim avoided cost methododolgy that involve adding costs related to a CT to high demand hours. TURN argues that these modifications are necessary to put low-load factor air conditioning measures on consistent footing with the valuation approaches being applied to both rate design and demand-response resources. In particular, TURN recommends that all utilities be required to add to the top 100 hours the CT capital costs, calculated using a real economic carrying charge rate, minus the energy savings created by the CT. TURN estimates that this approach would result in a residual capacity value in the vicinity of $20-$35 per kW-yr.

In DRA's view, the operational characteristics of dispatchable demand response programs are fundamentally different from those of energy efficiency and distributed generation programs, in the same way that peaking plants have different operational characteristics and economics from shoulder/baseload plants. Because of these differences, DRA argues that it is unreasonable to value demand response and energy efficiency programs using the same avoided costs even when both programs contribute to load reduction during the same hour. DRA supports valuing dispatchable programs targeted to reduce load during the critical peak period hours, such as demand response, based on the costs of a CT. However, DRA argues that non-dispatachable resources such as energy efficiency and distributed generation are more accurately valued using the current avoided cost methodology, which allows but does not ensure recovery of the fixed costs of a CT under all circumstances.

SDG&E and SoCalGas are opposed to changing the interim avoided cost methodology in this phase of the proceeding, but support a capacity adder of 10 percent for residential a/c and a 5 percent adder for commercial cooling, as suggested by TURN in the workshop. In their view, this is a reasonable settlement position in light of the lack of consensus on a method for calculating a precise CT-based adder, or on how to spread those capacity costs to hours in the year. In its reply comments, TURN supports this position.

SCE does not take a position on whether it is necessary at this time to include a CT-based adder to avoided costs, prior to full discussion of this matter in Phase 3 of this proceeding.

6.4. Final Report Recommendations

The 2006 Update consultants state that the threshold question is whether the resource balance condition requires entry of both CT and CCGT units. They believe that it is premature to conclude at this time that it does. In their view, this is a complex theoretical question that will be the subject of research efforts and future proceedings. They conclude that major revisions to the interim methodology should await the results of those proceedings.

The Final Report also presents a comparison between current avoided generation costs and the avoided costs derived from a CT-based adder. This adder was originally developed by E3 for the CEC Title-24 building standards investigation into the valuation of demand response measures. It was developed from the perspective looking at a CT as a "back stop" technology, that is, the technology that utilities would add for additional capacity if the market was not building enough to meet critical peak demand.

Using updated New York Mercantile Exchange (NYMEX) and natural gas fundamentals price forecasts, the Final Report presents calculations of a CT-based adder in the range of $28 to $34 per kW-yr range. Using spot gas prices, rather than an annual average value, increases the adder by approximately $10-yr.

6.5. Discussion

The debate in this phase of the proceeding over avoided cost methodology is not a new one. In Phase 1 of this proceeding, some parties argued that the use of a CCGT proxy for long-run avoided costs would misstate those costs for high-usage periods when CTs would be operating as the marginal units, and presented similar recommendations to modify the top end of the price shape to contain the explicit cost of a CT.44 We rejected these arguments, stating that:

"E3 is not providing a cost shape that assumes that CCGTs are the marginal plant for all 8,760 hours in the year. Rather, the CCGT is used to set the average annual market price. When this average price is applied to the hourly market shape, the result is that some hours will have costs higher than the CCGT annual average cost (when CTs would be on the margin) and some hours would have lower prices (when other baseload units would be on the margin)."45

It is clear that some parties are still dissatisfied with the methodogical basis for our interim avoided costs with respect to the evaluation of energy efficiency programs, demand-response programs, or both. However, we agree with the conclusions of the Final Report that the CT-adder approach for modifying avoided costs during peak hours raises theoretical issues concerning LRMC that are more appropriately addressed in Phase 3.

In particular, PG&E's proposed methodology for modifying the current avoided costs is based on the assumption that "there are two marginal capacity resources: a new CT in hours when prices are relatively high, and a new CCGT to meet baseload demand."46 However, as indicated above, the current methodology is not based on assumptions of what type of plant operates at the margin. Nor is it based on the theory that avoided costs must be constructed to always provide sufficient margin for CT owners to operate their plants in peak demand periods, or to reflect a backstop technology that the utilities might have to build during periods of short-term peak capacity shortages.

Rather, as discussed in Section 7.1 above, the current methodology is based on the approximation of an electricity supply curve at long-run market equilibrium. We believe that the adopted approach represents a fundamentally different approach to avoided costs than the CT-adder based methodologies underlying PG&E's and TURN's recommended modifications. It is beyond the scope of this 2006 Update to explore these theoretical differences sufficiently in order to carefully consider the proposals before us for modifying avoided costs, whether for energy efficiency, demand-response, or both. This type of exploration is more appropriate for Phase 3 of this proceeding.

Moreover, even if we agreed with the theories underlying a CT-adder, there would be numerous approaches and assumptions to consider and resolve before one could be adopted. There is no consensus on these matters, and we lack a sufficient record in this phase of the proceeding for resolving them.

For example, in concept the adder is the capital cost of the new CT minus the "margin" or profit that operating the CT would make its owners in the market. There is more than one approach for calculating this margin. We note that PG&E recommends one approach and TURN uses another. The record in this phase of the proceeding also indicates that the level of the adder would be affected by a number of assumptions, including heat rates, CT capital costs, the capital carrying charge rate (real or nominal), and the assumed years of the CT life.

There are also various approaches to consider for determining the hours in which the resulting increases to avoided costs should be made, and how to allocate the appropriate amount to each hour. In addition, one must decide whether to subtract the increased value in the peak hours from other hours so that the average avoided cost is no greater than a CCGT and if so, how to allocate those reductions to specific hours.

Finally, the comments of DRA and TURN raise the issue of whether it would be reasonable to apply the resulting adder to both energy efficiency programs and demand-response programs when they reduce demand in the same hour. These parties present arguments on both sides that need to be explored based on a more extensive record.

In sum, modifying current avoided costs using a CT-adder approach requires the resolution of complex theoretical issues, assumptions and methodological issues that are beyond the scope of this 2006 Update. As stated at the outset, this phase of the proceeding is not the forum for proposals that fundamentally reject or represent a major change to the interim method. Phase 3 of this rulemaking has been clearly designated as the forum for such proposals.

As discussed above, some parties have recommended a simple capacity adder of 10 percent for residential a/c and 5 percent for commercial cooling as an alternative to explicitly adopting a party's specific proposal to alter the hourly price profile with a CT-adder. However, this approach still assumes that the current hourly price profile fails to value avoided costs properly for low load-factor energy efficiency measures during peak hours. Until we examine further the underlying theoretical and methodological issues discussed above, we are not prepared at this juncture to accept this assumption. Moreover, as pointed out in the Final Report, the current price profiles may already reasonably capture the higher value of programs that save energy during peak demand periods, since they were developed based on market prices that were highly volatile during the 1998-2000 period.47

Finally, we disagree with TURN's contentions that such modifications are necessary to put low-load factor air conditioning measures on consistent footing with the valuation approaches being applied to rate design or the evaluation of other resource options. During the workshop, the assigned ALJ requested that TURN and the utilities identify the Commission proceedings in which a valuation approach utilizing a CT-adder approach was currently being applied, and to summarize the general method. Our review of these submittals indicates that a Commission decision has been rendered in only a few these proceedings. These proceedings address rate design, revenue allocation or demand response funding proposals in which the Commission has adopted settlements that resolve the issues without addressing the reasonableness of any party's proposed valuation methodology. In these instances the settling parties agreed that no specific assessment of a marginal cost or avoided cost methodology was required to resolve the issues in these proceedings, and the Commission found that the settlements were reasonable without resolving them.48

However, the Commission has recently considered methodological issues very similar to those raised by parties to this proceeding in the context of establishing the "market-price referent" (MPR) for the Renewals Portfolio Standard (RPS) program. One purpose of the MPR is to establish the market price at or below which the costs of long-term contracts entered into by the utilities with eligible renewable energy resources will be deemed reasonable and authorized in rates. To establish this market price, the Commission has developed a proxy plant to model the long-term costs associated with fixed-price electricity from new generating facilities. Under the adopted MPR methodology, time-of-delivery (TOD) factors are applied to the annual cost of the proxy plant in order to create a price shape that takes into account the value of different products.49

As SDG&E points out in its March 9, 2006 comments, the Commission's recent decision on a methodology for calculating the 2005 MPR would argue against making adjustments to the PX profile to allow the price shape to return the capital cost of a CT. We note that in the RPS proceeding PG&E and several other parties specifically recommended against the use of a CT proxy for calculating the MPR. Rather, they argued that the MPR for peak period energy should be established by applying TOD factors exclusively to a CCGT. The Commission adopted this approach, agreeing with PG&E that "the application of TOD factors to the baseload MPR took into account the value of different products, including baseload, peaking and as-available output."50

For all the reasons discussed above, we do not modify the interim avoided costs methodology in this 2006 Update. The methodological issues raised in this phase of the proceeding may, however, be appropriate topics to explore further during Phase 3.

41 Methodology and Forecast of Long-Term Avoided Costs for the Evaluation of California Energy Efficiency Programs, Prepared for the California Public Utilities Commission Energy Division, by Energy and Environmental Economics, Inc., October 25, 2004, pp. 47-48, 54-55.

42 A forward contract obligates the seller to sell and the buyer to buy at a specific price for a specific quantity delivered to a specific location, i.e., the energy deliveries are firm. As explained in D.05-04-024, forward price data is considered to reflect market prices, including capacity. As part of forward price determination, the market assigns a value to the capacity used to ensure firm delivery of the contracted energy. The value is small (large) to reflect the expected surplus (shortage) in the capacity used for firm delivery. This value does not necessarily track the historic fixed cost of capacity and, in the years prior to resource balance, forward prices do not cover the full cost of a new entrant. See D.05-04-024, pp. 31-32.

43 PG&E Comments, March 27, 2006, p. 9.

44 See D.05-04-024, mimeo., pp. 30-31.

45 Ibid., p. 32.

46 PG&E Comments, March 31, 2006, p. 7.

47 See Final Report, p. 18.

48 See D.05-11-005 in A.04-06-024 (PG&E's Rate Design Window), pp. 5, 20; D.05-12-003 in A.05-02-019 (SDG&E Rate Design Window) pp. 15-16; D.06-03-024 in A.05-06-008, A.05-06-006 and A.05-06-017 ( Demand-Response Program Plans and Funding for PG&E, SDG&E and SCE), p. 12 and Appendix A.

49 As discussed in the Final Report, SCE initially suggested that the TOD factors used for RPS replace the current PX hourly shapes in the interim avoided cost methodology. No parties currently suggest this modification to the interim methodology for the reasons discussed in the Final Report (pp. 23-24). However, the issue of consistency in the price shapes we use to evaluate various resource options may be an appropriate topic for Phase 3.

50 D.05-12-042 (as corrected by D.06-01-029), mimeo., p. 33.

Previous PageTop Of PageNext PageGo To First Page