D0209053 Interim Opinion on Procurement Issues: DWR Contract Allocation

5. Adopted Contract Allocation

Despite the different allocation principles and comparison metrics advocated by the parties, the preferred contract allocations presented by PG&E and (jointly) by SCE and SDG&E are identical in more respects than they are different. First, neither proposal allocates power to a utility from a contract of its affiliate. Second, the proposals reflect consensus that separately identifiable products in a contract can be split, but that dispatchable products should not be split. Third, the utilities' preferred allocations generally put all contracts with NP-15 specified delivery points with PG&E, and all those with SP-15 specified delivery points with SDG&E.³⁵ In fact, with the exception of only two of the contracts (Coral and Sunrise), the utilities' preferred contract allocations are identical. (See Attachment 2.)

Under the SCE/SDG&E joint proposal, both the Coral and Sunrise contracts are allocated to PG&E. Under PG&E's proposal, Sunrise is allocated to southern California (PG&E does not specify any allocations between SCE and SDG&E), and Coral is split between PG&E and southern California. PG&E proposes a split whereby 25% of the base and additional quantities are allocated to PG&E, with the remainder going to the south.

Each utility urges us to resolve this remaining allocation issue by selecting the allocation principles and comparison metric that it prefers to the exclusion of others. In doing so, we would need to select from among the following a single yardstick to use in measuring the results of each allocation option: (1) a current base period of net short, (2) a projection of net short into the future including estimates of direct access migration, or (3) a projection of net short excluding direct access migration. Attachment 3 compares the net-short yardsticks presented in this proceeding.

Once the yardstick is selected, the utilities would then have us choose the single most appropriate metric with which to measure the results of each allocation option: Should it be the quantity of power allocated to each utility, as PG&E prefers? Should it be the above-market costs that SCE calculates, or should we utilize the all-in costs that SDG&E measures? And how should the residual net short and other metrics that DWR developed be considered?

We do not believe that such an approach is appropriate, or necessary, to resolve the issues in this proceeding, for two major reasons. First, there appears to be no single, clear cut "right" framework for considering contract allocation proposals, or an ideal companion comparison metric. Take the issue of the net-short position. SCE and SDG&E argue that the only legitimate perspective is one that looks at the net-short of the utilities during the 2001-2002 period because that is the net-short position DWR was trying to fill when it entered into the contracts. There is some merit to this perspective to the extent that DWR may not have anticipated direct access load migration at that time, or taken into account other factors that now lead it to project significant changes to the net-short positions of the utilities in the coming years.

However, under this approach, SCE and SDG&E make the questionable assumption that DWR entered into long-term contracts without any consideration of future net short needs or longer-term hydro conditions. Moreover, eliminating any consideration of future net-short needs ignores the potential impact of contract allocation decisions on the utilities' ability to actively participate in the procurement process as they resume that responsibility. In fact, we observe that forecasts of residual net short must have been at least implicitly considered by the utilities, since each utility ended up proposing an allocation that gave it the highest residual net short percentage relative to any of the other allocation options.³⁶ (See Attachment 4.)

SCE urges us to reject any allocation principle that utilizes a forecast of net short, arguing that there is too much forecasting error contained in such projections. In particular, SCE argues that the PROSYMRun35 is not consistent with SCE's forecast of net short over the next seven years, has not been examined in a Commission proceeding and is likely to be the subject of contentious disputes over accuracy when it does come before the Commission in DWR's revenue requirements proceeding.³⁷ SDG&E also argues that DWR's forecast of net short contains inaccurate assumptions.³⁸

However, one could argue that the above-market cost metric, which SCE asks us to rely on in determining the ultimate reasonableness of contract allocation proposals, suffers from similar shortcomings. SCE's calculations of above-market costs, which first appear in this proceeding on July 19, 2002, have not been scrutinized by parties or ultimately adopted as reasonable in any Commission proceeding. In fact, SDG&E states that it did not even review SCE's calculations as it developed the joint proposal.³⁹ SCE's methodology relies on several calculations and projections that are based on subjective assumptions, including a forecast of future market prices (forward price curve) based on broker quotes and (after 2007) growth rate assumptions, and calculations of future hourly market prices that are derived from a regression analysis of the forward price curves and historical market prices.

Despite SCE's suggestion that the above-market cost metric is immune from forecasting inaccuracies, because the relative proportion of above-market costs among contracts do not depend upon the absolute level of the market price projection,⁴⁰ a closer examination of SCE's workpapers indicates that this is not the case. Under SCE's methodology, changes to the projection of forward prices will impact the relative amount of power dispatched from dispatchable contracts, vis-à-vis the must-take contracts. This, in turn, can alter the relative level of contract costs (and resulting proportion of costs above market) among dispatchable and must-take contracts. In this way, the results of SCE's above-market cost calculations are, in fact, sensitive to market price assumptions.

In sum, one reason to reject the notion of selecting a particular contract allocation approach over another is that there is simply no clear "winner" in the bunch. The second reason to reject this notion is that the proposals appear to be driven exclusively by the results, rather than by a commitment to underlying allocation principles. Had the reverse been true, one would expect to see some mixed results under the utility's preferred allocation in terms of the various comparison metrics presented in this proceeding.

However, as can be seen from the comparison tables presented in Attachment 4, PG&E's preferred allocation results in the lowest allocation of estimated costs to its customers, irrespective of what other proposal it is compared against or which cost metric is used.⁴¹ Similarly, PG&E finds itself with the least amount of contract energy and capacity allocated to its portfolio (and, correspondingly, the largest residual net short) under its preferred option, relative to all others. As a direct corollary, SCE and SDG&E are allocated the highest costs under any of the cost metrics, the most contract energy and capacity (and left with the lowest residual net short) under PG&E's proposal.⁴² It is not surprising to see similar results (in reverse) under SCE's and SDG&E's preferred contract allocation.

In sum, despite the very detailed arguments of each utility on why we should use a particular set of principles and methods for allocating the DWR contracts, the record in this proceeding indicates to us that those principles and methods were in fact developed to justify a specific set of results. Moreover, as discussed above, selecting a single allocation methodology as the ideal approach ignores the fact that the various allocation principles and associated comparison metrics presented in this proceeding have disadvantages when considered in isolation.

Instead, we will allocate Sunrise and Coral in a manner that strikes an appropriate balance between the competing proposals of PG&E and SDG&E/SCE and the various principles and comparison metrics considered. We believe that the ALJ alternate that allocates the Sunrise contract to SDG&E and the Coral contract to PG&E strikes such a balance, for several reasons.

First, despite the utilities' objections to this alternate, the results of such an allocation do reach a reasonable middle ground between the utilities' proposals with respect to the comparison metrics and net-short calculations considered in this proceeding. This can be seen by examining the tables in Attachment 4.

Whereas PG&E's proposal would result in a 38% share of contract energy (37% share of capacity) allocated to its portfolio, and SCE's proposal would allocate to PG&E a 44% share of energy (47% share of capacity), the ALJ alternate results in a 41% share of energy (43% share of capacity) going to PG&E. In terms of SCE's above-market cost metric, PG&E would be allocated 42% of the above-market costs calculated by SCE under the ALJ alternate, compared to 37% under its own proposal and 48% under the joint proposal of SCE and SDG&E. The contract costs (assuming either a pro rata or cost following allocation) for PG&E under the ALJ alternate also fall between the two utility proposals. The ALJ alternate results in an allocation to PG&E of 7% of the total residual net short, compared to 10% under PG&E's proposal and 4% under the SCE/SDG&E joint proposal.

Similarly, the ALJ alternate finds a middle-ground with respect to the allocation of power, cost and residual net short for SDG&E and SCE, as well. The ALJ alternate also produces results for the allocation of contract power and above-market costs that fall within the range of net short calculations presented in this proceeding as the yardstick for evaluating allocation options. This can be seen by comparing the percentage allocations associated with the ALJ Alternate in Attachment 4 for "allocated energy," "allocated capacity" and "allocated above-market costs" with the range of net-short percentages presented in Attachment 3.

Second, the ALJ alternate reaches an appropriate balance in allocating contracts among the utility portfolios that involve some delivery uncertainties. SCE is allocated the Sempra must-take contract, which has delivery point optionality for approximately 1500 MW. PG&E is allocated the must-take Coral and dispatchable PacificCorp contract quantities with delivery point optionality for combined total of 820 MW. SDG&E is allocated the dispatchable Sunrise contract (560 MW) with a delivery point in ZP-26, north of Path 26. For all three utilities, there may be potential transmission bottlenecks during certain times of the year in delivering power from these contracts to their service territories. However, in our judgment, the ALJ alternate allocation spreads the delivery risk associated with these contracts more equitably among all three utilities when compared to the utilities' proposals.

Third, the ALJ alternate also reaches a reasonable outcome in terms of SP-15 and NP-15 zonal allocation: PG&E takes scheduling and dispatch responsibility for all quantities that have NP-15 specified delivery points, and contracts with SP-15 specified delivery points are allocated to the southern utilities. This allocation of responsibility avoids necessitating an inter-zonal transfer or a sale of these quantities if a transmission path is congested.

In contrast, both PG&E's preferred proposal and the permutation of the Coral split that PG&E presented in its July 30 comments would allocate a significant portion of the Coral contract quantities designed for NP-15 to the south. Under PG&E's preferred proposal, where 25% of the base and 25% of the additional quantities are allocated to PG&E with the rest going south, a total of up to 288.75 MW designated for NP-15 delivery would be allocated to SDG&E and SCE. Under the permutation that PG&E developed in response to the ALJ's ruling, where all base quantities would go south, 125 MW of NP-15 designated quantities would be allocated to SDG&E and SCE.⁴³ Moreover, as SCE points out, the complexities of the Coral contract make it an undesirable candidate for splitting.

And finally, the ALJ alternate produces a resolution of the issues without requiring any utility to manage the contract of its affiliate. While our Affiliate Transaction Rules do not prohibit such an arrangement and provide sufficient safeguards against self-dealing and other potential market abuses, a contract allocation approach that does not create a utility-affiliate relationship simplifies our task in overseeing the administration of such contracts.⁴⁴

For these reasons, we adopt the allocation of DWR contracts under the ALJ alternate, as presented in Table 1.

The utilities urge us to specify in advance a trigger that would initiate either a reconsideration of the physical contract allocations we adopt today, or a financial adjustment to the allocation of DWR's revenue requirement among the utilities.For example, SCE recommends that we establish a trigger to reallocate the DWR revenue requirement among utilities if restructuring of any contract changes the above-market cost by the greater of 10 percent of the original contract's above-market costs, or $10 million. Under SCE's proposal, the Commission would reallocate the resulting change in above-market costs equitably among the utilities in proportion to the share of above-market costs associated with today's adopted contract allocation. PG&E proposes a trigger that would adjust the physical contract allocation adopted today based on an energy threshold, in the event of either contract restructuring or termination. SDG&E draws from PG&E's and SCE's proposals and recommends a trigger mechanism that is based on changes to the amount of energy, capacity or above-market costs.⁴⁵

We do not adopt any of the utilities' proposals for a trigger mechanism. As discussed above, we believe that the best way to coordinate DWR contracts with utility resources is to put them in the utilities' resource portfolios to be scheduled and dispatched by the utilities. This decision, along with a decision on the utilities' procurement plans, will enable the utilities to move forward with their procurement planning and determine the appropriate combination of long-term and short-term energy and capacity needed to meet their individual residual net short requirements.

In our view, adopting an advance trigger mechanism to reconsider the contract allocation we adopt today would insert an unacceptable level of additional uncertainty and complexity into the procurement process going forward. If a DWR contract were renegotiated and the trigger criteria were met, each utility would be placed in the position of trying to plan for future needs not knowing which DWR contracts might be removed from its current portfolio, added from another utility's portfolio, or a combination of both. Until the Commission took final action in response to the "triggered" reallocation proceeding, this uncertainty would persist. Based on our experience with the contract allocation process to date, revisiting this issue would take a significant amount of time and resources. Moreover, any such future reallocation process would be further complicated by the fact that the utilities would already have made procurement decisions and commitments based on the allocation we adopt today. While SCE's proposal to establish a financial adjustment trigger appears to avoid these particular problems, it relies upon an above-market cost metric (both in the trigger and response) that has not been carefully scrutinized or adopted by the Commission.

³⁵ The only exception is PG&E's treatment (split) of the Coral contract, as discussed further below. ³⁶ The only exception to this observation is SCE's proposal when compared to the DWR options that would allow affiliate contracts. See RT at 44-45. ³⁷ RT at 87; SCE's July 30, 2002 comments, pp. 8-10; SCE's August 5, 2002 Reply Comments, pp. 11-13. ³⁸ SDG&E's July 30, 2002 comments, pp. 3-4. ³⁹ RT at 59. ⁴⁰ RT at 87-88. ⁴¹ We do not present the all-in cost metric prepared by SDG&E in these tables because SDG&E submitted its analysis and the $/MWh results under seal. We presume this is because the calculations require the use of information regarding SDG&E's residual net short quantities and other information that SDG&E considers confidential. For similar reasons, SDG&E did not prepare all-in costs for SCE and PG&E, and SCE and PG&E indicated during workshops that they could not reproduce SDG&E's results for their systems without a delay in the schedule. Therefore, even if SDG&E's numbers were not under seal, we do not have a utility-by-utility comparison of this cost metric, by allocation option, as we do with the other cost metrics presented in this proceeding. However, based on our review of the calculations that SDG&E did provide in its filings, SDG&E's all-in costs are minimized under the same scenarios for which the other cost metrics are lowest for SDG&E, and vice versa. We therefore expect that the relative position of each utility with respect to all-in costs, had they been provided, would correlate with the other cost metrics in a similar manner. ⁴² There is less of a direct pattern to the results in terms of the must-take surplus, i.e., the utility's preferred scenario does not systematically result in the lowest surplus amounts relative to all other options that were considered in this proceeding. However, the utilities' preferred allocations do maintain the must-take surplus within a relatively low (2-4%) range. ⁴³ RT at 21. CERS Summary Diagram of Written Example Used at July 26, 2002 Workshop, dated July 30, 2002. The Coral contract specifies that the entire 350MW of "additional quantities" are to be delivered to NP-15. 75% of the additional quantities equal 262.5 MW. In addition, the seller may increase/decrease all capacities a year by 10%. 75% of this "put" option equals 26.25 MW, for a total of 288.75 MW of NP-15 quantities going to the south under PG&E's preferred allocation.

The contract also requires a minimum of 25% of the 500 MW in base quantities to be delivered to NP-15, so PG&E's alternate permutation would send 125MW of NP-15 designated quantities to the south.

⁴⁴ All interactions between a utility and its affiliates are subject to specific rules addressing nondiscrimination, disclosure and information, and separation. In particular, the rules generally require that all transactions between a utility and its affiliates be conducted at arms-length. (See D.97-12-088, Appendix A.) ⁴⁵ SCE's July 19, 2002 comments, pp. 27-29. PG&E's July 30, 2002 comments, pp. 7-8; SDG&E's August 5, 2002 comments, pp. 11-12.