VI. Litigation Positions, Past Allocations, and Fairness Metrics
Since we have rejected the proposed settlement agreement, we next consider the litigation proposals of the parties. Unfortunately, we also find the litigation proposals to be unsuitable for permanently allocating the costs of the DWR contracts.
The litigation proposals of PG&E and SCE, while not identical, share a number of flaws. Among other things, the PG&E and SCE proposals are based upon the inequitable CFC methodology, rely upon flawed estimates of above-market costs, and invite significant re-litigation. The proposals of ORA and SDG&E are based upon the allocation methodology adopted for 2003 in D.02-12-045. However, the ORA proposal is somewhat incomplete, and SDG&E incorporates additional self-serving resource assumptions in its proposal.
We will, however, take as our starting point the pro rata allocation methodology we adopted in D.02-12-045. This allocation methodology is generally advocated in the litigation positions of ORA and SDG&E, and partially supported by the litigation position of PG&E. In D.02-12-045, the adopted methodology pooled the costs of the DWR contracts, reflecting the fact that DWR's contracts were signed with the intent to meet the aggregate need of the utilities. (Id., pp. 11-13.) Aggregating the costs of the contracts accordingly matches the way the costs were actually incurred, and also spreads the pain of those contracts that are particularly expensive. (Id.)
After pooling the DWR contract costs, D.02-12-045 then allocated the costs to the customers in proportion to the quantity of energy supplied by DWR to each utility for 2003, as forecast by a modeling run performed by DWR. The use of supplied energy as the basis for the allocation was appropriate, as it allocated the costs of the contracts proportionally to the benefits received from the contracts.
D.02-12-045 used the most current forecast for which there was an adequate record. (Id., pp. 27-28.) Given that the forecast was for one year only, this approach had a reasonable probability of accuracy. However, in this proceeding we need to find a basis for allocating the costs of the DWR contracts through 2013, when the last of the contracts expire. Using a forecast of supplied energy over a relatively long time period, however, raises some of the same concerns we have expressed regarding the accuracy of the above-market cost forecast.
In addition, it is not clear what forecast should be used. One approach would use the most recent DWR forecasts of supplied energy for the relevant time period. (See, e.g., SDG&E Opening Brief, p. 9.) This approach, by using the most recent forecast reflecting the latest data and assumptions, would hopefully be the most accurate. Even so, its accuracy is uncertain at best. In addition, SCE argues that the use of current forecasts requires the use of confidential information, making it difficult for parties to review and confirm the results of the allocation methodology. (SCE Opening Brief, p. 32.)
On the other hand, it is also possible to use a forecast based on a model used by DWR back when it was entering into the contracts. (See, e.g., PG&E Opening Brief, p. 18.) This, as discussed above, is an attempt to mirror the state of mind of DWR at the time it executed the contracts, and tries to reflect DWR's then-current projections over the longer term. As a proxy for DWR's state of mind, this approach may not be bad, but its ultimate accuracy is likely to be worse than the more recent forecasts.8 Neither the current forecast nor the historical forecast is ideal, and while both have benefits and flaws, it is difficult to determine which would be the lesser of two evils. Ultimately, it is not clear that any forecast in the record of this proceeding is really any good, and we decline to base our allocation on the available forecasts.
We cannot predict the future, and in this case the past is also of little help, as the DWR contracts at issue were signed at a time of crisis, confusion, and uncertainty, rendering our traditional notions of cost causation inappropriate. In large part we are "spreading the pain" of a unique occurrence, for which our standard methods are ill-suited. Accordingly, we must find another way to reach a fair allocation. Fortunately, the parties have provided such a route.
As a guide to evaluating the various allocation methodologies, several parties recommended the use of a "fairness yardstick" or "fairness metric," against which allocation proposals could be measured.9 Not surprisingly, there was some divergence among the parties among what should be considered fair. Nevertheless, given the problems of the other methods proposed in this proceeding, the fairness metrics appear to provide the best avenue for developing a fair and equitable cost allocation.
In fact, it is quite informative to look at the fairness metrics as well as actual current and recent historical allocations.10 We note that the percentages recounted below represent total cost allocations, and not percentages of unavoidable costs, which are the only costs at issue in this decision, as described more fully in the next section.
Source/Method |
Allocation to: |
||
PG&E |
SCE |
SDG&E | |
D.02-12-04511 |
44% |
42% |
14% |
Interim 200412 |
40% |
47.3% |
12.7% |
PG&E Metric13 |
39% |
48.4% |
12.5% |
SCE Metric14 |
48% |
36% |
16% |
ORA Metric15 |
39.2% |
48.3% |
12.6% |
Settlement Metric |
43.6% |
42.6% |
13.8% |
It is particularly interesting to note that two parties other than SDG&E (PG&E and ORA) were willing to argue that an allocation as low as approximately 12.5% would be appropriate for SDG&E. With the exception of SCE's Metric, the range of allocations recommended or previously adopted for SDG&E is also relatively narrow. Since all of the principles upon which we might base an allocation methodology appear to be flawed, we will simply use the concurrence of the parties to establish fixed allocation percentages that will be applied to the cost of DWR's contracts.
SDG&E correctly argues that the percentages in the table above relate to total cost allocation, and not an allocation of fixed or unavoidable costs alone. SDG&E argues that in order to arrive at a result of 12.5% of the total costs to be borne by SDG&E (the lowest of the percentages reflected above), only 9.5% of the fixed costs being allocated in this decision should be attributed to their ratepayers.
We conclude that all of the proposals summarized above represent the range of reasonable possible outcomes. Since the decision we make today is inherently a zero-sum game that involves spreading costs fairly and evenly such that no one set of ratepayers is either harmed or subsidized unnecessarily, we choose to allocate to SDG&E the average of all of the "fairness metric" percentage allocations proposed, which equates to 13.6% of total costs. Using a ratio approach, if 12.5% of total costs equates to 9.5% of unavoidable costs to SDG&E, a 13.6% share of total costs equates to 10.3% of unavoidable costs. Thus, for the fixed costs we are allocating today, SDG&E's percentage will be 10.3%.
There is less concurrence regarding the relative shares to be allocated to PG&E and SCE, but considering the levels of PG&E's Metric, SCE's Metric, and the Settlement Metric, it is appropriate to allocate the remaining 89.7% on the basis that both utilities agreed to amongst themselves in the Settlement Agreement. To derive the appropriate percentages, we use another ratio approach. Under the Proposed Settlement, after subtracting SDG&E's share of the non-variable or unavoidable costs, the remainder is divided between PG&E and SCE 47% to 53%, respectively, over the total period 2004-2013. While we allocate a lower percentage to SDG&E as discussed above, we divide the remainder of the costs between PG&E and SCE in the same manner as results from their Proposed Settlement. Accordingly, after subtracting SDG&E's 10.3% share, we adopt an allocation of the remaining costs in the same relative proportion as results from the Settlement Agreement. Thus, PG&E's share will be 42.2%, while SCE's will be 47.5%.
Use of fixed allocation percentages is consistent with the recommendation of SDG&E. According to SDG&E, fixed percentages eliminate the ability of utilities to shift costs to each other via their dispatch decisions (SDG&E Opening Brief, pp. 17-19), and also reduce the motivation to relitigate the allocation methodology (SDG&E Reply Brief, pp. 23-24).
8 As noted above, while the best sources of information appear to be the Nichols Declaration and Prosym Run 19g, there was some controversy about the most appropriate source for the historical data and forecasts available to DWR. Accordingly, it is not clear what relative weight each of these sources should be given, if any.
9 In essence, the "Utility Allocation Percentages" in the proposed settlement also represent a fairness metric. (Motion of Settling Parties, p. 4.)
10 For purposes of this table, Prosym Run 43 was used, as that run provided the basis for the parties' litigation proposals, including their fairness metrics.
11 Allocation for 2003.
12 As adopted in D.04-01-028, applying the same method adopted in D.02-12-045.
13 Based on method adopted in D.02-12-045.
14 Based on forecasts of net short (SCE Opening Brief, p. 11).
15 Based on method adopted in D.02-12-045.
