Allocation of DWR's 2003 Revenue Requirement

The main issue in this proceeding is how to allocate DWR's 2003 revenue requirement among the customers of three major utilities. This is not a brand-new task; we have previously allocated DWR's revenue requirement, but not in the same context that we face today. Nevertheless, two of our recent decisions, D.02-02-052 and D.02-09-053, provide some guidance on this issue.

In D.02-02-052, we allocated DWR's revenue requirement for 2001 and 2002.⁴ In that decision, we evaluated a number of competing proposals, and ultimately adopted an allocation method proposed by SCE. We summarized that method and its basis as follows:

SCE characterizes the procurement costs of DWR fixed long term (90 days or longer) contracts as costs incurred to meet the joint net short position of all three utilities. Because these long-term contracts provided a benefit to the entire State of California by lowering electricity prices on the spot market, SCE proposes that such fixed contract costs be allocated pro rata based on each utility's net short position.

For short-term purchases (less than 90 days), however, SCE proposes that supply costs be allocated between PG&E and southern California utility customers based on the separate zonal cost of supplies using Path 15 as a dividing point. (Id., p.48.)

For 2003, however, DWR will not be making any short-term purchases. (See, Water Code section 80260.) Accordingly, we do not need to allocate the costs of short-term purchases for 2003, but we do need to again allocate the costs of the existing long-term contracts.

Subsequently, in D.02-09-053, we adopted a policy of allocating the variable costs of the existing DWR contracts to the three major utilities. As of January 1, 2003, the utilities will be placing those contracts into their resource portfolios to be scheduled and dispatched in a least-cost manner. This was done as part of the process of requiring the utilities to resume their procurement planning role. As we stated, "...the utilities will now perform all of the day-to-day scheduling, dispatch and administrative functions for the DWR contracts allocated to their portfolios, just as they will perform those functions for their existing resources and new procurements. Legal title, financial reporting and responsibility for the payment of contract-related bills will remain with DWR." (D.02-09-053, p.5.)

While D.02-09-053 established the policy that the variable costs of each contract should follow contract allocation (id., p.6), it left to this proceeding the determination of the proper allocation of the total DWR revenue requirement. As a practical matter, since we are not changing the allocation of variable costs from D.02-09-053, what remains to be allocated here are costs other than variable costs, which consist primarily of the fixed costs of the contracts and DWR's related administrative and general costs.

The contents of DWR's 2003 revenue requirement are shown in the following summary tables:

DWR's August 16^th Determination of 2003 Power Charge Expenses

Power Costs	$4,119,902,243
Administrative and General Expenses	$28,400,000
Ancillary services	$170,454,426
Increase in Operating Fund Balance	$517,399,690
Total DWR Power Charge Expense	$4,836,156,359
Less: Revenue from Sales of Excess Power	($128,885,940)
Less: Interest Earnings on Fund Balances	($59,007,505)
Total Ratepayer Revenue Requirement	$4,648,262,914

Source: DWR August 16^th Determination, Table A-1

The positions of a number of parties shifted during the course of this proceeding, as their understanding of other parties' positions (and their own positions) evolved. These migrations of position were likely exaggerated by the highly expedited schedule, as parties had a relatively short time to do discovery and analysis prior to testimony and hearings. While this movement has resulted in some increased alignment of positions, the four most active parties still presented four different allocation methodologies.

SCE

SCE has proposed that the allocation of fixed costs follow the methodology adopted in D.02-09-053 for allocation of variable costs. This approach, commonly referred to as "costs-follow-contracts," would result in the fixed costs of DWR contracts being allocated to the customers of the same utilities to which the variable costs of those contracts were allocated in D.02-09-053.

According to SCE, the advantage of this approach is its internal consistency, as it avoids the possibility of one utility receiving a large allocation of variable costs under one method and a large allocation of fixed costs under another method. In addition to avoiding a mix of different allocation methods, which SCE regards as potentially unfair, SCE argues that its proposal is unique in that it provides the only approach that does not require future proceedings to establish future year allocations. SCE argues in the alternative that if the Commission were to decide not to use the "costs-follow-contracts" approach, then the Commission should adopt ORA's recommended approach.

SCE's "costs-follow-contracts" proposal attempts to use the allocation of contracts adopted in D.02-09-053 to allocate fixed contract costs, but SCE has not established that doing so is appropriate. D.02-02-052 addressed fixed costs, and accordingly is the more directly applicable precedent than D.02-09-053, which focused on variable costs. SCE's proposal conflicts with D.02-02-052.⁵ While we are not necessarily bound to follow D.02-02-052, SCE has not persuaded us that there is a good reason for departing from that decision.

In D.02-02-052 we stated:

[W]e agree with the goal of allocating DWR costs in relation to the costs of providing service. We do not believe, however, that segregating disproportionately higher priced DWR power for allocation exclusively to northern California consumers is a proper or fair application of traditional cost-based ratemaking policies. (Id., p.4.)

One measure of cost causation in relation to the three separate utility service territories would be evidence that DWR had actually procured separate portfolios of supplies specifically targeted toward each respective utility's customers. If DWR had expressly procured a separate portfolio of supplies for each utility service territory, there would be a strong cause-and-effect relationship between location of supplies and specific utility service territory served. This, in fact, did not occur. (Id., pp. 59-60.)

We concluded that:

DWR thus has not maintained separate portfolios to meet the net short positions of each utility. Any allocation of power purchased under the DWR contracts and spot market purchases for each respective service area by assuming distinctly separate sources of supply for each utility is not consistent with the way DWR constructed its portfolio of supplies, and would not necessarily result in any more logical or accurate cost causation than a statewide pro rata approach. (Id., p. 60.)

Our allocation in D.02-02-052 recognized the primarily integrated nature of power procurement undertaken by DWR for California utility customers, but we also adjusted for utility-specific differences, where applicable. Utility-specific adjustments were determined to be appropriate only in the case of short-term purchases, which we allocated geographically. Short-term purchases are not present here, eliminating the need for corresponding utility-specific adjustments to the allocation methodology.

Since DWR signed contracts for a statewide need, allocating the fixed costs of contracts to utility service territories based upon geographic location does not match how or why those contracts were obtained. It would be arbitrary and unfair for one or more service territories to end up with a disproportionate number of high-priced contracts when DWR was not trying to balance costs among service territories.

TURN notes that SCE's approach has the "appeal of simplicity and finality." Nevertheless, TURN argues that while SCE's "costs-follow-contracts" approach may have potential as a long term or future methodology, it is too soon to adopt it, particularly while the contracts themselves are in the process of being renegotiated. (TURN Reply Brief, p. 5.)⁶ As TURN points out, with the contracts in active renegotiation, we cannot know how the cost of each contract may change in the future, and we have no way to evaluate the ultimate fairness of this allocation approach. PG&E similarly believes that "costs-follow-contracts" may be appropriate after contract renegotiations are concluded, but not before. (Ex. 1, p. 1-8.)

PG&E

PG&E proposes that DWR's revenue requirement be allocated to each utility in proportion to each utility's 2003 net short, adjusted to add back load loss from direct access and departing load customers, resulting in what PG&E calls "pre-load migration net short." From that initial allocation, PG&E would then subtract the variable costs that have been allocated to each utility. The remainder for each utility is the fixed cost component that gets remitted to DWR. According to PG&E, this method best takes into account direct access and departing load, and also best reflects the cost drivers of DWR's original contracting activities.

Compared with SCE's proposal, PG&E's is more consistent with DWR's original procurement focus, which was the aggregate net short position of the three utilities. Nevertheless, the passage of time, and the corresponding changes in the electricity market, render PG&E's proposal less appropriate than it may have been a year ago. When the Commission was examining the allocation of DWR's revenue requirement for 2001 and 2002, DWR was buying power for each utility's net short via a mix of short-term and longer-term contract purchases. For 2003, DWR is out of the procurement business, and the energy delivered to each utility's customers by long-term DWR contracts does not necessarily match that utility's net short.

These changes render PG&E's allocation unfair. As ORA points out:

[A]llocation on the basis of net short can lead to double counting and the imposition of inequitable costs for its residual net short. For example, a utility allocated a share of contract energy that is smaller than its share of net short will end up paying for its residual net short twice, once as part of the DWR revenue requirement, and a second time in the open market as the utility resumes responsibility for procurement of its residual net short. (ORA Reply Brief, p.3, citing to SCE's Opening Brief.)

TURN also makes the same point - PG&E's proposal could result in customers essentially paying twice for the same energy.

SDG&E

SDG&E proposes a "postage stamp" allocation, with DWR contract costs allocated to each utility in proportion to the quantity of energy supplied by DWR to each utility. However, SDG&E does not subtract out the variable costs the way that PG&E does.⁷ Instead, SDG&E allocates the fixed costs independently of the variable costs. According to SDG&E, variable costs have already been allocated in D.02-09-053, have no role in the allocation of fixed costs, and need not be considered here. SDG&E argues that its proposed allocation is the most consistent with D.02-02-052.

The initial part of SDG&E's approach, with its allocation by supplied energy, is relatively equitable. By pooling all of the costs, it reflects the fact that DWR purchases and contracts were intended to cover the aggregate net short position of the three utilities. SDG&E's approach is more consistent with DWR's actual practices than is SCE's approach, which disaggregates the costs to the customers of each utility. SDG&E keeps the costs and benefits more closely aligned than SCE or PG&E, because SDG&E starts with the costs aggregated (the way that DWR incurred them), and then allocates them on the basis of what the contracts will actually provide in 2003: energy. DWR is not providing for the utilities' net short. Supplied energy, as proposed by SDG&E, is the most appropriate criteria for allocating the fixed costs of the DWR contracts.

However, SDG&E's subsequent disregard of variable costs gives an unfair result, and is criticized by all other parties. ORA makes the basic point:

SDG&E's direct allocation of fixed costs can unfairly burden a utility with a disproportionate share of variable costs. SDG&E's method leaves utilities sharing fixed costs, but not sharing variable costs. A utility with a disproportionally large share of variable costs ends up paying all of their own variable costs as well as a greater than proportionate share of others' fixed costs. (ORA Opening Brief, p. 8.)

PG&E provides a hypothetical example to illustrate the problem:

Under the example there are two utilities, and two contracts. The two contracts are expected to have the same overall costs. One contract has all fixed costs, and is allocated to one of the utilities. The other contract has all variable costs, and is allocated to the other utility...[T]he only distinction between the circumstances the two utilities face is that one has been allocated a contract that is all variable costs, while the other has been allocated a contract that is all fixed costs...The example illustrates that under SDG&E's approach the utility to which the variable cost contract has been allocated would bear all of the variable costs, plus half of the fixed costs, resulting in an overall burden for it of three-quarters of the costs. The other utility would bear only half of the fixed costs, resulting in an overall burden for it of one-quarter of the costs. (PG&E Opening Brief, pp. 13-14.)

SCE and TURN agree with ORA and PG&E that SDG&E's proposal to ignore variable costs is unfair.

ORA

ORA's proposal for allocation is also a "postage stamp" allocation. ORA's proposal starts out somewhat similarly to SDG&E, with a pro-rata allocation of the DWR revenue requirement based on each utility's share of the total amount of DWR delivered energy. ORA then departs from SDG&E by subtracting out the variable costs that have been allocated to each utility, resulting in a residual (fixed cost) revenue requirement.

According to ORA, the advantages of this method are that all utility bundled customers would be charged the same rate, and the allocation derived is the fairest, because it most accurately associates energy costs with the energy that the utility customers are actually getting from DWR's contracts. (ORA Opening Brief, pp. 3-4.)

While it is not clear that all bundled customers would actually be charged the same "rate" for DWR energy, ORA's proposal does treat all bundled customers equitably. ORA's overall approach is in fact the fairest of those proposed. Like SDG&E's proposal, ORA's proposal allocates costs in a way that corresponds to the benefits received (energy), and spreads the pain of those DWR contracts that are particularly expensive. ORA's proposal to distribute the costs of DWR contracts statewide among all ratepayers is more equitable and less arbitrary than the proposals of SCE and PG&E. Furthermore, by subtracting out the variable costs that we allocated in D.02-09-053, ORA's proposal avoids the problems caused by SDG&E's proposal to allocate fixed costs independently from variable costs.

An integral part of ORA's proposal is its recommendation that the Commission apply what ORA calls a "pre-Direct Access metric." ORA argues that the Commission should adjust the allocation of DWR's revenue requirement to take into consideration direct access and departing load customers subject to the Cost Responsibility Surcharge (CRS) set in R.02-01-011. In its Reply Brief, ORA acknowledges that this adjustment requires the results of a "Direct Access-In" modeling run from DWR's consultant, which had not yet been performed. ORA anticipated that such a modeling run would be completed well prior to the issuance of a Proposed Decision in this proceeding, and accordingly could be incorporated here. (ORA Reply Brief, p. 4, fn. 2.) Unfortunately, that did not happen, and the modeling run could not be completed in time to be utilized in this proceeding.

ORA's proposed adjustment received broad support. In addition to TURN, even parties who proposed different allocation methodologies did not quarrel with ORA's proposed adjustment. PG&E generally agrees with ORA that the allocation of DWR's revenue requirement should take into account direct access migration, and that the allocation should be consistent with the treatment of direct access and departing load in the CRS proceeding. (PG&E Reply Brief, p.4.) While SCE indicates some reservations (due to its questioning of certain direct access and departing load data), it endorses ORA's allocation proposal - including direct access and departing load adjustments - as the next-best alternative to its own proposal. (SCE Opening Brief, p.9.)

The direct access adjustment proposed by ORA is appropriate.⁸ ORA's proposed departing load adjustment may also be appropriate, but it is not clear when the information necessary to perform that adjustment will be available, as resolution of that issue has been deferred in the CRS proceeding.

Nevertheless, despite its merits, we are unable to incorporate the direct access adjustment at this time, as the evidentiary record in this proceeding does not provide adequate support for that adjustment. Accordingly, for the time being we are adopting ORA's proposal without the direct access adjustments. As soon as we are able to incorporate the appropriate modeling runs, we will make the necessary adjustments to reflect direct access. DWR should incorporate a "Direct-Access-In" modeling run into any supplemental determination it submits, so that we can make the adjustments described above.

We do need to make several minor modifications to ORA's methodology. TURN, which supports ORA's approach,⁹ suggests two minor modifications. First, TURN argues that:

[T]he revenues associated with off-system sales of DWR power should not be "pooled" and then allocated among the three utilities. Rather, these revenues should directly offset the revenue requirement of the dispatching utility. Otherwise, the incentive for economic dispatch would be seriously distorted. (TURN Reply Brief, p. 3, emphasis in original.)

The three utilities concur on this point, and we also agree.¹⁰ Pooling would reduce the incentive for a utility to maximize the revenues from its sales of surplus energy. As discussed further below, revenues from sales of surplus energy should be credited to the portion of the revenue requirement allocated to the customers of the utility making the sale.

TURN also recommends that instead of using the numbers for "DWR Delivered Energy" (sometimes referred to as retail energy), as proposed by ORA, it would be more appropriate to use the numbers for "DWR Supplied Energy" (sometimes referred to as wholesale energy). The basic difference between these two is that Delivered Energy has line losses subtracted out, while Supplied Energy reflects total DWR supplies prior to the subtraction of line losses. TURN's recommendation (which is similar to SDG&E's position on this point) results in the allocation of the revenue requirements better reflecting the differing line losses of the utilities, because DWR does not need to send as much energy to the customers of those utilities with lower line losses. To allocate DWR's revenue requirement on the basis of the amount of energy received by each utility's customer would result in customers of utilities with low line losses paying for the energy lost by the systems of utilities with larger line losses. Accordingly, we will modify ORA's proposal as recommended by TURN, and use the amount of energy sent to each utility service territory, rather than the amount of energy received, as the basis for allocating DWR's revenue requirement.

In addition to TURN's recommendations, PG&E points out that ORA erroneously treats tolling charges associated with DWR must-take contracts as a variable cost. (PG&E Reply Brief, p. 7.) As defined by D.02-09-053, tolling charges are sunk or unavoidable if those costs cannot be avoided by dispatch decision. DWR and the other utilities concur with PG&E that tolling contracts associated with must-take contracts should be considered a fixed cost, as their costs are not avoidable by dispatch decision. To apply ORA's allocation method in a manner that is consistent with D.02-09-053, we will treat tolling charges associated with must take contracts as fixed costs, not variable costs.

Our adopted interim methodology for the allocation of costs gives the results shown on the following table, which also shows how those results compare with the results of the other proposed methodologies:

Table A: Proposed and Interim Adopted Allocations

	Allocations to IOU Customers
	PG&E		SCE		SDG&E		Total
PG&E Proposal	$1,846,000,000	41%	$1,824,000,000	41%	$808,000,000	18%	$4,478,000,000
SCE Proposal	$2,198,000,000	48%	$1,708,000,000	37%	$664,000,000	15%	$4,570,000,000
SDG&E Proposal	$1,995,000,000	44%	$1,890,000,000	42%	$690,000,000	15%	$4,575,000,000
ORA Proposal	$2,042,000,000	45%	$1,764,000,000	39%	$752,000,000	17%	$4,559,000,000
ORA*	$1,965,158,417	44%	$1,879,525,727	42%	$643,087,606	14%	$4,487,771,749

*As modified and adopted in this decision.

Note: Proposed allocations are shown as presented by the parties in the "Contract Cost Allocation Comparison Exhibit". In addition, due to rounding, sums may not equal totals.

The following table provides more detailed information on the adopted allocation:

TABLE B: Detailed Summary of Interim Adopted Allocation

In order to have consistent assumptions and inputs, neither table adjusts for direct access (i.e., they do not incorporate a "Direct Access-In" modeling run), and both tables reflect the use of the modeling run known as PROSYM 36.¹¹ For additional detail on our adopted methodology, please refer to Appendix A.