V. Resolution of CMUA Petition for Modification

A. Overview of Issues

Although the scope of the rehearing focuses on the CRS allocation applicable to "new load," we also address as a related issue in this order the implications of new disclosures provided by PG&E and DWR relating to PG&E's "transferred load" forecast associated with MDL bypass. This issue was brought to the Commission's attention by CMUA.

In its reply comments on rehearing issues, however, CMUA took note of new disclosures contained in the comments of PG&E and DWR concerning the point in time that PG&E delivered its load forecasts to DWR. CMUA cited DWR statements in its memorandum attached to the October 20, 2003 ALJ ruling on rehearing issues, indicating that on February 14, 2001, DWR started to use a multi-year forecast that had been provided by PG&E.

These comments indicated that PG&E's forecast was delivered to DWR earlier than PG&E had previously claimed, and conflicted with certain assumptions underlying D.03-07-028 concerning the timing of the load forecasts relative to DWR's procurement of power, including procurement associated with MDL. Specifically, in D.03-07-028, we relied on assertions made by PG&E that it did not provide a sales forecast to DWR until June 2001. Based on the fact that DWR had completed the contracting of the bulk of its power purchases by that date, we concluded in D.03-07-028 that even though PG&E's load forecast incorporated specific adjustments to exclude MDL bypass, the load data was delivered too late to be utilized by DWR in determining its power procurement requirements. In view of the lack of record of any specific exclusions of MDL bypass in load forecasts relied upon by DWR, therefore, we did not adopt a CRS exclusion for MDL customers in D.03-07-028, except for the limited "new load" exception attributable to existing publicly-owned utilities discussed above.

If DWR received PG&E's forecast as early as February 2001, CMUA argued, the Commission's findings in D.03-07-028 were based on wrong information that required correction. CMUA asked that this issue be considered in the rehearing phase of this proceeding, or in the alternative, sought leave to file a Petition for Modification of D.03-07-028 to address the issue. In its Petition for Modification, CMUA argued that hearings may be necessary to address the veracity and implications of these "new factual representations" regarding PG&E load forecasts provided to DWR.

In view of the implications arising from apparent discrepancies as to when PG&E delivered its forecast to DWR, we scheduled evidentiary hearings related to the limited scope of factual disputes as to the nature and timing of load data provided by PG&E to DWR. By ALJ ruling dated August 10, 2004, the following issues were identified and set for further evidentiary hearings:

How many years' worth of forecast data were provided in the initial load forecast that PG&E delivered to DWR? When was the forecast first delivered to DWR?

What was the amount of "transferred load" that PG&E incorporated from its Bypass Report into the initial load forecast provided to DWR?

How does the "transferred load" impact the power requirements that DWR procured for (1) existing IOU customers as of February 1, 2001 that subsequently became MDL and (2) new load, if any, added by municipalities or irrigation districts after February 1, 2001 that were in areas covered by the "transferred load" forecasts.

To the extent that DWR independently extrapolated additional years of forecast data beyond those provided by PG&E, what, if any, data relating to municipal load bypass incorporated in DWR's calculations?

Do the currently adopted requirements for MDL CRS obligations appropriately take into account the effects of PG&E "transferred load" incorporated into forecasts utilized by DWR? If not, what adjustments to the MDL CRS obligation need to be adopted in order to recognize the effects of PG&E's "transferred load"?

PG&E and DWR produced witnesses to testify concerning their assertions set forth in their previously filed pleadings as to the load forecast submissions provided to DWR between January 1, 2001, and June 30, 2001, and the extent to which such submissions contained assumptions concerning MDL. PG&E sponsored two witnesses, Roy M. Kuga, PG&E Vice President in charge of oversight of daily gas and electric procurement functions, and Dennis M. Keane, PG&E Manager responsible for analysts supporting various PG&E regulatory filings and customer retention efforts. DWR sponsored one witness, Craig McDonald, managing director of Navigant Consulting, Inc.

B. Implications of Assumptions Concerning Delivery Date of PG&E Load Forecast Data to DWR

At the start of the rehearing proceeding, parties were initially in dispute concerning when PG&E delivered its three-year load forecast data to DWR. In its data response (attached to PG&E's rehearing comments), DWR stated that the forecast received from PG&E on February 14, 2001 consisted of three years of data (for 2001-03). DWR independently extended the forecast to cover a 10-year period by taking PG&E's three-year forecast (2001-03) using PG&E FERC filing data.

In its opposition to CMUA's Petition to Modify, filed March 18, 2004, PG&E denied that its August 2000 Bypass Report was provided to DWR in February 2001. PG&E claimed that it has no record of sending DWR any forecast in February of 2001. In its March 18, 2004, response, PG&E did attach an email printout indicating that it provided at least a one-year 2001 forecast to DWR at the end of March 2001. PG&E acknowledged receipt of a DWR email dated March 30, 2001 from DWR to PG&E, indicating that DWR "had requested a breakdown of PG&E's monthly 2001 sales (January through December). PG&E also attached an email dated March 30, 2001 from PG&E confirming transmittal of PG&E's "standard test year 2001 sales forecast" to Navigant, DWR's consultant.

There remained a dispute over whether two additional years' worth of forecast data (for 2002-03) were concurrently provided to DWR, and thus whether the associated municipal bypassed or transferred load incorporated in such data was known by DWR at the time it made its power purchases. At the start of evidentiary hearings, PG&E announced that it had just become aware of an email record sent by PG&E employee Claudia Greif confirming that PG&E's three-year forecast had, in fact, been sent to DWR as of February 12, 2001.⁸ Accordingly, it is now undisputed that a three-year forecast (2001-03), incorporating MDL bypass data from the August 2000 Bypass Report, was provided by PG&E to DWR on February 12, 2001.

As discussed below, we conclude that in light of this now undisputed confirmation of an earlier delivery date of information regarding PG&E's load, modification is required to D.03-07-028. In that decision, we relied on PG&E's erroneous assertion that its sales forecast was not provided to DWR until June 2001-after DWR had contracted for the bulk of its power purchases as a basis to conclude that DWR could not have taken into account the data concerning MDL bypass in its procurement of power. The fact that we now know that DWR received load data from PG&E containing MDL assumptions on February 12, 2001 is relevant to our previous findings in D.03-07-028 concerning the effects of MDL bypass assumptions on DWR procurement. In view of our revised factual findings, we no longer draw the conclusion that PG&E's forecast was received too late to be relevant in DWR's procurement of power. DWR executed a material number of contracts during the period between April 2001 and September 13, 2001. ⁹

PG&E argues that even assuming the "new facts" concerning the Bypass Report are proven to be true and had not been known by the Commission at the time D.03-07-028 was issued, such facts are not inconsistent with the decision. As such, PG&E claims there would be no effect on the "new load" exception, and all new load within its service territory would be served by PG&E and thus still subject to a CRS.

CMUA challenges PG&E's claim that none of the "transferred load" involved loss of "new load," arguing that the affected geographic areas that are annexed will no longer be served at all by PG&E, and that all "new load" in such areas will be served by the municipal utility.

CMUA argues that top the extent that PG&E excluded transferred load from its forecast as identified in the Bypass Report, PG&E necessarily excluded from its forecast the new load in those same geographic areas subject to annexation or condemnation pursuant to the transfer of load.

PG&E made no explicit adjustment to exclude new load in the annexed service territory where the transferred load occurs. Therefore, the load forecasts explicitly identified in the Bypass Report relate exclusively to transferred load. Nonetheless, PG&E witness Keane testified that PG&E would not be expected to serve new load within the same geographical area in which a publicly-owned utility was to assume responsibility for serving transferred load. As CMUA notes, PG&E will have no further right, obligation, or expectation to serve within the geographic areas covered by the transferred load.¹⁰ Thus, any new load within the annexed area to be served by the publicly owned utility would not have been recognized in the DWR forecast. Accordingly, it is reasonable to conclude that PG&E's load forecasts did not incorporate a provision for new load in those geographic areas where existing load was forecasted to be transferred to a publicly-owned utility. Since there was no new load assumed in the forecast for these annexed areas, we can conclude that DWR did not procure power to serve new load in the annexed areas. On that basis, we shall grant a limited exception for new load limited to that occurring within the annexed or condemned geographic areas covered by the transferred load identified in PG&E's Bypass Report. No cost shifting will occur to the extent that any limited new load exeption is confined to the geographic areas that were subject to the transferred load in the Bypass report.

"Transferred load" has relevance, moreover, with respect to the CRS obligation for existing utility load as of February 1, 2001 that departs to municipalities or irrigation districts. Load assumptions with an exclusion of "transferred load" has a bearing on the total MW load as of February 1, 2001 for which DWR was required to procure power. If data from the Bypass Report was supplied to DWR as early as February 12, 2001, identifying "transferred load" that was expected to depart, then DWR procurement requirements did not incorporate this load. Thus, the issue arises as to whether D.03-07-028 should be modified to adjust the MDL obligation for CRS to reflect transferred municipal load for which DWR did not procure power.

To the extent that the "transferred load" does not strictly relate to the rehearing phase (which addresses "new load" issues), but only to transfer of existing utility load as of February 1, 2001, its disposition remains relevant within the context of CMUA's Petition for Modification.

C. Quantification of PG&E "transferred load" Reflected in DWR Forecast

In order to determine the extent of any CRS exclusion for PG&E's transferred load bypass adjustments, we must quantify the amount of the transferred load incorporated in the bypass report, and, in turn, ascertain how it was reflected in the load forecasts utilized by DWR. While the specific figures in the Bypass Report relating to irrigation district and municipalization are not in dispute, parties expressed differing views concerning how the bypass figures should be translated into a CRS exclusion.

CMUA expressed the applicable exclusion in terms of MW capacity which it calculated as about 260 MW assumed to depart over the full 10-year forecast period. As a basis for the 260 MW calculation, CMUA applies a 40% load factor to the MWh sales forecast in the Bypass Report. PG&E argues that any exclusion should not be translated into megawatts using CMUA's assumed load factor calculation. PG&E claims that CMUA's use of a 40% load factor to derive the MW equivalent grossly inflates the load bypass estimate.

PG&E Witness Kuga provided a summary description of the derivation of the numerical values incorporated in the Bypass Report for "transferred load."¹¹ The Bypass Report depicted figures for cumulative bypass per year fro 2000-2004 in annual megawatt-hours (MWH). The two categories in the Bypass Report that relate to MDL were those that identify bypass due to irrigation districts and to municipalization. The reported bypass due to irrigation districts consist of forecasts of bypass to Modesto and Merced IDs, as well as bypass to SSJID and Laguna ID. The forecast of bypass to Modesto and Merced ID was based on a time-trend linear regression using historical data on PG&E's existing customers that had departed to date. The forecast bypass to SSJID and Laguna ID is associated with efforts by the two IDs to condemn PG&E's facilities and to serve existing customers. Accordingly, all of the bypass attributable to IDs is composed of transferred load rather than forecasts of new load.

The forecast bypass due to municipalization is composed of two elements: The first element is based upon PG&E's account services representative's expectations of lost sales of existing PG&E customers associated with future annexations by three existing municipal utilities (i.e., Redding, Roseville, and Lodi). The second element is an "expected value" calculation whereby a probability of 10% is applied to a forecast of lost sales of existing PG&E customers associated with possible condemnation efforts by two potential municipalities (Davis and Brentwood).

The total MDL is broken into two components: (1) transferred load associated with irrigation districts for 2001,2002, and 2003, respectively, and (2) transferred load associated with municipalizations for 2001,2002, and 2003, respectively.¹² This level of bypass for the 2001 through 2003 period was incorporated into the sale forecast provided DWR, as relied upon by DWR in it power purchase activities.

PG&E argues that any exception should be limited by the amount of MDL for 2003 contained in the Bypass Report. As set forth in the table on Exhibit 9 (attached as Appendix 1), the 2003 bypass amounts attributable to irrigation districts were 588 GWh/yr and the amounts attributable to municipalization were 152 GWh/yr. PG&E argues any CRS exception should be further limited by the difference between these amounts and the actual bypass load that had already departed prior to February 1, 2001 (and would thus not be responsible for DWR power charges in any event). In its opening brief, PG&E claims that based on year 2000 actual data, 352 GWh/yr of bypass attributable to irrigation districts had occurred, and 101 GWh/yr of municipalization bypass had occurred.

Based on PG&E's calculations, summarized below, the resulting subtraction leaves only 237 GWh/yr exception applicable to irrigation districts and 51 GWh/yr applicable to municipalization.

PG&E-Proposed Method of Computing Amount of Any CRS Exception:

Source of Bypass	Forecast Sales From Bypass Report (in GWh) 2003 Forecast 2000 Actuals Exception
To Irrigation Districts	588	352	237
Municipalization	152	101	51

PG&E presented its above-described proposal concerning the netting of the 2003 forecast amount against year 2000 actual amounts for the first time in its reply brief. Since PG&E did not present this proposal in its witnesses' testimony or even in its opening brief, there was no opportunity for opposing parties to be heard concerning the merits of PG&E's proposed netting methodology. Particularly in the absence of opportunity to be heard by other parties, we are not persuaded that it is appropriate to reduce the value of the CRS exception by netting the 2003 forecast bypass figures by year 2000 actual figures.

The record evidence indicates that DWR relied upon PG&E's forecast data for each of the years presented in the forecast independently of how the forecasts may have deviated from actual load fluctuations prior to 2001. There is no record evidence indicating that DWR manipulated the amounts forecast by PG&E for any category of load for any given year to adjust for the effects of backward-looking actual data concerning what occurred in the year 2000. In the interests of consistency, therefore, we find no reason to conclude that DWR treated forecasts of MDL bypass any differently than forecasts attributable to any other load category. Thus, even assuming PG&E's claims concerning the actual year 2000 load are numerically accurate, we do not conclude that the actual figures should be applied as a reduction to the otherwise available CRS exception. The exception should be determined based upon the forecast amounts relied upon by DWR, rather than upon actual load fluctuations that were not considered by DWR.

PG&E further states that the transferred load that was subtracted from its load forecast provided to DWR was based only on the 2003 forecast amounts. For the forecast period 2004 through 2010, DWR merely carried forward the absolute MWh amount from 2003 without increasing it by the trended amount from PG&E's regression analysis that was applied to other categories of load.

DWR extended PG&E's forecast to cover the 2004-2010 period by applying annual growth rates based on data in PG&E's FERC Form 714 filing which was independent of PG&E's Bypass Report. Thus, DWR did not incorporate the "trend line" growth rate to MDL included in the Bypass Report. Thus, DWR's extension of PG&E's forecast for the 2004-2010 period merely retained the absolute MWh amount from 2003 without increasing it by the trended amount from PG&E's regression analysis. Thus, we conclude that the CRS exception should likewise carry forward the absolute MWh amount from the 2003 forecast through 2010 without applying any escalation factor.

Merced and Modesto ID, by contrast, calculated the applicable MW exception by applying an annual growth factor through 2010. For the forecast years 2001 through 2005, Merced simply used the bypass figures contained in the Bypass Report. For each year from 2006 through 2010, Merced applied an annual growth increment of 72,871 MWh which equals the growth increment that PG&E assumed in its forecasts covering the 2001-2005 period.

D. Allocation of Transferred Load Exception Among MDL Customers

Parties expressed various proposals concerning how any exception granted for the transferred load should be applied and allocated among the various entities subject to the MDL CRS. Certain parties proposed that any CRS exception should be applied to the specific irrigation districts and municipalities that were identified in the Bypass Report. CMUA generally agrees with such an approach, but only to the extent that it does not foreclose MDL served by publicly-owned utilities not specifically mentioned in the Bypass Report from obtaining MDL exceptions that otherwise would not be used by those entities specifically identified in the Bypass Report.

Merced and Modesto point out that the Bypass Report identifies a single aggregate of bypass applicable to both irrigation districts, but does not delineate how much applies to each entity. Merced and Modesto, in their opening briefs, offered a proposed allocation between the two that is mutually agreeable between them. We find this proposed division of load reasonable and shall adopt it.

If the Commission decides to provide a limited CRS exception based on the MDL bypass forecast of transferred load, PG&E indicates it would not object to allowing new as well as transferred load to come under the limited CRS exception. Even so, PG&E denies in principle that its transferred load bypass forecast had any effect on DWR forecasts relating to new load. PG&E also believes that further proceedings would be needed to determine appropriate tools to grant, track, and otherwise administer the application of any limited CRS exception among the publicly-owned utilities and irrigation districts that would be eligible for the exception.

We direct that the CRS exception for PG&E's transferred load be allocated in the following manner. As a first priority, the CRS exclusion shall be made available for use by those municipalities and irrigation districts that were specifically identified in the Bypass Report. In the case of Merced and Modesto IDs, we shall divide the aggregate amount allocated to them in the Bypass Report in accordance with proportionate shares to which they have mutually agreed, as indicated in their briefs. To the extent that one of those specifically identified entities does not utilize their allotted exception, the exception shall be made available to other MDL entities on a first-come, first-served basis. The priority for use of the DWR power charge exception for the entities named in the Bypass Report shall be determined on an annual basis utilizing the applicable amounts shown in the Bypass Report. Other entities authorized to utilize any portion of the CRS exception amount shall do so on an annual basis to the extent an unutilized portion is available. For determining the assignment of any unused portion of the allotted exception to such other MDL entities under the Bypass Report, priority shall first be given to load transferring specifically from PG&E bundled service. The MDL entities eligible to apply for the exception also must have been in existence and serving customers as of July 10, 2003, the date of
D.03-07-028. To the extent further details require resolution to implement the exclusion, those details shall be addressed in the billing and collection implementation phase of this proceeding.

In its comments on this decision, PG&E states that MDL is subject to Energy Recovery Bond charges imposed pursuant to Senate Bill (SB) 772, except for any new load that is excluded from paying CRS pursuant to this rehearing proceeding. The applicability and effects of SB772 is beyond the scope of this proceeding, and we make no findings in this order concerning it. We are issuing a separate order today dealing with the applicability of Energy Recovery Bond charges. To the extent that there are any remaining implementation issues relating to MDL cost responsibility for Energy Recovery Bond charges, parties may raise them in the billing and collection phase or other appropriate forum.

E. Extent to Which DWR Procurement was Influenced by Forecast Data

Parties dispute whether, or to what extent, DWR was influenced in its procurement actions by PG&E load forecast data, including the MDL bypass assumptions incorporated therein. PG&E argues that even assuming DWR received its forecast as early as February 2004, DWR's purchasing behavior was not materially impacted by the MDL bypass data. PG&E challenges the premise that commitments DWR made in the spring and summer of 2001 were closely tied to the forecasts that DWR received from the utilities. DWR witness McDonald testified that the "net short" forecasts DWR was working with were accurate to within no more than 10 percent.¹³ PG&E thus argues that this uncertainty "already swamps" any estimate of MDL bypass that was implicitly contained in the forecast PG&E provided to DWR. Further, DWR was continually adjusting its forecasts during this time frame.¹⁴ PG&E further argues that DWR's primary focus was on obtaining commitments for power for the summer of 2001, and that longer-term commitments arose out of that focus, as suppliers were unwilling to provide power at an acceptable price to DWR unless DWR was willing to make longer-term commitments. ¹⁵ PG&E claims that the longer-term commitments were apparently an unavoidable consequence of signing up power for the summer of 2001 irrespective of the MDL bypass assumptions.

PG&E thus argues that there was no "careful correlation" between the multi-year forecasts DWR received in early 2001 and the long-term commitments DWR ultimately made. As such, PG&E claims it is not reasonable to use those forecasts to excuse a portion of MDL customers from paying the CRS. Regardless of the fact that there was an estimate of transferred MDL implicit in its forecast given to DWR, PG&E argues that these customers should be obligated to pay the CRS.

We acknowledge the lack of a precise matching of the load forecasts utilized by DWR with the specific quantities of power procured for various categories of customers. Such lack of precision, however, does not justify completely disregarding the MDL bypass assumptions in considering the applicability of CRS. Our factual inquiry has to do with what categories of load were included or excluded from the forecast. Once we determine that a category was excluded, that means that DWR was not procuring power on behalf of the excluded category. Thus, irrespective of how imprecise the forecasts were or how well the supplies matched the demand forecasts, that matching exercise only had relevance with respect to those categories of load for which DWR was procuring power, as defined by the forecast provided to DWR. No matter how imprecise the load forecasts were, we are not persuaded that the imprecision proves that the transferred load was never included in the first place as a load category for which DWR was procuring power.

We applied a similar principle earlier in this proceeding in determining the applicability of a CRS for the other major category of departing load, namely customer generation. In that instance, there was a similar lack of precision in matching forecasts and procurement in the context of our considering whether to grant a CRS exclusion of any portion of Customer Generation Departing Load.¹⁶ Yet, that lack of precision did not prevent us from determining a reasonable approximation of affected load and adopting a CRS exclusion for Customer Generation Departing Load. On that basis, in D.03-04-030, we determined that the first 3000 MW of customer generation departing load to leave the IOUs' systems would not be required to pay the DWR ongoing power charge portion of the CRS. As we stated therein:

"It is clear that DWR, when negotiating long-term power contracts, assumed that a certain amount of customer generation departing load would occur every year and therefore did not procure long-term power for that portion of the load. In fact, such an assumption is based on common sense, since utilities have always faced departing load in various forms, including that caused by an economic downturn, improvements in energy efficiency and building codes, as well as installation of self-generation systems."¹⁷

Thus, while we concluded in D.03-04-030 that the Navigant model assumptions were not precise enough to set a year-by-year cap for purposes of excluding a portion of Customer Generation from the CRS, the overall rationale behind setting some form of cap still made sense as a way to avoid cost shifting. Thus, given the lack of precision to support a year-by-year cap, we adopted one overall cap for Customer Generation of 3,000 MW, representing an approximate cumulative forecast assumption over a 10-year period. (D.03-04-030 at 53.)

In similar fashion to Customer Generation, the lack of precision in the matching of forecasts and procurement is not a valid reason simply to ignore altogether the fact that the MDL bypass was never included in the load forecasts relied upon by DWR. Thus, whatever imprecision is present in power quantities that were procured, DWR was not acting to procure power for MDL load that was explicitly subtracted from the PG&E forecasts provided for use by DWR.

F. Timing of DWR Procurement Relative to Date of Receipt of Forecast

PG&E argues that even if the forecast was received as early as February 2001, DWR had already begun to enter into contractual commitments by then. PG&E argues that DWR had made a significant number of contractual commitments by the middle of March 2001, through negotiations that could not happen instantaneously.¹⁸ Therefore, PG&E minimizes the weight that its forecasts could have had on DWR's commitments, arguing that no strong cause-and-effect has been shown between the receipt of the forecast in February 2001 and consequent actions limiting procurement in response to MDL bypass. PG&E proposes that any CRS exception that might be granted be adjusted to account for contractual commitments made by DWR prior to receiving and relying upon the sales forecast from PG&E.

We find that although DWR had made certain limited contract commitments prior to February 12, 2001, the vast majority of commitments were finalized after that date. In Exhibit 18, in reference to the PG&E forecast, DWR stated that it "did rely on the information contained within this sales forecast in making decisions regarding power purchases."¹⁹ DWR Witness McDonald testified that "DWR started using this forecast in the mid-February timeframe."²⁰ The forecast was used by DWR in "the development of forecasts of net short..."²¹ The net short was used "to help the contracting team inform themselves about how much contract they should be entering into."²²

The fact that DWR had already executed a minimal number of power purchase agreements prior to receipt of the sales forecast data from PG&E (and possibly letter agreements with respect to others) does not negate the fact the DWR executed the vast bulk of its contractual commitments after February 12, 2001. Thus, we find nothing to convince us that the delivery of the load forecast on February 12, 2001 was too late to form the basis for the bulk of DWR procurement. Thus, we find nothing to preclude an exception from the ongoing power charge component of the CRS for transferred MDL reflected in those sales forecasts based on a claim that February 12, 2001 was too late in the procurement process.

An indicated by Exhibit 73 from the October 2002 hearings, only four contracts were executed before February 12, 2001, with a total capacity of 2150 MW. Three of the four contracts have already expired. Thus, the evidence indicates that very little, if any, long term DWR power purchases occurred prior to receipt of the PG&E load forecast on February 12, 2001.

Moreover, we made similar observations concerning the timing of contract purchases in the context of Customer (or Distributed) Generation in an earlier departing load phase of this proceeding. For example, preparation of the document entitled "Forecast for Distributed Generation in California,"²³ which DWR utilized to adjust the sales forecasts upon which its power purchase activities were based, was started in March 2001 and finalized in April 2001.²⁴ Although the forecast of how much load would depart the IOUs' systems for self generation was not finalized until April 2001, the Commission nevertheless provided an exception for this forecasted amount because DWR had not yet entered into the bulk of its power purchase contracts as of April 2001.²⁵ The Commission determined that "granting exceptions to certain portions of the CRS for customer generation up to 3,000 MW will not result in any cost shifting among customers, since costs for those MW were not incurred by DWR."²⁶ Thus, a similar finding applies in the context of MDL. Accordingly, we conclude that receipt of the forecast in February 2001 was early enough to have a material influence on DWR in terms of its procurement. Thus, granting CRS exceptions in recognition of the PG&E transferred load will not result in cost shifting among customers, since costs for that load were not incurred by DWR

G. Significance of Whether DWR, Itself, Makes the Adjustment

PG&E also argues that the granting of a CRS exceptions for MDL would be inconsistent with the Commission's treatment of customer generation departing load (CGDL). PG&E argues that the Commission granted exceptions to CGDL based on the forecasts of such load made by DWR, and not the utilities. Since DWR made no corresponding MDL forecast, PG&E argues that consistent treatment would disqualify MDL from an exception.

PG&E argues that the basis for the exclusion granted to CG DL in the form of a 3,000 MW cap is not analogous to the proposed basis for the CRS exclusion involving MDL transferred load. PG&E argues that in granting CG DL certain exceptions from the DWR power charge in D.03-04-030, the Commission was relying on "the DWR/Navigant model assumptions to set one overall cap of 3,000 MW (the approximate cumulative total (rounded) of DWR's annual assumptions over ten years)." (D.03-04-030, p. 54.) PG&E emphasizes that this 3,000 MW exception was based on DWR's separate, explicit specific reduction to its load forecasts to reflect anticipated load loss due to increased CGDL, above and beyond what might have been embedded in the utilities own forecasts.

PG&E quotes the MDL Decision where we stated: "While DWR actually forecasted a specific amount of departing load associated with new customer generation, it made no corresponding MDL forecast. The amount of customer generation departing load proposed to be exempt from the [power charge], by contrast, is directly tied to this DWR forecast." ²⁷

Thus, to the extent that any analogy is to be drawn with the Commission's decision on CGDL, PG&E claims the relevant inquiry in this case is whether DWR itself explicitly reduced the load forecasts it was using to as it made its power purchases for anticipated load loss to publicly- owned utilities.

We disagree with PG&E's claim that the only relevant inquiry is whether DWR, itself, (as opposed to PG&E) performed the forecast of a specific amount of MDL bypass. In the context of whether cost shifting is involved, the relevant inquiry is what the forecast, itself, includes or excludes. The particular entity (e.g., DWR or PG&E) that instigated the exception is merely incidental. The particular identity of the entity that makes the MDL bypass adjustment makes no difference in determining whether an adjustment exists, its magnitude, or how the resulting load forecast was utilized by DWR in procuring power.

In our earlier statement from D.03-07-028, quoted above by PG&E, we had already drawn the conclusion that there was no separate forecast of MDL bypass that had been specifically forecast by PG&E and provided to DWR. It was in that context that we focused on the separate forecasting actions performed by DWR as the only remaining relevant inquiry as to whether MDL had been taken into account in the load forecasts. Yet, since the subsequent additional evidence produced in this phase of the proceeding has proven that PG&E did, in fact, independently incorporate MDL bypass assumptions as part of the load forecast upon which DWR relied in making procurement decisions. In view of this additional evidence, therefore, our focus is not merely on the actions of DWR directly, but also on the actions of the IOU in preparing the forecast.

In terms of determining whether there is any cost shifting, the relevant issue is whether a particular component of load was included or excluded from the total load forecast relied upon by DWR in procuring power. The question of who made the adjustment to exclude such load--whether it was made by the IOU initially, or by DWR subsequent to receiving the IOU forecast-- is an incidental detail that has no bearing on the essential question of whether DWR procured power on behalf of such load.

Thus, we conclude that the load forecast provided to DWR by PG&E did, in fact, incorporate explicit assumptions concerning the bypass of MDL. Although we acknowledge that DWR did not make any separate subsequent adjustments to the forecast to reflect MDL bypass, it likewise did not add back any of the MDL bypass that had already been taken out of the load forecast by PG&E. Thus, the load forecasts utilized by DWR did not include the MDL bypass that had been incorporated from the PG&E 2000 Bypass Report.

While DWR did not make a specific adjustment to PG&E's sales forecast to account for municipal departing load as it had done with respect to customer generation, the adjustment that had already been made by PG&E carried through to the figures relied upon by DWR. Thus, in terms of the cause-and-effect relationship between load forecast assumptions and procurement decisions, DWR's procurement behavior with respect to MDL bypass assumptions was similar in effect as for Customer Generation bypass assumptions.

H. Analogies to the U.S. Navy Load CRS Requirement

PG&E also cites the Commission's action imposing a CRS requirement on U.S. Navy load as adopted in D.03-05-036, as providing an analogous situation to the MDL CRS issue before us here. In D.03-05-036, the Commission addressed whether the United States Navy should be granted an exception from the DA CRS for 80 MW of load that it obtained through a special contract with the Western Area Power administration (WAPA). In that decision, the Commission rejected the Navy's contention that it should receive a CRS exception, even though SDG&E's witness "testified that DWR did not buy contract power to serve the 80 MW of Navy Load...and was expressly excluded from the SDG&E load requirements provided to DWR."²⁸ The Commission justified its holding as follows:

"SDG&E points out that [Navy] was on bundled service as of February 1, 2001. Under the provisions of D.02-11-022, the Navy thus is obligated to pay the DA CRS on the same basis as other customers that meet that criterion.... Although the Navy procured power under the 80 MW independently of DWR, the power did not begin to flow under the special contract until after February 1, 2001....For at least some period on and after February 1, 2001 up until it began to be served under the special contract, the Navy would have been subject to bundled procurement for meeting its load demand. To the extent that DWR procured the net short for SDG&E bundled load during the period prior to April 1, 2001, some bundled power would have flowed to the Navy."²⁹

PG&E focuses on the fact that we applied the February 1, 2001 cut-off date for DWR cost responsibility, as mandated in AB 117, even though that the IOU forecast obtained by DWR reflected the loss of Navy load. PG&E claims that the manner in which SDG&E's and DWR's forecasts treated the Navy load are identical to the how PG&E's and DWR's forecasts treated the transferred MDL reflected in PG&E's August 2000 Bypass Report: PG&E argues that in both situations, the utility netted out some "transferred" load prior to providing its forecast to DWR, and in both situations DWR made no further explicit adjustments to the utility's forecast. PG&E thus argues that it would be consistent for the Commission to conclude that transferred MDL customers departing after February 1, 2001, owe the CRS, irrespective of the MDL "transferred" load that was subtracted from the load forecast provided to DWR.

We disagree that the U.S. Navy load treatment is analogous to the situation involved with the MDL transferred load. There are a number of differences between the two situations. For example, as explained in D.03-05-036, DWR/Navigant included the 80 MW in its modeling of SDG&E net short requirements, despite the fact that SDG&E had informed DWR/Navigant that this load should not be included because the Navy was procuring the load through its own separate contract. By contrast, DWR/Navigant expressly did not include any of the MDL transferred load in its forecasts or modeling thereof. Likewise, neither did it procure power on behalf of such PG&E load. Thus, DWR treated the load differently between the two situations with respect to whether it was included or excluded in forecasts and modeling calculations.

Moreover, the considerations involved here are based upon the principles we established in D.03-07-028 for the applicability of a CRS to MDL, where we expressly distinguished departing load from direct access, stating: "The dispute over the treatment of "new load" in the context of municipal customers raises issues different from those facing us in the DA phase of this proceeding."³⁰ We expressly considered the issue of whether the DWR load forecast incorporated a provision for MDL in concluding that a CRS exception should apply to new load. Thus, in D.03-07-028, we established a conceptual cause-and-effect link between the load forecast and the applicability of the MDL CRS to particular subcategories of MDL. While the rehearing order directed that a further factual record be developed, it did not disturb that underlying principle of cause-and-effect established in D.03-07-028, especially in the context of cost-shifting. We expressly stated that whether or not DWR procured on behalf of a segment of departing load was a relevant consideration in whether a CRS exception provision should apply.

In comments on the Proposed Decision, PG&E claims that granting a CRS exclusion for the transferred load would violate the cost-shifting prohibitions set forth in AB 117. PG&E argues that because the exclusion of transferred load from the MDL CRS would cause other groups of customers to pay more, an exclusion of transferred load constitutes impermissible cost shifting under AB 117. In making such a claim, however, PG&E misconstrues the requirements of AB 117. Cost shifting that is prohibited under AB117 specifically has to do with costs being imposed on one category of load where the costs were incurred on behalf of a different category of load. In the case of the MDL transferred load that was identified in the Bypass Report, however, no costs were incurred by DWR because the transferred load was never included within the load forecasts upon which DWR relied for purposes of procuring power. Since no DWR costs were ever incurred on behalf of the transferred load, there are now DWR costs to be shifted as a result of excluding the transferred load from the DWR power charge. Accordingly, there is no violation of AB 117 as a result of the transferred load exclusion. Rather, it would be a violation of AB 117 if transferred load was not excluded from the CRS.

⁸ See Ex. 5.

⁹ McDonald Testimony, Tr. 1474.

¹⁰ See Tr. 2488-2489, (PG&E/Keane).

¹¹ Ex. 3 (Kuga) at DMK-2-3

¹² Tr. 2551 (PG&E/Keane).

¹³ . See Tr. 2676-79 (DWR/McDonald).

¹⁴ See Tr. 2593 (DWR/McDonald).

¹⁵ See Tr. 2610-12/DWR McDonald.

¹⁶ DWR/McDonald RT 2688: 22-28.

¹⁷ D.03-04-030, mimeo. at p. 54.

¹⁸ . See, Tr. 2608 (DWR/McDonald).

¹⁹ Exh. 18, pg. 3, Response 1(g). As clarified by PG&E witness Keane, the forecast in question was a sales forecast (i.e., stated in MWh), not a load forecast (stated in MW capacity). 21 Tr/ 2497:7-9 (PG&E/Keane).

²⁰ 22 Tr. 2588:24-25 (DWR/McDonald)

²¹ 22 Tr. 2631:9-12 (DWR/McDonald)

²² 22 Tr. 26321:13-23 (DWR/McDonald)

²³ Exh. 72 ("Forecast for Distributed Generation in California").

²⁴ Tr. Vol. 12 (DWR/McDonald), p. 1473, line 10 to p. 1474, line 10 (testifying to the time frame over which the forecast for customer generation was prepared).

²⁵ Tr. Vol. 12 (DWR/McDonald), p. 1484 (testifying to the fact that DWR negotiated and executed a material number of contracts between April 2001 and September 13, 2001).

²⁶ See D.03-04-030, p. 64 (slip op.).

²⁷ D.03-07-028, pp. 36-37 (slip op.).

²⁸ D.03-05-036, p. 4 (slip op.).

²⁹ D.03-05-036, p. 7 (slip op).

³⁰ D.03-07-028, p. 57 (slip op.).