In the conventional ARA, forecast expenses are escalated to reach the attrition year revenue requirement. In this attrition proceeding, ORA notes that PG&E has spent less than authorized during 1999 and 2000 in all expense categories except Administrative and General (A&G), while showing no decline in safety and reliability. Concluding from this that PG&E does not need an increase in its expense categories, ORA recommends that attrition for 2001 be based on 1999 recorded figures (except for A&G expense). In ORA's opinion, basing attrition on recorded figures rather than authorized in areas experiencing significant underspending is consistent with the Commission's expressed desire to use this proceeding to assure itself and the public that the GRC decision "got it right."
ORA states that there are serious problems with PG&E's recorded amounts for both A&G Account 920 and the Account 923. The former includes Performance Incentive Plan (PIP) costs that were expressly disallowed by the GRC decision and assigned to shareholders. The latter includes allocations from the holding company to the utility that are over three times the amount found reasonable by the Commission. The difference exceeds $60 million. ORA asserts that PG&E should not be rewarded with an attrition increase to the extent it is based on costs that were either disallowed on policy grounds or are unreasonable on their face. In short, PG&E should not be rewarded for this type of overspending. Adoption of ORA's proposal for calculating expense attrition results in a $12 million decrease in expenses.
PG&E argues that ORA's position is directly contrary to the standard ARA mechanism - forecasted expenses escalated by formula - and is contrary to D.00-02-046. PG&E contends that Ordering Paragraph 15 is quite clear about the only two modifications to the adopted ARA mechanism the Commission will consider: i.e., "the results of the 1999 capital spending audit and to recognize amounts recorded in the VMBA." (Id., at 545.) ORA's efforts to circumvent the implementation of the adopted ARA mechanism by proposing a new ARA mechanism where the various expense components would be based on the lower of authorized or recorded expenses should be rejected as being both unfair and inconsistent with D.00-02-046. As PG&E points out, had ORA recommended that recorded numbers be used for all of the expense components, PG&E's ARA request would have increased.
D.00-02-046 tells us to take a closer look at PG&E's attrition filing than we would normally do had the filing been by way of an advice letter. That we have done. There have been four days of public hearing, 30 exhibits, and briefs by the active parties. But taking a closer look does not necessarily mean changing past practices. ORA would have us change from forecast expenses to recorded expenses when determining attrition relief, with an exception for higher recorded expenses with which ORA disagrees. For those expenses, we are asked to use lower forecast numbers. This approach is contrary to past ARA practice, takes on the appearance of a general rate case, and leads to inconsistent results as each party picks and chooses the numbers it is most comfortable with. Adopting ORA's expense calculations would still leave PG&E with a revenue shortfall to achieve its 9.12% authorized rate of return. ORA's primary recommendation would result in PG&E earning a rate of return of 8.22% in 2001. (Appendix A, Column H.)3
We conclude that in an ARA application filed in response to authority granted in a GRC, the better course is to use forecast expense and rate base unless the GRC decision specifies a different method. We do not disapprove of the use of recorded numbers or adjusted numbers to determine the need for attrition; but the history of this application and the order in D.00-02-046 do not support the proposal of ORA.
3 Regardless of how we decide the expense dispute ORA would still recommend no increase because of the perceived effect PG&E's bankruptcy filing has on PG&E's operations.