11. Electric Reliability Incentives for SDG&E

Decision 04-01-007 dated January 8, 2004 granted a petition to modify D.01-10-030⁸⁰ to extend the existing 2003 performance indicators for 2004 but specifically deferred to this proceeding the question of any financial incentives for 2004. This decision addresses whether or not any incentives should apply to 2004.

The parties identify 11 issues for electric reliability. We must determine whether or not to adopt the various mechanisms, and the right measurement targets, as performance incentives for SDG&E's electric distribution operations. The only reason to adopt the incentives would be to achieve better service over time than would occur without the incentives. We will resolve the following issues:

10. System Average Interruption Duration Index (SAIDI) target;

11. System Average Interruption Frequency Index (SAIFI) target;

12. Momentary Average Interruption Frequency Index (MAIFI) target;

13. Updating targets;

14. Maximum reliability reward or penalty;

15. Adoption of a Reliability Policy;

16. New indicators;

17. Cost reporting for electric reliability projects;

18. Adoption of a Reliability Standard Practice;

19. Benchmarking; and

20. Additions to electric reliability reporting requirements.

The intervenors propose higher standards for the previously existing incentives, SAIDI, SAIFI and MAFI (collectively, Electric Reliability Incentives) than requested by SDG&E, and Aglet opposes all reward and penalty mechanisms. In addition to the target, parties disagree on whether to have a deadband (a range of no penalty/reward) and how large a liveband (upper and lower limit to penalty/reward) to have. We also assume that the parties to the partial settlements in Phase 1 consistently litigated Phase 2 on the assumption of the Commission adopting the settlements. To the extent necessary, this decision will consider the outcome adopted in Phase 1 when adopting performance incentives. For example, the rate of cable outages was a significant factual dispute - a fundamental premise - in SDG&E's rebuttal Exhibit 165. As a result, the adopted forecasts for capital expenditures in Phase 1 related to cable maintenance and replacement, and other reliability-related expenditures, has a direct bearing on identifying the appropriate targets for the Electric Reliability Incentives.⁸¹ No party proposes to separate the measurements for underground cable and
non-cable performance.

Some of the issues will be discussed separately but they are inter-related; for example, whether we update SAIDI and others annually, or only once for the test year and post-test year period, may well affect the appropriate target for 2004 and 2005.

a. System Average Interruption Duration Index (SAIDI)

Parties propose an array of SAIDI goals and Aglet opposes the adoption of any incentive mechanism.

SAIDI Proposals
	SDG&E	ORA	CCUE	TURN	Aglet
Target	71	64	69	63	None
Deadband	none	7.7	none	None
Liveband	+/- 15	+/- 15	+/- 15	None
Reward/Penalty	$250,000	$125,000	$250,000	None
		No Reward
Range - Millions	+/- $3.75	-$1.875	+/- $3.75

b. System Average Interruption Frequency Index (SAIFI)

SAIFI Proposals
	SDG&E	ORA	CCUE	TURN	Aglet
Target	0.80	0.68	0.68	0.66	none
Deadband	none	0.07	none	none
Liveband	+/- 0.15	+/- 0.15	+/- 0.15	none
Reward/Penalty	$250,000	$125,000	$250,000	none
		No Reward
Range - Millions	+/- $3.75	-$1.875	+/- $3.75

c. Momentary Average Interruption Frequency Index (MAIFI)

MAIFI Proposals
	SDG&E	ORA	CCUE	TURN	Aglet
Target	0.97	0.77	0.77	0.77	none
Deadband	none	0.07	none	none
Liveband	+/- 0.30	+/- 0.30	+/- 0.15	none
Reward/Penalty	$50,000	$25,000	$50,000	none
		No Reward
Range - Millions	+/- $1.0	-$0.5	+/- $1.0

d. Deadbands

Only ORA proposes a deadband for the three Electric Reliability Incentives. A deadband is a range around the target where no incentive penalty or reward is assessed. It is attractive because the targets are only a reasonable estimate - it is highly unlikely that SDG&E could directly influence the precise outcome and so it could see a penalty or reward as a matter of chance. The ORA deadband narrows the range of penalties or rewards because ORA did not widen the overall liveband. One benefit of a deadband is that minor random variances in performance do not trigger an undeserved penalty or reward - undeserved in the sense that SDG&E's actions were not the likely cause of the variance from the target. ORA's deadbands are too large, in that there is no evidence in the record that would support the proposed range as likely to encompass the random influences compared to SDG&E's deliberate actions that affect the final result. Therefore we will adopt a deadband, but smaller than proposed by ORA, to eliminate the more random effect on penalties or rewards. ORA also structured its deadband for a penalty-only recommendation and did not expect that it would be exceeded on a regular basis. A narrower deadband will be more likely to invoke penalties or rewards and act as an incentive to SoGalGas and SDG&E.

e. Livebands

ORA and CCUE agree with SDG&E on two of the liveband sizes, CCUE would halve the MAIFI liveband. The justification for a liveband is to put an outer limit on both a penalty or a reward in the event of extraordinary results, because, again, SDG&E has little direct control on specific outages. The purpose of an incentive is to ensure proper attention, including expenditures on maintenance and capital improvements, is paid to electric reliability. We will adopt the liveband ranges as proposed:

+/-15 minutes for SAIDI,

+/-0.15 for SAIFI, and

+/-0.30 for MAIFI.

We agree with applicants that these livebands are large enough to provide an incentive without providing excessive rewards or penalties.

f. Reward and Penalty Targets

SDG&E proposes that the SAIDI, for example, would accrue a reward/penalty of $250,000 for every one-minute increment from the proposed target, up to a maximum +/- $3.750 million. With a proposed target of 71 minutes, the reward/penalty range would be from 56 minutes (good) to 86 minutes (bad). SDG&E proposes the10 most recent years' annual average of 71 minutes, rounded from 71.19.⁸² The problem with a 10-year average, especially when there have been incentives in place, is that any progress achieved over that time is diluted by earlier years' results. A fundamental principle underlying the utilization of incentives is that they lead to improvements, otherwise we would not impose the cost of the incentives on consumers.

ORA proposes using the most recent five-year average and a "rolling" average adjusted each year. The mechanical details would be dealt with in an advice letter.⁸³ We do not expect this rate cycle to be a long one, with the next rate case for both SoCalGas and SDG&E to have a TY 2008. We will not require an annual target adjustment, but we will use the most current five-year average as a part of the correct base for setting the targets.

ORA also proposes a penalty-only approach, and we find this to be inappropriate. The concept of an incentive mechanism, based only on a penalty, is not an incentive. ORA provides its perspective on "value of service"⁸⁴ that essentially concludes commercial customers face significant financial hardships from any outage and place a high value on avoiding outages. ORA argues that a "penalty-only structure will protect ratepayers from paying twice for the same performance, and protect the company from paying for outages beyond its control."⁸⁵ However, a reward or penalty cannot compensate or penalize SDG&E for the full cost expended or avoided for achieving the goals. The payments are rewards or penalties for a level of special performance, not the sole reimbursement for improving service reliability. In fact, if SDG&E were to consistently fail to spend the revenues provided in rates on the reliability projects adopted in the test year forecast, and implicit in the post-test years, then we can pursue other sanctions for its failure to meet its obligation to serve customers safely and reliably.

g. Reliability Policy

ORA proposes that the "Commission should adopt a policy to consistently value reliability which applies both to the determination of cost of service/revenue requirement and to penalties or rewards associated with reliability."⁸⁶ ORA does not make a specific policy proposal, although it does argue the allocation of the incentives is skewed between residential and commercial ratepayers. ORA concludes from its analysis that the incentive mechanism should be a penalty-only mechanism and the penalty amount should be half the size proposed by SDG&E.⁸⁷ If we accept ORA's analysis that commercial and industrial customers receive 97% of the benefit of a reduction in service interruptions, but only 46% of the costs, then we should consider a reallocation of the costs rather than a reduction in the penalty and an elimination of the reward.

Aglet argues that SDG&E has not met its burden of proof to show that the Electric Reliability Incentives are necessary or reasonable: "Despite approximately ten years of utility experience with PBR mechanisms in California, the applicants present no study on the causality between performance and financial incentives."⁸⁸ Aglet states⁸⁹ that SDG&E's proposed settlement in Phase 1 would allow the utility to meet its goal to maintain current levels of reliability; incentives are not necessary for safe and reliable service, and SDG&E has never studied the effectiveness of existing incentives. Aglet also argues that the value of service studies are out of date; that Phase 1 was SDG&E's opportunity to request ratepayer funding of cost-effective service quality improvements; and that management have salary incentives that ensure they are attentive to reliability.

Aglet also argues that SDG&E failed to prove that reliability is affected by the incentives. In fact, SDG&E argues that cable failures are rising and are the nature of the beast, at least the early underground cable installations, so Aglet raises a credible argument that the proposed incentives are not appropriate, and the "results" are as likely to be coincidental, citing the safety improvement at SoCalGas before a safety incentive was adopted.⁹⁰

Aglet concludes:

"Aglet believes that rewards gained through performance incentives like those proposed by SoCalGas and SDG&E depend more on the design of incentive formulas than incremental improvements in performance attributable to the incentives. The testimony in this proceeding focuses on the details of complex formulas, while giving little attention to more important issues of causality and overall company efficiency. The scorecard for SDG&E during the period from 1997 through 2002 shows 24 wins, 4 losses and 2 no-decisions. (Citations omitted.) SDG&E earned incentive rewards in 24 of 30 opportunities, and suffered penalties in only four. Based on review of recorded performance, Aglet cannot tell whether the incentives worked as promised or the financial outcomes merely reflect success in gaming the formulas."⁹¹

We are keenly aware of the SDG&E's record and we are not prepared to terminate the Electric Reliability Indicators without a more thorough analysis, because reliability has generally improved while incentives have been in effect. We do intend to adopt reasonable but challenging targets and not the 10-year status quo proposed by the applicant.

h. Adopted Electric Reliability Incentives.

We believe that ORA's proposed use of a five-year average, without the burden of annual adjustments, is the most reasonable base to set the Electric Reliability Incentives, but we are concerned that the use of averages does not sufficiently drive SDG&E to improve performance. Thus we believe that a small stretch factor, similar in concept to the stretch factor in the Base Margin escalation process, would be beneficial. It is also clear that SDG&E's control over reliability is not perfect and is not total; we believe that deadbands protect against unwarranted rewards or penalties.

A further ratepayer protection could be to lengthen the measurement period to two years - for example, if the measured SAIDI for 2005 and 2006 were 64 and 60 minutes respectively, the average would be 62. If the annual average target was set at 63, the effect would be that SDG&E beat the target by one minute. Any reward would need to be twice the annual level too. Ignoring any deadband, this example would allow the good year to partially offset the bad one, for a net incentive over a longer period of time. The record is clear that reliability improvement is a long-term exercise, dependent upon consistent maintenance and timely capital expenditures. Annual measurement artificially distorts the long-term commitment necessary for reliability improvements. We will not adopt a multi-year evaluation now without allowing the parties to consider its effects. We direct SDG&E to address this proposal in the next performance incentive proceeding. We also direct SDG&E to provide a detailed analysis that responds to Aglet's concerns. SDG&E must demonstrate that the incentives contribute to improving performance.

	Adopted Reliability Incentives
		SAIDI	SAIFI	MAIFI
	5 Year Base	64	0.68	0.77
	Stretch Factor	1	0.01	0.01
	Target	63	0.67	0.76
	Deadband	+/-2	+/-0.02	+/-0.02
	Liveband	+/-15	+/- 0.15	+/- 0.30
	Reward/Penalty	$250,000	$250,000	$50,000
	Range - Millions	$3.75	$3.75	$1.00

⁸⁰ Application 98-01-014 of San Diego Gas & Electric Company ("SDG&E") for Authority to Implement a Distribution Performance-Based Ratemaking Mechanism, Application 95-06-002 of Southern California Gas Company To Adopt PBR for Base Rates to be Effective January 1, 1997, and
Application 96-10-038 of Pacific Enterprises, Enova Corporation, Mineral Energy Company, B Mineral Energy Sub and G Mineral Energy Sub for Approval of a Plan of Merger of Pacific Enterprises and Enova Corporation with and into B Energy Sub and G Energy Sub, the Wholly-Owned Subsidiaries of a Newly Created Holding Company, Mineral Energy Company.

⁸¹ See Ex. 165, pp. CW-7 through CW-11, amongst other instances.

⁸² Ex. 159, p. CW-13, and ORA Opening Litigation Brief, pp. 24-25.

⁸³ Ex. 333, pp. 6-19 and 6-20.

⁸⁴ Ex. 333, pp. 6-25 to 6-27.

⁸⁵ ORA Opening brief, p. 23.

⁸⁶ Ex. 333, p. 6-6, lines 7-10.

⁸⁷ "Given the difficulty of determining any value that would be equitable to all classes, ORA recommends that the Commission adopt the revealed preference penalties of $125,000 per SAIDI and SAIFI unit and $25,000 per MAIFI unit as the most reasonably balanced on this record." ORA Opening Litigation Brief, p. 32.

⁸⁸ Aglet Opening Litigation Brief, p. 19 ff.

⁸⁹ Aglet cites Geier, 27 RT 2404:4, 2406:3-7, 2411:6; Geier, 27 RT 2455:11-17; and Petersilia, 30 RT 2726:10-12; Little, 30 RT 2809:2-7; Geier, 27 RT 2455:25-27, respectively.

⁹⁰ Aglet Opening Litigation Brief, p. 22.

⁹¹ Aglet Opening Litigation Brief, p. 23.