Discussion

We begin by pointing out that, contrary to some suggestions in Universal's briefs, this proceeding is not a tort case, but a tariff interpretation case in which two potentially conflicting provisions in an Edison tariff rule must be reconciled. SCE Rule 14 requires Edison to "exercise reasonable diligence to furnish/deliver a continuous and sufficient supply of electricity to its customers and to avoid any shortage or interruption of delivery thereof." However, the very next sentence of the rule states:

"[Edison] cannot, however, guarantee a continuous or sufficient supply [of electricity] or freedom from interruption. SCE will not be liable for interruption or shortage of supply, nor for any loss or damage occasioned thereby, if such interruption or shortage results from any cause not within its control."

Although Universal acknowledges that there is an inherent tension between reliability and price, it argues that Edison's power purchasing practices in the day-ahead market for June 27 did not meet the test for "reasonable diligence," and were the proximate cause of the Stage 2 alert called by the ISO on that day. After noting on page 6 of its Reply Brief that "over 2000 MWh were available at a market clearing price of $1,500 per MWh and additional supplies would have been available at prices above $1,500 per MWh" in the day-ahead market for the peak hour on June 27,¹¹ Universal concludes:

"In the abstract, it may be debatable where to find the balance between reliability and price, but this proceeding is not about where that balance might be. In the real world that issue has been resolved by Edison's Tariff Rule 14.

"Edison's Rule 14 obligates Edison to exercise reasonable diligence to furnish a continuous supply of power. That obligation is not couched in terms of price. Additional supplies were available to Edison in the PX's day ahead market. Economic curtailment was not an acceptable utility procurement practice under Rule 14." (Universal Reply Brief, p. 9.)

Thus, Universal's position seems to be that if energy was available in the day-ahead market, Edison was obliged to purchase it, no matter what the price, in order to meet its obligation to exercise reasonable diligence in delivering power to customers.

Universal's interpretation of Rule 14 is at odds with the evolution of interruptible rate programs that we described in D.01-04-006, our order in Phase I of R.00-10-002. In D.01-04-066, we decided to lift the suspension we had imposed in D.00-10-066 on the right of Edison's interruptible customers to make an annual election whether to opt out of SCE's I-6 tariff (or to adjust Firm Service Levels) partly because, owing to the dramatic increase in curtailments in 2000 and 2001, "the electricity system is operating outside any reasonable bounds, or any realistic assumption customers could have been expected to use" in making their opt-out decisions. (Id. at 14.)

In restoring opt-out rights, we noted that Commission expectations about the role of interruptible programs had changed in the late 1990s with the advent of electric restructuring. We pointed out that in 1998, because of the "transformation of the electricity market (e.g., deregulation, creation of ISO and Power Exchange)," we reduced from five years to one year the amount of notice that customers who wished to leave Edison's interruptible program were required to give. We also noted that after this change (and the curtailment requests I-6 customers experienced during the summer of 1998), SCE's interruptible customers "began to rely on the ability to reassess their situation annually." (Id. at 13-14.)

In the light of this history, we find no merit in Universal's argument that because there had been few curtailments prior to electric restructuring, "Universal reasonably anticipated that [its] interruptible status would remain secure." (Universal Opening Brief at 6.) Moreover, in view of the 1998 changes in the I-6 program, Universal is clearly exaggerating when it says that it "and other [I-6] customers could not know that their service reliability expectations had been severely compromised" by the changes in how Edison operated after the ISO assumed responsibility for what had been SCE's control area. (Universal Reply Brief at 2.)

Although it stops just short of saying so, Universal's position boils down to an assertion that by virtue of Rule 14, it became the beneficiary of a contract under which Edison guaranteed that, no matter what might happen, the frequency of curtailments would be no greater than when Universal first entered the interruptible program. In light of the language of Rule 14 itself and the above-noted changes to SCE's interruptible program, this is not a reasonable position. And as a corollary of this conclusion, the fact that a small additional amount of energy may have been available on June 27, 2000 does not mean that under Rule 14, Edison was obliged to purchase it regardless of price.

In any case, we do not think Universal has met its burden of proving that Edison could have purchased enough additional energy in the day-ahead market for June 27 to avert a Stage 2 alert. In order to prevail on its claim that Edison has not complied with Rule 14, Universal must show that (1) Edison could have met its forecast demand in the day-ahead market for June 27 by submitting a higher bid that would have been acceptable under the PX's rules, (2) Edison's meeting its forecast demand in the day-ahead market would have obviated the need for a curtailment request to I-6 customers, and (3) in view of all the circumstances, Edison acted unreasonably by failing to submit a higher bid.

Universal has not made that showing here. Edison's testimony clearly establishes that (1) based on the information available to it on June 26, SCE would have had to submit a vertical demand bid curve in the day-ahead market to have any assurance of obtaining enough energy to meet all of its forecast load for June 27, (2) the PX's rules did not allow Edison to submit a vertical demand bid curve, (3) even if the rules had allowed a vertical demand bid curve, the supply curve that has subsequently become available indicates that such a bid by Edison would have induced only an additional 65 MWh of total supply during the peak hour, (4) if Edison had submitted a vertical demand bid curve, the 250 MWh of additional supply it might have obtained would have been offset by a reduction of 175-185 MWh in the amounts awarded to other bidders, and (5) the Stage 2 alert called on June 27 came about as the result of a statewide energy shortage of at least 8,000 MWh. Since the most aggressive bidding by SCE that was theoretically possible would have increased total supply in the day-ahead market's peak hour by only 65 MWh, Edison's failure to submit a higher bid curve in that market cannot be considered the cause of the Stage 2 alert.

In its briefs and during the cross-examination of Gary Stern, Universal made several attempts to undermine his testimony on these points. For example, Universal did elicit an admission that the testimony of its witness did not specifically advocate the use of a vertical demand bid curve. (Tr. 171.) However, as Stern testified, that was the logical implication of Universal's testimony, and if it was advocating something less, then even less than 65 additional MWh would have become available:

"What [Universal] does is provide the implication that Edison was somehow unreasonable and in violation of Tariff 14 because of underscheduling. Thus, the logical inference is that Universal believes Edison should have been willing to pay any price in order to minimize underscheduling. The only way to pay . . . any price to minimize underscheduling would have been to bid a vertical demand curve at $2,500.

* * *

"And if Universal [meant] something less than this, then the resulting change would have been something less than 65 megawatt-hours. So what I intend to prove here is that there is nothing that was within Edison's capability to acquire sufficient power to avoid the underscheduling or purchases from the real-time market to any significant degree. And therefore, what it appears that Universal is implying that Edison should have done[,] was not possible." (Tr. 171-172.)

On page 9 of its Opening Brief, Universal asserts that "if Edison had submitted bids for its entire forecasted load at prices of at least $730.56 per MWh, Edison would have been able to purchase its full forecast hour 16 [i.e., peak] load." This assertion is based upon a statement in Stern's testimony that "assuming Edison had been successful in its vertical demand bid, and assuming there had been no congestion management, SCE would then have been able to purchase its full forecast hour 16 load of 13,938 MWh at the [market-clearing price] of $730.56, in the PX day-ahead market." (Ex. 6, p.23, lines 12-15.)

However, it is obvious from examining the section of his testimony in which this statement appears that Stern did not consider such a purchase a realistic possibility. First, as he had previously testified, the PX's rules did not allow Edison to submit a vertical demand bid curve. Second, Stern made clear that the amount of energy available to bidders at particular times in particular market zones was often reduced due to transmission constraints.¹² (Id. at p. 8, lines 1-17; p. 22, lines 9-13.) Thus, for Universal to suggest that Edison could have purchased enough energy to meet its forecast demand during the peak hour if only SCE had submitted a bid of at least $730.56 per MWh is a serious mischaracterization of Stern's testimony.

Universal's third line of attack is to argue that, since Edison assumed all of the other buyers' bids in the day-ahead market remained constant, Edison's analysis is unsound. The reason the analysis is unsound, Universal asserts, is because it assumes that Pacific Gas and Electric Company (PG&E) would also continue its underscheduling practices:

"[E]ven taken at face value[,] Edison's claims do not exonerate Edison. Edison's calculations assume that PG&E would continue to underschedule, which only confirms [Edison's and PG&E's] respective roles as concurrent tortfeasors . . . It is no defense for Edison that PG&E's actions also contributed to the putative supply failure." (Universal Reply Brief, p. 8.)

As stated below, we agree with Edison that during 2000, its difficulty in purchasing enough energy in the day-ahead market to meet forecasted load was a function of under-supply (i.e., sellers withholding from the PX energy they had offered in the past) rather than of underscheduling (i.e., utilities purchasing from the PX significantly less than their forecasted load in the hope of obtaining a better price in the ISO's real-time market). In any case, we agree with Stern that one cannot undertake the type of hypothetical analysis he conducted without keeping some factors constant, and that in this case one of the factors that needs to remain constant is the demand bid curves submitted by the other utilities:

"In this analysis, all other day-ahead bids submitted for June 27, 2000, both supply and demand, by other UDCs and market participants are assumed unchanged because SCE had no way to know, and certainly no way to change, their bids. This assumption, in fact, is necessary to show the impact that SCE's action ALONE would have had on the interruption." (Ex. 6, p. 22, lines 14-19.)

For all of these reasons, we conclude that none of Universal's criticisms undermine the validity of Stern's testimony. Even if the rules had allowed Edison to submit a vertical demand bid curve, the supply curves made available after-the-fact demonstrate that SCE could not have purchased enough energy in the June 27 day-ahead market to meet its forecasted load. Further, since the 5% reserve margin necessary to avoid a Stage 2 alert is computed on the basis of the ISO's system-wide demand rather than on the size of the real-time market alone, and the hypothetical use of a vertical demand bid curve would have elicited only an additional 65 MWh of supply, we find that SCE's scheduling practices in the June 27 day-ahead market did not cause or significantly contribute to the Stage 2 alert called on June 27.¹³ Thus, Edison has established that its power shortage on that day was due, in Rule 14's words, to a cause "not within [SCE's] control."

Perhaps sensing the weakness of its case about what was really possible in the day-ahead market for June 27, Universal has placed great emphasis on (1) the FERC's December 15, 2000 Order that prospectively imposed penalties on PG&E, Edison and San Diego Gas & Electric Company (SDG&E) if they did not schedule 95% of their forecasted load in the day-ahead market, and (2) statements made by the ISO and FERC during 2000 and 2001 that these three utilities were all guilty of underscheduling, and that this practice had created a reliability threat. Based on the December 15, 2000 FERC order and these statements (which complainant claims we cannot question due to principles of federal jurisdiction), Universal argues that it is clear Edison's own conduct was a proximate cause of the conditions leading to the June 27 Stage 2 alert, that the role played by other factors (such as market manipulation by Enron) does not absolve Edison of liability for its conduct, and therefore that Edison did not meet its obligations under Rule 14.

In our view, none of these arguments overcome the evidence Edison has presented. Universal's argument with respect to FERC's December 15, 2000 Order is particularly weak. As Stern noted in his testimony, the penalty that FERC prospectively imposed for "underscheduling" had precisely the opposite effect of what FERC had intended:

"By penalizing load [i.e., utilities] up to $100/MWh in the real-time market, the FERC was providing sellers the opportunity to raise their prices in the day-ahead market even more, since they now knew that utilities who refused their offers would be subject to this penalty. As SCE had warned in its pleadings to FERC, the penalty, along with the soft cap `price mitigation,' backfired, and the reliability problems and skyrocketing prices reached new heights during January and February 2001." (Ex. 6, p. 18.)

The nature of these reliability problems are set forth in Exhibit 5, which Universal sponsored. Edison vividly summarizes the data in Exhibit 5 as follows:

"Between the time the [FERC] order was issued in December 2000 until it was significantly modified in June 2001, California suffered some 85 Stage 1 alerts, 78 Stage 2 alerts, and 39 Stage 3 alerts, as well as 6 days of rolling blackouts. This level of alerts and blackouts had never been experienced in the entire history of the state, and reflects the degree of chaos that was unleashed on California by FERC's order." (Edison Reply Brief, p. 14; footnotes omitted.)

As Edison has emphasized (and Universal acknowledges), the resultant chaos caused FERC to make substantial modifications to the December 15, 2000 Order in June 2001, and then to eliminate the underscheduling penalty altogether in the December 19, 2001 Order. In light of this history, we agree with Stern that "the fact FERC made a serious error in December 2000, which it later rectified, does not support a conclusion that SCE acted inappropriately in its scheduling practices in June 2000." (Ex. 6, p. 18.)

We also find unpersuasive Universal's argument that, as a memo on trading strategies by Enron's counsel contended,¹⁴ supply withholding by generators was merely a response to utility underscheduling. (Universal Opening Brief, p. 12.) As Stern noted, an hour-by-hour comparison of the amounts of energy offered to the PX on August 25, 1999 with those offered on June 15 and June 27, 2000 shows that on average, about 10,000 MWh less was offered for the days in 2000, even though purchasers were willing to pay significantly more in 2000. (Ex. 6, p. 15.) Further, as Stern also points out, generators such as Enron had an incentive to mislead regulators into believing that highly-profitable trading strategies such as Fat Boy, Load Shift and Ricochet were merely defensive responses to utility underscheduling. (Id. at 15-16.) We also note that the guilty plea by former Enron vice president Timothy Belden (which is quoted in footnote 7) significantly undercuts the argument that Enron's trading strategies were merely responses to underscheduling.

Belden's guilty plea is not the only admission by an energy trader that supply withholding was happening in the day-ahead market during 2000. In January 2003, FERC approved a settlement between its staff and various companies in the Reliant Energy Wholesale Group (Reliant) in which the latter admitted that on June 20 and 21, 2000 -- i.e., within a week before the date at issue here -- two Reliant affiliates purposefully withheld available capacity from the PX day-ahead market for the express purpose of driving up the prices paid in that market.

In approving the staff settlement, FERC noted both the circumstances that had motivated such behavior and the evidence indicating it had occurred:

"The amounts bid into the CalPX day-ahead market reveal the extent of Reliant's withholding. For the weekdays of the prior
two-week period, Reliant bid an average of approximately 1130 MW in the CalPX day-ahead market with a high of 1800 MW and a low of 950 MW. On June 21 and 22, 2000, Reliant sold an average of approximately 130 MW in the CalPX day-ahead market with a high of 550 MW and a low of zero in 25 of the hours.

"Statements made by some of Reliant's traders demonstrate that the reason for reducing capacity bids was to increase CalPX day-ahead prices in order to mitigate losses in Reliant's existing forward positions. The transcripts clearly state the market manipulation strategy and goal, including the use of such terms as: (1) `market manipulation attempts on our part,' (2) `we decided that the prices were too low on the daily market so we shut down . . .;' and (3) `everybody thought it was really exciting that we were gonna play some market power.' Thus, Reliant employees intentionally withheld capacity from the day-ahead market to manipulate prices. Reliant has agreed to pay the CalPX customers who suffered financially as a result of the Reliant traders' actions and to ensure that similar trading activities will not recur." (Order Approving Stipulation and Consent Agreement, Docket No. PA02-2-001, 102 FERC ¶ 61, 108 at p. 61,287; footnote omitted.)

In evaluating what happened on June 27, 2000, Universal would have us give conclusive weight to statements made several years ago by FERC and the ISO that were critical of utility underscheduling, even though these statements were made well before the evidence about Reliant's and Enron's trading practices came to light. Such a result would make no sense, and the principles of federal jurisdiction cited by Universal do not require it.

We close with the observation that because of the complex events that occurred in the markets run by the PX and the ISO during 2000, we think that the formulation of proximate cause on which Universal ultimately relies in this case cuts much too broadly. Universal argues that the test for proximate cause is whether the conduct at issue can be considered a "substantial factor" in bringing about the harm complained of, and relies on the statement in Osborn v. Irwin Memorial Blood Bank (1992) 5 Cal.App.4^th 234, 253, that "[c]onduct can be considered a substantial factor in bringing about harm if it `has created a force or series of forces which are in continuous and active operation up to the time of the harm' (Rest. 2d Torts, § 433, subd. (b)), or stated another way, `the effects of the actor's negligent conduct actively and continuously operate to bring about harm to another,' (Rest.2d Torts, § 439, 433, com. e.)." Universal argues that because "Edison's actions [since 1998] had contributed to a series of forces that were in continuous and active operation up to" June 27, 2000, SCE should be held responsible for the fact it could not obtain enough energy to serve forecasted load in the day-ahead market for June 27, even though other factors may also have contributed to its difficulties in this regard. (Universal Opening Brief, pp. 21-22.)

Proximate cause is predominantly a tort concept, and has rarely if ever been applied by the Commission in tariff interpretation cases, which more nearly resemble contract disputes. Even in proceedings with a tort-like flavor (such as reasonableness reviews), the Commission has made it clear that we are not automatically bound to apply the same tort concepts that a civil court would employ. See, e.g., Re Southern California Edison Company, D.94-03-048, 53 CPUC2d 452, 466-67, 479. In any event, however, Universal's formulation of the proximate cause standard is too amorphous to apply to the complex events that took place in the power markets in 1998 and 1999. As Professors Prosser and Keeton have stated:

"'Proximate cause' - in itself an unfortunate term - is merely the limitation which the courts have placed upon the actor's responsibility for the consequences of the actor's conduct. In a philosophical sense, the consequences of an act go forward to eternity, and the causes of an event go back to the dawn of human events, and beyond. But any attempt to impose responsibility upon such a basis would result in infinite liability for all wrongful acts, and would `set society on edge and fill the courts with endless litigation.' As a practical matter, legal responsibility must be limited to those causes which are so closely connected with the result and of such significance that the law is justified in imposing liability." (Prosser and Keeton, THE LAW OF TORTS (5^th ed. 1984), § 41, p. 264; footnote omitted.)

The issue we are required to decide in this case is what obligations Rule 14 imposed on Edison on June 27, 2000, not during the two-year period leading up to that date. For example, while Edison may have had an incentive in 1998 and 1999 to minimize its purchased power costs to accelerate the recovery of Competitive Transition Charges -- as Universal alleged and Stern admitted (Universal Opening Brief, pp. 19-20; Tr. 127-128) -- that incentive had vanished by May 2000, when power costs in the PX market consistently began to exceed Edison's frozen retail rates. By November 2000, Edison's solvency was at stake. Stern is correct that if, in connection with June 27, 2000, Rule 14 were to be interpreted to impose on Edison the obligations that Universal claims, then the implication would be that Edison should have reached insolvency sooner than it did. (Ex. 6, pp. 27-28.)

Universal's proximate cause argument also ignores the changes that had come about in the demand-supply balance in the markets by June 2000. Stern's evidence is compelling that by that time, significantly less energy was being offered during the peak hour in the day-ahead market than had been offered the year before. (Id. at 15-16.) The guilty plea by Timothy Belden of Enron and the Reliant settlement with the FERC staff confirm this. Universal did not rebut Stern's evidence that even if SCE had been allowed to submit a vertical demand bid curve for the peak hour in the June 27 day-ahead market, the most it could have obtained would have been about 250 additional MWh, with other bidders' allocations being reduced by about 185 MWh. This clearly would not have been enough to avert the Stage 2 alert. ¹⁵

Thus, we conclude that Universal has not proven that Edison breached its duties under Rule 14 in connection with the PX's day-ahead power market for June 27, 2000. Accordingly, the $395,409.60 excess energy charge that Edison imposed on Universal is valid, and this penalty must now be paid to Edison.

"Edison's attempted proof of futility [in the day-ahead market] is deficient regardless of the legal standard applied. Edison presented no evidence regarding what might have been the result in the PX day[-]of market if it had been willing to pay a higher price. Without that additional evidence it is not possible for this Commission to find that Edison would not have been able to purchase its full requirements from the PX." (Universal Opening Brief, p. 22.)