V. SoCalGas Conduct During the Subject Period

SoCalGas submits that its conduct during the subject period should be assessed in light of expected future conditions at the time, as indicated by forward markets. SoCalGas also cautions that its actions should be viewed in the context of national and California conditions and the information available to it at the time. It maintains that, when its decisions to delay storage injection decisions were made in summer 2000, gas prices were thought to be temporarily high based on forward price curves. The electricity market crisis dominated the headlines. SoCalGas points out that the Carlsbad explosion had not occurred and low hydro conditions were not yet known. SoCalGas maintains that its decisions in the summer and fall of 2000 were made without knowledge that tight market conditions would prevail the following winter and that they therefore cannot be construed to reflect the exercise of market power.

Edison asserts that SoCalGas took actions to manipulate the price of natural gas at the California border during the subject period, in particular, that it engaged in hub activities and storage behavior that contributed significantly to the increased prices and volatility experienced during the winter of 2000-2001. Edison maintains that the existence of a "perfect storm" of forces facilitated SoCalGas' exercise of market power during the subject period. Edison maintains that SoCalGas took the opportunity to further constrain an already tight market and drive prices to their December 2000 heights.

Edison argues that the winter hub loan repayments, combined with low storage inventory going into the winter, increased demand for flowing supplies at the border and that the physical effect of these actions was equivalent to the direct withholding of pipeline capacity from the market. Edison argues further that SoCalGas took advantage of high prices and volatility throughout the subject period by engaging in after-market sales and financial derivative transactions to profit from the effects of its actions on the California border.

SoCalGas rebuts each of Edison's charges, and ORA agrees with SoCalGas that Edison has not provided evidence that SoCalGas manipulated the market at the California. ORA points to ratepayer benefits that accrued under the GCIM and notes that during the gas price spikes of 2000/2001, SoCalGas' ratepayers had the lowest average cost of gas of the four gas utilities serving California. ORA maintains that, if noncore customers were caught short of capacity and/or storage gas during the 2000/2001 period, they have no one to blame but themselves. It states that one of the main reasons why SoCalGas' ratepayers paid lower gas costs than other utilities was the core's substantial commitment to interstate capacity and core storage injections, which limited the need to buy gas supplies at the border. ORA argues that noncore customers had the same tools available to them, and suffered the consequences of not utilizing them.

SoCalGas maintains that its actions demonstrate that it did not know that California border prices were going to skyrocket in December 2000 and that it operated reasonably to protect core customers. SoCalGas reports that it nominated over 99% of its available firm interstate capacity during the subject period, and achieved an average delivery rate of 88%. SoCalGas also explains that it purchased additional capacity on adjoining pipelines to enhance core delivery performance on the El Paso pipeline system. Edison does not contest that SoCalGas took reasonable steps to maximize the use of its interstate capacity, and views such efforts as entirely rational since the gas purchased over that capacity was low-cost basin gas, and SoCalGas wanted as much gas as possible to loan to third parties and to sell, in order to increase its GCIM benefits.

SoCalGas argues that Edison ignores the fact that SDG&E was very dependent on border purchases and was affected "almost as strongly as Edison by the electricity crisis." SoCalGas maintains that increased profits at SoCalGas at the expense of SDG&E would do its shareholders "absolutely no good." The following table summarizes the sources of SoCalGas' GCIM savings, relative to benchmark costs, during the subject period.

Table 3

Sources of GCIM Savings Below the Benchmark

During the Subject Period

(Millions of Dollars)

Hub revenues 1.0 18.6 5.2 24.8

Baseload activity 0.4 3.5 1.5 5.4

After-market sales 1.5 95.2 4.1 100.8

After-market

purchases/other 0.2 3.6 21.5 25.3

Financial transactions 0.0 103.1 10.3 113.4

Adjustments 0.0 ( 0.4) 0.0 ( 0.4)

Total 3.0 223.6 42.6 269.2

A. Sales and Hub Loans of Core Gas to Noncore Customers

SoCalGas undertook sales and hub loans of core gas to noncore customers in every month during the subject period, but the balance shifted to loans in June 2000. The largest volume of after-market sales was in May 2000, and the largest volume of hub loans was in June 2000.

SoCalGas describes that it bought and stored border gas in the late winter and early spring of 2000 because border prices were lower than forward summer prices. As a result, by the beginning of May, it had almost 20 Bcf of purchased gas in storage, higher than typical at that time of year. In June 2000, border prices became slightly backwardated for the upcoming winter. The degree of backwardation increased significantly after the Carlsbad rupture in August. SoCalGas testified that it responded to this change in forward prices by selling gas and by loaning gas already in storage that was not needed until winter. SoCalGas maintains that loans accomplish the same price protection as physical storage or standard financial hedges, but at lower cost than physical storage and none of the downside GCIM risk. SoCalGas contends that increasing storage inventory during the summer and fall of 2000 in the face of backwardated prices would have been irrational because the expectation was that this would have increased core customers' prices.

After-market sales. SoCalGas' after-market sales yielded GCIM savings of $100.8 million during the subject period. The GCIM mechanism provides incentives to sell gas in the daily market when there are perceived temporary price spikes, and to purchase spot gas when prices fall below the bidweek price. SoCalGas asserts that both of these incentives have market benefits because they make additional supplies available during periods with a tighter supply/demand balance and have a stabilizing effect on the daily markets.

SoCalGas reports that it was a net seller in the spot market every day when there was a price spike. SoCalGas states that in some cases its spot sales were done in conjunction with hedges or physical swing purchases to ensure that replacement supplies, either later the same month or in forward months, were at a lower cost.

SoCalGas made about 6.9 million MMBtu of after-market sales in March 2000, for a GCIM profit of $1.5 million. This was the second-largest volume of after-market sales during the subject period.

During May 2000, border prices rose to almost $5.00 per MMBtu, and forward border prices were roughly the same level as May spot prices. Viewing the price increase as temporary, SoCalGas sold 11.4 million MMBtu in the spot market in May, at a GCIM profit of $10.0 million. May 2000 after-market sales exceeded by far its after-market sales during any other month in the subject period, in volume although not in profits.

Edison recognizes that SoCalGas' sales of gas to noncore customers in May 2000 served to mitigate the border price increase that month. Edison argues, however, that the large May sales increased gas costs for core customers well beyond the price mitigation that may have occurred in May, since gas prices were higher than the May bidweek price for the remainder of the year. If SoCalGas had injected the gas into storage, core customers would have paid the May price plus transport for the gas, while foregoing their share of the sales margin. Alternatively, SoCalGas could have sold the gas (as it did) and locked in the price of replacement gas for later purchase (which it did not do). Edison's view is that GCIM incentives drove SoCalGas' decisions to sell gas in May rather than inject into storage and to forego locking in the price of replacement gas. The GCIM mechanism does not penalize SoCalGas for failing to inject into storage and then paying higher prices later and, because of the trading risk, offers no incentive to fix forward prices.

SoCalGas made after-market sales of 2.4 million MMBtu in August, with a GCIM profit of $3.3 million, and 1.1 million MMBtu in September essentially at the GCIM benchmark. October after-market sales were minimal. In November, after-market sales of 3.3 million MMBtu yielded GCIM savings of $18.8 million. Edison criticizes SoCalGas for continuing to sell gas in the period following the Carlsbad rupture rather than injecting it into storage.

The largest GCIM savings due to after-market sales were achieved in December 2000, when gas prices were at their highest. After-market sales of 3.5 million MMBtu resulted in a GCIM benefit of $44.7 million. SoCalGas described that, during December, reduced core demand (due to more moderate weather) and high after-market prices led it to sell gas in the after-market. SoCalGas states that, because system inventory was above peak day minimums throughout December, these sales did not have an adverse impact on system reliability. SoCalGas reports that on December 1, 2000, it sold 90,000 MMBtu of net physical gas at the border (delivering 76,257 MMBtu after cuts) and purchased a December swing swap for 60,000 MMBtu. SoCalGas states that it drew down physical storage by about 2 Bcf during the week of the greatest price spikes in December, allowing it to sell gas and likely moderate the price spike. SoCalGas also reports that SoCalGas Topock prices rose above prices at other receipt points due to the dominance of Enron and EnronOnline, and that it arbitraged the difference by buying about 0.4 Bcf of gas at a non-Topock receipt point and selling a similar volume at Topock at a profit of over $5 per MMBtu.

Edison's view is that SoCalGas profited from opportunistic after-market sales in November 2000 and also in December 2000 when California border prices were at their peak. Edison asserts that, while SoCalGas' December after-market sales likely moderated the price spikes, they were basically profit-taking due to SoCalGas' earlier exercise of market power.

SoCalGas made lesser but still relatively significant after-market sales in February, March, and April 2001.

SoCalGas argues that Edison's criticisms are contradictory, pointing to Edison's criticisms that SoCalGas sold gas when prices spiked in November rather than maintaining storage levels and that SoCalGas should have sold more gas in December. SoCalGas submits that, "If prudent management in November was to keep gas in storage for an unexpected price shock in December, applying the same principles of prudent management would argue for similar restraint in storage withdrawals in December, given the exceptionally low levels of inventories entering December and obvious increase in volatility in late November."

Hub loans with winter repayments. SoCalGas asserts that hub loans are advantageous because they lock in cost savings for customers while providing a planned level of storage and protection of core gas costs in winter. Beginning in June 2000, SoCalGas decided to focus on hub loans rather than gas sales to noncore customers. SoCalGas originated the largest volume of hub loans in the subject period during June (10.0 million MMBtu). Of these loans, 3.1 Bcf was to be repaid in December 2000. By the end of July 2000, SoCalGas had about 15.0 Bcf in outstanding loans, with about 5.8 Bcf of the loan repayments scheduled for December. SoCalGas continued to make hub loans throughout the subject period.

The following table contains selected information regarding hub activities during the subject period.

Table 4

SoCalGas Hub Loans

During the Subject Period

New Loan Scheduled Winter Cumulative

(mil. MMBtu) (Bcf) (mil. MMBtu) (Bcf)

March 2000 0.3 4.8

April 1.3 4.5

May 0.7 3.3 0.6

June 10.0 12.0 3.6

July 6.7 15.0 5.6

August 3.8 17.7 6.1

September 1.4 18.7 6.1

October 1.8 18.5 8.6

November 2.4 17.3 3.2 8.9

December 2.4 10.6 8.9

January 2001 3.3 10.0 3.5

February 4.4 8.7 5.1

March 4.1 6.6 6.5

April 2.8 7.0 5.2

May 2.9 7.2 5.9

Of the net amount of hub loans (hub loans net of parked gas) outstanding on October 31, 2000, 9.2 Bcf were scheduled to return in November and December, in order to fulfill SoCalGas' storage target commitment based on its use of a purchased storage target. The net flow into the hub (loan repayments net of parks) in December was 7.9 Bcf, or an average of 255 MMcfd.

The subject period was not the first time SoCalGas undertook hub loans. Edison describes that SoCalGas engaged in a similar, but less significant, hub loan program in 1996 and had entered the 1996/1997 winter with a net loan position of 6.0 Bcf. However, the 1996/1997 and 2000/2001 winters appear to be the only two winters in which SoCalGas was in a net loan position at the beginning of the winter withdrawal season.7

SoCalGas explains that most of the hub loans with November and December 2000 repayment dates were scheduled for December because core load was normally higher in December than November, SoCalGas was less likely to have a high system OFO event in December, and the December returns would help avoid the potential for exceeding the core's physical inventory rights during early November when there often are injections. In addition, during June and July, the difference in the forward price between November and December was minimal. Later in the summer, the forward November price became greater than the December forward price, further justifying repayments for December instead of November.

Edison reports that the vast majority of loans during the subject period were made from storage gas, not flowing supplies, thus drawing down storage levels. Edison maintains that SoCalGas was fully cognizant of the impact that the hub loans and reduced core storage levels would have on flowing supplies at the border at the time of repayment. Edison cites the Southwest Flow Model's forecasts of December 2000 flows as an indication that SoCalGas would have expected a high likelihood of winter price spikes and congestion on the southwest pipelines.

An internal SoCalGas memo cited that "Hub loans to Winter are an effective strategy for capturing Winter premiums" (Ex. 90, Att. 3-2). SoCalGas asserts that this language "identified Hub loans as a potential way to protect the core from winter price spikes."

Edison acknowledges that hub loans mitigated price spikes in summer 2000. It asserts, however, that by scheduling 8.9 million MMBtu of loan repayments for December, SoCalGas knowingly constrained the market at that time. Pointing out that in 5 of the 6 years between 1994 and 1999 the peak demand month for SoCalGas was December, Edison maintains that SoCalGas understood from the beginning that December repayments would occur during the month most likely to have the tightest supply-demand balance. Edison states that the loans SoCalGas made for December 2000 repayment, which equated to a demand for border gas of about 250 MMcfd in December (hub loan repayments net of park activities), were sufficient to cause price spikes. Edison asserts that prices in December would have been lower if, instead of obtaining the gas from flowing loan repayments, SoCalGas had withdrawn these quantities from storage. Edison argues that, if SoCalGas had loaned less gas, storage inventory levels would have been higher at the beginning of winter and SoCalGas could have withdrawn more gas in December.

In Edison's view, the Carlsbad rupture in August 2000 provided additional indications that the supply/demand balance was tight and would continue to be tight during the winter. The Southwest Flow Model forecasts following the Carlsbad rupture indicated December flows of about 3,500 MMcfd, higher than actual September flows, and including December hub inflows of 150 MMcfd. After the Carlsbad rupture, high basis differentials indicated that the market was already constrained. Yet between September and November, SoCalGas continued to loan additional gas for winter repayments, increasing total December loan repayments from about 6 Bcf at the time of the Carlsbad rupture to 8.9 Bcf at the end of November, an increase of about 100 MMcfd. In particular, SoCalGas negotiated hub loans in late October requiring 2.0 Bcf of December repayments. Edison reiterates its view that by September 2000 SoCalGas fully expected price spikes in the coming winter months, pointing in particular to the Winter Hedging Strategy document (Ex. 92, Att. 9) and draft Gas Acquisition Committee Meeting notes of September 22, 2000 (Ex. 75) stating that "Gas will be coming back in winter counter cyclically and will set up loans during price spikes." Edison asserts that SoCalGas' loan behavior reflects a knowing exercise of market power, and that the loans are equivalent to the direct withholding of interstate pipeline capacity from the market.

SoCalGas responds that any company borrowing gas that anticipated that gas prices would be high at repayment time had the ability to store gas ahead of time for loan repayment purposes. SoCalGas asserts that the availability of gas storage to noncore customers would have, in itself, prevented SoCalGas from using gas loans to distort winter prices. Edison counters that the noncore holders of storage capacity had no way to know ahead of time that the SoCalGas was planning to start the winter with such a low level of core storage.

SoCalGas defends its decision to make new loans in the latter part of November 2000 for December delivery. SoCalGas takes issue with Edison's focus on loans under which gas was withdrawn in November for December repayment without reporting loans arranged in November under which gas was withdrawn in December for later repayment. "The hub made loans in early November for December repayment, but in latter parts of the month the hub arranged loans that would be taken out in December, thus offsetting some of the December loan repayments. Moreover, the California hub received two parks (negotiated November 17 and November 27) in which gas was injected in November so that it could be withdrawn in December. Thus, ...between November 1 and November 30, the hub reduced the amount of gas due to flow into the hub during December by about 150 MMcf." (Ex. 7 (Montgomery) at 33.)

SoCalGas makes seemingly contradictory statements on the question of whether hub loans affected border prices. SoCalGas' preferred argument appears to be that the schedule of loan repayments did not increase the winter demand for flowing gas supplies and did not affect border prices during the subject period. When taking this view, SoCalGas maintains that the hub loans were financial transactions that determined who paid for gas delivered to SoCalGas during the winter, but that they had no effect on physical volumes flowing over the border. SoCalGas posits that a hub loan made in July would make a difference in December only if one assumes that the alternative is to inject the gas into storage in July and to withdraw it in December. It maintains that it would have been a financial burden on core customers to add to storage inventory at that time, that the reasonable alternative to making hub loans would have been to sell the gas, and that the supply/demand balance would have been the same with that alternative as with a hub loan.

While claiming that hub loans did not affect border prices, SoCalGas also acknowledges that, "to the extent the loans may have led to decreased total system storage levels in the winter, these decreased total system storage levels may have put upward pressure on border prices." SoCalGas' witness stated:

It is total physical storage on the SoCalGas system on October 31, 2000, that determines the amount of gas available for withdrawal in the subsequent five winter months. This level of storage was arguably lower than it would have been absent hub loans. With lower storage levels entering the winter, withdrawals during the Winter were also arguably lower than they would have been without the loans, resulting in greater demand for flowing gas and therefore more upward pressure on prices. Since virtually all the working gas in storage on the SoCalGas system was withdrawn by March 2001, it is reasonable to suppose that if there had been additional gas in storage, more would have been withdrawn. If it had been possible to withdraw more gas from storage, supplies would have been less tight during the winter and directionally border prices would have been lower.

Therefore, the lower level of storage due to summer loans and reduced injections into storage had the entirely unforeseen consequence of contributing to price increases in the winter. (Ex. 3 at IX-130 to IX-131.)

Consistent with this assessment, Edison asserts that, had SoCalGas not made hub loans for winter repayment, SoCalGas would have put more gas in storage and would have withdrawn more gas during the winter months, which would have reduced flowing supplies at the border and lowered border prices.

B. Management of Storage

SoCalGas' Gas Operations Department acts as the system operator and is responsible for operating SoCalGas' storage fields as well as redelivery of supplies for both core and noncore customers. Remedial measures are in place to prevent Gas Acquisition from obtaining knowledge of noncore activities available to Gas Operations in its role as the system operator.

Storage allows transportation assets to be sized to meet average system needs rather than peak demand, with gas purchased and stored during periods of relatively low demand (the injection season) available during periods of higher demand (the withdrawal season). In addition, storage in excess of amounts needed to maintain reliability provides a valuable physical hedge against future price increases.

If there is excess transmission capacity from the producing basins, filling storage during the injection season may be less critical than if such transmission is likely to be constrained during high demand periods. One of the central issues in Phase I.A is whether SoCalGas acted properly, in light of what it knew about likely winter conditions, in managing its storage system to rely on flowing repayment of hub loans during the winter of 2000/2001 rather than having that gas in storage at the beginning of the winter withdrawal season.

Edison submitted evidence that SoCalGas began reducing its planned storage injections in May 2000, and notes that this was about the time SoCalGas noticed that the supply/demand balance at the border was becoming tight and basis differentials were increasing because of EPME's control of 1.3 Bcf of El Paso pipeline capacity. Edison asserts that SoCalGas cut back on its planned injections in order to enhance future market tightening, and that low core storage levels at the beginning of winter followed by inadequate withdrawals were a primary contributor to border spikes during the 2000/2001 winter.

SoCalGas maintains that, at most, storage decisions played only a minor part in the winter price spikes. It acknowledges that counting loan repayments toward the core storage target may have led to somewhat lower levels of physical storage, which in its view became one small part of the upward pressure on prices in winter. SoCalGas contends, however, that its level of physical storage entering the winter season and its reliance on hub loan repayments were reasonable. In SoCalGas' view, purchasing more gas for inventory during the summer and fall of 2000 in the face of backwardated prices would have been irrational because of the expectation that such an action would increase core customers' prices.

SoCalGas asserts that it is the total system physical storage level that matters, not the level of hub loans, and emphasizes that SoCalGas put far more physical gas in storage during the subject period than any other market participant in southern California. According to SoCalGas' explanation, low system storage levels were primarily attributable to the backwardated market and the behavior of noncore customers and marketers, not SoCalGas.

SoCalGas has four operating storage fields with a combined capacity of 105.6 Bcf. This capacity is allocated to core (70 Bcf), balancing (5.3 Bcf), and unbundled/noncore (30.3 Bcf). About 2 Bcf of core's 70 Bcf was allocated to the Core Aggregation Transportation (CAT) program.

SoCalGas ceased operations at its fifth storage field at Montebello in early 1997 and filed an application in January 1998 for authority to sell the field. A settlement approved in June 2001 provided that Montebello's working gas (3 Bcf) and cushion gas would be withdrawn and sold. Withdrawals from the Montebello field commenced in July 2001.

1. SoCalGas October 31, 2000 Core Storage Level

Edison asserts that SoGalGas entered the 2000/2001 winter heating season 13.6 Bcf short of its October 31 target storage level, setting the stage to constrain market supplies of gas throughout the winter.8 SoCalGas maintains to the contrary that its core physical storage was only 5.6 Bcf short of the October 31 target and was reasonable.

Edison argues that SoCalGas' hub and sales activities in 2000 ignored the fundamental objective of getting gas across the border and into storage fields during the April-October injection season for use during the winter. In Edison's view, SoCalGas, through its use of hub loans and sales rather than storage injections, turned what should have been storage withdrawals during the winter period, especially December, into a situation where core demand had to be met by flowing supplies, in particular hub repayments, exactly the situation storage is designed to avoid. Edison points out that there is no profit incentive under GCIM to maximize storage fill. Edison criticizes SoCalGas in particular for continuing to sell and loan gas in the period following the Carlsbad rupture rather than using the gas to build storage.

During the subject period, SoCalGas was required to fill core storage to 70 Bcf, plus or minus 5 Bcf, by October 31, 2000. The core's inventory target was a physical target when it was developed for the 1992/1993 winter. SoCalGas' March 2000 injection plan anticipated that SoCalGas would inject gas to obtain 70 Bcf in core storage, including 1.4 Bcf of parked gas, by October 31, 2000.9 SoCalGas' actual injection levels were lower than anticipated in the March 2000 plan, such that on October 31, SoCalGas had 56.4 Bcf of core gas in the four operating storage fields, including 1.7 Bcf of CAT storage and 1.6 Bcf of parked gas. Excluding parked gas, core's purchased gas in the 4 operating storage fields on October 31, 2000 was 54.8 Bcf, which was 15.2 Bcf below core's allocation in the fields.

Physical core storage inventory was less at the start of the 2000/2001 winter than at the same time in any other year in the 1994 - 2002 period, as indicated in the following table.

Table 5

October 31 Physical Storage Levels*

1994 - 2002

(Bcf)

As indicated in Table 5, the total storage level of 65.2 Bcf (excluding Montebello) on October 31, 2000 was the lowest during the 1994 through 2001 period except for 1996 (56.7 Bcf), which was also unusually low. Of the other 6 years during the period, the lowest total storage level was 89.1 Bcf (1999); the average for those 6 years was 98.0 Bcf. By December 1, 2000, total storage inventory had dropped to an historic low for the date, 50.0 Bcf excluding Montebello.

The following table provides more detail on October 31 storage levels during the 1998 - 2000 period (Ex. 85, Figure 3-24).

Table 6

October 31 Retail Core Storage Levels+

1998 - 2002

(Bcf)

Retail core purchased inventory 58.8 62.9 71.9 64.5 62.3

Net gas parked/loaned to noncore 3.6 0.9 (17.2) 3.0 3.5

Retail core physical gas in storage 62.4 63.8 54.7 67.5 65.8

Net loan repayments by 12/31 ** - 9.2 - -

Core aggregation inventory n/a* 2.0 1.7 0.9 0.7

Total core storage 62.4 65.8 65.6 68.4 66.5

SoCalGas reports that retail core storage inventory reached 85% of total capacity by the end of October (59.4 Bcf of gas in storage, including CAT, parked, and Montebello gas) compared to the total core allocation of 70 Bcf. SoCalGas compares this storage level with a nationwide average of 83% in November 2000 and the storage level of its noncore customers other than SDG&E, which averaged 12% (3 Bcf) of the 24 Bcf available to them. SDG&E entered the winter season with its 6 Bcf of contracted storage nearly filled. SoCalGas states that it met its storage target with 59.4 Bcf of physical storage inventory and 9.2 Bcf of net returning hub loans, and that it was well positioned at the commencement of the winter season.

Issues in dispute regarding the October 31 core storage level include whether the core target level should be viewed as 70 Bcf or 65 Bcf, whether Montebello gas should be counted toward meeting the target, whether SoCalGas' reliance on hub loan repayments in lieu of physical storage was approved and reasonable, and whether the physical storage level achieved by October 31, 2000 was reasonable. We also address the proper treatment of parked gas in assessing core storage levels.

Edison asserts that core storage was short by 13.6 Bcf on October 31, while SoCalGas maintains that the physical shortfall was only 5.6 Bcf and was more than offset by hub loan repayments. The two numerical assessments differ because Edison uses the 70 Bcf core storage capacity whereas SoCalGas uses the 65 Bcf lower limit of the established 70 Bcf plus or minus 5 Bcf target range, and Edison excludes whereas SoCalGas includes the 3 Bcf of working gas in the Montebello storage field. Both parties include parked gas in their calculations of October 31 storage levels.

We view use of 70 Bcf or a 65 Bcf target as a matter of semantics, since the target was established as a range. Achievement of either 65 Bcf or 70 Bcf would fall within the approved target range. However, parked gas should not be included in assessing SoCalGas's compliance with the storage target, or in assessing core storage adequacy and reliability. Parked gas has to be returned and, particularly if the return is before or during the winter, would not be available when needed to maintain reliability.10 SoCalGas has recognized that parked gas should not be included in assessing compliance with either physical or purchased storage targets (Tr. at 1539). At the end of October 2000, there were 18.5 Bcf of loans and 1.6 Bcf of parks, for a net park/loan amount of -17.2 Bcf, as indicated in Table 6. Excluding parks, 53.1 Bcf of gas was in the four operating storage fields on October 31, 2000.

Montebello Working Gas. SoCalGas reports that the facilities necessary to allow withdrawal of the working gas in the Montebello storage field were maintained and that the Montebello gas could have been withdrawn if necessary. However, SoCalGas did not utilize this gas during the extremely adverse conditions of the 2000/2001 winter, even though it knew that use of the Montebello gas would have decreased California border prices (Ex. 90, Att. 2-5).

During the subject period, SoCalGas did not treat Montebello as a working storage field. SoCalGas operated its storage assets essentially as if the Montebello working gas did not exist. SoCalGas explains that Gas Operations did not include Montebello in its operating plans because the inventory would not be available for cycling while the Commission was considering the Montebello settlement, which had been submitted and was under consideration during the 2000/2001winter.11 SoCalGas excluded Montebello gas in its GasSelect postings of total system storage volume starting in September 1998.

SoCalGas established peak day minimum requirements for the 2000/2001 winter excluding Montebello. As a result, its winter noncore balancing rules were pegged to physical storage levels excluding Montebello. SoCalGas maintains that this exclusion was reasonable because the working gas in Montebello was not expected to be withdrawn during the 2000-2001 winter.

SoCalGas' Seasonal Operations Plan for its storage facilities (excerpted in Exhibit 130) stated that there would be no planned use of the Montebello storage field during the 2000/2001 winter. For contingencies, the operations plan specified curtailment procedures but did not address use of Montebello.

SoCalGas' August 31, 2000 advice letter requesting authorization to implement winter balancing rules on October 1 rather than November 1, 2000 specified that, if storage volumes were drawn down excessively, SoCalGas' only two options would be to purchase expensive gas for core customers or to interrupt service to noncore customers. The advice letter did not mention withdrawals from Montebello as a third alternative.

In a November 29, 2000 internal meeting shortly after the Montebello settlement agreement was signed, SoCalGas addressed closure of the field. Meeting notes indicate that "(t)he remaining working gas in Montebello will be available for withdrawal by January 31, 2001" and that "(t)he working gas in the field will only be withdrawn subsequent to the 851 approval or in case of a supply upset that threatens core customers." (Ex. 92, Att. 32). SoCalGas asserts that the engineers at the meeting were considering the best way to withdraw all of the gas in Montebello, including both working and cushion gas, and that the cited January 31, 2001 date was the target date for work to be completed to allow blow down of the entire field to occur at the maximum rate possible.

In much of the prepared testimony in this proceeding, SoCalGas witnesses exclude Montebello working gas when reporting storage levels and conditions. SoCalGas does not include Montebello gas when it describes the conditions that transpired during the winter, e.g., that total storage levels came within 0.6 Bcf of the peak day minimum on March 5, 2001.

Edison asserts that Montebello gas should not be included in tallies of stored gas available during the 2000/2001 winter and alternatively that, if the Montebello gas was indeed available, it should have been withdrawn in light of system conditions during the winter.

SoCalGas' testimony is convincing that the 3 Bcf of working gas in the Montebello field could have been withdrawn during the 2000/2001 winter. However, SoCalGas operated its system as if the gas was not available. While SoCalGas states that it would have withdrawn the Montebello working gas to avoid curtailment situations, it did not use the gas during the dire conditions that transpired and, in fact, it imposed peak day balancing requirements without taking the gas into account. SoCalGas did not want to use the gas pending consideration of the settlement regarding its ownership and disposition. We conclude that, for all practical purposes, SoCalGas withheld the 3 Bcf of Montebello working gas during the 2000/2001 winter. The record was not sufficiently developed to allow us to draw conclusions regarding whether some of the approximately 20 Bcf of cushion gas in Montebello could have been withdrawn to ease system constraints.

Purchased vs. Physical Storage Requirement. A primary justification SoCalGas gives for its level of physical core storage on October 31, 2000 is that net hub loan repayments were planned to provide 9.2 Bcf of additional gas in November and December 2000. SoCalGas interprets its core storage target as a purchased target rather than a physical target, and counts net hub loan repayments (loan repayments net of incoming hub parks) due during November and December toward meeting the purchased target. SoCalGas' use of a purchased inventory target is based on its interpretation of D.97-11-070, in which the Commission approved SoCalGas' petition to modify the gas balancing rules adopted in D.90-09-089.

The Commission has never examined SoCalGas' reliance on hub loan repayments to satisfy core storage requirements. In SoCalGas' July 18, 1997 petition to modify the gas balancing rules adopted in D.90-09-089, in a section entitled "The Commission Should Not Be Distracted by Side Issues," SoCalGas responded to criticisms that SoCalGas did not meet its core storage target during the 1996-97 winter, stating that,

[I]n order to decrease the likelihood and/or duration of the 70 percent or 90 percent balancing regimes, SoCalGas will ensure that the core (retail plus CAT) suppliers meet their purchased inventory target of 70 Bcf (+/- 5 Bcf) as of October 31st. If a Hub loan position exists as of October 31st, only the Hub loan position scheduled for payback by December 31st will be applied towards satisfying the retail core purchased inventory amount.

Nowhere in its petition did SoCalGas acknowledge that use of a purchased inventory target was contrary to the Commission's existing standard, nor did it ask that Commission policy be changed to allow use of a purchased inventory target. Instead, in an offhand manner, SoCalGas promised to "ensure" that operations would comply with a standard that was contrary to the Commission's physical inventory standard. In D.97-11-070, the Commission approved SoCalGas' two requested modifications to the gas balancing rules (contained in Appendix A to the petition, which did not address the inventory target), without mention of SoCalGas' passing reference to a purchased inventory target. Since the Commission did not address the storage target at all in D.97-11-070, it was incorrect for SoCalGas to rely on that order as granting authority for it to use a purchased inventory target.12

In D.02-06-023, we recognized that, "(t)he Gas Acquisition Department has a Commission-established storage inventory capacity of 70 Bcf and aimed to get within 5 Bcf of full capacity by November 1, including gas repayable by the end of December." In that decision, we did not address the propriety of SoCalGas' reliance on repayments, but specified that the October 31 storage target in Year 9 and thereafter would include only physical gas in storage, consistent with the submitted settlement.

As Edison suggests, it would be inappropriate to mechanistically use a purchased inventory target as a benchmark of the reasonableness of SoCalGas' efforts to meet winter storage needs. Indeed, while SoCalGas contends it had Commission authorization to use a purchased target, it recognized a responsibility to make reasonable decisions regarding the amount of gas physically in storage. SoCalGas' internal hub operating guidelines (Ex. 90, Att. 3-23) state that "Hub transactions will not be undertaken if, based on current and expected conditions, they subject core customers to higher cost or jeopardize supply reliability." While emphasizing that SoCalGas inappropriately switched to a purchased inventory target without Commission authorization, we also assess SoCalGas' hub loan program against this standard of reasonableness.

2. Winter Operation of Storage

Edison argues that SoCalGas kept the market for gas further constrained in December 2000 by not drawing down storage sufficiently even though, in Edison's opinion, such a draw down could have mitigated or eliminated the December border price spikes. SoCalGas believes that it managed its storage reasonably throughout the subject period.

SoCalGas started the winter season with 56.4 Bcf of gas (including CAT and hub parks, excluding Montebello gas) in core storage, compared to its March 2000 plan of 70 Bcf. There was 65.2 Bcf in total storage excluding Montebello. SoCalGas bought 3.3 million MMBtu in the after-market and similarly sold 3.3 million MMBtu in the after-market in November. Departing from a mid-September plan to have net core storage injections of 1.2 Bcf in November, SoCalGas withdrew 8.4 Bcf of core gas; 15.0 Bcf of total gas was withdrawn in November. SoCalGas states that very cold weather increased core demand in November by 11 Bcf over the BCAP forecast and by 9 Bcf over the prior year's demand. SoCalGas ended November with 46.3 Bcf of core gas (excluding CAT and Montebello) in storage and 50.0 Bcf of total system inventory. SoCalGas reported that this was an historic low for the date. By comparison, the five-year average operational total inventory for December 1 was 95 Bcf.

In December 2000, SoCalGas withdrew 2.8 Bcf of gas from storage. It justified this minimal level of withdrawals on the basis of the low level of gas in storage entering the month and the need to preserve storage for remaining winter reliability, and also because gas demand was 8 Bcf lower than expected in December due to warmer than normal weather. Because of the low storage withdrawals, SoCalGas depended almost entirely on flowing gas to meet core demand in December. SoCalGas reports that all of December's storage withdrawal was during the week of December price spikes, which served to moderate the price spikes. SoCalGas loaned 2.4 Bcf of gas in December, explaining that only 0.3 Bcf was negotiated in December, with 1.1 Bcf negotiated in November and 1.1 Bcf negotiated the prior spring. SoCalGas bought 3.0 million MMBtu of after-market gas and sold 3.5 million MMBtu of after-market gas. Noncore injected 2.8 Bcf of gas into storage in December. SoCalGas ended December with 43.5 Bcf in core storage (excluding CAT and Montebello) and 50.0 Bcf of total system inventory.

Colder than normal weather returned in January and February 2001. Core sendout each of these months was higher than in any other month in the subject period. SoCalGas withdrew 20.1 Bcf of core gas (excluding CAT) in January. It made 4.3 million MMBtu of spot purchases in January and sold 0.8 million MMBtu in the after-market. Total storage inventory dropped to the core peak day minimum plus 20 Bcf, and the 70% daily balancing requirement was triggered on January 18, 2001. As SoCalGas reports, the 70% balancing requirement appeared to have increased deliveries by about 20 MMcfd. Border spot prices spiked to exceed $15 per MMBtu. The 70% daily balancing trigger had only been reached once before, in the winter of 1997-1998. SoCalGas ended January with 23.4 Bcf of core storage (excluding CAT and Montebello) and 27.8 Bcf of total storage.

In February 2001, with cold weather continuing, SoCalGas withdrew 15.2 Bcf of core gas (excluding CAT) from storage. It made 3.8 million MMBtu of spot market purchases and sold 1.1 million MMBtu in the spot market. Total storage inventory was depleted to the core peak day minimum plus 5 Bcf on February 12, so that customers were required to deliver a minimum of 90% of their gas usage on a daily basis. The 90% daily balancing requirement had never been imposed before, and had a noticeable effect on core delivery patterns. Deliveries increased by almost 100 MMcfd over the rest of the month, and border gas prices spiked to exceed $36 per MMBtu. SoCalGas ended February with about 8.5 Bcf of core storage (excluding CAT and Montebello) and 13.7 Bcf of total storage.

SoCalGas' storage inventory reached a low of 12.3 Bcf on March 5, 2001, the second-lowest total ever recorded and only 0.6 Bcf above that day's system peak day minimum requirement (the minimum level needed to provide core peak day reliability plus noncore firm withdrawal and balancing services). SoCalGas describes that it aggressively purchased gas in March in order to stabilize core storage gas and accelerate the core summer storage injection program. SoCalGas purchased 16.0 million MMBtu of border gas in March, over 50% more than in any other month in the subject period. Price spikes of over $30 per MMBtu accompanied the border purchasing program. With relatively mild weather, the core ended March with a physical storage level (excluding CAT) of 22.5 Bcf.

SoCalGas reports that, because of continuing concerns about the potential for high gas demand due to expected electric demand, SoCalGas accelerated gas purchases and storage injections throughout the remainder of the subject period. SoCalGas injected about 5 Bcf in April 2001 and about 10 Bcf in May 2001 into core storage.

Edison asserts that SoCalGas' opportunistic sales of approximately 8 Bcf from November 2000 through January 2001 and its continuing to loan gas during the same period strongly indicate that SoCalGas was not concerned at all about core reliability. Edison particularly criticizes SoCalGas' November sale of 3.3 million MMBtu of after-market gas as one of the reasons for the large November storage withdrawal. SoCalGas counters that in November it purchased 3.3 MMBtu of after-market gas, which offset its after-market sales.

At the same time, Edison maintains that SoCalGas should have, and would have if it had not been exercising market power, withdrawn additional storage volumes in December 2000 in order to reduce gas prices. Edison maintains that SoCalGas' claim that additional withdrawals were not possible that month is contradicted by its withdrawal levels in prior years and by a prior claim that it only needed 35 Bcf of core storage for reliability purposes. SoCalGas responds that its statement that only about 35 Bcf is necessary for meeting peak day reliability had been made in conjunction with the proposed unbundling of core storage under certain system rules and conditions being contemplated in the GIR proceeding-rules and conditions which were not in place during the winter of 2000/2001. Edison argues that SoCalGas cannot justify the low storage withdrawals in December 2000 when the price spikes were at their most devastating by pointing to greater withdrawals in January 2001 when prices moderated significantly. Edison believes that December prices would have been moderated with a steady pattern of storage withdrawals in December and January, rather than 2.8 Bcf in December and 20.1 Bcf in January.

SoCalGas responds that Edison's arguments are contradictory, pointing to Edison's criticisms that SoCalGas sold gas when prices spiked in November rather than maintaining storage levels and that SoCalGas should have sold more gas in December. SoCalGas submits that, "If prudent management in November was to keep gas in storage for an unexpected price shock in December, applying the same principles of prudent management would argue for similar restraint in storage withdrawals in December, given the exceptionally low levels of inventories entering December and obvious increase in volatility in late November." SoCalGas argues that larger December withdrawals would have been irresponsible in light of the already low storage levels and with most of the winter season still ahead. SoCalGas points out that more aggressive core drawdown in December would have triggered the tighter winter balancing rules earlier. SoCalGas also contends that drawing down more storage to reduce price spikes in December would have left less gas in storage to deal with the February and March price spikes, which would have been higher.

Edison concludes that, if SoCalGas' low storage level throughout the winter and into March 2001 was a problem, "it was a problem created by SoCalGas," pointing out that SoCalGas still had an 8.7 Bcf net loan position at the end of February 2001 and had sold substantial amounts of gas in the border spot gas market during November (2.2 Bcf), December (3.3 Bcf), January (0.5 Bcf), and February (0.7 Bcf).

3. Storage Held by Noncore Customers

SoCalGas reports that noncore storage customers, excluding SDG&E, had 24 Bcf of storage capacity available to them, with nearly 23 Bcf under contract. These customers had approximately 11 Bcf in storage at the end of July 2000 but withdrew gas such that they had only 3 Bcf in storage at the end of October 2000. They had close to zero inventory stored in mid-December but increased their storage position by 1.1 Bcf in December, which SoCalGas cites as a factor increasing demand for flowing supplies in December. SDG&E had 6 Bcf of storage rights and, unlike noncore marketers, increased its gas in storage from 2.4 Bcf at the end of July 2000 to almost its full contractual amount.

SoCalGas and Edison suggest that noncore customers (other than SDG&E) did not fill their storage because of market backwardation. SoCalGas explains its view that, in depleting inventory during the summer months, noncore customers were acting rationally in their own short-term self-interest, with traders willing to take the risk expecting that gas would be cheaper in the future.

SoCalGas maintains that, to the extent that total storage was a factor in the high winter border prices, it was due to noncore storage behavior. SoCalGas points out that it is not allowed to buy and store gas for noncore customers and that, aside from its responsibility to ensure peak day minimum storage levels, Gas Acquisition has no responsibility to noncore customers. SoCalGas maintains that it would have been irresponsible to further draw down core gas storage in December in order to attempt to insulate noncore customers from the consequences of their storage and hedging decisions.

Edison rebuts that there is a clear distinction between the situations of individual noncore storage holders and SoCalGas. Unlike the core with 70 Bcf of storage reservation, the noncore consists of relatively small holders of storage capacity. While each small holder acted in what it perceived to be its own economic self-interest during the subject period, there was nothing that a noncore holder could have done individually to significantly affect storage fill. Further, the noncore holders of storage capacity had no way to know that the core was going to start the winter with such a low level of storage.

SoCalGas had non-public knowledge throughout the subject period regarding planned repayment of hub loans. SoCalGas posted total storage inventory on its GasSelect bulletin board, with no breakdown between core and noncore storage and no indication as to the amount of loans/parks or the timing of loan or park repayments. SoCalGas argues that total physical storage is "the only figure needed to complete an equation that determines the potential tightness of markets in the winter" and that "knowledge of the loan repayment schedule is of no value in predicting future gas prices." However, the timing of flows into and out of the hub can affect border prices and noncore customers would have benefited from the knowledge that SoCalGas had scheduled loan repayments in winter periods, especially December.

It appears that SoCalGas was aware of the value of information regarding its hub loan activity and deliberately withheld that information from the market. We note in particular an e-mail sequence in which an employee of TXU Energy asked SoCalGas about the relationship between SoCalGas' loan activity and storage levels and SoCalGas decided not to post responsive information on GasSelect (Ex. 92, Att. 27).

C. Financial Positions

Through financial hedges, SoCalGas lowered gas costs relative to the benchmark by $113.4 million during the subject period. The GCIM formula does not provide incentives to engage in financial hedges as a means to protect customers against price volatility. SoCalGas characterizes its hedge program as insurance, which is viewed as not likely to pay off in the form of net gains and which constitutes a significant shareholder risk.

Edison criticizes the financial positions that SoCalGas took at the California border. Edison argues that SoCalGas took these border positions based on its knowledge that the market was misled about the backwardated price curves and, as a result, profited from the high gas prices it was creating. SoCalGas responds that hedging does not require a belief that forward markets are incorrect regarding the expected value of gas on some future date, but reflects an unwillingness to bear the risk of the uncertainty associated with future prices.

SoCalGas took financial positions other than at the California border, including lucrative hedges involving New York Mercantile Exchange (NYMEX) Henry Hub call options. Edison does not take issue with SoCalGas' NYMEX hedges, explaining its view that, unlike the California border hedges, SoCalGas could not exercise market power regarding the NYMEX hedges.

SoCalGas' California border hedges (over-the-counter (OTC) basis, price, and swing swaps) added over $43 million to SoCalGas' GCIM gains during the subject period, with more than $34.7 million in profit in December 2000 and January 2001. Monthly results of SoCalGas' border hedges during the subject period are summarized in the following table, based on information in Figure 4-2 in Exhibit 85:

Table 7

SoCalGas Border Hedges

OTC Basis, Swing, and Price Swaps

Gains/(Losses)

SoCalGas took short financial positions at the southern California border through August 2000, but switched to long positions starting on September 1, 2000. SoCalGas explains that after the Carlsbad rupture it began to be concerned about its exposure to the border market and began building a long hedge using basis swaps. Edison suggests an alternative explanation: that SoCalGas shifted its border position because of its knowledge regarding the effects of its significant hub loan activity in the June-August period, which required loans to be repaid in the winter. Edison explains its view that, in early August 2000, SoCalGas had started to form an opinion that the November basis was likely to expand significantly from the level it was trading at in August. In an August 4, 2000 e-mail initiating November long border/basin positions, Gas Acquisitions' Energy Economics Manager indicated that, "Based on our flow analysis, our expectation is that they [border/basis differentials] could reach the current market spreads of around $1." SoCalGas soon thereafter started to take a net long position at the border.

SoCalGas states that it added border hedges in November "when it became apparent that more supplies would be purchased at the border because of extreme cold weather in early November and a reduction of in-state supplies from Exxon under the long-term POPCO contract." Edison argues that SoCalGas established the additional border positions in November because it recognized that the repayment of hub loans would contribute to record demand for southwest gas.

In September, SoCalGas started buying OTC border basis swaps with winter settlement dates. By September 15, it had a net buy position of 35,000 MMBtud for OTC border basis swaps for December gas. Thereafter, SoCalGas maintained a net buy position for December, although it cashed out most of its position prior to December. It had a net gain of $7.9 million on its December border basis swaps.

Primarily in late November and early December, SoCalGas bought January basis swaps, with a net buy position of 165,000 MMBtud by the end of December. SoCalGas held most of the January basis swaps through January and reaped a net gain of $18.0 million.

SoCalGas' border basis swaps for February and March of 2001 were for even larger positions but with less successful results. SoCalGas "cashed out" most of its long border basis positions for February and March by the end of January. SoCalGas bought more border basis swaps for March in February but lost money on most of them. For May, the settlement price was $10.049, even higher than December.

SoCalGas also engaged in OTC border swing swaps. It entered into significant positions for November and December gas starting in mid-November, with the majority of its November and December border swing swaps executed between November 13 and November 20. Edison notes that this was the same time that the November 15, 2000 Gas Acquisition Committee meeting minutes indicated SoCalGas' expectation that, "(i)n December, California demand for southwest gas supplies is projected to be at record levels due to increased core purchases and large Hub paybacks. The current forecast implies a potential shortage of delivery for the expected demand" (Ex. 90, Att. 3-27). The border swing swap transactions for November gas led to gains of $2.8 million, while those for December gas led to gains of $16.3 million.

SoCalGas submits that it acquired the November and December swing swaps because it increased its December purchase plan to replace gas withdrawn in November due to unexpected weather-related demand. Because it expected that much of the increased purchases would have to be in the after-market, SoCalGas purchased 2.2 Bcf of December swing swaps, in addition to the 1.3 Bcf in December basis swaps then outstanding. When weather moderated in December, the open swing swap positions were reduced. During the first week of December, reduced core demand and high after-market prices led to the decision to sell 2.3 Bcf of gas in the after-market. The remaining swing swaps provided price protection if core demand rose again and border purchases were required later in the month.

SoCalGas describes that it established a long summer border basis position, limited to 40 Bcf, in March 2001 to protect the core against a price run-up as it attempted to fill storage. As system storage levels increased, this financial position was reduced. SoCalGas established a short border basis position for winter 2001 to help protect the value of the high cost of gas injected during the early storage build. SoCalGas describes that, as the gas demand for generation decreased from expected levels, the basis differential moderated and the positions partially offset the higher cost of the early storage build.

SoCalGas argues that for its hedges to be consistent with actions of a market manipulator, it would have had to first establish a financial position at the California border and then subsequently establish the physical position. Edison responds that the financial position was established in the fall of 2000 before the physical flows into the hub (winter loan repayments) occurred, advance knowledge of which was not available to other market participants. Further, SoCalGas made significant increases in its hub loans with December repayments during October and November, after it took financial border positions, which in Edison's opinion satisfies SoCalGas' criteria for a market manipulation.

Edison argues that SoCalGas' entering into border hedges based on information regarding loan repayments and storage decisions that only SoCalGas possessed was a conflict of interest and an exercise of market power. In addition, Edison asserts that it was a conflict of interest for SoCalGas to increase the amount of hub loans with December repayments while it had a long December border position. Edison also argues that SoCalGas' purchase of a large amount of border gas during November for December delivery while it had a long financial position was a conflict of interest and that SoCalGas should have relied on storage instead.

SoCalGas submits that it took border positions to protect its core customers against price risk and to support the purchase of low cost gas. SoCalGas maintains that it almost never held open border positions that exceeded the amount of gas it planned to purchase at the border13 and that it rode out most (but not all) basis and swing swaps to the end of the month in order to provide continuing protection against price volatility. While it achieved an aggregate gain on border positions, this included positions with losses as well as gains. SoCalGas argues that these losses contradict any assertion that it somehow knew the behavior of the market and therefore was able to profit from that knowledge. SoCalGas also asserts that, had it been relying on some kind of inside information, it could have taken positions earlier than it did. SoCalGas also argues that it would have been extremely counterproductive to take any steps that would deliberately lead to higher gas prices, because its limited financial hedges did not cover all of forecasted demand.

Edison notes that the GCIM does not provide incentives to engage in NYMEX-based hedging transactions. If gas prices did not increase, the NYMEX hedges would expire worthless and their cost would count against GCIM. Edison points out SoCalGas modified Gas Acquisition's employee incentive compensation plan to exclude the impacts of the NYMEX winter hedge program, whereas border transactions were included in the incentive compensation plan.

D. Discussion

The Commission has allowed SoCalGas to procure and sell gas to noncore customers and to engage in hub activities, including hub loans, as means to improve the use of core assets and lower core gas costs. In addition, financial hedges can be a valuable tool to provide core customer protection from volatility in gas prices. Many of SoCalGas' actions during the subject period may have been consistent with these goals. However, we find that SoCalGas misused these tools at times with the intent to boost GCIM earnings rather than to protect core customers and, in the most benign interpretation, with disregard for the resulting harm to the broader gas market.

In particular, we conclude that SoCalGas' unprecedented concentration of winter repayments for hub loans and after-market sales undertaken when it should have been filling core storage deferred the acquisition of gas needed for core customers' winter use, thus requiring more expensive replacement purchases later in the yearly cycle. These actions also increased constraints on the ability of border gas supplies to meet demand, thus contributing to price increases and spikes, particularly in December 2000.

SoCalGas' actions during the subject period, in the depths of the energy crisis, must be assessed with recognition that they were made, as SoCalGas suggests regarding its winter storage withdrawals, under conditions of immense uncertainty about where markets were going. The uncertainty that existed throughout the subject period argues for thoughtful approaches to utility operations with particular attention given to risk management. SoCalGas emphasizes its reliance on financial hedges, undertaken counter to market expectations as reflected in forward prices, as a step to protect core consumers from price risk. At the same time, SoCalGas argues inexplicably that its storage and hub loan decisions should be assessed solely by reference to futures prices at the time they were undertaken.

Particularly in light of the unprecedented market conditions at the time, we would not expect SoCalGas to base its operations on a simplistic assumption that futures prices were an accurate and reliable prediction of market outcomes,14 and we do not find this after-the-fact justification convincing now. While SoCalGas represents that forward prices reflect a consensus of market participants regarding expected prices, that is not always the case. Futures prices provide some information about market expectations, but may be skewed by speculators or participants with market power seeking quick profits independently of their expectations regarding future market conditions. SoCalGas recognized the susceptibility of futures prices to manipulation at least by early August, 2000 (Ex. 90, Att. A-5, see also Ex. 90, Att. 1-1). Further, even the best market predictions are often wrong. The experience in 1996/1997 should have provided SoCalGas with ample insight that futures prices are unreliable.

SoCalGas asserts that it did not foresee the price increases or related system constraints that occurred during the 2000/2001 winter. We agree that there is no indication that SoCalGas anticipated price increases or spikes of the magnitude that actually occurred, with spot prices reaching $60 per MMBtu. However, as we discuss in Section IV, we find convincing evidence that SoCalGas was aware that the market was likely to be constrained during the winter.

There is also convincing evidence that SoCalGas knew its actions in the summer and fall of 2000 would contribute to border constraints later in the year. Commencing in June 2000, SoCalGas planned to reduce physical core storage to unprecedented low levels and to enter into hub loans with large repayments during the winter months. It continued to enter into hub loans with winter returns throughout the fall of 2000, and chose to buy and sell after-market gas in November rather than bolster storage, even though it was indisputably clear to SoCalGas at that point that the market would be constrained during the winter months.

SoCalGas' objectionable actions were driven by its desire to reap shareholder profits through its GCIM. Gas Acquisition set "stretch goals" totaling $38 million for GCIM Year 7, including goals for hub activities ($8.5 million), after-market sales ($18.0 million), and financial positions ($3.0 million) (Ex. 90, Att. 5-5). These goals compare to average yearly GCIM savings of $12.6 million during the prior 6 years, with the largest yearly achievement being $24.2 million in GCIM Year 6. At the planning conference in April 2000 where the GCIM Year 7 goals were discussed, Gas Acquisition urged its staff to be more aggressive in leveraging physical assets. Participants were encouraged to be more active in basis trading and to increase position risks; discussions addressed catching the "Big Wave." Participants were reminded that compensation would reward performance. After the planning conference, Gas Acquisition commenced to work toward the GCIM Year 7 goals.

SoCalGas affirmatively planned to, and did, benefit from the high gas prices and volatility that were occurring. As SoCalGas points out, Gas Acquisition could have made more money if it had increased storage withdrawals in December, increased after-market sales, hedged more, and cashed out hedges at peak market conditions. However, there were multiple factors affecting the market on a day-to-day basis and SoCalGas' foresight was limited. Nevertheless, as Edison comments, SoCalGas "made plenty of money" knowing that prices were trending upward and projecting that record demand would occur, particularly in December because of the large hub loan repayments.

The fact that core customers have received $192.7 million in "benefits" due to GCIM Year 7 sharing does not tell the whole story regarding the effect of SoCalGas' actions on core gas prices and core customers. Acquisition of gas for core needs that could have been achieved through relatively low-cost purchases during the summer injection period was instead deferred to later periods when the only gas available was high-priced border gas.15 The GCIM does not penalize SoCalGas for such costly shifts in needed purchases. Additionally, because SoCalGas' actions contributed to bidweek border price increases during the 2000/2001 winter, the GCIM benchmark was set higher and the "savings" calculation for border purchases is not an accurate reflection of whether SoCalGas saved money on border gas purchases. While SoCalGas' core customers were protected from the higher cost of winter gas needed for loan repayments, they were exposed to the unhedged portion of SoCalGas' cost of the direct border purchases. To determine the true impact on core customers, GCIM customer "benefits" would have to be netted against these increases in core costs. While we cannot establish the impact definitively, it is clear that the net effect was an increase in core customer bills.

We recognize that, of the four gas utilities serving California, SoCalGas had the lowest average cost of core gas during the subject period. This was due, in part, to SoCalGas' relatively large interstate capacity, which allowed it to rely more heavily on basin gas purchases and to be less reliant on border purchases compared to other utilities or noncore gas users without such direct access to basin gas. Thus, while SoCalGas' core gas costs were higher than the prior winter, its core customers were spared the worst of the price spikes.

The fact that SoCalGas' core gas costs were lower than those of other California utilities does not excuse SoCalGas' behavior, which caused the cost of border gas purchased for core customers to be higher than it would have been otherwise and had rippling effects throughout the gas and electricity markets. Noncore gas customers and electricity customers, who received no benefit from GCIM, were directly harmed by the increased border prices. As Edison noted, industrial gas customers with contracts indexed to border prices suffered severe financial harm when border prices were high. SoCalGas actions that increased border gas prices also increased electricity prices. Thus, in addition to higher gas bills as a result of SoCalGas' actions, their electricity prices increased.

In the remainder of this section, we provide a more thorough discussion of our analysis that led us to these conclusions.

After-market Sales, Hub Loans, and Storage Management. After-market sales and hub loans can be profitable for core customers, e.g., in situations where gas purchased for core customers turns out to be in excess of current core needs. However, if misused, they can raise gas costs for core customers and, indeed, other gas purchasers as well. After-market sales and hub loans can raise overall gas costs for core customers if gas in excess of core needs is purchased for speculative purposes and market conditions do not materialize to make such sales or loans attractive. The record does not indicate that this scenario occurred during the subject period. To the contrary, excess gas purchased in the spring of 2000 was sold or loaned profitably as market conditions tightened. SoCalGas' core storage inventory after the May 2000 sales of excess gas was still well above average for that time of year. We do not take issue with SoCalGas' May 2000 sales under the GCIM mechanism.

After-market sales and hub loans can also increase core customer costs if gas needed for core purposes is sold or loaned for short-term GCIM profits such that the gas has to be replaced later with higher-cost purchases, or such that operational constraints or reliability concerns arise. This is what happened commencing in June 2000. SoCalGas pursued GCIM profits by selling and loaning gas while shortchanging core storage needs. SoCalGas thus contributed to higher winter prices, to the detriment of core customers and the market as a whole. While drawing down core storage to unprecedented lows, SoCalGas purchased after-market gas in several months, including August, September, and November, and sold the gas rather than injecting it. SoCalGas explained that it hedged the cost of some of the replacement supplies. However, such price protections did not provide sorely needed operational flexibility as winter conditions worsened and gas supplies became scarce at any price.

Even when it may be reasonable for SoCalGas to provide temporarily excess core gas to noncore customers, hub loans do not minimize core gas costs compared to other options. Selling the excess gas and locking in the price of needed replacement gas through a futures contract would yield the same gas flows as occur with hub loans, but typically with higher profits for the core because a sell/buy arrangement would capture the entirety of the spread as savings for core customers. The current GCIM structure discourages transactions that lock in forward prices, however, because such behavior is "judged" for GCIM purposes based on bidweek gas prices in the month in which the gas is delivered. In addition, forward purchases for November and December 2000 deliveries would not have helped meet SoCalGas' interpretation of its October 31 storage requirement as a purchased target. SoCalGas' preference for hub loans may be understandable because of the current GCIM incentives but it is counter to customer interests. In Section VI, we adopt steps to modify GCIM to better align its incentives with customer interests in this regard.

While SoCalGas' hub loans may have served to dampen gas prices somewhat in the summer of 2000, winter repayments undoubtedly increased price and volatility to a much greater extent due to the more extreme conditions during the 2000/2001 winter.

As we establish in Section V.B, SoCalGas entered the winter withdrawal season with 16.9 Bcf of unfilled core storage capacity, 11.9 Bcf less than the lower bound of the 70 Bcf plus or minus 5 Bcf core storage target established by the Commission. If SoCalGas had sold less gas to noncore customers during the injection period or had loaned less gas for winter repayment, SoCalGas could have filled its core storage before the beginning of the winter withdrawal season. In November, with historically low storage levels, SoCalGas purchased and sold 3.3 Bcf of after-market gas rather than using it to meet core gas needs and conserve stored gas supplies. In combination, these decisions made core winter demand heavily dependent on flowing supplies at the California border. This deleterious result is obtained whether the flowing supplies come from hub loan repayments or from direct SoCalGas purchases. SoCalGas' failure to fill its core storage adequately limited the ability of storage to fulfill its intended purpose during winter demand conditions.

Edison argues that, with yearly peak demand days occurring frequently during December, SoCalGas' reliance on December hub loan repayments set up the situation where core demand had to be met by flowing supplies. While December 2000 was relatively mild and did not contain the winter's peak demand days, an exceptionally cold November had the same effect, because already tight storage was drawn down precipitously, limiting SoCalGas' ability to withdraw gas in December.

In December 2000, SoCalGas relied on 8.9 Bcf of hub loan repayments (7.9 Bcf net of hub outs) and 4.3 Bcf of border gas (net of noncore sales) to meet core customer needs. The net hub loan repayments and core border purchases required border flows averaging almost 300 MMcfd. This constituted over 30% of core burn in December, and over 11% of total SoCalGas sendout that month. With the tight supply/demand conditions in the winter of 2000/2001, there is no doubt that these volumes were sufficient, as SoCalGas acknowledges, to affect border prices, with the higher prices borne by all border purchasers. These price effects occurred even though, as SoCalGas points out, the volumes of winter hub loan repayments were small relative to the demand and supply shocks that occurred throughout the subject period and were an even smaller portion of total gas flows during the period.

Edison contends that SoCalGas should have withdrawn more gas from storage in December in order to mitigate the December border price spikes. In light of the depleted storage levels at that time, we are not convinced that such a step would have been desirable. Additionally, because of the concurrent price increases elsewhere in the region, Edison's assertion is not credible that SoCalGas' actions, by themselves, caused the entirety of the winter border price spikes. However, SoCalGas' actions may well explain why the December spikes were larger at the California border than elsewhere.

Core storage levels reached historic lows in January and February 2001, with the imposition of first 70% and then 90% daily balancing requirements. Price spikes occurred in both months and in March 2001, but without reaching December heights. SoCalGas sold after-market gas and reaped GCIM profits during each period of price spikes in those months.

SoCalGas' interpretation of the core storage target to include hub loans with repayments by the end of December had the effect of moving the storage target deadline from the end of October to the end of December.16 Such a step was counter to the fundamental purpose of a storage target to ensure sufficient gas supplies before the start of the winter season. The fact that SoCalGas did not have the storage reserves to weather one month of unexpectedly high withdrawals (November) and respond adequately to tight conditions in December confirms that SoCalGas' physical storage was insufficient entering the winter season. This demonstrates that reliance on winter loan repayments to meet storage targets is contrary to the hedging function of storage.

The 2000/2001 winter was the only year in which SoCalGas relied on December hub loan repayments to meet its storage targets. In D.02-06-023, the Commission approved a settlement that specified the use of a physical storage target for SoCalGas in future GCIM years.

We agree with SoCalGas that if total storage had been filled before the winter withdrawal season (total storage was 30 Bcf below capacity on October 31, 2000), winter prices would have been lower. We note, however, that noncore customers may have been inclined to fill more of their storage and may have made different choices regarding the use of financial derivatives if they had had access to information regarding SoCalGas' hub loan activities and repayment schedules, or if they had known that SoCalGas would not meet its Commission-established storage target.

We conclude that SoCalGas undertook concerted actions commencing in June 2000 and continuing throughout the remainder of the GCIM Year 7 cycle which were detrimental to core customers and the broader gas market. SoCalGas' hub loans and after-market sales undertaken when it should have been building its storage inventory for winter use deferred the acquisition of core needed for core customers' winter use, thus requiring more expensive replacement purchases later in the yearly cycle. These actions also acted to constrain market supplies during the 2000/2001 winter and increased winter gas prices at the California border.

It is not clear whether SoCalGas undertook these actions solely to profit from the arbitrage opportunity--with SoCalGas viewing the prospect that its actions would increase prices as an acceptable consequence--or whether SoCalGas had an express intent to constrain the winter gas market in order to increase border prices and further enhance its arbitrage profits.

By scheduling significant volumes of winter hub loan repayments, which caused increased reliance on flowing gas supplies during peak winter periods; by failing to fill core storage adequately during the injection season; and by choosing not to withdraw the core's working gas in the Montebello storage field even during extremely tight system conditions, SoCalGas knowingly contributed to supply constraints and effectively withheld gas supply during peak winter months. These actions caused gas price increases and volatility at the California border. By selling gas during the resulting price spikes, SoCalGas profited from the price increases caused by its actions. As a result, we conclude that SoCalGas' actions went beyond GCIM-incented profiteering into the realm of market manipulation and an exercise of market power.

Because SoCalGas' GCIM profits between June 2000 and March 2001 were the result of market manipulation, we conclude that these profits were not reasonable and were received in violation of Section 451. As a result, we require that SoCalGas refund all profits that shareholders received due to operation of the GCIM mechanism during those months. The refund will total approximately $28.8 million, plus interest.

While this refund of GCIM profits will return SoCalGas' ill-gotten gains to core customers, we do not address potential culpability for harm to market participants other than core customers. As a result, and in light of our findings that SoCalGas exercised market power and manipulated the gas market during the subject period, we refer our Phase I.A findings to the Attorney General of the State of California, who is currently investigating the activities of Sempra and its subsidiaries, including SoCalGas, during the energy crisis, or to other appropriate law enforcement agencies. If requested, the Commission will cooperate fully with any law enforcement agency regarding this matter.

The record does contain a full analysis of market conditions and SoCalGas' actions during April and May 2001, the first two months of the next yearly GCIM cycle. As a result, we do not have sufficient information to assess whether SoCalGas' actions during those months raise concerns comparable to those we found for GCIM Year 7 with commencement of the hub loan program starting in June 2000.

Financial Derivatives. Like hub activities, the use of financial derivatives can be beneficial to core customers. In particular, financial transactions can be used to hedge against price increases for needed core gas purchases, consistent with SoCalGas' explanation for its extensive financial activities during the subject period. Hedging undertaken for other purposes, however, can be detrimental to core customer interests.

Financial transactions can be useful as a risk management tool, to stabilize both gas prices for core customers and internal cash flow for the company. A National Regulatory Research Institute (NRRI) report (Ex. 84) characterizes hedging used for risk management purposes as "bona fide hedging." It cautions that bona fide hedging will not reduce the average cost of gas purchases over time, but can best be viewed as price insurance. Another analysis (Ex. 102) explains that hedging actually increases expected costs due to transaction costs and, because of this, hedging and risk management are not related to least-cost planning.

Financial hedges can also be undertaken for speculative purposes unrelated to providing customer value due to price stabilization. Speculators hope to profit from price movement or volatility. As NRRI explains, speculators assume the risk shifted from bona fide hedgers and hope to profit from future movements in the market that are not reflected in forward prices. NRRI recognizes that the boundary between bona fide hedging and speculation is not always clear, but describes that attempts to "time" or "beat" the market are generally speculative. We agree with NRRI that local distribution companies, as regulated entities providing service to core customers, should refrain from market speculation.

In addition to price stabilization and speculation, the record contains several other possible justifications for hedging activities. First, as Edison alleges regarding SoCalGas' activities during the subject period, a firm can profit from financial transactions based on market manipulation and/or non-public information. Notably, Gas Acquisitions' Energy Risk Manager argued against SoCalGas' use of border financial derivatives, due in part to Enron's ability to manipulate prices (Ex. 90, Att. 1-1).

An internal SoCalGas document provides additional possible explanations for SoCalGas' use of financial derivatives. An undated "Business case for using derivatives" presentation to Sempra's Vice President of Risk Management lists the following justifications for derivatives use:

1. Price Discovery. Active trading in basis markets facilitates price discovery. This enables (1) the Hub to better define the value of parking and loaning transactions and (2) Gas Supply group to value forward month and spot purchases and sales.

2. Great liquidity of derivatives market eases position initiation. The liquidity and anonymity of the derivatives market enables Gas Acquisition to quietly establish a significant position in the market.

3. Hedge physical position. Derivatives enable Gas Acquisition to hedge the GCIM risk of a physical purchase or sale if the group finds the position risk uncomfortable. (Ex. 92, Att., 28, emphasis added.)

Of particular note is that the third justification listed above is tied to GCIM risk rather than customer risk. Because the GCIM calculations are based on bidweek prices and actual purchase volumes, GCIM risk arises primarily if after-market purchases are necessary at prices exceeding the benchmark. Customer protection is notably absent from this presentation.

The use of physical gas storage or hub loans can be viewed as alternatives to financial derivatives as means to protect core customers from price increases or volatility, as SoCalGas recognized when it was assessing possible hedging programs for the 2000/2001 winter. A significant difference, as SoCalGas has recognized, is that use of storage can provide operational flexibility and reduce price volatility, with benefits for the entire market, whereas financial derivatives only protect SoCalGas and core customers.

It appears that many of SoCalGas' financial transactions were undertaken to hedge against increasing prices for expected purchases, in particular its winter NYMEX-based hedges. In addition, as SoCalGas explains, it entered into hedges in conjunction with some after-market sales, to lock in the price of replacement gas if needed. Such uses of hedges appear advantageous in terms of providing customer protection from unexpected price increases.

However, the record indicates that at least some of SoCalGas' financial transactions were taken for reasons other than customer protection. As mentioned above, hedging undertaken to protect against price increases is not expected to reduce gas costs but rather, over time, increases average gas costs by the transaction costs of the hedges. Yet SoCalGas' "stretch goals" for GCIM Year 7 included a $3.0 million gain due to financial positions (Ex. 90, Att. 5-5). The Gas Acquisition planning conference in April 2000 urged participants to be more active in basis trading and increase position risks; discussions addressed catching the "Big Wave." Lacking from the conference agenda is any discussion of hedging to protect against increasing prices, which would not be expected to yield GCIM profits. The establishment of a goal for GCIM profits due to financial transactions, coupled with the manner in which this goal was presented at the conference, leads to a reasonable inference that SoCalGas expected its Gas Acquisition employees to engage in hedging activities other than "bona fide" hedging, to use NRRI's term.

A SoCalGas document discusses using SoCalGas' system demand and Southwest Flow Model forecasts to anticipate tight supply conditions and cautions that "use of this information for speculative purposes is at your own risk" (Ex, 92, Att. 13).

There is anecdotal evidence in the record that some of SoCalGas' financial positions were undertaken for speculative purposes and/or based on SoCalGas non-public information regarding market conditions. As one example, SoCalGas lost $347,000 due to a speculative August 2000 short border basis position betting against the existing border premium (Ex. 90, Att. A-5). In addition, a number of e-mails (Ex. 90, Att. 4-1) providing instructions for placing financial positions described speculative trades, including the following:

· March 13, 2000: "These [NYMEX] positions will take advantage of a winter contract price run-up into the $3.50-$4.00 range, which the analytic group believes is highly possible given the abnormally low U.S. storage level forecast for the end of this injection season."

· May 24, 2000: "We expect this [NYMEX spread] to result in a 20 cent gain in the next few months as the spread is presently much wider than normal."

· July 18, 2000: "The [short put] strategy is to take advantage of high market volatility during market declines..."

· August 10, 2000: "The [San Juan/Permian] spread will profit from any narrowing due to expected flow improvements out of the SJ into bidweek. ... it is highly volatile."

· August 24, 2000: "Expectation is that these [January long calls] could double in price with additional bullish storage reports or hurricanes. Strategy is to buy more of these on market weakness."

· September 1, 2000: San Juan/Permian swing "strategy is to take advantage of rising SJ prices from reduced flows into SJ with cooler Rockies/PNW weather." Long November and December call options "in anticipation of stronger prices into fall."

Other border positions described in the e-mails were based on SoCalGas' internal forecasts of winter flow conditions at the border:

· August 4, 2000: "Based on our flow analysis, our expectation is that [the November basin/border spreads] could reach the current market spreads of around $1."

· September 8, 2000: September Remainder of Month Permian/border spread "in anticipation of the spread widening with additional SW gas demand."

· October 18, 2000: Long November San Juan/border spread to "take advantage of forecast high flows to CA."

We are troubled by the indications that SoCalGas traders engaged in market speculation, and particularly by the California border hedges undertaken based on SoCalGas' forecasts of border flow conditions. These border hedges indicate an intent, while anecdotal, to profit from the tight winter conditions occurring, in part, due to SoCalGas' own hub loan and storage activities. The fact that SoCalGas undertook these hedges buttresses our finding that SoCalGas knowingly manipulated the gas market and exerted market power through its hub loan program and other activities during the subject period.

In Section VI, we adopt plans to evaluate appropriate risk management activities for SoCalGas, including whether such activities should be included in the GCIM calculation or undertaken in a manner that complements the procurement incentive mechanism.

7 SoCalGas' hub went into a net loan position during November 1999 for a single month, but that amount was not used to meet its October 31 storage target for 1999.

8 SoCalGas assert that Edison's testimony in this proceeding regarding the inadequacy of SoCalGas' core storage levels during the subject period is inconsistent with its position as a signatory to the Comprehensive Settlement Agreement in April 2000. Edison disputes, however, SoCaGas' representation that Edison agreed in April 2000 that SoCalGas' core storage reservation should be reduced from 70 Bcf to 55 Bcf. Edison argues that while it did support, in principle, a reduction in SoCalGas' core storage capacity, it did not recommend a reduction to 55 Bcf and had not analyzed the profitability and market implications of reducing the core storage reservation to 55 Bcf, as SoCalGas had done in the May 1999 Winter Price Spike Analysis.

9 The March 30, 200 plan for October 31 storage levels (Ex. 90, Att. 3-5) appears to exclude Montebello and include CAT gas, while October 31 targets in summer injection plans commencing in May 2000 (Ex. 90, Att. 3-16) include Montebello gas and exclude CAT gas. 10 Of the 1.6 Bcf of parked gas at the end of October 2000, 0.7 Bcf was returned by the end of November and 1.3 Bcf was returned by the end of December. 11 The Montebello settlement provided for withdrawal of all gas, with proceeds from working gas going to customers and proceeds from cushion gas belonging to shareholders.

12 Canales v. Alviso (1970) 3 Cal.3d 118, 127 n.2: "'It is axiomatic that cases are not authority for propositions not considered.' (People v. Gilbert, 1 Cal.3d 475, 482, fn. 7). `Questions which merely lurk in the record, neither brought to the attention of the court nor ruled upon, are not to be considered as having been so decided as to constitute precedents.' (Webster v. Fall (1925) 266 U.S. 507, 511)."

13 SoCalGas reports that it held long basis positions entering January 2001 equal to about 115% of the border purchases actually made during that month. It explains, however, that this situation arose because border purchases turned out to be less than expected in January. 14 Indeed, Exhibit 115 indicates that SoCalGas prepared internal predictions of future gas prices. However, SoCalGas' price forecasts for the subject period are not in the Phase I.A record. 15 During the subject period, SoCalGas bought a record 17% of its gas purchases at the border. It is not clear how much of these purchases was due to core needs and how much was due to noncore sales undertaken for GCIM profits. 16 SoCalGas has provided no rationale, either in its petition to modify D.90-09-089 or subsequently, for counting hub loan repayments but not forward purchases with November or December delivery dates toward a purchased storage goal.

Previous PageTop Of PageNext PageGo To First Page