IV. SoCalGas Knowledge and Expectations Regarding the 2000/2001 Gas Market

A. SoCalGas Ability to Affect Border Prices and Wield Market Power

SoCalGas acknowledges that some of its actions during the subject period affected border gas prices. It points out, however, that in any real world working market, market participants' consumption and output decisions may have some impact on the market price. SoCalGas asserts that a market is still competitive if the cost of withholding from the competitive level would be greater than the benefit. SoCalGas concludes that evidence of an exercise of market power requires more than simply high prices, since high prices may simply indicate increases in costs and/or demand and may represent scarcity pricing consistent with the competitive price level.

Market power is the ability to move price above the competitive level for a sustained period of time and profit from doing so. The exercise of market power by a supplier requires a withholding that causes a significant and sustained increase in the price of the product withheld, and the withholding must be profitable. As SoCalGas explained, the withholding could be through either removal of supply from the market (physical withholding) or the offering of prices that are too high (economic withholding).

SoCalGas and Edison disagree regarding the relevant product and geographic markets for assessing whether SoCalGas has market power to affect border gas prices. Theoretical market constructs aside, we believe that testimony by SoCalGas witness Montgomery is largely accurate in describing the southern California gas market during the subject period:

Gas coming across the border or from intrastate sources is clearly a substitute for gas withdrawn from storage. ...[A] hypothetical monopolist who controlled hub services and storage ... could not cause actual delivered gas prices to rise as long as there was free movement of gas across the border...

The only conditions under which restricting supply or increasing demand in southern California might affect the price of gas is when total system demand approaches system capacity (in this case, the sum of gas available from storage and intrastate sources plus the capacity of interstate pipelines to deliver gas). If demand does not approach capacity, increased supplies of natural gas ...will flow into California whenever border prices begin to rise (even slightly) relative to prices outside California, as buyers seek to minimize purchase costs and suppliers seek to maximize revenues. These increased supplies will offset any attempted withholding and defeat any possible price increases, as long as demand is below that capacity. (Ex. 7 (Montgomery) at 39-40.)

We would refine this description somewhat, because a comparison of demand and system capacity must recognize that factors such as capacity withholding may prevent the interstate capacity from being utilized fully. SoCalGas' own analysis of market conditions concludes that the pipelines were constrained during periods of price spikes. Thus, by SoCalGas' own analysis, the conditions under which Montgomery hypothesizes that control of hub services and storage could affect border prices existed during the subject period. While the pipelines had capacity to deliver more gas to California, well-documented factors prevented the importation of increased supplies of natural gas to counter the effects of SoCalGas' hub loan repayments and storage decisions. Thus, we conclude that SoCalGas had the ability during the subject period to move the price of gas above the competitive level by effectively restricting the supply of of gas through its hub services and gas storage activities.

An inquiry regarding whether SoCalGas wielded market power entails an examination of whether SoCalGas knew prior to the winter of 2000/2001 that it had the ability to affect border prices, which we examine in Section IV.B, whether SoCalGas knowingly undertook actions with the effect of withholding capacity with the intent of profiting from such actions, and whether such actions were profitable, which we examine in Section V.

B. SoCalGas Knowledge Regarding 2000/2001 Market Conditions

As SoCalGas notes, any alleged exercise of market power must be judged on the basis of information available at the time. SoCalGas asserts that it did not know and could not have known that there would be border congestion in the 2000-2001 winter and, therefore, that its decisions earlier in the year could not constitute an exercise of market power.

SoCalGas acknowledges that its low levels of physical storage going into the 2000/2001 winter put upward pressure on border prices. However, it maintains that it had no reason to anticipate that outcome, because the market expectation during the summer of 2000 was that gas supplies during the winter would not be tight. SoCalGas' Exhibit 107 indicates that the forward basis differential for uniform gas delivery throughout the winter (November through March) from the Permian basin was less than the maximum cost of transportation every day between March 1 and November 1 except for the two weeks after the Carlsbad rupture. This winter basis differential did not remain above the cost of transportation consistently until after November 6. Comparable forward basis differentials for the month of December showed a similar pattern. SoCalGas concludes that the forward winter basis differentials indicated that any attempt to make the market short of gas in winter by keeping core physical storage low would have been defeated.

Edison asserts to the contrary that SoCalGas was aware that increasing system constraints during the winter of 2000 would increase gas prices at the California border. Edison maintains that SoCalGas planned to take advantage of the tight supply/demand balance at the border after it became apparent in the spring of 2000 that EPME was withholding pipeline capacity and other conditions were developing that would contribute to a tightening of the supply/demand balance. Edison argues that, in undertaking its hub loan program, SoCalGas was aware that the winter loan repayments were likely to contribute to border constraints and higher and more volatile winter border prices. In Edison's view, SoCalGas similarly understood when it was taking financial positions at the California border that the market backwardation was incorrect, based on its non-public knowledge regarding winter hub loan repayments.

We agree that what SoCalGas knew, and when SoCalGas knew it, are essential areas of inquiry in assessing whether SoCalGas knowingly manipulated the gas market during the subject period. We describe the record chronologically regarding SoCalGas' knowledge regarding market conditions. Each individual document or event may not provide a definitive indication regarding SoCalGas' knowledge. However, the record as a whole demonstrates that SoCalGas had analyzed the border dynamics over a number of years and understood clearly during the summer and fall of 2000 that increased reliance on flowing gas supplies during the upcoming winter and decreased availability of storage withdrawals would contribute to market constraints and put upward pressure on border prices.

First, SoCalGas' experiences during the 1996/1997 winter would have provided an understanding that low levels of storage in 2000/2001 would affect winter prices. There are strong parallels between the two periods. The market became backwardated in July 1996 and continued in backwardation through November of that year. Noncore inventory declined as a result and, unlike in 2000/2001, actually was negative (- 2 Bcf) entering the winter season. SoCalGas made hub loans and entered the 1996/1997 winter season with 59 Bcf of physical core storage and 6 Bcf of net hub loans. Contrary to the earlier backwardation, winter gas prices then rose above the prior summer's levels. The January 1997 San Juan and Permian indices topped $4 per MMBtu, and winter border spot prices were, on average, over $2 per MMBtu higher than summer border spot prices. Edison provided two contemporaneous trade press articles attributing the regionwide 1996-1997 price run-ups to low storage in the west.

While there are similarities between conditions in 1996/1997 and 2000/2001, differences between the two periods would have indicated that prices in the 2000/2001 winter likely would be higher than those experienced in the earlier period. Bidweek basis differentials in the summer of 2000 were already indicating constraints on the interstate pipeline system, whereas the monthly basis differentials never exceeded $0.23 per MMBtu in 1996. Interstate deliveries into the SoCalGas system were much higher throughout 2000 than the level of deliveries in 1996. Additionally, SoCalGas' net hub inflows were planned to be much larger in December 2000 (net December 2000 hub inflows were 7.9 Bcf) than in December 1996 (when they totaled 1.0 Bcf).

In 1998 and 1999, SoCalGas followed the market manipulations of Natural Gas Clearinghouse (NGC), which later changed its name to Dynegy. At the beginning of January 1998, NGC obtained 1.5 Bcfd, roughly 30% of the total pipeline capacity on the EPNG and Transwestern pipelines to the California border. SoCalGas proactively analyzed the potential effects of NGC/Dynegy's control of interstate capacity. In mid-January 1998, a SoCalGas analyst quantified potential effects on pipeline capacity prices of possible NGC withholding of capacity (Ex. 115) and offered the following opinion:

With regard to the NGC strategy of using its pipeline capacity on EPNG's system, it seems that they are testing the market to get it to reveal the "Price Elasticity of Demand" and EPNG seems reluctant at this point to "compete" by offering Interruptible Capacity for anything less than "full reservation charges."

The string of e-mails in Exhibit 115 characterized that the "spot price of pipeline capacity" increased substantially in January after NGC acquired its pipeline capacity and also indicated that SoCalGas considered this analysis in its internal forecast of California border prices.

The border/basin differential began to rise immediately after NGC acquired its capacity. Minutes of the January 23, 1998 meeting of SoCalGas' Gas Acquisition Committee⁴ (Ex. 70, Att. 1) attributed the increase in basis differentials to "rumors that NGC is not releasing any of its capacity and that EPNG is no longer discounting [interruptible transmission]." Edison asserts that the minutes' statement that "Gas Acquisition will continue to monitor market and regulatory changes surrounding this issue in search of opportunities" indicates that SoCalGas believed that there were profit opportunities to be had from NGC/Dynegy's anticompetitive behavior, whereas SoCalGas counters that the quote merely reflects SoCalGas' commitment to seek low cost gas.

A March 1998 Gas Acquisition briefing indicated that NGC had publicly identified that it did not intend to release any of its acquired capacity to third parties and expressed SoCalGas' understanding that NGC had the potential to establish the marginal price of gas sold at the border and that "(b)order vs. basis differentials will continue to be influenced by NGC's control of 1.5 Bcfd on EPNG" (Ex. 70, Att. 8).

SoCalGas predicted that NGC could affect basis differentials and border gas prices in high demand months whenever other parties needed NGC capacity (Ex. 70, Att. 6). An early version of SoCalGas' Southwest Flow Model forecasted demand for southwest gas for each remaining month in 1998 and identified that NGC capacity could be needed to meet demand in August and September but that sufficient non-NGC capacity was likely to be available to meet demand in the other months of 1998 (Ex. 70, Att. 10). Foreshadowing SoCalGas' activities in 2000/2001, SoCalGas recognized that, whereas NGC for the most part could not by itself dictate the market price by withholding capacity, SoCalGas' "(s)torage dynamics and the core's use of its transportation rights can impact these results."

SoCalGas prepared an internal Winter Price Spike Analysis in May 1999 in conjunction with Gas Industry Restructuring (GIR) settlement discussions. Among other things, GIR proposals would tighten winter balancing rules and reduce the core storage inventory reservation from 70 Bcf to 55 Bcf. The one-page analysis indicated that GIR proposals would increase the magnitude of winter price spikes, resulting in increased gas costs for customers and higher GCIM profits. Edison asserts that the Winter Price Spike Analysis demonstrates that SoCalGas, one year prior to making its storage and hub loan decisions during the subject period, already had a reason to expect that those decisions would have a physical effect on the market and that lower levels of core storage inventory would contribute to price increases. Edison points out that SoCalGas reduced the planned amount of core storage entering the 2000/2001 winter from 70 Bcf in a March 2000 plan⁵ to an actual 55 Bcf (excluding CAT and Montebello gas) as of October 31, 2000,⁶ just as it asserts was contemplated in the Winter Price Spike Analysis.

SoCalGas responds that the Winter Price Spike Analysis was a very rough analysis put together in a few hours, with the increase in price spikes assumed to be due only to the proposed tighter balancing rules in the GIR proposal. It maintains that the analysis assumed that the lower core inventory level under GIR would reduce SoCalGas' ability to sell gas (i.e., 3 Bcf rather than 4 Bcf) during a price spike but would not affect the magnitude of the price spike. SoCalGas argues that Edison's conclusion that the table implies that SoCalGas believed that lower core storage would lead to higher price spikes is inconsistent with explicit language in the table.

Edison responds that SoCalGas would have understood in May 1999 the relationship between storage levels and winter daily balancing rules. Edison's logic is that because (1) balancing rules become more stringent as storage levels decrease and (2) demand for gas increases when noncore customers have to comply with the stricter balancing rules, SoCalGas knew that, by entering winter with lower core storage as required by the GIR proposal, the more stringent daily balancing rules would occur sooner and would lead to price spikes such as shown in the Winter Price Spike Analysis.

Interpreting SoCalGas' Winter Price Spike Analysis Edison's way, i.e., that the analysis assumes that SoCalGas' lower storage inventory would contribute to the hypothetical price spikes is not clearly supported by the explicit language in the table, but is not precluded. Edison's interpretation assumes that lower core storage would mean lower total storage and/or that limiting core sales would affect the magnitude of the price spikes. While we cannot know whether the analyst preparing the Winter Price Spike Analysis made these assumptions, SoCalGas' analysis of market conditions related to NGC behavior indicated that SoCalGas understood that the use of storage could affect market prices during constrained conditions. In our view, possibly equally important is that the Winter Price Spike Analysis concludes that the "(n)et impact is an increase of $1 million/year in winter sales opportunities," or $4 million in increased profits in the year of the price spike. Thus, it is clear that SoCalGas was looking at the possibility of increased winter price spikes due to the GIR proposal as a GCIM profit opportunity.

After Dynegy's contract with EPNG expired on December 31, 1999, EPME acquired 1.2 Bcfd of the Dynegy capacity in March 2000 for a 15-month term. Consistent with its tracking of NGC/Dynegy behavior, SoCalGas paid attention to EPME's behavior after it took over the NGC/Dynegy capacity. SoCalGas was undoubtedly aware of the Commission's complaint against EPNG and EPME filed on April 4, 2000 at FERC. Edison cites an April 27, 2000 SoCalGas e-mail that, in Edison's view, confirms that SoCalGas was aware that EPME was withholding capacity and engaging in sham postings of pipeline capacity (Ex. 92, Att. 19). The e-mail describes that EPME posted a capacity release at a bid rate higher than allowed. The e-mail characterized this posting as an "error." SoCalGas responds that "Edison has taken upon itself to interpret the word `error' to mean `sham.'" However, it is clear that the e-mail author understood that EPME did not desire to sell the capacity, since the e-mail explained that EPME did not plan to modify the posting and did not expect anyone to bid on the capacity.

A May 2000 presentation to ORA establishes that SoCalGas was aware at that time that basis differentials were increasing because of EPME's behavior and that the supply/demand balance at the border was becoming tight. In this presentation, SoCalGas stated as a single "key market factor" for GCIM Year 7 that EPME had obtained the Dynegy capacity on the El Paso pipeline and that border price differentials "have increased from last year and are expected to stay wide during the summer injection season" (Ex. 90, Att. 3-1). A June 2, 2000 e-mail from Gas Acquisition to senior Sempra management noted similarly that basis differentials had increased in May and commented that "The El Paso merchant group had a good month" (Ex. 90, Att. 3-2).

By June 2000, SoCalGas knew that tightening of the system was occurring, as indicated by increasing basis differentials. Internal SoCalGas documents indicate that in early June 2000, SoCalGas "anticipate[d] additional spikes this summer" and that the Gas Acquisition group was "getting ready for the next increase in volatility" and planning to "benefit from high volatility summer markets" (Ex. 90, Att. 3-2). Another June 2000 document indicated that Gas Acquisition was planning to "(m)aintain adequate storage through summer to take advantage of potential additional price spikes" (Ex. 90, Att. 3-4).

Edison points to language in SoCalGas weekly strategy sheets as additional proof that SoCalGas intended to use hub loans in lieu of physical core storage injections to achieve a core physical storage inventory of roughly 55 Bcf at the end of the 2000 injection season. In particular, the June 13, 2000 strategy sheet (Ex. 114) states the following as one of SoCalGas' injection season strategies:

7. Hub Outs to Offset Injections?

Assuming no GIR settlement, develop a contingency plan for using hub outs to offset injections to take advantage of any baseload/swing purchase opportunities. If there is a settlement, assume Z99 withdrawals offset deliveries/injections.

SoCalGas submits that the language in Item 7 describes a method Gas Acquisition can use to obtain additional supplies during Operational Flow Order (OFO) events. SoCalGas explains its view that having hub outs scheduled during OFO events allows SoCalGas to buy additional low-cost gas while staying within the retail core's injection rights. Edison responds that such an interpretation is not apparent in the language and, if that is indeed what the language meant, it relies on a questionable interpretation of SoCalGas' balancing rules and provides additional evidence regarding SoCalGas' manipulation of hub services as a means of flexibly meeting its storage obligations and creating GCIM profit opportunities.

Neither of the conflicting possible interpretations of Item 7 is obvious from its wording. We note that SoCalGas' periodic 2000 Summer Injection Plan sheets until May 30, 2000 (Ex. 90, Atts. 3-5 and 3-16) address injections during OFO events: "Plan Uses Retail Core Firm Injection rights of 317 MMcfd max. for purchases (as available used to handle improved delivery performance on OFO days)." On its face, this language does not appear to support SoCalGas' interpretation of Item 7. Regarding Item 7, we agree with SoCalGas' witness that "the wording is a little sloppy" (Tr. at 1533), but cannot draw conclusions beyond that at this time.

SoCalGas monitored and forecasted east-of-California gas flows, which were a factor leading to EPNG capacity cuts to California. SoCalGas correlated daily basis differentials with the corresponding California southwest and east-of-California flows. It compared the results to forecasted California southwest and east-of-California flows, thus gaining insight into what basis differentials might be expected during the winter of 2000/2001. In an early October 2000 comparison, the forecasted December flows were comparable to the highest combined flows experienced to date; by mid-November the forecasted December flows were higher than any recorded to date. These analyses (Ex. 79) confirm that SoCalGas understood that basis differentials were likely to be high in December. By mid-November, SoCalGas' comparisons indicated that the December basis differentials could be "off the charts."

SoCalGas knew that Enron was manipulating the southern California gas market. SoCalGas describes how EOL's dominance increased SoCalGas-Topock prices and how SoCalGas profited by arbitraging gas at Topock (Ex. 2 at VI-30). Separately, a SoCalGas internal memo dated August 4, 2000 produced during discovery described a border basis swap (short border/long Permian basis) undertaken in mid-July with the belief that the border premium would not increase beyond the $0.55 El Paso interruptible transportation rate. With El Paso constraints and no interruptible transportation, SoCalGas described its perception that "a border long" Enron ran the border basis up. SoCalGas' position resulted in a $347,000 loss. (Ex. 90, Att. A-5.)

SoCalGas incorporated the effects of supply shocks such as the August 19, 2000 Carlsbad rupture and the October 1, 2000 Exxon contract termination into its forecasts of winter supply conditions.

SoCalGas made an advice letter filing at the end of August 2000 seeking permission to implement winter balancing rules one month early, on October 1, 2000 rather than on November 1, 2000 as its tariff permitted, if SoCalGas' total system storage (excluding Montebello inventory) should fall below 47 Bcf during September. The filing stated that "unusually high consumption relative to deliveries ... has led to a net withdrawal of more than 8 Bcf from May 1^st to date..." It recognized uncertainties associated with El Paso's delivery capacity, national gas supply storage levels, and electric generator demand, and expressed concern that gas would continue to be underdelivered and storage volumes drawn down too low to protect core customers in the winter, so that SoCalGas would be required to purchase expensive gas supplies and core rates would "rise to an even greater level than that which is currently expected from recent high gas prices." SoCalGas withdrew the advice letter when it determined that storage levels were unlikely to fall to a level to warrant winter balancing rules during October. While this advice letter specifically asked for changes in its October operations, the bases for those changes would portend continuing supply problems throughout the winter. Thus, SoCalGas was aware that market conditions were developing that could lead to higher winter prices, contrary to the market backwardation that still existed.

Edison cites a September 2000 speech (Ex. 92, Att. 21) and an August 2000 e-mail to the president of SDG&E (Ex. 92, Att. 22) which it asserts, along with the advice letter filing, "suggest that SoCalGas understood the risks of the upcoming winter and the likelihood of price spikes." Edison also cites draft Gas Acquisition Committee meeting notes from September 22, 2000 (Ex. 75), which state regarding the 17 Bcf loan position that "Gas will be coming back in winter counter cyclically and will set up loans during price spikes."

SoCalGas tracked the electricity market closely and prepared gas demand forecasts that considered the effect of hydro conditions, nuclear outages, and other factors on gas demand. It understood during the summer and fall of 2000 that gas demand for electricity generation was high and would remain high through the winter. A September 28, 2000 Winter Hedging Strategy draft document (Ex. 92, Att. 9) provides insight regarding SoCalGas' view of the market at that time. The document indicated that:

This report reviews a potential winter hedging strategy for SoCal winter gas volumes which are subject to volatile increases in price. The principal working assumption...is that there exits (sic) a potential for significant winter price spikes. ...

This document does note (sic) provide an exhaustive review of the market conditions which prompted this exploration of hedging strategies. It does, however, highlight certain bullish market factors which combined with other bullish factors push prices higher this winter.

In fact, the single market factor described in the document is the gas price risk related to SoCalGas' expectation that gas-fired electric generation would remain strong during the 2000/2001 winter. It cited a PIRA forecast that WSCC demand for gas would increase by 1.4 Bcfd, a 71% increase over the average of the prior two winters, and that much of that increase would be in California. It stated that the futures market for electricity provided support for the view that gas demand by electric generators would be exceptionally strong in that winter, in particular, that "NYMEX Palo Verde futures suggest WSCC electricity prices will record new highs in the coming months" with the conclusion that, "Current high forward market power prices create a significant incentive for gas fired generators to operate at high levels of capacity this winter." (Emphases in original.)

SoCalGas' internal forecasts of expected winter conditions, combined with its knowledge about current market conditions, provided SoCalGas with knowledge when it was making its hub loans that there was a high likelihood of winter price spikes and congestion and that winter hub loan repayments would push prices even higher. SoCalGas developed its Southwest Flow Model, a spreadsheet computer model that forecasts gas demand and monthly flows into California (SoCalGas, PG&E, and Mojave) on the EPNG and Transwestern pipelines. The forecasts are based on prior year flows adjusted for forecasted differences in hydro conditions, nuclear plant availability, in-state gas production, and other factors. A simple version of the Southwest Flow Model was used in 1998 (Ex. 70, Att. 10) to estimate the portion of load that would require NGC capacity each month during that year. During the subject period, more elaborate versions of the Southwest Flow Model were used, which include additional adjustments for weather, hub activity, SoCalGas net border purchases or sales, noncore storage behavior, and other factors. The model assumes that hub loan repayments are made from gas flowing into California.

During the subject period, a separate module in the Southwest Flow Model was used at times to forecast the extent to which EPME capacity would be required to serve the forecasted demand. This calculation adjusted the basic flow model outputs for expected EPNG pipeline cuts.

In June and July 2000, border/basin basis differentials rose above transportation costs, indicating that the system was becoming constrained. Actual flows into California in those months were comparable to the flows projected by the Southwest Flow Model for December 2000. Because actual basis differentials were elevated in June and July, the Southwest Flow Model results provided an indication that system constraint conditions were also likely to occur in December at comparable levels of flow.

The following table presents, for June through November 2000, the average monthly flows on the EPNG and Transwestern pipelines, contemporaneous forecasts from the Southwest Flow Model of December demand for interstate gas, the amount of December net hub loan repayments assumed in those forecasts, the average amount of December repayments actually scheduled, and average basis differentials (Ex. 92, Atts. 12, 14, and 15).

Table 2

Average Southwest Flows, December Forecasts,

(MMcfd)

		SW Model Dec. Forecast			Basis Differentials ($/MMBtu)
Month	Monthly Flows	Flow	Net Hub In	Hub Loans w. Dec. Repay	S.J.	Permian
June	3,029	2,922	46	90	0.66	0.47
July	3,179	2,957	46	143	0.97	0.63
August	3,237	3,187	142	188	1.83	0.92
September	3,411	3,368	152	197	1.82	1.11
October	3,525	3,466	160	221	1.00	0.66
November	3,387	3,461	168	283	4.76	4.56
December	--	--	--	--	17.04	16.38

In reality, net hub loan repayments averaged about 255 MMcfd in December. To the extent that SoCalGas had already scheduled or was planning December repayments larger than those reflected in the forecasts, it would have been aware of their impacts on December flows even though they were not included in the Southwest Flow Model results indicated above.

SoCalGas prepared a separate forecast in November 2000 of December 2000 conditions, which predicted that California demand for southwest gas during December would be in the range of 3,500 to 3,700 MMcfd, higher than the indicated in the Southwest Flow Model. This estimate formed the basis for a statement in minutes of the November 15, 2000 Gas Acquisition Committee meeting (Ex. 90, Att. 3-27) that, "In December, California demand for southwest gas supplies is projected to be at record levels due to increased core purchases and large Hub paybacks." The minutes also describe November maintenance on both Transwestern and EPNG pipelines and state that southwest deliveries would be reduced as a result. The minutes state that the current forecast implies a potential shortage of delivery for the expected demand.

Throughout the summer and fall of 2000, SoCalGas could compare actual southwest flows occurring each month and contemporaneous forecasts of December demand. Taking into account current conditions and planned winter hub loan repayments, the flow forecasts indicated a high likelihood of continued and increasing winter congestion on the southwest pipelines, based on actual operational conditions on the pipeline system. A reasonable inference is that SoCalGas expected, based on its own forecasts, that average basis differentials would be higher than the forward markets were indicating. SoCalGas would also have known that there was a high likelihood of daily price spikes that would provide particularly lucrative sales opportunities, such as modeled in the May 1999 Winter Price Spike Analysis. SoCalGas would also have known that the addition of incremental flowing supplies, e.g., in the form of hub loan repayments, would have an atypically large price effect at the California border.